39
Review of Finance (2015) 19: pp. 1415–1453 doi:10.1093/rof/rfu032 Advance Access publication: August 25, 2014 The Profits–Leverage Puzzle Revisited* MURRAY Z. FRANK 1 and VIDHAN K. GOYAL 2 1 University of Minnesota and 2 Hong Kong University of Science and Technology Abstract. The inverse relation between leverage and profitability is widely regarded as a serious defect of the trade-off theory. We show that the defect is not with the theory but with the use of a leverage ratio in which profitability affects both the numerator and the denominator. Profitability directly increases the value of equity. Firms do take the predicted offsetting actions. They issue debt and repurchase equity when profitability rises, and retire debt and issue equity when profitability falls. Consistent with variable transactions costs, the adjustment is not generally sufficient to fully undo the profitability shocks. Accordingly, on average the leverage ratio falls as profitability rises. JEL Classification: G32 1. Introduction The trade-off theory of capital structure predicts that more profitable firms ought to borrow more and have higher leverage. 1 The profits–leverage puzzle is the empirical evidence that the predicted sign is backwards. “The most telling evidence against the static trade-off theory is the strong inverse * We are especially grateful for helpful comments from our referee. We also thank Franklin Allen, Raj Aggarwal, Ralf Elsas, Mark Flannery, David Florysiak, Fangjian Fu, Catharina Klepsch, Paul Povel, Jay Ritter, Philip Strahan, Ilya Strebulaev, Michael Roberts, and workshop participants at the American Finance Association, Universita` Bocconi, Boston College, City University of Hong Kong, University of Florida, Imperial College, Korea University, University of Minnesota, University of Munich, University of Pittsburgh, Oxford University, and Singapore Management University for helpful comments. Murray Z. Frank thanks Piper Jaffray for financial support. Vidhan K. Goyal thanks the Research Grants Council of Hong Kong for financial support (Project #641608). We alone are responsible for any errors. 1 The term “trade-off theory” is used in different ways by different authors. For some authors it means that bankruptcy and taxes are being balanced as in Fischer, Heinkel, and Zechner (1989) and Leland (1994). For other authors, such as Fama and French (2002), it includes agency-based arguments. Some authors such as Strebulaev (2007) simplify by assuming that investment is unaffected, even though the cost of capital is changed by the leverage choice. Other authors such as Hackbarth and Mauer (2012) analyze the impact of the leverage choice on investment. A review is provided by Frank and Goyal (2008). V C The Author(s) 2014. Published by Oxford University Press on behalf of the European Finance Association. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons. org/licenses/by/4.0/), which permits unrestricted reuse, distribution, and reproduction in any medium, provided the original work is properly cited. Downloaded from https://academic.oup.com/rof/article/19/4/1415/1568525 by guest on 02 August 2022

The Profits–Leverage Puzzle Revisited* - Oxford Academic

Embed Size (px)

Citation preview

Review of Finance (2015) 19 pp 1415ndash1453doi101093rofrfu032Advance Access publication August 25 2014

The ProfitsndashLeverage Puzzle Revisited

MURRAY Z FRANK1 and VIDHAN K GOYAL2

1University of Minnesota and 2Hong Kong University of Science andTechnology

Abstract The inverse relation between leverage and profitability is widely regarded as aserious defect of the trade-off theory We show that the defect is not with the theory but

with the use of a leverage ratio in which profitability affects both the numerator and thedenominator Profitability directly increases the value of equity Firms do take the predictedoffsetting actions They issue debt and repurchase equity when profitability rises and retiredebt and issue equity when profitability falls Consistent with variable transactions costs

the adjustment is not generally sufficient to fully undo the profitability shocks Accordinglyon average the leverage ratio falls as profitability rises

JEL Classification G32

1 Introduction

The trade-off theory of capital structure predicts that more profitable firmsought to borrow more and have higher leverage1 The profitsndashleveragepuzzle is the empirical evidence that the predicted sign is backwards ldquoThemost telling evidence against the static trade-off theory is the strong inverse

We are especially grateful for helpful comments from our referee We also thankFranklin Allen Raj Aggarwal Ralf Elsas Mark Flannery David Florysiak Fangjian

Fu Catharina Klepsch Paul Povel Jay Ritter Philip Strahan Ilya Strebulaev MichaelRoberts and workshop participants at the American Finance Association UniversitaBocconi Boston College City University of Hong Kong University of Florida Imperial

College Korea University University of Minnesota University of Munich University ofPittsburgh Oxford University and Singapore Management University for helpfulcomments Murray Z Frank thanks Piper Jaffray for financial support Vidhan K

Goyal thanks the Research Grants Council of Hong Kong for financial support (Project641608) We alone are responsible for any errors1 The term ldquotrade-off theoryrdquo is used in different ways by different authors For someauthors it means that bankruptcy and taxes are being balanced as in Fischer Heinkel and

Zechner (1989) and Leland (1994) For other authors such as Fama and French (2002) itincludes agency-based arguments Some authors such as Strebulaev (2007) simplify byassuming that investment is unaffected even though the cost of capital is changed by the

leverage choice Other authors such as Hackbarth and Mauer (2012) analyze the impact ofthe leverage choice on investment A review is provided by Frank and Goyal (2008)

VC The Author(s) 2014 Published by Oxford University Press on behalf of the European Finance Association

This is an Open Access article distributed under the terms of the Creative Commons Attribution License (httpcreativecommons

orglicensesby40) which permits unrestricted reuse distribution and reproduction in any medium provided the original work is

properly cited

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

correlation between profitability and financial leverage Within an industrythe most profitable firms borrow less the least profitable borrow morerdquo(Myers 1993 p 6) This puzzling relationship between corporate profitsand leverage is widely regarded as a particularly serious defect of thetrade-off theory (see eg Fama and French (2002))2

In this article we revisit the profitsndashleverage relationship and show thatthe usual interpretation of the evidence is wrong What really happens is thatfirms only actively adjust leverage occasionally and the leverage adjustmentsthat follow profitability changes are typically not big enough to fully offsetthe effect of the profitability shocks This result does not arise in a one-period trade-off model nor in the usual dynamic trade-off models thathave fixed adjustment costs but no variable adjustment costs Howeverthis result is easy to understand in a trade-off model with both fixed andvariable costs of adjustment Since both kinds of costs are realistic theprofitsndashleverage puzzle does not provide telling evidence against the trade-off theoryIn a standard static trade-off theory model free of any transactions costs

the firm chooses to be at the optimal leverage target By construction thereare no dynamic effects in a one-period model On the other hand in a con-tinuous time trade-off models without fixed costs as in Kane Marcus andMcDonald (1984) and Brennan and Schwartz (1984) firms actively releveralmost continuously to maintain very high leverage to avoid taxes while juststaying out of bankruptcy Since the almost continuous issuance andrepurchasing of debt is not observed in practice the literature switched toconsideration of fixed adjustment costs For simplicity the variable adjust-ment costs were generally dropped to focus on the fact that fixed costsinduced periods without rebalancing Fischer Heinkel and Zechner (1989)found that due to the fixed costs much of the time the firm passively acceptsleverage shocks However if leverage becomes too high (or too low) the firmrebalances leverage to a target level The idea that firms are passive much ofthe time and only occasionally rebalance to a single leverage target is also inGoldstein Ju and Leland (2001) Strebulaev (2007) and Morellec Nikolovand Schurhoff (2012)Leary and Roberts (2005) explicitly recognize that both fixed and variable

transactions costs matter for leverage adjustments The implications of

2 A partial list of papers documenting an inverse relation between leverage and profitability

include Auerbach (1985) Graham and Tucker (2006) Long and Malitz (1985) Titman andWessels (1988) Fischer Heinkel and Zechner (1989) Rajan and Zingales (1995) andBooth et al (2001) Frank and Goyal (2009) show that the inverse leveragendashprofitabilityrelation has become weaker in the recent decades

1416 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

having both types of costs for the leverage puzzle can be readily understoodDue to the fixed costs small shocks are ignored since it will not pay to undotheir effect As in the pure fixed cost models there is a zone of inactionIf leverage becomes too high (or too low) it will pay for the firm to takeaction How strongly will the firm react If there are no variable costs thenaction from either boundary will be to the same target value But withvariable costs the firm must consider how much rebalancing is worthwhileThe result will be partial moves away from the boundary to the point atwhich the cost and the benefit of further moves are just balanced As a resultthere are two rebalancing points or ldquotargetsrdquomdashan upper target and a lowertarget3

From the perspective of a static model or a single target model the firmunder-adjusts to the profitability shock in the sense that the impact of theshock is only partly undone But this is not suboptimal behavior It is arecognition by the firm that fully undoing the shock would be more costlythan it is worth Variable costs count To go all the way to the staticoptimum would be excessively costly and so firms only go part wayEmpirically we find that the predicted ldquounder-adjustmentrdquo is observedThe nonlinearity of debt and equity contracts may also be important formagnitudes Because debt is concave it will tend to be less responsive toprofitability shocks In contrast equity is convex which makes it more re-sponsive to profitability shocksUnder-adjustment due to variable costs of adjusting is important when

interpreting empirical evidence of cross-sectional variation in leverage ratiosLet D denote corporate debt and E corporate equity then leverage isL frac14 D

DthornE Consider a positive shock to profits which causes E to increaseHowever because of under-adjustment D increases by less In other wordsthe denominator increases by more than the numerator in the leverage ratioand so the ratio falls Next consider a negative shock to profit In this case Ewill drop and due to under-adjustment D drops by less Leverage increasesThus a negative relationship will be observed between profits and leveragedespite the fact that firms are actively rebalancing in the direction predictedby the trade-off theory

3 In models with both fixed and variable costs of adjusting there are two kinds of first-order conditions known as ldquovalue matchingrdquo and ldquosmooth pastingrdquo The key analytical

ideas have been understood since Constantinides and Richard (1978) and are well explainedby Dixit (1993) This result from Constantinides and Richard (1978) is depicted by Learyand Roberts (2005) in Figure 1 panel C where the conditions are explicitly interpreted asleverage

PROFITSndashLEVERAGE PUZZLEREVISITED 1417

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

The main empirical findings are as given below

(1) When a firm makes extra profits the book value of equity increasesunless the firm takes some sort of offsetting action Similarly when a

firm makes extra profits the market value of the firmrsquos equity in-creases unless there is some sort of offsetting action Thus moreprofitable firms automatically have more book equity and more

market equity unless the firm takes offsetting actions

(2) Among large firms the highest profit firms increase their debt themost Those with high profits experience large increases in both thebook and the market value of equity The highest profit firms tend torepurchase equity while the lowest profit firms tend to issue moreequity

(3) Among small firms profit seems to have only a very minor effect ondebt Those with high profits experience some increases in both thebook and the market value of equity Those with low profits experi-ence negative effects on market equity There is a tendency to issueequity with the lowest profit firms issuing the most equity

(4) When firms adjust leverage the magnitude of the adjustment isnot sufficient to fully undo the impact of the underlying shocksFirms do not return to a unique static optimum They seem toldquounderadjustrdquo

There is a huge prior literature on our topic For a review of the literaturesee Frank and Goyal (2008) The inverse relationship between profitabilityand leverage ratios has generated a variety of responses from scholars A

common response articulated by Strebulaev and Whited (2012) is to arguethat the trade-off theory predictions could be more complex in a dynamicmodel Hennessy and Whited (2005) for example specify a dynamic model

of investment and financing under uncertainty and show that a trade-offmodel can explain a number of empirically observed stylized facts includingthe negative leveragendashprofitability relation In their model highly profitablefirms find it unattractive to issue debt because additional debt in such firms

finances a distribution to shareholders rather than replace equityOther scholars focus on adjustment costs and how a proper consideration

of these costs could reconcile the trade-off theory with empirical evidenceMyers (1984) discusses how adjustment costs induce lags in adjusting to the

optimum which could explain the observed wide variation in leverage ratiosFischer Heinkel and Zechner (1989) provide an explicit model of the ideaFama and French (2002) speculate that the negative relation betweenleverage and profitability is perhaps picking-up transitory variation in

1418 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

leverage rather than target variation in leverage A more formal analysis oftransitory debt is provided by DeAngelo DeAngelo and Whited (2011)Leary and Roberts (2005) consider both fixed and variable adjustment

costs and show that firms really do rebalance leverage Firms do notignore leverage ratios when making financing decisions and they do notjust passively accept the impact of equity market shocks Leary andRoberts (2005) do not study the implications of the form of the adjustmentcosts for the profitsndashleverage puzzle which is our focus Their focus is on thespeed of adjustment toward the target and on what that implies about thepersistence of leverageStrebulaev (2007) calibrates a dynamic trade-off model in which all costs

are proportional to the firm value A large firm is a rescaled version of asmall firm and so the target leverage ratio is independent of firm size Thisstructure is not designed to consider the differences between the financingactions of large versus small firms that we observe in the data Because he isstudying a trade-off model he finds that more profitable firms issue moredebt In simulated data Strebulaev (2007) shows that increases in profitabil-ity are correlated with future profitability which in turn increases the valueof the firm and lowers leverage (see p 1772) But expected future profit-ability increases do not always lead to current refinancing His cost structuredoes not lead to the under-adjustment phenomenon of the sort that wedocumentSeveral papers attribute the negative relation between leverage and prof-

itability to firms passively accumulating profits (see eg Fama and French(2002) and Kayhan and Titman (2007)) However the existing evidence isoften indirect and provides mix findings for this idea The implication that atthe time of rebalancing leverage should be positively related to profitabilityis also examined in several earlier studies Mackie-Mason (1990) shows thatcompanies with tax loss carryforwards are more likely to issue equity Incontrast Jung Kim and Stulz (1996) report finding no relation between thelikelihood of equity issuance and profitability Hovakimian Opler andTitman (2001) show that security issuance decisions of firms are afunction of the deviation from target leverage and other regressors includingfirm profitability They show that debt issuances are positively related toprofitability The present article while supporting aspects of their findingsexamines both issuing decisions and magnitudes as a function of firm sizeand profits In doing so we provide a more granular understanding of theeffect of profitability on changes in equity that derive from real operationsand changes in equity that are the result of active issuances and repurchasesof securities Furthermore we provide a sharply different interpretationalong with evidence in support of our interpretation We stress the

PROFITSndashLEVERAGE PUZZLEREVISITED 1419

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

distinction between automatic changes in equity that stem from operationsas opposed to actual issues or repurchases of equity This distinction is em-pirically importantThis article is organized as follows Section 2 provides theoretical motiv-

ation for our empirical strategy Section 3 describes the construction of thedata and variables and provides summary statistics Section 4 providesresults from the fixed effects estimates of leverage ratios Section 5presents the frequency and magnitudes of financing activity for sorts onfirm size and profits Section 6 provides the main results on debt andequity issuance regressions Section 7 explores the debt and equityissuance responses of firms during good and bad times Section 8examines the debt and equity issuances when scaled by total issuances andtotal capital Section 9 concludes the article

2 Empirical Strategy

The key idea in the trade-off theory is that firms choose capital structure bytrading off the various costs and benefits The considerations may includebankruptcy costs tax benefits agency costs transactions costs etc Differentversions of the theory include slightly different elements From our perspec-tive the leverage adjustment costs are of particular importanceA decision to issue debt or to redeem it or to issue shares or to repurchase

them requires managerial thought and consideration There are fixed organ-izational costs There is also the possibility of managerial inertia as docu-mented in Bertrand and Mullainathan (2003) These are fixed costs that areapproximately independent of the magnitude of the adjustments Adjustingalso has marginal costs Depending on which action is taken there may befees for underwriters bankers lawyers etc These may have a fixed compo-nent too Commonly they have a variable cost component which dependson the size of the transaction So both fixed and variable costs of adjustingleverage are normally involved The exact magnitudes vary from situation tosituation What is worse these are likely to have time-varying components4

Why does the form of the transactions costs matter Because the existenceof such fees implies that there will not be a unique leverage target This is anapplication of an idea that goes back to Constantinides and Richard (1978)and Constantinides (1979) It should be stressed that there is nothing newhere in terms of the formal structure This is just a reinterpretation of

4 If there were no time-varying component then it would be easy to identify the exact upperand lower targets for each firm In each case the firm would go exactly to the appropriatetarget In reality with time variation in adjustment costs such precision is not possible

1420 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

standard dynamic optimization conditions For a reader who wants a moreexplicit treatment of the mathematics we recommend Dixit (1993) The ob-servation that this mathematics can be interpreted in terms of corporateleverage in which firm size matters is also not novel It can be found inLeary and Roberts (2005) who provide a simulation with both fixed costsand a weakly convex adjustment costs What is new here is the analysis of theimplication for the impact of profits on leverageSuppose that without active rebalancing leverage would just wander

about due to changes in the stock market value of equity According tothe trade-off theory excessively high leverage is costly (bankruptcy andagency costs) as is excessively low leverage (lost tax benefits and free cashflow problems) If there are no transactions costs then under conventionaltrade-off theory assumptions there will be a unique leverage target that willbe maintained at all timesDue to the fixed costs such frequent rebalancing would be too costly As a

result the firm permits the leverage to drift However if the drift is tooextreme in either direction then the firm will actively rebalance in the direc-tion of the static optimum If the rebalanced firm (inclusive of rebalancingcosts) is worth more then rebalancing is worth it If the rebalanced firmwould be worth less then rebalancing is not worth it Rebalancing takesplace at the point at which the values just matchWhen the firm is rebalancing it also needs to worry about how far to

move in the desired direction When each step has costs the firm needs toensure that the marginal cost of the last step is just equal to the marginalbenefit As a result both from excessively high leverage and from excessivelylow leverage the movements are only part wayTo fix ideas consider an off-the-shelf model of the leverage ratio The

focus is on the impact of the adjustment cost structure The formal structureis essentially the cash management problem studied by Constantinides andRichard (1978) and Dixit (1993) We follow the presentation and notation inDixit (1993) very closely to show that this is just an application of well-established ideas to the leverage problemLet xt denote the leverage ratio at date t is the interest rate Ku is a fixed

cost of actively increasing leverage Kd is the fixed cost of actively reducingleverage mu is the marginal cost of each unit of active leverage increase md isthe marginal cost of each unit of active leverage decrease f(x) is the flowpayoff if leverage is at x F(x) is the expected present value of the payoff ifleverage is at x x is the lower adjustment threshold x is the lower targetleverage (when adjusting from below) x is the upper adjustment thresholdand x is the upper target leverage (when adjusting from above) Accordingto the trade-off theory F 0 gt 0 F 00 lt 0

PROFITSndashLEVERAGE PUZZLEREVISITED 1421

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

In the absence of active adjustments the leverage ratio xt follows astandard process

dx frac14 ethx tTHORNdtthorn ethx tTHORNdw

where is the trend coefficient is the volatility coefficient and w is anormal shock process with zero mean and unit variance If x lt x lt x thenfinancial inaction is optimal If x drops to x then it is reset to x and if xrises to x then it is reset to xThe trade-off theory is frequently thought of as a class of theories in which

firm value first increases in leverage and ultimately drops if leverage gets toohigh This can be motivated by taxes and bankruptcy or by various agencyconcerns A range of ideas can be allowed for when we write that the flowpayoff (dividends andor profits) is fethxtTHORN and it depends on the leverage statext

5 If the firm starts at x0frac14 x then F(x) is the expected present value of theflow payoffs The discount rate is which allows us to write

FethxTHORN frac14 E

Z 10

fethxtTHORNetdtjx0 frac14 x

To maximize the firm value it must choose a leverage policy that sayswhen to rebalance and how much to rebalance These are given by thefamous value matching and smooth pasting conditionsThe value matching conditions are as follows

FethxTHORN FethxTHORN frac14 Ku thornmuethx xTHORN

FethxTHORN FethxTHORN frac14 Kd thornmdethx xTHORN

The smooth pasting conditions are as follows

F 0ethxTHORN frac14 F 0ethxTHORN frac14 mu

F 0ethxTHORN frac14 F 0ethxTHORN frac14 md

As explained by Constantinides and Richard (1978) with standard as-sumptions about the shape of F there are a pair of lines at heights mu andmd such that they intersect the function F 0ethxTHORN in such a way as to define thepoints x lt x lt x lt xTo sum up in this trade-off model with both fixed and variable

rebalancing costs there are four critical leverage points x gt x gt x gt x

5 For simplicity we suppose that the flow payoff is a continuous function of leverageThere is no discrete exit or bankruptcy point This allows us to simply interpret theConstantinides and Richard model much as in Leary and Roberts (2005)

1422 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

If leverage is between x (upper boundary) and x (lower boundary) the firmdoes nothing active Changes in leverage reflect various shocks to revenuesand costs that hit the firm If leverage drifts up to x (or above) then the firmrebalances to x If leverage drifts down to x (or below) the firm rebalancesto xWhat does this imply empirically A firm that does nothing can be

anywhere within the interval between x and x A firm that is rebalancingis a firm that has just hit one of the outer boundaries How does profitabilityaffect the likelihood of hitting a boundary High profits increase the value ofequity This increases the denominator of the leverage ratio and thus makesit more likely that the firm hits the lower boundary Low profits reduce thevalue of equity and thus make it more likely that the firm hits the upperboundaryTo be more specific suppose that the costs are symmetric and the leverage

changes are also following an ordinary Brownian motion Then the median

inert firm will have leverage ofxthornx

2 A firm that just actively reduced leverage

will be at x gtxthornx

2 A firm that just actively increased leverage will be at

x ltxthornx

2 A firm that just reduced leverage will generally be a firm that had

very high leverage To the extent that this reflects a lack of activerefinancing such firm will have low value of equity The low value ofequity will generally reflect low profits A firm that just increased leveragewill generally be a firm that had very low leverage In the same way a firmthat just increased its leverage will tend to have had high value of equity dueto high operating profits6

In this setting it is natural to focus on firms that are actively adjusting andto compare them to each other and to firms that are inactive It is alsohelpful to look directly at financing decisions and profitability rather thanfocusing on a leverage ratioOur approach shares the perspective of Welch (2004) Leary and Roberts

(2005) and Strebulaev (2007) that it is very important to focus on activeadjustments by firms Welch (2004) transforms the data to put everythingback into a ratio form His ldquoimplied debt ratiosrdquo still have both debt andequity in the denominator Similarly Leary and Roberts (2005) andStrebulaev (2007) pay a great deal of attention to leverage ratios whenstudying financing decisions Strebulaev (2007) does not have fixed costssince he makes use of a rescaling property Empirically we find differentadjusting behavior by large and by small firms So the rescaling property is

6 If we knew the values of the critical leverage points then it would be possible to derivemore refined tests Unfortunately those critical values are not observable

PROFITSndashLEVERAGE PUZZLEREVISITED 1423

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

an issue of concern His model does not account for the negative relationshipbetween current leverage and past market-to-book ratioWe depart from the earlier literature by arguing that the use of leverage

ratios can be particularly misleading for profitability since the use of aleverage ratio causes identification problems It is hard to be sure whethera leverage change reflects changes in the numerator or the denominator Thisis particularly problematic when as here the rebalancing is partialIt should be noted that much empirical capital structure literature debates

about whether to use a book leverage ratio or a market leverage ratio Bookleverage is used based on the argument that managers borrow against assetsin place which are better reflected by book assets Others such as Welch(2004) argue that book values are essentially an accounting fiction and thatonly market leverage should count Empirical papers regularly attempt tosidestep this debate by carrying out tests on both book leverage and marketleverage Our point is that both kinds of leverage ratios are potentially mis-leading when examining the impact of profits on capital structure

3 Data

We use conventional data sources starting with the merged Compustat-CRSP data The data are annual and are converted into constant year2000 dollars using the GDP deflator We exclude financial companies (SIC6000-6999) firms involved in major mergers (Compustat footnote code AB)firms reporting financial data in currencies other than the US dollars andfirms with missing data on our key variables7 The ratio variables aretrimmed at the 1 level in both tails of the distribution This serves toremove outliers and the most extremely misrecorded data The finalsample consists of 179021 firm-year observations from 1971 to 2009Table I provides definitions of financial variables and reports summary

statistics The average debt (in constant US$) is about $653 million while themedian is $24 million A significant fraction of firms have zero debt (the 10thpercentile is 0) Book equity is slightly larger than book debt Market equityis almost three times larger than book debt Book assets average $2191million although the medians are considerably smallerIf issuing or retiring securities incurs no fixed costs then we would expect

to see many small actions and very few large actions (Leary and Roberts2005) If there were significant fixed costs involved in issuing or retiringoutstanding securities then small issues might not be worthwhile Table I

7 These include debt book value of equity market value of equity assets book andmarket leverage profitability market-to-book assets ratio and tangibility

1424 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

shows that although most firms issue little debt or equity in a given year theaverages are large In other words when firms actually enter debt and equitymarkets they intervene massivelyThe mean constant dollar debt issue is $165 million (the median is $2

million) In unreported tables we find that the average debt issue is about81 of assets (the median is 18) About 38 of the firms issue no debt8 issue between 0 and 1 of the value of their assets as debt another

Table I Data description

Data sources The sample comes from the annual Compustat files The sample period is

1971ndash2009 We exclude financial companies (SIC 6000-6999) firms involved in majormergers (Compustat footnote code AB) firms reporting financial data in currencies otherthan the US dollar and firms with missing data on the key variables All financial variables

are deflated to year 2000 using the GDP deflator N is the number of observations SD isthe standard deviation D BVE MVE Assets Debt issuance Equity issuance Debt re-payment Equity repayment are all measured in $ millions The ratio variables are trimmed

at the 1 level in both tails of the distribution

Variable definitions Debt (D)frac14 long-term debt (dltt)thorn short-term debt (dlc) book equity(BVE)frac14 common shareholder equity (ceq) market equity (MVE)frac14 no of outstanding

shares (csho) closing price ( prcc f ) assetsfrac14 book assets (at) debt issuancefrac14 issuance oflong-term debt (Maxethdltis 0THORN)thorn increase in current debt (Maxethdlcch 0THORN) equity issu-ancefrac14 sale of common stock (Maxethsstk 0THORN) debt repaymentfrac14 reduction of long-termdebt (Maxethdltr 0THORN)thorn decrease in current debt (Minethdlcch 0THORN) equity repurchasefrac14 purchase

of common stock (Maxethprstkc 0THORN) book leveragefrac14D(DthornBVE) market leveragefrac14D(DthornMVE) profitabilityfrac14EBITDA (oibdp)assets market-to-book ratio M

B frac14Marketvalue of assets (MVA)assets where MVAfrac14DthornMVEthorn preferred-liq value (pstkl)

deferred taxes (txditc) tangibilityfrac14 net property plant and equipment (ppent)assets

Variable N Mean SD

Distribution

10th 50th 90th

Debt (D) 179021 653 4592 0 24 1048

Book equity (BVE) 179021 815 4122 3 70 1373

Market equity (MVE) 179021 1784 9589 9 118 2698

Assets 179021 2191 11530 10 153 3575

Debt issuance 179021 165 1335 0 2 245

Equity issuance 179021 26 186 0 0 40

Debt repayment 179021 132 1123 0 3 181

Equity repurchase 179021 24 257 0 0 9

Book leverage 179021 036 034 000 031 072

Market leverage 179021 027 025 000 020 066

Profitability 179021 005 027 018 011 023MB 179021 165 206 052 102 324

Tangibility 179021 031 024 005 026 070

PROFITSndashLEVERAGE PUZZLEREVISITED 1425

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

16 issue between 1 and 5 of the value of their assets as debt and theremaining 38 issue debt in excess of 5 of the value of their assetsThe mean constant dollar equity issue is about $26 million (the median is

about $04 million) As a fraction of assets the mean and median equityissues are about 7 and 02 respectively About 33 of the firms issue noequity 34 of the firms issue between 0 and 1 of the value of theirassets as equity another 14 issue between 1 and 5 of the value of theirassets as equity and the remaining 19 of the firms issue equity that is inexcess of 5 of the value of their assets Average debt repayments are largerthan equity repurchases This perhaps reflects the finite maturity of debt andits contractual repayment The median firm does not repurchase equityWe construct both book and market leverage ratios Book leverage is

defined as debt over debt plus book equity Market leverage is defined asdebt over debt plus market equity8 The median book leverage is 036 (theaverage is 031) The median market leverage is 027 (the average is 020)Profitability is defined as the ratio of operating income before depreci-

ation to assets While the average firm is profitable (the ratio of EBITDA toassets is 005) the median firm is even more profitable (with a profitabilityratio of 011) But the sample also includes a large number of unprofitablefirms as the 10th percentile is 018 The table also reports descriptive stat-istics on the market-to-book ratio and the tangibility ratio The market-to-book ratio (MB) defined as the ratio of the market value of assets to bookassets averages at about 165 Tangibility defined as the ratio of netproperty plant and equipment to assets averages at about 31

4 Leverage Ratio Regressions

The previous literature focuses on estimates obtained using leverage ratiosHence we start with a similar estimation to check whether our results matchthose of previous studies Table II presents the results for book leverage Weobtain similar results for market leverage and they are presented inAppendix Table AI The regressions include leverage factors following theprevious capital structure literature (see eg Lemmon Roberts and Zender(2008) and Frank and Goyal (2009)) The factors include (i) profitability (ii)industry median leverage (iii) market-to-book assets ratio (iv) tangibility ofassets and (v) firm size Industry median leverage is constructed as themedian leverage of all other firms in the industry excluding the firm under

8 Welch (2011) stresses the idea that nonfinancial liabilities should not be implicitly mis-treated as if they were equity by paying excessively narrow attention to financial liabilitiesin a leverage ratio We have adopted his approach in the empirical work reported here

1426 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

consideration Firm size is defined as the natural log of assets Other factorsare defined in Section 3 We cluster standard errors by firm and estimatethese regressions both with and without fixed effectsIn Table II we run conventional cross-section leverage regressions Since

we employ the usual data it is not surprising that our results match thosereported in the existing capital structure literature Profitability has anegative sign in both the book leverage regressions and in the marketleverage regressions The coefficients on other factors largely match thosereported in earlier studies Firms operating in industries with high leverage

Table II Leverage and profitability

The table presents parameter estimates for book leverage regressions The sample consists

of nonfinancial companies in the annual Compustat files during the period 1971ndash2009 Theleverage is estimated as the ratio of debt over debt plus book equity The explanatoryvariables Profitabilityt1

MB

t1

Tangibilityt1 and LnethAssetsTHORNt1 are described in

Table I IndLevt1 is estimated as the median book leverage of all other firms in thesame industry The industry is defined at the level of the four-digit SIC code All specifi-cations include the year fixed effects The specifications in column (2) additionally include

the firm fixed effects We report t-statistics where the standard errors are clustered at thefirm level in parentheses

aSignificant at the 1 level

Quantile regressions

(1) (2) (3) (4) (5)

Profitabilityt1 026a 020a 011a 017a 041a

(293) (192) (378) (518) (1015)

IndLevt1 042a 012a 030a 052a 052a

(369) (107) (926) (1252) (950)MB

t1

002a 001a 001a 002a 002a

(203) (102) (392) (426) (418)

Tangibilityt1 018a 025a 024a 019a 012a

(206) (154) (964) (580) (276)

LnethAssetsTHORNt1 001a 002a 002a 002a 001a

(114) (76) (899) (478) (96)

Constant 012a 010a 012a 002a 029a

(156) (60) (284) (27) (387)

Year FE Yes Yes Yes Yes Yes

Firm FE No Yes No No No

Regression type OLS FE 25th tile Median 75thtile

Clustered SE Yes Yes No No No

R2 013 054

Pseudo R2 016 015 009

N 158824 158824 158824 158824 158824

PROFITSndashLEVERAGE PUZZLEREVISITED 1427

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

tend to have high leverage A higher market-to-book ratio is associated withlower leverage Larger firms are typically more highly levered The coeffi-cients on tangibility are positive The conclusion from Table II is clear Theleverage regression results match those reported in previous studiesAt times one may be concerned that in a regression what is being

estimated is a conditional mean while some parts of the distributionmight behave differently from other parts If so then ldquoaccidentsrdquo likesample selection issues or asymmetries in the underlying distributionsmight play a greater role than is usually recognized In other words thesource of identification could be misinterpreted To guard against this po-tential problem we estimate the same model using quantile regressions9

Columns 3 4 and 5 report what happens at the 25th 50th and 75th per-centiles The basic model results are robust in terms of both the sign and thestatistical significanceThe impact of profits does appear to be somewhat stronger among the

high leverage firms For instance in a book leverage regression the coeffi-cient on profits for the first quartile is 011 with a t-ratio of 378 For thethird quartile the coefficient on profits is 041 with a t-ratio of 1015This difference is curious and might deserve further study However the factthat in each case we get the negative sign and statistical significance is suf-ficient for the purposes of the present article So the results on profits arerather robust across the distributionTable III sorts firms into those that are actively adjusting their leverage

and those that are passive In the presence of both fixed and variable trans-acts costs this distinction is important We expect the coefficient on profit-ability in leverage ratio regressions to be more negative in firms that arepassive and less negative in firms that are actively adjustingWe define active firms as those that are either issuing net debt or net

equity in excess of 5 of their assets Passive firms are those that neitherissue debt nor equity in nontrivial amounts (ie greater than 5 of theirassets) Consistent with our expectations we find in Table III that the coef-ficients on profits are significantly more negative for passive firms than theyare for active firms With firm and year fixed effects the coefficient onprofits is 008 for active firms and 017 for the passive firms The twocoefficients are significantly different from each other (plt 001) InTable III the standard control variables are used They have the usualsigns and for the most part do not differ much between the active and thepassive firms

9 Cameron and Trivedi (2010) provide an extensive discussion of quantile regressions inStata

1428 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

5 Financing Activity

51 FREQUENCY OF DEBT AND EQUITY ISSUES

Table IV reports the percentage of firms issuing or repurchasing debt orequity for annual sorts based on lagged profitability In Panel A weemploy the conventional 5 cutoff rule to exclude minor fluctuations anddefine firms ldquoissuing debtrdquo as those that issue debt (both short-term andlong-term) in excess of 5 of the value of their assets Other decisions aresimilarly defined using a 5 cutoff Appendix Table AII examines the dis-tributional properties across profitability (and size) quintiles for ldquolargerdquo is-suances by increasing the 5 cutoff to higher levels such as 10 and 15How does security issuance behavior vary with firm profitability To

answer this question Table IV sorts firms by profitability and reports the

Table III Leverage and profitability for active and passive firms

The table presents estimates of leverage on firm characteristics for the subsamples of firms

that are active in either debt or equity markets Active firms are defined as those that issueeither net debt in excess of 5 of the value of assets or issue net equity in excess of 5 ofthe value of their assets The passive firms do not issue significant (more than 5) of either

debt or equity The sample comes from the annual Compustat files during the period 1971ndash2009 Financial firms are excluded The variables are described in Tables I and II Wereport t-statistics where the standard errors are clustered at the firm level in parentheses

aSignificant at the 1 level

Active Passive

(1) (2) (3) (4)

Profitabilityt1 006a 008a 022a 017a

(126) (148) (254) (188)

IndLevt1 044a 019a 041a 013a

(424) (149) (326) (127)MB

t1

002a 001a 003a 001a

(362) (196) (327) (172)

Tangibilityt1 010a 014a 014a 019a

(146) (118) (145) (130)

LnethAssetsTHORNt1 001a 004a 001a 004a

(121) (179) (147) (169)

Constant 012a 008a 003a 007a

(120) (57) (45) (42)

Year FE Yes Yes Yes Yes

Firm FE No Yes No Yes

R2 034 064 033 075

N 43395 43395 58943 58943

PROFITSndashLEVERAGE PUZZLEREVISITED 1429

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

percentage of firms issuing or repurchasing debt or equity The table alsoreports test statistics from tests that examine whether the percentage differ-ences between the high- and low-profit firms are significantly different fromeach other The results in Column (1) show that the likelihood of issuing debt

Table IV Profitability sorts for debt and equity issuers

The sample contains nonfinancial firms listed on the annual Compustat files for the period

from 1971 to 2009 The table presents the percentage of firms issuing and retiring (orrepurchasing) debt and equity The firms are classified as ldquoissuing debtrdquo if they issuedebt in excess of 5 of the value of their assets as ldquoissuing equityrdquo if they issue equity

in excess of 5 of the value of their assets as ldquoretiring debtrdquo if they retire debt in excess of5 of the value of their assets and as ldquorepurchasing equityrdquo if they repurchase equity inexcess of 5 of the value of their assets In addition we also report net debt issuers which

are firms that issue net debt over 5 of the value of their assets and net equity issuerswhich are firms that issue net equity in excess of 5 of the value of their assets Weannually sort firms on lagged profitability and report the percentage of firms in each ofthese categories The bottom part of each panel reports the percentage of firms issuing or

retiring securities by profitability within the smallest and largest asset quintiles The tablealso reports results from t-test for the difference in percentage of firms issuing or retiringsecurities in the lowest and the highest profit quintile

Percentage of firms

Iss Ret Iss Iss Rep Iss Both IssD IssE

D D NetD E E NetE DampE RepE RetD

(1) (2) (3) (4) (5) (6) (7) (8) (9)

Allfirm

s

Low 276 245 186 301 13 297 71 01 38

2 319 299 181 117 21 112 27 02 20

3 385 344 194 86 29 80 21 05 15

4 386 330 213 81 43 75 20 07 14

High 329 273 198 92 82 81 17 13 14

t Highfrac14Low 150 84 39 723 414 763 358 176 200

Smallfirm

s Low 275 204 201 411 11 405 109 01 47

2 291 233 213 379 10 378 99 01 45

3 281 246 204 265 14 261 66 01 41

4 275 272 186 162 15 159 34 02 35

High 275 281 178 136 27 131 25 02 26

t Highfrac14Low 01 84 28 316 55 318 171 18 55

Largefirm

s

Low 414 363 177 87 22 79 28 03 11

2 449 379 180 62 30 57 17 07 09

3 439 360 186 56 41 48 16 09 07

4 403 320 194 46 67 39 13 12 06

High 374 264 207 50 129 35 10 28 05

t Highfrac14Low 50 130 45 89 244 116 77 120 47

1430 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

is mostly independent of firm profitability In the aggregate there is a smalldifference but it is sensitive to sorting on firm size Similarly in Column (2)there is only a weak relation between the profitability and the likelihood thatthe firm retires debt To a first approximation the fraction of firms issuing orrepurchasing debt is largely independent of profits Large firms are somewhatmore likely than small firms to issue debt and also to retire debtUnlike debt issues the probability of issuing equity is strongly related to

profitability as can be seen in Column (4) Low-profit firms are much morelikely to issue equity than are high-profit firms In the lowest profit quintile301 of the firms issued equity In the highest profit quintile it was only92 The lower panel reports separate sorts for small and large firms Similarpatterns are found for both groups of firms that is the likelihood of equityissuance declines as profitability increases The differences between the quin-tiles are quite stark for the group of small firms All quintiles of small firmsissue equity much more frequently than the corresponding large firm quintilesHigh-profit firms are much more likely to repurchase equity as shown in

Column (5) In the overall sample only 13 of low-profit firmsrepurchased equity compared with 82 of the high-profit firms Whenwe separate the small and large firms again a difference emerges Ingeneral small firms do not repurchase much equity while a greaterfraction of the large firms do soGiven these facts it is natural that low profitability firms are much more

likely to be net issuers of equity than are high profitability firms as shown inColumn (6) Among the firms in the lowest quintiles of profits roughly 30of firms issue net equity that exceeds 5 percent of their assets In contrastfor the most profitable quintile only about 8 issue net equityFirm size is an important variable in the recent literaturemdashit is often used

as a proxy for access to capital markets (as in Faulkender and Petersen(2006) and Leary (2009)) Small firms are bank-dependent risky and infor-mationally opaque They have restricted access to public debt markets andconsequently face more severe supply constraints in their ability to issuedebt Thus we expect small firms to be more sluggish in adjusting theirdebt and equity in response to shocks to profitability Large firms incontrast have much easier access to public debt markets and they facefewer obstacles in accessing securities marketsTo examine how size and profitability interact the bottom part of Table IV

examines profitability sorts for the smallest and largest firms We first sortfirms annually by firm size and then within each size quintile we sort them onprofitability We do these two-way sorts to ensure that we have similarnumber of firms in profitability subgroups for both small and large firmsamples Unconditionally small firms are generally less profitable while

PROFITSndashLEVERAGE PUZZLEREVISITED 1431

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

large firms are relatively more profitable The two-way sorts allow us to havea uniform distribution of firms in various profitability quintiles for both smalland large firm samples The bottom part of Table IV reports results for thesmallest and the largest size quintiles Among the small firms there is littlerelation between profitability and the likelihood of issuing debt Howeverlarger firms exhibit a small increase in the likelihood of net issuance of debtwith increasing profitabilityIn contrast to debt the effects of profitability on equity issuance and

repurchases are much more consistent across size quintiles As profitabilityincreases firms are generally less likely to issue equity and more likely torepurchase it Across the two extreme size groups we note a striking differ-ence between the proportion of firms issuing equity and those issuing debtAmong low-profit small firms almost 41 are equity issuers whereas onlyabout 9 of low-profit large firms are equity issuers However regardless ofsize we note a monotonic reduction in the likelihood of issuing equity asprofitability increasesColumn (7) reports the fraction of firms that issue both debt and equity in

excess of 5 of assets both as a function of profits in the upper panel andas a function of firm size and profits in the lower panel In the upper panelwe see that low-profit firms are much more likely to issue both debt andequity (71) while the high-profit firms are less likely to do so (17) Thelower panel shows that small firms are much more likely to be issuers of bothin contrast to large firms In both firm size categories the profit pattern is thesame with low profits being more frequent issuers of both simultaneouslyImportantly Column (8) shows that the likelihood of issuing debt and

simultaneously repurchasing equity increases with profitability Converselyin Column (9) we find that the likelihood of doing the reverse that is issuingequity and retiring debt declines with profitability The effects of sorting onfirm size and firm profitability mirror those reported for all firms Overallthe results show that low-profit firms are less likely to issue debt and repur-chase equity they are instead more likely to issue equity and retire debtFirms with high profitability exhibit the reverse patternIn Appendix Table AII we present results from profitability (and size)

quintiles for large issuances defined using the 10 cutoff (in Panel A) andthe 15 cutoff (in Panel B) The results from these panels are largely con-sistent with those described above

52 MAGNITUDES OF ISSUANCES AND EQUITY CHANGES

In the previous section we considered the probability of having a nontriviallevel of debt or equity activity The next question is how large are the dollar

1432 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

values involved In Table V we sort firms according to profits and thentabulate the levels and changes in both debt and equity We do this firstfor all firms and then for small versus large firmsFor the analysis on all firms we observe in Column (1) that debt peaks at

the middle of the distribution This is because firms with medium profitabil-ity are also the largest as seen in Column (8) which reports average assetvalues for different profitability quintiles When we sort first by firm size andthen examine firms with different levels of profitability within size quintileswe find that for small firms debt is roughly independent of profits while forlarge firms less profitable firms have a higher level of debtAs expected Columns (3) and (5) show that more profitable firms have

higher equity values Columns (2) and (7) consider the relationship betweenissuances of debt and equity and firm profitability Debt issuances are sig-nificantly larger for more profitable firms In contrast equity issuances arelarger for low-profit firms Small low-profit firms issue substantially lessequity compared with small high-profit firms Among large firms the low-profit ones tend to issue equity while the high-profit ones tend to repurchaseequity The fact that more profitable firms issue debt and repurchase equitywhile the least profitable firms retire debt and issue equity is consistent withthe predicted relation between profitability and financing decisions under thetrade-off theoryWe also examine two-way sorts by size and profitability and report results

for the smallest and largest quintile of firms For the smallest quintile offirms the change in debt is largely unrelated to profits But for large firmsthere is a positive relation between profits and debt issuances High-profitfirms have a big positive change in debt Low-profit large firms have anegative change in debtColumns (4) and (6) provide an explanation of why the leverage ratio

regression results contradict those from the basic profitability sorts pre-sented here As we can see profitability indirectly affects leverage ratiosby increasing equity values Changes in both the book value of equity andthe market value of equity are positive and large for highly profitable firmsIn contrast these changes are negative for less profitable firms This evidenceillustrates an important issue concerning the use of leverage ratios Suchratios are often interpreted as essentially reflecting the use of debt by thefirm This interpretation while common is empirically misleadingFor the typical firm the change in the value of equity is larger than the

change in debt For example in the third profit quintile for large firmscolumn (7) the mean equity issue is just $9 million but the change in themarket value of equity is $242 million At the same time the mean change indebt is $27 million This suggests that a fair bit of the observed variation in

PROFITSndashLEVERAGE PUZZLEREVISITED 1433

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

the leverage ratios is primarily driven by the changes in the market value ofequity in the denominator rather than by the changes in debt in the numer-ator Since equity issues are often small this implies that the variation in theleverage ratio is primarily driven by internal operations rather than byexternal financing actions This again points to the fact that leverageratios can provide a misleading account of actual patterns in the dataThere is always a potential concern that the averages may be misleading

due to the impact of outliers To address this concern in AppendixTable AIII we present median values of the profitability sorts Theseresults generally reinforce the findings in Table VTable VI again considers the magnitudes of financing activity but this

time the issues are scaled by a measure of firm size That way it is easier to

Table V Magnitude of financing activity

The sample contains nonfinancial firms listed on the annual Compustat files for the period

from 1971 to 2009 The table presents average debt and equity changes in debt and equityand equity issuances (in millions of dollars) for firms sorted on profitability within sizeclasses The table reports information for all firms sorted on profitability and for profit-

ability sorts within the smallest and largest firms The sorts are done annually The tablealso reports results from t-test for the difference in percentage of firms issuing or retiringsecurities in the lowest and the highest profit quintile

D D BVE BVE MVE MVE EqyIss Assets

(1) (2) (3) (4) (5) (6) (7) (8)

Allfirm

s

Low 83 5 78 10 120 5 7 234

2 605 13 509 24 721 8 17 1691

3 1227 9 1123 22 1834 3 9 3551

4 833 23 1166 7 2408 74 9 2994

High 570 53 1171 71 3817 68 36 2581

t Highfrac14Low 31 9 36 9 38 2 16 36

Smallfirm

s Low 25 04 42 01 332 28 33 97

2 29 06 55 02 278 16 21 116

3 31 07 63 01 212 05 12 127

4 30 06 70 05 179 10 08 136

High 24 06 84 14 234 17 07 142

t Highfrac14Low 04 15 153 85 73 11 201 107

Largefirm

s Low 3929 55 2990 152 4671 11 63 10846

2 3826 19 3178 95 5122 66 11 10721

3 3074 27 3498 41 6173 242 9 9697

4 2623 62 3915 45 8473 94 57 9855

High 2180 193 4642 261 14924 202 148 10171

t Highfrac14Low 94 64 108 87 246 12 169 16

1434 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

see whether large firms behave disproportionately when compared withsmall firms Column (1) shows that low-profit firms have more debtas reported in many previous studies In Columns (4) and (6) we seethat low-profit firms have a major decline in both the book value ofequity and the market value of equity The high-profit firms have thereverse experienceMore interesting are the active decisions of the firms In Column (2) we see

that in contrast to the usual impression the low-profit firms are reducingdebt while the high-profit firms are increasing it In Column (7) we see thatthe low-profit firms are also reducing leverage by issuing equity The high-profit firms issue little equity In the lower panel similar results are foundwhen we separate small and large firms Equity issuing is particularly strong

Table VI Magnitude of financing activitymdashscaled levels and changes

The sample contains nonfinancial firms listed on the annual Compustat files for the period

from 1971 to 2009 The table presents average debt and equity changes in debt and equityand equity issuances scaled by assets for firms sorted on profitability within size classes Thetable reports information for all firms sorted on profitability and for profitability sorts

within the smallest and largest firms The sorts are done annually The table also reportsresults from t-test for the difference in percentage of firms issuing or retiring securities inthe lowest and the highest profit quintile

D D BVE BVE MVE MVE EqyIss

(1) (2) (3) (4) (5) (6) (7)

Allfirm

s

Low 386 122 12 871 3427 1868 217

2 285 58 407 119 940 157 28

3 293 12 425 14 742 03 12

4 269 26 461 02 955 27 10

High 192 15 551 47 1675 74 07

t Highfrac14Low 35 49 50 47 127 42 157

Smallfirm

s Low 545 154 841 2145 9688 9128 696

2 684 49 712 1915 5014 2318 288

3 331 253 141 969 3063 954 129

4 267 181 357 175 1673 221 55

High 194 05 543 26 1758 147 41

t Highfrac14Low 49 17 51 36 88 27 99

Largefirm

s Low 350 30 306 58 511 55 09

2 346 14 335 13 523 02 04

3 353 165 332 71 711 29 01

4 280 01 414 04 909 09 05

High 227 15 477 19 1643 27 14

t Highfrac14Low 371 112 453 181 570 57 221

PROFITSndashLEVERAGE PUZZLEREVISITED 1435

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

among the small low-profit firms that are also experiencing sharp reductionsin the value of equity In column (2) we see that both for the large and for thesmall firms net debt increases are primarily found in the highest quintile offirm profitability

6 Debt and Equity Regressions

So far we have demonstrated that in our data the conventional leverageregressions have the usual signs We then demonstrated that in simple sortsof the data strikingly different active financing patterns emerge The nexttask is to reconcile these differencesTable VII presents simple regressions predicting changes in debt book

equity and market equity as well as equity issuances Consistent with thesorts we control for changes in firm size and we also include firm and yeardummies There may be a concern about the speed of a firmrsquos responsesAccordingly we consider the lagged change in profits as the main case butwe also include the contemporaneous change in profits Whether the contem-poraneous change in profits is included or not does not in fact matter Wereport panel-robust standard errors adjusted for clustering at the firm levelThe results in Table VII support the findings in the sorts Firms that have

an increase in profits in 1 year increase debt both in the same year and in thenext year Firms that have an increase in profits experience an increase inboth the book value of equity and the market value of equity Importantlyhowever firms that experience and increase in profits reduce their equityissues All these effects are statistically significant (except for the marketvalue of equity in the column 5 specification) and seem intuitively reason-able In Appendix Table AIV similar regressions to those presented aboveare reported but without firm-fixed effects We find that empirically verylittle changes whether we include firm fixed effects or leave them out Thusthe targeting behavior to the extent that it happens does not account forwhat we are reportingThe next step is to bring the sorts and the simple regressions together into

a common setting We therefore include factors in addition to profitability(i) median industry leverage (ii) the market-to-book assets ratio (iii) tangi-bility of assets and (iv) firm size (measured by the log of assets) Rajan andZingales (1995) show that these factors are related to leverage in G7countries A number of studies have used these factors to estimateleverage targets Frank and Goyal (2009) show that these factors arerobustly related to leverage in the USA In these regressions we use indica-tors for the quintile that the firm is in for each factor

1436 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

Table VIII reports these results In Columns (1) and (2) we examine debtissues and again show that profits positively affect debt issuances The effectsare large and statistically significant at the 1 level Columns (3) and (4)which examine book equity show a strong effect of profits This effect is onlyslightly reduced by the inclusion of other factors Columns (5) and (6) illus-trate that the impact of profits on the change in market equity is also robustto the inclusion of the conventional factors Columns (7) and (8) examineequity issuances Here we again find that profits have a significant negativeimpact on equity issuances As before the results are robust to the inclusionof the conventional factorsIn unreported tables we estimate debt and equity changes using quantile

regressions In terms of the signs and significance of the profitability variable

Table VII Debt and equity changes

The sample contains nonfinancial firms listed on the annual Compustat files for the period

from 1971 to 2009 The table presents fixed effect estimates from regressions of changes indebt (D) changes in the book value of equity (BVE) changes in the market value ofequity (MVE) and net equity issuances (EqyIss) The explanatory variables include

current and lagged change in operating income before depreciation (Profits andProfitst1) and lagged change in the book value of assets (Assetst1) In addition tofixed firm effects the regressions include year indicator variables The reported t-statistics

are corrected for clustering at the firm level

aMean significant at the 1 level

bMean significant at the 5 level

cMean significant at the 10 level

D BVE MVE EqyIss

(1) (2) (3) (4) (5) (6) (7) (8)

Profitst1 022a 025a 015b 035a 004 060b 005b 007a

(33) (31) (25) (35) (03) (25) (25) (53)

Profits 050b 078a 140a 003a

(24) (32) (48) (34)

Assetst1 000 007b 025a 001b

(00) (21) (35) (25)

Constant 167 283a 34 51 1145a 1182a 605a 595a

(14) (30) (04) (05) (32) (29) (69) (74)

Year FE Yes Yes Yes Yes Yes Yes Yes Yes

Firm FE Yes Yes Yes Yes Yes Yes Yes Yes

R2 001 004 001 010 001 005 002 003

N 162056 162056 162130 162130 157550 157550 162154 162154

PROFITSndashLEVERAGE PUZZLEREVISITED 1437

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

the results are qualitatively similar to the OLS results reported in Table VIIIProfitability positively affects debt issuances and negatively affects equityissuances across various quantiles Overall the results in Table VIII showthat there is nothing that is special about using sorts or regressions to explaindebt and equity issues In either case we find that more profitable firms tendto increase their debt experience an increase in the value of equity andrepurchase shares Thus the control factors are not responsible for theusual rejection of the trade-off theory It would appear that what mattersis the size of the issuing activity relative to the organic increases in equityvalue due to profits In other words what matters in the leverage regressionseems to be coming from the presence of E in D

DthornE and not from D

Table VIII Debt and equity issuances profitability and leverage factor quintiles

The sample contains nonfinancial firms listed on the annual Compustat files for the period

from 1971 to 2009 The table presents estimates from regressions of change in debt (D)change in the book value of equity (BVE) change in the market value of equity (MVE)and net equity issuances (EquityIss) The explanatory variables include lagged profitability

quintiles and lagged leverage factor quintiles In addition to fixed firm effects the regres-sions include year indicator variables The reported t-statistics are corrected for clustering atthe firm level

aMean significant at the 1 level

bMean significant at the 5 level

cMean significant at the 10 level

D BVE MVE EqyIss

(1) (2) (3) (4) (5) (6) (7) (8)

ProfitabilityQint 214a 167a 283a 175a 80 429a 53a 55a

(70) (54) (66) (59) (08) (48) (55) (62)

IndLevQint 104a 43 99 20

(29) (15) (10) (10)

Ln(Assets)Qint52 17 1026a 53a

(16) (05) (77) (35)MB

Qint158a 372a 1155a 09

(45) (56) (106) (06)

TangibilityQint 05 177 436c 64a

(01) (15) (18) (31)

Constant 167 96 656b 235 4613a 8610a 113 570a

(08) (04) (26) (04) (61) (58) (11) (31)

Year FE Yes Yes Yes Yes Yes Yes Yes Yes

Firm FE Yes Yes Yes Yes Yes Yes Yes Yes

R2 0002 0002 0003 0004 0013 0014 0013 0014

N 155587 155037 155587 155037 155587 155037 155587 155037

1438 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

7 Are Financial Market Conditions Important

According to the trade-off theory capital structure is determined by con-sidering the impact of costs and benefits of debt The time variation inthe costs and benefits of debt imply that issuance decisions would alsovary over time Furthermore market conditions affect the transactioncosts of adjusting Adverse selection is a greater problem in a cold marketthan in a hot one Accordingly it may be easier to adjust in good marketconditions Related ideas have been developed by Huang and Ritter(2009) Faulkender et al (2012) and Halling Yu and Zechner (2012)Faulkender Flannery Hankins and Smithrsquos paper examines cross-sectionalvariation in adjustment speeds including market valuations andhow they may affect adjustment speeds It is natural to think that ingood market conditions it will be less costly to issue both debt and equityand adjusting leverage to respond to profitability shocks would be mucheasierTo test the importance of market conditions we require a definition of

good times and bad times Our empirical strategy is to estimate good timesversus bad times at the four-digit industry level We define an industry ashaving ldquogood timesrdquo if the median firm in that industry has a market-to-book ratio that is higher than the 67th percentile of the time-series distribu-tion of the industry median market-to-book ratios Conversely an industryis defined as having ldquobad timesrdquo if the median firm in the industry has amarket-to-book ratio that is lower than the 33rd percentile of the time-seriesdistribution of the industry median market-to-book ratiosPanel A of Table IX tabulates issuance activity for profitability sorts in

both good times and bad times As might be expected active debt and equityissuances are larger during good times There is also somewhat more activeswaps between debt and equity during good timesPanel B reports the issuance activity for the smallest and the largest firms

during the good times and bad times Within each size quintile firms aresorted on profitability This is a three-way sort As before we find that thereis much more active use of external markets during good times Small low-profit firms are more likely to issue equity in good times than in bad timesDebt issuances are significantly higher in good times Large high-profitfirms are significantly more likely to issue debt and repurchase equity ingood times than in bad timesPanel C reports the magnitudes of the financial variables rather than the

frequencies During bad times less profitable large firms retire substantialamounts of debt and they show a tendency to issue equity Small firms donot seem to engage in similar debt reduction activities However like the

PROFITSndashLEVERAGE PUZZLEREVISITED 1439

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

Table IX Debt and equity issues in good and bad times

The table reports the frequency and magnitude of financing activity for sorts on profits for subsamples

of firms in good and bad times An industry is defined as having ldquogood timesrdquo if the median firm in

that industry has a market-to-book ratio that is higher than the 67th percentile of the time-series

distribution of industry median market-to-book ratios Conversely an industry is defined as having

bad times if the median firm in that industry has a market-to-book ratio that is lower than the 33rd

percentile of the time-series distribution of industry median market-to-book ratios Panels A and B

report the percentage of firms (i) issuing net debt in excess of 5 of the value of their assets (ii) issuing

net equity in excess of 5 of the value of their assets (iii) issuing debt and repurchasing equity both in

excess of 5 of the value of their assets and (iv) issuing equity and retiring debt both in excess of 5

of the value of their assets Panel C tabulates the mean debt and equity levels and the mean changes in

debt and equity issuances The sample contains nonfinancial firms listed on the annual Compustat files

for the period from 1971 to 2009

Iss Iss Iss D Iss ENet D Net E Rep E Ret D(1) (2) (3) (4)

Panel A sorts on profitability

Badtimes

Low 145 131 01 202 170 49 02 083 173 46 04 074 182 47 03 10

High 177 51 06 10t Highfrac14Low 41 136 33 01

Goodtimes

Low 206 381 01 462 197 148 02 233 212 103 04 184 234 92 10 18

High 216 92 16 14t Highfrac14Low 23 658 138 23

Panel B sorts on firm size and profitability

Badtimes

Smallfirm

s Low 151 318 00 312 182 276 02 253 151 140 02 294 183 73 00 19

High 177 80 00 11t Highfrac14Low 10 90 NA 20

Largefirm

s Low 172 54 01 042 148 39 04 053 150 43 04 044 156 33 03 06

High 147 28 05 03t Highfrac14Low 17 32 20 02

Goodtimes Smallfirm

s Low 213 448 01 542 230 443 01 473 222 325 00 474 203 197 02 38

High 190 153 02 31t Highfrac14Low 19 230 10 39

Largefirm

s Low 203 96 03 122 208 66 09 133 232 54 16 064 215 39 19 05

High 234 36 38 04t Highfrac14Low 29 99 92 38

(continued)

1440 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

larger firms there is a tendency for small firms to issue more equitymdashat leastto the extent possibleDuring good times large profitable firms raise significant amounts of debt

as they experience an increase in the value of their equity Such firms alsoengage in share repurchases Debt issues by smaller firms are much lessaffected by their own profitability during good times and their change indebt is more or less independent of their profits However equity issuanceis primarily found among the low-profit small firms during good timesTable IX shows that empirically profitability has a time-varying impacton leverage choicesIn Table X the basic regressions for changes in debt and for equity

issuance are presented for both good times and bad times Comparinggood times with bad times we see that the effects of profits are muchstronger in good times In bad times the impacts are rather weak statisticallyThis difference is interesting and deserves further study However for ourpurposes recall that the static trade-off theory implies time-varying capital

Table IX Continued

D D BVE BVE MVE MVE EqyIss Assets

(1) (2) (3) (4) (5) (6) (7) (8)

Panel C magnitude of financing activity

Badtimes Smallfirm

s Low 40 03 58 14 305 06 22 140

2 44 05 82 07 240 67 17 180

3 45 04 94 02 176 18 05 187

4 43 09 98 02 137 04 03 188

High 36 10 114 13 209 16 04 199

Largefirm

s Low 3365 73 2523 169 2304 163 94 8972

2 3031 95 2983 35 3284 278 56 9068

3 2930 5 3360 14 3947 308 57 9225

4 2281 6 3817 19 5213 442 7 8902

High 1790 62 4767 13 8799 58 31 9573

Goodtimes

Smallfirm

s Low 24 06 38 05 339 23 39 90

2 26 08 48 01 301 04 25 104

3 30 09 60 04 246 03 17 125

4 28 07 66 08 206 18 12 130

High 23 06 85 16 261 13 08 141

Largefirm

s Low 4532 36 3477 103 6557 260 53 12465

2 4580 58 3527 139 6911 421 23 12522

3 2952 111 3485 2 7586 279 32 9535

4 2752 124 3939 72 10686 33 105 10226

High 2429 228 4819 377 18565 268 222 10830

PROFITSndashLEVERAGE PUZZLEREVISITED 1441

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

structure choices even when the target is time invariant Empirically we doobserve time-varying choices

8 Is Scaling Important

In the preceding section we provided results for unscaled debt and equityissues This seems appropriate to us However it involves making twochanges to the standard regression We would like to determine if both

Table X Debt and equity changes in good and bad timesmdashregressions

The sample contains nonfinancial firms listed on the annual Compustat files for the period

from 1971 to 2009 The table presents estimates from regressions of change in debt (D)and net equity issuances (EqyIss) in good and bad times The explanatory variables includechange in profits lagged change in profits and lagged change in assets The regressions

include both firm and year fixed effects An industry is defined as having ldquogood timesrdquo ifthe median firm in that industry has a market-to-book ratio that is higher than the 67thpercentile of the time-series distribution of industry median market-to-book ratios

Conversely an industry is defined as having bad times if the median firm in thatindustry has a market-to-book ratio that is lower than the 33rd percentile of the time-series distribution of industry median market-to-book ratios The reported t-statistics arecorrected for clustering at the firm level

aMean significant at the 1 level

bMean significant at the 5 level

cMean significant at the 10 level

Good times Bad times

D EqyIss D EqyIss

(1) (2) (3) (4) (5) (6) (7) (8)

Profits 087a 091a 006b 009a 060 065 003 003b

(30) (29) (20) (28) (15) (16) (16) (20)

Profitst1 029c 015a 034c 001

(19) (37) (17) (04)

Assetst1 000 002b 000 001

(02) (22) (00) (07)

Constant 1337 3989 8691a 10001a 1351 1851 4099a 4055a

(03) (09) (55) (37) (09) (13) (80) (78)

Year FE Yes Yes Yes Yes Yes Yes Yes Yes

Firm FE Yes Yes Yes Yes Yes Yes Yes Yes

R2 0071 0080 0018 0045 0178 0212 0034 0041

N 74827 68729 74891 68787 26965 26013 26971 26018

1442 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

changes are crucial Accordingly in this section we take a step back toconsider prior literature in which variables are scaled The questions arehow does scaling affect the results and whyTable XI examines alternative scalings Columns (1) and (2) scale debt and

equity issuances by the total firm net issuance Columns (3) and (4) scaledebt and equity issuances by the sum of debt and book equity Columns (5)and (6) consider the change in debt scaled by the sum of debt and bookequity Columns (7) and (8) consider the change in leverage due to externalfinancing Since we are examining changes we use the changes in thestandard regression factors as controls The primary interest is in the coef-ficient on the change in profitsWe find that a change in profits is associated with an increase in debt

issues But the sign reverses when we scale the debt issues by the sum of debtplus book equity This is telling What is really driving the results is thechange in equity in the denominator not the change in debt in the numer-ator This result is further substantiated in Columns (5) and (6) where thechange in debt to capital ratio is used instead of debt issuanceTo further explore the impact of active issuance relative to passive

experiencing of changes we decompose leverage by separating out theeffect of debt and equity issues from those due to changes in retainedearnings Thus we define leverage change due to external financing as

LevEF frac14Dt1 thorn d

Dt1 thorn Et1 thorn dthorn e

Dt1

Dt1 thorn Et1 eth1THORN

where D is debt E is book value of equity d is debt issues net of retirementsand e is equity issues net of repurchases Garvey and Hanka (1999) use asimilar measure in their study This measure directly captures the netleverage effect of debt and equity issuancesColumns (7) and (8) report results from regressions of the net leverage

effect of debt and equity issues on changes in profitability after controllingfor other leverage factors in differences and year indicator variables Theeffect of profitability on leverage change due to external financing is positiveand statistically significant This comparison suggests that changes in totaldebt and equity are only partly a result of debt and equity issuance decisionsOther balance sheet adjustments complicate the inferences from leverageratio regressionsFor most of our analysis we follow the literature in treating debt as homo-

genous It is of course potentially interesting to ask what happens when wedistinguish between short-term debt and long-term debt Which part of thefirmrsquos debt maturity structure drives the under-adjustment Table XII

PROFITSndashLEVERAGE PUZZLEREVISITED 1443

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

provides results from such decompositions Alternative scalings areconsidered The main results of interest are that the distinction betweenshort- and long-term debt is important Short-term debt issuance is largelyindependent of changes in profits The negative sign on profit change stemsfrom the action of long-term debt This seems intuitively reasonable

Table XI Scaled issuances

The sample contains nonfinancial firms listed on the annual Compustat files for the period

from 1971 to 2009 The table presents estimates from regressions of debt issuance to totalissuance (in columns (1) and (2)) debt issuance to total capitalization (in columns (3) and(4)) change in debt to capitalization ratio (in columns (5) and (6)) and change in leverage

ratio considering only external financing (in columns (7) and (8)) The change in leverageratio considering only external financing is defined as

LevEF frac14Dt1 thorn d

Dt1 thorn Et1 thorn dthorn e

Dt1

Dt1 thorn Et1

where D is debt E is book value of equity d is debt issues net of retirements and e is equity

issues net of repurchases The regressions include lagged leverage factors and year indicatorvariables Some of the specifications include firm fixed effects The reported t-statistics arecorrected for clustering at the firm level

aMean significant at the 1 level

bMean significant at the 5 level

cMean significant at the 10 level

DebtIssTotIss

DebtIssethDthornETHORN D

DthornE LevEF

(1) (2) (3) (4) (5) (6) (7) (8)

Profitt1 004a 000 002a 001c 003a 002 001b 002a

(35) (03) (34) (18) (35) (14) (23) (26)

IndLevt1 004c 001 001 001c 000 000 000 000

(19) (06) (15) (19) (02) (04) (03) (05)

Market=bookt1 001a 001a 000a 000a 000a 000a 000 000c

(47) (85) (68) (60) (55) (36) (05) (18)

Tangt1 004 005 013a 011a 011a 006a 003a 003a

(12) (15) (144) (112) (79) (38) (44) (42)

Assetst1 008a 004a 006a 004a 002a 001a 003a 002a

(97) (45) (283) (161) (53) (35) (147) (101)

Constant 079a 074a 002a 003a 000 000 000b 002a

(497) (442) (91) (111) (00) (05) (22) (93)

Year FE Yes Yes Yes Yes Yes Yes Yes Yes

Firm FE No Yes No Yes No Yes No Yes

R2 0015 0098 0022 0076 0009 0070 0012 0086

N 133905 133905 140552 140552 138229 138229 140572 140572

1444 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

inasmuch as short-term debt is probably more associated with day-to-dayoperations rather than longer term planning considerationsIn Table XIII we examine change in leverage ( D

DthornE) and change inleverage due to external financing (LEVEF) during good and bad timesseparately The traditional negative effect of a change in profits on achange in leverage is found both in good times and in bad timesHowever when we decompose the impact in order to isolate the externalcomponent as before the result disappears In good times the externalfinancing component of a change in profits is weak but still positiveDuring bad times it is not statistically significant This implies that distin-guishing between good and bad times does not fundamentally alter the basic

Table XII Short- and long-term debt issuances

The sample contains nonfinancial firms listed on the annual Compustat files for the period

from 1971 to 2009 The table presents estimates from regressions of short-term debtissuance to total issuance (in columns (1) and (2)) long-term debt issuance to totalissuance (in columns (3) and (4)) short-term debt issuance to total capitalization (in

columns (5) and (6)) and long-term debt to total capitalization (in columns (7) and (8))Some of the specifications include firm fixed effects The reported t-statistics are correctedfor clustering at the firm level

aMean significant at the 1 level

bMean significant at the 5 level

cMean significant at the 10 level

STDebtIssTotIss

LTDebtIssTotIss

STDebtIssethDthornETHORN

LTDebtIssethDthornETHORN

(1) (2) (3) (4) (5) (6) (7) (8)

Profitabilityt1 0002 0003 0042a 0008 0001 0001 0019a 0012a

(02) (04) (30) (05) (14) (10) (45) (28)

IndLevt1 0009 0007 0070b 0033 0001 0002c 0014a 0005

(06) (05) (26) (11) (08) (17) (26) (09)

M=Bt1 0003a 0003a 0003b 0010a 0000 0000 0003a 0002a

(48) (35) (22) (65) (10) (01) (84) (42)

Tangibilityt1 0020 0007 0011 0067 0005a 0006a 0097a 0081a

(12) (03) (03) (16) (34) (33) (108) (89)

Assetst1 0039a 0011b 0034a 0024b 0001a 0001 0046a 0020a

(89) (22) (38) (23) (36) (13) (190) (82)

Constant 0001b 0153a 0798a 0591a 0000c 0006a 0057a 0096a

(25) (144) (470) (288) (17) (124) (292) (234)

Year FE Yes Yes Yes Yes Yes Yes Yes Yes

Firm FE No Yes No Yes No Yes No Yes

R2 0025 0089 0018 0073 0030 0147 0016 0298

N 136146 136146 136140 136140 143031 143031 142847 142847

PROFITSndashLEVERAGE PUZZLEREVISITED 1445

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

Table XIII Externally financed leverage changes in good and bad times

The sample contains nonfinancial firms listed on the annual Compustat files for the period

from 1971 to 2009 The table presents estimates from regressions of change in debt tocapitalization ratio (in columns (1) (2) (5) and (6)) and change in leverage ratio consider-ing only external financing (in columns (3) (4) (7) and (8)) The change in leverage ratio

considering only external financing is defined as

LevEF frac14Dt1 thorn d

Dt1 thorn Et1 thorn dthorn e

Dt1

Dt1 thorn Et1

where D is debt E is book value of equity d is debt issues net of retirements and e is equityissues net of repurchases The regressions include lagged leverage factors and year indicator

variables The reported t-statistics are corrected for clustering at the firm level An industryis defined as having ldquogood timesrdquo if the median firm in that industry has a market-to-bookratio that is higher than the 67th percentile of the time-series distribution of industry

median market-to-book ratios Conversely an industry is defined as having ldquobad timesrdquoif the median firm in that industry has a market-to-book ratio that is lower than the 33rdpercentile of the time-series distribution of industry median market-to-book ratios

aMean significant at the 1 level

bMean significant at the 5 level

cMean significant at the 10 level

Good times Bad times

DDthornE LevEF D

DthornE LevEF

(1) (2) (3) (4) (5) (6) (7) (8)

Profitabiltyt1 0022c 0005 0016c 0025b 0073b 0041 0007 0007

(17) (03) (18) (24) (25) (11) (04) (03)

IndLevt1 0000 0011 0001 0009 0010 0018 0001 0000

(00) (09) (01) (13) (07) (10) (02) (00)

MB

t1

0004a 0003b 0000 0001 0008a 0001 0000 0000

(40) (24) (05) (08) (28) (04) (03) (03)

Tangt1 0097a 0041 0033a 0038b 0127a 0048 0059a 0033

(45) (16) (27) (25) (49) (14) (43) (16)

Assetst1 0014a 0013b 0024a 0019a 0037a 0023c 0036a 0025a

(30) (20) (84) (48) (45) (17) (88) (43)

Constant 0030b 0036 0000 0003 0028c 0014 0021 0001

(20) (16) (00) (07) (17) (06) (11) (01)

Year FE Yes Yes Yes Yes Yes Yes Yes Yes

Firm FE No Yes No Yes No Yes No Yes

R2 0006 0088 0009 0117 0026 0251 0023 0202

N 56820 56820 58200 58200 23462 23462 23511 23511

1446 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

conclusions The traditional leverage result on profits is driven by increasesin equity that are experienced and then partially but not completely offsetby firmsTo the extent that firms are engaging in active financing they are following

the prescriptions of the trade-off theory Thus we have a coherent picture ofhow firms are responding to profits Basically more profitable firms experiencean increase in the value of their equity They respond to this by issuing debtand repurchasing equity But the magnitude of the increase in equity is largerthan the magnitudes of the active steps taken in the debt and equity marketsThis means that the firmsrsquo actual actions are as predicted by the trade-offtheory But it leaves open the question of why they do not act more stronglyThe capital structure literature has long been interested in dynamics and

capital structure adjustment paths In recent years this interest has increas-ingly resulted in explicit dynamic optimizing models such as those ofGoldstein Ju and Leland (2001) Hennessy and Whited (2005) andStrebulaev (2007) These models commonly make specialized assumptionsleading some to question their generalityWe argue that both fixed transactions costs and variable transactions costs

are plausible and realistic in models that pin down firm size The simultaneouspresence of both is important Due to fixed costs firms only rebalance inter-mittently Due to variable adjustment costs the extent of the adjustment isreduced to the point at which an extra unit to adjustment is balanced againstthe extra unit transactions costs This means that both leverage increases andleverage reductions will only go part way toward a static optimumThe key point is that with both fixed and variable transactions costs we

should expect to only see intermittent leverage adjustments When leverage isadjusted it should ldquounder adjustrdquo when compared to a frictionless modelFurther discussion of transactions costs in a capital structure setting isprovided by Leary and Roberts (2005)

9 Conclusion

The connection between corporate profits and capital structure has beenvery influential in the assessment of the trade-off theory The conventionalinterpretation of the evidence has pushed the literature away from the trade-off theory and toward much more complex dynamic models and ideas As aresult it is important to make sure that the evidence has been correctlyinterpreted We show that the problem is with the interpretation of thetest results and a failure to take into account the costs associated with re-sponding to a profit shock

PROFITSndashLEVERAGE PUZZLEREVISITED 1447

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

We document that more profitable firms really do borrow more (not less)They tend to repurchase equity They experience an increase in both thebook value of equity and the market value of equity Less profitable firmsreally do tend to reduce their debt and to issue equityFirm size and market conditions also matter Larger firms tend to be more

active in the debt markets while smaller firms tend to be relatively moreactive in the equity markets During good times there is more use ofexternal financingThe usual profitsndashleverage puzzle result is primarily driven by the increase

in equity that is experienced by the more profitable firms Accordingly thepuzzle should be restated as asking why do firms not take sufficiently largeoffsetting actions to fully undo the change in equity What limits the mag-nitudes of the typical leverage response to profit shocksIn a frictionless model the partial response appears puzzling But there is

good empirical reason to believe that rebalancing entails both fixed andvariable costs and that firm size matters The rebalancing costs can befully avoided by doing nothing Accordingly the firm must decide whetherany given shock is big enough to be worth responding to If it is then thefirm must decide how big a response is called for These technical conditionsare known as ldquovalue matchingrdquo and ldquosmooth pastingrdquo Thus optimizationimplies that some shocks will be ignored Even if the shock is not ignoredthe optimal response will only partially undo the shock The magnitude ofthe leverage response must balance the marginal cost and the marginalbenefit of an extra unit of leverage Since the marginal cost of adjustingleverage is strictly positive the adjustment toward that static leverageoptimum will only go part way This is true both for leverage increasesand for leverage reductionsOne might conjecture that the variable costs would be time varying and

that such costs might have declined over time Since we do not have a fullaccounting of all of the variable costs it is of course impossible to be sureHowever there is some evidence suggesting that at least the directunderwriting fees are remarkably stable over time10

Surely nobody ought to be surprised that more profitable firms are moreinclined to repurchase equity while unprofitable firms tend not to do so Thefact that more profitable firms typically experience an increase in the value of

10 For IPOs Jay Ritter provides important summary evidence as httpbearwarringtonufleduritterIPOs2013Underwritingpdf He finds that for medium-sized deals a remark-

ably large fraction of the deals have a 7 fee year after year Whether this is due tobehavioral factors or some sort of a coordination device is not clear The extent towhich similar intertemporal stability carries over to other forms of external financemerits further study

1448 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

equity is equally natural Were it not for the contrary claims in the literature(ie the profitsndashleverage puzzle) we would argue that nobody ought to findit surprising that more profitable firms are more inclined to issue debt Thefact that leverage adjustments typically only partially offset profit shocksshould also sound quite natural These facts all fit together easily within thetrade-off theory once we take account of both fixed and variable costs ofactively adjusting capital structure

Appendix

Table AI Market leverage regressions

The table presents estimates of the regressions of market leverage on firm characteristics

The sample comes from the annual Compustat files during the period 1971ndash2009 Financialfirms are excluded The leverage is estimated as the ratio of debt over debt plus marketequity The explanatory variables Profitabilityt1

MB

t1

Tangibilityt1 and LnethAssetsTHORNt1are described in Table 1 IndMedianLevt1 is estimated as the median market leverage of allother firms in the same industry The industry is defined at the level of the four-digit SICcode All specifications include the year fixed effects The specifications in column (2)

additionally include the firm fixed effects We report t-statistics where the standard errorsare clustered at the firm level in parentheses

aSignificant at the 1 level

Quantile regressions

(1) (2) (3) (4) (5)

Profitabilityt1 013a 013a 007a 011a 020a

(275) (262) (349) (425) (538)

IndLevt1 046a 020a 032a 058a 063a

(506) (239) (1299) (1630) (1042)MB

t1

003a 001a 001a 002a 004a

(509) (303) (743) (709) (534)

Tangibilityt1 013a 018a 013a 014a 010a

(191) (181) (762) (571) (232)

LnethAssetsTHORNt1 001a 005a 001a 001a lt001a

(112) (258) (682) (342) (29)

Constant 010a 011a 005a 005a 029a

(171) (98) (173) (118) (398)

Year FE Yes Yes Yes Yes Yes

Firm FE No Yes No No No

Regression type OLS FE 25th tile Median 75thtile

Clustered SE Yes Yes No No No

R2 030 067

Pseudo R2 016 023 0201

N 158578 158578 158578 158578 158578

PROFITSndashLEVERAGE PUZZLEREVISITED 1449

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

Table AII Profitability sorts for firms issuing large amounts

Percentage of firms

Iss Ret Iss Iss Rep Iss Iss IssD IssE

D D NetD E E NetE DampE RepE RetD

(1) (2) (3) (4) (5) (6) (7) (8) (9)

Panel A 10 cutoff to define issuing or retiring activity

Allfirm

s

Low 205 150 134 260 05 257 45 00 23

2 207 175 103 77 06 75 11 01 10

3 243 196 101 50 08 47 08 01 09

4 252 186 113 50 13 48 08 02 08

High 227 161 118 60 27 57 07 04 09

t Highfrac14Low 70 38 62 755 226 768 320 90 152

Smallfirm

s Low 230 132 162 380 05 375 82 00 32

2 228 143 165 334 04 333 65 00 26

3 211 146 147 216 03 214 38 00 23

4 190 154 120 126 04 124 15 01 20

High 194 162 114 99 09 97 12 01 15

t Highfrac14Low 42 40 68 342 20 340 169 07 131

Largefirm

s Low 231 200 82 31 06 29 08 01 04

2 231 187 67 17 06 14 03 01 02

3 234 177 77 17 09 14 03 02 03

4 218 158 75 16 17 14 02 03 02

High 218 135 99 19 41 15 03 07 02

t Highfrac14Low 19 106 36 46 135 57 39 58 23

Percentage of firms

Iss Ret Iss Iss Rep Iss Both Iss D Iss E

D D Net D E E Net E DampE Rep E Ret D

(1) (2) (3) (4) (5) (6) (7) (8) (9)

Panel B 15 cutoff to define issuing or retiring activity

Allfirm

s

Low 158 101 100 230 01 228 27 00 15

2 142 115 63 57 01 56 05 00 06

3 165 128 57 32 01 31 03 00 05

4 173 118 65 35 01 34 03 00 05

High 160 102 70 46 04 44 03 00 05

t Highfrac14Low 08 06 141 738 87 746 277 24 131

Smallfirm

s Low 195 95 132 354 01 350 56 00 24

2 185 96 131 303 01 301 44 00 18

3 160 96 104 179 00 176 18 00 15

4 134 99 78 99 01 98 08 00 12

High 133 95 73 82 01 80 06 00 10

t Highfrac14Low 80 00 95 345 12 344 149 12 52

Largefirm

s Low 145 123 42 14 01 12 03 00 01

2 137 114 32 07 01 06 01 00 01

3 142 104 35 06 01 05 00 00 01

4 130 91 36 08 03 07 01 01 01

High 134 81 48 09 06 07 01 01 01

t Highfrac14Low 19 85 17 26 54 33 34 17 11

1450 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

Table AIV Debt and equity changesmdashwithout fixed effects

The sample contains nonfinancial firms listed on the annual Compustat files for the period from 1971

to 2009 The table presents estimates from regressions of change in debt (D) change in the book

value of equity (BVE) change in the market value of equity (MVE) and net equity issuances

(EqyIss) The explanatory variables include current and lagged change in operating income before

depreciation (Profits and Profitst1) and lagged change in book value of assets (Assetst1) The

regressions include year indicator variables The reported t-statistics are corrected for clustering at the

firm level

aMean significant at the 1 level

bMean significant at the 5 level

cMean significant at the 10 level

D BVE MVE EqyIss

(1) (2) (3) (4) (5) (6) (7) (8)

Profitst1 025a 024a 023a 038a 007 066a 007a 009a

(40) (32) (36) (39) (06) (29) (42) (51)Profits 051b 082a 149a 004a

(25) (33) (49) (33)Assetst1 002 005c 024a 001b

(08) (17) (34) (23)Constant 93 176a 45 71 1552a 1704a 130a 140a

(12) (28) (07) (09) (49) (49) (45) (45)Year FE Yes Yes Yes Yes Yes Yes Yes YesR2 001 005 001 011 001 005 002 002N 162056 162056 162130 162130 157550 157550 162154 162154

Table AIII Median profitability sorts

The sample contains nonfinancial firms listed on the annual Compustat files for the period from 1971

to 2009 The table presents median debt and equity median changes in debt and equity and median

equity issuances (EqyIss) in millions of dollars for all firms sorted on profitability and for profitability

sorts within the smallest and largest firms The sorts are done annually

D D BVE BVE MVE MVE EqyIss Assets(1) (2) (3) (4) (5) (6) (7) (8)

Allfirm

s

Low 21 00 75 17 256 20 01 1922 186 02 449 18 598 19 00 11033 816 06 1274 03 1637 05 00 32244 787 02 1541 18 2488 03 00 3556High 273 00 1361 66 3300 14 00 2578

Smallfirm

s Low 05 00 13 07 140 14 06 402 06 00 25 08 114 12 01 563 09 00 32 05 85 07 00 674 12 00 39 01 68 03 00 79High 10 00 51 04 98 01 00 90

Largefirm

s Low 11485 381 10269 371 13227 202 03 333902 13051 393 12786 140 17571 169 05 392463 11554 287 13429 24 21623 305 00 370314 9963 209 15052 194 29424 308 00 36705High 7361 00 16380 613 49142 439 00 34976

PROFITSndashLEVERAGE PUZZLEREVISITED 1451

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

References

Auerbach A J (1985) Real determinants of corporate leverage in B M Friedman (ed)

Corporate Capital Structures in the United States University of Chicago Press Chicago

IllinoisBertrand M and Mullainathan S (2003) Enjoying the quiet life corporate governance and

managerial preferences Journal of Political Economy 111 1043ndash1075Booth L Varouj A Demirguc-Kunt A and Maksimovic V (2001) Capital structures in

developing countries Journal of Finance 56 87ndash130Brennan M J and Schwartz E S (1984) Optimal financial policy and firm valuation

Journal of Finance 39 593ndash607Cameron A C and Trivedi P K (2010) Microeconometrics using Stata Stata Press TexasConstantinides G M (1979) Multiperiod consumption and investment behavior with

convex transactions costs Management Science 25 1127ndash1137Constantinides G M and Richard S F (1978) Existence of optimal simple policies for

discounted-cost inventory and cash management in continuous time Operations Research

26 620ndash636DeAngelo H DeAngelo L and Whited T M (2011) Capital structure dynamics and

transitory debt Journal of Financial Economics 99 235ndash261Dixit A K (1993) The Art of Smooth Pasting Routledge London and New YorkFama E and French K R (2002) Testing trade-off and pecking order predictions about

dividends and debt Review of Financial Studies 15 1ndash33Faulkender M and Petersen M A (2006) Does the source of capital affect capital struc-

ture Review of Financial Studies 19 45ndash79Faulkender M Flannery M J Hankins K W and Smith J M (2012) Cash Flows and

Leverage Adjustments Journal of Financial Economics 103 632ndash646Fischer E Heinkel R and Zechner J (1989) Dynamic capital structure choice theory

and tests Journal of Finance 44 19ndash40Frank M Z and Goyal V K (2008) Trade-off and pecking order theories of

debt in B E Eckbo (ed) Handbook of Corporate Finance Empirical Corporate Finance

Vol 2 Handbook of Finance Series Chapter 12 ElsevierNorth-Holland Amsterdam pp

135ndash202

Frank M Z and Goyal V K (2009) Capital structure decisions which factors are reliably

important Financial Management 38 1ndash37

Garvey G T and Hanka G (1999) Capital structure and corporate control the effect of

antitakeover statutes on firm leverage Journal of Finance 54 519ndash546

Goldstein R Ju N and Leland H (2001) An EBIT-based model of dynamic capital

structure Journal of Business 74 483ndash512

Graham J R and Tucker A L (2006) Tax shelters and corporate debt policy Journal of

Financial Economics 81 563ndash594

Hackbarth D and Mauer D C (2012) Optimal priority structure capital structure and

investment Review of Financial Studies 25 747ndash796

Halling M Yu J and Zechner J (2012) Leverage dynamics over the business cycle

Working paper University of Utah University of New South Wales and Vienna

University of Economics and Business AdministrationHennessy C A and Whited T A (2005) Debt dynamics Journal of Finance 60 1129ndash1165

Huang R and Ritter J R (2009) Testing theories of capital structure and estimating the

speed of adjustment Journal of Financial and Qualitative Analysis 44 237ndash271

1452 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

Hovakimian A Opler T and Titman S (2001) The debt-equity choice Journal of

Financial and Quantitative Analysis 36 1ndash24Jung K Kim Y and Stulz R M (1996) Timing investment opportunities managerial

discretion and the security issue decisions Journal of Financial Economics 42 159ndash185Kane A Marcus A J and McDonald R L (1984) How big is the tax advantage to

debt Journal of Finance 39 841ndash853Kayhan A and Titman S (2007) Firmsrsquo histories and their capital structures Journal of

Financial Economics 83 1ndash32Leary M T (2009) Bank loan supply lender choice and corporate capital structure

Journal of Finance 64 1143ndash1185Leary M T and Roberts M R (2005) Do firms rebalance their capital structures

Journal of Finance 60 2575ndash2619Leland H E (1994) Corporate debt value bond covenants and optimal capital structure

Journal of finance 49 1213ndash1252Lemmon M L Roberts M R and Zender J F (2008) Back to the beginning persist-

ence and the cross-section of corporate capital structure Journal of Finance 63 1575ndash1608Long M S and Malitz I B (1985) The investment-financing nexus some empirical

evidence Midland Corporate Finance Journal 3 53ndash59Mackie-Mason J K (1990) Do taxes affect corporate financing decisions Journal of

Finance 45 1471ndash1493Morellec E Nikolov B and Schurhoff N (2012) Corporate governance and capital

structure dynamics Journal of Finance 67 803ndash848Myers S C (1984) The capital structure puzzle Journal of Finance 39 575ndash592Myers S C (1993) Still searching for optimal capital structure Journal of Applied

Corporate Finance 61 4ndash14Rajan R and Zingales L (1995) What do we know about capital structure some evidence

from international data Journal of Finance 50 1421ndash1460Strebulaev I A (2007) Do tests of capital structure theory mean what they say Journal of

Finance 62 1747ndash1787

Strebulaev I A and Whited T M (2012) Dynamic Models and Structural Estimation in

Corporate Finance Now Publisher Boston and Delft

Titman S and Wessels R (1988) The determinants of capital structure choice Journal of

Finance 43 1ndash21

Welch I (2004) Capital structure and stock returns Journal of Political Economy 112

106ndash131

Welch I (2011) Two common problems in capital structure research the financial-debt-

to-asset ratio and issuing activity versus leverage changes International Review of Finance

11 1ndash17

PROFITSndashLEVERAGE PUZZLEREVISITED 1453

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

correlation between profitability and financial leverage Within an industrythe most profitable firms borrow less the least profitable borrow morerdquo(Myers 1993 p 6) This puzzling relationship between corporate profitsand leverage is widely regarded as a particularly serious defect of thetrade-off theory (see eg Fama and French (2002))2

In this article we revisit the profitsndashleverage relationship and show thatthe usual interpretation of the evidence is wrong What really happens is thatfirms only actively adjust leverage occasionally and the leverage adjustmentsthat follow profitability changes are typically not big enough to fully offsetthe effect of the profitability shocks This result does not arise in a one-period trade-off model nor in the usual dynamic trade-off models thathave fixed adjustment costs but no variable adjustment costs Howeverthis result is easy to understand in a trade-off model with both fixed andvariable costs of adjustment Since both kinds of costs are realistic theprofitsndashleverage puzzle does not provide telling evidence against the trade-off theoryIn a standard static trade-off theory model free of any transactions costs

the firm chooses to be at the optimal leverage target By construction thereare no dynamic effects in a one-period model On the other hand in a con-tinuous time trade-off models without fixed costs as in Kane Marcus andMcDonald (1984) and Brennan and Schwartz (1984) firms actively releveralmost continuously to maintain very high leverage to avoid taxes while juststaying out of bankruptcy Since the almost continuous issuance andrepurchasing of debt is not observed in practice the literature switched toconsideration of fixed adjustment costs For simplicity the variable adjust-ment costs were generally dropped to focus on the fact that fixed costsinduced periods without rebalancing Fischer Heinkel and Zechner (1989)found that due to the fixed costs much of the time the firm passively acceptsleverage shocks However if leverage becomes too high (or too low) the firmrebalances leverage to a target level The idea that firms are passive much ofthe time and only occasionally rebalance to a single leverage target is also inGoldstein Ju and Leland (2001) Strebulaev (2007) and Morellec Nikolovand Schurhoff (2012)Leary and Roberts (2005) explicitly recognize that both fixed and variable

transactions costs matter for leverage adjustments The implications of

2 A partial list of papers documenting an inverse relation between leverage and profitability

include Auerbach (1985) Graham and Tucker (2006) Long and Malitz (1985) Titman andWessels (1988) Fischer Heinkel and Zechner (1989) Rajan and Zingales (1995) andBooth et al (2001) Frank and Goyal (2009) show that the inverse leveragendashprofitabilityrelation has become weaker in the recent decades

1416 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

having both types of costs for the leverage puzzle can be readily understoodDue to the fixed costs small shocks are ignored since it will not pay to undotheir effect As in the pure fixed cost models there is a zone of inactionIf leverage becomes too high (or too low) it will pay for the firm to takeaction How strongly will the firm react If there are no variable costs thenaction from either boundary will be to the same target value But withvariable costs the firm must consider how much rebalancing is worthwhileThe result will be partial moves away from the boundary to the point atwhich the cost and the benefit of further moves are just balanced As a resultthere are two rebalancing points or ldquotargetsrdquomdashan upper target and a lowertarget3

From the perspective of a static model or a single target model the firmunder-adjusts to the profitability shock in the sense that the impact of theshock is only partly undone But this is not suboptimal behavior It is arecognition by the firm that fully undoing the shock would be more costlythan it is worth Variable costs count To go all the way to the staticoptimum would be excessively costly and so firms only go part wayEmpirically we find that the predicted ldquounder-adjustmentrdquo is observedThe nonlinearity of debt and equity contracts may also be important formagnitudes Because debt is concave it will tend to be less responsive toprofitability shocks In contrast equity is convex which makes it more re-sponsive to profitability shocksUnder-adjustment due to variable costs of adjusting is important when

interpreting empirical evidence of cross-sectional variation in leverage ratiosLet D denote corporate debt and E corporate equity then leverage isL frac14 D

DthornE Consider a positive shock to profits which causes E to increaseHowever because of under-adjustment D increases by less In other wordsthe denominator increases by more than the numerator in the leverage ratioand so the ratio falls Next consider a negative shock to profit In this case Ewill drop and due to under-adjustment D drops by less Leverage increasesThus a negative relationship will be observed between profits and leveragedespite the fact that firms are actively rebalancing in the direction predictedby the trade-off theory

3 In models with both fixed and variable costs of adjusting there are two kinds of first-order conditions known as ldquovalue matchingrdquo and ldquosmooth pastingrdquo The key analytical

ideas have been understood since Constantinides and Richard (1978) and are well explainedby Dixit (1993) This result from Constantinides and Richard (1978) is depicted by Learyand Roberts (2005) in Figure 1 panel C where the conditions are explicitly interpreted asleverage

PROFITSndashLEVERAGE PUZZLEREVISITED 1417

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

The main empirical findings are as given below

(1) When a firm makes extra profits the book value of equity increasesunless the firm takes some sort of offsetting action Similarly when a

firm makes extra profits the market value of the firmrsquos equity in-creases unless there is some sort of offsetting action Thus moreprofitable firms automatically have more book equity and more

market equity unless the firm takes offsetting actions

(2) Among large firms the highest profit firms increase their debt themost Those with high profits experience large increases in both thebook and the market value of equity The highest profit firms tend torepurchase equity while the lowest profit firms tend to issue moreequity

(3) Among small firms profit seems to have only a very minor effect ondebt Those with high profits experience some increases in both thebook and the market value of equity Those with low profits experi-ence negative effects on market equity There is a tendency to issueequity with the lowest profit firms issuing the most equity

(4) When firms adjust leverage the magnitude of the adjustment isnot sufficient to fully undo the impact of the underlying shocksFirms do not return to a unique static optimum They seem toldquounderadjustrdquo

There is a huge prior literature on our topic For a review of the literaturesee Frank and Goyal (2008) The inverse relationship between profitabilityand leverage ratios has generated a variety of responses from scholars A

common response articulated by Strebulaev and Whited (2012) is to arguethat the trade-off theory predictions could be more complex in a dynamicmodel Hennessy and Whited (2005) for example specify a dynamic model

of investment and financing under uncertainty and show that a trade-offmodel can explain a number of empirically observed stylized facts includingthe negative leveragendashprofitability relation In their model highly profitablefirms find it unattractive to issue debt because additional debt in such firms

finances a distribution to shareholders rather than replace equityOther scholars focus on adjustment costs and how a proper consideration

of these costs could reconcile the trade-off theory with empirical evidenceMyers (1984) discusses how adjustment costs induce lags in adjusting to the

optimum which could explain the observed wide variation in leverage ratiosFischer Heinkel and Zechner (1989) provide an explicit model of the ideaFama and French (2002) speculate that the negative relation betweenleverage and profitability is perhaps picking-up transitory variation in

1418 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

leverage rather than target variation in leverage A more formal analysis oftransitory debt is provided by DeAngelo DeAngelo and Whited (2011)Leary and Roberts (2005) consider both fixed and variable adjustment

costs and show that firms really do rebalance leverage Firms do notignore leverage ratios when making financing decisions and they do notjust passively accept the impact of equity market shocks Leary andRoberts (2005) do not study the implications of the form of the adjustmentcosts for the profitsndashleverage puzzle which is our focus Their focus is on thespeed of adjustment toward the target and on what that implies about thepersistence of leverageStrebulaev (2007) calibrates a dynamic trade-off model in which all costs

are proportional to the firm value A large firm is a rescaled version of asmall firm and so the target leverage ratio is independent of firm size Thisstructure is not designed to consider the differences between the financingactions of large versus small firms that we observe in the data Because he isstudying a trade-off model he finds that more profitable firms issue moredebt In simulated data Strebulaev (2007) shows that increases in profitabil-ity are correlated with future profitability which in turn increases the valueof the firm and lowers leverage (see p 1772) But expected future profit-ability increases do not always lead to current refinancing His cost structuredoes not lead to the under-adjustment phenomenon of the sort that wedocumentSeveral papers attribute the negative relation between leverage and prof-

itability to firms passively accumulating profits (see eg Fama and French(2002) and Kayhan and Titman (2007)) However the existing evidence isoften indirect and provides mix findings for this idea The implication that atthe time of rebalancing leverage should be positively related to profitabilityis also examined in several earlier studies Mackie-Mason (1990) shows thatcompanies with tax loss carryforwards are more likely to issue equity Incontrast Jung Kim and Stulz (1996) report finding no relation between thelikelihood of equity issuance and profitability Hovakimian Opler andTitman (2001) show that security issuance decisions of firms are afunction of the deviation from target leverage and other regressors includingfirm profitability They show that debt issuances are positively related toprofitability The present article while supporting aspects of their findingsexamines both issuing decisions and magnitudes as a function of firm sizeand profits In doing so we provide a more granular understanding of theeffect of profitability on changes in equity that derive from real operationsand changes in equity that are the result of active issuances and repurchasesof securities Furthermore we provide a sharply different interpretationalong with evidence in support of our interpretation We stress the

PROFITSndashLEVERAGE PUZZLEREVISITED 1419

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

distinction between automatic changes in equity that stem from operationsas opposed to actual issues or repurchases of equity This distinction is em-pirically importantThis article is organized as follows Section 2 provides theoretical motiv-

ation for our empirical strategy Section 3 describes the construction of thedata and variables and provides summary statistics Section 4 providesresults from the fixed effects estimates of leverage ratios Section 5presents the frequency and magnitudes of financing activity for sorts onfirm size and profits Section 6 provides the main results on debt andequity issuance regressions Section 7 explores the debt and equityissuance responses of firms during good and bad times Section 8examines the debt and equity issuances when scaled by total issuances andtotal capital Section 9 concludes the article

2 Empirical Strategy

The key idea in the trade-off theory is that firms choose capital structure bytrading off the various costs and benefits The considerations may includebankruptcy costs tax benefits agency costs transactions costs etc Differentversions of the theory include slightly different elements From our perspec-tive the leverage adjustment costs are of particular importanceA decision to issue debt or to redeem it or to issue shares or to repurchase

them requires managerial thought and consideration There are fixed organ-izational costs There is also the possibility of managerial inertia as docu-mented in Bertrand and Mullainathan (2003) These are fixed costs that areapproximately independent of the magnitude of the adjustments Adjustingalso has marginal costs Depending on which action is taken there may befees for underwriters bankers lawyers etc These may have a fixed compo-nent too Commonly they have a variable cost component which dependson the size of the transaction So both fixed and variable costs of adjustingleverage are normally involved The exact magnitudes vary from situation tosituation What is worse these are likely to have time-varying components4

Why does the form of the transactions costs matter Because the existenceof such fees implies that there will not be a unique leverage target This is anapplication of an idea that goes back to Constantinides and Richard (1978)and Constantinides (1979) It should be stressed that there is nothing newhere in terms of the formal structure This is just a reinterpretation of

4 If there were no time-varying component then it would be easy to identify the exact upperand lower targets for each firm In each case the firm would go exactly to the appropriatetarget In reality with time variation in adjustment costs such precision is not possible

1420 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

standard dynamic optimization conditions For a reader who wants a moreexplicit treatment of the mathematics we recommend Dixit (1993) The ob-servation that this mathematics can be interpreted in terms of corporateleverage in which firm size matters is also not novel It can be found inLeary and Roberts (2005) who provide a simulation with both fixed costsand a weakly convex adjustment costs What is new here is the analysis of theimplication for the impact of profits on leverageSuppose that without active rebalancing leverage would just wander

about due to changes in the stock market value of equity According tothe trade-off theory excessively high leverage is costly (bankruptcy andagency costs) as is excessively low leverage (lost tax benefits and free cashflow problems) If there are no transactions costs then under conventionaltrade-off theory assumptions there will be a unique leverage target that willbe maintained at all timesDue to the fixed costs such frequent rebalancing would be too costly As a

result the firm permits the leverage to drift However if the drift is tooextreme in either direction then the firm will actively rebalance in the direc-tion of the static optimum If the rebalanced firm (inclusive of rebalancingcosts) is worth more then rebalancing is worth it If the rebalanced firmwould be worth less then rebalancing is not worth it Rebalancing takesplace at the point at which the values just matchWhen the firm is rebalancing it also needs to worry about how far to

move in the desired direction When each step has costs the firm needs toensure that the marginal cost of the last step is just equal to the marginalbenefit As a result both from excessively high leverage and from excessivelylow leverage the movements are only part wayTo fix ideas consider an off-the-shelf model of the leverage ratio The

focus is on the impact of the adjustment cost structure The formal structureis essentially the cash management problem studied by Constantinides andRichard (1978) and Dixit (1993) We follow the presentation and notation inDixit (1993) very closely to show that this is just an application of well-established ideas to the leverage problemLet xt denote the leverage ratio at date t is the interest rate Ku is a fixed

cost of actively increasing leverage Kd is the fixed cost of actively reducingleverage mu is the marginal cost of each unit of active leverage increase md isthe marginal cost of each unit of active leverage decrease f(x) is the flowpayoff if leverage is at x F(x) is the expected present value of the payoff ifleverage is at x x is the lower adjustment threshold x is the lower targetleverage (when adjusting from below) x is the upper adjustment thresholdand x is the upper target leverage (when adjusting from above) Accordingto the trade-off theory F 0 gt 0 F 00 lt 0

PROFITSndashLEVERAGE PUZZLEREVISITED 1421

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

In the absence of active adjustments the leverage ratio xt follows astandard process

dx frac14 ethx tTHORNdtthorn ethx tTHORNdw

where is the trend coefficient is the volatility coefficient and w is anormal shock process with zero mean and unit variance If x lt x lt x thenfinancial inaction is optimal If x drops to x then it is reset to x and if xrises to x then it is reset to xThe trade-off theory is frequently thought of as a class of theories in which

firm value first increases in leverage and ultimately drops if leverage gets toohigh This can be motivated by taxes and bankruptcy or by various agencyconcerns A range of ideas can be allowed for when we write that the flowpayoff (dividends andor profits) is fethxtTHORN and it depends on the leverage statext

5 If the firm starts at x0frac14 x then F(x) is the expected present value of theflow payoffs The discount rate is which allows us to write

FethxTHORN frac14 E

Z 10

fethxtTHORNetdtjx0 frac14 x

To maximize the firm value it must choose a leverage policy that sayswhen to rebalance and how much to rebalance These are given by thefamous value matching and smooth pasting conditionsThe value matching conditions are as follows

FethxTHORN FethxTHORN frac14 Ku thornmuethx xTHORN

FethxTHORN FethxTHORN frac14 Kd thornmdethx xTHORN

The smooth pasting conditions are as follows

F 0ethxTHORN frac14 F 0ethxTHORN frac14 mu

F 0ethxTHORN frac14 F 0ethxTHORN frac14 md

As explained by Constantinides and Richard (1978) with standard as-sumptions about the shape of F there are a pair of lines at heights mu andmd such that they intersect the function F 0ethxTHORN in such a way as to define thepoints x lt x lt x lt xTo sum up in this trade-off model with both fixed and variable

rebalancing costs there are four critical leverage points x gt x gt x gt x

5 For simplicity we suppose that the flow payoff is a continuous function of leverageThere is no discrete exit or bankruptcy point This allows us to simply interpret theConstantinides and Richard model much as in Leary and Roberts (2005)

1422 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

If leverage is between x (upper boundary) and x (lower boundary) the firmdoes nothing active Changes in leverage reflect various shocks to revenuesand costs that hit the firm If leverage drifts up to x (or above) then the firmrebalances to x If leverage drifts down to x (or below) the firm rebalancesto xWhat does this imply empirically A firm that does nothing can be

anywhere within the interval between x and x A firm that is rebalancingis a firm that has just hit one of the outer boundaries How does profitabilityaffect the likelihood of hitting a boundary High profits increase the value ofequity This increases the denominator of the leverage ratio and thus makesit more likely that the firm hits the lower boundary Low profits reduce thevalue of equity and thus make it more likely that the firm hits the upperboundaryTo be more specific suppose that the costs are symmetric and the leverage

changes are also following an ordinary Brownian motion Then the median

inert firm will have leverage ofxthornx

2 A firm that just actively reduced leverage

will be at x gtxthornx

2 A firm that just actively increased leverage will be at

x ltxthornx

2 A firm that just reduced leverage will generally be a firm that had

very high leverage To the extent that this reflects a lack of activerefinancing such firm will have low value of equity The low value ofequity will generally reflect low profits A firm that just increased leveragewill generally be a firm that had very low leverage In the same way a firmthat just increased its leverage will tend to have had high value of equity dueto high operating profits6

In this setting it is natural to focus on firms that are actively adjusting andto compare them to each other and to firms that are inactive It is alsohelpful to look directly at financing decisions and profitability rather thanfocusing on a leverage ratioOur approach shares the perspective of Welch (2004) Leary and Roberts

(2005) and Strebulaev (2007) that it is very important to focus on activeadjustments by firms Welch (2004) transforms the data to put everythingback into a ratio form His ldquoimplied debt ratiosrdquo still have both debt andequity in the denominator Similarly Leary and Roberts (2005) andStrebulaev (2007) pay a great deal of attention to leverage ratios whenstudying financing decisions Strebulaev (2007) does not have fixed costssince he makes use of a rescaling property Empirically we find differentadjusting behavior by large and by small firms So the rescaling property is

6 If we knew the values of the critical leverage points then it would be possible to derivemore refined tests Unfortunately those critical values are not observable

PROFITSndashLEVERAGE PUZZLEREVISITED 1423

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

an issue of concern His model does not account for the negative relationshipbetween current leverage and past market-to-book ratioWe depart from the earlier literature by arguing that the use of leverage

ratios can be particularly misleading for profitability since the use of aleverage ratio causes identification problems It is hard to be sure whethera leverage change reflects changes in the numerator or the denominator Thisis particularly problematic when as here the rebalancing is partialIt should be noted that much empirical capital structure literature debates

about whether to use a book leverage ratio or a market leverage ratio Bookleverage is used based on the argument that managers borrow against assetsin place which are better reflected by book assets Others such as Welch(2004) argue that book values are essentially an accounting fiction and thatonly market leverage should count Empirical papers regularly attempt tosidestep this debate by carrying out tests on both book leverage and marketleverage Our point is that both kinds of leverage ratios are potentially mis-leading when examining the impact of profits on capital structure

3 Data

We use conventional data sources starting with the merged Compustat-CRSP data The data are annual and are converted into constant year2000 dollars using the GDP deflator We exclude financial companies (SIC6000-6999) firms involved in major mergers (Compustat footnote code AB)firms reporting financial data in currencies other than the US dollars andfirms with missing data on our key variables7 The ratio variables aretrimmed at the 1 level in both tails of the distribution This serves toremove outliers and the most extremely misrecorded data The finalsample consists of 179021 firm-year observations from 1971 to 2009Table I provides definitions of financial variables and reports summary

statistics The average debt (in constant US$) is about $653 million while themedian is $24 million A significant fraction of firms have zero debt (the 10thpercentile is 0) Book equity is slightly larger than book debt Market equityis almost three times larger than book debt Book assets average $2191million although the medians are considerably smallerIf issuing or retiring securities incurs no fixed costs then we would expect

to see many small actions and very few large actions (Leary and Roberts2005) If there were significant fixed costs involved in issuing or retiringoutstanding securities then small issues might not be worthwhile Table I

7 These include debt book value of equity market value of equity assets book andmarket leverage profitability market-to-book assets ratio and tangibility

1424 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

shows that although most firms issue little debt or equity in a given year theaverages are large In other words when firms actually enter debt and equitymarkets they intervene massivelyThe mean constant dollar debt issue is $165 million (the median is $2

million) In unreported tables we find that the average debt issue is about81 of assets (the median is 18) About 38 of the firms issue no debt8 issue between 0 and 1 of the value of their assets as debt another

Table I Data description

Data sources The sample comes from the annual Compustat files The sample period is

1971ndash2009 We exclude financial companies (SIC 6000-6999) firms involved in majormergers (Compustat footnote code AB) firms reporting financial data in currencies otherthan the US dollar and firms with missing data on the key variables All financial variables

are deflated to year 2000 using the GDP deflator N is the number of observations SD isthe standard deviation D BVE MVE Assets Debt issuance Equity issuance Debt re-payment Equity repayment are all measured in $ millions The ratio variables are trimmed

at the 1 level in both tails of the distribution

Variable definitions Debt (D)frac14 long-term debt (dltt)thorn short-term debt (dlc) book equity(BVE)frac14 common shareholder equity (ceq) market equity (MVE)frac14 no of outstanding

shares (csho) closing price ( prcc f ) assetsfrac14 book assets (at) debt issuancefrac14 issuance oflong-term debt (Maxethdltis 0THORN)thorn increase in current debt (Maxethdlcch 0THORN) equity issu-ancefrac14 sale of common stock (Maxethsstk 0THORN) debt repaymentfrac14 reduction of long-termdebt (Maxethdltr 0THORN)thorn decrease in current debt (Minethdlcch 0THORN) equity repurchasefrac14 purchase

of common stock (Maxethprstkc 0THORN) book leveragefrac14D(DthornBVE) market leveragefrac14D(DthornMVE) profitabilityfrac14EBITDA (oibdp)assets market-to-book ratio M

B frac14Marketvalue of assets (MVA)assets where MVAfrac14DthornMVEthorn preferred-liq value (pstkl)

deferred taxes (txditc) tangibilityfrac14 net property plant and equipment (ppent)assets

Variable N Mean SD

Distribution

10th 50th 90th

Debt (D) 179021 653 4592 0 24 1048

Book equity (BVE) 179021 815 4122 3 70 1373

Market equity (MVE) 179021 1784 9589 9 118 2698

Assets 179021 2191 11530 10 153 3575

Debt issuance 179021 165 1335 0 2 245

Equity issuance 179021 26 186 0 0 40

Debt repayment 179021 132 1123 0 3 181

Equity repurchase 179021 24 257 0 0 9

Book leverage 179021 036 034 000 031 072

Market leverage 179021 027 025 000 020 066

Profitability 179021 005 027 018 011 023MB 179021 165 206 052 102 324

Tangibility 179021 031 024 005 026 070

PROFITSndashLEVERAGE PUZZLEREVISITED 1425

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

16 issue between 1 and 5 of the value of their assets as debt and theremaining 38 issue debt in excess of 5 of the value of their assetsThe mean constant dollar equity issue is about $26 million (the median is

about $04 million) As a fraction of assets the mean and median equityissues are about 7 and 02 respectively About 33 of the firms issue noequity 34 of the firms issue between 0 and 1 of the value of theirassets as equity another 14 issue between 1 and 5 of the value of theirassets as equity and the remaining 19 of the firms issue equity that is inexcess of 5 of the value of their assets Average debt repayments are largerthan equity repurchases This perhaps reflects the finite maturity of debt andits contractual repayment The median firm does not repurchase equityWe construct both book and market leverage ratios Book leverage is

defined as debt over debt plus book equity Market leverage is defined asdebt over debt plus market equity8 The median book leverage is 036 (theaverage is 031) The median market leverage is 027 (the average is 020)Profitability is defined as the ratio of operating income before depreci-

ation to assets While the average firm is profitable (the ratio of EBITDA toassets is 005) the median firm is even more profitable (with a profitabilityratio of 011) But the sample also includes a large number of unprofitablefirms as the 10th percentile is 018 The table also reports descriptive stat-istics on the market-to-book ratio and the tangibility ratio The market-to-book ratio (MB) defined as the ratio of the market value of assets to bookassets averages at about 165 Tangibility defined as the ratio of netproperty plant and equipment to assets averages at about 31

4 Leverage Ratio Regressions

The previous literature focuses on estimates obtained using leverage ratiosHence we start with a similar estimation to check whether our results matchthose of previous studies Table II presents the results for book leverage Weobtain similar results for market leverage and they are presented inAppendix Table AI The regressions include leverage factors following theprevious capital structure literature (see eg Lemmon Roberts and Zender(2008) and Frank and Goyal (2009)) The factors include (i) profitability (ii)industry median leverage (iii) market-to-book assets ratio (iv) tangibility ofassets and (v) firm size Industry median leverage is constructed as themedian leverage of all other firms in the industry excluding the firm under

8 Welch (2011) stresses the idea that nonfinancial liabilities should not be implicitly mis-treated as if they were equity by paying excessively narrow attention to financial liabilitiesin a leverage ratio We have adopted his approach in the empirical work reported here

1426 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

consideration Firm size is defined as the natural log of assets Other factorsare defined in Section 3 We cluster standard errors by firm and estimatethese regressions both with and without fixed effectsIn Table II we run conventional cross-section leverage regressions Since

we employ the usual data it is not surprising that our results match thosereported in the existing capital structure literature Profitability has anegative sign in both the book leverage regressions and in the marketleverage regressions The coefficients on other factors largely match thosereported in earlier studies Firms operating in industries with high leverage

Table II Leverage and profitability

The table presents parameter estimates for book leverage regressions The sample consists

of nonfinancial companies in the annual Compustat files during the period 1971ndash2009 Theleverage is estimated as the ratio of debt over debt plus book equity The explanatoryvariables Profitabilityt1

MB

t1

Tangibilityt1 and LnethAssetsTHORNt1 are described in

Table I IndLevt1 is estimated as the median book leverage of all other firms in thesame industry The industry is defined at the level of the four-digit SIC code All specifi-cations include the year fixed effects The specifications in column (2) additionally include

the firm fixed effects We report t-statistics where the standard errors are clustered at thefirm level in parentheses

aSignificant at the 1 level

Quantile regressions

(1) (2) (3) (4) (5)

Profitabilityt1 026a 020a 011a 017a 041a

(293) (192) (378) (518) (1015)

IndLevt1 042a 012a 030a 052a 052a

(369) (107) (926) (1252) (950)MB

t1

002a 001a 001a 002a 002a

(203) (102) (392) (426) (418)

Tangibilityt1 018a 025a 024a 019a 012a

(206) (154) (964) (580) (276)

LnethAssetsTHORNt1 001a 002a 002a 002a 001a

(114) (76) (899) (478) (96)

Constant 012a 010a 012a 002a 029a

(156) (60) (284) (27) (387)

Year FE Yes Yes Yes Yes Yes

Firm FE No Yes No No No

Regression type OLS FE 25th tile Median 75thtile

Clustered SE Yes Yes No No No

R2 013 054

Pseudo R2 016 015 009

N 158824 158824 158824 158824 158824

PROFITSndashLEVERAGE PUZZLEREVISITED 1427

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

tend to have high leverage A higher market-to-book ratio is associated withlower leverage Larger firms are typically more highly levered The coeffi-cients on tangibility are positive The conclusion from Table II is clear Theleverage regression results match those reported in previous studiesAt times one may be concerned that in a regression what is being

estimated is a conditional mean while some parts of the distributionmight behave differently from other parts If so then ldquoaccidentsrdquo likesample selection issues or asymmetries in the underlying distributionsmight play a greater role than is usually recognized In other words thesource of identification could be misinterpreted To guard against this po-tential problem we estimate the same model using quantile regressions9

Columns 3 4 and 5 report what happens at the 25th 50th and 75th per-centiles The basic model results are robust in terms of both the sign and thestatistical significanceThe impact of profits does appear to be somewhat stronger among the

high leverage firms For instance in a book leverage regression the coeffi-cient on profits for the first quartile is 011 with a t-ratio of 378 For thethird quartile the coefficient on profits is 041 with a t-ratio of 1015This difference is curious and might deserve further study However the factthat in each case we get the negative sign and statistical significance is suf-ficient for the purposes of the present article So the results on profits arerather robust across the distributionTable III sorts firms into those that are actively adjusting their leverage

and those that are passive In the presence of both fixed and variable trans-acts costs this distinction is important We expect the coefficient on profit-ability in leverage ratio regressions to be more negative in firms that arepassive and less negative in firms that are actively adjustingWe define active firms as those that are either issuing net debt or net

equity in excess of 5 of their assets Passive firms are those that neitherissue debt nor equity in nontrivial amounts (ie greater than 5 of theirassets) Consistent with our expectations we find in Table III that the coef-ficients on profits are significantly more negative for passive firms than theyare for active firms With firm and year fixed effects the coefficient onprofits is 008 for active firms and 017 for the passive firms The twocoefficients are significantly different from each other (plt 001) InTable III the standard control variables are used They have the usualsigns and for the most part do not differ much between the active and thepassive firms

9 Cameron and Trivedi (2010) provide an extensive discussion of quantile regressions inStata

1428 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

5 Financing Activity

51 FREQUENCY OF DEBT AND EQUITY ISSUES

Table IV reports the percentage of firms issuing or repurchasing debt orequity for annual sorts based on lagged profitability In Panel A weemploy the conventional 5 cutoff rule to exclude minor fluctuations anddefine firms ldquoissuing debtrdquo as those that issue debt (both short-term andlong-term) in excess of 5 of the value of their assets Other decisions aresimilarly defined using a 5 cutoff Appendix Table AII examines the dis-tributional properties across profitability (and size) quintiles for ldquolargerdquo is-suances by increasing the 5 cutoff to higher levels such as 10 and 15How does security issuance behavior vary with firm profitability To

answer this question Table IV sorts firms by profitability and reports the

Table III Leverage and profitability for active and passive firms

The table presents estimates of leverage on firm characteristics for the subsamples of firms

that are active in either debt or equity markets Active firms are defined as those that issueeither net debt in excess of 5 of the value of assets or issue net equity in excess of 5 ofthe value of their assets The passive firms do not issue significant (more than 5) of either

debt or equity The sample comes from the annual Compustat files during the period 1971ndash2009 Financial firms are excluded The variables are described in Tables I and II Wereport t-statistics where the standard errors are clustered at the firm level in parentheses

aSignificant at the 1 level

Active Passive

(1) (2) (3) (4)

Profitabilityt1 006a 008a 022a 017a

(126) (148) (254) (188)

IndLevt1 044a 019a 041a 013a

(424) (149) (326) (127)MB

t1

002a 001a 003a 001a

(362) (196) (327) (172)

Tangibilityt1 010a 014a 014a 019a

(146) (118) (145) (130)

LnethAssetsTHORNt1 001a 004a 001a 004a

(121) (179) (147) (169)

Constant 012a 008a 003a 007a

(120) (57) (45) (42)

Year FE Yes Yes Yes Yes

Firm FE No Yes No Yes

R2 034 064 033 075

N 43395 43395 58943 58943

PROFITSndashLEVERAGE PUZZLEREVISITED 1429

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

percentage of firms issuing or repurchasing debt or equity The table alsoreports test statistics from tests that examine whether the percentage differ-ences between the high- and low-profit firms are significantly different fromeach other The results in Column (1) show that the likelihood of issuing debt

Table IV Profitability sorts for debt and equity issuers

The sample contains nonfinancial firms listed on the annual Compustat files for the period

from 1971 to 2009 The table presents the percentage of firms issuing and retiring (orrepurchasing) debt and equity The firms are classified as ldquoissuing debtrdquo if they issuedebt in excess of 5 of the value of their assets as ldquoissuing equityrdquo if they issue equity

in excess of 5 of the value of their assets as ldquoretiring debtrdquo if they retire debt in excess of5 of the value of their assets and as ldquorepurchasing equityrdquo if they repurchase equity inexcess of 5 of the value of their assets In addition we also report net debt issuers which

are firms that issue net debt over 5 of the value of their assets and net equity issuerswhich are firms that issue net equity in excess of 5 of the value of their assets Weannually sort firms on lagged profitability and report the percentage of firms in each ofthese categories The bottom part of each panel reports the percentage of firms issuing or

retiring securities by profitability within the smallest and largest asset quintiles The tablealso reports results from t-test for the difference in percentage of firms issuing or retiringsecurities in the lowest and the highest profit quintile

Percentage of firms

Iss Ret Iss Iss Rep Iss Both IssD IssE

D D NetD E E NetE DampE RepE RetD

(1) (2) (3) (4) (5) (6) (7) (8) (9)

Allfirm

s

Low 276 245 186 301 13 297 71 01 38

2 319 299 181 117 21 112 27 02 20

3 385 344 194 86 29 80 21 05 15

4 386 330 213 81 43 75 20 07 14

High 329 273 198 92 82 81 17 13 14

t Highfrac14Low 150 84 39 723 414 763 358 176 200

Smallfirm

s Low 275 204 201 411 11 405 109 01 47

2 291 233 213 379 10 378 99 01 45

3 281 246 204 265 14 261 66 01 41

4 275 272 186 162 15 159 34 02 35

High 275 281 178 136 27 131 25 02 26

t Highfrac14Low 01 84 28 316 55 318 171 18 55

Largefirm

s

Low 414 363 177 87 22 79 28 03 11

2 449 379 180 62 30 57 17 07 09

3 439 360 186 56 41 48 16 09 07

4 403 320 194 46 67 39 13 12 06

High 374 264 207 50 129 35 10 28 05

t Highfrac14Low 50 130 45 89 244 116 77 120 47

1430 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

is mostly independent of firm profitability In the aggregate there is a smalldifference but it is sensitive to sorting on firm size Similarly in Column (2)there is only a weak relation between the profitability and the likelihood thatthe firm retires debt To a first approximation the fraction of firms issuing orrepurchasing debt is largely independent of profits Large firms are somewhatmore likely than small firms to issue debt and also to retire debtUnlike debt issues the probability of issuing equity is strongly related to

profitability as can be seen in Column (4) Low-profit firms are much morelikely to issue equity than are high-profit firms In the lowest profit quintile301 of the firms issued equity In the highest profit quintile it was only92 The lower panel reports separate sorts for small and large firms Similarpatterns are found for both groups of firms that is the likelihood of equityissuance declines as profitability increases The differences between the quin-tiles are quite stark for the group of small firms All quintiles of small firmsissue equity much more frequently than the corresponding large firm quintilesHigh-profit firms are much more likely to repurchase equity as shown in

Column (5) In the overall sample only 13 of low-profit firmsrepurchased equity compared with 82 of the high-profit firms Whenwe separate the small and large firms again a difference emerges Ingeneral small firms do not repurchase much equity while a greaterfraction of the large firms do soGiven these facts it is natural that low profitability firms are much more

likely to be net issuers of equity than are high profitability firms as shown inColumn (6) Among the firms in the lowest quintiles of profits roughly 30of firms issue net equity that exceeds 5 percent of their assets In contrastfor the most profitable quintile only about 8 issue net equityFirm size is an important variable in the recent literaturemdashit is often used

as a proxy for access to capital markets (as in Faulkender and Petersen(2006) and Leary (2009)) Small firms are bank-dependent risky and infor-mationally opaque They have restricted access to public debt markets andconsequently face more severe supply constraints in their ability to issuedebt Thus we expect small firms to be more sluggish in adjusting theirdebt and equity in response to shocks to profitability Large firms incontrast have much easier access to public debt markets and they facefewer obstacles in accessing securities marketsTo examine how size and profitability interact the bottom part of Table IV

examines profitability sorts for the smallest and largest firms We first sortfirms annually by firm size and then within each size quintile we sort them onprofitability We do these two-way sorts to ensure that we have similarnumber of firms in profitability subgroups for both small and large firmsamples Unconditionally small firms are generally less profitable while

PROFITSndashLEVERAGE PUZZLEREVISITED 1431

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

large firms are relatively more profitable The two-way sorts allow us to havea uniform distribution of firms in various profitability quintiles for both smalland large firm samples The bottom part of Table IV reports results for thesmallest and the largest size quintiles Among the small firms there is littlerelation between profitability and the likelihood of issuing debt Howeverlarger firms exhibit a small increase in the likelihood of net issuance of debtwith increasing profitabilityIn contrast to debt the effects of profitability on equity issuance and

repurchases are much more consistent across size quintiles As profitabilityincreases firms are generally less likely to issue equity and more likely torepurchase it Across the two extreme size groups we note a striking differ-ence between the proportion of firms issuing equity and those issuing debtAmong low-profit small firms almost 41 are equity issuers whereas onlyabout 9 of low-profit large firms are equity issuers However regardless ofsize we note a monotonic reduction in the likelihood of issuing equity asprofitability increasesColumn (7) reports the fraction of firms that issue both debt and equity in

excess of 5 of assets both as a function of profits in the upper panel andas a function of firm size and profits in the lower panel In the upper panelwe see that low-profit firms are much more likely to issue both debt andequity (71) while the high-profit firms are less likely to do so (17) Thelower panel shows that small firms are much more likely to be issuers of bothin contrast to large firms In both firm size categories the profit pattern is thesame with low profits being more frequent issuers of both simultaneouslyImportantly Column (8) shows that the likelihood of issuing debt and

simultaneously repurchasing equity increases with profitability Converselyin Column (9) we find that the likelihood of doing the reverse that is issuingequity and retiring debt declines with profitability The effects of sorting onfirm size and firm profitability mirror those reported for all firms Overallthe results show that low-profit firms are less likely to issue debt and repur-chase equity they are instead more likely to issue equity and retire debtFirms with high profitability exhibit the reverse patternIn Appendix Table AII we present results from profitability (and size)

quintiles for large issuances defined using the 10 cutoff (in Panel A) andthe 15 cutoff (in Panel B) The results from these panels are largely con-sistent with those described above

52 MAGNITUDES OF ISSUANCES AND EQUITY CHANGES

In the previous section we considered the probability of having a nontriviallevel of debt or equity activity The next question is how large are the dollar

1432 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

values involved In Table V we sort firms according to profits and thentabulate the levels and changes in both debt and equity We do this firstfor all firms and then for small versus large firmsFor the analysis on all firms we observe in Column (1) that debt peaks at

the middle of the distribution This is because firms with medium profitabil-ity are also the largest as seen in Column (8) which reports average assetvalues for different profitability quintiles When we sort first by firm size andthen examine firms with different levels of profitability within size quintileswe find that for small firms debt is roughly independent of profits while forlarge firms less profitable firms have a higher level of debtAs expected Columns (3) and (5) show that more profitable firms have

higher equity values Columns (2) and (7) consider the relationship betweenissuances of debt and equity and firm profitability Debt issuances are sig-nificantly larger for more profitable firms In contrast equity issuances arelarger for low-profit firms Small low-profit firms issue substantially lessequity compared with small high-profit firms Among large firms the low-profit ones tend to issue equity while the high-profit ones tend to repurchaseequity The fact that more profitable firms issue debt and repurchase equitywhile the least profitable firms retire debt and issue equity is consistent withthe predicted relation between profitability and financing decisions under thetrade-off theoryWe also examine two-way sorts by size and profitability and report results

for the smallest and largest quintile of firms For the smallest quintile offirms the change in debt is largely unrelated to profits But for large firmsthere is a positive relation between profits and debt issuances High-profitfirms have a big positive change in debt Low-profit large firms have anegative change in debtColumns (4) and (6) provide an explanation of why the leverage ratio

regression results contradict those from the basic profitability sorts pre-sented here As we can see profitability indirectly affects leverage ratiosby increasing equity values Changes in both the book value of equity andthe market value of equity are positive and large for highly profitable firmsIn contrast these changes are negative for less profitable firms This evidenceillustrates an important issue concerning the use of leverage ratios Suchratios are often interpreted as essentially reflecting the use of debt by thefirm This interpretation while common is empirically misleadingFor the typical firm the change in the value of equity is larger than the

change in debt For example in the third profit quintile for large firmscolumn (7) the mean equity issue is just $9 million but the change in themarket value of equity is $242 million At the same time the mean change indebt is $27 million This suggests that a fair bit of the observed variation in

PROFITSndashLEVERAGE PUZZLEREVISITED 1433

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

the leverage ratios is primarily driven by the changes in the market value ofequity in the denominator rather than by the changes in debt in the numer-ator Since equity issues are often small this implies that the variation in theleverage ratio is primarily driven by internal operations rather than byexternal financing actions This again points to the fact that leverageratios can provide a misleading account of actual patterns in the dataThere is always a potential concern that the averages may be misleading

due to the impact of outliers To address this concern in AppendixTable AIII we present median values of the profitability sorts Theseresults generally reinforce the findings in Table VTable VI again considers the magnitudes of financing activity but this

time the issues are scaled by a measure of firm size That way it is easier to

Table V Magnitude of financing activity

The sample contains nonfinancial firms listed on the annual Compustat files for the period

from 1971 to 2009 The table presents average debt and equity changes in debt and equityand equity issuances (in millions of dollars) for firms sorted on profitability within sizeclasses The table reports information for all firms sorted on profitability and for profit-

ability sorts within the smallest and largest firms The sorts are done annually The tablealso reports results from t-test for the difference in percentage of firms issuing or retiringsecurities in the lowest and the highest profit quintile

D D BVE BVE MVE MVE EqyIss Assets

(1) (2) (3) (4) (5) (6) (7) (8)

Allfirm

s

Low 83 5 78 10 120 5 7 234

2 605 13 509 24 721 8 17 1691

3 1227 9 1123 22 1834 3 9 3551

4 833 23 1166 7 2408 74 9 2994

High 570 53 1171 71 3817 68 36 2581

t Highfrac14Low 31 9 36 9 38 2 16 36

Smallfirm

s Low 25 04 42 01 332 28 33 97

2 29 06 55 02 278 16 21 116

3 31 07 63 01 212 05 12 127

4 30 06 70 05 179 10 08 136

High 24 06 84 14 234 17 07 142

t Highfrac14Low 04 15 153 85 73 11 201 107

Largefirm

s Low 3929 55 2990 152 4671 11 63 10846

2 3826 19 3178 95 5122 66 11 10721

3 3074 27 3498 41 6173 242 9 9697

4 2623 62 3915 45 8473 94 57 9855

High 2180 193 4642 261 14924 202 148 10171

t Highfrac14Low 94 64 108 87 246 12 169 16

1434 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

see whether large firms behave disproportionately when compared withsmall firms Column (1) shows that low-profit firms have more debtas reported in many previous studies In Columns (4) and (6) we seethat low-profit firms have a major decline in both the book value ofequity and the market value of equity The high-profit firms have thereverse experienceMore interesting are the active decisions of the firms In Column (2) we see

that in contrast to the usual impression the low-profit firms are reducingdebt while the high-profit firms are increasing it In Column (7) we see thatthe low-profit firms are also reducing leverage by issuing equity The high-profit firms issue little equity In the lower panel similar results are foundwhen we separate small and large firms Equity issuing is particularly strong

Table VI Magnitude of financing activitymdashscaled levels and changes

The sample contains nonfinancial firms listed on the annual Compustat files for the period

from 1971 to 2009 The table presents average debt and equity changes in debt and equityand equity issuances scaled by assets for firms sorted on profitability within size classes Thetable reports information for all firms sorted on profitability and for profitability sorts

within the smallest and largest firms The sorts are done annually The table also reportsresults from t-test for the difference in percentage of firms issuing or retiring securities inthe lowest and the highest profit quintile

D D BVE BVE MVE MVE EqyIss

(1) (2) (3) (4) (5) (6) (7)

Allfirm

s

Low 386 122 12 871 3427 1868 217

2 285 58 407 119 940 157 28

3 293 12 425 14 742 03 12

4 269 26 461 02 955 27 10

High 192 15 551 47 1675 74 07

t Highfrac14Low 35 49 50 47 127 42 157

Smallfirm

s Low 545 154 841 2145 9688 9128 696

2 684 49 712 1915 5014 2318 288

3 331 253 141 969 3063 954 129

4 267 181 357 175 1673 221 55

High 194 05 543 26 1758 147 41

t Highfrac14Low 49 17 51 36 88 27 99

Largefirm

s Low 350 30 306 58 511 55 09

2 346 14 335 13 523 02 04

3 353 165 332 71 711 29 01

4 280 01 414 04 909 09 05

High 227 15 477 19 1643 27 14

t Highfrac14Low 371 112 453 181 570 57 221

PROFITSndashLEVERAGE PUZZLEREVISITED 1435

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

among the small low-profit firms that are also experiencing sharp reductionsin the value of equity In column (2) we see that both for the large and for thesmall firms net debt increases are primarily found in the highest quintile offirm profitability

6 Debt and Equity Regressions

So far we have demonstrated that in our data the conventional leverageregressions have the usual signs We then demonstrated that in simple sortsof the data strikingly different active financing patterns emerge The nexttask is to reconcile these differencesTable VII presents simple regressions predicting changes in debt book

equity and market equity as well as equity issuances Consistent with thesorts we control for changes in firm size and we also include firm and yeardummies There may be a concern about the speed of a firmrsquos responsesAccordingly we consider the lagged change in profits as the main case butwe also include the contemporaneous change in profits Whether the contem-poraneous change in profits is included or not does not in fact matter Wereport panel-robust standard errors adjusted for clustering at the firm levelThe results in Table VII support the findings in the sorts Firms that have

an increase in profits in 1 year increase debt both in the same year and in thenext year Firms that have an increase in profits experience an increase inboth the book value of equity and the market value of equity Importantlyhowever firms that experience and increase in profits reduce their equityissues All these effects are statistically significant (except for the marketvalue of equity in the column 5 specification) and seem intuitively reason-able In Appendix Table AIV similar regressions to those presented aboveare reported but without firm-fixed effects We find that empirically verylittle changes whether we include firm fixed effects or leave them out Thusthe targeting behavior to the extent that it happens does not account forwhat we are reportingThe next step is to bring the sorts and the simple regressions together into

a common setting We therefore include factors in addition to profitability(i) median industry leverage (ii) the market-to-book assets ratio (iii) tangi-bility of assets and (iv) firm size (measured by the log of assets) Rajan andZingales (1995) show that these factors are related to leverage in G7countries A number of studies have used these factors to estimateleverage targets Frank and Goyal (2009) show that these factors arerobustly related to leverage in the USA In these regressions we use indica-tors for the quintile that the firm is in for each factor

1436 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

Table VIII reports these results In Columns (1) and (2) we examine debtissues and again show that profits positively affect debt issuances The effectsare large and statistically significant at the 1 level Columns (3) and (4)which examine book equity show a strong effect of profits This effect is onlyslightly reduced by the inclusion of other factors Columns (5) and (6) illus-trate that the impact of profits on the change in market equity is also robustto the inclusion of the conventional factors Columns (7) and (8) examineequity issuances Here we again find that profits have a significant negativeimpact on equity issuances As before the results are robust to the inclusionof the conventional factorsIn unreported tables we estimate debt and equity changes using quantile

regressions In terms of the signs and significance of the profitability variable

Table VII Debt and equity changes

The sample contains nonfinancial firms listed on the annual Compustat files for the period

from 1971 to 2009 The table presents fixed effect estimates from regressions of changes indebt (D) changes in the book value of equity (BVE) changes in the market value ofequity (MVE) and net equity issuances (EqyIss) The explanatory variables include

current and lagged change in operating income before depreciation (Profits andProfitst1) and lagged change in the book value of assets (Assetst1) In addition tofixed firm effects the regressions include year indicator variables The reported t-statistics

are corrected for clustering at the firm level

aMean significant at the 1 level

bMean significant at the 5 level

cMean significant at the 10 level

D BVE MVE EqyIss

(1) (2) (3) (4) (5) (6) (7) (8)

Profitst1 022a 025a 015b 035a 004 060b 005b 007a

(33) (31) (25) (35) (03) (25) (25) (53)

Profits 050b 078a 140a 003a

(24) (32) (48) (34)

Assetst1 000 007b 025a 001b

(00) (21) (35) (25)

Constant 167 283a 34 51 1145a 1182a 605a 595a

(14) (30) (04) (05) (32) (29) (69) (74)

Year FE Yes Yes Yes Yes Yes Yes Yes Yes

Firm FE Yes Yes Yes Yes Yes Yes Yes Yes

R2 001 004 001 010 001 005 002 003

N 162056 162056 162130 162130 157550 157550 162154 162154

PROFITSndashLEVERAGE PUZZLEREVISITED 1437

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

the results are qualitatively similar to the OLS results reported in Table VIIIProfitability positively affects debt issuances and negatively affects equityissuances across various quantiles Overall the results in Table VIII showthat there is nothing that is special about using sorts or regressions to explaindebt and equity issues In either case we find that more profitable firms tendto increase their debt experience an increase in the value of equity andrepurchase shares Thus the control factors are not responsible for theusual rejection of the trade-off theory It would appear that what mattersis the size of the issuing activity relative to the organic increases in equityvalue due to profits In other words what matters in the leverage regressionseems to be coming from the presence of E in D

DthornE and not from D

Table VIII Debt and equity issuances profitability and leverage factor quintiles

The sample contains nonfinancial firms listed on the annual Compustat files for the period

from 1971 to 2009 The table presents estimates from regressions of change in debt (D)change in the book value of equity (BVE) change in the market value of equity (MVE)and net equity issuances (EquityIss) The explanatory variables include lagged profitability

quintiles and lagged leverage factor quintiles In addition to fixed firm effects the regres-sions include year indicator variables The reported t-statistics are corrected for clustering atthe firm level

aMean significant at the 1 level

bMean significant at the 5 level

cMean significant at the 10 level

D BVE MVE EqyIss

(1) (2) (3) (4) (5) (6) (7) (8)

ProfitabilityQint 214a 167a 283a 175a 80 429a 53a 55a

(70) (54) (66) (59) (08) (48) (55) (62)

IndLevQint 104a 43 99 20

(29) (15) (10) (10)

Ln(Assets)Qint52 17 1026a 53a

(16) (05) (77) (35)MB

Qint158a 372a 1155a 09

(45) (56) (106) (06)

TangibilityQint 05 177 436c 64a

(01) (15) (18) (31)

Constant 167 96 656b 235 4613a 8610a 113 570a

(08) (04) (26) (04) (61) (58) (11) (31)

Year FE Yes Yes Yes Yes Yes Yes Yes Yes

Firm FE Yes Yes Yes Yes Yes Yes Yes Yes

R2 0002 0002 0003 0004 0013 0014 0013 0014

N 155587 155037 155587 155037 155587 155037 155587 155037

1438 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

7 Are Financial Market Conditions Important

According to the trade-off theory capital structure is determined by con-sidering the impact of costs and benefits of debt The time variation inthe costs and benefits of debt imply that issuance decisions would alsovary over time Furthermore market conditions affect the transactioncosts of adjusting Adverse selection is a greater problem in a cold marketthan in a hot one Accordingly it may be easier to adjust in good marketconditions Related ideas have been developed by Huang and Ritter(2009) Faulkender et al (2012) and Halling Yu and Zechner (2012)Faulkender Flannery Hankins and Smithrsquos paper examines cross-sectionalvariation in adjustment speeds including market valuations andhow they may affect adjustment speeds It is natural to think that ingood market conditions it will be less costly to issue both debt and equityand adjusting leverage to respond to profitability shocks would be mucheasierTo test the importance of market conditions we require a definition of

good times and bad times Our empirical strategy is to estimate good timesversus bad times at the four-digit industry level We define an industry ashaving ldquogood timesrdquo if the median firm in that industry has a market-to-book ratio that is higher than the 67th percentile of the time-series distribu-tion of the industry median market-to-book ratios Conversely an industryis defined as having ldquobad timesrdquo if the median firm in the industry has amarket-to-book ratio that is lower than the 33rd percentile of the time-seriesdistribution of the industry median market-to-book ratiosPanel A of Table IX tabulates issuance activity for profitability sorts in

both good times and bad times As might be expected active debt and equityissuances are larger during good times There is also somewhat more activeswaps between debt and equity during good timesPanel B reports the issuance activity for the smallest and the largest firms

during the good times and bad times Within each size quintile firms aresorted on profitability This is a three-way sort As before we find that thereis much more active use of external markets during good times Small low-profit firms are more likely to issue equity in good times than in bad timesDebt issuances are significantly higher in good times Large high-profitfirms are significantly more likely to issue debt and repurchase equity ingood times than in bad timesPanel C reports the magnitudes of the financial variables rather than the

frequencies During bad times less profitable large firms retire substantialamounts of debt and they show a tendency to issue equity Small firms donot seem to engage in similar debt reduction activities However like the

PROFITSndashLEVERAGE PUZZLEREVISITED 1439

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

Table IX Debt and equity issues in good and bad times

The table reports the frequency and magnitude of financing activity for sorts on profits for subsamples

of firms in good and bad times An industry is defined as having ldquogood timesrdquo if the median firm in

that industry has a market-to-book ratio that is higher than the 67th percentile of the time-series

distribution of industry median market-to-book ratios Conversely an industry is defined as having

bad times if the median firm in that industry has a market-to-book ratio that is lower than the 33rd

percentile of the time-series distribution of industry median market-to-book ratios Panels A and B

report the percentage of firms (i) issuing net debt in excess of 5 of the value of their assets (ii) issuing

net equity in excess of 5 of the value of their assets (iii) issuing debt and repurchasing equity both in

excess of 5 of the value of their assets and (iv) issuing equity and retiring debt both in excess of 5

of the value of their assets Panel C tabulates the mean debt and equity levels and the mean changes in

debt and equity issuances The sample contains nonfinancial firms listed on the annual Compustat files

for the period from 1971 to 2009

Iss Iss Iss D Iss ENet D Net E Rep E Ret D(1) (2) (3) (4)

Panel A sorts on profitability

Badtimes

Low 145 131 01 202 170 49 02 083 173 46 04 074 182 47 03 10

High 177 51 06 10t Highfrac14Low 41 136 33 01

Goodtimes

Low 206 381 01 462 197 148 02 233 212 103 04 184 234 92 10 18

High 216 92 16 14t Highfrac14Low 23 658 138 23

Panel B sorts on firm size and profitability

Badtimes

Smallfirm

s Low 151 318 00 312 182 276 02 253 151 140 02 294 183 73 00 19

High 177 80 00 11t Highfrac14Low 10 90 NA 20

Largefirm

s Low 172 54 01 042 148 39 04 053 150 43 04 044 156 33 03 06

High 147 28 05 03t Highfrac14Low 17 32 20 02

Goodtimes Smallfirm

s Low 213 448 01 542 230 443 01 473 222 325 00 474 203 197 02 38

High 190 153 02 31t Highfrac14Low 19 230 10 39

Largefirm

s Low 203 96 03 122 208 66 09 133 232 54 16 064 215 39 19 05

High 234 36 38 04t Highfrac14Low 29 99 92 38

(continued)

1440 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

larger firms there is a tendency for small firms to issue more equitymdashat leastto the extent possibleDuring good times large profitable firms raise significant amounts of debt

as they experience an increase in the value of their equity Such firms alsoengage in share repurchases Debt issues by smaller firms are much lessaffected by their own profitability during good times and their change indebt is more or less independent of their profits However equity issuanceis primarily found among the low-profit small firms during good timesTable IX shows that empirically profitability has a time-varying impacton leverage choicesIn Table X the basic regressions for changes in debt and for equity

issuance are presented for both good times and bad times Comparinggood times with bad times we see that the effects of profits are muchstronger in good times In bad times the impacts are rather weak statisticallyThis difference is interesting and deserves further study However for ourpurposes recall that the static trade-off theory implies time-varying capital

Table IX Continued

D D BVE BVE MVE MVE EqyIss Assets

(1) (2) (3) (4) (5) (6) (7) (8)

Panel C magnitude of financing activity

Badtimes Smallfirm

s Low 40 03 58 14 305 06 22 140

2 44 05 82 07 240 67 17 180

3 45 04 94 02 176 18 05 187

4 43 09 98 02 137 04 03 188

High 36 10 114 13 209 16 04 199

Largefirm

s Low 3365 73 2523 169 2304 163 94 8972

2 3031 95 2983 35 3284 278 56 9068

3 2930 5 3360 14 3947 308 57 9225

4 2281 6 3817 19 5213 442 7 8902

High 1790 62 4767 13 8799 58 31 9573

Goodtimes

Smallfirm

s Low 24 06 38 05 339 23 39 90

2 26 08 48 01 301 04 25 104

3 30 09 60 04 246 03 17 125

4 28 07 66 08 206 18 12 130

High 23 06 85 16 261 13 08 141

Largefirm

s Low 4532 36 3477 103 6557 260 53 12465

2 4580 58 3527 139 6911 421 23 12522

3 2952 111 3485 2 7586 279 32 9535

4 2752 124 3939 72 10686 33 105 10226

High 2429 228 4819 377 18565 268 222 10830

PROFITSndashLEVERAGE PUZZLEREVISITED 1441

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

structure choices even when the target is time invariant Empirically we doobserve time-varying choices

8 Is Scaling Important

In the preceding section we provided results for unscaled debt and equityissues This seems appropriate to us However it involves making twochanges to the standard regression We would like to determine if both

Table X Debt and equity changes in good and bad timesmdashregressions

The sample contains nonfinancial firms listed on the annual Compustat files for the period

from 1971 to 2009 The table presents estimates from regressions of change in debt (D)and net equity issuances (EqyIss) in good and bad times The explanatory variables includechange in profits lagged change in profits and lagged change in assets The regressions

include both firm and year fixed effects An industry is defined as having ldquogood timesrdquo ifthe median firm in that industry has a market-to-book ratio that is higher than the 67thpercentile of the time-series distribution of industry median market-to-book ratios

Conversely an industry is defined as having bad times if the median firm in thatindustry has a market-to-book ratio that is lower than the 33rd percentile of the time-series distribution of industry median market-to-book ratios The reported t-statistics arecorrected for clustering at the firm level

aMean significant at the 1 level

bMean significant at the 5 level

cMean significant at the 10 level

Good times Bad times

D EqyIss D EqyIss

(1) (2) (3) (4) (5) (6) (7) (8)

Profits 087a 091a 006b 009a 060 065 003 003b

(30) (29) (20) (28) (15) (16) (16) (20)

Profitst1 029c 015a 034c 001

(19) (37) (17) (04)

Assetst1 000 002b 000 001

(02) (22) (00) (07)

Constant 1337 3989 8691a 10001a 1351 1851 4099a 4055a

(03) (09) (55) (37) (09) (13) (80) (78)

Year FE Yes Yes Yes Yes Yes Yes Yes Yes

Firm FE Yes Yes Yes Yes Yes Yes Yes Yes

R2 0071 0080 0018 0045 0178 0212 0034 0041

N 74827 68729 74891 68787 26965 26013 26971 26018

1442 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

changes are crucial Accordingly in this section we take a step back toconsider prior literature in which variables are scaled The questions arehow does scaling affect the results and whyTable XI examines alternative scalings Columns (1) and (2) scale debt and

equity issuances by the total firm net issuance Columns (3) and (4) scaledebt and equity issuances by the sum of debt and book equity Columns (5)and (6) consider the change in debt scaled by the sum of debt and bookequity Columns (7) and (8) consider the change in leverage due to externalfinancing Since we are examining changes we use the changes in thestandard regression factors as controls The primary interest is in the coef-ficient on the change in profitsWe find that a change in profits is associated with an increase in debt

issues But the sign reverses when we scale the debt issues by the sum of debtplus book equity This is telling What is really driving the results is thechange in equity in the denominator not the change in debt in the numer-ator This result is further substantiated in Columns (5) and (6) where thechange in debt to capital ratio is used instead of debt issuanceTo further explore the impact of active issuance relative to passive

experiencing of changes we decompose leverage by separating out theeffect of debt and equity issues from those due to changes in retainedearnings Thus we define leverage change due to external financing as

LevEF frac14Dt1 thorn d

Dt1 thorn Et1 thorn dthorn e

Dt1

Dt1 thorn Et1 eth1THORN

where D is debt E is book value of equity d is debt issues net of retirementsand e is equity issues net of repurchases Garvey and Hanka (1999) use asimilar measure in their study This measure directly captures the netleverage effect of debt and equity issuancesColumns (7) and (8) report results from regressions of the net leverage

effect of debt and equity issues on changes in profitability after controllingfor other leverage factors in differences and year indicator variables Theeffect of profitability on leverage change due to external financing is positiveand statistically significant This comparison suggests that changes in totaldebt and equity are only partly a result of debt and equity issuance decisionsOther balance sheet adjustments complicate the inferences from leverageratio regressionsFor most of our analysis we follow the literature in treating debt as homo-

genous It is of course potentially interesting to ask what happens when wedistinguish between short-term debt and long-term debt Which part of thefirmrsquos debt maturity structure drives the under-adjustment Table XII

PROFITSndashLEVERAGE PUZZLEREVISITED 1443

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

provides results from such decompositions Alternative scalings areconsidered The main results of interest are that the distinction betweenshort- and long-term debt is important Short-term debt issuance is largelyindependent of changes in profits The negative sign on profit change stemsfrom the action of long-term debt This seems intuitively reasonable

Table XI Scaled issuances

The sample contains nonfinancial firms listed on the annual Compustat files for the period

from 1971 to 2009 The table presents estimates from regressions of debt issuance to totalissuance (in columns (1) and (2)) debt issuance to total capitalization (in columns (3) and(4)) change in debt to capitalization ratio (in columns (5) and (6)) and change in leverage

ratio considering only external financing (in columns (7) and (8)) The change in leverageratio considering only external financing is defined as

LevEF frac14Dt1 thorn d

Dt1 thorn Et1 thorn dthorn e

Dt1

Dt1 thorn Et1

where D is debt E is book value of equity d is debt issues net of retirements and e is equity

issues net of repurchases The regressions include lagged leverage factors and year indicatorvariables Some of the specifications include firm fixed effects The reported t-statistics arecorrected for clustering at the firm level

aMean significant at the 1 level

bMean significant at the 5 level

cMean significant at the 10 level

DebtIssTotIss

DebtIssethDthornETHORN D

DthornE LevEF

(1) (2) (3) (4) (5) (6) (7) (8)

Profitt1 004a 000 002a 001c 003a 002 001b 002a

(35) (03) (34) (18) (35) (14) (23) (26)

IndLevt1 004c 001 001 001c 000 000 000 000

(19) (06) (15) (19) (02) (04) (03) (05)

Market=bookt1 001a 001a 000a 000a 000a 000a 000 000c

(47) (85) (68) (60) (55) (36) (05) (18)

Tangt1 004 005 013a 011a 011a 006a 003a 003a

(12) (15) (144) (112) (79) (38) (44) (42)

Assetst1 008a 004a 006a 004a 002a 001a 003a 002a

(97) (45) (283) (161) (53) (35) (147) (101)

Constant 079a 074a 002a 003a 000 000 000b 002a

(497) (442) (91) (111) (00) (05) (22) (93)

Year FE Yes Yes Yes Yes Yes Yes Yes Yes

Firm FE No Yes No Yes No Yes No Yes

R2 0015 0098 0022 0076 0009 0070 0012 0086

N 133905 133905 140552 140552 138229 138229 140572 140572

1444 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

inasmuch as short-term debt is probably more associated with day-to-dayoperations rather than longer term planning considerationsIn Table XIII we examine change in leverage ( D

DthornE) and change inleverage due to external financing (LEVEF) during good and bad timesseparately The traditional negative effect of a change in profits on achange in leverage is found both in good times and in bad timesHowever when we decompose the impact in order to isolate the externalcomponent as before the result disappears In good times the externalfinancing component of a change in profits is weak but still positiveDuring bad times it is not statistically significant This implies that distin-guishing between good and bad times does not fundamentally alter the basic

Table XII Short- and long-term debt issuances

The sample contains nonfinancial firms listed on the annual Compustat files for the period

from 1971 to 2009 The table presents estimates from regressions of short-term debtissuance to total issuance (in columns (1) and (2)) long-term debt issuance to totalissuance (in columns (3) and (4)) short-term debt issuance to total capitalization (in

columns (5) and (6)) and long-term debt to total capitalization (in columns (7) and (8))Some of the specifications include firm fixed effects The reported t-statistics are correctedfor clustering at the firm level

aMean significant at the 1 level

bMean significant at the 5 level

cMean significant at the 10 level

STDebtIssTotIss

LTDebtIssTotIss

STDebtIssethDthornETHORN

LTDebtIssethDthornETHORN

(1) (2) (3) (4) (5) (6) (7) (8)

Profitabilityt1 0002 0003 0042a 0008 0001 0001 0019a 0012a

(02) (04) (30) (05) (14) (10) (45) (28)

IndLevt1 0009 0007 0070b 0033 0001 0002c 0014a 0005

(06) (05) (26) (11) (08) (17) (26) (09)

M=Bt1 0003a 0003a 0003b 0010a 0000 0000 0003a 0002a

(48) (35) (22) (65) (10) (01) (84) (42)

Tangibilityt1 0020 0007 0011 0067 0005a 0006a 0097a 0081a

(12) (03) (03) (16) (34) (33) (108) (89)

Assetst1 0039a 0011b 0034a 0024b 0001a 0001 0046a 0020a

(89) (22) (38) (23) (36) (13) (190) (82)

Constant 0001b 0153a 0798a 0591a 0000c 0006a 0057a 0096a

(25) (144) (470) (288) (17) (124) (292) (234)

Year FE Yes Yes Yes Yes Yes Yes Yes Yes

Firm FE No Yes No Yes No Yes No Yes

R2 0025 0089 0018 0073 0030 0147 0016 0298

N 136146 136146 136140 136140 143031 143031 142847 142847

PROFITSndashLEVERAGE PUZZLEREVISITED 1445

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

Table XIII Externally financed leverage changes in good and bad times

The sample contains nonfinancial firms listed on the annual Compustat files for the period

from 1971 to 2009 The table presents estimates from regressions of change in debt tocapitalization ratio (in columns (1) (2) (5) and (6)) and change in leverage ratio consider-ing only external financing (in columns (3) (4) (7) and (8)) The change in leverage ratio

considering only external financing is defined as

LevEF frac14Dt1 thorn d

Dt1 thorn Et1 thorn dthorn e

Dt1

Dt1 thorn Et1

where D is debt E is book value of equity d is debt issues net of retirements and e is equityissues net of repurchases The regressions include lagged leverage factors and year indicator

variables The reported t-statistics are corrected for clustering at the firm level An industryis defined as having ldquogood timesrdquo if the median firm in that industry has a market-to-bookratio that is higher than the 67th percentile of the time-series distribution of industry

median market-to-book ratios Conversely an industry is defined as having ldquobad timesrdquoif the median firm in that industry has a market-to-book ratio that is lower than the 33rdpercentile of the time-series distribution of industry median market-to-book ratios

aMean significant at the 1 level

bMean significant at the 5 level

cMean significant at the 10 level

Good times Bad times

DDthornE LevEF D

DthornE LevEF

(1) (2) (3) (4) (5) (6) (7) (8)

Profitabiltyt1 0022c 0005 0016c 0025b 0073b 0041 0007 0007

(17) (03) (18) (24) (25) (11) (04) (03)

IndLevt1 0000 0011 0001 0009 0010 0018 0001 0000

(00) (09) (01) (13) (07) (10) (02) (00)

MB

t1

0004a 0003b 0000 0001 0008a 0001 0000 0000

(40) (24) (05) (08) (28) (04) (03) (03)

Tangt1 0097a 0041 0033a 0038b 0127a 0048 0059a 0033

(45) (16) (27) (25) (49) (14) (43) (16)

Assetst1 0014a 0013b 0024a 0019a 0037a 0023c 0036a 0025a

(30) (20) (84) (48) (45) (17) (88) (43)

Constant 0030b 0036 0000 0003 0028c 0014 0021 0001

(20) (16) (00) (07) (17) (06) (11) (01)

Year FE Yes Yes Yes Yes Yes Yes Yes Yes

Firm FE No Yes No Yes No Yes No Yes

R2 0006 0088 0009 0117 0026 0251 0023 0202

N 56820 56820 58200 58200 23462 23462 23511 23511

1446 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

conclusions The traditional leverage result on profits is driven by increasesin equity that are experienced and then partially but not completely offsetby firmsTo the extent that firms are engaging in active financing they are following

the prescriptions of the trade-off theory Thus we have a coherent picture ofhow firms are responding to profits Basically more profitable firms experiencean increase in the value of their equity They respond to this by issuing debtand repurchasing equity But the magnitude of the increase in equity is largerthan the magnitudes of the active steps taken in the debt and equity marketsThis means that the firmsrsquo actual actions are as predicted by the trade-offtheory But it leaves open the question of why they do not act more stronglyThe capital structure literature has long been interested in dynamics and

capital structure adjustment paths In recent years this interest has increas-ingly resulted in explicit dynamic optimizing models such as those ofGoldstein Ju and Leland (2001) Hennessy and Whited (2005) andStrebulaev (2007) These models commonly make specialized assumptionsleading some to question their generalityWe argue that both fixed transactions costs and variable transactions costs

are plausible and realistic in models that pin down firm size The simultaneouspresence of both is important Due to fixed costs firms only rebalance inter-mittently Due to variable adjustment costs the extent of the adjustment isreduced to the point at which an extra unit to adjustment is balanced againstthe extra unit transactions costs This means that both leverage increases andleverage reductions will only go part way toward a static optimumThe key point is that with both fixed and variable transactions costs we

should expect to only see intermittent leverage adjustments When leverage isadjusted it should ldquounder adjustrdquo when compared to a frictionless modelFurther discussion of transactions costs in a capital structure setting isprovided by Leary and Roberts (2005)

9 Conclusion

The connection between corporate profits and capital structure has beenvery influential in the assessment of the trade-off theory The conventionalinterpretation of the evidence has pushed the literature away from the trade-off theory and toward much more complex dynamic models and ideas As aresult it is important to make sure that the evidence has been correctlyinterpreted We show that the problem is with the interpretation of thetest results and a failure to take into account the costs associated with re-sponding to a profit shock

PROFITSndashLEVERAGE PUZZLEREVISITED 1447

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

We document that more profitable firms really do borrow more (not less)They tend to repurchase equity They experience an increase in both thebook value of equity and the market value of equity Less profitable firmsreally do tend to reduce their debt and to issue equityFirm size and market conditions also matter Larger firms tend to be more

active in the debt markets while smaller firms tend to be relatively moreactive in the equity markets During good times there is more use ofexternal financingThe usual profitsndashleverage puzzle result is primarily driven by the increase

in equity that is experienced by the more profitable firms Accordingly thepuzzle should be restated as asking why do firms not take sufficiently largeoffsetting actions to fully undo the change in equity What limits the mag-nitudes of the typical leverage response to profit shocksIn a frictionless model the partial response appears puzzling But there is

good empirical reason to believe that rebalancing entails both fixed andvariable costs and that firm size matters The rebalancing costs can befully avoided by doing nothing Accordingly the firm must decide whetherany given shock is big enough to be worth responding to If it is then thefirm must decide how big a response is called for These technical conditionsare known as ldquovalue matchingrdquo and ldquosmooth pastingrdquo Thus optimizationimplies that some shocks will be ignored Even if the shock is not ignoredthe optimal response will only partially undo the shock The magnitude ofthe leverage response must balance the marginal cost and the marginalbenefit of an extra unit of leverage Since the marginal cost of adjustingleverage is strictly positive the adjustment toward that static leverageoptimum will only go part way This is true both for leverage increasesand for leverage reductionsOne might conjecture that the variable costs would be time varying and

that such costs might have declined over time Since we do not have a fullaccounting of all of the variable costs it is of course impossible to be sureHowever there is some evidence suggesting that at least the directunderwriting fees are remarkably stable over time10

Surely nobody ought to be surprised that more profitable firms are moreinclined to repurchase equity while unprofitable firms tend not to do so Thefact that more profitable firms typically experience an increase in the value of

10 For IPOs Jay Ritter provides important summary evidence as httpbearwarringtonufleduritterIPOs2013Underwritingpdf He finds that for medium-sized deals a remark-

ably large fraction of the deals have a 7 fee year after year Whether this is due tobehavioral factors or some sort of a coordination device is not clear The extent towhich similar intertemporal stability carries over to other forms of external financemerits further study

1448 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

equity is equally natural Were it not for the contrary claims in the literature(ie the profitsndashleverage puzzle) we would argue that nobody ought to findit surprising that more profitable firms are more inclined to issue debt Thefact that leverage adjustments typically only partially offset profit shocksshould also sound quite natural These facts all fit together easily within thetrade-off theory once we take account of both fixed and variable costs ofactively adjusting capital structure

Appendix

Table AI Market leverage regressions

The table presents estimates of the regressions of market leverage on firm characteristics

The sample comes from the annual Compustat files during the period 1971ndash2009 Financialfirms are excluded The leverage is estimated as the ratio of debt over debt plus marketequity The explanatory variables Profitabilityt1

MB

t1

Tangibilityt1 and LnethAssetsTHORNt1are described in Table 1 IndMedianLevt1 is estimated as the median market leverage of allother firms in the same industry The industry is defined at the level of the four-digit SICcode All specifications include the year fixed effects The specifications in column (2)

additionally include the firm fixed effects We report t-statistics where the standard errorsare clustered at the firm level in parentheses

aSignificant at the 1 level

Quantile regressions

(1) (2) (3) (4) (5)

Profitabilityt1 013a 013a 007a 011a 020a

(275) (262) (349) (425) (538)

IndLevt1 046a 020a 032a 058a 063a

(506) (239) (1299) (1630) (1042)MB

t1

003a 001a 001a 002a 004a

(509) (303) (743) (709) (534)

Tangibilityt1 013a 018a 013a 014a 010a

(191) (181) (762) (571) (232)

LnethAssetsTHORNt1 001a 005a 001a 001a lt001a

(112) (258) (682) (342) (29)

Constant 010a 011a 005a 005a 029a

(171) (98) (173) (118) (398)

Year FE Yes Yes Yes Yes Yes

Firm FE No Yes No No No

Regression type OLS FE 25th tile Median 75thtile

Clustered SE Yes Yes No No No

R2 030 067

Pseudo R2 016 023 0201

N 158578 158578 158578 158578 158578

PROFITSndashLEVERAGE PUZZLEREVISITED 1449

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

Table AII Profitability sorts for firms issuing large amounts

Percentage of firms

Iss Ret Iss Iss Rep Iss Iss IssD IssE

D D NetD E E NetE DampE RepE RetD

(1) (2) (3) (4) (5) (6) (7) (8) (9)

Panel A 10 cutoff to define issuing or retiring activity

Allfirm

s

Low 205 150 134 260 05 257 45 00 23

2 207 175 103 77 06 75 11 01 10

3 243 196 101 50 08 47 08 01 09

4 252 186 113 50 13 48 08 02 08

High 227 161 118 60 27 57 07 04 09

t Highfrac14Low 70 38 62 755 226 768 320 90 152

Smallfirm

s Low 230 132 162 380 05 375 82 00 32

2 228 143 165 334 04 333 65 00 26

3 211 146 147 216 03 214 38 00 23

4 190 154 120 126 04 124 15 01 20

High 194 162 114 99 09 97 12 01 15

t Highfrac14Low 42 40 68 342 20 340 169 07 131

Largefirm

s Low 231 200 82 31 06 29 08 01 04

2 231 187 67 17 06 14 03 01 02

3 234 177 77 17 09 14 03 02 03

4 218 158 75 16 17 14 02 03 02

High 218 135 99 19 41 15 03 07 02

t Highfrac14Low 19 106 36 46 135 57 39 58 23

Percentage of firms

Iss Ret Iss Iss Rep Iss Both Iss D Iss E

D D Net D E E Net E DampE Rep E Ret D

(1) (2) (3) (4) (5) (6) (7) (8) (9)

Panel B 15 cutoff to define issuing or retiring activity

Allfirm

s

Low 158 101 100 230 01 228 27 00 15

2 142 115 63 57 01 56 05 00 06

3 165 128 57 32 01 31 03 00 05

4 173 118 65 35 01 34 03 00 05

High 160 102 70 46 04 44 03 00 05

t Highfrac14Low 08 06 141 738 87 746 277 24 131

Smallfirm

s Low 195 95 132 354 01 350 56 00 24

2 185 96 131 303 01 301 44 00 18

3 160 96 104 179 00 176 18 00 15

4 134 99 78 99 01 98 08 00 12

High 133 95 73 82 01 80 06 00 10

t Highfrac14Low 80 00 95 345 12 344 149 12 52

Largefirm

s Low 145 123 42 14 01 12 03 00 01

2 137 114 32 07 01 06 01 00 01

3 142 104 35 06 01 05 00 00 01

4 130 91 36 08 03 07 01 01 01

High 134 81 48 09 06 07 01 01 01

t Highfrac14Low 19 85 17 26 54 33 34 17 11

1450 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

Table AIV Debt and equity changesmdashwithout fixed effects

The sample contains nonfinancial firms listed on the annual Compustat files for the period from 1971

to 2009 The table presents estimates from regressions of change in debt (D) change in the book

value of equity (BVE) change in the market value of equity (MVE) and net equity issuances

(EqyIss) The explanatory variables include current and lagged change in operating income before

depreciation (Profits and Profitst1) and lagged change in book value of assets (Assetst1) The

regressions include year indicator variables The reported t-statistics are corrected for clustering at the

firm level

aMean significant at the 1 level

bMean significant at the 5 level

cMean significant at the 10 level

D BVE MVE EqyIss

(1) (2) (3) (4) (5) (6) (7) (8)

Profitst1 025a 024a 023a 038a 007 066a 007a 009a

(40) (32) (36) (39) (06) (29) (42) (51)Profits 051b 082a 149a 004a

(25) (33) (49) (33)Assetst1 002 005c 024a 001b

(08) (17) (34) (23)Constant 93 176a 45 71 1552a 1704a 130a 140a

(12) (28) (07) (09) (49) (49) (45) (45)Year FE Yes Yes Yes Yes Yes Yes Yes YesR2 001 005 001 011 001 005 002 002N 162056 162056 162130 162130 157550 157550 162154 162154

Table AIII Median profitability sorts

The sample contains nonfinancial firms listed on the annual Compustat files for the period from 1971

to 2009 The table presents median debt and equity median changes in debt and equity and median

equity issuances (EqyIss) in millions of dollars for all firms sorted on profitability and for profitability

sorts within the smallest and largest firms The sorts are done annually

D D BVE BVE MVE MVE EqyIss Assets(1) (2) (3) (4) (5) (6) (7) (8)

Allfirm

s

Low 21 00 75 17 256 20 01 1922 186 02 449 18 598 19 00 11033 816 06 1274 03 1637 05 00 32244 787 02 1541 18 2488 03 00 3556High 273 00 1361 66 3300 14 00 2578

Smallfirm

s Low 05 00 13 07 140 14 06 402 06 00 25 08 114 12 01 563 09 00 32 05 85 07 00 674 12 00 39 01 68 03 00 79High 10 00 51 04 98 01 00 90

Largefirm

s Low 11485 381 10269 371 13227 202 03 333902 13051 393 12786 140 17571 169 05 392463 11554 287 13429 24 21623 305 00 370314 9963 209 15052 194 29424 308 00 36705High 7361 00 16380 613 49142 439 00 34976

PROFITSndashLEVERAGE PUZZLEREVISITED 1451

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

References

Auerbach A J (1985) Real determinants of corporate leverage in B M Friedman (ed)

Corporate Capital Structures in the United States University of Chicago Press Chicago

IllinoisBertrand M and Mullainathan S (2003) Enjoying the quiet life corporate governance and

managerial preferences Journal of Political Economy 111 1043ndash1075Booth L Varouj A Demirguc-Kunt A and Maksimovic V (2001) Capital structures in

developing countries Journal of Finance 56 87ndash130Brennan M J and Schwartz E S (1984) Optimal financial policy and firm valuation

Journal of Finance 39 593ndash607Cameron A C and Trivedi P K (2010) Microeconometrics using Stata Stata Press TexasConstantinides G M (1979) Multiperiod consumption and investment behavior with

convex transactions costs Management Science 25 1127ndash1137Constantinides G M and Richard S F (1978) Existence of optimal simple policies for

discounted-cost inventory and cash management in continuous time Operations Research

26 620ndash636DeAngelo H DeAngelo L and Whited T M (2011) Capital structure dynamics and

transitory debt Journal of Financial Economics 99 235ndash261Dixit A K (1993) The Art of Smooth Pasting Routledge London and New YorkFama E and French K R (2002) Testing trade-off and pecking order predictions about

dividends and debt Review of Financial Studies 15 1ndash33Faulkender M and Petersen M A (2006) Does the source of capital affect capital struc-

ture Review of Financial Studies 19 45ndash79Faulkender M Flannery M J Hankins K W and Smith J M (2012) Cash Flows and

Leverage Adjustments Journal of Financial Economics 103 632ndash646Fischer E Heinkel R and Zechner J (1989) Dynamic capital structure choice theory

and tests Journal of Finance 44 19ndash40Frank M Z and Goyal V K (2008) Trade-off and pecking order theories of

debt in B E Eckbo (ed) Handbook of Corporate Finance Empirical Corporate Finance

Vol 2 Handbook of Finance Series Chapter 12 ElsevierNorth-Holland Amsterdam pp

135ndash202

Frank M Z and Goyal V K (2009) Capital structure decisions which factors are reliably

important Financial Management 38 1ndash37

Garvey G T and Hanka G (1999) Capital structure and corporate control the effect of

antitakeover statutes on firm leverage Journal of Finance 54 519ndash546

Goldstein R Ju N and Leland H (2001) An EBIT-based model of dynamic capital

structure Journal of Business 74 483ndash512

Graham J R and Tucker A L (2006) Tax shelters and corporate debt policy Journal of

Financial Economics 81 563ndash594

Hackbarth D and Mauer D C (2012) Optimal priority structure capital structure and

investment Review of Financial Studies 25 747ndash796

Halling M Yu J and Zechner J (2012) Leverage dynamics over the business cycle

Working paper University of Utah University of New South Wales and Vienna

University of Economics and Business AdministrationHennessy C A and Whited T A (2005) Debt dynamics Journal of Finance 60 1129ndash1165

Huang R and Ritter J R (2009) Testing theories of capital structure and estimating the

speed of adjustment Journal of Financial and Qualitative Analysis 44 237ndash271

1452 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

Hovakimian A Opler T and Titman S (2001) The debt-equity choice Journal of

Financial and Quantitative Analysis 36 1ndash24Jung K Kim Y and Stulz R M (1996) Timing investment opportunities managerial

discretion and the security issue decisions Journal of Financial Economics 42 159ndash185Kane A Marcus A J and McDonald R L (1984) How big is the tax advantage to

debt Journal of Finance 39 841ndash853Kayhan A and Titman S (2007) Firmsrsquo histories and their capital structures Journal of

Financial Economics 83 1ndash32Leary M T (2009) Bank loan supply lender choice and corporate capital structure

Journal of Finance 64 1143ndash1185Leary M T and Roberts M R (2005) Do firms rebalance their capital structures

Journal of Finance 60 2575ndash2619Leland H E (1994) Corporate debt value bond covenants and optimal capital structure

Journal of finance 49 1213ndash1252Lemmon M L Roberts M R and Zender J F (2008) Back to the beginning persist-

ence and the cross-section of corporate capital structure Journal of Finance 63 1575ndash1608Long M S and Malitz I B (1985) The investment-financing nexus some empirical

evidence Midland Corporate Finance Journal 3 53ndash59Mackie-Mason J K (1990) Do taxes affect corporate financing decisions Journal of

Finance 45 1471ndash1493Morellec E Nikolov B and Schurhoff N (2012) Corporate governance and capital

structure dynamics Journal of Finance 67 803ndash848Myers S C (1984) The capital structure puzzle Journal of Finance 39 575ndash592Myers S C (1993) Still searching for optimal capital structure Journal of Applied

Corporate Finance 61 4ndash14Rajan R and Zingales L (1995) What do we know about capital structure some evidence

from international data Journal of Finance 50 1421ndash1460Strebulaev I A (2007) Do tests of capital structure theory mean what they say Journal of

Finance 62 1747ndash1787

Strebulaev I A and Whited T M (2012) Dynamic Models and Structural Estimation in

Corporate Finance Now Publisher Boston and Delft

Titman S and Wessels R (1988) The determinants of capital structure choice Journal of

Finance 43 1ndash21

Welch I (2004) Capital structure and stock returns Journal of Political Economy 112

106ndash131

Welch I (2011) Two common problems in capital structure research the financial-debt-

to-asset ratio and issuing activity versus leverage changes International Review of Finance

11 1ndash17

PROFITSndashLEVERAGE PUZZLEREVISITED 1453

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

having both types of costs for the leverage puzzle can be readily understoodDue to the fixed costs small shocks are ignored since it will not pay to undotheir effect As in the pure fixed cost models there is a zone of inactionIf leverage becomes too high (or too low) it will pay for the firm to takeaction How strongly will the firm react If there are no variable costs thenaction from either boundary will be to the same target value But withvariable costs the firm must consider how much rebalancing is worthwhileThe result will be partial moves away from the boundary to the point atwhich the cost and the benefit of further moves are just balanced As a resultthere are two rebalancing points or ldquotargetsrdquomdashan upper target and a lowertarget3

From the perspective of a static model or a single target model the firmunder-adjusts to the profitability shock in the sense that the impact of theshock is only partly undone But this is not suboptimal behavior It is arecognition by the firm that fully undoing the shock would be more costlythan it is worth Variable costs count To go all the way to the staticoptimum would be excessively costly and so firms only go part wayEmpirically we find that the predicted ldquounder-adjustmentrdquo is observedThe nonlinearity of debt and equity contracts may also be important formagnitudes Because debt is concave it will tend to be less responsive toprofitability shocks In contrast equity is convex which makes it more re-sponsive to profitability shocksUnder-adjustment due to variable costs of adjusting is important when

interpreting empirical evidence of cross-sectional variation in leverage ratiosLet D denote corporate debt and E corporate equity then leverage isL frac14 D

DthornE Consider a positive shock to profits which causes E to increaseHowever because of under-adjustment D increases by less In other wordsthe denominator increases by more than the numerator in the leverage ratioand so the ratio falls Next consider a negative shock to profit In this case Ewill drop and due to under-adjustment D drops by less Leverage increasesThus a negative relationship will be observed between profits and leveragedespite the fact that firms are actively rebalancing in the direction predictedby the trade-off theory

3 In models with both fixed and variable costs of adjusting there are two kinds of first-order conditions known as ldquovalue matchingrdquo and ldquosmooth pastingrdquo The key analytical

ideas have been understood since Constantinides and Richard (1978) and are well explainedby Dixit (1993) This result from Constantinides and Richard (1978) is depicted by Learyand Roberts (2005) in Figure 1 panel C where the conditions are explicitly interpreted asleverage

PROFITSndashLEVERAGE PUZZLEREVISITED 1417

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

The main empirical findings are as given below

(1) When a firm makes extra profits the book value of equity increasesunless the firm takes some sort of offsetting action Similarly when a

firm makes extra profits the market value of the firmrsquos equity in-creases unless there is some sort of offsetting action Thus moreprofitable firms automatically have more book equity and more

market equity unless the firm takes offsetting actions

(2) Among large firms the highest profit firms increase their debt themost Those with high profits experience large increases in both thebook and the market value of equity The highest profit firms tend torepurchase equity while the lowest profit firms tend to issue moreequity

(3) Among small firms profit seems to have only a very minor effect ondebt Those with high profits experience some increases in both thebook and the market value of equity Those with low profits experi-ence negative effects on market equity There is a tendency to issueequity with the lowest profit firms issuing the most equity

(4) When firms adjust leverage the magnitude of the adjustment isnot sufficient to fully undo the impact of the underlying shocksFirms do not return to a unique static optimum They seem toldquounderadjustrdquo

There is a huge prior literature on our topic For a review of the literaturesee Frank and Goyal (2008) The inverse relationship between profitabilityand leverage ratios has generated a variety of responses from scholars A

common response articulated by Strebulaev and Whited (2012) is to arguethat the trade-off theory predictions could be more complex in a dynamicmodel Hennessy and Whited (2005) for example specify a dynamic model

of investment and financing under uncertainty and show that a trade-offmodel can explain a number of empirically observed stylized facts includingthe negative leveragendashprofitability relation In their model highly profitablefirms find it unattractive to issue debt because additional debt in such firms

finances a distribution to shareholders rather than replace equityOther scholars focus on adjustment costs and how a proper consideration

of these costs could reconcile the trade-off theory with empirical evidenceMyers (1984) discusses how adjustment costs induce lags in adjusting to the

optimum which could explain the observed wide variation in leverage ratiosFischer Heinkel and Zechner (1989) provide an explicit model of the ideaFama and French (2002) speculate that the negative relation betweenleverage and profitability is perhaps picking-up transitory variation in

1418 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

leverage rather than target variation in leverage A more formal analysis oftransitory debt is provided by DeAngelo DeAngelo and Whited (2011)Leary and Roberts (2005) consider both fixed and variable adjustment

costs and show that firms really do rebalance leverage Firms do notignore leverage ratios when making financing decisions and they do notjust passively accept the impact of equity market shocks Leary andRoberts (2005) do not study the implications of the form of the adjustmentcosts for the profitsndashleverage puzzle which is our focus Their focus is on thespeed of adjustment toward the target and on what that implies about thepersistence of leverageStrebulaev (2007) calibrates a dynamic trade-off model in which all costs

are proportional to the firm value A large firm is a rescaled version of asmall firm and so the target leverage ratio is independent of firm size Thisstructure is not designed to consider the differences between the financingactions of large versus small firms that we observe in the data Because he isstudying a trade-off model he finds that more profitable firms issue moredebt In simulated data Strebulaev (2007) shows that increases in profitabil-ity are correlated with future profitability which in turn increases the valueof the firm and lowers leverage (see p 1772) But expected future profit-ability increases do not always lead to current refinancing His cost structuredoes not lead to the under-adjustment phenomenon of the sort that wedocumentSeveral papers attribute the negative relation between leverage and prof-

itability to firms passively accumulating profits (see eg Fama and French(2002) and Kayhan and Titman (2007)) However the existing evidence isoften indirect and provides mix findings for this idea The implication that atthe time of rebalancing leverage should be positively related to profitabilityis also examined in several earlier studies Mackie-Mason (1990) shows thatcompanies with tax loss carryforwards are more likely to issue equity Incontrast Jung Kim and Stulz (1996) report finding no relation between thelikelihood of equity issuance and profitability Hovakimian Opler andTitman (2001) show that security issuance decisions of firms are afunction of the deviation from target leverage and other regressors includingfirm profitability They show that debt issuances are positively related toprofitability The present article while supporting aspects of their findingsexamines both issuing decisions and magnitudes as a function of firm sizeand profits In doing so we provide a more granular understanding of theeffect of profitability on changes in equity that derive from real operationsand changes in equity that are the result of active issuances and repurchasesof securities Furthermore we provide a sharply different interpretationalong with evidence in support of our interpretation We stress the

PROFITSndashLEVERAGE PUZZLEREVISITED 1419

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

distinction between automatic changes in equity that stem from operationsas opposed to actual issues or repurchases of equity This distinction is em-pirically importantThis article is organized as follows Section 2 provides theoretical motiv-

ation for our empirical strategy Section 3 describes the construction of thedata and variables and provides summary statistics Section 4 providesresults from the fixed effects estimates of leverage ratios Section 5presents the frequency and magnitudes of financing activity for sorts onfirm size and profits Section 6 provides the main results on debt andequity issuance regressions Section 7 explores the debt and equityissuance responses of firms during good and bad times Section 8examines the debt and equity issuances when scaled by total issuances andtotal capital Section 9 concludes the article

2 Empirical Strategy

The key idea in the trade-off theory is that firms choose capital structure bytrading off the various costs and benefits The considerations may includebankruptcy costs tax benefits agency costs transactions costs etc Differentversions of the theory include slightly different elements From our perspec-tive the leverage adjustment costs are of particular importanceA decision to issue debt or to redeem it or to issue shares or to repurchase

them requires managerial thought and consideration There are fixed organ-izational costs There is also the possibility of managerial inertia as docu-mented in Bertrand and Mullainathan (2003) These are fixed costs that areapproximately independent of the magnitude of the adjustments Adjustingalso has marginal costs Depending on which action is taken there may befees for underwriters bankers lawyers etc These may have a fixed compo-nent too Commonly they have a variable cost component which dependson the size of the transaction So both fixed and variable costs of adjustingleverage are normally involved The exact magnitudes vary from situation tosituation What is worse these are likely to have time-varying components4

Why does the form of the transactions costs matter Because the existenceof such fees implies that there will not be a unique leverage target This is anapplication of an idea that goes back to Constantinides and Richard (1978)and Constantinides (1979) It should be stressed that there is nothing newhere in terms of the formal structure This is just a reinterpretation of

4 If there were no time-varying component then it would be easy to identify the exact upperand lower targets for each firm In each case the firm would go exactly to the appropriatetarget In reality with time variation in adjustment costs such precision is not possible

1420 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

standard dynamic optimization conditions For a reader who wants a moreexplicit treatment of the mathematics we recommend Dixit (1993) The ob-servation that this mathematics can be interpreted in terms of corporateleverage in which firm size matters is also not novel It can be found inLeary and Roberts (2005) who provide a simulation with both fixed costsand a weakly convex adjustment costs What is new here is the analysis of theimplication for the impact of profits on leverageSuppose that without active rebalancing leverage would just wander

about due to changes in the stock market value of equity According tothe trade-off theory excessively high leverage is costly (bankruptcy andagency costs) as is excessively low leverage (lost tax benefits and free cashflow problems) If there are no transactions costs then under conventionaltrade-off theory assumptions there will be a unique leverage target that willbe maintained at all timesDue to the fixed costs such frequent rebalancing would be too costly As a

result the firm permits the leverage to drift However if the drift is tooextreme in either direction then the firm will actively rebalance in the direc-tion of the static optimum If the rebalanced firm (inclusive of rebalancingcosts) is worth more then rebalancing is worth it If the rebalanced firmwould be worth less then rebalancing is not worth it Rebalancing takesplace at the point at which the values just matchWhen the firm is rebalancing it also needs to worry about how far to

move in the desired direction When each step has costs the firm needs toensure that the marginal cost of the last step is just equal to the marginalbenefit As a result both from excessively high leverage and from excessivelylow leverage the movements are only part wayTo fix ideas consider an off-the-shelf model of the leverage ratio The

focus is on the impact of the adjustment cost structure The formal structureis essentially the cash management problem studied by Constantinides andRichard (1978) and Dixit (1993) We follow the presentation and notation inDixit (1993) very closely to show that this is just an application of well-established ideas to the leverage problemLet xt denote the leverage ratio at date t is the interest rate Ku is a fixed

cost of actively increasing leverage Kd is the fixed cost of actively reducingleverage mu is the marginal cost of each unit of active leverage increase md isthe marginal cost of each unit of active leverage decrease f(x) is the flowpayoff if leverage is at x F(x) is the expected present value of the payoff ifleverage is at x x is the lower adjustment threshold x is the lower targetleverage (when adjusting from below) x is the upper adjustment thresholdand x is the upper target leverage (when adjusting from above) Accordingto the trade-off theory F 0 gt 0 F 00 lt 0

PROFITSndashLEVERAGE PUZZLEREVISITED 1421

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

In the absence of active adjustments the leverage ratio xt follows astandard process

dx frac14 ethx tTHORNdtthorn ethx tTHORNdw

where is the trend coefficient is the volatility coefficient and w is anormal shock process with zero mean and unit variance If x lt x lt x thenfinancial inaction is optimal If x drops to x then it is reset to x and if xrises to x then it is reset to xThe trade-off theory is frequently thought of as a class of theories in which

firm value first increases in leverage and ultimately drops if leverage gets toohigh This can be motivated by taxes and bankruptcy or by various agencyconcerns A range of ideas can be allowed for when we write that the flowpayoff (dividends andor profits) is fethxtTHORN and it depends on the leverage statext

5 If the firm starts at x0frac14 x then F(x) is the expected present value of theflow payoffs The discount rate is which allows us to write

FethxTHORN frac14 E

Z 10

fethxtTHORNetdtjx0 frac14 x

To maximize the firm value it must choose a leverage policy that sayswhen to rebalance and how much to rebalance These are given by thefamous value matching and smooth pasting conditionsThe value matching conditions are as follows

FethxTHORN FethxTHORN frac14 Ku thornmuethx xTHORN

FethxTHORN FethxTHORN frac14 Kd thornmdethx xTHORN

The smooth pasting conditions are as follows

F 0ethxTHORN frac14 F 0ethxTHORN frac14 mu

F 0ethxTHORN frac14 F 0ethxTHORN frac14 md

As explained by Constantinides and Richard (1978) with standard as-sumptions about the shape of F there are a pair of lines at heights mu andmd such that they intersect the function F 0ethxTHORN in such a way as to define thepoints x lt x lt x lt xTo sum up in this trade-off model with both fixed and variable

rebalancing costs there are four critical leverage points x gt x gt x gt x

5 For simplicity we suppose that the flow payoff is a continuous function of leverageThere is no discrete exit or bankruptcy point This allows us to simply interpret theConstantinides and Richard model much as in Leary and Roberts (2005)

1422 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

If leverage is between x (upper boundary) and x (lower boundary) the firmdoes nothing active Changes in leverage reflect various shocks to revenuesand costs that hit the firm If leverage drifts up to x (or above) then the firmrebalances to x If leverage drifts down to x (or below) the firm rebalancesto xWhat does this imply empirically A firm that does nothing can be

anywhere within the interval between x and x A firm that is rebalancingis a firm that has just hit one of the outer boundaries How does profitabilityaffect the likelihood of hitting a boundary High profits increase the value ofequity This increases the denominator of the leverage ratio and thus makesit more likely that the firm hits the lower boundary Low profits reduce thevalue of equity and thus make it more likely that the firm hits the upperboundaryTo be more specific suppose that the costs are symmetric and the leverage

changes are also following an ordinary Brownian motion Then the median

inert firm will have leverage ofxthornx

2 A firm that just actively reduced leverage

will be at x gtxthornx

2 A firm that just actively increased leverage will be at

x ltxthornx

2 A firm that just reduced leverage will generally be a firm that had

very high leverage To the extent that this reflects a lack of activerefinancing such firm will have low value of equity The low value ofequity will generally reflect low profits A firm that just increased leveragewill generally be a firm that had very low leverage In the same way a firmthat just increased its leverage will tend to have had high value of equity dueto high operating profits6

In this setting it is natural to focus on firms that are actively adjusting andto compare them to each other and to firms that are inactive It is alsohelpful to look directly at financing decisions and profitability rather thanfocusing on a leverage ratioOur approach shares the perspective of Welch (2004) Leary and Roberts

(2005) and Strebulaev (2007) that it is very important to focus on activeadjustments by firms Welch (2004) transforms the data to put everythingback into a ratio form His ldquoimplied debt ratiosrdquo still have both debt andequity in the denominator Similarly Leary and Roberts (2005) andStrebulaev (2007) pay a great deal of attention to leverage ratios whenstudying financing decisions Strebulaev (2007) does not have fixed costssince he makes use of a rescaling property Empirically we find differentadjusting behavior by large and by small firms So the rescaling property is

6 If we knew the values of the critical leverage points then it would be possible to derivemore refined tests Unfortunately those critical values are not observable

PROFITSndashLEVERAGE PUZZLEREVISITED 1423

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

an issue of concern His model does not account for the negative relationshipbetween current leverage and past market-to-book ratioWe depart from the earlier literature by arguing that the use of leverage

ratios can be particularly misleading for profitability since the use of aleverage ratio causes identification problems It is hard to be sure whethera leverage change reflects changes in the numerator or the denominator Thisis particularly problematic when as here the rebalancing is partialIt should be noted that much empirical capital structure literature debates

about whether to use a book leverage ratio or a market leverage ratio Bookleverage is used based on the argument that managers borrow against assetsin place which are better reflected by book assets Others such as Welch(2004) argue that book values are essentially an accounting fiction and thatonly market leverage should count Empirical papers regularly attempt tosidestep this debate by carrying out tests on both book leverage and marketleverage Our point is that both kinds of leverage ratios are potentially mis-leading when examining the impact of profits on capital structure

3 Data

We use conventional data sources starting with the merged Compustat-CRSP data The data are annual and are converted into constant year2000 dollars using the GDP deflator We exclude financial companies (SIC6000-6999) firms involved in major mergers (Compustat footnote code AB)firms reporting financial data in currencies other than the US dollars andfirms with missing data on our key variables7 The ratio variables aretrimmed at the 1 level in both tails of the distribution This serves toremove outliers and the most extremely misrecorded data The finalsample consists of 179021 firm-year observations from 1971 to 2009Table I provides definitions of financial variables and reports summary

statistics The average debt (in constant US$) is about $653 million while themedian is $24 million A significant fraction of firms have zero debt (the 10thpercentile is 0) Book equity is slightly larger than book debt Market equityis almost three times larger than book debt Book assets average $2191million although the medians are considerably smallerIf issuing or retiring securities incurs no fixed costs then we would expect

to see many small actions and very few large actions (Leary and Roberts2005) If there were significant fixed costs involved in issuing or retiringoutstanding securities then small issues might not be worthwhile Table I

7 These include debt book value of equity market value of equity assets book andmarket leverage profitability market-to-book assets ratio and tangibility

1424 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

shows that although most firms issue little debt or equity in a given year theaverages are large In other words when firms actually enter debt and equitymarkets they intervene massivelyThe mean constant dollar debt issue is $165 million (the median is $2

million) In unreported tables we find that the average debt issue is about81 of assets (the median is 18) About 38 of the firms issue no debt8 issue between 0 and 1 of the value of their assets as debt another

Table I Data description

Data sources The sample comes from the annual Compustat files The sample period is

1971ndash2009 We exclude financial companies (SIC 6000-6999) firms involved in majormergers (Compustat footnote code AB) firms reporting financial data in currencies otherthan the US dollar and firms with missing data on the key variables All financial variables

are deflated to year 2000 using the GDP deflator N is the number of observations SD isthe standard deviation D BVE MVE Assets Debt issuance Equity issuance Debt re-payment Equity repayment are all measured in $ millions The ratio variables are trimmed

at the 1 level in both tails of the distribution

Variable definitions Debt (D)frac14 long-term debt (dltt)thorn short-term debt (dlc) book equity(BVE)frac14 common shareholder equity (ceq) market equity (MVE)frac14 no of outstanding

shares (csho) closing price ( prcc f ) assetsfrac14 book assets (at) debt issuancefrac14 issuance oflong-term debt (Maxethdltis 0THORN)thorn increase in current debt (Maxethdlcch 0THORN) equity issu-ancefrac14 sale of common stock (Maxethsstk 0THORN) debt repaymentfrac14 reduction of long-termdebt (Maxethdltr 0THORN)thorn decrease in current debt (Minethdlcch 0THORN) equity repurchasefrac14 purchase

of common stock (Maxethprstkc 0THORN) book leveragefrac14D(DthornBVE) market leveragefrac14D(DthornMVE) profitabilityfrac14EBITDA (oibdp)assets market-to-book ratio M

B frac14Marketvalue of assets (MVA)assets where MVAfrac14DthornMVEthorn preferred-liq value (pstkl)

deferred taxes (txditc) tangibilityfrac14 net property plant and equipment (ppent)assets

Variable N Mean SD

Distribution

10th 50th 90th

Debt (D) 179021 653 4592 0 24 1048

Book equity (BVE) 179021 815 4122 3 70 1373

Market equity (MVE) 179021 1784 9589 9 118 2698

Assets 179021 2191 11530 10 153 3575

Debt issuance 179021 165 1335 0 2 245

Equity issuance 179021 26 186 0 0 40

Debt repayment 179021 132 1123 0 3 181

Equity repurchase 179021 24 257 0 0 9

Book leverage 179021 036 034 000 031 072

Market leverage 179021 027 025 000 020 066

Profitability 179021 005 027 018 011 023MB 179021 165 206 052 102 324

Tangibility 179021 031 024 005 026 070

PROFITSndashLEVERAGE PUZZLEREVISITED 1425

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

16 issue between 1 and 5 of the value of their assets as debt and theremaining 38 issue debt in excess of 5 of the value of their assetsThe mean constant dollar equity issue is about $26 million (the median is

about $04 million) As a fraction of assets the mean and median equityissues are about 7 and 02 respectively About 33 of the firms issue noequity 34 of the firms issue between 0 and 1 of the value of theirassets as equity another 14 issue between 1 and 5 of the value of theirassets as equity and the remaining 19 of the firms issue equity that is inexcess of 5 of the value of their assets Average debt repayments are largerthan equity repurchases This perhaps reflects the finite maturity of debt andits contractual repayment The median firm does not repurchase equityWe construct both book and market leverage ratios Book leverage is

defined as debt over debt plus book equity Market leverage is defined asdebt over debt plus market equity8 The median book leverage is 036 (theaverage is 031) The median market leverage is 027 (the average is 020)Profitability is defined as the ratio of operating income before depreci-

ation to assets While the average firm is profitable (the ratio of EBITDA toassets is 005) the median firm is even more profitable (with a profitabilityratio of 011) But the sample also includes a large number of unprofitablefirms as the 10th percentile is 018 The table also reports descriptive stat-istics on the market-to-book ratio and the tangibility ratio The market-to-book ratio (MB) defined as the ratio of the market value of assets to bookassets averages at about 165 Tangibility defined as the ratio of netproperty plant and equipment to assets averages at about 31

4 Leverage Ratio Regressions

The previous literature focuses on estimates obtained using leverage ratiosHence we start with a similar estimation to check whether our results matchthose of previous studies Table II presents the results for book leverage Weobtain similar results for market leverage and they are presented inAppendix Table AI The regressions include leverage factors following theprevious capital structure literature (see eg Lemmon Roberts and Zender(2008) and Frank and Goyal (2009)) The factors include (i) profitability (ii)industry median leverage (iii) market-to-book assets ratio (iv) tangibility ofassets and (v) firm size Industry median leverage is constructed as themedian leverage of all other firms in the industry excluding the firm under

8 Welch (2011) stresses the idea that nonfinancial liabilities should not be implicitly mis-treated as if they were equity by paying excessively narrow attention to financial liabilitiesin a leverage ratio We have adopted his approach in the empirical work reported here

1426 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

consideration Firm size is defined as the natural log of assets Other factorsare defined in Section 3 We cluster standard errors by firm and estimatethese regressions both with and without fixed effectsIn Table II we run conventional cross-section leverage regressions Since

we employ the usual data it is not surprising that our results match thosereported in the existing capital structure literature Profitability has anegative sign in both the book leverage regressions and in the marketleverage regressions The coefficients on other factors largely match thosereported in earlier studies Firms operating in industries with high leverage

Table II Leverage and profitability

The table presents parameter estimates for book leverage regressions The sample consists

of nonfinancial companies in the annual Compustat files during the period 1971ndash2009 Theleverage is estimated as the ratio of debt over debt plus book equity The explanatoryvariables Profitabilityt1

MB

t1

Tangibilityt1 and LnethAssetsTHORNt1 are described in

Table I IndLevt1 is estimated as the median book leverage of all other firms in thesame industry The industry is defined at the level of the four-digit SIC code All specifi-cations include the year fixed effects The specifications in column (2) additionally include

the firm fixed effects We report t-statistics where the standard errors are clustered at thefirm level in parentheses

aSignificant at the 1 level

Quantile regressions

(1) (2) (3) (4) (5)

Profitabilityt1 026a 020a 011a 017a 041a

(293) (192) (378) (518) (1015)

IndLevt1 042a 012a 030a 052a 052a

(369) (107) (926) (1252) (950)MB

t1

002a 001a 001a 002a 002a

(203) (102) (392) (426) (418)

Tangibilityt1 018a 025a 024a 019a 012a

(206) (154) (964) (580) (276)

LnethAssetsTHORNt1 001a 002a 002a 002a 001a

(114) (76) (899) (478) (96)

Constant 012a 010a 012a 002a 029a

(156) (60) (284) (27) (387)

Year FE Yes Yes Yes Yes Yes

Firm FE No Yes No No No

Regression type OLS FE 25th tile Median 75thtile

Clustered SE Yes Yes No No No

R2 013 054

Pseudo R2 016 015 009

N 158824 158824 158824 158824 158824

PROFITSndashLEVERAGE PUZZLEREVISITED 1427

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

tend to have high leverage A higher market-to-book ratio is associated withlower leverage Larger firms are typically more highly levered The coeffi-cients on tangibility are positive The conclusion from Table II is clear Theleverage regression results match those reported in previous studiesAt times one may be concerned that in a regression what is being

estimated is a conditional mean while some parts of the distributionmight behave differently from other parts If so then ldquoaccidentsrdquo likesample selection issues or asymmetries in the underlying distributionsmight play a greater role than is usually recognized In other words thesource of identification could be misinterpreted To guard against this po-tential problem we estimate the same model using quantile regressions9

Columns 3 4 and 5 report what happens at the 25th 50th and 75th per-centiles The basic model results are robust in terms of both the sign and thestatistical significanceThe impact of profits does appear to be somewhat stronger among the

high leverage firms For instance in a book leverage regression the coeffi-cient on profits for the first quartile is 011 with a t-ratio of 378 For thethird quartile the coefficient on profits is 041 with a t-ratio of 1015This difference is curious and might deserve further study However the factthat in each case we get the negative sign and statistical significance is suf-ficient for the purposes of the present article So the results on profits arerather robust across the distributionTable III sorts firms into those that are actively adjusting their leverage

and those that are passive In the presence of both fixed and variable trans-acts costs this distinction is important We expect the coefficient on profit-ability in leverage ratio regressions to be more negative in firms that arepassive and less negative in firms that are actively adjustingWe define active firms as those that are either issuing net debt or net

equity in excess of 5 of their assets Passive firms are those that neitherissue debt nor equity in nontrivial amounts (ie greater than 5 of theirassets) Consistent with our expectations we find in Table III that the coef-ficients on profits are significantly more negative for passive firms than theyare for active firms With firm and year fixed effects the coefficient onprofits is 008 for active firms and 017 for the passive firms The twocoefficients are significantly different from each other (plt 001) InTable III the standard control variables are used They have the usualsigns and for the most part do not differ much between the active and thepassive firms

9 Cameron and Trivedi (2010) provide an extensive discussion of quantile regressions inStata

1428 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

5 Financing Activity

51 FREQUENCY OF DEBT AND EQUITY ISSUES

Table IV reports the percentage of firms issuing or repurchasing debt orequity for annual sorts based on lagged profitability In Panel A weemploy the conventional 5 cutoff rule to exclude minor fluctuations anddefine firms ldquoissuing debtrdquo as those that issue debt (both short-term andlong-term) in excess of 5 of the value of their assets Other decisions aresimilarly defined using a 5 cutoff Appendix Table AII examines the dis-tributional properties across profitability (and size) quintiles for ldquolargerdquo is-suances by increasing the 5 cutoff to higher levels such as 10 and 15How does security issuance behavior vary with firm profitability To

answer this question Table IV sorts firms by profitability and reports the

Table III Leverage and profitability for active and passive firms

The table presents estimates of leverage on firm characteristics for the subsamples of firms

that are active in either debt or equity markets Active firms are defined as those that issueeither net debt in excess of 5 of the value of assets or issue net equity in excess of 5 ofthe value of their assets The passive firms do not issue significant (more than 5) of either

debt or equity The sample comes from the annual Compustat files during the period 1971ndash2009 Financial firms are excluded The variables are described in Tables I and II Wereport t-statistics where the standard errors are clustered at the firm level in parentheses

aSignificant at the 1 level

Active Passive

(1) (2) (3) (4)

Profitabilityt1 006a 008a 022a 017a

(126) (148) (254) (188)

IndLevt1 044a 019a 041a 013a

(424) (149) (326) (127)MB

t1

002a 001a 003a 001a

(362) (196) (327) (172)

Tangibilityt1 010a 014a 014a 019a

(146) (118) (145) (130)

LnethAssetsTHORNt1 001a 004a 001a 004a

(121) (179) (147) (169)

Constant 012a 008a 003a 007a

(120) (57) (45) (42)

Year FE Yes Yes Yes Yes

Firm FE No Yes No Yes

R2 034 064 033 075

N 43395 43395 58943 58943

PROFITSndashLEVERAGE PUZZLEREVISITED 1429

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

percentage of firms issuing or repurchasing debt or equity The table alsoreports test statistics from tests that examine whether the percentage differ-ences between the high- and low-profit firms are significantly different fromeach other The results in Column (1) show that the likelihood of issuing debt

Table IV Profitability sorts for debt and equity issuers

The sample contains nonfinancial firms listed on the annual Compustat files for the period

from 1971 to 2009 The table presents the percentage of firms issuing and retiring (orrepurchasing) debt and equity The firms are classified as ldquoissuing debtrdquo if they issuedebt in excess of 5 of the value of their assets as ldquoissuing equityrdquo if they issue equity

in excess of 5 of the value of their assets as ldquoretiring debtrdquo if they retire debt in excess of5 of the value of their assets and as ldquorepurchasing equityrdquo if they repurchase equity inexcess of 5 of the value of their assets In addition we also report net debt issuers which

are firms that issue net debt over 5 of the value of their assets and net equity issuerswhich are firms that issue net equity in excess of 5 of the value of their assets Weannually sort firms on lagged profitability and report the percentage of firms in each ofthese categories The bottom part of each panel reports the percentage of firms issuing or

retiring securities by profitability within the smallest and largest asset quintiles The tablealso reports results from t-test for the difference in percentage of firms issuing or retiringsecurities in the lowest and the highest profit quintile

Percentage of firms

Iss Ret Iss Iss Rep Iss Both IssD IssE

D D NetD E E NetE DampE RepE RetD

(1) (2) (3) (4) (5) (6) (7) (8) (9)

Allfirm

s

Low 276 245 186 301 13 297 71 01 38

2 319 299 181 117 21 112 27 02 20

3 385 344 194 86 29 80 21 05 15

4 386 330 213 81 43 75 20 07 14

High 329 273 198 92 82 81 17 13 14

t Highfrac14Low 150 84 39 723 414 763 358 176 200

Smallfirm

s Low 275 204 201 411 11 405 109 01 47

2 291 233 213 379 10 378 99 01 45

3 281 246 204 265 14 261 66 01 41

4 275 272 186 162 15 159 34 02 35

High 275 281 178 136 27 131 25 02 26

t Highfrac14Low 01 84 28 316 55 318 171 18 55

Largefirm

s

Low 414 363 177 87 22 79 28 03 11

2 449 379 180 62 30 57 17 07 09

3 439 360 186 56 41 48 16 09 07

4 403 320 194 46 67 39 13 12 06

High 374 264 207 50 129 35 10 28 05

t Highfrac14Low 50 130 45 89 244 116 77 120 47

1430 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

is mostly independent of firm profitability In the aggregate there is a smalldifference but it is sensitive to sorting on firm size Similarly in Column (2)there is only a weak relation between the profitability and the likelihood thatthe firm retires debt To a first approximation the fraction of firms issuing orrepurchasing debt is largely independent of profits Large firms are somewhatmore likely than small firms to issue debt and also to retire debtUnlike debt issues the probability of issuing equity is strongly related to

profitability as can be seen in Column (4) Low-profit firms are much morelikely to issue equity than are high-profit firms In the lowest profit quintile301 of the firms issued equity In the highest profit quintile it was only92 The lower panel reports separate sorts for small and large firms Similarpatterns are found for both groups of firms that is the likelihood of equityissuance declines as profitability increases The differences between the quin-tiles are quite stark for the group of small firms All quintiles of small firmsissue equity much more frequently than the corresponding large firm quintilesHigh-profit firms are much more likely to repurchase equity as shown in

Column (5) In the overall sample only 13 of low-profit firmsrepurchased equity compared with 82 of the high-profit firms Whenwe separate the small and large firms again a difference emerges Ingeneral small firms do not repurchase much equity while a greaterfraction of the large firms do soGiven these facts it is natural that low profitability firms are much more

likely to be net issuers of equity than are high profitability firms as shown inColumn (6) Among the firms in the lowest quintiles of profits roughly 30of firms issue net equity that exceeds 5 percent of their assets In contrastfor the most profitable quintile only about 8 issue net equityFirm size is an important variable in the recent literaturemdashit is often used

as a proxy for access to capital markets (as in Faulkender and Petersen(2006) and Leary (2009)) Small firms are bank-dependent risky and infor-mationally opaque They have restricted access to public debt markets andconsequently face more severe supply constraints in their ability to issuedebt Thus we expect small firms to be more sluggish in adjusting theirdebt and equity in response to shocks to profitability Large firms incontrast have much easier access to public debt markets and they facefewer obstacles in accessing securities marketsTo examine how size and profitability interact the bottom part of Table IV

examines profitability sorts for the smallest and largest firms We first sortfirms annually by firm size and then within each size quintile we sort them onprofitability We do these two-way sorts to ensure that we have similarnumber of firms in profitability subgroups for both small and large firmsamples Unconditionally small firms are generally less profitable while

PROFITSndashLEVERAGE PUZZLEREVISITED 1431

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

large firms are relatively more profitable The two-way sorts allow us to havea uniform distribution of firms in various profitability quintiles for both smalland large firm samples The bottom part of Table IV reports results for thesmallest and the largest size quintiles Among the small firms there is littlerelation between profitability and the likelihood of issuing debt Howeverlarger firms exhibit a small increase in the likelihood of net issuance of debtwith increasing profitabilityIn contrast to debt the effects of profitability on equity issuance and

repurchases are much more consistent across size quintiles As profitabilityincreases firms are generally less likely to issue equity and more likely torepurchase it Across the two extreme size groups we note a striking differ-ence between the proportion of firms issuing equity and those issuing debtAmong low-profit small firms almost 41 are equity issuers whereas onlyabout 9 of low-profit large firms are equity issuers However regardless ofsize we note a monotonic reduction in the likelihood of issuing equity asprofitability increasesColumn (7) reports the fraction of firms that issue both debt and equity in

excess of 5 of assets both as a function of profits in the upper panel andas a function of firm size and profits in the lower panel In the upper panelwe see that low-profit firms are much more likely to issue both debt andequity (71) while the high-profit firms are less likely to do so (17) Thelower panel shows that small firms are much more likely to be issuers of bothin contrast to large firms In both firm size categories the profit pattern is thesame with low profits being more frequent issuers of both simultaneouslyImportantly Column (8) shows that the likelihood of issuing debt and

simultaneously repurchasing equity increases with profitability Converselyin Column (9) we find that the likelihood of doing the reverse that is issuingequity and retiring debt declines with profitability The effects of sorting onfirm size and firm profitability mirror those reported for all firms Overallthe results show that low-profit firms are less likely to issue debt and repur-chase equity they are instead more likely to issue equity and retire debtFirms with high profitability exhibit the reverse patternIn Appendix Table AII we present results from profitability (and size)

quintiles for large issuances defined using the 10 cutoff (in Panel A) andthe 15 cutoff (in Panel B) The results from these panels are largely con-sistent with those described above

52 MAGNITUDES OF ISSUANCES AND EQUITY CHANGES

In the previous section we considered the probability of having a nontriviallevel of debt or equity activity The next question is how large are the dollar

1432 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

values involved In Table V we sort firms according to profits and thentabulate the levels and changes in both debt and equity We do this firstfor all firms and then for small versus large firmsFor the analysis on all firms we observe in Column (1) that debt peaks at

the middle of the distribution This is because firms with medium profitabil-ity are also the largest as seen in Column (8) which reports average assetvalues for different profitability quintiles When we sort first by firm size andthen examine firms with different levels of profitability within size quintileswe find that for small firms debt is roughly independent of profits while forlarge firms less profitable firms have a higher level of debtAs expected Columns (3) and (5) show that more profitable firms have

higher equity values Columns (2) and (7) consider the relationship betweenissuances of debt and equity and firm profitability Debt issuances are sig-nificantly larger for more profitable firms In contrast equity issuances arelarger for low-profit firms Small low-profit firms issue substantially lessequity compared with small high-profit firms Among large firms the low-profit ones tend to issue equity while the high-profit ones tend to repurchaseequity The fact that more profitable firms issue debt and repurchase equitywhile the least profitable firms retire debt and issue equity is consistent withthe predicted relation between profitability and financing decisions under thetrade-off theoryWe also examine two-way sorts by size and profitability and report results

for the smallest and largest quintile of firms For the smallest quintile offirms the change in debt is largely unrelated to profits But for large firmsthere is a positive relation between profits and debt issuances High-profitfirms have a big positive change in debt Low-profit large firms have anegative change in debtColumns (4) and (6) provide an explanation of why the leverage ratio

regression results contradict those from the basic profitability sorts pre-sented here As we can see profitability indirectly affects leverage ratiosby increasing equity values Changes in both the book value of equity andthe market value of equity are positive and large for highly profitable firmsIn contrast these changes are negative for less profitable firms This evidenceillustrates an important issue concerning the use of leverage ratios Suchratios are often interpreted as essentially reflecting the use of debt by thefirm This interpretation while common is empirically misleadingFor the typical firm the change in the value of equity is larger than the

change in debt For example in the third profit quintile for large firmscolumn (7) the mean equity issue is just $9 million but the change in themarket value of equity is $242 million At the same time the mean change indebt is $27 million This suggests that a fair bit of the observed variation in

PROFITSndashLEVERAGE PUZZLEREVISITED 1433

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

the leverage ratios is primarily driven by the changes in the market value ofequity in the denominator rather than by the changes in debt in the numer-ator Since equity issues are often small this implies that the variation in theleverage ratio is primarily driven by internal operations rather than byexternal financing actions This again points to the fact that leverageratios can provide a misleading account of actual patterns in the dataThere is always a potential concern that the averages may be misleading

due to the impact of outliers To address this concern in AppendixTable AIII we present median values of the profitability sorts Theseresults generally reinforce the findings in Table VTable VI again considers the magnitudes of financing activity but this

time the issues are scaled by a measure of firm size That way it is easier to

Table V Magnitude of financing activity

The sample contains nonfinancial firms listed on the annual Compustat files for the period

from 1971 to 2009 The table presents average debt and equity changes in debt and equityand equity issuances (in millions of dollars) for firms sorted on profitability within sizeclasses The table reports information for all firms sorted on profitability and for profit-

ability sorts within the smallest and largest firms The sorts are done annually The tablealso reports results from t-test for the difference in percentage of firms issuing or retiringsecurities in the lowest and the highest profit quintile

D D BVE BVE MVE MVE EqyIss Assets

(1) (2) (3) (4) (5) (6) (7) (8)

Allfirm

s

Low 83 5 78 10 120 5 7 234

2 605 13 509 24 721 8 17 1691

3 1227 9 1123 22 1834 3 9 3551

4 833 23 1166 7 2408 74 9 2994

High 570 53 1171 71 3817 68 36 2581

t Highfrac14Low 31 9 36 9 38 2 16 36

Smallfirm

s Low 25 04 42 01 332 28 33 97

2 29 06 55 02 278 16 21 116

3 31 07 63 01 212 05 12 127

4 30 06 70 05 179 10 08 136

High 24 06 84 14 234 17 07 142

t Highfrac14Low 04 15 153 85 73 11 201 107

Largefirm

s Low 3929 55 2990 152 4671 11 63 10846

2 3826 19 3178 95 5122 66 11 10721

3 3074 27 3498 41 6173 242 9 9697

4 2623 62 3915 45 8473 94 57 9855

High 2180 193 4642 261 14924 202 148 10171

t Highfrac14Low 94 64 108 87 246 12 169 16

1434 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

see whether large firms behave disproportionately when compared withsmall firms Column (1) shows that low-profit firms have more debtas reported in many previous studies In Columns (4) and (6) we seethat low-profit firms have a major decline in both the book value ofequity and the market value of equity The high-profit firms have thereverse experienceMore interesting are the active decisions of the firms In Column (2) we see

that in contrast to the usual impression the low-profit firms are reducingdebt while the high-profit firms are increasing it In Column (7) we see thatthe low-profit firms are also reducing leverage by issuing equity The high-profit firms issue little equity In the lower panel similar results are foundwhen we separate small and large firms Equity issuing is particularly strong

Table VI Magnitude of financing activitymdashscaled levels and changes

The sample contains nonfinancial firms listed on the annual Compustat files for the period

from 1971 to 2009 The table presents average debt and equity changes in debt and equityand equity issuances scaled by assets for firms sorted on profitability within size classes Thetable reports information for all firms sorted on profitability and for profitability sorts

within the smallest and largest firms The sorts are done annually The table also reportsresults from t-test for the difference in percentage of firms issuing or retiring securities inthe lowest and the highest profit quintile

D D BVE BVE MVE MVE EqyIss

(1) (2) (3) (4) (5) (6) (7)

Allfirm

s

Low 386 122 12 871 3427 1868 217

2 285 58 407 119 940 157 28

3 293 12 425 14 742 03 12

4 269 26 461 02 955 27 10

High 192 15 551 47 1675 74 07

t Highfrac14Low 35 49 50 47 127 42 157

Smallfirm

s Low 545 154 841 2145 9688 9128 696

2 684 49 712 1915 5014 2318 288

3 331 253 141 969 3063 954 129

4 267 181 357 175 1673 221 55

High 194 05 543 26 1758 147 41

t Highfrac14Low 49 17 51 36 88 27 99

Largefirm

s Low 350 30 306 58 511 55 09

2 346 14 335 13 523 02 04

3 353 165 332 71 711 29 01

4 280 01 414 04 909 09 05

High 227 15 477 19 1643 27 14

t Highfrac14Low 371 112 453 181 570 57 221

PROFITSndashLEVERAGE PUZZLEREVISITED 1435

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

among the small low-profit firms that are also experiencing sharp reductionsin the value of equity In column (2) we see that both for the large and for thesmall firms net debt increases are primarily found in the highest quintile offirm profitability

6 Debt and Equity Regressions

So far we have demonstrated that in our data the conventional leverageregressions have the usual signs We then demonstrated that in simple sortsof the data strikingly different active financing patterns emerge The nexttask is to reconcile these differencesTable VII presents simple regressions predicting changes in debt book

equity and market equity as well as equity issuances Consistent with thesorts we control for changes in firm size and we also include firm and yeardummies There may be a concern about the speed of a firmrsquos responsesAccordingly we consider the lagged change in profits as the main case butwe also include the contemporaneous change in profits Whether the contem-poraneous change in profits is included or not does not in fact matter Wereport panel-robust standard errors adjusted for clustering at the firm levelThe results in Table VII support the findings in the sorts Firms that have

an increase in profits in 1 year increase debt both in the same year and in thenext year Firms that have an increase in profits experience an increase inboth the book value of equity and the market value of equity Importantlyhowever firms that experience and increase in profits reduce their equityissues All these effects are statistically significant (except for the marketvalue of equity in the column 5 specification) and seem intuitively reason-able In Appendix Table AIV similar regressions to those presented aboveare reported but without firm-fixed effects We find that empirically verylittle changes whether we include firm fixed effects or leave them out Thusthe targeting behavior to the extent that it happens does not account forwhat we are reportingThe next step is to bring the sorts and the simple regressions together into

a common setting We therefore include factors in addition to profitability(i) median industry leverage (ii) the market-to-book assets ratio (iii) tangi-bility of assets and (iv) firm size (measured by the log of assets) Rajan andZingales (1995) show that these factors are related to leverage in G7countries A number of studies have used these factors to estimateleverage targets Frank and Goyal (2009) show that these factors arerobustly related to leverage in the USA In these regressions we use indica-tors for the quintile that the firm is in for each factor

1436 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

Table VIII reports these results In Columns (1) and (2) we examine debtissues and again show that profits positively affect debt issuances The effectsare large and statistically significant at the 1 level Columns (3) and (4)which examine book equity show a strong effect of profits This effect is onlyslightly reduced by the inclusion of other factors Columns (5) and (6) illus-trate that the impact of profits on the change in market equity is also robustto the inclusion of the conventional factors Columns (7) and (8) examineequity issuances Here we again find that profits have a significant negativeimpact on equity issuances As before the results are robust to the inclusionof the conventional factorsIn unreported tables we estimate debt and equity changes using quantile

regressions In terms of the signs and significance of the profitability variable

Table VII Debt and equity changes

The sample contains nonfinancial firms listed on the annual Compustat files for the period

from 1971 to 2009 The table presents fixed effect estimates from regressions of changes indebt (D) changes in the book value of equity (BVE) changes in the market value ofequity (MVE) and net equity issuances (EqyIss) The explanatory variables include

current and lagged change in operating income before depreciation (Profits andProfitst1) and lagged change in the book value of assets (Assetst1) In addition tofixed firm effects the regressions include year indicator variables The reported t-statistics

are corrected for clustering at the firm level

aMean significant at the 1 level

bMean significant at the 5 level

cMean significant at the 10 level

D BVE MVE EqyIss

(1) (2) (3) (4) (5) (6) (7) (8)

Profitst1 022a 025a 015b 035a 004 060b 005b 007a

(33) (31) (25) (35) (03) (25) (25) (53)

Profits 050b 078a 140a 003a

(24) (32) (48) (34)

Assetst1 000 007b 025a 001b

(00) (21) (35) (25)

Constant 167 283a 34 51 1145a 1182a 605a 595a

(14) (30) (04) (05) (32) (29) (69) (74)

Year FE Yes Yes Yes Yes Yes Yes Yes Yes

Firm FE Yes Yes Yes Yes Yes Yes Yes Yes

R2 001 004 001 010 001 005 002 003

N 162056 162056 162130 162130 157550 157550 162154 162154

PROFITSndashLEVERAGE PUZZLEREVISITED 1437

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

the results are qualitatively similar to the OLS results reported in Table VIIIProfitability positively affects debt issuances and negatively affects equityissuances across various quantiles Overall the results in Table VIII showthat there is nothing that is special about using sorts or regressions to explaindebt and equity issues In either case we find that more profitable firms tendto increase their debt experience an increase in the value of equity andrepurchase shares Thus the control factors are not responsible for theusual rejection of the trade-off theory It would appear that what mattersis the size of the issuing activity relative to the organic increases in equityvalue due to profits In other words what matters in the leverage regressionseems to be coming from the presence of E in D

DthornE and not from D

Table VIII Debt and equity issuances profitability and leverage factor quintiles

The sample contains nonfinancial firms listed on the annual Compustat files for the period

from 1971 to 2009 The table presents estimates from regressions of change in debt (D)change in the book value of equity (BVE) change in the market value of equity (MVE)and net equity issuances (EquityIss) The explanatory variables include lagged profitability

quintiles and lagged leverage factor quintiles In addition to fixed firm effects the regres-sions include year indicator variables The reported t-statistics are corrected for clustering atthe firm level

aMean significant at the 1 level

bMean significant at the 5 level

cMean significant at the 10 level

D BVE MVE EqyIss

(1) (2) (3) (4) (5) (6) (7) (8)

ProfitabilityQint 214a 167a 283a 175a 80 429a 53a 55a

(70) (54) (66) (59) (08) (48) (55) (62)

IndLevQint 104a 43 99 20

(29) (15) (10) (10)

Ln(Assets)Qint52 17 1026a 53a

(16) (05) (77) (35)MB

Qint158a 372a 1155a 09

(45) (56) (106) (06)

TangibilityQint 05 177 436c 64a

(01) (15) (18) (31)

Constant 167 96 656b 235 4613a 8610a 113 570a

(08) (04) (26) (04) (61) (58) (11) (31)

Year FE Yes Yes Yes Yes Yes Yes Yes Yes

Firm FE Yes Yes Yes Yes Yes Yes Yes Yes

R2 0002 0002 0003 0004 0013 0014 0013 0014

N 155587 155037 155587 155037 155587 155037 155587 155037

1438 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

7 Are Financial Market Conditions Important

According to the trade-off theory capital structure is determined by con-sidering the impact of costs and benefits of debt The time variation inthe costs and benefits of debt imply that issuance decisions would alsovary over time Furthermore market conditions affect the transactioncosts of adjusting Adverse selection is a greater problem in a cold marketthan in a hot one Accordingly it may be easier to adjust in good marketconditions Related ideas have been developed by Huang and Ritter(2009) Faulkender et al (2012) and Halling Yu and Zechner (2012)Faulkender Flannery Hankins and Smithrsquos paper examines cross-sectionalvariation in adjustment speeds including market valuations andhow they may affect adjustment speeds It is natural to think that ingood market conditions it will be less costly to issue both debt and equityand adjusting leverage to respond to profitability shocks would be mucheasierTo test the importance of market conditions we require a definition of

good times and bad times Our empirical strategy is to estimate good timesversus bad times at the four-digit industry level We define an industry ashaving ldquogood timesrdquo if the median firm in that industry has a market-to-book ratio that is higher than the 67th percentile of the time-series distribu-tion of the industry median market-to-book ratios Conversely an industryis defined as having ldquobad timesrdquo if the median firm in the industry has amarket-to-book ratio that is lower than the 33rd percentile of the time-seriesdistribution of the industry median market-to-book ratiosPanel A of Table IX tabulates issuance activity for profitability sorts in

both good times and bad times As might be expected active debt and equityissuances are larger during good times There is also somewhat more activeswaps between debt and equity during good timesPanel B reports the issuance activity for the smallest and the largest firms

during the good times and bad times Within each size quintile firms aresorted on profitability This is a three-way sort As before we find that thereis much more active use of external markets during good times Small low-profit firms are more likely to issue equity in good times than in bad timesDebt issuances are significantly higher in good times Large high-profitfirms are significantly more likely to issue debt and repurchase equity ingood times than in bad timesPanel C reports the magnitudes of the financial variables rather than the

frequencies During bad times less profitable large firms retire substantialamounts of debt and they show a tendency to issue equity Small firms donot seem to engage in similar debt reduction activities However like the

PROFITSndashLEVERAGE PUZZLEREVISITED 1439

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

Table IX Debt and equity issues in good and bad times

The table reports the frequency and magnitude of financing activity for sorts on profits for subsamples

of firms in good and bad times An industry is defined as having ldquogood timesrdquo if the median firm in

that industry has a market-to-book ratio that is higher than the 67th percentile of the time-series

distribution of industry median market-to-book ratios Conversely an industry is defined as having

bad times if the median firm in that industry has a market-to-book ratio that is lower than the 33rd

percentile of the time-series distribution of industry median market-to-book ratios Panels A and B

report the percentage of firms (i) issuing net debt in excess of 5 of the value of their assets (ii) issuing

net equity in excess of 5 of the value of their assets (iii) issuing debt and repurchasing equity both in

excess of 5 of the value of their assets and (iv) issuing equity and retiring debt both in excess of 5

of the value of their assets Panel C tabulates the mean debt and equity levels and the mean changes in

debt and equity issuances The sample contains nonfinancial firms listed on the annual Compustat files

for the period from 1971 to 2009

Iss Iss Iss D Iss ENet D Net E Rep E Ret D(1) (2) (3) (4)

Panel A sorts on profitability

Badtimes

Low 145 131 01 202 170 49 02 083 173 46 04 074 182 47 03 10

High 177 51 06 10t Highfrac14Low 41 136 33 01

Goodtimes

Low 206 381 01 462 197 148 02 233 212 103 04 184 234 92 10 18

High 216 92 16 14t Highfrac14Low 23 658 138 23

Panel B sorts on firm size and profitability

Badtimes

Smallfirm

s Low 151 318 00 312 182 276 02 253 151 140 02 294 183 73 00 19

High 177 80 00 11t Highfrac14Low 10 90 NA 20

Largefirm

s Low 172 54 01 042 148 39 04 053 150 43 04 044 156 33 03 06

High 147 28 05 03t Highfrac14Low 17 32 20 02

Goodtimes Smallfirm

s Low 213 448 01 542 230 443 01 473 222 325 00 474 203 197 02 38

High 190 153 02 31t Highfrac14Low 19 230 10 39

Largefirm

s Low 203 96 03 122 208 66 09 133 232 54 16 064 215 39 19 05

High 234 36 38 04t Highfrac14Low 29 99 92 38

(continued)

1440 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

larger firms there is a tendency for small firms to issue more equitymdashat leastto the extent possibleDuring good times large profitable firms raise significant amounts of debt

as they experience an increase in the value of their equity Such firms alsoengage in share repurchases Debt issues by smaller firms are much lessaffected by their own profitability during good times and their change indebt is more or less independent of their profits However equity issuanceis primarily found among the low-profit small firms during good timesTable IX shows that empirically profitability has a time-varying impacton leverage choicesIn Table X the basic regressions for changes in debt and for equity

issuance are presented for both good times and bad times Comparinggood times with bad times we see that the effects of profits are muchstronger in good times In bad times the impacts are rather weak statisticallyThis difference is interesting and deserves further study However for ourpurposes recall that the static trade-off theory implies time-varying capital

Table IX Continued

D D BVE BVE MVE MVE EqyIss Assets

(1) (2) (3) (4) (5) (6) (7) (8)

Panel C magnitude of financing activity

Badtimes Smallfirm

s Low 40 03 58 14 305 06 22 140

2 44 05 82 07 240 67 17 180

3 45 04 94 02 176 18 05 187

4 43 09 98 02 137 04 03 188

High 36 10 114 13 209 16 04 199

Largefirm

s Low 3365 73 2523 169 2304 163 94 8972

2 3031 95 2983 35 3284 278 56 9068

3 2930 5 3360 14 3947 308 57 9225

4 2281 6 3817 19 5213 442 7 8902

High 1790 62 4767 13 8799 58 31 9573

Goodtimes

Smallfirm

s Low 24 06 38 05 339 23 39 90

2 26 08 48 01 301 04 25 104

3 30 09 60 04 246 03 17 125

4 28 07 66 08 206 18 12 130

High 23 06 85 16 261 13 08 141

Largefirm

s Low 4532 36 3477 103 6557 260 53 12465

2 4580 58 3527 139 6911 421 23 12522

3 2952 111 3485 2 7586 279 32 9535

4 2752 124 3939 72 10686 33 105 10226

High 2429 228 4819 377 18565 268 222 10830

PROFITSndashLEVERAGE PUZZLEREVISITED 1441

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

structure choices even when the target is time invariant Empirically we doobserve time-varying choices

8 Is Scaling Important

In the preceding section we provided results for unscaled debt and equityissues This seems appropriate to us However it involves making twochanges to the standard regression We would like to determine if both

Table X Debt and equity changes in good and bad timesmdashregressions

The sample contains nonfinancial firms listed on the annual Compustat files for the period

from 1971 to 2009 The table presents estimates from regressions of change in debt (D)and net equity issuances (EqyIss) in good and bad times The explanatory variables includechange in profits lagged change in profits and lagged change in assets The regressions

include both firm and year fixed effects An industry is defined as having ldquogood timesrdquo ifthe median firm in that industry has a market-to-book ratio that is higher than the 67thpercentile of the time-series distribution of industry median market-to-book ratios

Conversely an industry is defined as having bad times if the median firm in thatindustry has a market-to-book ratio that is lower than the 33rd percentile of the time-series distribution of industry median market-to-book ratios The reported t-statistics arecorrected for clustering at the firm level

aMean significant at the 1 level

bMean significant at the 5 level

cMean significant at the 10 level

Good times Bad times

D EqyIss D EqyIss

(1) (2) (3) (4) (5) (6) (7) (8)

Profits 087a 091a 006b 009a 060 065 003 003b

(30) (29) (20) (28) (15) (16) (16) (20)

Profitst1 029c 015a 034c 001

(19) (37) (17) (04)

Assetst1 000 002b 000 001

(02) (22) (00) (07)

Constant 1337 3989 8691a 10001a 1351 1851 4099a 4055a

(03) (09) (55) (37) (09) (13) (80) (78)

Year FE Yes Yes Yes Yes Yes Yes Yes Yes

Firm FE Yes Yes Yes Yes Yes Yes Yes Yes

R2 0071 0080 0018 0045 0178 0212 0034 0041

N 74827 68729 74891 68787 26965 26013 26971 26018

1442 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

changes are crucial Accordingly in this section we take a step back toconsider prior literature in which variables are scaled The questions arehow does scaling affect the results and whyTable XI examines alternative scalings Columns (1) and (2) scale debt and

equity issuances by the total firm net issuance Columns (3) and (4) scaledebt and equity issuances by the sum of debt and book equity Columns (5)and (6) consider the change in debt scaled by the sum of debt and bookequity Columns (7) and (8) consider the change in leverage due to externalfinancing Since we are examining changes we use the changes in thestandard regression factors as controls The primary interest is in the coef-ficient on the change in profitsWe find that a change in profits is associated with an increase in debt

issues But the sign reverses when we scale the debt issues by the sum of debtplus book equity This is telling What is really driving the results is thechange in equity in the denominator not the change in debt in the numer-ator This result is further substantiated in Columns (5) and (6) where thechange in debt to capital ratio is used instead of debt issuanceTo further explore the impact of active issuance relative to passive

experiencing of changes we decompose leverage by separating out theeffect of debt and equity issues from those due to changes in retainedearnings Thus we define leverage change due to external financing as

LevEF frac14Dt1 thorn d

Dt1 thorn Et1 thorn dthorn e

Dt1

Dt1 thorn Et1 eth1THORN

where D is debt E is book value of equity d is debt issues net of retirementsand e is equity issues net of repurchases Garvey and Hanka (1999) use asimilar measure in their study This measure directly captures the netleverage effect of debt and equity issuancesColumns (7) and (8) report results from regressions of the net leverage

effect of debt and equity issues on changes in profitability after controllingfor other leverage factors in differences and year indicator variables Theeffect of profitability on leverage change due to external financing is positiveand statistically significant This comparison suggests that changes in totaldebt and equity are only partly a result of debt and equity issuance decisionsOther balance sheet adjustments complicate the inferences from leverageratio regressionsFor most of our analysis we follow the literature in treating debt as homo-

genous It is of course potentially interesting to ask what happens when wedistinguish between short-term debt and long-term debt Which part of thefirmrsquos debt maturity structure drives the under-adjustment Table XII

PROFITSndashLEVERAGE PUZZLEREVISITED 1443

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

provides results from such decompositions Alternative scalings areconsidered The main results of interest are that the distinction betweenshort- and long-term debt is important Short-term debt issuance is largelyindependent of changes in profits The negative sign on profit change stemsfrom the action of long-term debt This seems intuitively reasonable

Table XI Scaled issuances

The sample contains nonfinancial firms listed on the annual Compustat files for the period

from 1971 to 2009 The table presents estimates from regressions of debt issuance to totalissuance (in columns (1) and (2)) debt issuance to total capitalization (in columns (3) and(4)) change in debt to capitalization ratio (in columns (5) and (6)) and change in leverage

ratio considering only external financing (in columns (7) and (8)) The change in leverageratio considering only external financing is defined as

LevEF frac14Dt1 thorn d

Dt1 thorn Et1 thorn dthorn e

Dt1

Dt1 thorn Et1

where D is debt E is book value of equity d is debt issues net of retirements and e is equity

issues net of repurchases The regressions include lagged leverage factors and year indicatorvariables Some of the specifications include firm fixed effects The reported t-statistics arecorrected for clustering at the firm level

aMean significant at the 1 level

bMean significant at the 5 level

cMean significant at the 10 level

DebtIssTotIss

DebtIssethDthornETHORN D

DthornE LevEF

(1) (2) (3) (4) (5) (6) (7) (8)

Profitt1 004a 000 002a 001c 003a 002 001b 002a

(35) (03) (34) (18) (35) (14) (23) (26)

IndLevt1 004c 001 001 001c 000 000 000 000

(19) (06) (15) (19) (02) (04) (03) (05)

Market=bookt1 001a 001a 000a 000a 000a 000a 000 000c

(47) (85) (68) (60) (55) (36) (05) (18)

Tangt1 004 005 013a 011a 011a 006a 003a 003a

(12) (15) (144) (112) (79) (38) (44) (42)

Assetst1 008a 004a 006a 004a 002a 001a 003a 002a

(97) (45) (283) (161) (53) (35) (147) (101)

Constant 079a 074a 002a 003a 000 000 000b 002a

(497) (442) (91) (111) (00) (05) (22) (93)

Year FE Yes Yes Yes Yes Yes Yes Yes Yes

Firm FE No Yes No Yes No Yes No Yes

R2 0015 0098 0022 0076 0009 0070 0012 0086

N 133905 133905 140552 140552 138229 138229 140572 140572

1444 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

inasmuch as short-term debt is probably more associated with day-to-dayoperations rather than longer term planning considerationsIn Table XIII we examine change in leverage ( D

DthornE) and change inleverage due to external financing (LEVEF) during good and bad timesseparately The traditional negative effect of a change in profits on achange in leverage is found both in good times and in bad timesHowever when we decompose the impact in order to isolate the externalcomponent as before the result disappears In good times the externalfinancing component of a change in profits is weak but still positiveDuring bad times it is not statistically significant This implies that distin-guishing between good and bad times does not fundamentally alter the basic

Table XII Short- and long-term debt issuances

The sample contains nonfinancial firms listed on the annual Compustat files for the period

from 1971 to 2009 The table presents estimates from regressions of short-term debtissuance to total issuance (in columns (1) and (2)) long-term debt issuance to totalissuance (in columns (3) and (4)) short-term debt issuance to total capitalization (in

columns (5) and (6)) and long-term debt to total capitalization (in columns (7) and (8))Some of the specifications include firm fixed effects The reported t-statistics are correctedfor clustering at the firm level

aMean significant at the 1 level

bMean significant at the 5 level

cMean significant at the 10 level

STDebtIssTotIss

LTDebtIssTotIss

STDebtIssethDthornETHORN

LTDebtIssethDthornETHORN

(1) (2) (3) (4) (5) (6) (7) (8)

Profitabilityt1 0002 0003 0042a 0008 0001 0001 0019a 0012a

(02) (04) (30) (05) (14) (10) (45) (28)

IndLevt1 0009 0007 0070b 0033 0001 0002c 0014a 0005

(06) (05) (26) (11) (08) (17) (26) (09)

M=Bt1 0003a 0003a 0003b 0010a 0000 0000 0003a 0002a

(48) (35) (22) (65) (10) (01) (84) (42)

Tangibilityt1 0020 0007 0011 0067 0005a 0006a 0097a 0081a

(12) (03) (03) (16) (34) (33) (108) (89)

Assetst1 0039a 0011b 0034a 0024b 0001a 0001 0046a 0020a

(89) (22) (38) (23) (36) (13) (190) (82)

Constant 0001b 0153a 0798a 0591a 0000c 0006a 0057a 0096a

(25) (144) (470) (288) (17) (124) (292) (234)

Year FE Yes Yes Yes Yes Yes Yes Yes Yes

Firm FE No Yes No Yes No Yes No Yes

R2 0025 0089 0018 0073 0030 0147 0016 0298

N 136146 136146 136140 136140 143031 143031 142847 142847

PROFITSndashLEVERAGE PUZZLEREVISITED 1445

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

Table XIII Externally financed leverage changes in good and bad times

The sample contains nonfinancial firms listed on the annual Compustat files for the period

from 1971 to 2009 The table presents estimates from regressions of change in debt tocapitalization ratio (in columns (1) (2) (5) and (6)) and change in leverage ratio consider-ing only external financing (in columns (3) (4) (7) and (8)) The change in leverage ratio

considering only external financing is defined as

LevEF frac14Dt1 thorn d

Dt1 thorn Et1 thorn dthorn e

Dt1

Dt1 thorn Et1

where D is debt E is book value of equity d is debt issues net of retirements and e is equityissues net of repurchases The regressions include lagged leverage factors and year indicator

variables The reported t-statistics are corrected for clustering at the firm level An industryis defined as having ldquogood timesrdquo if the median firm in that industry has a market-to-bookratio that is higher than the 67th percentile of the time-series distribution of industry

median market-to-book ratios Conversely an industry is defined as having ldquobad timesrdquoif the median firm in that industry has a market-to-book ratio that is lower than the 33rdpercentile of the time-series distribution of industry median market-to-book ratios

aMean significant at the 1 level

bMean significant at the 5 level

cMean significant at the 10 level

Good times Bad times

DDthornE LevEF D

DthornE LevEF

(1) (2) (3) (4) (5) (6) (7) (8)

Profitabiltyt1 0022c 0005 0016c 0025b 0073b 0041 0007 0007

(17) (03) (18) (24) (25) (11) (04) (03)

IndLevt1 0000 0011 0001 0009 0010 0018 0001 0000

(00) (09) (01) (13) (07) (10) (02) (00)

MB

t1

0004a 0003b 0000 0001 0008a 0001 0000 0000

(40) (24) (05) (08) (28) (04) (03) (03)

Tangt1 0097a 0041 0033a 0038b 0127a 0048 0059a 0033

(45) (16) (27) (25) (49) (14) (43) (16)

Assetst1 0014a 0013b 0024a 0019a 0037a 0023c 0036a 0025a

(30) (20) (84) (48) (45) (17) (88) (43)

Constant 0030b 0036 0000 0003 0028c 0014 0021 0001

(20) (16) (00) (07) (17) (06) (11) (01)

Year FE Yes Yes Yes Yes Yes Yes Yes Yes

Firm FE No Yes No Yes No Yes No Yes

R2 0006 0088 0009 0117 0026 0251 0023 0202

N 56820 56820 58200 58200 23462 23462 23511 23511

1446 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

conclusions The traditional leverage result on profits is driven by increasesin equity that are experienced and then partially but not completely offsetby firmsTo the extent that firms are engaging in active financing they are following

the prescriptions of the trade-off theory Thus we have a coherent picture ofhow firms are responding to profits Basically more profitable firms experiencean increase in the value of their equity They respond to this by issuing debtand repurchasing equity But the magnitude of the increase in equity is largerthan the magnitudes of the active steps taken in the debt and equity marketsThis means that the firmsrsquo actual actions are as predicted by the trade-offtheory But it leaves open the question of why they do not act more stronglyThe capital structure literature has long been interested in dynamics and

capital structure adjustment paths In recent years this interest has increas-ingly resulted in explicit dynamic optimizing models such as those ofGoldstein Ju and Leland (2001) Hennessy and Whited (2005) andStrebulaev (2007) These models commonly make specialized assumptionsleading some to question their generalityWe argue that both fixed transactions costs and variable transactions costs

are plausible and realistic in models that pin down firm size The simultaneouspresence of both is important Due to fixed costs firms only rebalance inter-mittently Due to variable adjustment costs the extent of the adjustment isreduced to the point at which an extra unit to adjustment is balanced againstthe extra unit transactions costs This means that both leverage increases andleverage reductions will only go part way toward a static optimumThe key point is that with both fixed and variable transactions costs we

should expect to only see intermittent leverage adjustments When leverage isadjusted it should ldquounder adjustrdquo when compared to a frictionless modelFurther discussion of transactions costs in a capital structure setting isprovided by Leary and Roberts (2005)

9 Conclusion

The connection between corporate profits and capital structure has beenvery influential in the assessment of the trade-off theory The conventionalinterpretation of the evidence has pushed the literature away from the trade-off theory and toward much more complex dynamic models and ideas As aresult it is important to make sure that the evidence has been correctlyinterpreted We show that the problem is with the interpretation of thetest results and a failure to take into account the costs associated with re-sponding to a profit shock

PROFITSndashLEVERAGE PUZZLEREVISITED 1447

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

We document that more profitable firms really do borrow more (not less)They tend to repurchase equity They experience an increase in both thebook value of equity and the market value of equity Less profitable firmsreally do tend to reduce their debt and to issue equityFirm size and market conditions also matter Larger firms tend to be more

active in the debt markets while smaller firms tend to be relatively moreactive in the equity markets During good times there is more use ofexternal financingThe usual profitsndashleverage puzzle result is primarily driven by the increase

in equity that is experienced by the more profitable firms Accordingly thepuzzle should be restated as asking why do firms not take sufficiently largeoffsetting actions to fully undo the change in equity What limits the mag-nitudes of the typical leverage response to profit shocksIn a frictionless model the partial response appears puzzling But there is

good empirical reason to believe that rebalancing entails both fixed andvariable costs and that firm size matters The rebalancing costs can befully avoided by doing nothing Accordingly the firm must decide whetherany given shock is big enough to be worth responding to If it is then thefirm must decide how big a response is called for These technical conditionsare known as ldquovalue matchingrdquo and ldquosmooth pastingrdquo Thus optimizationimplies that some shocks will be ignored Even if the shock is not ignoredthe optimal response will only partially undo the shock The magnitude ofthe leverage response must balance the marginal cost and the marginalbenefit of an extra unit of leverage Since the marginal cost of adjustingleverage is strictly positive the adjustment toward that static leverageoptimum will only go part way This is true both for leverage increasesand for leverage reductionsOne might conjecture that the variable costs would be time varying and

that such costs might have declined over time Since we do not have a fullaccounting of all of the variable costs it is of course impossible to be sureHowever there is some evidence suggesting that at least the directunderwriting fees are remarkably stable over time10

Surely nobody ought to be surprised that more profitable firms are moreinclined to repurchase equity while unprofitable firms tend not to do so Thefact that more profitable firms typically experience an increase in the value of

10 For IPOs Jay Ritter provides important summary evidence as httpbearwarringtonufleduritterIPOs2013Underwritingpdf He finds that for medium-sized deals a remark-

ably large fraction of the deals have a 7 fee year after year Whether this is due tobehavioral factors or some sort of a coordination device is not clear The extent towhich similar intertemporal stability carries over to other forms of external financemerits further study

1448 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

equity is equally natural Were it not for the contrary claims in the literature(ie the profitsndashleverage puzzle) we would argue that nobody ought to findit surprising that more profitable firms are more inclined to issue debt Thefact that leverage adjustments typically only partially offset profit shocksshould also sound quite natural These facts all fit together easily within thetrade-off theory once we take account of both fixed and variable costs ofactively adjusting capital structure

Appendix

Table AI Market leverage regressions

The table presents estimates of the regressions of market leverage on firm characteristics

The sample comes from the annual Compustat files during the period 1971ndash2009 Financialfirms are excluded The leverage is estimated as the ratio of debt over debt plus marketequity The explanatory variables Profitabilityt1

MB

t1

Tangibilityt1 and LnethAssetsTHORNt1are described in Table 1 IndMedianLevt1 is estimated as the median market leverage of allother firms in the same industry The industry is defined at the level of the four-digit SICcode All specifications include the year fixed effects The specifications in column (2)

additionally include the firm fixed effects We report t-statistics where the standard errorsare clustered at the firm level in parentheses

aSignificant at the 1 level

Quantile regressions

(1) (2) (3) (4) (5)

Profitabilityt1 013a 013a 007a 011a 020a

(275) (262) (349) (425) (538)

IndLevt1 046a 020a 032a 058a 063a

(506) (239) (1299) (1630) (1042)MB

t1

003a 001a 001a 002a 004a

(509) (303) (743) (709) (534)

Tangibilityt1 013a 018a 013a 014a 010a

(191) (181) (762) (571) (232)

LnethAssetsTHORNt1 001a 005a 001a 001a lt001a

(112) (258) (682) (342) (29)

Constant 010a 011a 005a 005a 029a

(171) (98) (173) (118) (398)

Year FE Yes Yes Yes Yes Yes

Firm FE No Yes No No No

Regression type OLS FE 25th tile Median 75thtile

Clustered SE Yes Yes No No No

R2 030 067

Pseudo R2 016 023 0201

N 158578 158578 158578 158578 158578

PROFITSndashLEVERAGE PUZZLEREVISITED 1449

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

Table AII Profitability sorts for firms issuing large amounts

Percentage of firms

Iss Ret Iss Iss Rep Iss Iss IssD IssE

D D NetD E E NetE DampE RepE RetD

(1) (2) (3) (4) (5) (6) (7) (8) (9)

Panel A 10 cutoff to define issuing or retiring activity

Allfirm

s

Low 205 150 134 260 05 257 45 00 23

2 207 175 103 77 06 75 11 01 10

3 243 196 101 50 08 47 08 01 09

4 252 186 113 50 13 48 08 02 08

High 227 161 118 60 27 57 07 04 09

t Highfrac14Low 70 38 62 755 226 768 320 90 152

Smallfirm

s Low 230 132 162 380 05 375 82 00 32

2 228 143 165 334 04 333 65 00 26

3 211 146 147 216 03 214 38 00 23

4 190 154 120 126 04 124 15 01 20

High 194 162 114 99 09 97 12 01 15

t Highfrac14Low 42 40 68 342 20 340 169 07 131

Largefirm

s Low 231 200 82 31 06 29 08 01 04

2 231 187 67 17 06 14 03 01 02

3 234 177 77 17 09 14 03 02 03

4 218 158 75 16 17 14 02 03 02

High 218 135 99 19 41 15 03 07 02

t Highfrac14Low 19 106 36 46 135 57 39 58 23

Percentage of firms

Iss Ret Iss Iss Rep Iss Both Iss D Iss E

D D Net D E E Net E DampE Rep E Ret D

(1) (2) (3) (4) (5) (6) (7) (8) (9)

Panel B 15 cutoff to define issuing or retiring activity

Allfirm

s

Low 158 101 100 230 01 228 27 00 15

2 142 115 63 57 01 56 05 00 06

3 165 128 57 32 01 31 03 00 05

4 173 118 65 35 01 34 03 00 05

High 160 102 70 46 04 44 03 00 05

t Highfrac14Low 08 06 141 738 87 746 277 24 131

Smallfirm

s Low 195 95 132 354 01 350 56 00 24

2 185 96 131 303 01 301 44 00 18

3 160 96 104 179 00 176 18 00 15

4 134 99 78 99 01 98 08 00 12

High 133 95 73 82 01 80 06 00 10

t Highfrac14Low 80 00 95 345 12 344 149 12 52

Largefirm

s Low 145 123 42 14 01 12 03 00 01

2 137 114 32 07 01 06 01 00 01

3 142 104 35 06 01 05 00 00 01

4 130 91 36 08 03 07 01 01 01

High 134 81 48 09 06 07 01 01 01

t Highfrac14Low 19 85 17 26 54 33 34 17 11

1450 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

Table AIV Debt and equity changesmdashwithout fixed effects

The sample contains nonfinancial firms listed on the annual Compustat files for the period from 1971

to 2009 The table presents estimates from regressions of change in debt (D) change in the book

value of equity (BVE) change in the market value of equity (MVE) and net equity issuances

(EqyIss) The explanatory variables include current and lagged change in operating income before

depreciation (Profits and Profitst1) and lagged change in book value of assets (Assetst1) The

regressions include year indicator variables The reported t-statistics are corrected for clustering at the

firm level

aMean significant at the 1 level

bMean significant at the 5 level

cMean significant at the 10 level

D BVE MVE EqyIss

(1) (2) (3) (4) (5) (6) (7) (8)

Profitst1 025a 024a 023a 038a 007 066a 007a 009a

(40) (32) (36) (39) (06) (29) (42) (51)Profits 051b 082a 149a 004a

(25) (33) (49) (33)Assetst1 002 005c 024a 001b

(08) (17) (34) (23)Constant 93 176a 45 71 1552a 1704a 130a 140a

(12) (28) (07) (09) (49) (49) (45) (45)Year FE Yes Yes Yes Yes Yes Yes Yes YesR2 001 005 001 011 001 005 002 002N 162056 162056 162130 162130 157550 157550 162154 162154

Table AIII Median profitability sorts

The sample contains nonfinancial firms listed on the annual Compustat files for the period from 1971

to 2009 The table presents median debt and equity median changes in debt and equity and median

equity issuances (EqyIss) in millions of dollars for all firms sorted on profitability and for profitability

sorts within the smallest and largest firms The sorts are done annually

D D BVE BVE MVE MVE EqyIss Assets(1) (2) (3) (4) (5) (6) (7) (8)

Allfirm

s

Low 21 00 75 17 256 20 01 1922 186 02 449 18 598 19 00 11033 816 06 1274 03 1637 05 00 32244 787 02 1541 18 2488 03 00 3556High 273 00 1361 66 3300 14 00 2578

Smallfirm

s Low 05 00 13 07 140 14 06 402 06 00 25 08 114 12 01 563 09 00 32 05 85 07 00 674 12 00 39 01 68 03 00 79High 10 00 51 04 98 01 00 90

Largefirm

s Low 11485 381 10269 371 13227 202 03 333902 13051 393 12786 140 17571 169 05 392463 11554 287 13429 24 21623 305 00 370314 9963 209 15052 194 29424 308 00 36705High 7361 00 16380 613 49142 439 00 34976

PROFITSndashLEVERAGE PUZZLEREVISITED 1451

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

References

Auerbach A J (1985) Real determinants of corporate leverage in B M Friedman (ed)

Corporate Capital Structures in the United States University of Chicago Press Chicago

IllinoisBertrand M and Mullainathan S (2003) Enjoying the quiet life corporate governance and

managerial preferences Journal of Political Economy 111 1043ndash1075Booth L Varouj A Demirguc-Kunt A and Maksimovic V (2001) Capital structures in

developing countries Journal of Finance 56 87ndash130Brennan M J and Schwartz E S (1984) Optimal financial policy and firm valuation

Journal of Finance 39 593ndash607Cameron A C and Trivedi P K (2010) Microeconometrics using Stata Stata Press TexasConstantinides G M (1979) Multiperiod consumption and investment behavior with

convex transactions costs Management Science 25 1127ndash1137Constantinides G M and Richard S F (1978) Existence of optimal simple policies for

discounted-cost inventory and cash management in continuous time Operations Research

26 620ndash636DeAngelo H DeAngelo L and Whited T M (2011) Capital structure dynamics and

transitory debt Journal of Financial Economics 99 235ndash261Dixit A K (1993) The Art of Smooth Pasting Routledge London and New YorkFama E and French K R (2002) Testing trade-off and pecking order predictions about

dividends and debt Review of Financial Studies 15 1ndash33Faulkender M and Petersen M A (2006) Does the source of capital affect capital struc-

ture Review of Financial Studies 19 45ndash79Faulkender M Flannery M J Hankins K W and Smith J M (2012) Cash Flows and

Leverage Adjustments Journal of Financial Economics 103 632ndash646Fischer E Heinkel R and Zechner J (1989) Dynamic capital structure choice theory

and tests Journal of Finance 44 19ndash40Frank M Z and Goyal V K (2008) Trade-off and pecking order theories of

debt in B E Eckbo (ed) Handbook of Corporate Finance Empirical Corporate Finance

Vol 2 Handbook of Finance Series Chapter 12 ElsevierNorth-Holland Amsterdam pp

135ndash202

Frank M Z and Goyal V K (2009) Capital structure decisions which factors are reliably

important Financial Management 38 1ndash37

Garvey G T and Hanka G (1999) Capital structure and corporate control the effect of

antitakeover statutes on firm leverage Journal of Finance 54 519ndash546

Goldstein R Ju N and Leland H (2001) An EBIT-based model of dynamic capital

structure Journal of Business 74 483ndash512

Graham J R and Tucker A L (2006) Tax shelters and corporate debt policy Journal of

Financial Economics 81 563ndash594

Hackbarth D and Mauer D C (2012) Optimal priority structure capital structure and

investment Review of Financial Studies 25 747ndash796

Halling M Yu J and Zechner J (2012) Leverage dynamics over the business cycle

Working paper University of Utah University of New South Wales and Vienna

University of Economics and Business AdministrationHennessy C A and Whited T A (2005) Debt dynamics Journal of Finance 60 1129ndash1165

Huang R and Ritter J R (2009) Testing theories of capital structure and estimating the

speed of adjustment Journal of Financial and Qualitative Analysis 44 237ndash271

1452 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

Hovakimian A Opler T and Titman S (2001) The debt-equity choice Journal of

Financial and Quantitative Analysis 36 1ndash24Jung K Kim Y and Stulz R M (1996) Timing investment opportunities managerial

discretion and the security issue decisions Journal of Financial Economics 42 159ndash185Kane A Marcus A J and McDonald R L (1984) How big is the tax advantage to

debt Journal of Finance 39 841ndash853Kayhan A and Titman S (2007) Firmsrsquo histories and their capital structures Journal of

Financial Economics 83 1ndash32Leary M T (2009) Bank loan supply lender choice and corporate capital structure

Journal of Finance 64 1143ndash1185Leary M T and Roberts M R (2005) Do firms rebalance their capital structures

Journal of Finance 60 2575ndash2619Leland H E (1994) Corporate debt value bond covenants and optimal capital structure

Journal of finance 49 1213ndash1252Lemmon M L Roberts M R and Zender J F (2008) Back to the beginning persist-

ence and the cross-section of corporate capital structure Journal of Finance 63 1575ndash1608Long M S and Malitz I B (1985) The investment-financing nexus some empirical

evidence Midland Corporate Finance Journal 3 53ndash59Mackie-Mason J K (1990) Do taxes affect corporate financing decisions Journal of

Finance 45 1471ndash1493Morellec E Nikolov B and Schurhoff N (2012) Corporate governance and capital

structure dynamics Journal of Finance 67 803ndash848Myers S C (1984) The capital structure puzzle Journal of Finance 39 575ndash592Myers S C (1993) Still searching for optimal capital structure Journal of Applied

Corporate Finance 61 4ndash14Rajan R and Zingales L (1995) What do we know about capital structure some evidence

from international data Journal of Finance 50 1421ndash1460Strebulaev I A (2007) Do tests of capital structure theory mean what they say Journal of

Finance 62 1747ndash1787

Strebulaev I A and Whited T M (2012) Dynamic Models and Structural Estimation in

Corporate Finance Now Publisher Boston and Delft

Titman S and Wessels R (1988) The determinants of capital structure choice Journal of

Finance 43 1ndash21

Welch I (2004) Capital structure and stock returns Journal of Political Economy 112

106ndash131

Welch I (2011) Two common problems in capital structure research the financial-debt-

to-asset ratio and issuing activity versus leverage changes International Review of Finance

11 1ndash17

PROFITSndashLEVERAGE PUZZLEREVISITED 1453

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

The main empirical findings are as given below

(1) When a firm makes extra profits the book value of equity increasesunless the firm takes some sort of offsetting action Similarly when a

firm makes extra profits the market value of the firmrsquos equity in-creases unless there is some sort of offsetting action Thus moreprofitable firms automatically have more book equity and more

market equity unless the firm takes offsetting actions

(2) Among large firms the highest profit firms increase their debt themost Those with high profits experience large increases in both thebook and the market value of equity The highest profit firms tend torepurchase equity while the lowest profit firms tend to issue moreequity

(3) Among small firms profit seems to have only a very minor effect ondebt Those with high profits experience some increases in both thebook and the market value of equity Those with low profits experi-ence negative effects on market equity There is a tendency to issueequity with the lowest profit firms issuing the most equity

(4) When firms adjust leverage the magnitude of the adjustment isnot sufficient to fully undo the impact of the underlying shocksFirms do not return to a unique static optimum They seem toldquounderadjustrdquo

There is a huge prior literature on our topic For a review of the literaturesee Frank and Goyal (2008) The inverse relationship between profitabilityand leverage ratios has generated a variety of responses from scholars A

common response articulated by Strebulaev and Whited (2012) is to arguethat the trade-off theory predictions could be more complex in a dynamicmodel Hennessy and Whited (2005) for example specify a dynamic model

of investment and financing under uncertainty and show that a trade-offmodel can explain a number of empirically observed stylized facts includingthe negative leveragendashprofitability relation In their model highly profitablefirms find it unattractive to issue debt because additional debt in such firms

finances a distribution to shareholders rather than replace equityOther scholars focus on adjustment costs and how a proper consideration

of these costs could reconcile the trade-off theory with empirical evidenceMyers (1984) discusses how adjustment costs induce lags in adjusting to the

optimum which could explain the observed wide variation in leverage ratiosFischer Heinkel and Zechner (1989) provide an explicit model of the ideaFama and French (2002) speculate that the negative relation betweenleverage and profitability is perhaps picking-up transitory variation in

1418 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

leverage rather than target variation in leverage A more formal analysis oftransitory debt is provided by DeAngelo DeAngelo and Whited (2011)Leary and Roberts (2005) consider both fixed and variable adjustment

costs and show that firms really do rebalance leverage Firms do notignore leverage ratios when making financing decisions and they do notjust passively accept the impact of equity market shocks Leary andRoberts (2005) do not study the implications of the form of the adjustmentcosts for the profitsndashleverage puzzle which is our focus Their focus is on thespeed of adjustment toward the target and on what that implies about thepersistence of leverageStrebulaev (2007) calibrates a dynamic trade-off model in which all costs

are proportional to the firm value A large firm is a rescaled version of asmall firm and so the target leverage ratio is independent of firm size Thisstructure is not designed to consider the differences between the financingactions of large versus small firms that we observe in the data Because he isstudying a trade-off model he finds that more profitable firms issue moredebt In simulated data Strebulaev (2007) shows that increases in profitabil-ity are correlated with future profitability which in turn increases the valueof the firm and lowers leverage (see p 1772) But expected future profit-ability increases do not always lead to current refinancing His cost structuredoes not lead to the under-adjustment phenomenon of the sort that wedocumentSeveral papers attribute the negative relation between leverage and prof-

itability to firms passively accumulating profits (see eg Fama and French(2002) and Kayhan and Titman (2007)) However the existing evidence isoften indirect and provides mix findings for this idea The implication that atthe time of rebalancing leverage should be positively related to profitabilityis also examined in several earlier studies Mackie-Mason (1990) shows thatcompanies with tax loss carryforwards are more likely to issue equity Incontrast Jung Kim and Stulz (1996) report finding no relation between thelikelihood of equity issuance and profitability Hovakimian Opler andTitman (2001) show that security issuance decisions of firms are afunction of the deviation from target leverage and other regressors includingfirm profitability They show that debt issuances are positively related toprofitability The present article while supporting aspects of their findingsexamines both issuing decisions and magnitudes as a function of firm sizeand profits In doing so we provide a more granular understanding of theeffect of profitability on changes in equity that derive from real operationsand changes in equity that are the result of active issuances and repurchasesof securities Furthermore we provide a sharply different interpretationalong with evidence in support of our interpretation We stress the

PROFITSndashLEVERAGE PUZZLEREVISITED 1419

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

distinction between automatic changes in equity that stem from operationsas opposed to actual issues or repurchases of equity This distinction is em-pirically importantThis article is organized as follows Section 2 provides theoretical motiv-

ation for our empirical strategy Section 3 describes the construction of thedata and variables and provides summary statistics Section 4 providesresults from the fixed effects estimates of leverage ratios Section 5presents the frequency and magnitudes of financing activity for sorts onfirm size and profits Section 6 provides the main results on debt andequity issuance regressions Section 7 explores the debt and equityissuance responses of firms during good and bad times Section 8examines the debt and equity issuances when scaled by total issuances andtotal capital Section 9 concludes the article

2 Empirical Strategy

The key idea in the trade-off theory is that firms choose capital structure bytrading off the various costs and benefits The considerations may includebankruptcy costs tax benefits agency costs transactions costs etc Differentversions of the theory include slightly different elements From our perspec-tive the leverage adjustment costs are of particular importanceA decision to issue debt or to redeem it or to issue shares or to repurchase

them requires managerial thought and consideration There are fixed organ-izational costs There is also the possibility of managerial inertia as docu-mented in Bertrand and Mullainathan (2003) These are fixed costs that areapproximately independent of the magnitude of the adjustments Adjustingalso has marginal costs Depending on which action is taken there may befees for underwriters bankers lawyers etc These may have a fixed compo-nent too Commonly they have a variable cost component which dependson the size of the transaction So both fixed and variable costs of adjustingleverage are normally involved The exact magnitudes vary from situation tosituation What is worse these are likely to have time-varying components4

Why does the form of the transactions costs matter Because the existenceof such fees implies that there will not be a unique leverage target This is anapplication of an idea that goes back to Constantinides and Richard (1978)and Constantinides (1979) It should be stressed that there is nothing newhere in terms of the formal structure This is just a reinterpretation of

4 If there were no time-varying component then it would be easy to identify the exact upperand lower targets for each firm In each case the firm would go exactly to the appropriatetarget In reality with time variation in adjustment costs such precision is not possible

1420 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

standard dynamic optimization conditions For a reader who wants a moreexplicit treatment of the mathematics we recommend Dixit (1993) The ob-servation that this mathematics can be interpreted in terms of corporateleverage in which firm size matters is also not novel It can be found inLeary and Roberts (2005) who provide a simulation with both fixed costsand a weakly convex adjustment costs What is new here is the analysis of theimplication for the impact of profits on leverageSuppose that without active rebalancing leverage would just wander

about due to changes in the stock market value of equity According tothe trade-off theory excessively high leverage is costly (bankruptcy andagency costs) as is excessively low leverage (lost tax benefits and free cashflow problems) If there are no transactions costs then under conventionaltrade-off theory assumptions there will be a unique leverage target that willbe maintained at all timesDue to the fixed costs such frequent rebalancing would be too costly As a

result the firm permits the leverage to drift However if the drift is tooextreme in either direction then the firm will actively rebalance in the direc-tion of the static optimum If the rebalanced firm (inclusive of rebalancingcosts) is worth more then rebalancing is worth it If the rebalanced firmwould be worth less then rebalancing is not worth it Rebalancing takesplace at the point at which the values just matchWhen the firm is rebalancing it also needs to worry about how far to

move in the desired direction When each step has costs the firm needs toensure that the marginal cost of the last step is just equal to the marginalbenefit As a result both from excessively high leverage and from excessivelylow leverage the movements are only part wayTo fix ideas consider an off-the-shelf model of the leverage ratio The

focus is on the impact of the adjustment cost structure The formal structureis essentially the cash management problem studied by Constantinides andRichard (1978) and Dixit (1993) We follow the presentation and notation inDixit (1993) very closely to show that this is just an application of well-established ideas to the leverage problemLet xt denote the leverage ratio at date t is the interest rate Ku is a fixed

cost of actively increasing leverage Kd is the fixed cost of actively reducingleverage mu is the marginal cost of each unit of active leverage increase md isthe marginal cost of each unit of active leverage decrease f(x) is the flowpayoff if leverage is at x F(x) is the expected present value of the payoff ifleverage is at x x is the lower adjustment threshold x is the lower targetleverage (when adjusting from below) x is the upper adjustment thresholdand x is the upper target leverage (when adjusting from above) Accordingto the trade-off theory F 0 gt 0 F 00 lt 0

PROFITSndashLEVERAGE PUZZLEREVISITED 1421

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

In the absence of active adjustments the leverage ratio xt follows astandard process

dx frac14 ethx tTHORNdtthorn ethx tTHORNdw

where is the trend coefficient is the volatility coefficient and w is anormal shock process with zero mean and unit variance If x lt x lt x thenfinancial inaction is optimal If x drops to x then it is reset to x and if xrises to x then it is reset to xThe trade-off theory is frequently thought of as a class of theories in which

firm value first increases in leverage and ultimately drops if leverage gets toohigh This can be motivated by taxes and bankruptcy or by various agencyconcerns A range of ideas can be allowed for when we write that the flowpayoff (dividends andor profits) is fethxtTHORN and it depends on the leverage statext

5 If the firm starts at x0frac14 x then F(x) is the expected present value of theflow payoffs The discount rate is which allows us to write

FethxTHORN frac14 E

Z 10

fethxtTHORNetdtjx0 frac14 x

To maximize the firm value it must choose a leverage policy that sayswhen to rebalance and how much to rebalance These are given by thefamous value matching and smooth pasting conditionsThe value matching conditions are as follows

FethxTHORN FethxTHORN frac14 Ku thornmuethx xTHORN

FethxTHORN FethxTHORN frac14 Kd thornmdethx xTHORN

The smooth pasting conditions are as follows

F 0ethxTHORN frac14 F 0ethxTHORN frac14 mu

F 0ethxTHORN frac14 F 0ethxTHORN frac14 md

As explained by Constantinides and Richard (1978) with standard as-sumptions about the shape of F there are a pair of lines at heights mu andmd such that they intersect the function F 0ethxTHORN in such a way as to define thepoints x lt x lt x lt xTo sum up in this trade-off model with both fixed and variable

rebalancing costs there are four critical leverage points x gt x gt x gt x

5 For simplicity we suppose that the flow payoff is a continuous function of leverageThere is no discrete exit or bankruptcy point This allows us to simply interpret theConstantinides and Richard model much as in Leary and Roberts (2005)

1422 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

If leverage is between x (upper boundary) and x (lower boundary) the firmdoes nothing active Changes in leverage reflect various shocks to revenuesand costs that hit the firm If leverage drifts up to x (or above) then the firmrebalances to x If leverage drifts down to x (or below) the firm rebalancesto xWhat does this imply empirically A firm that does nothing can be

anywhere within the interval between x and x A firm that is rebalancingis a firm that has just hit one of the outer boundaries How does profitabilityaffect the likelihood of hitting a boundary High profits increase the value ofequity This increases the denominator of the leverage ratio and thus makesit more likely that the firm hits the lower boundary Low profits reduce thevalue of equity and thus make it more likely that the firm hits the upperboundaryTo be more specific suppose that the costs are symmetric and the leverage

changes are also following an ordinary Brownian motion Then the median

inert firm will have leverage ofxthornx

2 A firm that just actively reduced leverage

will be at x gtxthornx

2 A firm that just actively increased leverage will be at

x ltxthornx

2 A firm that just reduced leverage will generally be a firm that had

very high leverage To the extent that this reflects a lack of activerefinancing such firm will have low value of equity The low value ofequity will generally reflect low profits A firm that just increased leveragewill generally be a firm that had very low leverage In the same way a firmthat just increased its leverage will tend to have had high value of equity dueto high operating profits6

In this setting it is natural to focus on firms that are actively adjusting andto compare them to each other and to firms that are inactive It is alsohelpful to look directly at financing decisions and profitability rather thanfocusing on a leverage ratioOur approach shares the perspective of Welch (2004) Leary and Roberts

(2005) and Strebulaev (2007) that it is very important to focus on activeadjustments by firms Welch (2004) transforms the data to put everythingback into a ratio form His ldquoimplied debt ratiosrdquo still have both debt andequity in the denominator Similarly Leary and Roberts (2005) andStrebulaev (2007) pay a great deal of attention to leverage ratios whenstudying financing decisions Strebulaev (2007) does not have fixed costssince he makes use of a rescaling property Empirically we find differentadjusting behavior by large and by small firms So the rescaling property is

6 If we knew the values of the critical leverage points then it would be possible to derivemore refined tests Unfortunately those critical values are not observable

PROFITSndashLEVERAGE PUZZLEREVISITED 1423

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

an issue of concern His model does not account for the negative relationshipbetween current leverage and past market-to-book ratioWe depart from the earlier literature by arguing that the use of leverage

ratios can be particularly misleading for profitability since the use of aleverage ratio causes identification problems It is hard to be sure whethera leverage change reflects changes in the numerator or the denominator Thisis particularly problematic when as here the rebalancing is partialIt should be noted that much empirical capital structure literature debates

about whether to use a book leverage ratio or a market leverage ratio Bookleverage is used based on the argument that managers borrow against assetsin place which are better reflected by book assets Others such as Welch(2004) argue that book values are essentially an accounting fiction and thatonly market leverage should count Empirical papers regularly attempt tosidestep this debate by carrying out tests on both book leverage and marketleverage Our point is that both kinds of leverage ratios are potentially mis-leading when examining the impact of profits on capital structure

3 Data

We use conventional data sources starting with the merged Compustat-CRSP data The data are annual and are converted into constant year2000 dollars using the GDP deflator We exclude financial companies (SIC6000-6999) firms involved in major mergers (Compustat footnote code AB)firms reporting financial data in currencies other than the US dollars andfirms with missing data on our key variables7 The ratio variables aretrimmed at the 1 level in both tails of the distribution This serves toremove outliers and the most extremely misrecorded data The finalsample consists of 179021 firm-year observations from 1971 to 2009Table I provides definitions of financial variables and reports summary

statistics The average debt (in constant US$) is about $653 million while themedian is $24 million A significant fraction of firms have zero debt (the 10thpercentile is 0) Book equity is slightly larger than book debt Market equityis almost three times larger than book debt Book assets average $2191million although the medians are considerably smallerIf issuing or retiring securities incurs no fixed costs then we would expect

to see many small actions and very few large actions (Leary and Roberts2005) If there were significant fixed costs involved in issuing or retiringoutstanding securities then small issues might not be worthwhile Table I

7 These include debt book value of equity market value of equity assets book andmarket leverage profitability market-to-book assets ratio and tangibility

1424 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

shows that although most firms issue little debt or equity in a given year theaverages are large In other words when firms actually enter debt and equitymarkets they intervene massivelyThe mean constant dollar debt issue is $165 million (the median is $2

million) In unreported tables we find that the average debt issue is about81 of assets (the median is 18) About 38 of the firms issue no debt8 issue between 0 and 1 of the value of their assets as debt another

Table I Data description

Data sources The sample comes from the annual Compustat files The sample period is

1971ndash2009 We exclude financial companies (SIC 6000-6999) firms involved in majormergers (Compustat footnote code AB) firms reporting financial data in currencies otherthan the US dollar and firms with missing data on the key variables All financial variables

are deflated to year 2000 using the GDP deflator N is the number of observations SD isthe standard deviation D BVE MVE Assets Debt issuance Equity issuance Debt re-payment Equity repayment are all measured in $ millions The ratio variables are trimmed

at the 1 level in both tails of the distribution

Variable definitions Debt (D)frac14 long-term debt (dltt)thorn short-term debt (dlc) book equity(BVE)frac14 common shareholder equity (ceq) market equity (MVE)frac14 no of outstanding

shares (csho) closing price ( prcc f ) assetsfrac14 book assets (at) debt issuancefrac14 issuance oflong-term debt (Maxethdltis 0THORN)thorn increase in current debt (Maxethdlcch 0THORN) equity issu-ancefrac14 sale of common stock (Maxethsstk 0THORN) debt repaymentfrac14 reduction of long-termdebt (Maxethdltr 0THORN)thorn decrease in current debt (Minethdlcch 0THORN) equity repurchasefrac14 purchase

of common stock (Maxethprstkc 0THORN) book leveragefrac14D(DthornBVE) market leveragefrac14D(DthornMVE) profitabilityfrac14EBITDA (oibdp)assets market-to-book ratio M

B frac14Marketvalue of assets (MVA)assets where MVAfrac14DthornMVEthorn preferred-liq value (pstkl)

deferred taxes (txditc) tangibilityfrac14 net property plant and equipment (ppent)assets

Variable N Mean SD

Distribution

10th 50th 90th

Debt (D) 179021 653 4592 0 24 1048

Book equity (BVE) 179021 815 4122 3 70 1373

Market equity (MVE) 179021 1784 9589 9 118 2698

Assets 179021 2191 11530 10 153 3575

Debt issuance 179021 165 1335 0 2 245

Equity issuance 179021 26 186 0 0 40

Debt repayment 179021 132 1123 0 3 181

Equity repurchase 179021 24 257 0 0 9

Book leverage 179021 036 034 000 031 072

Market leverage 179021 027 025 000 020 066

Profitability 179021 005 027 018 011 023MB 179021 165 206 052 102 324

Tangibility 179021 031 024 005 026 070

PROFITSndashLEVERAGE PUZZLEREVISITED 1425

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

16 issue between 1 and 5 of the value of their assets as debt and theremaining 38 issue debt in excess of 5 of the value of their assetsThe mean constant dollar equity issue is about $26 million (the median is

about $04 million) As a fraction of assets the mean and median equityissues are about 7 and 02 respectively About 33 of the firms issue noequity 34 of the firms issue between 0 and 1 of the value of theirassets as equity another 14 issue between 1 and 5 of the value of theirassets as equity and the remaining 19 of the firms issue equity that is inexcess of 5 of the value of their assets Average debt repayments are largerthan equity repurchases This perhaps reflects the finite maturity of debt andits contractual repayment The median firm does not repurchase equityWe construct both book and market leverage ratios Book leverage is

defined as debt over debt plus book equity Market leverage is defined asdebt over debt plus market equity8 The median book leverage is 036 (theaverage is 031) The median market leverage is 027 (the average is 020)Profitability is defined as the ratio of operating income before depreci-

ation to assets While the average firm is profitable (the ratio of EBITDA toassets is 005) the median firm is even more profitable (with a profitabilityratio of 011) But the sample also includes a large number of unprofitablefirms as the 10th percentile is 018 The table also reports descriptive stat-istics on the market-to-book ratio and the tangibility ratio The market-to-book ratio (MB) defined as the ratio of the market value of assets to bookassets averages at about 165 Tangibility defined as the ratio of netproperty plant and equipment to assets averages at about 31

4 Leverage Ratio Regressions

The previous literature focuses on estimates obtained using leverage ratiosHence we start with a similar estimation to check whether our results matchthose of previous studies Table II presents the results for book leverage Weobtain similar results for market leverage and they are presented inAppendix Table AI The regressions include leverage factors following theprevious capital structure literature (see eg Lemmon Roberts and Zender(2008) and Frank and Goyal (2009)) The factors include (i) profitability (ii)industry median leverage (iii) market-to-book assets ratio (iv) tangibility ofassets and (v) firm size Industry median leverage is constructed as themedian leverage of all other firms in the industry excluding the firm under

8 Welch (2011) stresses the idea that nonfinancial liabilities should not be implicitly mis-treated as if they were equity by paying excessively narrow attention to financial liabilitiesin a leverage ratio We have adopted his approach in the empirical work reported here

1426 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

consideration Firm size is defined as the natural log of assets Other factorsare defined in Section 3 We cluster standard errors by firm and estimatethese regressions both with and without fixed effectsIn Table II we run conventional cross-section leverage regressions Since

we employ the usual data it is not surprising that our results match thosereported in the existing capital structure literature Profitability has anegative sign in both the book leverage regressions and in the marketleverage regressions The coefficients on other factors largely match thosereported in earlier studies Firms operating in industries with high leverage

Table II Leverage and profitability

The table presents parameter estimates for book leverage regressions The sample consists

of nonfinancial companies in the annual Compustat files during the period 1971ndash2009 Theleverage is estimated as the ratio of debt over debt plus book equity The explanatoryvariables Profitabilityt1

MB

t1

Tangibilityt1 and LnethAssetsTHORNt1 are described in

Table I IndLevt1 is estimated as the median book leverage of all other firms in thesame industry The industry is defined at the level of the four-digit SIC code All specifi-cations include the year fixed effects The specifications in column (2) additionally include

the firm fixed effects We report t-statistics where the standard errors are clustered at thefirm level in parentheses

aSignificant at the 1 level

Quantile regressions

(1) (2) (3) (4) (5)

Profitabilityt1 026a 020a 011a 017a 041a

(293) (192) (378) (518) (1015)

IndLevt1 042a 012a 030a 052a 052a

(369) (107) (926) (1252) (950)MB

t1

002a 001a 001a 002a 002a

(203) (102) (392) (426) (418)

Tangibilityt1 018a 025a 024a 019a 012a

(206) (154) (964) (580) (276)

LnethAssetsTHORNt1 001a 002a 002a 002a 001a

(114) (76) (899) (478) (96)

Constant 012a 010a 012a 002a 029a

(156) (60) (284) (27) (387)

Year FE Yes Yes Yes Yes Yes

Firm FE No Yes No No No

Regression type OLS FE 25th tile Median 75thtile

Clustered SE Yes Yes No No No

R2 013 054

Pseudo R2 016 015 009

N 158824 158824 158824 158824 158824

PROFITSndashLEVERAGE PUZZLEREVISITED 1427

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

tend to have high leverage A higher market-to-book ratio is associated withlower leverage Larger firms are typically more highly levered The coeffi-cients on tangibility are positive The conclusion from Table II is clear Theleverage regression results match those reported in previous studiesAt times one may be concerned that in a regression what is being

estimated is a conditional mean while some parts of the distributionmight behave differently from other parts If so then ldquoaccidentsrdquo likesample selection issues or asymmetries in the underlying distributionsmight play a greater role than is usually recognized In other words thesource of identification could be misinterpreted To guard against this po-tential problem we estimate the same model using quantile regressions9

Columns 3 4 and 5 report what happens at the 25th 50th and 75th per-centiles The basic model results are robust in terms of both the sign and thestatistical significanceThe impact of profits does appear to be somewhat stronger among the

high leverage firms For instance in a book leverage regression the coeffi-cient on profits for the first quartile is 011 with a t-ratio of 378 For thethird quartile the coefficient on profits is 041 with a t-ratio of 1015This difference is curious and might deserve further study However the factthat in each case we get the negative sign and statistical significance is suf-ficient for the purposes of the present article So the results on profits arerather robust across the distributionTable III sorts firms into those that are actively adjusting their leverage

and those that are passive In the presence of both fixed and variable trans-acts costs this distinction is important We expect the coefficient on profit-ability in leverage ratio regressions to be more negative in firms that arepassive and less negative in firms that are actively adjustingWe define active firms as those that are either issuing net debt or net

equity in excess of 5 of their assets Passive firms are those that neitherissue debt nor equity in nontrivial amounts (ie greater than 5 of theirassets) Consistent with our expectations we find in Table III that the coef-ficients on profits are significantly more negative for passive firms than theyare for active firms With firm and year fixed effects the coefficient onprofits is 008 for active firms and 017 for the passive firms The twocoefficients are significantly different from each other (plt 001) InTable III the standard control variables are used They have the usualsigns and for the most part do not differ much between the active and thepassive firms

9 Cameron and Trivedi (2010) provide an extensive discussion of quantile regressions inStata

1428 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

5 Financing Activity

51 FREQUENCY OF DEBT AND EQUITY ISSUES

Table IV reports the percentage of firms issuing or repurchasing debt orequity for annual sorts based on lagged profitability In Panel A weemploy the conventional 5 cutoff rule to exclude minor fluctuations anddefine firms ldquoissuing debtrdquo as those that issue debt (both short-term andlong-term) in excess of 5 of the value of their assets Other decisions aresimilarly defined using a 5 cutoff Appendix Table AII examines the dis-tributional properties across profitability (and size) quintiles for ldquolargerdquo is-suances by increasing the 5 cutoff to higher levels such as 10 and 15How does security issuance behavior vary with firm profitability To

answer this question Table IV sorts firms by profitability and reports the

Table III Leverage and profitability for active and passive firms

The table presents estimates of leverage on firm characteristics for the subsamples of firms

that are active in either debt or equity markets Active firms are defined as those that issueeither net debt in excess of 5 of the value of assets or issue net equity in excess of 5 ofthe value of their assets The passive firms do not issue significant (more than 5) of either

debt or equity The sample comes from the annual Compustat files during the period 1971ndash2009 Financial firms are excluded The variables are described in Tables I and II Wereport t-statistics where the standard errors are clustered at the firm level in parentheses

aSignificant at the 1 level

Active Passive

(1) (2) (3) (4)

Profitabilityt1 006a 008a 022a 017a

(126) (148) (254) (188)

IndLevt1 044a 019a 041a 013a

(424) (149) (326) (127)MB

t1

002a 001a 003a 001a

(362) (196) (327) (172)

Tangibilityt1 010a 014a 014a 019a

(146) (118) (145) (130)

LnethAssetsTHORNt1 001a 004a 001a 004a

(121) (179) (147) (169)

Constant 012a 008a 003a 007a

(120) (57) (45) (42)

Year FE Yes Yes Yes Yes

Firm FE No Yes No Yes

R2 034 064 033 075

N 43395 43395 58943 58943

PROFITSndashLEVERAGE PUZZLEREVISITED 1429

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

percentage of firms issuing or repurchasing debt or equity The table alsoreports test statistics from tests that examine whether the percentage differ-ences between the high- and low-profit firms are significantly different fromeach other The results in Column (1) show that the likelihood of issuing debt

Table IV Profitability sorts for debt and equity issuers

The sample contains nonfinancial firms listed on the annual Compustat files for the period

from 1971 to 2009 The table presents the percentage of firms issuing and retiring (orrepurchasing) debt and equity The firms are classified as ldquoissuing debtrdquo if they issuedebt in excess of 5 of the value of their assets as ldquoissuing equityrdquo if they issue equity

in excess of 5 of the value of their assets as ldquoretiring debtrdquo if they retire debt in excess of5 of the value of their assets and as ldquorepurchasing equityrdquo if they repurchase equity inexcess of 5 of the value of their assets In addition we also report net debt issuers which

are firms that issue net debt over 5 of the value of their assets and net equity issuerswhich are firms that issue net equity in excess of 5 of the value of their assets Weannually sort firms on lagged profitability and report the percentage of firms in each ofthese categories The bottom part of each panel reports the percentage of firms issuing or

retiring securities by profitability within the smallest and largest asset quintiles The tablealso reports results from t-test for the difference in percentage of firms issuing or retiringsecurities in the lowest and the highest profit quintile

Percentage of firms

Iss Ret Iss Iss Rep Iss Both IssD IssE

D D NetD E E NetE DampE RepE RetD

(1) (2) (3) (4) (5) (6) (7) (8) (9)

Allfirm

s

Low 276 245 186 301 13 297 71 01 38

2 319 299 181 117 21 112 27 02 20

3 385 344 194 86 29 80 21 05 15

4 386 330 213 81 43 75 20 07 14

High 329 273 198 92 82 81 17 13 14

t Highfrac14Low 150 84 39 723 414 763 358 176 200

Smallfirm

s Low 275 204 201 411 11 405 109 01 47

2 291 233 213 379 10 378 99 01 45

3 281 246 204 265 14 261 66 01 41

4 275 272 186 162 15 159 34 02 35

High 275 281 178 136 27 131 25 02 26

t Highfrac14Low 01 84 28 316 55 318 171 18 55

Largefirm

s

Low 414 363 177 87 22 79 28 03 11

2 449 379 180 62 30 57 17 07 09

3 439 360 186 56 41 48 16 09 07

4 403 320 194 46 67 39 13 12 06

High 374 264 207 50 129 35 10 28 05

t Highfrac14Low 50 130 45 89 244 116 77 120 47

1430 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

is mostly independent of firm profitability In the aggregate there is a smalldifference but it is sensitive to sorting on firm size Similarly in Column (2)there is only a weak relation between the profitability and the likelihood thatthe firm retires debt To a first approximation the fraction of firms issuing orrepurchasing debt is largely independent of profits Large firms are somewhatmore likely than small firms to issue debt and also to retire debtUnlike debt issues the probability of issuing equity is strongly related to

profitability as can be seen in Column (4) Low-profit firms are much morelikely to issue equity than are high-profit firms In the lowest profit quintile301 of the firms issued equity In the highest profit quintile it was only92 The lower panel reports separate sorts for small and large firms Similarpatterns are found for both groups of firms that is the likelihood of equityissuance declines as profitability increases The differences between the quin-tiles are quite stark for the group of small firms All quintiles of small firmsissue equity much more frequently than the corresponding large firm quintilesHigh-profit firms are much more likely to repurchase equity as shown in

Column (5) In the overall sample only 13 of low-profit firmsrepurchased equity compared with 82 of the high-profit firms Whenwe separate the small and large firms again a difference emerges Ingeneral small firms do not repurchase much equity while a greaterfraction of the large firms do soGiven these facts it is natural that low profitability firms are much more

likely to be net issuers of equity than are high profitability firms as shown inColumn (6) Among the firms in the lowest quintiles of profits roughly 30of firms issue net equity that exceeds 5 percent of their assets In contrastfor the most profitable quintile only about 8 issue net equityFirm size is an important variable in the recent literaturemdashit is often used

as a proxy for access to capital markets (as in Faulkender and Petersen(2006) and Leary (2009)) Small firms are bank-dependent risky and infor-mationally opaque They have restricted access to public debt markets andconsequently face more severe supply constraints in their ability to issuedebt Thus we expect small firms to be more sluggish in adjusting theirdebt and equity in response to shocks to profitability Large firms incontrast have much easier access to public debt markets and they facefewer obstacles in accessing securities marketsTo examine how size and profitability interact the bottom part of Table IV

examines profitability sorts for the smallest and largest firms We first sortfirms annually by firm size and then within each size quintile we sort them onprofitability We do these two-way sorts to ensure that we have similarnumber of firms in profitability subgroups for both small and large firmsamples Unconditionally small firms are generally less profitable while

PROFITSndashLEVERAGE PUZZLEREVISITED 1431

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

large firms are relatively more profitable The two-way sorts allow us to havea uniform distribution of firms in various profitability quintiles for both smalland large firm samples The bottom part of Table IV reports results for thesmallest and the largest size quintiles Among the small firms there is littlerelation between profitability and the likelihood of issuing debt Howeverlarger firms exhibit a small increase in the likelihood of net issuance of debtwith increasing profitabilityIn contrast to debt the effects of profitability on equity issuance and

repurchases are much more consistent across size quintiles As profitabilityincreases firms are generally less likely to issue equity and more likely torepurchase it Across the two extreme size groups we note a striking differ-ence between the proportion of firms issuing equity and those issuing debtAmong low-profit small firms almost 41 are equity issuers whereas onlyabout 9 of low-profit large firms are equity issuers However regardless ofsize we note a monotonic reduction in the likelihood of issuing equity asprofitability increasesColumn (7) reports the fraction of firms that issue both debt and equity in

excess of 5 of assets both as a function of profits in the upper panel andas a function of firm size and profits in the lower panel In the upper panelwe see that low-profit firms are much more likely to issue both debt andequity (71) while the high-profit firms are less likely to do so (17) Thelower panel shows that small firms are much more likely to be issuers of bothin contrast to large firms In both firm size categories the profit pattern is thesame with low profits being more frequent issuers of both simultaneouslyImportantly Column (8) shows that the likelihood of issuing debt and

simultaneously repurchasing equity increases with profitability Converselyin Column (9) we find that the likelihood of doing the reverse that is issuingequity and retiring debt declines with profitability The effects of sorting onfirm size and firm profitability mirror those reported for all firms Overallthe results show that low-profit firms are less likely to issue debt and repur-chase equity they are instead more likely to issue equity and retire debtFirms with high profitability exhibit the reverse patternIn Appendix Table AII we present results from profitability (and size)

quintiles for large issuances defined using the 10 cutoff (in Panel A) andthe 15 cutoff (in Panel B) The results from these panels are largely con-sistent with those described above

52 MAGNITUDES OF ISSUANCES AND EQUITY CHANGES

In the previous section we considered the probability of having a nontriviallevel of debt or equity activity The next question is how large are the dollar

1432 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

values involved In Table V we sort firms according to profits and thentabulate the levels and changes in both debt and equity We do this firstfor all firms and then for small versus large firmsFor the analysis on all firms we observe in Column (1) that debt peaks at

the middle of the distribution This is because firms with medium profitabil-ity are also the largest as seen in Column (8) which reports average assetvalues for different profitability quintiles When we sort first by firm size andthen examine firms with different levels of profitability within size quintileswe find that for small firms debt is roughly independent of profits while forlarge firms less profitable firms have a higher level of debtAs expected Columns (3) and (5) show that more profitable firms have

higher equity values Columns (2) and (7) consider the relationship betweenissuances of debt and equity and firm profitability Debt issuances are sig-nificantly larger for more profitable firms In contrast equity issuances arelarger for low-profit firms Small low-profit firms issue substantially lessequity compared with small high-profit firms Among large firms the low-profit ones tend to issue equity while the high-profit ones tend to repurchaseequity The fact that more profitable firms issue debt and repurchase equitywhile the least profitable firms retire debt and issue equity is consistent withthe predicted relation between profitability and financing decisions under thetrade-off theoryWe also examine two-way sorts by size and profitability and report results

for the smallest and largest quintile of firms For the smallest quintile offirms the change in debt is largely unrelated to profits But for large firmsthere is a positive relation between profits and debt issuances High-profitfirms have a big positive change in debt Low-profit large firms have anegative change in debtColumns (4) and (6) provide an explanation of why the leverage ratio

regression results contradict those from the basic profitability sorts pre-sented here As we can see profitability indirectly affects leverage ratiosby increasing equity values Changes in both the book value of equity andthe market value of equity are positive and large for highly profitable firmsIn contrast these changes are negative for less profitable firms This evidenceillustrates an important issue concerning the use of leverage ratios Suchratios are often interpreted as essentially reflecting the use of debt by thefirm This interpretation while common is empirically misleadingFor the typical firm the change in the value of equity is larger than the

change in debt For example in the third profit quintile for large firmscolumn (7) the mean equity issue is just $9 million but the change in themarket value of equity is $242 million At the same time the mean change indebt is $27 million This suggests that a fair bit of the observed variation in

PROFITSndashLEVERAGE PUZZLEREVISITED 1433

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

the leverage ratios is primarily driven by the changes in the market value ofequity in the denominator rather than by the changes in debt in the numer-ator Since equity issues are often small this implies that the variation in theleverage ratio is primarily driven by internal operations rather than byexternal financing actions This again points to the fact that leverageratios can provide a misleading account of actual patterns in the dataThere is always a potential concern that the averages may be misleading

due to the impact of outliers To address this concern in AppendixTable AIII we present median values of the profitability sorts Theseresults generally reinforce the findings in Table VTable VI again considers the magnitudes of financing activity but this

time the issues are scaled by a measure of firm size That way it is easier to

Table V Magnitude of financing activity

The sample contains nonfinancial firms listed on the annual Compustat files for the period

from 1971 to 2009 The table presents average debt and equity changes in debt and equityand equity issuances (in millions of dollars) for firms sorted on profitability within sizeclasses The table reports information for all firms sorted on profitability and for profit-

ability sorts within the smallest and largest firms The sorts are done annually The tablealso reports results from t-test for the difference in percentage of firms issuing or retiringsecurities in the lowest and the highest profit quintile

D D BVE BVE MVE MVE EqyIss Assets

(1) (2) (3) (4) (5) (6) (7) (8)

Allfirm

s

Low 83 5 78 10 120 5 7 234

2 605 13 509 24 721 8 17 1691

3 1227 9 1123 22 1834 3 9 3551

4 833 23 1166 7 2408 74 9 2994

High 570 53 1171 71 3817 68 36 2581

t Highfrac14Low 31 9 36 9 38 2 16 36

Smallfirm

s Low 25 04 42 01 332 28 33 97

2 29 06 55 02 278 16 21 116

3 31 07 63 01 212 05 12 127

4 30 06 70 05 179 10 08 136

High 24 06 84 14 234 17 07 142

t Highfrac14Low 04 15 153 85 73 11 201 107

Largefirm

s Low 3929 55 2990 152 4671 11 63 10846

2 3826 19 3178 95 5122 66 11 10721

3 3074 27 3498 41 6173 242 9 9697

4 2623 62 3915 45 8473 94 57 9855

High 2180 193 4642 261 14924 202 148 10171

t Highfrac14Low 94 64 108 87 246 12 169 16

1434 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

see whether large firms behave disproportionately when compared withsmall firms Column (1) shows that low-profit firms have more debtas reported in many previous studies In Columns (4) and (6) we seethat low-profit firms have a major decline in both the book value ofequity and the market value of equity The high-profit firms have thereverse experienceMore interesting are the active decisions of the firms In Column (2) we see

that in contrast to the usual impression the low-profit firms are reducingdebt while the high-profit firms are increasing it In Column (7) we see thatthe low-profit firms are also reducing leverage by issuing equity The high-profit firms issue little equity In the lower panel similar results are foundwhen we separate small and large firms Equity issuing is particularly strong

Table VI Magnitude of financing activitymdashscaled levels and changes

The sample contains nonfinancial firms listed on the annual Compustat files for the period

from 1971 to 2009 The table presents average debt and equity changes in debt and equityand equity issuances scaled by assets for firms sorted on profitability within size classes Thetable reports information for all firms sorted on profitability and for profitability sorts

within the smallest and largest firms The sorts are done annually The table also reportsresults from t-test for the difference in percentage of firms issuing or retiring securities inthe lowest and the highest profit quintile

D D BVE BVE MVE MVE EqyIss

(1) (2) (3) (4) (5) (6) (7)

Allfirm

s

Low 386 122 12 871 3427 1868 217

2 285 58 407 119 940 157 28

3 293 12 425 14 742 03 12

4 269 26 461 02 955 27 10

High 192 15 551 47 1675 74 07

t Highfrac14Low 35 49 50 47 127 42 157

Smallfirm

s Low 545 154 841 2145 9688 9128 696

2 684 49 712 1915 5014 2318 288

3 331 253 141 969 3063 954 129

4 267 181 357 175 1673 221 55

High 194 05 543 26 1758 147 41

t Highfrac14Low 49 17 51 36 88 27 99

Largefirm

s Low 350 30 306 58 511 55 09

2 346 14 335 13 523 02 04

3 353 165 332 71 711 29 01

4 280 01 414 04 909 09 05

High 227 15 477 19 1643 27 14

t Highfrac14Low 371 112 453 181 570 57 221

PROFITSndashLEVERAGE PUZZLEREVISITED 1435

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

among the small low-profit firms that are also experiencing sharp reductionsin the value of equity In column (2) we see that both for the large and for thesmall firms net debt increases are primarily found in the highest quintile offirm profitability

6 Debt and Equity Regressions

So far we have demonstrated that in our data the conventional leverageregressions have the usual signs We then demonstrated that in simple sortsof the data strikingly different active financing patterns emerge The nexttask is to reconcile these differencesTable VII presents simple regressions predicting changes in debt book

equity and market equity as well as equity issuances Consistent with thesorts we control for changes in firm size and we also include firm and yeardummies There may be a concern about the speed of a firmrsquos responsesAccordingly we consider the lagged change in profits as the main case butwe also include the contemporaneous change in profits Whether the contem-poraneous change in profits is included or not does not in fact matter Wereport panel-robust standard errors adjusted for clustering at the firm levelThe results in Table VII support the findings in the sorts Firms that have

an increase in profits in 1 year increase debt both in the same year and in thenext year Firms that have an increase in profits experience an increase inboth the book value of equity and the market value of equity Importantlyhowever firms that experience and increase in profits reduce their equityissues All these effects are statistically significant (except for the marketvalue of equity in the column 5 specification) and seem intuitively reason-able In Appendix Table AIV similar regressions to those presented aboveare reported but without firm-fixed effects We find that empirically verylittle changes whether we include firm fixed effects or leave them out Thusthe targeting behavior to the extent that it happens does not account forwhat we are reportingThe next step is to bring the sorts and the simple regressions together into

a common setting We therefore include factors in addition to profitability(i) median industry leverage (ii) the market-to-book assets ratio (iii) tangi-bility of assets and (iv) firm size (measured by the log of assets) Rajan andZingales (1995) show that these factors are related to leverage in G7countries A number of studies have used these factors to estimateleverage targets Frank and Goyal (2009) show that these factors arerobustly related to leverage in the USA In these regressions we use indica-tors for the quintile that the firm is in for each factor

1436 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

Table VIII reports these results In Columns (1) and (2) we examine debtissues and again show that profits positively affect debt issuances The effectsare large and statistically significant at the 1 level Columns (3) and (4)which examine book equity show a strong effect of profits This effect is onlyslightly reduced by the inclusion of other factors Columns (5) and (6) illus-trate that the impact of profits on the change in market equity is also robustto the inclusion of the conventional factors Columns (7) and (8) examineequity issuances Here we again find that profits have a significant negativeimpact on equity issuances As before the results are robust to the inclusionof the conventional factorsIn unreported tables we estimate debt and equity changes using quantile

regressions In terms of the signs and significance of the profitability variable

Table VII Debt and equity changes

The sample contains nonfinancial firms listed on the annual Compustat files for the period

from 1971 to 2009 The table presents fixed effect estimates from regressions of changes indebt (D) changes in the book value of equity (BVE) changes in the market value ofequity (MVE) and net equity issuances (EqyIss) The explanatory variables include

current and lagged change in operating income before depreciation (Profits andProfitst1) and lagged change in the book value of assets (Assetst1) In addition tofixed firm effects the regressions include year indicator variables The reported t-statistics

are corrected for clustering at the firm level

aMean significant at the 1 level

bMean significant at the 5 level

cMean significant at the 10 level

D BVE MVE EqyIss

(1) (2) (3) (4) (5) (6) (7) (8)

Profitst1 022a 025a 015b 035a 004 060b 005b 007a

(33) (31) (25) (35) (03) (25) (25) (53)

Profits 050b 078a 140a 003a

(24) (32) (48) (34)

Assetst1 000 007b 025a 001b

(00) (21) (35) (25)

Constant 167 283a 34 51 1145a 1182a 605a 595a

(14) (30) (04) (05) (32) (29) (69) (74)

Year FE Yes Yes Yes Yes Yes Yes Yes Yes

Firm FE Yes Yes Yes Yes Yes Yes Yes Yes

R2 001 004 001 010 001 005 002 003

N 162056 162056 162130 162130 157550 157550 162154 162154

PROFITSndashLEVERAGE PUZZLEREVISITED 1437

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

the results are qualitatively similar to the OLS results reported in Table VIIIProfitability positively affects debt issuances and negatively affects equityissuances across various quantiles Overall the results in Table VIII showthat there is nothing that is special about using sorts or regressions to explaindebt and equity issues In either case we find that more profitable firms tendto increase their debt experience an increase in the value of equity andrepurchase shares Thus the control factors are not responsible for theusual rejection of the trade-off theory It would appear that what mattersis the size of the issuing activity relative to the organic increases in equityvalue due to profits In other words what matters in the leverage regressionseems to be coming from the presence of E in D

DthornE and not from D

Table VIII Debt and equity issuances profitability and leverage factor quintiles

The sample contains nonfinancial firms listed on the annual Compustat files for the period

from 1971 to 2009 The table presents estimates from regressions of change in debt (D)change in the book value of equity (BVE) change in the market value of equity (MVE)and net equity issuances (EquityIss) The explanatory variables include lagged profitability

quintiles and lagged leverage factor quintiles In addition to fixed firm effects the regres-sions include year indicator variables The reported t-statistics are corrected for clustering atthe firm level

aMean significant at the 1 level

bMean significant at the 5 level

cMean significant at the 10 level

D BVE MVE EqyIss

(1) (2) (3) (4) (5) (6) (7) (8)

ProfitabilityQint 214a 167a 283a 175a 80 429a 53a 55a

(70) (54) (66) (59) (08) (48) (55) (62)

IndLevQint 104a 43 99 20

(29) (15) (10) (10)

Ln(Assets)Qint52 17 1026a 53a

(16) (05) (77) (35)MB

Qint158a 372a 1155a 09

(45) (56) (106) (06)

TangibilityQint 05 177 436c 64a

(01) (15) (18) (31)

Constant 167 96 656b 235 4613a 8610a 113 570a

(08) (04) (26) (04) (61) (58) (11) (31)

Year FE Yes Yes Yes Yes Yes Yes Yes Yes

Firm FE Yes Yes Yes Yes Yes Yes Yes Yes

R2 0002 0002 0003 0004 0013 0014 0013 0014

N 155587 155037 155587 155037 155587 155037 155587 155037

1438 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

7 Are Financial Market Conditions Important

According to the trade-off theory capital structure is determined by con-sidering the impact of costs and benefits of debt The time variation inthe costs and benefits of debt imply that issuance decisions would alsovary over time Furthermore market conditions affect the transactioncosts of adjusting Adverse selection is a greater problem in a cold marketthan in a hot one Accordingly it may be easier to adjust in good marketconditions Related ideas have been developed by Huang and Ritter(2009) Faulkender et al (2012) and Halling Yu and Zechner (2012)Faulkender Flannery Hankins and Smithrsquos paper examines cross-sectionalvariation in adjustment speeds including market valuations andhow they may affect adjustment speeds It is natural to think that ingood market conditions it will be less costly to issue both debt and equityand adjusting leverage to respond to profitability shocks would be mucheasierTo test the importance of market conditions we require a definition of

good times and bad times Our empirical strategy is to estimate good timesversus bad times at the four-digit industry level We define an industry ashaving ldquogood timesrdquo if the median firm in that industry has a market-to-book ratio that is higher than the 67th percentile of the time-series distribu-tion of the industry median market-to-book ratios Conversely an industryis defined as having ldquobad timesrdquo if the median firm in the industry has amarket-to-book ratio that is lower than the 33rd percentile of the time-seriesdistribution of the industry median market-to-book ratiosPanel A of Table IX tabulates issuance activity for profitability sorts in

both good times and bad times As might be expected active debt and equityissuances are larger during good times There is also somewhat more activeswaps between debt and equity during good timesPanel B reports the issuance activity for the smallest and the largest firms

during the good times and bad times Within each size quintile firms aresorted on profitability This is a three-way sort As before we find that thereis much more active use of external markets during good times Small low-profit firms are more likely to issue equity in good times than in bad timesDebt issuances are significantly higher in good times Large high-profitfirms are significantly more likely to issue debt and repurchase equity ingood times than in bad timesPanel C reports the magnitudes of the financial variables rather than the

frequencies During bad times less profitable large firms retire substantialamounts of debt and they show a tendency to issue equity Small firms donot seem to engage in similar debt reduction activities However like the

PROFITSndashLEVERAGE PUZZLEREVISITED 1439

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

Table IX Debt and equity issues in good and bad times

The table reports the frequency and magnitude of financing activity for sorts on profits for subsamples

of firms in good and bad times An industry is defined as having ldquogood timesrdquo if the median firm in

that industry has a market-to-book ratio that is higher than the 67th percentile of the time-series

distribution of industry median market-to-book ratios Conversely an industry is defined as having

bad times if the median firm in that industry has a market-to-book ratio that is lower than the 33rd

percentile of the time-series distribution of industry median market-to-book ratios Panels A and B

report the percentage of firms (i) issuing net debt in excess of 5 of the value of their assets (ii) issuing

net equity in excess of 5 of the value of their assets (iii) issuing debt and repurchasing equity both in

excess of 5 of the value of their assets and (iv) issuing equity and retiring debt both in excess of 5

of the value of their assets Panel C tabulates the mean debt and equity levels and the mean changes in

debt and equity issuances The sample contains nonfinancial firms listed on the annual Compustat files

for the period from 1971 to 2009

Iss Iss Iss D Iss ENet D Net E Rep E Ret D(1) (2) (3) (4)

Panel A sorts on profitability

Badtimes

Low 145 131 01 202 170 49 02 083 173 46 04 074 182 47 03 10

High 177 51 06 10t Highfrac14Low 41 136 33 01

Goodtimes

Low 206 381 01 462 197 148 02 233 212 103 04 184 234 92 10 18

High 216 92 16 14t Highfrac14Low 23 658 138 23

Panel B sorts on firm size and profitability

Badtimes

Smallfirm

s Low 151 318 00 312 182 276 02 253 151 140 02 294 183 73 00 19

High 177 80 00 11t Highfrac14Low 10 90 NA 20

Largefirm

s Low 172 54 01 042 148 39 04 053 150 43 04 044 156 33 03 06

High 147 28 05 03t Highfrac14Low 17 32 20 02

Goodtimes Smallfirm

s Low 213 448 01 542 230 443 01 473 222 325 00 474 203 197 02 38

High 190 153 02 31t Highfrac14Low 19 230 10 39

Largefirm

s Low 203 96 03 122 208 66 09 133 232 54 16 064 215 39 19 05

High 234 36 38 04t Highfrac14Low 29 99 92 38

(continued)

1440 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

larger firms there is a tendency for small firms to issue more equitymdashat leastto the extent possibleDuring good times large profitable firms raise significant amounts of debt

as they experience an increase in the value of their equity Such firms alsoengage in share repurchases Debt issues by smaller firms are much lessaffected by their own profitability during good times and their change indebt is more or less independent of their profits However equity issuanceis primarily found among the low-profit small firms during good timesTable IX shows that empirically profitability has a time-varying impacton leverage choicesIn Table X the basic regressions for changes in debt and for equity

issuance are presented for both good times and bad times Comparinggood times with bad times we see that the effects of profits are muchstronger in good times In bad times the impacts are rather weak statisticallyThis difference is interesting and deserves further study However for ourpurposes recall that the static trade-off theory implies time-varying capital

Table IX Continued

D D BVE BVE MVE MVE EqyIss Assets

(1) (2) (3) (4) (5) (6) (7) (8)

Panel C magnitude of financing activity

Badtimes Smallfirm

s Low 40 03 58 14 305 06 22 140

2 44 05 82 07 240 67 17 180

3 45 04 94 02 176 18 05 187

4 43 09 98 02 137 04 03 188

High 36 10 114 13 209 16 04 199

Largefirm

s Low 3365 73 2523 169 2304 163 94 8972

2 3031 95 2983 35 3284 278 56 9068

3 2930 5 3360 14 3947 308 57 9225

4 2281 6 3817 19 5213 442 7 8902

High 1790 62 4767 13 8799 58 31 9573

Goodtimes

Smallfirm

s Low 24 06 38 05 339 23 39 90

2 26 08 48 01 301 04 25 104

3 30 09 60 04 246 03 17 125

4 28 07 66 08 206 18 12 130

High 23 06 85 16 261 13 08 141

Largefirm

s Low 4532 36 3477 103 6557 260 53 12465

2 4580 58 3527 139 6911 421 23 12522

3 2952 111 3485 2 7586 279 32 9535

4 2752 124 3939 72 10686 33 105 10226

High 2429 228 4819 377 18565 268 222 10830

PROFITSndashLEVERAGE PUZZLEREVISITED 1441

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

structure choices even when the target is time invariant Empirically we doobserve time-varying choices

8 Is Scaling Important

In the preceding section we provided results for unscaled debt and equityissues This seems appropriate to us However it involves making twochanges to the standard regression We would like to determine if both

Table X Debt and equity changes in good and bad timesmdashregressions

The sample contains nonfinancial firms listed on the annual Compustat files for the period

from 1971 to 2009 The table presents estimates from regressions of change in debt (D)and net equity issuances (EqyIss) in good and bad times The explanatory variables includechange in profits lagged change in profits and lagged change in assets The regressions

include both firm and year fixed effects An industry is defined as having ldquogood timesrdquo ifthe median firm in that industry has a market-to-book ratio that is higher than the 67thpercentile of the time-series distribution of industry median market-to-book ratios

Conversely an industry is defined as having bad times if the median firm in thatindustry has a market-to-book ratio that is lower than the 33rd percentile of the time-series distribution of industry median market-to-book ratios The reported t-statistics arecorrected for clustering at the firm level

aMean significant at the 1 level

bMean significant at the 5 level

cMean significant at the 10 level

Good times Bad times

D EqyIss D EqyIss

(1) (2) (3) (4) (5) (6) (7) (8)

Profits 087a 091a 006b 009a 060 065 003 003b

(30) (29) (20) (28) (15) (16) (16) (20)

Profitst1 029c 015a 034c 001

(19) (37) (17) (04)

Assetst1 000 002b 000 001

(02) (22) (00) (07)

Constant 1337 3989 8691a 10001a 1351 1851 4099a 4055a

(03) (09) (55) (37) (09) (13) (80) (78)

Year FE Yes Yes Yes Yes Yes Yes Yes Yes

Firm FE Yes Yes Yes Yes Yes Yes Yes Yes

R2 0071 0080 0018 0045 0178 0212 0034 0041

N 74827 68729 74891 68787 26965 26013 26971 26018

1442 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

changes are crucial Accordingly in this section we take a step back toconsider prior literature in which variables are scaled The questions arehow does scaling affect the results and whyTable XI examines alternative scalings Columns (1) and (2) scale debt and

equity issuances by the total firm net issuance Columns (3) and (4) scaledebt and equity issuances by the sum of debt and book equity Columns (5)and (6) consider the change in debt scaled by the sum of debt and bookequity Columns (7) and (8) consider the change in leverage due to externalfinancing Since we are examining changes we use the changes in thestandard regression factors as controls The primary interest is in the coef-ficient on the change in profitsWe find that a change in profits is associated with an increase in debt

issues But the sign reverses when we scale the debt issues by the sum of debtplus book equity This is telling What is really driving the results is thechange in equity in the denominator not the change in debt in the numer-ator This result is further substantiated in Columns (5) and (6) where thechange in debt to capital ratio is used instead of debt issuanceTo further explore the impact of active issuance relative to passive

experiencing of changes we decompose leverage by separating out theeffect of debt and equity issues from those due to changes in retainedearnings Thus we define leverage change due to external financing as

LevEF frac14Dt1 thorn d

Dt1 thorn Et1 thorn dthorn e

Dt1

Dt1 thorn Et1 eth1THORN

where D is debt E is book value of equity d is debt issues net of retirementsand e is equity issues net of repurchases Garvey and Hanka (1999) use asimilar measure in their study This measure directly captures the netleverage effect of debt and equity issuancesColumns (7) and (8) report results from regressions of the net leverage

effect of debt and equity issues on changes in profitability after controllingfor other leverage factors in differences and year indicator variables Theeffect of profitability on leverage change due to external financing is positiveand statistically significant This comparison suggests that changes in totaldebt and equity are only partly a result of debt and equity issuance decisionsOther balance sheet adjustments complicate the inferences from leverageratio regressionsFor most of our analysis we follow the literature in treating debt as homo-

genous It is of course potentially interesting to ask what happens when wedistinguish between short-term debt and long-term debt Which part of thefirmrsquos debt maturity structure drives the under-adjustment Table XII

PROFITSndashLEVERAGE PUZZLEREVISITED 1443

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

provides results from such decompositions Alternative scalings areconsidered The main results of interest are that the distinction betweenshort- and long-term debt is important Short-term debt issuance is largelyindependent of changes in profits The negative sign on profit change stemsfrom the action of long-term debt This seems intuitively reasonable

Table XI Scaled issuances

The sample contains nonfinancial firms listed on the annual Compustat files for the period

from 1971 to 2009 The table presents estimates from regressions of debt issuance to totalissuance (in columns (1) and (2)) debt issuance to total capitalization (in columns (3) and(4)) change in debt to capitalization ratio (in columns (5) and (6)) and change in leverage

ratio considering only external financing (in columns (7) and (8)) The change in leverageratio considering only external financing is defined as

LevEF frac14Dt1 thorn d

Dt1 thorn Et1 thorn dthorn e

Dt1

Dt1 thorn Et1

where D is debt E is book value of equity d is debt issues net of retirements and e is equity

issues net of repurchases The regressions include lagged leverage factors and year indicatorvariables Some of the specifications include firm fixed effects The reported t-statistics arecorrected for clustering at the firm level

aMean significant at the 1 level

bMean significant at the 5 level

cMean significant at the 10 level

DebtIssTotIss

DebtIssethDthornETHORN D

DthornE LevEF

(1) (2) (3) (4) (5) (6) (7) (8)

Profitt1 004a 000 002a 001c 003a 002 001b 002a

(35) (03) (34) (18) (35) (14) (23) (26)

IndLevt1 004c 001 001 001c 000 000 000 000

(19) (06) (15) (19) (02) (04) (03) (05)

Market=bookt1 001a 001a 000a 000a 000a 000a 000 000c

(47) (85) (68) (60) (55) (36) (05) (18)

Tangt1 004 005 013a 011a 011a 006a 003a 003a

(12) (15) (144) (112) (79) (38) (44) (42)

Assetst1 008a 004a 006a 004a 002a 001a 003a 002a

(97) (45) (283) (161) (53) (35) (147) (101)

Constant 079a 074a 002a 003a 000 000 000b 002a

(497) (442) (91) (111) (00) (05) (22) (93)

Year FE Yes Yes Yes Yes Yes Yes Yes Yes

Firm FE No Yes No Yes No Yes No Yes

R2 0015 0098 0022 0076 0009 0070 0012 0086

N 133905 133905 140552 140552 138229 138229 140572 140572

1444 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

inasmuch as short-term debt is probably more associated with day-to-dayoperations rather than longer term planning considerationsIn Table XIII we examine change in leverage ( D

DthornE) and change inleverage due to external financing (LEVEF) during good and bad timesseparately The traditional negative effect of a change in profits on achange in leverage is found both in good times and in bad timesHowever when we decompose the impact in order to isolate the externalcomponent as before the result disappears In good times the externalfinancing component of a change in profits is weak but still positiveDuring bad times it is not statistically significant This implies that distin-guishing between good and bad times does not fundamentally alter the basic

Table XII Short- and long-term debt issuances

The sample contains nonfinancial firms listed on the annual Compustat files for the period

from 1971 to 2009 The table presents estimates from regressions of short-term debtissuance to total issuance (in columns (1) and (2)) long-term debt issuance to totalissuance (in columns (3) and (4)) short-term debt issuance to total capitalization (in

columns (5) and (6)) and long-term debt to total capitalization (in columns (7) and (8))Some of the specifications include firm fixed effects The reported t-statistics are correctedfor clustering at the firm level

aMean significant at the 1 level

bMean significant at the 5 level

cMean significant at the 10 level

STDebtIssTotIss

LTDebtIssTotIss

STDebtIssethDthornETHORN

LTDebtIssethDthornETHORN

(1) (2) (3) (4) (5) (6) (7) (8)

Profitabilityt1 0002 0003 0042a 0008 0001 0001 0019a 0012a

(02) (04) (30) (05) (14) (10) (45) (28)

IndLevt1 0009 0007 0070b 0033 0001 0002c 0014a 0005

(06) (05) (26) (11) (08) (17) (26) (09)

M=Bt1 0003a 0003a 0003b 0010a 0000 0000 0003a 0002a

(48) (35) (22) (65) (10) (01) (84) (42)

Tangibilityt1 0020 0007 0011 0067 0005a 0006a 0097a 0081a

(12) (03) (03) (16) (34) (33) (108) (89)

Assetst1 0039a 0011b 0034a 0024b 0001a 0001 0046a 0020a

(89) (22) (38) (23) (36) (13) (190) (82)

Constant 0001b 0153a 0798a 0591a 0000c 0006a 0057a 0096a

(25) (144) (470) (288) (17) (124) (292) (234)

Year FE Yes Yes Yes Yes Yes Yes Yes Yes

Firm FE No Yes No Yes No Yes No Yes

R2 0025 0089 0018 0073 0030 0147 0016 0298

N 136146 136146 136140 136140 143031 143031 142847 142847

PROFITSndashLEVERAGE PUZZLEREVISITED 1445

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

Table XIII Externally financed leverage changes in good and bad times

The sample contains nonfinancial firms listed on the annual Compustat files for the period

from 1971 to 2009 The table presents estimates from regressions of change in debt tocapitalization ratio (in columns (1) (2) (5) and (6)) and change in leverage ratio consider-ing only external financing (in columns (3) (4) (7) and (8)) The change in leverage ratio

considering only external financing is defined as

LevEF frac14Dt1 thorn d

Dt1 thorn Et1 thorn dthorn e

Dt1

Dt1 thorn Et1

where D is debt E is book value of equity d is debt issues net of retirements and e is equityissues net of repurchases The regressions include lagged leverage factors and year indicator

variables The reported t-statistics are corrected for clustering at the firm level An industryis defined as having ldquogood timesrdquo if the median firm in that industry has a market-to-bookratio that is higher than the 67th percentile of the time-series distribution of industry

median market-to-book ratios Conversely an industry is defined as having ldquobad timesrdquoif the median firm in that industry has a market-to-book ratio that is lower than the 33rdpercentile of the time-series distribution of industry median market-to-book ratios

aMean significant at the 1 level

bMean significant at the 5 level

cMean significant at the 10 level

Good times Bad times

DDthornE LevEF D

DthornE LevEF

(1) (2) (3) (4) (5) (6) (7) (8)

Profitabiltyt1 0022c 0005 0016c 0025b 0073b 0041 0007 0007

(17) (03) (18) (24) (25) (11) (04) (03)

IndLevt1 0000 0011 0001 0009 0010 0018 0001 0000

(00) (09) (01) (13) (07) (10) (02) (00)

MB

t1

0004a 0003b 0000 0001 0008a 0001 0000 0000

(40) (24) (05) (08) (28) (04) (03) (03)

Tangt1 0097a 0041 0033a 0038b 0127a 0048 0059a 0033

(45) (16) (27) (25) (49) (14) (43) (16)

Assetst1 0014a 0013b 0024a 0019a 0037a 0023c 0036a 0025a

(30) (20) (84) (48) (45) (17) (88) (43)

Constant 0030b 0036 0000 0003 0028c 0014 0021 0001

(20) (16) (00) (07) (17) (06) (11) (01)

Year FE Yes Yes Yes Yes Yes Yes Yes Yes

Firm FE No Yes No Yes No Yes No Yes

R2 0006 0088 0009 0117 0026 0251 0023 0202

N 56820 56820 58200 58200 23462 23462 23511 23511

1446 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

conclusions The traditional leverage result on profits is driven by increasesin equity that are experienced and then partially but not completely offsetby firmsTo the extent that firms are engaging in active financing they are following

the prescriptions of the trade-off theory Thus we have a coherent picture ofhow firms are responding to profits Basically more profitable firms experiencean increase in the value of their equity They respond to this by issuing debtand repurchasing equity But the magnitude of the increase in equity is largerthan the magnitudes of the active steps taken in the debt and equity marketsThis means that the firmsrsquo actual actions are as predicted by the trade-offtheory But it leaves open the question of why they do not act more stronglyThe capital structure literature has long been interested in dynamics and

capital structure adjustment paths In recent years this interest has increas-ingly resulted in explicit dynamic optimizing models such as those ofGoldstein Ju and Leland (2001) Hennessy and Whited (2005) andStrebulaev (2007) These models commonly make specialized assumptionsleading some to question their generalityWe argue that both fixed transactions costs and variable transactions costs

are plausible and realistic in models that pin down firm size The simultaneouspresence of both is important Due to fixed costs firms only rebalance inter-mittently Due to variable adjustment costs the extent of the adjustment isreduced to the point at which an extra unit to adjustment is balanced againstthe extra unit transactions costs This means that both leverage increases andleverage reductions will only go part way toward a static optimumThe key point is that with both fixed and variable transactions costs we

should expect to only see intermittent leverage adjustments When leverage isadjusted it should ldquounder adjustrdquo when compared to a frictionless modelFurther discussion of transactions costs in a capital structure setting isprovided by Leary and Roberts (2005)

9 Conclusion

The connection between corporate profits and capital structure has beenvery influential in the assessment of the trade-off theory The conventionalinterpretation of the evidence has pushed the literature away from the trade-off theory and toward much more complex dynamic models and ideas As aresult it is important to make sure that the evidence has been correctlyinterpreted We show that the problem is with the interpretation of thetest results and a failure to take into account the costs associated with re-sponding to a profit shock

PROFITSndashLEVERAGE PUZZLEREVISITED 1447

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

We document that more profitable firms really do borrow more (not less)They tend to repurchase equity They experience an increase in both thebook value of equity and the market value of equity Less profitable firmsreally do tend to reduce their debt and to issue equityFirm size and market conditions also matter Larger firms tend to be more

active in the debt markets while smaller firms tend to be relatively moreactive in the equity markets During good times there is more use ofexternal financingThe usual profitsndashleverage puzzle result is primarily driven by the increase

in equity that is experienced by the more profitable firms Accordingly thepuzzle should be restated as asking why do firms not take sufficiently largeoffsetting actions to fully undo the change in equity What limits the mag-nitudes of the typical leverage response to profit shocksIn a frictionless model the partial response appears puzzling But there is

good empirical reason to believe that rebalancing entails both fixed andvariable costs and that firm size matters The rebalancing costs can befully avoided by doing nothing Accordingly the firm must decide whetherany given shock is big enough to be worth responding to If it is then thefirm must decide how big a response is called for These technical conditionsare known as ldquovalue matchingrdquo and ldquosmooth pastingrdquo Thus optimizationimplies that some shocks will be ignored Even if the shock is not ignoredthe optimal response will only partially undo the shock The magnitude ofthe leverage response must balance the marginal cost and the marginalbenefit of an extra unit of leverage Since the marginal cost of adjustingleverage is strictly positive the adjustment toward that static leverageoptimum will only go part way This is true both for leverage increasesand for leverage reductionsOne might conjecture that the variable costs would be time varying and

that such costs might have declined over time Since we do not have a fullaccounting of all of the variable costs it is of course impossible to be sureHowever there is some evidence suggesting that at least the directunderwriting fees are remarkably stable over time10

Surely nobody ought to be surprised that more profitable firms are moreinclined to repurchase equity while unprofitable firms tend not to do so Thefact that more profitable firms typically experience an increase in the value of

10 For IPOs Jay Ritter provides important summary evidence as httpbearwarringtonufleduritterIPOs2013Underwritingpdf He finds that for medium-sized deals a remark-

ably large fraction of the deals have a 7 fee year after year Whether this is due tobehavioral factors or some sort of a coordination device is not clear The extent towhich similar intertemporal stability carries over to other forms of external financemerits further study

1448 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

equity is equally natural Were it not for the contrary claims in the literature(ie the profitsndashleverage puzzle) we would argue that nobody ought to findit surprising that more profitable firms are more inclined to issue debt Thefact that leverage adjustments typically only partially offset profit shocksshould also sound quite natural These facts all fit together easily within thetrade-off theory once we take account of both fixed and variable costs ofactively adjusting capital structure

Appendix

Table AI Market leverage regressions

The table presents estimates of the regressions of market leverage on firm characteristics

The sample comes from the annual Compustat files during the period 1971ndash2009 Financialfirms are excluded The leverage is estimated as the ratio of debt over debt plus marketequity The explanatory variables Profitabilityt1

MB

t1

Tangibilityt1 and LnethAssetsTHORNt1are described in Table 1 IndMedianLevt1 is estimated as the median market leverage of allother firms in the same industry The industry is defined at the level of the four-digit SICcode All specifications include the year fixed effects The specifications in column (2)

additionally include the firm fixed effects We report t-statistics where the standard errorsare clustered at the firm level in parentheses

aSignificant at the 1 level

Quantile regressions

(1) (2) (3) (4) (5)

Profitabilityt1 013a 013a 007a 011a 020a

(275) (262) (349) (425) (538)

IndLevt1 046a 020a 032a 058a 063a

(506) (239) (1299) (1630) (1042)MB

t1

003a 001a 001a 002a 004a

(509) (303) (743) (709) (534)

Tangibilityt1 013a 018a 013a 014a 010a

(191) (181) (762) (571) (232)

LnethAssetsTHORNt1 001a 005a 001a 001a lt001a

(112) (258) (682) (342) (29)

Constant 010a 011a 005a 005a 029a

(171) (98) (173) (118) (398)

Year FE Yes Yes Yes Yes Yes

Firm FE No Yes No No No

Regression type OLS FE 25th tile Median 75thtile

Clustered SE Yes Yes No No No

R2 030 067

Pseudo R2 016 023 0201

N 158578 158578 158578 158578 158578

PROFITSndashLEVERAGE PUZZLEREVISITED 1449

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

Table AII Profitability sorts for firms issuing large amounts

Percentage of firms

Iss Ret Iss Iss Rep Iss Iss IssD IssE

D D NetD E E NetE DampE RepE RetD

(1) (2) (3) (4) (5) (6) (7) (8) (9)

Panel A 10 cutoff to define issuing or retiring activity

Allfirm

s

Low 205 150 134 260 05 257 45 00 23

2 207 175 103 77 06 75 11 01 10

3 243 196 101 50 08 47 08 01 09

4 252 186 113 50 13 48 08 02 08

High 227 161 118 60 27 57 07 04 09

t Highfrac14Low 70 38 62 755 226 768 320 90 152

Smallfirm

s Low 230 132 162 380 05 375 82 00 32

2 228 143 165 334 04 333 65 00 26

3 211 146 147 216 03 214 38 00 23

4 190 154 120 126 04 124 15 01 20

High 194 162 114 99 09 97 12 01 15

t Highfrac14Low 42 40 68 342 20 340 169 07 131

Largefirm

s Low 231 200 82 31 06 29 08 01 04

2 231 187 67 17 06 14 03 01 02

3 234 177 77 17 09 14 03 02 03

4 218 158 75 16 17 14 02 03 02

High 218 135 99 19 41 15 03 07 02

t Highfrac14Low 19 106 36 46 135 57 39 58 23

Percentage of firms

Iss Ret Iss Iss Rep Iss Both Iss D Iss E

D D Net D E E Net E DampE Rep E Ret D

(1) (2) (3) (4) (5) (6) (7) (8) (9)

Panel B 15 cutoff to define issuing or retiring activity

Allfirm

s

Low 158 101 100 230 01 228 27 00 15

2 142 115 63 57 01 56 05 00 06

3 165 128 57 32 01 31 03 00 05

4 173 118 65 35 01 34 03 00 05

High 160 102 70 46 04 44 03 00 05

t Highfrac14Low 08 06 141 738 87 746 277 24 131

Smallfirm

s Low 195 95 132 354 01 350 56 00 24

2 185 96 131 303 01 301 44 00 18

3 160 96 104 179 00 176 18 00 15

4 134 99 78 99 01 98 08 00 12

High 133 95 73 82 01 80 06 00 10

t Highfrac14Low 80 00 95 345 12 344 149 12 52

Largefirm

s Low 145 123 42 14 01 12 03 00 01

2 137 114 32 07 01 06 01 00 01

3 142 104 35 06 01 05 00 00 01

4 130 91 36 08 03 07 01 01 01

High 134 81 48 09 06 07 01 01 01

t Highfrac14Low 19 85 17 26 54 33 34 17 11

1450 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

Table AIV Debt and equity changesmdashwithout fixed effects

The sample contains nonfinancial firms listed on the annual Compustat files for the period from 1971

to 2009 The table presents estimates from regressions of change in debt (D) change in the book

value of equity (BVE) change in the market value of equity (MVE) and net equity issuances

(EqyIss) The explanatory variables include current and lagged change in operating income before

depreciation (Profits and Profitst1) and lagged change in book value of assets (Assetst1) The

regressions include year indicator variables The reported t-statistics are corrected for clustering at the

firm level

aMean significant at the 1 level

bMean significant at the 5 level

cMean significant at the 10 level

D BVE MVE EqyIss

(1) (2) (3) (4) (5) (6) (7) (8)

Profitst1 025a 024a 023a 038a 007 066a 007a 009a

(40) (32) (36) (39) (06) (29) (42) (51)Profits 051b 082a 149a 004a

(25) (33) (49) (33)Assetst1 002 005c 024a 001b

(08) (17) (34) (23)Constant 93 176a 45 71 1552a 1704a 130a 140a

(12) (28) (07) (09) (49) (49) (45) (45)Year FE Yes Yes Yes Yes Yes Yes Yes YesR2 001 005 001 011 001 005 002 002N 162056 162056 162130 162130 157550 157550 162154 162154

Table AIII Median profitability sorts

The sample contains nonfinancial firms listed on the annual Compustat files for the period from 1971

to 2009 The table presents median debt and equity median changes in debt and equity and median

equity issuances (EqyIss) in millions of dollars for all firms sorted on profitability and for profitability

sorts within the smallest and largest firms The sorts are done annually

D D BVE BVE MVE MVE EqyIss Assets(1) (2) (3) (4) (5) (6) (7) (8)

Allfirm

s

Low 21 00 75 17 256 20 01 1922 186 02 449 18 598 19 00 11033 816 06 1274 03 1637 05 00 32244 787 02 1541 18 2488 03 00 3556High 273 00 1361 66 3300 14 00 2578

Smallfirm

s Low 05 00 13 07 140 14 06 402 06 00 25 08 114 12 01 563 09 00 32 05 85 07 00 674 12 00 39 01 68 03 00 79High 10 00 51 04 98 01 00 90

Largefirm

s Low 11485 381 10269 371 13227 202 03 333902 13051 393 12786 140 17571 169 05 392463 11554 287 13429 24 21623 305 00 370314 9963 209 15052 194 29424 308 00 36705High 7361 00 16380 613 49142 439 00 34976

PROFITSndashLEVERAGE PUZZLEREVISITED 1451

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

References

Auerbach A J (1985) Real determinants of corporate leverage in B M Friedman (ed)

Corporate Capital Structures in the United States University of Chicago Press Chicago

IllinoisBertrand M and Mullainathan S (2003) Enjoying the quiet life corporate governance and

managerial preferences Journal of Political Economy 111 1043ndash1075Booth L Varouj A Demirguc-Kunt A and Maksimovic V (2001) Capital structures in

developing countries Journal of Finance 56 87ndash130Brennan M J and Schwartz E S (1984) Optimal financial policy and firm valuation

Journal of Finance 39 593ndash607Cameron A C and Trivedi P K (2010) Microeconometrics using Stata Stata Press TexasConstantinides G M (1979) Multiperiod consumption and investment behavior with

convex transactions costs Management Science 25 1127ndash1137Constantinides G M and Richard S F (1978) Existence of optimal simple policies for

discounted-cost inventory and cash management in continuous time Operations Research

26 620ndash636DeAngelo H DeAngelo L and Whited T M (2011) Capital structure dynamics and

transitory debt Journal of Financial Economics 99 235ndash261Dixit A K (1993) The Art of Smooth Pasting Routledge London and New YorkFama E and French K R (2002) Testing trade-off and pecking order predictions about

dividends and debt Review of Financial Studies 15 1ndash33Faulkender M and Petersen M A (2006) Does the source of capital affect capital struc-

ture Review of Financial Studies 19 45ndash79Faulkender M Flannery M J Hankins K W and Smith J M (2012) Cash Flows and

Leverage Adjustments Journal of Financial Economics 103 632ndash646Fischer E Heinkel R and Zechner J (1989) Dynamic capital structure choice theory

and tests Journal of Finance 44 19ndash40Frank M Z and Goyal V K (2008) Trade-off and pecking order theories of

debt in B E Eckbo (ed) Handbook of Corporate Finance Empirical Corporate Finance

Vol 2 Handbook of Finance Series Chapter 12 ElsevierNorth-Holland Amsterdam pp

135ndash202

Frank M Z and Goyal V K (2009) Capital structure decisions which factors are reliably

important Financial Management 38 1ndash37

Garvey G T and Hanka G (1999) Capital structure and corporate control the effect of

antitakeover statutes on firm leverage Journal of Finance 54 519ndash546

Goldstein R Ju N and Leland H (2001) An EBIT-based model of dynamic capital

structure Journal of Business 74 483ndash512

Graham J R and Tucker A L (2006) Tax shelters and corporate debt policy Journal of

Financial Economics 81 563ndash594

Hackbarth D and Mauer D C (2012) Optimal priority structure capital structure and

investment Review of Financial Studies 25 747ndash796

Halling M Yu J and Zechner J (2012) Leverage dynamics over the business cycle

Working paper University of Utah University of New South Wales and Vienna

University of Economics and Business AdministrationHennessy C A and Whited T A (2005) Debt dynamics Journal of Finance 60 1129ndash1165

Huang R and Ritter J R (2009) Testing theories of capital structure and estimating the

speed of adjustment Journal of Financial and Qualitative Analysis 44 237ndash271

1452 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

Hovakimian A Opler T and Titman S (2001) The debt-equity choice Journal of

Financial and Quantitative Analysis 36 1ndash24Jung K Kim Y and Stulz R M (1996) Timing investment opportunities managerial

discretion and the security issue decisions Journal of Financial Economics 42 159ndash185Kane A Marcus A J and McDonald R L (1984) How big is the tax advantage to

debt Journal of Finance 39 841ndash853Kayhan A and Titman S (2007) Firmsrsquo histories and their capital structures Journal of

Financial Economics 83 1ndash32Leary M T (2009) Bank loan supply lender choice and corporate capital structure

Journal of Finance 64 1143ndash1185Leary M T and Roberts M R (2005) Do firms rebalance their capital structures

Journal of Finance 60 2575ndash2619Leland H E (1994) Corporate debt value bond covenants and optimal capital structure

Journal of finance 49 1213ndash1252Lemmon M L Roberts M R and Zender J F (2008) Back to the beginning persist-

ence and the cross-section of corporate capital structure Journal of Finance 63 1575ndash1608Long M S and Malitz I B (1985) The investment-financing nexus some empirical

evidence Midland Corporate Finance Journal 3 53ndash59Mackie-Mason J K (1990) Do taxes affect corporate financing decisions Journal of

Finance 45 1471ndash1493Morellec E Nikolov B and Schurhoff N (2012) Corporate governance and capital

structure dynamics Journal of Finance 67 803ndash848Myers S C (1984) The capital structure puzzle Journal of Finance 39 575ndash592Myers S C (1993) Still searching for optimal capital structure Journal of Applied

Corporate Finance 61 4ndash14Rajan R and Zingales L (1995) What do we know about capital structure some evidence

from international data Journal of Finance 50 1421ndash1460Strebulaev I A (2007) Do tests of capital structure theory mean what they say Journal of

Finance 62 1747ndash1787

Strebulaev I A and Whited T M (2012) Dynamic Models and Structural Estimation in

Corporate Finance Now Publisher Boston and Delft

Titman S and Wessels R (1988) The determinants of capital structure choice Journal of

Finance 43 1ndash21

Welch I (2004) Capital structure and stock returns Journal of Political Economy 112

106ndash131

Welch I (2011) Two common problems in capital structure research the financial-debt-

to-asset ratio and issuing activity versus leverage changes International Review of Finance

11 1ndash17

PROFITSndashLEVERAGE PUZZLEREVISITED 1453

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

leverage rather than target variation in leverage A more formal analysis oftransitory debt is provided by DeAngelo DeAngelo and Whited (2011)Leary and Roberts (2005) consider both fixed and variable adjustment

costs and show that firms really do rebalance leverage Firms do notignore leverage ratios when making financing decisions and they do notjust passively accept the impact of equity market shocks Leary andRoberts (2005) do not study the implications of the form of the adjustmentcosts for the profitsndashleverage puzzle which is our focus Their focus is on thespeed of adjustment toward the target and on what that implies about thepersistence of leverageStrebulaev (2007) calibrates a dynamic trade-off model in which all costs

are proportional to the firm value A large firm is a rescaled version of asmall firm and so the target leverage ratio is independent of firm size Thisstructure is not designed to consider the differences between the financingactions of large versus small firms that we observe in the data Because he isstudying a trade-off model he finds that more profitable firms issue moredebt In simulated data Strebulaev (2007) shows that increases in profitabil-ity are correlated with future profitability which in turn increases the valueof the firm and lowers leverage (see p 1772) But expected future profit-ability increases do not always lead to current refinancing His cost structuredoes not lead to the under-adjustment phenomenon of the sort that wedocumentSeveral papers attribute the negative relation between leverage and prof-

itability to firms passively accumulating profits (see eg Fama and French(2002) and Kayhan and Titman (2007)) However the existing evidence isoften indirect and provides mix findings for this idea The implication that atthe time of rebalancing leverage should be positively related to profitabilityis also examined in several earlier studies Mackie-Mason (1990) shows thatcompanies with tax loss carryforwards are more likely to issue equity Incontrast Jung Kim and Stulz (1996) report finding no relation between thelikelihood of equity issuance and profitability Hovakimian Opler andTitman (2001) show that security issuance decisions of firms are afunction of the deviation from target leverage and other regressors includingfirm profitability They show that debt issuances are positively related toprofitability The present article while supporting aspects of their findingsexamines both issuing decisions and magnitudes as a function of firm sizeand profits In doing so we provide a more granular understanding of theeffect of profitability on changes in equity that derive from real operationsand changes in equity that are the result of active issuances and repurchasesof securities Furthermore we provide a sharply different interpretationalong with evidence in support of our interpretation We stress the

PROFITSndashLEVERAGE PUZZLEREVISITED 1419

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

distinction between automatic changes in equity that stem from operationsas opposed to actual issues or repurchases of equity This distinction is em-pirically importantThis article is organized as follows Section 2 provides theoretical motiv-

ation for our empirical strategy Section 3 describes the construction of thedata and variables and provides summary statistics Section 4 providesresults from the fixed effects estimates of leverage ratios Section 5presents the frequency and magnitudes of financing activity for sorts onfirm size and profits Section 6 provides the main results on debt andequity issuance regressions Section 7 explores the debt and equityissuance responses of firms during good and bad times Section 8examines the debt and equity issuances when scaled by total issuances andtotal capital Section 9 concludes the article

2 Empirical Strategy

The key idea in the trade-off theory is that firms choose capital structure bytrading off the various costs and benefits The considerations may includebankruptcy costs tax benefits agency costs transactions costs etc Differentversions of the theory include slightly different elements From our perspec-tive the leverage adjustment costs are of particular importanceA decision to issue debt or to redeem it or to issue shares or to repurchase

them requires managerial thought and consideration There are fixed organ-izational costs There is also the possibility of managerial inertia as docu-mented in Bertrand and Mullainathan (2003) These are fixed costs that areapproximately independent of the magnitude of the adjustments Adjustingalso has marginal costs Depending on which action is taken there may befees for underwriters bankers lawyers etc These may have a fixed compo-nent too Commonly they have a variable cost component which dependson the size of the transaction So both fixed and variable costs of adjustingleverage are normally involved The exact magnitudes vary from situation tosituation What is worse these are likely to have time-varying components4

Why does the form of the transactions costs matter Because the existenceof such fees implies that there will not be a unique leverage target This is anapplication of an idea that goes back to Constantinides and Richard (1978)and Constantinides (1979) It should be stressed that there is nothing newhere in terms of the formal structure This is just a reinterpretation of

4 If there were no time-varying component then it would be easy to identify the exact upperand lower targets for each firm In each case the firm would go exactly to the appropriatetarget In reality with time variation in adjustment costs such precision is not possible

1420 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

standard dynamic optimization conditions For a reader who wants a moreexplicit treatment of the mathematics we recommend Dixit (1993) The ob-servation that this mathematics can be interpreted in terms of corporateleverage in which firm size matters is also not novel It can be found inLeary and Roberts (2005) who provide a simulation with both fixed costsand a weakly convex adjustment costs What is new here is the analysis of theimplication for the impact of profits on leverageSuppose that without active rebalancing leverage would just wander

about due to changes in the stock market value of equity According tothe trade-off theory excessively high leverage is costly (bankruptcy andagency costs) as is excessively low leverage (lost tax benefits and free cashflow problems) If there are no transactions costs then under conventionaltrade-off theory assumptions there will be a unique leverage target that willbe maintained at all timesDue to the fixed costs such frequent rebalancing would be too costly As a

result the firm permits the leverage to drift However if the drift is tooextreme in either direction then the firm will actively rebalance in the direc-tion of the static optimum If the rebalanced firm (inclusive of rebalancingcosts) is worth more then rebalancing is worth it If the rebalanced firmwould be worth less then rebalancing is not worth it Rebalancing takesplace at the point at which the values just matchWhen the firm is rebalancing it also needs to worry about how far to

move in the desired direction When each step has costs the firm needs toensure that the marginal cost of the last step is just equal to the marginalbenefit As a result both from excessively high leverage and from excessivelylow leverage the movements are only part wayTo fix ideas consider an off-the-shelf model of the leverage ratio The

focus is on the impact of the adjustment cost structure The formal structureis essentially the cash management problem studied by Constantinides andRichard (1978) and Dixit (1993) We follow the presentation and notation inDixit (1993) very closely to show that this is just an application of well-established ideas to the leverage problemLet xt denote the leverage ratio at date t is the interest rate Ku is a fixed

cost of actively increasing leverage Kd is the fixed cost of actively reducingleverage mu is the marginal cost of each unit of active leverage increase md isthe marginal cost of each unit of active leverage decrease f(x) is the flowpayoff if leverage is at x F(x) is the expected present value of the payoff ifleverage is at x x is the lower adjustment threshold x is the lower targetleverage (when adjusting from below) x is the upper adjustment thresholdand x is the upper target leverage (when adjusting from above) Accordingto the trade-off theory F 0 gt 0 F 00 lt 0

PROFITSndashLEVERAGE PUZZLEREVISITED 1421

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

In the absence of active adjustments the leverage ratio xt follows astandard process

dx frac14 ethx tTHORNdtthorn ethx tTHORNdw

where is the trend coefficient is the volatility coefficient and w is anormal shock process with zero mean and unit variance If x lt x lt x thenfinancial inaction is optimal If x drops to x then it is reset to x and if xrises to x then it is reset to xThe trade-off theory is frequently thought of as a class of theories in which

firm value first increases in leverage and ultimately drops if leverage gets toohigh This can be motivated by taxes and bankruptcy or by various agencyconcerns A range of ideas can be allowed for when we write that the flowpayoff (dividends andor profits) is fethxtTHORN and it depends on the leverage statext

5 If the firm starts at x0frac14 x then F(x) is the expected present value of theflow payoffs The discount rate is which allows us to write

FethxTHORN frac14 E

Z 10

fethxtTHORNetdtjx0 frac14 x

To maximize the firm value it must choose a leverage policy that sayswhen to rebalance and how much to rebalance These are given by thefamous value matching and smooth pasting conditionsThe value matching conditions are as follows

FethxTHORN FethxTHORN frac14 Ku thornmuethx xTHORN

FethxTHORN FethxTHORN frac14 Kd thornmdethx xTHORN

The smooth pasting conditions are as follows

F 0ethxTHORN frac14 F 0ethxTHORN frac14 mu

F 0ethxTHORN frac14 F 0ethxTHORN frac14 md

As explained by Constantinides and Richard (1978) with standard as-sumptions about the shape of F there are a pair of lines at heights mu andmd such that they intersect the function F 0ethxTHORN in such a way as to define thepoints x lt x lt x lt xTo sum up in this trade-off model with both fixed and variable

rebalancing costs there are four critical leverage points x gt x gt x gt x

5 For simplicity we suppose that the flow payoff is a continuous function of leverageThere is no discrete exit or bankruptcy point This allows us to simply interpret theConstantinides and Richard model much as in Leary and Roberts (2005)

1422 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

If leverage is between x (upper boundary) and x (lower boundary) the firmdoes nothing active Changes in leverage reflect various shocks to revenuesand costs that hit the firm If leverage drifts up to x (or above) then the firmrebalances to x If leverage drifts down to x (or below) the firm rebalancesto xWhat does this imply empirically A firm that does nothing can be

anywhere within the interval between x and x A firm that is rebalancingis a firm that has just hit one of the outer boundaries How does profitabilityaffect the likelihood of hitting a boundary High profits increase the value ofequity This increases the denominator of the leverage ratio and thus makesit more likely that the firm hits the lower boundary Low profits reduce thevalue of equity and thus make it more likely that the firm hits the upperboundaryTo be more specific suppose that the costs are symmetric and the leverage

changes are also following an ordinary Brownian motion Then the median

inert firm will have leverage ofxthornx

2 A firm that just actively reduced leverage

will be at x gtxthornx

2 A firm that just actively increased leverage will be at

x ltxthornx

2 A firm that just reduced leverage will generally be a firm that had

very high leverage To the extent that this reflects a lack of activerefinancing such firm will have low value of equity The low value ofequity will generally reflect low profits A firm that just increased leveragewill generally be a firm that had very low leverage In the same way a firmthat just increased its leverage will tend to have had high value of equity dueto high operating profits6

In this setting it is natural to focus on firms that are actively adjusting andto compare them to each other and to firms that are inactive It is alsohelpful to look directly at financing decisions and profitability rather thanfocusing on a leverage ratioOur approach shares the perspective of Welch (2004) Leary and Roberts

(2005) and Strebulaev (2007) that it is very important to focus on activeadjustments by firms Welch (2004) transforms the data to put everythingback into a ratio form His ldquoimplied debt ratiosrdquo still have both debt andequity in the denominator Similarly Leary and Roberts (2005) andStrebulaev (2007) pay a great deal of attention to leverage ratios whenstudying financing decisions Strebulaev (2007) does not have fixed costssince he makes use of a rescaling property Empirically we find differentadjusting behavior by large and by small firms So the rescaling property is

6 If we knew the values of the critical leverage points then it would be possible to derivemore refined tests Unfortunately those critical values are not observable

PROFITSndashLEVERAGE PUZZLEREVISITED 1423

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

an issue of concern His model does not account for the negative relationshipbetween current leverage and past market-to-book ratioWe depart from the earlier literature by arguing that the use of leverage

ratios can be particularly misleading for profitability since the use of aleverage ratio causes identification problems It is hard to be sure whethera leverage change reflects changes in the numerator or the denominator Thisis particularly problematic when as here the rebalancing is partialIt should be noted that much empirical capital structure literature debates

about whether to use a book leverage ratio or a market leverage ratio Bookleverage is used based on the argument that managers borrow against assetsin place which are better reflected by book assets Others such as Welch(2004) argue that book values are essentially an accounting fiction and thatonly market leverage should count Empirical papers regularly attempt tosidestep this debate by carrying out tests on both book leverage and marketleverage Our point is that both kinds of leverage ratios are potentially mis-leading when examining the impact of profits on capital structure

3 Data

We use conventional data sources starting with the merged Compustat-CRSP data The data are annual and are converted into constant year2000 dollars using the GDP deflator We exclude financial companies (SIC6000-6999) firms involved in major mergers (Compustat footnote code AB)firms reporting financial data in currencies other than the US dollars andfirms with missing data on our key variables7 The ratio variables aretrimmed at the 1 level in both tails of the distribution This serves toremove outliers and the most extremely misrecorded data The finalsample consists of 179021 firm-year observations from 1971 to 2009Table I provides definitions of financial variables and reports summary

statistics The average debt (in constant US$) is about $653 million while themedian is $24 million A significant fraction of firms have zero debt (the 10thpercentile is 0) Book equity is slightly larger than book debt Market equityis almost three times larger than book debt Book assets average $2191million although the medians are considerably smallerIf issuing or retiring securities incurs no fixed costs then we would expect

to see many small actions and very few large actions (Leary and Roberts2005) If there were significant fixed costs involved in issuing or retiringoutstanding securities then small issues might not be worthwhile Table I

7 These include debt book value of equity market value of equity assets book andmarket leverage profitability market-to-book assets ratio and tangibility

1424 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

shows that although most firms issue little debt or equity in a given year theaverages are large In other words when firms actually enter debt and equitymarkets they intervene massivelyThe mean constant dollar debt issue is $165 million (the median is $2

million) In unreported tables we find that the average debt issue is about81 of assets (the median is 18) About 38 of the firms issue no debt8 issue between 0 and 1 of the value of their assets as debt another

Table I Data description

Data sources The sample comes from the annual Compustat files The sample period is

1971ndash2009 We exclude financial companies (SIC 6000-6999) firms involved in majormergers (Compustat footnote code AB) firms reporting financial data in currencies otherthan the US dollar and firms with missing data on the key variables All financial variables

are deflated to year 2000 using the GDP deflator N is the number of observations SD isthe standard deviation D BVE MVE Assets Debt issuance Equity issuance Debt re-payment Equity repayment are all measured in $ millions The ratio variables are trimmed

at the 1 level in both tails of the distribution

Variable definitions Debt (D)frac14 long-term debt (dltt)thorn short-term debt (dlc) book equity(BVE)frac14 common shareholder equity (ceq) market equity (MVE)frac14 no of outstanding

shares (csho) closing price ( prcc f ) assetsfrac14 book assets (at) debt issuancefrac14 issuance oflong-term debt (Maxethdltis 0THORN)thorn increase in current debt (Maxethdlcch 0THORN) equity issu-ancefrac14 sale of common stock (Maxethsstk 0THORN) debt repaymentfrac14 reduction of long-termdebt (Maxethdltr 0THORN)thorn decrease in current debt (Minethdlcch 0THORN) equity repurchasefrac14 purchase

of common stock (Maxethprstkc 0THORN) book leveragefrac14D(DthornBVE) market leveragefrac14D(DthornMVE) profitabilityfrac14EBITDA (oibdp)assets market-to-book ratio M

B frac14Marketvalue of assets (MVA)assets where MVAfrac14DthornMVEthorn preferred-liq value (pstkl)

deferred taxes (txditc) tangibilityfrac14 net property plant and equipment (ppent)assets

Variable N Mean SD

Distribution

10th 50th 90th

Debt (D) 179021 653 4592 0 24 1048

Book equity (BVE) 179021 815 4122 3 70 1373

Market equity (MVE) 179021 1784 9589 9 118 2698

Assets 179021 2191 11530 10 153 3575

Debt issuance 179021 165 1335 0 2 245

Equity issuance 179021 26 186 0 0 40

Debt repayment 179021 132 1123 0 3 181

Equity repurchase 179021 24 257 0 0 9

Book leverage 179021 036 034 000 031 072

Market leverage 179021 027 025 000 020 066

Profitability 179021 005 027 018 011 023MB 179021 165 206 052 102 324

Tangibility 179021 031 024 005 026 070

PROFITSndashLEVERAGE PUZZLEREVISITED 1425

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

16 issue between 1 and 5 of the value of their assets as debt and theremaining 38 issue debt in excess of 5 of the value of their assetsThe mean constant dollar equity issue is about $26 million (the median is

about $04 million) As a fraction of assets the mean and median equityissues are about 7 and 02 respectively About 33 of the firms issue noequity 34 of the firms issue between 0 and 1 of the value of theirassets as equity another 14 issue between 1 and 5 of the value of theirassets as equity and the remaining 19 of the firms issue equity that is inexcess of 5 of the value of their assets Average debt repayments are largerthan equity repurchases This perhaps reflects the finite maturity of debt andits contractual repayment The median firm does not repurchase equityWe construct both book and market leverage ratios Book leverage is

defined as debt over debt plus book equity Market leverage is defined asdebt over debt plus market equity8 The median book leverage is 036 (theaverage is 031) The median market leverage is 027 (the average is 020)Profitability is defined as the ratio of operating income before depreci-

ation to assets While the average firm is profitable (the ratio of EBITDA toassets is 005) the median firm is even more profitable (with a profitabilityratio of 011) But the sample also includes a large number of unprofitablefirms as the 10th percentile is 018 The table also reports descriptive stat-istics on the market-to-book ratio and the tangibility ratio The market-to-book ratio (MB) defined as the ratio of the market value of assets to bookassets averages at about 165 Tangibility defined as the ratio of netproperty plant and equipment to assets averages at about 31

4 Leverage Ratio Regressions

The previous literature focuses on estimates obtained using leverage ratiosHence we start with a similar estimation to check whether our results matchthose of previous studies Table II presents the results for book leverage Weobtain similar results for market leverage and they are presented inAppendix Table AI The regressions include leverage factors following theprevious capital structure literature (see eg Lemmon Roberts and Zender(2008) and Frank and Goyal (2009)) The factors include (i) profitability (ii)industry median leverage (iii) market-to-book assets ratio (iv) tangibility ofassets and (v) firm size Industry median leverage is constructed as themedian leverage of all other firms in the industry excluding the firm under

8 Welch (2011) stresses the idea that nonfinancial liabilities should not be implicitly mis-treated as if they were equity by paying excessively narrow attention to financial liabilitiesin a leverage ratio We have adopted his approach in the empirical work reported here

1426 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

consideration Firm size is defined as the natural log of assets Other factorsare defined in Section 3 We cluster standard errors by firm and estimatethese regressions both with and without fixed effectsIn Table II we run conventional cross-section leverage regressions Since

we employ the usual data it is not surprising that our results match thosereported in the existing capital structure literature Profitability has anegative sign in both the book leverage regressions and in the marketleverage regressions The coefficients on other factors largely match thosereported in earlier studies Firms operating in industries with high leverage

Table II Leverage and profitability

The table presents parameter estimates for book leverage regressions The sample consists

of nonfinancial companies in the annual Compustat files during the period 1971ndash2009 Theleverage is estimated as the ratio of debt over debt plus book equity The explanatoryvariables Profitabilityt1

MB

t1

Tangibilityt1 and LnethAssetsTHORNt1 are described in

Table I IndLevt1 is estimated as the median book leverage of all other firms in thesame industry The industry is defined at the level of the four-digit SIC code All specifi-cations include the year fixed effects The specifications in column (2) additionally include

the firm fixed effects We report t-statistics where the standard errors are clustered at thefirm level in parentheses

aSignificant at the 1 level

Quantile regressions

(1) (2) (3) (4) (5)

Profitabilityt1 026a 020a 011a 017a 041a

(293) (192) (378) (518) (1015)

IndLevt1 042a 012a 030a 052a 052a

(369) (107) (926) (1252) (950)MB

t1

002a 001a 001a 002a 002a

(203) (102) (392) (426) (418)

Tangibilityt1 018a 025a 024a 019a 012a

(206) (154) (964) (580) (276)

LnethAssetsTHORNt1 001a 002a 002a 002a 001a

(114) (76) (899) (478) (96)

Constant 012a 010a 012a 002a 029a

(156) (60) (284) (27) (387)

Year FE Yes Yes Yes Yes Yes

Firm FE No Yes No No No

Regression type OLS FE 25th tile Median 75thtile

Clustered SE Yes Yes No No No

R2 013 054

Pseudo R2 016 015 009

N 158824 158824 158824 158824 158824

PROFITSndashLEVERAGE PUZZLEREVISITED 1427

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

tend to have high leverage A higher market-to-book ratio is associated withlower leverage Larger firms are typically more highly levered The coeffi-cients on tangibility are positive The conclusion from Table II is clear Theleverage regression results match those reported in previous studiesAt times one may be concerned that in a regression what is being

estimated is a conditional mean while some parts of the distributionmight behave differently from other parts If so then ldquoaccidentsrdquo likesample selection issues or asymmetries in the underlying distributionsmight play a greater role than is usually recognized In other words thesource of identification could be misinterpreted To guard against this po-tential problem we estimate the same model using quantile regressions9

Columns 3 4 and 5 report what happens at the 25th 50th and 75th per-centiles The basic model results are robust in terms of both the sign and thestatistical significanceThe impact of profits does appear to be somewhat stronger among the

high leverage firms For instance in a book leverage regression the coeffi-cient on profits for the first quartile is 011 with a t-ratio of 378 For thethird quartile the coefficient on profits is 041 with a t-ratio of 1015This difference is curious and might deserve further study However the factthat in each case we get the negative sign and statistical significance is suf-ficient for the purposes of the present article So the results on profits arerather robust across the distributionTable III sorts firms into those that are actively adjusting their leverage

and those that are passive In the presence of both fixed and variable trans-acts costs this distinction is important We expect the coefficient on profit-ability in leverage ratio regressions to be more negative in firms that arepassive and less negative in firms that are actively adjustingWe define active firms as those that are either issuing net debt or net

equity in excess of 5 of their assets Passive firms are those that neitherissue debt nor equity in nontrivial amounts (ie greater than 5 of theirassets) Consistent with our expectations we find in Table III that the coef-ficients on profits are significantly more negative for passive firms than theyare for active firms With firm and year fixed effects the coefficient onprofits is 008 for active firms and 017 for the passive firms The twocoefficients are significantly different from each other (plt 001) InTable III the standard control variables are used They have the usualsigns and for the most part do not differ much between the active and thepassive firms

9 Cameron and Trivedi (2010) provide an extensive discussion of quantile regressions inStata

1428 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

5 Financing Activity

51 FREQUENCY OF DEBT AND EQUITY ISSUES

Table IV reports the percentage of firms issuing or repurchasing debt orequity for annual sorts based on lagged profitability In Panel A weemploy the conventional 5 cutoff rule to exclude minor fluctuations anddefine firms ldquoissuing debtrdquo as those that issue debt (both short-term andlong-term) in excess of 5 of the value of their assets Other decisions aresimilarly defined using a 5 cutoff Appendix Table AII examines the dis-tributional properties across profitability (and size) quintiles for ldquolargerdquo is-suances by increasing the 5 cutoff to higher levels such as 10 and 15How does security issuance behavior vary with firm profitability To

answer this question Table IV sorts firms by profitability and reports the

Table III Leverage and profitability for active and passive firms

The table presents estimates of leverage on firm characteristics for the subsamples of firms

that are active in either debt or equity markets Active firms are defined as those that issueeither net debt in excess of 5 of the value of assets or issue net equity in excess of 5 ofthe value of their assets The passive firms do not issue significant (more than 5) of either

debt or equity The sample comes from the annual Compustat files during the period 1971ndash2009 Financial firms are excluded The variables are described in Tables I and II Wereport t-statistics where the standard errors are clustered at the firm level in parentheses

aSignificant at the 1 level

Active Passive

(1) (2) (3) (4)

Profitabilityt1 006a 008a 022a 017a

(126) (148) (254) (188)

IndLevt1 044a 019a 041a 013a

(424) (149) (326) (127)MB

t1

002a 001a 003a 001a

(362) (196) (327) (172)

Tangibilityt1 010a 014a 014a 019a

(146) (118) (145) (130)

LnethAssetsTHORNt1 001a 004a 001a 004a

(121) (179) (147) (169)

Constant 012a 008a 003a 007a

(120) (57) (45) (42)

Year FE Yes Yes Yes Yes

Firm FE No Yes No Yes

R2 034 064 033 075

N 43395 43395 58943 58943

PROFITSndashLEVERAGE PUZZLEREVISITED 1429

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

percentage of firms issuing or repurchasing debt or equity The table alsoreports test statistics from tests that examine whether the percentage differ-ences between the high- and low-profit firms are significantly different fromeach other The results in Column (1) show that the likelihood of issuing debt

Table IV Profitability sorts for debt and equity issuers

The sample contains nonfinancial firms listed on the annual Compustat files for the period

from 1971 to 2009 The table presents the percentage of firms issuing and retiring (orrepurchasing) debt and equity The firms are classified as ldquoissuing debtrdquo if they issuedebt in excess of 5 of the value of their assets as ldquoissuing equityrdquo if they issue equity

in excess of 5 of the value of their assets as ldquoretiring debtrdquo if they retire debt in excess of5 of the value of their assets and as ldquorepurchasing equityrdquo if they repurchase equity inexcess of 5 of the value of their assets In addition we also report net debt issuers which

are firms that issue net debt over 5 of the value of their assets and net equity issuerswhich are firms that issue net equity in excess of 5 of the value of their assets Weannually sort firms on lagged profitability and report the percentage of firms in each ofthese categories The bottom part of each panel reports the percentage of firms issuing or

retiring securities by profitability within the smallest and largest asset quintiles The tablealso reports results from t-test for the difference in percentage of firms issuing or retiringsecurities in the lowest and the highest profit quintile

Percentage of firms

Iss Ret Iss Iss Rep Iss Both IssD IssE

D D NetD E E NetE DampE RepE RetD

(1) (2) (3) (4) (5) (6) (7) (8) (9)

Allfirm

s

Low 276 245 186 301 13 297 71 01 38

2 319 299 181 117 21 112 27 02 20

3 385 344 194 86 29 80 21 05 15

4 386 330 213 81 43 75 20 07 14

High 329 273 198 92 82 81 17 13 14

t Highfrac14Low 150 84 39 723 414 763 358 176 200

Smallfirm

s Low 275 204 201 411 11 405 109 01 47

2 291 233 213 379 10 378 99 01 45

3 281 246 204 265 14 261 66 01 41

4 275 272 186 162 15 159 34 02 35

High 275 281 178 136 27 131 25 02 26

t Highfrac14Low 01 84 28 316 55 318 171 18 55

Largefirm

s

Low 414 363 177 87 22 79 28 03 11

2 449 379 180 62 30 57 17 07 09

3 439 360 186 56 41 48 16 09 07

4 403 320 194 46 67 39 13 12 06

High 374 264 207 50 129 35 10 28 05

t Highfrac14Low 50 130 45 89 244 116 77 120 47

1430 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

is mostly independent of firm profitability In the aggregate there is a smalldifference but it is sensitive to sorting on firm size Similarly in Column (2)there is only a weak relation between the profitability and the likelihood thatthe firm retires debt To a first approximation the fraction of firms issuing orrepurchasing debt is largely independent of profits Large firms are somewhatmore likely than small firms to issue debt and also to retire debtUnlike debt issues the probability of issuing equity is strongly related to

profitability as can be seen in Column (4) Low-profit firms are much morelikely to issue equity than are high-profit firms In the lowest profit quintile301 of the firms issued equity In the highest profit quintile it was only92 The lower panel reports separate sorts for small and large firms Similarpatterns are found for both groups of firms that is the likelihood of equityissuance declines as profitability increases The differences between the quin-tiles are quite stark for the group of small firms All quintiles of small firmsissue equity much more frequently than the corresponding large firm quintilesHigh-profit firms are much more likely to repurchase equity as shown in

Column (5) In the overall sample only 13 of low-profit firmsrepurchased equity compared with 82 of the high-profit firms Whenwe separate the small and large firms again a difference emerges Ingeneral small firms do not repurchase much equity while a greaterfraction of the large firms do soGiven these facts it is natural that low profitability firms are much more

likely to be net issuers of equity than are high profitability firms as shown inColumn (6) Among the firms in the lowest quintiles of profits roughly 30of firms issue net equity that exceeds 5 percent of their assets In contrastfor the most profitable quintile only about 8 issue net equityFirm size is an important variable in the recent literaturemdashit is often used

as a proxy for access to capital markets (as in Faulkender and Petersen(2006) and Leary (2009)) Small firms are bank-dependent risky and infor-mationally opaque They have restricted access to public debt markets andconsequently face more severe supply constraints in their ability to issuedebt Thus we expect small firms to be more sluggish in adjusting theirdebt and equity in response to shocks to profitability Large firms incontrast have much easier access to public debt markets and they facefewer obstacles in accessing securities marketsTo examine how size and profitability interact the bottom part of Table IV

examines profitability sorts for the smallest and largest firms We first sortfirms annually by firm size and then within each size quintile we sort them onprofitability We do these two-way sorts to ensure that we have similarnumber of firms in profitability subgroups for both small and large firmsamples Unconditionally small firms are generally less profitable while

PROFITSndashLEVERAGE PUZZLEREVISITED 1431

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

large firms are relatively more profitable The two-way sorts allow us to havea uniform distribution of firms in various profitability quintiles for both smalland large firm samples The bottom part of Table IV reports results for thesmallest and the largest size quintiles Among the small firms there is littlerelation between profitability and the likelihood of issuing debt Howeverlarger firms exhibit a small increase in the likelihood of net issuance of debtwith increasing profitabilityIn contrast to debt the effects of profitability on equity issuance and

repurchases are much more consistent across size quintiles As profitabilityincreases firms are generally less likely to issue equity and more likely torepurchase it Across the two extreme size groups we note a striking differ-ence between the proportion of firms issuing equity and those issuing debtAmong low-profit small firms almost 41 are equity issuers whereas onlyabout 9 of low-profit large firms are equity issuers However regardless ofsize we note a monotonic reduction in the likelihood of issuing equity asprofitability increasesColumn (7) reports the fraction of firms that issue both debt and equity in

excess of 5 of assets both as a function of profits in the upper panel andas a function of firm size and profits in the lower panel In the upper panelwe see that low-profit firms are much more likely to issue both debt andequity (71) while the high-profit firms are less likely to do so (17) Thelower panel shows that small firms are much more likely to be issuers of bothin contrast to large firms In both firm size categories the profit pattern is thesame with low profits being more frequent issuers of both simultaneouslyImportantly Column (8) shows that the likelihood of issuing debt and

simultaneously repurchasing equity increases with profitability Converselyin Column (9) we find that the likelihood of doing the reverse that is issuingequity and retiring debt declines with profitability The effects of sorting onfirm size and firm profitability mirror those reported for all firms Overallthe results show that low-profit firms are less likely to issue debt and repur-chase equity they are instead more likely to issue equity and retire debtFirms with high profitability exhibit the reverse patternIn Appendix Table AII we present results from profitability (and size)

quintiles for large issuances defined using the 10 cutoff (in Panel A) andthe 15 cutoff (in Panel B) The results from these panels are largely con-sistent with those described above

52 MAGNITUDES OF ISSUANCES AND EQUITY CHANGES

In the previous section we considered the probability of having a nontriviallevel of debt or equity activity The next question is how large are the dollar

1432 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

values involved In Table V we sort firms according to profits and thentabulate the levels and changes in both debt and equity We do this firstfor all firms and then for small versus large firmsFor the analysis on all firms we observe in Column (1) that debt peaks at

the middle of the distribution This is because firms with medium profitabil-ity are also the largest as seen in Column (8) which reports average assetvalues for different profitability quintiles When we sort first by firm size andthen examine firms with different levels of profitability within size quintileswe find that for small firms debt is roughly independent of profits while forlarge firms less profitable firms have a higher level of debtAs expected Columns (3) and (5) show that more profitable firms have

higher equity values Columns (2) and (7) consider the relationship betweenissuances of debt and equity and firm profitability Debt issuances are sig-nificantly larger for more profitable firms In contrast equity issuances arelarger for low-profit firms Small low-profit firms issue substantially lessequity compared with small high-profit firms Among large firms the low-profit ones tend to issue equity while the high-profit ones tend to repurchaseequity The fact that more profitable firms issue debt and repurchase equitywhile the least profitable firms retire debt and issue equity is consistent withthe predicted relation between profitability and financing decisions under thetrade-off theoryWe also examine two-way sorts by size and profitability and report results

for the smallest and largest quintile of firms For the smallest quintile offirms the change in debt is largely unrelated to profits But for large firmsthere is a positive relation between profits and debt issuances High-profitfirms have a big positive change in debt Low-profit large firms have anegative change in debtColumns (4) and (6) provide an explanation of why the leverage ratio

regression results contradict those from the basic profitability sorts pre-sented here As we can see profitability indirectly affects leverage ratiosby increasing equity values Changes in both the book value of equity andthe market value of equity are positive and large for highly profitable firmsIn contrast these changes are negative for less profitable firms This evidenceillustrates an important issue concerning the use of leverage ratios Suchratios are often interpreted as essentially reflecting the use of debt by thefirm This interpretation while common is empirically misleadingFor the typical firm the change in the value of equity is larger than the

change in debt For example in the third profit quintile for large firmscolumn (7) the mean equity issue is just $9 million but the change in themarket value of equity is $242 million At the same time the mean change indebt is $27 million This suggests that a fair bit of the observed variation in

PROFITSndashLEVERAGE PUZZLEREVISITED 1433

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

the leverage ratios is primarily driven by the changes in the market value ofequity in the denominator rather than by the changes in debt in the numer-ator Since equity issues are often small this implies that the variation in theleverage ratio is primarily driven by internal operations rather than byexternal financing actions This again points to the fact that leverageratios can provide a misleading account of actual patterns in the dataThere is always a potential concern that the averages may be misleading

due to the impact of outliers To address this concern in AppendixTable AIII we present median values of the profitability sorts Theseresults generally reinforce the findings in Table VTable VI again considers the magnitudes of financing activity but this

time the issues are scaled by a measure of firm size That way it is easier to

Table V Magnitude of financing activity

The sample contains nonfinancial firms listed on the annual Compustat files for the period

from 1971 to 2009 The table presents average debt and equity changes in debt and equityand equity issuances (in millions of dollars) for firms sorted on profitability within sizeclasses The table reports information for all firms sorted on profitability and for profit-

ability sorts within the smallest and largest firms The sorts are done annually The tablealso reports results from t-test for the difference in percentage of firms issuing or retiringsecurities in the lowest and the highest profit quintile

D D BVE BVE MVE MVE EqyIss Assets

(1) (2) (3) (4) (5) (6) (7) (8)

Allfirm

s

Low 83 5 78 10 120 5 7 234

2 605 13 509 24 721 8 17 1691

3 1227 9 1123 22 1834 3 9 3551

4 833 23 1166 7 2408 74 9 2994

High 570 53 1171 71 3817 68 36 2581

t Highfrac14Low 31 9 36 9 38 2 16 36

Smallfirm

s Low 25 04 42 01 332 28 33 97

2 29 06 55 02 278 16 21 116

3 31 07 63 01 212 05 12 127

4 30 06 70 05 179 10 08 136

High 24 06 84 14 234 17 07 142

t Highfrac14Low 04 15 153 85 73 11 201 107

Largefirm

s Low 3929 55 2990 152 4671 11 63 10846

2 3826 19 3178 95 5122 66 11 10721

3 3074 27 3498 41 6173 242 9 9697

4 2623 62 3915 45 8473 94 57 9855

High 2180 193 4642 261 14924 202 148 10171

t Highfrac14Low 94 64 108 87 246 12 169 16

1434 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

see whether large firms behave disproportionately when compared withsmall firms Column (1) shows that low-profit firms have more debtas reported in many previous studies In Columns (4) and (6) we seethat low-profit firms have a major decline in both the book value ofequity and the market value of equity The high-profit firms have thereverse experienceMore interesting are the active decisions of the firms In Column (2) we see

that in contrast to the usual impression the low-profit firms are reducingdebt while the high-profit firms are increasing it In Column (7) we see thatthe low-profit firms are also reducing leverage by issuing equity The high-profit firms issue little equity In the lower panel similar results are foundwhen we separate small and large firms Equity issuing is particularly strong

Table VI Magnitude of financing activitymdashscaled levels and changes

The sample contains nonfinancial firms listed on the annual Compustat files for the period

from 1971 to 2009 The table presents average debt and equity changes in debt and equityand equity issuances scaled by assets for firms sorted on profitability within size classes Thetable reports information for all firms sorted on profitability and for profitability sorts

within the smallest and largest firms The sorts are done annually The table also reportsresults from t-test for the difference in percentage of firms issuing or retiring securities inthe lowest and the highest profit quintile

D D BVE BVE MVE MVE EqyIss

(1) (2) (3) (4) (5) (6) (7)

Allfirm

s

Low 386 122 12 871 3427 1868 217

2 285 58 407 119 940 157 28

3 293 12 425 14 742 03 12

4 269 26 461 02 955 27 10

High 192 15 551 47 1675 74 07

t Highfrac14Low 35 49 50 47 127 42 157

Smallfirm

s Low 545 154 841 2145 9688 9128 696

2 684 49 712 1915 5014 2318 288

3 331 253 141 969 3063 954 129

4 267 181 357 175 1673 221 55

High 194 05 543 26 1758 147 41

t Highfrac14Low 49 17 51 36 88 27 99

Largefirm

s Low 350 30 306 58 511 55 09

2 346 14 335 13 523 02 04

3 353 165 332 71 711 29 01

4 280 01 414 04 909 09 05

High 227 15 477 19 1643 27 14

t Highfrac14Low 371 112 453 181 570 57 221

PROFITSndashLEVERAGE PUZZLEREVISITED 1435

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

among the small low-profit firms that are also experiencing sharp reductionsin the value of equity In column (2) we see that both for the large and for thesmall firms net debt increases are primarily found in the highest quintile offirm profitability

6 Debt and Equity Regressions

So far we have demonstrated that in our data the conventional leverageregressions have the usual signs We then demonstrated that in simple sortsof the data strikingly different active financing patterns emerge The nexttask is to reconcile these differencesTable VII presents simple regressions predicting changes in debt book

equity and market equity as well as equity issuances Consistent with thesorts we control for changes in firm size and we also include firm and yeardummies There may be a concern about the speed of a firmrsquos responsesAccordingly we consider the lagged change in profits as the main case butwe also include the contemporaneous change in profits Whether the contem-poraneous change in profits is included or not does not in fact matter Wereport panel-robust standard errors adjusted for clustering at the firm levelThe results in Table VII support the findings in the sorts Firms that have

an increase in profits in 1 year increase debt both in the same year and in thenext year Firms that have an increase in profits experience an increase inboth the book value of equity and the market value of equity Importantlyhowever firms that experience and increase in profits reduce their equityissues All these effects are statistically significant (except for the marketvalue of equity in the column 5 specification) and seem intuitively reason-able In Appendix Table AIV similar regressions to those presented aboveare reported but without firm-fixed effects We find that empirically verylittle changes whether we include firm fixed effects or leave them out Thusthe targeting behavior to the extent that it happens does not account forwhat we are reportingThe next step is to bring the sorts and the simple regressions together into

a common setting We therefore include factors in addition to profitability(i) median industry leverage (ii) the market-to-book assets ratio (iii) tangi-bility of assets and (iv) firm size (measured by the log of assets) Rajan andZingales (1995) show that these factors are related to leverage in G7countries A number of studies have used these factors to estimateleverage targets Frank and Goyal (2009) show that these factors arerobustly related to leverage in the USA In these regressions we use indica-tors for the quintile that the firm is in for each factor

1436 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

Table VIII reports these results In Columns (1) and (2) we examine debtissues and again show that profits positively affect debt issuances The effectsare large and statistically significant at the 1 level Columns (3) and (4)which examine book equity show a strong effect of profits This effect is onlyslightly reduced by the inclusion of other factors Columns (5) and (6) illus-trate that the impact of profits on the change in market equity is also robustto the inclusion of the conventional factors Columns (7) and (8) examineequity issuances Here we again find that profits have a significant negativeimpact on equity issuances As before the results are robust to the inclusionof the conventional factorsIn unreported tables we estimate debt and equity changes using quantile

regressions In terms of the signs and significance of the profitability variable

Table VII Debt and equity changes

The sample contains nonfinancial firms listed on the annual Compustat files for the period

from 1971 to 2009 The table presents fixed effect estimates from regressions of changes indebt (D) changes in the book value of equity (BVE) changes in the market value ofequity (MVE) and net equity issuances (EqyIss) The explanatory variables include

current and lagged change in operating income before depreciation (Profits andProfitst1) and lagged change in the book value of assets (Assetst1) In addition tofixed firm effects the regressions include year indicator variables The reported t-statistics

are corrected for clustering at the firm level

aMean significant at the 1 level

bMean significant at the 5 level

cMean significant at the 10 level

D BVE MVE EqyIss

(1) (2) (3) (4) (5) (6) (7) (8)

Profitst1 022a 025a 015b 035a 004 060b 005b 007a

(33) (31) (25) (35) (03) (25) (25) (53)

Profits 050b 078a 140a 003a

(24) (32) (48) (34)

Assetst1 000 007b 025a 001b

(00) (21) (35) (25)

Constant 167 283a 34 51 1145a 1182a 605a 595a

(14) (30) (04) (05) (32) (29) (69) (74)

Year FE Yes Yes Yes Yes Yes Yes Yes Yes

Firm FE Yes Yes Yes Yes Yes Yes Yes Yes

R2 001 004 001 010 001 005 002 003

N 162056 162056 162130 162130 157550 157550 162154 162154

PROFITSndashLEVERAGE PUZZLEREVISITED 1437

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

the results are qualitatively similar to the OLS results reported in Table VIIIProfitability positively affects debt issuances and negatively affects equityissuances across various quantiles Overall the results in Table VIII showthat there is nothing that is special about using sorts or regressions to explaindebt and equity issues In either case we find that more profitable firms tendto increase their debt experience an increase in the value of equity andrepurchase shares Thus the control factors are not responsible for theusual rejection of the trade-off theory It would appear that what mattersis the size of the issuing activity relative to the organic increases in equityvalue due to profits In other words what matters in the leverage regressionseems to be coming from the presence of E in D

DthornE and not from D

Table VIII Debt and equity issuances profitability and leverage factor quintiles

The sample contains nonfinancial firms listed on the annual Compustat files for the period

from 1971 to 2009 The table presents estimates from regressions of change in debt (D)change in the book value of equity (BVE) change in the market value of equity (MVE)and net equity issuances (EquityIss) The explanatory variables include lagged profitability

quintiles and lagged leverage factor quintiles In addition to fixed firm effects the regres-sions include year indicator variables The reported t-statistics are corrected for clustering atthe firm level

aMean significant at the 1 level

bMean significant at the 5 level

cMean significant at the 10 level

D BVE MVE EqyIss

(1) (2) (3) (4) (5) (6) (7) (8)

ProfitabilityQint 214a 167a 283a 175a 80 429a 53a 55a

(70) (54) (66) (59) (08) (48) (55) (62)

IndLevQint 104a 43 99 20

(29) (15) (10) (10)

Ln(Assets)Qint52 17 1026a 53a

(16) (05) (77) (35)MB

Qint158a 372a 1155a 09

(45) (56) (106) (06)

TangibilityQint 05 177 436c 64a

(01) (15) (18) (31)

Constant 167 96 656b 235 4613a 8610a 113 570a

(08) (04) (26) (04) (61) (58) (11) (31)

Year FE Yes Yes Yes Yes Yes Yes Yes Yes

Firm FE Yes Yes Yes Yes Yes Yes Yes Yes

R2 0002 0002 0003 0004 0013 0014 0013 0014

N 155587 155037 155587 155037 155587 155037 155587 155037

1438 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

7 Are Financial Market Conditions Important

According to the trade-off theory capital structure is determined by con-sidering the impact of costs and benefits of debt The time variation inthe costs and benefits of debt imply that issuance decisions would alsovary over time Furthermore market conditions affect the transactioncosts of adjusting Adverse selection is a greater problem in a cold marketthan in a hot one Accordingly it may be easier to adjust in good marketconditions Related ideas have been developed by Huang and Ritter(2009) Faulkender et al (2012) and Halling Yu and Zechner (2012)Faulkender Flannery Hankins and Smithrsquos paper examines cross-sectionalvariation in adjustment speeds including market valuations andhow they may affect adjustment speeds It is natural to think that ingood market conditions it will be less costly to issue both debt and equityand adjusting leverage to respond to profitability shocks would be mucheasierTo test the importance of market conditions we require a definition of

good times and bad times Our empirical strategy is to estimate good timesversus bad times at the four-digit industry level We define an industry ashaving ldquogood timesrdquo if the median firm in that industry has a market-to-book ratio that is higher than the 67th percentile of the time-series distribu-tion of the industry median market-to-book ratios Conversely an industryis defined as having ldquobad timesrdquo if the median firm in the industry has amarket-to-book ratio that is lower than the 33rd percentile of the time-seriesdistribution of the industry median market-to-book ratiosPanel A of Table IX tabulates issuance activity for profitability sorts in

both good times and bad times As might be expected active debt and equityissuances are larger during good times There is also somewhat more activeswaps between debt and equity during good timesPanel B reports the issuance activity for the smallest and the largest firms

during the good times and bad times Within each size quintile firms aresorted on profitability This is a three-way sort As before we find that thereis much more active use of external markets during good times Small low-profit firms are more likely to issue equity in good times than in bad timesDebt issuances are significantly higher in good times Large high-profitfirms are significantly more likely to issue debt and repurchase equity ingood times than in bad timesPanel C reports the magnitudes of the financial variables rather than the

frequencies During bad times less profitable large firms retire substantialamounts of debt and they show a tendency to issue equity Small firms donot seem to engage in similar debt reduction activities However like the

PROFITSndashLEVERAGE PUZZLEREVISITED 1439

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

Table IX Debt and equity issues in good and bad times

The table reports the frequency and magnitude of financing activity for sorts on profits for subsamples

of firms in good and bad times An industry is defined as having ldquogood timesrdquo if the median firm in

that industry has a market-to-book ratio that is higher than the 67th percentile of the time-series

distribution of industry median market-to-book ratios Conversely an industry is defined as having

bad times if the median firm in that industry has a market-to-book ratio that is lower than the 33rd

percentile of the time-series distribution of industry median market-to-book ratios Panels A and B

report the percentage of firms (i) issuing net debt in excess of 5 of the value of their assets (ii) issuing

net equity in excess of 5 of the value of their assets (iii) issuing debt and repurchasing equity both in

excess of 5 of the value of their assets and (iv) issuing equity and retiring debt both in excess of 5

of the value of their assets Panel C tabulates the mean debt and equity levels and the mean changes in

debt and equity issuances The sample contains nonfinancial firms listed on the annual Compustat files

for the period from 1971 to 2009

Iss Iss Iss D Iss ENet D Net E Rep E Ret D(1) (2) (3) (4)

Panel A sorts on profitability

Badtimes

Low 145 131 01 202 170 49 02 083 173 46 04 074 182 47 03 10

High 177 51 06 10t Highfrac14Low 41 136 33 01

Goodtimes

Low 206 381 01 462 197 148 02 233 212 103 04 184 234 92 10 18

High 216 92 16 14t Highfrac14Low 23 658 138 23

Panel B sorts on firm size and profitability

Badtimes

Smallfirm

s Low 151 318 00 312 182 276 02 253 151 140 02 294 183 73 00 19

High 177 80 00 11t Highfrac14Low 10 90 NA 20

Largefirm

s Low 172 54 01 042 148 39 04 053 150 43 04 044 156 33 03 06

High 147 28 05 03t Highfrac14Low 17 32 20 02

Goodtimes Smallfirm

s Low 213 448 01 542 230 443 01 473 222 325 00 474 203 197 02 38

High 190 153 02 31t Highfrac14Low 19 230 10 39

Largefirm

s Low 203 96 03 122 208 66 09 133 232 54 16 064 215 39 19 05

High 234 36 38 04t Highfrac14Low 29 99 92 38

(continued)

1440 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

larger firms there is a tendency for small firms to issue more equitymdashat leastto the extent possibleDuring good times large profitable firms raise significant amounts of debt

as they experience an increase in the value of their equity Such firms alsoengage in share repurchases Debt issues by smaller firms are much lessaffected by their own profitability during good times and their change indebt is more or less independent of their profits However equity issuanceis primarily found among the low-profit small firms during good timesTable IX shows that empirically profitability has a time-varying impacton leverage choicesIn Table X the basic regressions for changes in debt and for equity

issuance are presented for both good times and bad times Comparinggood times with bad times we see that the effects of profits are muchstronger in good times In bad times the impacts are rather weak statisticallyThis difference is interesting and deserves further study However for ourpurposes recall that the static trade-off theory implies time-varying capital

Table IX Continued

D D BVE BVE MVE MVE EqyIss Assets

(1) (2) (3) (4) (5) (6) (7) (8)

Panel C magnitude of financing activity

Badtimes Smallfirm

s Low 40 03 58 14 305 06 22 140

2 44 05 82 07 240 67 17 180

3 45 04 94 02 176 18 05 187

4 43 09 98 02 137 04 03 188

High 36 10 114 13 209 16 04 199

Largefirm

s Low 3365 73 2523 169 2304 163 94 8972

2 3031 95 2983 35 3284 278 56 9068

3 2930 5 3360 14 3947 308 57 9225

4 2281 6 3817 19 5213 442 7 8902

High 1790 62 4767 13 8799 58 31 9573

Goodtimes

Smallfirm

s Low 24 06 38 05 339 23 39 90

2 26 08 48 01 301 04 25 104

3 30 09 60 04 246 03 17 125

4 28 07 66 08 206 18 12 130

High 23 06 85 16 261 13 08 141

Largefirm

s Low 4532 36 3477 103 6557 260 53 12465

2 4580 58 3527 139 6911 421 23 12522

3 2952 111 3485 2 7586 279 32 9535

4 2752 124 3939 72 10686 33 105 10226

High 2429 228 4819 377 18565 268 222 10830

PROFITSndashLEVERAGE PUZZLEREVISITED 1441

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

structure choices even when the target is time invariant Empirically we doobserve time-varying choices

8 Is Scaling Important

In the preceding section we provided results for unscaled debt and equityissues This seems appropriate to us However it involves making twochanges to the standard regression We would like to determine if both

Table X Debt and equity changes in good and bad timesmdashregressions

The sample contains nonfinancial firms listed on the annual Compustat files for the period

from 1971 to 2009 The table presents estimates from regressions of change in debt (D)and net equity issuances (EqyIss) in good and bad times The explanatory variables includechange in profits lagged change in profits and lagged change in assets The regressions

include both firm and year fixed effects An industry is defined as having ldquogood timesrdquo ifthe median firm in that industry has a market-to-book ratio that is higher than the 67thpercentile of the time-series distribution of industry median market-to-book ratios

Conversely an industry is defined as having bad times if the median firm in thatindustry has a market-to-book ratio that is lower than the 33rd percentile of the time-series distribution of industry median market-to-book ratios The reported t-statistics arecorrected for clustering at the firm level

aMean significant at the 1 level

bMean significant at the 5 level

cMean significant at the 10 level

Good times Bad times

D EqyIss D EqyIss

(1) (2) (3) (4) (5) (6) (7) (8)

Profits 087a 091a 006b 009a 060 065 003 003b

(30) (29) (20) (28) (15) (16) (16) (20)

Profitst1 029c 015a 034c 001

(19) (37) (17) (04)

Assetst1 000 002b 000 001

(02) (22) (00) (07)

Constant 1337 3989 8691a 10001a 1351 1851 4099a 4055a

(03) (09) (55) (37) (09) (13) (80) (78)

Year FE Yes Yes Yes Yes Yes Yes Yes Yes

Firm FE Yes Yes Yes Yes Yes Yes Yes Yes

R2 0071 0080 0018 0045 0178 0212 0034 0041

N 74827 68729 74891 68787 26965 26013 26971 26018

1442 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

changes are crucial Accordingly in this section we take a step back toconsider prior literature in which variables are scaled The questions arehow does scaling affect the results and whyTable XI examines alternative scalings Columns (1) and (2) scale debt and

equity issuances by the total firm net issuance Columns (3) and (4) scaledebt and equity issuances by the sum of debt and book equity Columns (5)and (6) consider the change in debt scaled by the sum of debt and bookequity Columns (7) and (8) consider the change in leverage due to externalfinancing Since we are examining changes we use the changes in thestandard regression factors as controls The primary interest is in the coef-ficient on the change in profitsWe find that a change in profits is associated with an increase in debt

issues But the sign reverses when we scale the debt issues by the sum of debtplus book equity This is telling What is really driving the results is thechange in equity in the denominator not the change in debt in the numer-ator This result is further substantiated in Columns (5) and (6) where thechange in debt to capital ratio is used instead of debt issuanceTo further explore the impact of active issuance relative to passive

experiencing of changes we decompose leverage by separating out theeffect of debt and equity issues from those due to changes in retainedearnings Thus we define leverage change due to external financing as

LevEF frac14Dt1 thorn d

Dt1 thorn Et1 thorn dthorn e

Dt1

Dt1 thorn Et1 eth1THORN

where D is debt E is book value of equity d is debt issues net of retirementsand e is equity issues net of repurchases Garvey and Hanka (1999) use asimilar measure in their study This measure directly captures the netleverage effect of debt and equity issuancesColumns (7) and (8) report results from regressions of the net leverage

effect of debt and equity issues on changes in profitability after controllingfor other leverage factors in differences and year indicator variables Theeffect of profitability on leverage change due to external financing is positiveand statistically significant This comparison suggests that changes in totaldebt and equity are only partly a result of debt and equity issuance decisionsOther balance sheet adjustments complicate the inferences from leverageratio regressionsFor most of our analysis we follow the literature in treating debt as homo-

genous It is of course potentially interesting to ask what happens when wedistinguish between short-term debt and long-term debt Which part of thefirmrsquos debt maturity structure drives the under-adjustment Table XII

PROFITSndashLEVERAGE PUZZLEREVISITED 1443

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

provides results from such decompositions Alternative scalings areconsidered The main results of interest are that the distinction betweenshort- and long-term debt is important Short-term debt issuance is largelyindependent of changes in profits The negative sign on profit change stemsfrom the action of long-term debt This seems intuitively reasonable

Table XI Scaled issuances

The sample contains nonfinancial firms listed on the annual Compustat files for the period

from 1971 to 2009 The table presents estimates from regressions of debt issuance to totalissuance (in columns (1) and (2)) debt issuance to total capitalization (in columns (3) and(4)) change in debt to capitalization ratio (in columns (5) and (6)) and change in leverage

ratio considering only external financing (in columns (7) and (8)) The change in leverageratio considering only external financing is defined as

LevEF frac14Dt1 thorn d

Dt1 thorn Et1 thorn dthorn e

Dt1

Dt1 thorn Et1

where D is debt E is book value of equity d is debt issues net of retirements and e is equity

issues net of repurchases The regressions include lagged leverage factors and year indicatorvariables Some of the specifications include firm fixed effects The reported t-statistics arecorrected for clustering at the firm level

aMean significant at the 1 level

bMean significant at the 5 level

cMean significant at the 10 level

DebtIssTotIss

DebtIssethDthornETHORN D

DthornE LevEF

(1) (2) (3) (4) (5) (6) (7) (8)

Profitt1 004a 000 002a 001c 003a 002 001b 002a

(35) (03) (34) (18) (35) (14) (23) (26)

IndLevt1 004c 001 001 001c 000 000 000 000

(19) (06) (15) (19) (02) (04) (03) (05)

Market=bookt1 001a 001a 000a 000a 000a 000a 000 000c

(47) (85) (68) (60) (55) (36) (05) (18)

Tangt1 004 005 013a 011a 011a 006a 003a 003a

(12) (15) (144) (112) (79) (38) (44) (42)

Assetst1 008a 004a 006a 004a 002a 001a 003a 002a

(97) (45) (283) (161) (53) (35) (147) (101)

Constant 079a 074a 002a 003a 000 000 000b 002a

(497) (442) (91) (111) (00) (05) (22) (93)

Year FE Yes Yes Yes Yes Yes Yes Yes Yes

Firm FE No Yes No Yes No Yes No Yes

R2 0015 0098 0022 0076 0009 0070 0012 0086

N 133905 133905 140552 140552 138229 138229 140572 140572

1444 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

inasmuch as short-term debt is probably more associated with day-to-dayoperations rather than longer term planning considerationsIn Table XIII we examine change in leverage ( D

DthornE) and change inleverage due to external financing (LEVEF) during good and bad timesseparately The traditional negative effect of a change in profits on achange in leverage is found both in good times and in bad timesHowever when we decompose the impact in order to isolate the externalcomponent as before the result disappears In good times the externalfinancing component of a change in profits is weak but still positiveDuring bad times it is not statistically significant This implies that distin-guishing between good and bad times does not fundamentally alter the basic

Table XII Short- and long-term debt issuances

The sample contains nonfinancial firms listed on the annual Compustat files for the period

from 1971 to 2009 The table presents estimates from regressions of short-term debtissuance to total issuance (in columns (1) and (2)) long-term debt issuance to totalissuance (in columns (3) and (4)) short-term debt issuance to total capitalization (in

columns (5) and (6)) and long-term debt to total capitalization (in columns (7) and (8))Some of the specifications include firm fixed effects The reported t-statistics are correctedfor clustering at the firm level

aMean significant at the 1 level

bMean significant at the 5 level

cMean significant at the 10 level

STDebtIssTotIss

LTDebtIssTotIss

STDebtIssethDthornETHORN

LTDebtIssethDthornETHORN

(1) (2) (3) (4) (5) (6) (7) (8)

Profitabilityt1 0002 0003 0042a 0008 0001 0001 0019a 0012a

(02) (04) (30) (05) (14) (10) (45) (28)

IndLevt1 0009 0007 0070b 0033 0001 0002c 0014a 0005

(06) (05) (26) (11) (08) (17) (26) (09)

M=Bt1 0003a 0003a 0003b 0010a 0000 0000 0003a 0002a

(48) (35) (22) (65) (10) (01) (84) (42)

Tangibilityt1 0020 0007 0011 0067 0005a 0006a 0097a 0081a

(12) (03) (03) (16) (34) (33) (108) (89)

Assetst1 0039a 0011b 0034a 0024b 0001a 0001 0046a 0020a

(89) (22) (38) (23) (36) (13) (190) (82)

Constant 0001b 0153a 0798a 0591a 0000c 0006a 0057a 0096a

(25) (144) (470) (288) (17) (124) (292) (234)

Year FE Yes Yes Yes Yes Yes Yes Yes Yes

Firm FE No Yes No Yes No Yes No Yes

R2 0025 0089 0018 0073 0030 0147 0016 0298

N 136146 136146 136140 136140 143031 143031 142847 142847

PROFITSndashLEVERAGE PUZZLEREVISITED 1445

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

Table XIII Externally financed leverage changes in good and bad times

The sample contains nonfinancial firms listed on the annual Compustat files for the period

from 1971 to 2009 The table presents estimates from regressions of change in debt tocapitalization ratio (in columns (1) (2) (5) and (6)) and change in leverage ratio consider-ing only external financing (in columns (3) (4) (7) and (8)) The change in leverage ratio

considering only external financing is defined as

LevEF frac14Dt1 thorn d

Dt1 thorn Et1 thorn dthorn e

Dt1

Dt1 thorn Et1

where D is debt E is book value of equity d is debt issues net of retirements and e is equityissues net of repurchases The regressions include lagged leverage factors and year indicator

variables The reported t-statistics are corrected for clustering at the firm level An industryis defined as having ldquogood timesrdquo if the median firm in that industry has a market-to-bookratio that is higher than the 67th percentile of the time-series distribution of industry

median market-to-book ratios Conversely an industry is defined as having ldquobad timesrdquoif the median firm in that industry has a market-to-book ratio that is lower than the 33rdpercentile of the time-series distribution of industry median market-to-book ratios

aMean significant at the 1 level

bMean significant at the 5 level

cMean significant at the 10 level

Good times Bad times

DDthornE LevEF D

DthornE LevEF

(1) (2) (3) (4) (5) (6) (7) (8)

Profitabiltyt1 0022c 0005 0016c 0025b 0073b 0041 0007 0007

(17) (03) (18) (24) (25) (11) (04) (03)

IndLevt1 0000 0011 0001 0009 0010 0018 0001 0000

(00) (09) (01) (13) (07) (10) (02) (00)

MB

t1

0004a 0003b 0000 0001 0008a 0001 0000 0000

(40) (24) (05) (08) (28) (04) (03) (03)

Tangt1 0097a 0041 0033a 0038b 0127a 0048 0059a 0033

(45) (16) (27) (25) (49) (14) (43) (16)

Assetst1 0014a 0013b 0024a 0019a 0037a 0023c 0036a 0025a

(30) (20) (84) (48) (45) (17) (88) (43)

Constant 0030b 0036 0000 0003 0028c 0014 0021 0001

(20) (16) (00) (07) (17) (06) (11) (01)

Year FE Yes Yes Yes Yes Yes Yes Yes Yes

Firm FE No Yes No Yes No Yes No Yes

R2 0006 0088 0009 0117 0026 0251 0023 0202

N 56820 56820 58200 58200 23462 23462 23511 23511

1446 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

conclusions The traditional leverage result on profits is driven by increasesin equity that are experienced and then partially but not completely offsetby firmsTo the extent that firms are engaging in active financing they are following

the prescriptions of the trade-off theory Thus we have a coherent picture ofhow firms are responding to profits Basically more profitable firms experiencean increase in the value of their equity They respond to this by issuing debtand repurchasing equity But the magnitude of the increase in equity is largerthan the magnitudes of the active steps taken in the debt and equity marketsThis means that the firmsrsquo actual actions are as predicted by the trade-offtheory But it leaves open the question of why they do not act more stronglyThe capital structure literature has long been interested in dynamics and

capital structure adjustment paths In recent years this interest has increas-ingly resulted in explicit dynamic optimizing models such as those ofGoldstein Ju and Leland (2001) Hennessy and Whited (2005) andStrebulaev (2007) These models commonly make specialized assumptionsleading some to question their generalityWe argue that both fixed transactions costs and variable transactions costs

are plausible and realistic in models that pin down firm size The simultaneouspresence of both is important Due to fixed costs firms only rebalance inter-mittently Due to variable adjustment costs the extent of the adjustment isreduced to the point at which an extra unit to adjustment is balanced againstthe extra unit transactions costs This means that both leverage increases andleverage reductions will only go part way toward a static optimumThe key point is that with both fixed and variable transactions costs we

should expect to only see intermittent leverage adjustments When leverage isadjusted it should ldquounder adjustrdquo when compared to a frictionless modelFurther discussion of transactions costs in a capital structure setting isprovided by Leary and Roberts (2005)

9 Conclusion

The connection between corporate profits and capital structure has beenvery influential in the assessment of the trade-off theory The conventionalinterpretation of the evidence has pushed the literature away from the trade-off theory and toward much more complex dynamic models and ideas As aresult it is important to make sure that the evidence has been correctlyinterpreted We show that the problem is with the interpretation of thetest results and a failure to take into account the costs associated with re-sponding to a profit shock

PROFITSndashLEVERAGE PUZZLEREVISITED 1447

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

We document that more profitable firms really do borrow more (not less)They tend to repurchase equity They experience an increase in both thebook value of equity and the market value of equity Less profitable firmsreally do tend to reduce their debt and to issue equityFirm size and market conditions also matter Larger firms tend to be more

active in the debt markets while smaller firms tend to be relatively moreactive in the equity markets During good times there is more use ofexternal financingThe usual profitsndashleverage puzzle result is primarily driven by the increase

in equity that is experienced by the more profitable firms Accordingly thepuzzle should be restated as asking why do firms not take sufficiently largeoffsetting actions to fully undo the change in equity What limits the mag-nitudes of the typical leverage response to profit shocksIn a frictionless model the partial response appears puzzling But there is

good empirical reason to believe that rebalancing entails both fixed andvariable costs and that firm size matters The rebalancing costs can befully avoided by doing nothing Accordingly the firm must decide whetherany given shock is big enough to be worth responding to If it is then thefirm must decide how big a response is called for These technical conditionsare known as ldquovalue matchingrdquo and ldquosmooth pastingrdquo Thus optimizationimplies that some shocks will be ignored Even if the shock is not ignoredthe optimal response will only partially undo the shock The magnitude ofthe leverage response must balance the marginal cost and the marginalbenefit of an extra unit of leverage Since the marginal cost of adjustingleverage is strictly positive the adjustment toward that static leverageoptimum will only go part way This is true both for leverage increasesand for leverage reductionsOne might conjecture that the variable costs would be time varying and

that such costs might have declined over time Since we do not have a fullaccounting of all of the variable costs it is of course impossible to be sureHowever there is some evidence suggesting that at least the directunderwriting fees are remarkably stable over time10

Surely nobody ought to be surprised that more profitable firms are moreinclined to repurchase equity while unprofitable firms tend not to do so Thefact that more profitable firms typically experience an increase in the value of

10 For IPOs Jay Ritter provides important summary evidence as httpbearwarringtonufleduritterIPOs2013Underwritingpdf He finds that for medium-sized deals a remark-

ably large fraction of the deals have a 7 fee year after year Whether this is due tobehavioral factors or some sort of a coordination device is not clear The extent towhich similar intertemporal stability carries over to other forms of external financemerits further study

1448 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

equity is equally natural Were it not for the contrary claims in the literature(ie the profitsndashleverage puzzle) we would argue that nobody ought to findit surprising that more profitable firms are more inclined to issue debt Thefact that leverage adjustments typically only partially offset profit shocksshould also sound quite natural These facts all fit together easily within thetrade-off theory once we take account of both fixed and variable costs ofactively adjusting capital structure

Appendix

Table AI Market leverage regressions

The table presents estimates of the regressions of market leverage on firm characteristics

The sample comes from the annual Compustat files during the period 1971ndash2009 Financialfirms are excluded The leverage is estimated as the ratio of debt over debt plus marketequity The explanatory variables Profitabilityt1

MB

t1

Tangibilityt1 and LnethAssetsTHORNt1are described in Table 1 IndMedianLevt1 is estimated as the median market leverage of allother firms in the same industry The industry is defined at the level of the four-digit SICcode All specifications include the year fixed effects The specifications in column (2)

additionally include the firm fixed effects We report t-statistics where the standard errorsare clustered at the firm level in parentheses

aSignificant at the 1 level

Quantile regressions

(1) (2) (3) (4) (5)

Profitabilityt1 013a 013a 007a 011a 020a

(275) (262) (349) (425) (538)

IndLevt1 046a 020a 032a 058a 063a

(506) (239) (1299) (1630) (1042)MB

t1

003a 001a 001a 002a 004a

(509) (303) (743) (709) (534)

Tangibilityt1 013a 018a 013a 014a 010a

(191) (181) (762) (571) (232)

LnethAssetsTHORNt1 001a 005a 001a 001a lt001a

(112) (258) (682) (342) (29)

Constant 010a 011a 005a 005a 029a

(171) (98) (173) (118) (398)

Year FE Yes Yes Yes Yes Yes

Firm FE No Yes No No No

Regression type OLS FE 25th tile Median 75thtile

Clustered SE Yes Yes No No No

R2 030 067

Pseudo R2 016 023 0201

N 158578 158578 158578 158578 158578

PROFITSndashLEVERAGE PUZZLEREVISITED 1449

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

Table AII Profitability sorts for firms issuing large amounts

Percentage of firms

Iss Ret Iss Iss Rep Iss Iss IssD IssE

D D NetD E E NetE DampE RepE RetD

(1) (2) (3) (4) (5) (6) (7) (8) (9)

Panel A 10 cutoff to define issuing or retiring activity

Allfirm

s

Low 205 150 134 260 05 257 45 00 23

2 207 175 103 77 06 75 11 01 10

3 243 196 101 50 08 47 08 01 09

4 252 186 113 50 13 48 08 02 08

High 227 161 118 60 27 57 07 04 09

t Highfrac14Low 70 38 62 755 226 768 320 90 152

Smallfirm

s Low 230 132 162 380 05 375 82 00 32

2 228 143 165 334 04 333 65 00 26

3 211 146 147 216 03 214 38 00 23

4 190 154 120 126 04 124 15 01 20

High 194 162 114 99 09 97 12 01 15

t Highfrac14Low 42 40 68 342 20 340 169 07 131

Largefirm

s Low 231 200 82 31 06 29 08 01 04

2 231 187 67 17 06 14 03 01 02

3 234 177 77 17 09 14 03 02 03

4 218 158 75 16 17 14 02 03 02

High 218 135 99 19 41 15 03 07 02

t Highfrac14Low 19 106 36 46 135 57 39 58 23

Percentage of firms

Iss Ret Iss Iss Rep Iss Both Iss D Iss E

D D Net D E E Net E DampE Rep E Ret D

(1) (2) (3) (4) (5) (6) (7) (8) (9)

Panel B 15 cutoff to define issuing or retiring activity

Allfirm

s

Low 158 101 100 230 01 228 27 00 15

2 142 115 63 57 01 56 05 00 06

3 165 128 57 32 01 31 03 00 05

4 173 118 65 35 01 34 03 00 05

High 160 102 70 46 04 44 03 00 05

t Highfrac14Low 08 06 141 738 87 746 277 24 131

Smallfirm

s Low 195 95 132 354 01 350 56 00 24

2 185 96 131 303 01 301 44 00 18

3 160 96 104 179 00 176 18 00 15

4 134 99 78 99 01 98 08 00 12

High 133 95 73 82 01 80 06 00 10

t Highfrac14Low 80 00 95 345 12 344 149 12 52

Largefirm

s Low 145 123 42 14 01 12 03 00 01

2 137 114 32 07 01 06 01 00 01

3 142 104 35 06 01 05 00 00 01

4 130 91 36 08 03 07 01 01 01

High 134 81 48 09 06 07 01 01 01

t Highfrac14Low 19 85 17 26 54 33 34 17 11

1450 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

Table AIV Debt and equity changesmdashwithout fixed effects

The sample contains nonfinancial firms listed on the annual Compustat files for the period from 1971

to 2009 The table presents estimates from regressions of change in debt (D) change in the book

value of equity (BVE) change in the market value of equity (MVE) and net equity issuances

(EqyIss) The explanatory variables include current and lagged change in operating income before

depreciation (Profits and Profitst1) and lagged change in book value of assets (Assetst1) The

regressions include year indicator variables The reported t-statistics are corrected for clustering at the

firm level

aMean significant at the 1 level

bMean significant at the 5 level

cMean significant at the 10 level

D BVE MVE EqyIss

(1) (2) (3) (4) (5) (6) (7) (8)

Profitst1 025a 024a 023a 038a 007 066a 007a 009a

(40) (32) (36) (39) (06) (29) (42) (51)Profits 051b 082a 149a 004a

(25) (33) (49) (33)Assetst1 002 005c 024a 001b

(08) (17) (34) (23)Constant 93 176a 45 71 1552a 1704a 130a 140a

(12) (28) (07) (09) (49) (49) (45) (45)Year FE Yes Yes Yes Yes Yes Yes Yes YesR2 001 005 001 011 001 005 002 002N 162056 162056 162130 162130 157550 157550 162154 162154

Table AIII Median profitability sorts

The sample contains nonfinancial firms listed on the annual Compustat files for the period from 1971

to 2009 The table presents median debt and equity median changes in debt and equity and median

equity issuances (EqyIss) in millions of dollars for all firms sorted on profitability and for profitability

sorts within the smallest and largest firms The sorts are done annually

D D BVE BVE MVE MVE EqyIss Assets(1) (2) (3) (4) (5) (6) (7) (8)

Allfirm

s

Low 21 00 75 17 256 20 01 1922 186 02 449 18 598 19 00 11033 816 06 1274 03 1637 05 00 32244 787 02 1541 18 2488 03 00 3556High 273 00 1361 66 3300 14 00 2578

Smallfirm

s Low 05 00 13 07 140 14 06 402 06 00 25 08 114 12 01 563 09 00 32 05 85 07 00 674 12 00 39 01 68 03 00 79High 10 00 51 04 98 01 00 90

Largefirm

s Low 11485 381 10269 371 13227 202 03 333902 13051 393 12786 140 17571 169 05 392463 11554 287 13429 24 21623 305 00 370314 9963 209 15052 194 29424 308 00 36705High 7361 00 16380 613 49142 439 00 34976

PROFITSndashLEVERAGE PUZZLEREVISITED 1451

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

References

Auerbach A J (1985) Real determinants of corporate leverage in B M Friedman (ed)

Corporate Capital Structures in the United States University of Chicago Press Chicago

IllinoisBertrand M and Mullainathan S (2003) Enjoying the quiet life corporate governance and

managerial preferences Journal of Political Economy 111 1043ndash1075Booth L Varouj A Demirguc-Kunt A and Maksimovic V (2001) Capital structures in

developing countries Journal of Finance 56 87ndash130Brennan M J and Schwartz E S (1984) Optimal financial policy and firm valuation

Journal of Finance 39 593ndash607Cameron A C and Trivedi P K (2010) Microeconometrics using Stata Stata Press TexasConstantinides G M (1979) Multiperiod consumption and investment behavior with

convex transactions costs Management Science 25 1127ndash1137Constantinides G M and Richard S F (1978) Existence of optimal simple policies for

discounted-cost inventory and cash management in continuous time Operations Research

26 620ndash636DeAngelo H DeAngelo L and Whited T M (2011) Capital structure dynamics and

transitory debt Journal of Financial Economics 99 235ndash261Dixit A K (1993) The Art of Smooth Pasting Routledge London and New YorkFama E and French K R (2002) Testing trade-off and pecking order predictions about

dividends and debt Review of Financial Studies 15 1ndash33Faulkender M and Petersen M A (2006) Does the source of capital affect capital struc-

ture Review of Financial Studies 19 45ndash79Faulkender M Flannery M J Hankins K W and Smith J M (2012) Cash Flows and

Leverage Adjustments Journal of Financial Economics 103 632ndash646Fischer E Heinkel R and Zechner J (1989) Dynamic capital structure choice theory

and tests Journal of Finance 44 19ndash40Frank M Z and Goyal V K (2008) Trade-off and pecking order theories of

debt in B E Eckbo (ed) Handbook of Corporate Finance Empirical Corporate Finance

Vol 2 Handbook of Finance Series Chapter 12 ElsevierNorth-Holland Amsterdam pp

135ndash202

Frank M Z and Goyal V K (2009) Capital structure decisions which factors are reliably

important Financial Management 38 1ndash37

Garvey G T and Hanka G (1999) Capital structure and corporate control the effect of

antitakeover statutes on firm leverage Journal of Finance 54 519ndash546

Goldstein R Ju N and Leland H (2001) An EBIT-based model of dynamic capital

structure Journal of Business 74 483ndash512

Graham J R and Tucker A L (2006) Tax shelters and corporate debt policy Journal of

Financial Economics 81 563ndash594

Hackbarth D and Mauer D C (2012) Optimal priority structure capital structure and

investment Review of Financial Studies 25 747ndash796

Halling M Yu J and Zechner J (2012) Leverage dynamics over the business cycle

Working paper University of Utah University of New South Wales and Vienna

University of Economics and Business AdministrationHennessy C A and Whited T A (2005) Debt dynamics Journal of Finance 60 1129ndash1165

Huang R and Ritter J R (2009) Testing theories of capital structure and estimating the

speed of adjustment Journal of Financial and Qualitative Analysis 44 237ndash271

1452 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

Hovakimian A Opler T and Titman S (2001) The debt-equity choice Journal of

Financial and Quantitative Analysis 36 1ndash24Jung K Kim Y and Stulz R M (1996) Timing investment opportunities managerial

discretion and the security issue decisions Journal of Financial Economics 42 159ndash185Kane A Marcus A J and McDonald R L (1984) How big is the tax advantage to

debt Journal of Finance 39 841ndash853Kayhan A and Titman S (2007) Firmsrsquo histories and their capital structures Journal of

Financial Economics 83 1ndash32Leary M T (2009) Bank loan supply lender choice and corporate capital structure

Journal of Finance 64 1143ndash1185Leary M T and Roberts M R (2005) Do firms rebalance their capital structures

Journal of Finance 60 2575ndash2619Leland H E (1994) Corporate debt value bond covenants and optimal capital structure

Journal of finance 49 1213ndash1252Lemmon M L Roberts M R and Zender J F (2008) Back to the beginning persist-

ence and the cross-section of corporate capital structure Journal of Finance 63 1575ndash1608Long M S and Malitz I B (1985) The investment-financing nexus some empirical

evidence Midland Corporate Finance Journal 3 53ndash59Mackie-Mason J K (1990) Do taxes affect corporate financing decisions Journal of

Finance 45 1471ndash1493Morellec E Nikolov B and Schurhoff N (2012) Corporate governance and capital

structure dynamics Journal of Finance 67 803ndash848Myers S C (1984) The capital structure puzzle Journal of Finance 39 575ndash592Myers S C (1993) Still searching for optimal capital structure Journal of Applied

Corporate Finance 61 4ndash14Rajan R and Zingales L (1995) What do we know about capital structure some evidence

from international data Journal of Finance 50 1421ndash1460Strebulaev I A (2007) Do tests of capital structure theory mean what they say Journal of

Finance 62 1747ndash1787

Strebulaev I A and Whited T M (2012) Dynamic Models and Structural Estimation in

Corporate Finance Now Publisher Boston and Delft

Titman S and Wessels R (1988) The determinants of capital structure choice Journal of

Finance 43 1ndash21

Welch I (2004) Capital structure and stock returns Journal of Political Economy 112

106ndash131

Welch I (2011) Two common problems in capital structure research the financial-debt-

to-asset ratio and issuing activity versus leverage changes International Review of Finance

11 1ndash17

PROFITSndashLEVERAGE PUZZLEREVISITED 1453

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

distinction between automatic changes in equity that stem from operationsas opposed to actual issues or repurchases of equity This distinction is em-pirically importantThis article is organized as follows Section 2 provides theoretical motiv-

ation for our empirical strategy Section 3 describes the construction of thedata and variables and provides summary statistics Section 4 providesresults from the fixed effects estimates of leverage ratios Section 5presents the frequency and magnitudes of financing activity for sorts onfirm size and profits Section 6 provides the main results on debt andequity issuance regressions Section 7 explores the debt and equityissuance responses of firms during good and bad times Section 8examines the debt and equity issuances when scaled by total issuances andtotal capital Section 9 concludes the article

2 Empirical Strategy

The key idea in the trade-off theory is that firms choose capital structure bytrading off the various costs and benefits The considerations may includebankruptcy costs tax benefits agency costs transactions costs etc Differentversions of the theory include slightly different elements From our perspec-tive the leverage adjustment costs are of particular importanceA decision to issue debt or to redeem it or to issue shares or to repurchase

them requires managerial thought and consideration There are fixed organ-izational costs There is also the possibility of managerial inertia as docu-mented in Bertrand and Mullainathan (2003) These are fixed costs that areapproximately independent of the magnitude of the adjustments Adjustingalso has marginal costs Depending on which action is taken there may befees for underwriters bankers lawyers etc These may have a fixed compo-nent too Commonly they have a variable cost component which dependson the size of the transaction So both fixed and variable costs of adjustingleverage are normally involved The exact magnitudes vary from situation tosituation What is worse these are likely to have time-varying components4

Why does the form of the transactions costs matter Because the existenceof such fees implies that there will not be a unique leverage target This is anapplication of an idea that goes back to Constantinides and Richard (1978)and Constantinides (1979) It should be stressed that there is nothing newhere in terms of the formal structure This is just a reinterpretation of

4 If there were no time-varying component then it would be easy to identify the exact upperand lower targets for each firm In each case the firm would go exactly to the appropriatetarget In reality with time variation in adjustment costs such precision is not possible

1420 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

standard dynamic optimization conditions For a reader who wants a moreexplicit treatment of the mathematics we recommend Dixit (1993) The ob-servation that this mathematics can be interpreted in terms of corporateleverage in which firm size matters is also not novel It can be found inLeary and Roberts (2005) who provide a simulation with both fixed costsand a weakly convex adjustment costs What is new here is the analysis of theimplication for the impact of profits on leverageSuppose that without active rebalancing leverage would just wander

about due to changes in the stock market value of equity According tothe trade-off theory excessively high leverage is costly (bankruptcy andagency costs) as is excessively low leverage (lost tax benefits and free cashflow problems) If there are no transactions costs then under conventionaltrade-off theory assumptions there will be a unique leverage target that willbe maintained at all timesDue to the fixed costs such frequent rebalancing would be too costly As a

result the firm permits the leverage to drift However if the drift is tooextreme in either direction then the firm will actively rebalance in the direc-tion of the static optimum If the rebalanced firm (inclusive of rebalancingcosts) is worth more then rebalancing is worth it If the rebalanced firmwould be worth less then rebalancing is not worth it Rebalancing takesplace at the point at which the values just matchWhen the firm is rebalancing it also needs to worry about how far to

move in the desired direction When each step has costs the firm needs toensure that the marginal cost of the last step is just equal to the marginalbenefit As a result both from excessively high leverage and from excessivelylow leverage the movements are only part wayTo fix ideas consider an off-the-shelf model of the leverage ratio The

focus is on the impact of the adjustment cost structure The formal structureis essentially the cash management problem studied by Constantinides andRichard (1978) and Dixit (1993) We follow the presentation and notation inDixit (1993) very closely to show that this is just an application of well-established ideas to the leverage problemLet xt denote the leverage ratio at date t is the interest rate Ku is a fixed

cost of actively increasing leverage Kd is the fixed cost of actively reducingleverage mu is the marginal cost of each unit of active leverage increase md isthe marginal cost of each unit of active leverage decrease f(x) is the flowpayoff if leverage is at x F(x) is the expected present value of the payoff ifleverage is at x x is the lower adjustment threshold x is the lower targetleverage (when adjusting from below) x is the upper adjustment thresholdand x is the upper target leverage (when adjusting from above) Accordingto the trade-off theory F 0 gt 0 F 00 lt 0

PROFITSndashLEVERAGE PUZZLEREVISITED 1421

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

In the absence of active adjustments the leverage ratio xt follows astandard process

dx frac14 ethx tTHORNdtthorn ethx tTHORNdw

where is the trend coefficient is the volatility coefficient and w is anormal shock process with zero mean and unit variance If x lt x lt x thenfinancial inaction is optimal If x drops to x then it is reset to x and if xrises to x then it is reset to xThe trade-off theory is frequently thought of as a class of theories in which

firm value first increases in leverage and ultimately drops if leverage gets toohigh This can be motivated by taxes and bankruptcy or by various agencyconcerns A range of ideas can be allowed for when we write that the flowpayoff (dividends andor profits) is fethxtTHORN and it depends on the leverage statext

5 If the firm starts at x0frac14 x then F(x) is the expected present value of theflow payoffs The discount rate is which allows us to write

FethxTHORN frac14 E

Z 10

fethxtTHORNetdtjx0 frac14 x

To maximize the firm value it must choose a leverage policy that sayswhen to rebalance and how much to rebalance These are given by thefamous value matching and smooth pasting conditionsThe value matching conditions are as follows

FethxTHORN FethxTHORN frac14 Ku thornmuethx xTHORN

FethxTHORN FethxTHORN frac14 Kd thornmdethx xTHORN

The smooth pasting conditions are as follows

F 0ethxTHORN frac14 F 0ethxTHORN frac14 mu

F 0ethxTHORN frac14 F 0ethxTHORN frac14 md

As explained by Constantinides and Richard (1978) with standard as-sumptions about the shape of F there are a pair of lines at heights mu andmd such that they intersect the function F 0ethxTHORN in such a way as to define thepoints x lt x lt x lt xTo sum up in this trade-off model with both fixed and variable

rebalancing costs there are four critical leverage points x gt x gt x gt x

5 For simplicity we suppose that the flow payoff is a continuous function of leverageThere is no discrete exit or bankruptcy point This allows us to simply interpret theConstantinides and Richard model much as in Leary and Roberts (2005)

1422 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

If leverage is between x (upper boundary) and x (lower boundary) the firmdoes nothing active Changes in leverage reflect various shocks to revenuesand costs that hit the firm If leverage drifts up to x (or above) then the firmrebalances to x If leverage drifts down to x (or below) the firm rebalancesto xWhat does this imply empirically A firm that does nothing can be

anywhere within the interval between x and x A firm that is rebalancingis a firm that has just hit one of the outer boundaries How does profitabilityaffect the likelihood of hitting a boundary High profits increase the value ofequity This increases the denominator of the leverage ratio and thus makesit more likely that the firm hits the lower boundary Low profits reduce thevalue of equity and thus make it more likely that the firm hits the upperboundaryTo be more specific suppose that the costs are symmetric and the leverage

changes are also following an ordinary Brownian motion Then the median

inert firm will have leverage ofxthornx

2 A firm that just actively reduced leverage

will be at x gtxthornx

2 A firm that just actively increased leverage will be at

x ltxthornx

2 A firm that just reduced leverage will generally be a firm that had

very high leverage To the extent that this reflects a lack of activerefinancing such firm will have low value of equity The low value ofequity will generally reflect low profits A firm that just increased leveragewill generally be a firm that had very low leverage In the same way a firmthat just increased its leverage will tend to have had high value of equity dueto high operating profits6

In this setting it is natural to focus on firms that are actively adjusting andto compare them to each other and to firms that are inactive It is alsohelpful to look directly at financing decisions and profitability rather thanfocusing on a leverage ratioOur approach shares the perspective of Welch (2004) Leary and Roberts

(2005) and Strebulaev (2007) that it is very important to focus on activeadjustments by firms Welch (2004) transforms the data to put everythingback into a ratio form His ldquoimplied debt ratiosrdquo still have both debt andequity in the denominator Similarly Leary and Roberts (2005) andStrebulaev (2007) pay a great deal of attention to leverage ratios whenstudying financing decisions Strebulaev (2007) does not have fixed costssince he makes use of a rescaling property Empirically we find differentadjusting behavior by large and by small firms So the rescaling property is

6 If we knew the values of the critical leverage points then it would be possible to derivemore refined tests Unfortunately those critical values are not observable

PROFITSndashLEVERAGE PUZZLEREVISITED 1423

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

an issue of concern His model does not account for the negative relationshipbetween current leverage and past market-to-book ratioWe depart from the earlier literature by arguing that the use of leverage

ratios can be particularly misleading for profitability since the use of aleverage ratio causes identification problems It is hard to be sure whethera leverage change reflects changes in the numerator or the denominator Thisis particularly problematic when as here the rebalancing is partialIt should be noted that much empirical capital structure literature debates

about whether to use a book leverage ratio or a market leverage ratio Bookleverage is used based on the argument that managers borrow against assetsin place which are better reflected by book assets Others such as Welch(2004) argue that book values are essentially an accounting fiction and thatonly market leverage should count Empirical papers regularly attempt tosidestep this debate by carrying out tests on both book leverage and marketleverage Our point is that both kinds of leverage ratios are potentially mis-leading when examining the impact of profits on capital structure

3 Data

We use conventional data sources starting with the merged Compustat-CRSP data The data are annual and are converted into constant year2000 dollars using the GDP deflator We exclude financial companies (SIC6000-6999) firms involved in major mergers (Compustat footnote code AB)firms reporting financial data in currencies other than the US dollars andfirms with missing data on our key variables7 The ratio variables aretrimmed at the 1 level in both tails of the distribution This serves toremove outliers and the most extremely misrecorded data The finalsample consists of 179021 firm-year observations from 1971 to 2009Table I provides definitions of financial variables and reports summary

statistics The average debt (in constant US$) is about $653 million while themedian is $24 million A significant fraction of firms have zero debt (the 10thpercentile is 0) Book equity is slightly larger than book debt Market equityis almost three times larger than book debt Book assets average $2191million although the medians are considerably smallerIf issuing or retiring securities incurs no fixed costs then we would expect

to see many small actions and very few large actions (Leary and Roberts2005) If there were significant fixed costs involved in issuing or retiringoutstanding securities then small issues might not be worthwhile Table I

7 These include debt book value of equity market value of equity assets book andmarket leverage profitability market-to-book assets ratio and tangibility

1424 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

shows that although most firms issue little debt or equity in a given year theaverages are large In other words when firms actually enter debt and equitymarkets they intervene massivelyThe mean constant dollar debt issue is $165 million (the median is $2

million) In unreported tables we find that the average debt issue is about81 of assets (the median is 18) About 38 of the firms issue no debt8 issue between 0 and 1 of the value of their assets as debt another

Table I Data description

Data sources The sample comes from the annual Compustat files The sample period is

1971ndash2009 We exclude financial companies (SIC 6000-6999) firms involved in majormergers (Compustat footnote code AB) firms reporting financial data in currencies otherthan the US dollar and firms with missing data on the key variables All financial variables

are deflated to year 2000 using the GDP deflator N is the number of observations SD isthe standard deviation D BVE MVE Assets Debt issuance Equity issuance Debt re-payment Equity repayment are all measured in $ millions The ratio variables are trimmed

at the 1 level in both tails of the distribution

Variable definitions Debt (D)frac14 long-term debt (dltt)thorn short-term debt (dlc) book equity(BVE)frac14 common shareholder equity (ceq) market equity (MVE)frac14 no of outstanding

shares (csho) closing price ( prcc f ) assetsfrac14 book assets (at) debt issuancefrac14 issuance oflong-term debt (Maxethdltis 0THORN)thorn increase in current debt (Maxethdlcch 0THORN) equity issu-ancefrac14 sale of common stock (Maxethsstk 0THORN) debt repaymentfrac14 reduction of long-termdebt (Maxethdltr 0THORN)thorn decrease in current debt (Minethdlcch 0THORN) equity repurchasefrac14 purchase

of common stock (Maxethprstkc 0THORN) book leveragefrac14D(DthornBVE) market leveragefrac14D(DthornMVE) profitabilityfrac14EBITDA (oibdp)assets market-to-book ratio M

B frac14Marketvalue of assets (MVA)assets where MVAfrac14DthornMVEthorn preferred-liq value (pstkl)

deferred taxes (txditc) tangibilityfrac14 net property plant and equipment (ppent)assets

Variable N Mean SD

Distribution

10th 50th 90th

Debt (D) 179021 653 4592 0 24 1048

Book equity (BVE) 179021 815 4122 3 70 1373

Market equity (MVE) 179021 1784 9589 9 118 2698

Assets 179021 2191 11530 10 153 3575

Debt issuance 179021 165 1335 0 2 245

Equity issuance 179021 26 186 0 0 40

Debt repayment 179021 132 1123 0 3 181

Equity repurchase 179021 24 257 0 0 9

Book leverage 179021 036 034 000 031 072

Market leverage 179021 027 025 000 020 066

Profitability 179021 005 027 018 011 023MB 179021 165 206 052 102 324

Tangibility 179021 031 024 005 026 070

PROFITSndashLEVERAGE PUZZLEREVISITED 1425

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

16 issue between 1 and 5 of the value of their assets as debt and theremaining 38 issue debt in excess of 5 of the value of their assetsThe mean constant dollar equity issue is about $26 million (the median is

about $04 million) As a fraction of assets the mean and median equityissues are about 7 and 02 respectively About 33 of the firms issue noequity 34 of the firms issue between 0 and 1 of the value of theirassets as equity another 14 issue between 1 and 5 of the value of theirassets as equity and the remaining 19 of the firms issue equity that is inexcess of 5 of the value of their assets Average debt repayments are largerthan equity repurchases This perhaps reflects the finite maturity of debt andits contractual repayment The median firm does not repurchase equityWe construct both book and market leverage ratios Book leverage is

defined as debt over debt plus book equity Market leverage is defined asdebt over debt plus market equity8 The median book leverage is 036 (theaverage is 031) The median market leverage is 027 (the average is 020)Profitability is defined as the ratio of operating income before depreci-

ation to assets While the average firm is profitable (the ratio of EBITDA toassets is 005) the median firm is even more profitable (with a profitabilityratio of 011) But the sample also includes a large number of unprofitablefirms as the 10th percentile is 018 The table also reports descriptive stat-istics on the market-to-book ratio and the tangibility ratio The market-to-book ratio (MB) defined as the ratio of the market value of assets to bookassets averages at about 165 Tangibility defined as the ratio of netproperty plant and equipment to assets averages at about 31

4 Leverage Ratio Regressions

The previous literature focuses on estimates obtained using leverage ratiosHence we start with a similar estimation to check whether our results matchthose of previous studies Table II presents the results for book leverage Weobtain similar results for market leverage and they are presented inAppendix Table AI The regressions include leverage factors following theprevious capital structure literature (see eg Lemmon Roberts and Zender(2008) and Frank and Goyal (2009)) The factors include (i) profitability (ii)industry median leverage (iii) market-to-book assets ratio (iv) tangibility ofassets and (v) firm size Industry median leverage is constructed as themedian leverage of all other firms in the industry excluding the firm under

8 Welch (2011) stresses the idea that nonfinancial liabilities should not be implicitly mis-treated as if they were equity by paying excessively narrow attention to financial liabilitiesin a leverage ratio We have adopted his approach in the empirical work reported here

1426 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

consideration Firm size is defined as the natural log of assets Other factorsare defined in Section 3 We cluster standard errors by firm and estimatethese regressions both with and without fixed effectsIn Table II we run conventional cross-section leverage regressions Since

we employ the usual data it is not surprising that our results match thosereported in the existing capital structure literature Profitability has anegative sign in both the book leverage regressions and in the marketleverage regressions The coefficients on other factors largely match thosereported in earlier studies Firms operating in industries with high leverage

Table II Leverage and profitability

The table presents parameter estimates for book leverage regressions The sample consists

of nonfinancial companies in the annual Compustat files during the period 1971ndash2009 Theleverage is estimated as the ratio of debt over debt plus book equity The explanatoryvariables Profitabilityt1

MB

t1

Tangibilityt1 and LnethAssetsTHORNt1 are described in

Table I IndLevt1 is estimated as the median book leverage of all other firms in thesame industry The industry is defined at the level of the four-digit SIC code All specifi-cations include the year fixed effects The specifications in column (2) additionally include

the firm fixed effects We report t-statistics where the standard errors are clustered at thefirm level in parentheses

aSignificant at the 1 level

Quantile regressions

(1) (2) (3) (4) (5)

Profitabilityt1 026a 020a 011a 017a 041a

(293) (192) (378) (518) (1015)

IndLevt1 042a 012a 030a 052a 052a

(369) (107) (926) (1252) (950)MB

t1

002a 001a 001a 002a 002a

(203) (102) (392) (426) (418)

Tangibilityt1 018a 025a 024a 019a 012a

(206) (154) (964) (580) (276)

LnethAssetsTHORNt1 001a 002a 002a 002a 001a

(114) (76) (899) (478) (96)

Constant 012a 010a 012a 002a 029a

(156) (60) (284) (27) (387)

Year FE Yes Yes Yes Yes Yes

Firm FE No Yes No No No

Regression type OLS FE 25th tile Median 75thtile

Clustered SE Yes Yes No No No

R2 013 054

Pseudo R2 016 015 009

N 158824 158824 158824 158824 158824

PROFITSndashLEVERAGE PUZZLEREVISITED 1427

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

tend to have high leverage A higher market-to-book ratio is associated withlower leverage Larger firms are typically more highly levered The coeffi-cients on tangibility are positive The conclusion from Table II is clear Theleverage regression results match those reported in previous studiesAt times one may be concerned that in a regression what is being

estimated is a conditional mean while some parts of the distributionmight behave differently from other parts If so then ldquoaccidentsrdquo likesample selection issues or asymmetries in the underlying distributionsmight play a greater role than is usually recognized In other words thesource of identification could be misinterpreted To guard against this po-tential problem we estimate the same model using quantile regressions9

Columns 3 4 and 5 report what happens at the 25th 50th and 75th per-centiles The basic model results are robust in terms of both the sign and thestatistical significanceThe impact of profits does appear to be somewhat stronger among the

high leverage firms For instance in a book leverage regression the coeffi-cient on profits for the first quartile is 011 with a t-ratio of 378 For thethird quartile the coefficient on profits is 041 with a t-ratio of 1015This difference is curious and might deserve further study However the factthat in each case we get the negative sign and statistical significance is suf-ficient for the purposes of the present article So the results on profits arerather robust across the distributionTable III sorts firms into those that are actively adjusting their leverage

and those that are passive In the presence of both fixed and variable trans-acts costs this distinction is important We expect the coefficient on profit-ability in leverage ratio regressions to be more negative in firms that arepassive and less negative in firms that are actively adjustingWe define active firms as those that are either issuing net debt or net

equity in excess of 5 of their assets Passive firms are those that neitherissue debt nor equity in nontrivial amounts (ie greater than 5 of theirassets) Consistent with our expectations we find in Table III that the coef-ficients on profits are significantly more negative for passive firms than theyare for active firms With firm and year fixed effects the coefficient onprofits is 008 for active firms and 017 for the passive firms The twocoefficients are significantly different from each other (plt 001) InTable III the standard control variables are used They have the usualsigns and for the most part do not differ much between the active and thepassive firms

9 Cameron and Trivedi (2010) provide an extensive discussion of quantile regressions inStata

1428 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

5 Financing Activity

51 FREQUENCY OF DEBT AND EQUITY ISSUES

Table IV reports the percentage of firms issuing or repurchasing debt orequity for annual sorts based on lagged profitability In Panel A weemploy the conventional 5 cutoff rule to exclude minor fluctuations anddefine firms ldquoissuing debtrdquo as those that issue debt (both short-term andlong-term) in excess of 5 of the value of their assets Other decisions aresimilarly defined using a 5 cutoff Appendix Table AII examines the dis-tributional properties across profitability (and size) quintiles for ldquolargerdquo is-suances by increasing the 5 cutoff to higher levels such as 10 and 15How does security issuance behavior vary with firm profitability To

answer this question Table IV sorts firms by profitability and reports the

Table III Leverage and profitability for active and passive firms

The table presents estimates of leverage on firm characteristics for the subsamples of firms

that are active in either debt or equity markets Active firms are defined as those that issueeither net debt in excess of 5 of the value of assets or issue net equity in excess of 5 ofthe value of their assets The passive firms do not issue significant (more than 5) of either

debt or equity The sample comes from the annual Compustat files during the period 1971ndash2009 Financial firms are excluded The variables are described in Tables I and II Wereport t-statistics where the standard errors are clustered at the firm level in parentheses

aSignificant at the 1 level

Active Passive

(1) (2) (3) (4)

Profitabilityt1 006a 008a 022a 017a

(126) (148) (254) (188)

IndLevt1 044a 019a 041a 013a

(424) (149) (326) (127)MB

t1

002a 001a 003a 001a

(362) (196) (327) (172)

Tangibilityt1 010a 014a 014a 019a

(146) (118) (145) (130)

LnethAssetsTHORNt1 001a 004a 001a 004a

(121) (179) (147) (169)

Constant 012a 008a 003a 007a

(120) (57) (45) (42)

Year FE Yes Yes Yes Yes

Firm FE No Yes No Yes

R2 034 064 033 075

N 43395 43395 58943 58943

PROFITSndashLEVERAGE PUZZLEREVISITED 1429

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

percentage of firms issuing or repurchasing debt or equity The table alsoreports test statistics from tests that examine whether the percentage differ-ences between the high- and low-profit firms are significantly different fromeach other The results in Column (1) show that the likelihood of issuing debt

Table IV Profitability sorts for debt and equity issuers

The sample contains nonfinancial firms listed on the annual Compustat files for the period

from 1971 to 2009 The table presents the percentage of firms issuing and retiring (orrepurchasing) debt and equity The firms are classified as ldquoissuing debtrdquo if they issuedebt in excess of 5 of the value of their assets as ldquoissuing equityrdquo if they issue equity

in excess of 5 of the value of their assets as ldquoretiring debtrdquo if they retire debt in excess of5 of the value of their assets and as ldquorepurchasing equityrdquo if they repurchase equity inexcess of 5 of the value of their assets In addition we also report net debt issuers which

are firms that issue net debt over 5 of the value of their assets and net equity issuerswhich are firms that issue net equity in excess of 5 of the value of their assets Weannually sort firms on lagged profitability and report the percentage of firms in each ofthese categories The bottom part of each panel reports the percentage of firms issuing or

retiring securities by profitability within the smallest and largest asset quintiles The tablealso reports results from t-test for the difference in percentage of firms issuing or retiringsecurities in the lowest and the highest profit quintile

Percentage of firms

Iss Ret Iss Iss Rep Iss Both IssD IssE

D D NetD E E NetE DampE RepE RetD

(1) (2) (3) (4) (5) (6) (7) (8) (9)

Allfirm

s

Low 276 245 186 301 13 297 71 01 38

2 319 299 181 117 21 112 27 02 20

3 385 344 194 86 29 80 21 05 15

4 386 330 213 81 43 75 20 07 14

High 329 273 198 92 82 81 17 13 14

t Highfrac14Low 150 84 39 723 414 763 358 176 200

Smallfirm

s Low 275 204 201 411 11 405 109 01 47

2 291 233 213 379 10 378 99 01 45

3 281 246 204 265 14 261 66 01 41

4 275 272 186 162 15 159 34 02 35

High 275 281 178 136 27 131 25 02 26

t Highfrac14Low 01 84 28 316 55 318 171 18 55

Largefirm

s

Low 414 363 177 87 22 79 28 03 11

2 449 379 180 62 30 57 17 07 09

3 439 360 186 56 41 48 16 09 07

4 403 320 194 46 67 39 13 12 06

High 374 264 207 50 129 35 10 28 05

t Highfrac14Low 50 130 45 89 244 116 77 120 47

1430 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

is mostly independent of firm profitability In the aggregate there is a smalldifference but it is sensitive to sorting on firm size Similarly in Column (2)there is only a weak relation between the profitability and the likelihood thatthe firm retires debt To a first approximation the fraction of firms issuing orrepurchasing debt is largely independent of profits Large firms are somewhatmore likely than small firms to issue debt and also to retire debtUnlike debt issues the probability of issuing equity is strongly related to

profitability as can be seen in Column (4) Low-profit firms are much morelikely to issue equity than are high-profit firms In the lowest profit quintile301 of the firms issued equity In the highest profit quintile it was only92 The lower panel reports separate sorts for small and large firms Similarpatterns are found for both groups of firms that is the likelihood of equityissuance declines as profitability increases The differences between the quin-tiles are quite stark for the group of small firms All quintiles of small firmsissue equity much more frequently than the corresponding large firm quintilesHigh-profit firms are much more likely to repurchase equity as shown in

Column (5) In the overall sample only 13 of low-profit firmsrepurchased equity compared with 82 of the high-profit firms Whenwe separate the small and large firms again a difference emerges Ingeneral small firms do not repurchase much equity while a greaterfraction of the large firms do soGiven these facts it is natural that low profitability firms are much more

likely to be net issuers of equity than are high profitability firms as shown inColumn (6) Among the firms in the lowest quintiles of profits roughly 30of firms issue net equity that exceeds 5 percent of their assets In contrastfor the most profitable quintile only about 8 issue net equityFirm size is an important variable in the recent literaturemdashit is often used

as a proxy for access to capital markets (as in Faulkender and Petersen(2006) and Leary (2009)) Small firms are bank-dependent risky and infor-mationally opaque They have restricted access to public debt markets andconsequently face more severe supply constraints in their ability to issuedebt Thus we expect small firms to be more sluggish in adjusting theirdebt and equity in response to shocks to profitability Large firms incontrast have much easier access to public debt markets and they facefewer obstacles in accessing securities marketsTo examine how size and profitability interact the bottom part of Table IV

examines profitability sorts for the smallest and largest firms We first sortfirms annually by firm size and then within each size quintile we sort them onprofitability We do these two-way sorts to ensure that we have similarnumber of firms in profitability subgroups for both small and large firmsamples Unconditionally small firms are generally less profitable while

PROFITSndashLEVERAGE PUZZLEREVISITED 1431

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

large firms are relatively more profitable The two-way sorts allow us to havea uniform distribution of firms in various profitability quintiles for both smalland large firm samples The bottom part of Table IV reports results for thesmallest and the largest size quintiles Among the small firms there is littlerelation between profitability and the likelihood of issuing debt Howeverlarger firms exhibit a small increase in the likelihood of net issuance of debtwith increasing profitabilityIn contrast to debt the effects of profitability on equity issuance and

repurchases are much more consistent across size quintiles As profitabilityincreases firms are generally less likely to issue equity and more likely torepurchase it Across the two extreme size groups we note a striking differ-ence between the proportion of firms issuing equity and those issuing debtAmong low-profit small firms almost 41 are equity issuers whereas onlyabout 9 of low-profit large firms are equity issuers However regardless ofsize we note a monotonic reduction in the likelihood of issuing equity asprofitability increasesColumn (7) reports the fraction of firms that issue both debt and equity in

excess of 5 of assets both as a function of profits in the upper panel andas a function of firm size and profits in the lower panel In the upper panelwe see that low-profit firms are much more likely to issue both debt andequity (71) while the high-profit firms are less likely to do so (17) Thelower panel shows that small firms are much more likely to be issuers of bothin contrast to large firms In both firm size categories the profit pattern is thesame with low profits being more frequent issuers of both simultaneouslyImportantly Column (8) shows that the likelihood of issuing debt and

simultaneously repurchasing equity increases with profitability Converselyin Column (9) we find that the likelihood of doing the reverse that is issuingequity and retiring debt declines with profitability The effects of sorting onfirm size and firm profitability mirror those reported for all firms Overallthe results show that low-profit firms are less likely to issue debt and repur-chase equity they are instead more likely to issue equity and retire debtFirms with high profitability exhibit the reverse patternIn Appendix Table AII we present results from profitability (and size)

quintiles for large issuances defined using the 10 cutoff (in Panel A) andthe 15 cutoff (in Panel B) The results from these panels are largely con-sistent with those described above

52 MAGNITUDES OF ISSUANCES AND EQUITY CHANGES

In the previous section we considered the probability of having a nontriviallevel of debt or equity activity The next question is how large are the dollar

1432 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

values involved In Table V we sort firms according to profits and thentabulate the levels and changes in both debt and equity We do this firstfor all firms and then for small versus large firmsFor the analysis on all firms we observe in Column (1) that debt peaks at

the middle of the distribution This is because firms with medium profitabil-ity are also the largest as seen in Column (8) which reports average assetvalues for different profitability quintiles When we sort first by firm size andthen examine firms with different levels of profitability within size quintileswe find that for small firms debt is roughly independent of profits while forlarge firms less profitable firms have a higher level of debtAs expected Columns (3) and (5) show that more profitable firms have

higher equity values Columns (2) and (7) consider the relationship betweenissuances of debt and equity and firm profitability Debt issuances are sig-nificantly larger for more profitable firms In contrast equity issuances arelarger for low-profit firms Small low-profit firms issue substantially lessequity compared with small high-profit firms Among large firms the low-profit ones tend to issue equity while the high-profit ones tend to repurchaseequity The fact that more profitable firms issue debt and repurchase equitywhile the least profitable firms retire debt and issue equity is consistent withthe predicted relation between profitability and financing decisions under thetrade-off theoryWe also examine two-way sorts by size and profitability and report results

for the smallest and largest quintile of firms For the smallest quintile offirms the change in debt is largely unrelated to profits But for large firmsthere is a positive relation between profits and debt issuances High-profitfirms have a big positive change in debt Low-profit large firms have anegative change in debtColumns (4) and (6) provide an explanation of why the leverage ratio

regression results contradict those from the basic profitability sorts pre-sented here As we can see profitability indirectly affects leverage ratiosby increasing equity values Changes in both the book value of equity andthe market value of equity are positive and large for highly profitable firmsIn contrast these changes are negative for less profitable firms This evidenceillustrates an important issue concerning the use of leverage ratios Suchratios are often interpreted as essentially reflecting the use of debt by thefirm This interpretation while common is empirically misleadingFor the typical firm the change in the value of equity is larger than the

change in debt For example in the third profit quintile for large firmscolumn (7) the mean equity issue is just $9 million but the change in themarket value of equity is $242 million At the same time the mean change indebt is $27 million This suggests that a fair bit of the observed variation in

PROFITSndashLEVERAGE PUZZLEREVISITED 1433

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

the leverage ratios is primarily driven by the changes in the market value ofequity in the denominator rather than by the changes in debt in the numer-ator Since equity issues are often small this implies that the variation in theleverage ratio is primarily driven by internal operations rather than byexternal financing actions This again points to the fact that leverageratios can provide a misleading account of actual patterns in the dataThere is always a potential concern that the averages may be misleading

due to the impact of outliers To address this concern in AppendixTable AIII we present median values of the profitability sorts Theseresults generally reinforce the findings in Table VTable VI again considers the magnitudes of financing activity but this

time the issues are scaled by a measure of firm size That way it is easier to

Table V Magnitude of financing activity

The sample contains nonfinancial firms listed on the annual Compustat files for the period

from 1971 to 2009 The table presents average debt and equity changes in debt and equityand equity issuances (in millions of dollars) for firms sorted on profitability within sizeclasses The table reports information for all firms sorted on profitability and for profit-

ability sorts within the smallest and largest firms The sorts are done annually The tablealso reports results from t-test for the difference in percentage of firms issuing or retiringsecurities in the lowest and the highest profit quintile

D D BVE BVE MVE MVE EqyIss Assets

(1) (2) (3) (4) (5) (6) (7) (8)

Allfirm

s

Low 83 5 78 10 120 5 7 234

2 605 13 509 24 721 8 17 1691

3 1227 9 1123 22 1834 3 9 3551

4 833 23 1166 7 2408 74 9 2994

High 570 53 1171 71 3817 68 36 2581

t Highfrac14Low 31 9 36 9 38 2 16 36

Smallfirm

s Low 25 04 42 01 332 28 33 97

2 29 06 55 02 278 16 21 116

3 31 07 63 01 212 05 12 127

4 30 06 70 05 179 10 08 136

High 24 06 84 14 234 17 07 142

t Highfrac14Low 04 15 153 85 73 11 201 107

Largefirm

s Low 3929 55 2990 152 4671 11 63 10846

2 3826 19 3178 95 5122 66 11 10721

3 3074 27 3498 41 6173 242 9 9697

4 2623 62 3915 45 8473 94 57 9855

High 2180 193 4642 261 14924 202 148 10171

t Highfrac14Low 94 64 108 87 246 12 169 16

1434 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

see whether large firms behave disproportionately when compared withsmall firms Column (1) shows that low-profit firms have more debtas reported in many previous studies In Columns (4) and (6) we seethat low-profit firms have a major decline in both the book value ofequity and the market value of equity The high-profit firms have thereverse experienceMore interesting are the active decisions of the firms In Column (2) we see

that in contrast to the usual impression the low-profit firms are reducingdebt while the high-profit firms are increasing it In Column (7) we see thatthe low-profit firms are also reducing leverage by issuing equity The high-profit firms issue little equity In the lower panel similar results are foundwhen we separate small and large firms Equity issuing is particularly strong

Table VI Magnitude of financing activitymdashscaled levels and changes

The sample contains nonfinancial firms listed on the annual Compustat files for the period

from 1971 to 2009 The table presents average debt and equity changes in debt and equityand equity issuances scaled by assets for firms sorted on profitability within size classes Thetable reports information for all firms sorted on profitability and for profitability sorts

within the smallest and largest firms The sorts are done annually The table also reportsresults from t-test for the difference in percentage of firms issuing or retiring securities inthe lowest and the highest profit quintile

D D BVE BVE MVE MVE EqyIss

(1) (2) (3) (4) (5) (6) (7)

Allfirm

s

Low 386 122 12 871 3427 1868 217

2 285 58 407 119 940 157 28

3 293 12 425 14 742 03 12

4 269 26 461 02 955 27 10

High 192 15 551 47 1675 74 07

t Highfrac14Low 35 49 50 47 127 42 157

Smallfirm

s Low 545 154 841 2145 9688 9128 696

2 684 49 712 1915 5014 2318 288

3 331 253 141 969 3063 954 129

4 267 181 357 175 1673 221 55

High 194 05 543 26 1758 147 41

t Highfrac14Low 49 17 51 36 88 27 99

Largefirm

s Low 350 30 306 58 511 55 09

2 346 14 335 13 523 02 04

3 353 165 332 71 711 29 01

4 280 01 414 04 909 09 05

High 227 15 477 19 1643 27 14

t Highfrac14Low 371 112 453 181 570 57 221

PROFITSndashLEVERAGE PUZZLEREVISITED 1435

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

among the small low-profit firms that are also experiencing sharp reductionsin the value of equity In column (2) we see that both for the large and for thesmall firms net debt increases are primarily found in the highest quintile offirm profitability

6 Debt and Equity Regressions

So far we have demonstrated that in our data the conventional leverageregressions have the usual signs We then demonstrated that in simple sortsof the data strikingly different active financing patterns emerge The nexttask is to reconcile these differencesTable VII presents simple regressions predicting changes in debt book

equity and market equity as well as equity issuances Consistent with thesorts we control for changes in firm size and we also include firm and yeardummies There may be a concern about the speed of a firmrsquos responsesAccordingly we consider the lagged change in profits as the main case butwe also include the contemporaneous change in profits Whether the contem-poraneous change in profits is included or not does not in fact matter Wereport panel-robust standard errors adjusted for clustering at the firm levelThe results in Table VII support the findings in the sorts Firms that have

an increase in profits in 1 year increase debt both in the same year and in thenext year Firms that have an increase in profits experience an increase inboth the book value of equity and the market value of equity Importantlyhowever firms that experience and increase in profits reduce their equityissues All these effects are statistically significant (except for the marketvalue of equity in the column 5 specification) and seem intuitively reason-able In Appendix Table AIV similar regressions to those presented aboveare reported but without firm-fixed effects We find that empirically verylittle changes whether we include firm fixed effects or leave them out Thusthe targeting behavior to the extent that it happens does not account forwhat we are reportingThe next step is to bring the sorts and the simple regressions together into

a common setting We therefore include factors in addition to profitability(i) median industry leverage (ii) the market-to-book assets ratio (iii) tangi-bility of assets and (iv) firm size (measured by the log of assets) Rajan andZingales (1995) show that these factors are related to leverage in G7countries A number of studies have used these factors to estimateleverage targets Frank and Goyal (2009) show that these factors arerobustly related to leverage in the USA In these regressions we use indica-tors for the quintile that the firm is in for each factor

1436 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

Table VIII reports these results In Columns (1) and (2) we examine debtissues and again show that profits positively affect debt issuances The effectsare large and statistically significant at the 1 level Columns (3) and (4)which examine book equity show a strong effect of profits This effect is onlyslightly reduced by the inclusion of other factors Columns (5) and (6) illus-trate that the impact of profits on the change in market equity is also robustto the inclusion of the conventional factors Columns (7) and (8) examineequity issuances Here we again find that profits have a significant negativeimpact on equity issuances As before the results are robust to the inclusionof the conventional factorsIn unreported tables we estimate debt and equity changes using quantile

regressions In terms of the signs and significance of the profitability variable

Table VII Debt and equity changes

The sample contains nonfinancial firms listed on the annual Compustat files for the period

from 1971 to 2009 The table presents fixed effect estimates from regressions of changes indebt (D) changes in the book value of equity (BVE) changes in the market value ofequity (MVE) and net equity issuances (EqyIss) The explanatory variables include

current and lagged change in operating income before depreciation (Profits andProfitst1) and lagged change in the book value of assets (Assetst1) In addition tofixed firm effects the regressions include year indicator variables The reported t-statistics

are corrected for clustering at the firm level

aMean significant at the 1 level

bMean significant at the 5 level

cMean significant at the 10 level

D BVE MVE EqyIss

(1) (2) (3) (4) (5) (6) (7) (8)

Profitst1 022a 025a 015b 035a 004 060b 005b 007a

(33) (31) (25) (35) (03) (25) (25) (53)

Profits 050b 078a 140a 003a

(24) (32) (48) (34)

Assetst1 000 007b 025a 001b

(00) (21) (35) (25)

Constant 167 283a 34 51 1145a 1182a 605a 595a

(14) (30) (04) (05) (32) (29) (69) (74)

Year FE Yes Yes Yes Yes Yes Yes Yes Yes

Firm FE Yes Yes Yes Yes Yes Yes Yes Yes

R2 001 004 001 010 001 005 002 003

N 162056 162056 162130 162130 157550 157550 162154 162154

PROFITSndashLEVERAGE PUZZLEREVISITED 1437

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

the results are qualitatively similar to the OLS results reported in Table VIIIProfitability positively affects debt issuances and negatively affects equityissuances across various quantiles Overall the results in Table VIII showthat there is nothing that is special about using sorts or regressions to explaindebt and equity issues In either case we find that more profitable firms tendto increase their debt experience an increase in the value of equity andrepurchase shares Thus the control factors are not responsible for theusual rejection of the trade-off theory It would appear that what mattersis the size of the issuing activity relative to the organic increases in equityvalue due to profits In other words what matters in the leverage regressionseems to be coming from the presence of E in D

DthornE and not from D

Table VIII Debt and equity issuances profitability and leverage factor quintiles

The sample contains nonfinancial firms listed on the annual Compustat files for the period

from 1971 to 2009 The table presents estimates from regressions of change in debt (D)change in the book value of equity (BVE) change in the market value of equity (MVE)and net equity issuances (EquityIss) The explanatory variables include lagged profitability

quintiles and lagged leverage factor quintiles In addition to fixed firm effects the regres-sions include year indicator variables The reported t-statistics are corrected for clustering atthe firm level

aMean significant at the 1 level

bMean significant at the 5 level

cMean significant at the 10 level

D BVE MVE EqyIss

(1) (2) (3) (4) (5) (6) (7) (8)

ProfitabilityQint 214a 167a 283a 175a 80 429a 53a 55a

(70) (54) (66) (59) (08) (48) (55) (62)

IndLevQint 104a 43 99 20

(29) (15) (10) (10)

Ln(Assets)Qint52 17 1026a 53a

(16) (05) (77) (35)MB

Qint158a 372a 1155a 09

(45) (56) (106) (06)

TangibilityQint 05 177 436c 64a

(01) (15) (18) (31)

Constant 167 96 656b 235 4613a 8610a 113 570a

(08) (04) (26) (04) (61) (58) (11) (31)

Year FE Yes Yes Yes Yes Yes Yes Yes Yes

Firm FE Yes Yes Yes Yes Yes Yes Yes Yes

R2 0002 0002 0003 0004 0013 0014 0013 0014

N 155587 155037 155587 155037 155587 155037 155587 155037

1438 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

7 Are Financial Market Conditions Important

According to the trade-off theory capital structure is determined by con-sidering the impact of costs and benefits of debt The time variation inthe costs and benefits of debt imply that issuance decisions would alsovary over time Furthermore market conditions affect the transactioncosts of adjusting Adverse selection is a greater problem in a cold marketthan in a hot one Accordingly it may be easier to adjust in good marketconditions Related ideas have been developed by Huang and Ritter(2009) Faulkender et al (2012) and Halling Yu and Zechner (2012)Faulkender Flannery Hankins and Smithrsquos paper examines cross-sectionalvariation in adjustment speeds including market valuations andhow they may affect adjustment speeds It is natural to think that ingood market conditions it will be less costly to issue both debt and equityand adjusting leverage to respond to profitability shocks would be mucheasierTo test the importance of market conditions we require a definition of

good times and bad times Our empirical strategy is to estimate good timesversus bad times at the four-digit industry level We define an industry ashaving ldquogood timesrdquo if the median firm in that industry has a market-to-book ratio that is higher than the 67th percentile of the time-series distribu-tion of the industry median market-to-book ratios Conversely an industryis defined as having ldquobad timesrdquo if the median firm in the industry has amarket-to-book ratio that is lower than the 33rd percentile of the time-seriesdistribution of the industry median market-to-book ratiosPanel A of Table IX tabulates issuance activity for profitability sorts in

both good times and bad times As might be expected active debt and equityissuances are larger during good times There is also somewhat more activeswaps between debt and equity during good timesPanel B reports the issuance activity for the smallest and the largest firms

during the good times and bad times Within each size quintile firms aresorted on profitability This is a three-way sort As before we find that thereis much more active use of external markets during good times Small low-profit firms are more likely to issue equity in good times than in bad timesDebt issuances are significantly higher in good times Large high-profitfirms are significantly more likely to issue debt and repurchase equity ingood times than in bad timesPanel C reports the magnitudes of the financial variables rather than the

frequencies During bad times less profitable large firms retire substantialamounts of debt and they show a tendency to issue equity Small firms donot seem to engage in similar debt reduction activities However like the

PROFITSndashLEVERAGE PUZZLEREVISITED 1439

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

Table IX Debt and equity issues in good and bad times

The table reports the frequency and magnitude of financing activity for sorts on profits for subsamples

of firms in good and bad times An industry is defined as having ldquogood timesrdquo if the median firm in

that industry has a market-to-book ratio that is higher than the 67th percentile of the time-series

distribution of industry median market-to-book ratios Conversely an industry is defined as having

bad times if the median firm in that industry has a market-to-book ratio that is lower than the 33rd

percentile of the time-series distribution of industry median market-to-book ratios Panels A and B

report the percentage of firms (i) issuing net debt in excess of 5 of the value of their assets (ii) issuing

net equity in excess of 5 of the value of their assets (iii) issuing debt and repurchasing equity both in

excess of 5 of the value of their assets and (iv) issuing equity and retiring debt both in excess of 5

of the value of their assets Panel C tabulates the mean debt and equity levels and the mean changes in

debt and equity issuances The sample contains nonfinancial firms listed on the annual Compustat files

for the period from 1971 to 2009

Iss Iss Iss D Iss ENet D Net E Rep E Ret D(1) (2) (3) (4)

Panel A sorts on profitability

Badtimes

Low 145 131 01 202 170 49 02 083 173 46 04 074 182 47 03 10

High 177 51 06 10t Highfrac14Low 41 136 33 01

Goodtimes

Low 206 381 01 462 197 148 02 233 212 103 04 184 234 92 10 18

High 216 92 16 14t Highfrac14Low 23 658 138 23

Panel B sorts on firm size and profitability

Badtimes

Smallfirm

s Low 151 318 00 312 182 276 02 253 151 140 02 294 183 73 00 19

High 177 80 00 11t Highfrac14Low 10 90 NA 20

Largefirm

s Low 172 54 01 042 148 39 04 053 150 43 04 044 156 33 03 06

High 147 28 05 03t Highfrac14Low 17 32 20 02

Goodtimes Smallfirm

s Low 213 448 01 542 230 443 01 473 222 325 00 474 203 197 02 38

High 190 153 02 31t Highfrac14Low 19 230 10 39

Largefirm

s Low 203 96 03 122 208 66 09 133 232 54 16 064 215 39 19 05

High 234 36 38 04t Highfrac14Low 29 99 92 38

(continued)

1440 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

larger firms there is a tendency for small firms to issue more equitymdashat leastto the extent possibleDuring good times large profitable firms raise significant amounts of debt

as they experience an increase in the value of their equity Such firms alsoengage in share repurchases Debt issues by smaller firms are much lessaffected by their own profitability during good times and their change indebt is more or less independent of their profits However equity issuanceis primarily found among the low-profit small firms during good timesTable IX shows that empirically profitability has a time-varying impacton leverage choicesIn Table X the basic regressions for changes in debt and for equity

issuance are presented for both good times and bad times Comparinggood times with bad times we see that the effects of profits are muchstronger in good times In bad times the impacts are rather weak statisticallyThis difference is interesting and deserves further study However for ourpurposes recall that the static trade-off theory implies time-varying capital

Table IX Continued

D D BVE BVE MVE MVE EqyIss Assets

(1) (2) (3) (4) (5) (6) (7) (8)

Panel C magnitude of financing activity

Badtimes Smallfirm

s Low 40 03 58 14 305 06 22 140

2 44 05 82 07 240 67 17 180

3 45 04 94 02 176 18 05 187

4 43 09 98 02 137 04 03 188

High 36 10 114 13 209 16 04 199

Largefirm

s Low 3365 73 2523 169 2304 163 94 8972

2 3031 95 2983 35 3284 278 56 9068

3 2930 5 3360 14 3947 308 57 9225

4 2281 6 3817 19 5213 442 7 8902

High 1790 62 4767 13 8799 58 31 9573

Goodtimes

Smallfirm

s Low 24 06 38 05 339 23 39 90

2 26 08 48 01 301 04 25 104

3 30 09 60 04 246 03 17 125

4 28 07 66 08 206 18 12 130

High 23 06 85 16 261 13 08 141

Largefirm

s Low 4532 36 3477 103 6557 260 53 12465

2 4580 58 3527 139 6911 421 23 12522

3 2952 111 3485 2 7586 279 32 9535

4 2752 124 3939 72 10686 33 105 10226

High 2429 228 4819 377 18565 268 222 10830

PROFITSndashLEVERAGE PUZZLEREVISITED 1441

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

structure choices even when the target is time invariant Empirically we doobserve time-varying choices

8 Is Scaling Important

In the preceding section we provided results for unscaled debt and equityissues This seems appropriate to us However it involves making twochanges to the standard regression We would like to determine if both

Table X Debt and equity changes in good and bad timesmdashregressions

The sample contains nonfinancial firms listed on the annual Compustat files for the period

from 1971 to 2009 The table presents estimates from regressions of change in debt (D)and net equity issuances (EqyIss) in good and bad times The explanatory variables includechange in profits lagged change in profits and lagged change in assets The regressions

include both firm and year fixed effects An industry is defined as having ldquogood timesrdquo ifthe median firm in that industry has a market-to-book ratio that is higher than the 67thpercentile of the time-series distribution of industry median market-to-book ratios

Conversely an industry is defined as having bad times if the median firm in thatindustry has a market-to-book ratio that is lower than the 33rd percentile of the time-series distribution of industry median market-to-book ratios The reported t-statistics arecorrected for clustering at the firm level

aMean significant at the 1 level

bMean significant at the 5 level

cMean significant at the 10 level

Good times Bad times

D EqyIss D EqyIss

(1) (2) (3) (4) (5) (6) (7) (8)

Profits 087a 091a 006b 009a 060 065 003 003b

(30) (29) (20) (28) (15) (16) (16) (20)

Profitst1 029c 015a 034c 001

(19) (37) (17) (04)

Assetst1 000 002b 000 001

(02) (22) (00) (07)

Constant 1337 3989 8691a 10001a 1351 1851 4099a 4055a

(03) (09) (55) (37) (09) (13) (80) (78)

Year FE Yes Yes Yes Yes Yes Yes Yes Yes

Firm FE Yes Yes Yes Yes Yes Yes Yes Yes

R2 0071 0080 0018 0045 0178 0212 0034 0041

N 74827 68729 74891 68787 26965 26013 26971 26018

1442 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

changes are crucial Accordingly in this section we take a step back toconsider prior literature in which variables are scaled The questions arehow does scaling affect the results and whyTable XI examines alternative scalings Columns (1) and (2) scale debt and

equity issuances by the total firm net issuance Columns (3) and (4) scaledebt and equity issuances by the sum of debt and book equity Columns (5)and (6) consider the change in debt scaled by the sum of debt and bookequity Columns (7) and (8) consider the change in leverage due to externalfinancing Since we are examining changes we use the changes in thestandard regression factors as controls The primary interest is in the coef-ficient on the change in profitsWe find that a change in profits is associated with an increase in debt

issues But the sign reverses when we scale the debt issues by the sum of debtplus book equity This is telling What is really driving the results is thechange in equity in the denominator not the change in debt in the numer-ator This result is further substantiated in Columns (5) and (6) where thechange in debt to capital ratio is used instead of debt issuanceTo further explore the impact of active issuance relative to passive

experiencing of changes we decompose leverage by separating out theeffect of debt and equity issues from those due to changes in retainedearnings Thus we define leverage change due to external financing as

LevEF frac14Dt1 thorn d

Dt1 thorn Et1 thorn dthorn e

Dt1

Dt1 thorn Et1 eth1THORN

where D is debt E is book value of equity d is debt issues net of retirementsand e is equity issues net of repurchases Garvey and Hanka (1999) use asimilar measure in their study This measure directly captures the netleverage effect of debt and equity issuancesColumns (7) and (8) report results from regressions of the net leverage

effect of debt and equity issues on changes in profitability after controllingfor other leverage factors in differences and year indicator variables Theeffect of profitability on leverage change due to external financing is positiveand statistically significant This comparison suggests that changes in totaldebt and equity are only partly a result of debt and equity issuance decisionsOther balance sheet adjustments complicate the inferences from leverageratio regressionsFor most of our analysis we follow the literature in treating debt as homo-

genous It is of course potentially interesting to ask what happens when wedistinguish between short-term debt and long-term debt Which part of thefirmrsquos debt maturity structure drives the under-adjustment Table XII

PROFITSndashLEVERAGE PUZZLEREVISITED 1443

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

provides results from such decompositions Alternative scalings areconsidered The main results of interest are that the distinction betweenshort- and long-term debt is important Short-term debt issuance is largelyindependent of changes in profits The negative sign on profit change stemsfrom the action of long-term debt This seems intuitively reasonable

Table XI Scaled issuances

The sample contains nonfinancial firms listed on the annual Compustat files for the period

from 1971 to 2009 The table presents estimates from regressions of debt issuance to totalissuance (in columns (1) and (2)) debt issuance to total capitalization (in columns (3) and(4)) change in debt to capitalization ratio (in columns (5) and (6)) and change in leverage

ratio considering only external financing (in columns (7) and (8)) The change in leverageratio considering only external financing is defined as

LevEF frac14Dt1 thorn d

Dt1 thorn Et1 thorn dthorn e

Dt1

Dt1 thorn Et1

where D is debt E is book value of equity d is debt issues net of retirements and e is equity

issues net of repurchases The regressions include lagged leverage factors and year indicatorvariables Some of the specifications include firm fixed effects The reported t-statistics arecorrected for clustering at the firm level

aMean significant at the 1 level

bMean significant at the 5 level

cMean significant at the 10 level

DebtIssTotIss

DebtIssethDthornETHORN D

DthornE LevEF

(1) (2) (3) (4) (5) (6) (7) (8)

Profitt1 004a 000 002a 001c 003a 002 001b 002a

(35) (03) (34) (18) (35) (14) (23) (26)

IndLevt1 004c 001 001 001c 000 000 000 000

(19) (06) (15) (19) (02) (04) (03) (05)

Market=bookt1 001a 001a 000a 000a 000a 000a 000 000c

(47) (85) (68) (60) (55) (36) (05) (18)

Tangt1 004 005 013a 011a 011a 006a 003a 003a

(12) (15) (144) (112) (79) (38) (44) (42)

Assetst1 008a 004a 006a 004a 002a 001a 003a 002a

(97) (45) (283) (161) (53) (35) (147) (101)

Constant 079a 074a 002a 003a 000 000 000b 002a

(497) (442) (91) (111) (00) (05) (22) (93)

Year FE Yes Yes Yes Yes Yes Yes Yes Yes

Firm FE No Yes No Yes No Yes No Yes

R2 0015 0098 0022 0076 0009 0070 0012 0086

N 133905 133905 140552 140552 138229 138229 140572 140572

1444 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

inasmuch as short-term debt is probably more associated with day-to-dayoperations rather than longer term planning considerationsIn Table XIII we examine change in leverage ( D

DthornE) and change inleverage due to external financing (LEVEF) during good and bad timesseparately The traditional negative effect of a change in profits on achange in leverage is found both in good times and in bad timesHowever when we decompose the impact in order to isolate the externalcomponent as before the result disappears In good times the externalfinancing component of a change in profits is weak but still positiveDuring bad times it is not statistically significant This implies that distin-guishing between good and bad times does not fundamentally alter the basic

Table XII Short- and long-term debt issuances

The sample contains nonfinancial firms listed on the annual Compustat files for the period

from 1971 to 2009 The table presents estimates from regressions of short-term debtissuance to total issuance (in columns (1) and (2)) long-term debt issuance to totalissuance (in columns (3) and (4)) short-term debt issuance to total capitalization (in

columns (5) and (6)) and long-term debt to total capitalization (in columns (7) and (8))Some of the specifications include firm fixed effects The reported t-statistics are correctedfor clustering at the firm level

aMean significant at the 1 level

bMean significant at the 5 level

cMean significant at the 10 level

STDebtIssTotIss

LTDebtIssTotIss

STDebtIssethDthornETHORN

LTDebtIssethDthornETHORN

(1) (2) (3) (4) (5) (6) (7) (8)

Profitabilityt1 0002 0003 0042a 0008 0001 0001 0019a 0012a

(02) (04) (30) (05) (14) (10) (45) (28)

IndLevt1 0009 0007 0070b 0033 0001 0002c 0014a 0005

(06) (05) (26) (11) (08) (17) (26) (09)

M=Bt1 0003a 0003a 0003b 0010a 0000 0000 0003a 0002a

(48) (35) (22) (65) (10) (01) (84) (42)

Tangibilityt1 0020 0007 0011 0067 0005a 0006a 0097a 0081a

(12) (03) (03) (16) (34) (33) (108) (89)

Assetst1 0039a 0011b 0034a 0024b 0001a 0001 0046a 0020a

(89) (22) (38) (23) (36) (13) (190) (82)

Constant 0001b 0153a 0798a 0591a 0000c 0006a 0057a 0096a

(25) (144) (470) (288) (17) (124) (292) (234)

Year FE Yes Yes Yes Yes Yes Yes Yes Yes

Firm FE No Yes No Yes No Yes No Yes

R2 0025 0089 0018 0073 0030 0147 0016 0298

N 136146 136146 136140 136140 143031 143031 142847 142847

PROFITSndashLEVERAGE PUZZLEREVISITED 1445

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

Table XIII Externally financed leverage changes in good and bad times

The sample contains nonfinancial firms listed on the annual Compustat files for the period

from 1971 to 2009 The table presents estimates from regressions of change in debt tocapitalization ratio (in columns (1) (2) (5) and (6)) and change in leverage ratio consider-ing only external financing (in columns (3) (4) (7) and (8)) The change in leverage ratio

considering only external financing is defined as

LevEF frac14Dt1 thorn d

Dt1 thorn Et1 thorn dthorn e

Dt1

Dt1 thorn Et1

where D is debt E is book value of equity d is debt issues net of retirements and e is equityissues net of repurchases The regressions include lagged leverage factors and year indicator

variables The reported t-statistics are corrected for clustering at the firm level An industryis defined as having ldquogood timesrdquo if the median firm in that industry has a market-to-bookratio that is higher than the 67th percentile of the time-series distribution of industry

median market-to-book ratios Conversely an industry is defined as having ldquobad timesrdquoif the median firm in that industry has a market-to-book ratio that is lower than the 33rdpercentile of the time-series distribution of industry median market-to-book ratios

aMean significant at the 1 level

bMean significant at the 5 level

cMean significant at the 10 level

Good times Bad times

DDthornE LevEF D

DthornE LevEF

(1) (2) (3) (4) (5) (6) (7) (8)

Profitabiltyt1 0022c 0005 0016c 0025b 0073b 0041 0007 0007

(17) (03) (18) (24) (25) (11) (04) (03)

IndLevt1 0000 0011 0001 0009 0010 0018 0001 0000

(00) (09) (01) (13) (07) (10) (02) (00)

MB

t1

0004a 0003b 0000 0001 0008a 0001 0000 0000

(40) (24) (05) (08) (28) (04) (03) (03)

Tangt1 0097a 0041 0033a 0038b 0127a 0048 0059a 0033

(45) (16) (27) (25) (49) (14) (43) (16)

Assetst1 0014a 0013b 0024a 0019a 0037a 0023c 0036a 0025a

(30) (20) (84) (48) (45) (17) (88) (43)

Constant 0030b 0036 0000 0003 0028c 0014 0021 0001

(20) (16) (00) (07) (17) (06) (11) (01)

Year FE Yes Yes Yes Yes Yes Yes Yes Yes

Firm FE No Yes No Yes No Yes No Yes

R2 0006 0088 0009 0117 0026 0251 0023 0202

N 56820 56820 58200 58200 23462 23462 23511 23511

1446 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

conclusions The traditional leverage result on profits is driven by increasesin equity that are experienced and then partially but not completely offsetby firmsTo the extent that firms are engaging in active financing they are following

the prescriptions of the trade-off theory Thus we have a coherent picture ofhow firms are responding to profits Basically more profitable firms experiencean increase in the value of their equity They respond to this by issuing debtand repurchasing equity But the magnitude of the increase in equity is largerthan the magnitudes of the active steps taken in the debt and equity marketsThis means that the firmsrsquo actual actions are as predicted by the trade-offtheory But it leaves open the question of why they do not act more stronglyThe capital structure literature has long been interested in dynamics and

capital structure adjustment paths In recent years this interest has increas-ingly resulted in explicit dynamic optimizing models such as those ofGoldstein Ju and Leland (2001) Hennessy and Whited (2005) andStrebulaev (2007) These models commonly make specialized assumptionsleading some to question their generalityWe argue that both fixed transactions costs and variable transactions costs

are plausible and realistic in models that pin down firm size The simultaneouspresence of both is important Due to fixed costs firms only rebalance inter-mittently Due to variable adjustment costs the extent of the adjustment isreduced to the point at which an extra unit to adjustment is balanced againstthe extra unit transactions costs This means that both leverage increases andleverage reductions will only go part way toward a static optimumThe key point is that with both fixed and variable transactions costs we

should expect to only see intermittent leverage adjustments When leverage isadjusted it should ldquounder adjustrdquo when compared to a frictionless modelFurther discussion of transactions costs in a capital structure setting isprovided by Leary and Roberts (2005)

9 Conclusion

The connection between corporate profits and capital structure has beenvery influential in the assessment of the trade-off theory The conventionalinterpretation of the evidence has pushed the literature away from the trade-off theory and toward much more complex dynamic models and ideas As aresult it is important to make sure that the evidence has been correctlyinterpreted We show that the problem is with the interpretation of thetest results and a failure to take into account the costs associated with re-sponding to a profit shock

PROFITSndashLEVERAGE PUZZLEREVISITED 1447

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

We document that more profitable firms really do borrow more (not less)They tend to repurchase equity They experience an increase in both thebook value of equity and the market value of equity Less profitable firmsreally do tend to reduce their debt and to issue equityFirm size and market conditions also matter Larger firms tend to be more

active in the debt markets while smaller firms tend to be relatively moreactive in the equity markets During good times there is more use ofexternal financingThe usual profitsndashleverage puzzle result is primarily driven by the increase

in equity that is experienced by the more profitable firms Accordingly thepuzzle should be restated as asking why do firms not take sufficiently largeoffsetting actions to fully undo the change in equity What limits the mag-nitudes of the typical leverage response to profit shocksIn a frictionless model the partial response appears puzzling But there is

good empirical reason to believe that rebalancing entails both fixed andvariable costs and that firm size matters The rebalancing costs can befully avoided by doing nothing Accordingly the firm must decide whetherany given shock is big enough to be worth responding to If it is then thefirm must decide how big a response is called for These technical conditionsare known as ldquovalue matchingrdquo and ldquosmooth pastingrdquo Thus optimizationimplies that some shocks will be ignored Even if the shock is not ignoredthe optimal response will only partially undo the shock The magnitude ofthe leverage response must balance the marginal cost and the marginalbenefit of an extra unit of leverage Since the marginal cost of adjustingleverage is strictly positive the adjustment toward that static leverageoptimum will only go part way This is true both for leverage increasesand for leverage reductionsOne might conjecture that the variable costs would be time varying and

that such costs might have declined over time Since we do not have a fullaccounting of all of the variable costs it is of course impossible to be sureHowever there is some evidence suggesting that at least the directunderwriting fees are remarkably stable over time10

Surely nobody ought to be surprised that more profitable firms are moreinclined to repurchase equity while unprofitable firms tend not to do so Thefact that more profitable firms typically experience an increase in the value of

10 For IPOs Jay Ritter provides important summary evidence as httpbearwarringtonufleduritterIPOs2013Underwritingpdf He finds that for medium-sized deals a remark-

ably large fraction of the deals have a 7 fee year after year Whether this is due tobehavioral factors or some sort of a coordination device is not clear The extent towhich similar intertemporal stability carries over to other forms of external financemerits further study

1448 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

equity is equally natural Were it not for the contrary claims in the literature(ie the profitsndashleverage puzzle) we would argue that nobody ought to findit surprising that more profitable firms are more inclined to issue debt Thefact that leverage adjustments typically only partially offset profit shocksshould also sound quite natural These facts all fit together easily within thetrade-off theory once we take account of both fixed and variable costs ofactively adjusting capital structure

Appendix

Table AI Market leverage regressions

The table presents estimates of the regressions of market leverage on firm characteristics

The sample comes from the annual Compustat files during the period 1971ndash2009 Financialfirms are excluded The leverage is estimated as the ratio of debt over debt plus marketequity The explanatory variables Profitabilityt1

MB

t1

Tangibilityt1 and LnethAssetsTHORNt1are described in Table 1 IndMedianLevt1 is estimated as the median market leverage of allother firms in the same industry The industry is defined at the level of the four-digit SICcode All specifications include the year fixed effects The specifications in column (2)

additionally include the firm fixed effects We report t-statistics where the standard errorsare clustered at the firm level in parentheses

aSignificant at the 1 level

Quantile regressions

(1) (2) (3) (4) (5)

Profitabilityt1 013a 013a 007a 011a 020a

(275) (262) (349) (425) (538)

IndLevt1 046a 020a 032a 058a 063a

(506) (239) (1299) (1630) (1042)MB

t1

003a 001a 001a 002a 004a

(509) (303) (743) (709) (534)

Tangibilityt1 013a 018a 013a 014a 010a

(191) (181) (762) (571) (232)

LnethAssetsTHORNt1 001a 005a 001a 001a lt001a

(112) (258) (682) (342) (29)

Constant 010a 011a 005a 005a 029a

(171) (98) (173) (118) (398)

Year FE Yes Yes Yes Yes Yes

Firm FE No Yes No No No

Regression type OLS FE 25th tile Median 75thtile

Clustered SE Yes Yes No No No

R2 030 067

Pseudo R2 016 023 0201

N 158578 158578 158578 158578 158578

PROFITSndashLEVERAGE PUZZLEREVISITED 1449

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

Table AII Profitability sorts for firms issuing large amounts

Percentage of firms

Iss Ret Iss Iss Rep Iss Iss IssD IssE

D D NetD E E NetE DampE RepE RetD

(1) (2) (3) (4) (5) (6) (7) (8) (9)

Panel A 10 cutoff to define issuing or retiring activity

Allfirm

s

Low 205 150 134 260 05 257 45 00 23

2 207 175 103 77 06 75 11 01 10

3 243 196 101 50 08 47 08 01 09

4 252 186 113 50 13 48 08 02 08

High 227 161 118 60 27 57 07 04 09

t Highfrac14Low 70 38 62 755 226 768 320 90 152

Smallfirm

s Low 230 132 162 380 05 375 82 00 32

2 228 143 165 334 04 333 65 00 26

3 211 146 147 216 03 214 38 00 23

4 190 154 120 126 04 124 15 01 20

High 194 162 114 99 09 97 12 01 15

t Highfrac14Low 42 40 68 342 20 340 169 07 131

Largefirm

s Low 231 200 82 31 06 29 08 01 04

2 231 187 67 17 06 14 03 01 02

3 234 177 77 17 09 14 03 02 03

4 218 158 75 16 17 14 02 03 02

High 218 135 99 19 41 15 03 07 02

t Highfrac14Low 19 106 36 46 135 57 39 58 23

Percentage of firms

Iss Ret Iss Iss Rep Iss Both Iss D Iss E

D D Net D E E Net E DampE Rep E Ret D

(1) (2) (3) (4) (5) (6) (7) (8) (9)

Panel B 15 cutoff to define issuing or retiring activity

Allfirm

s

Low 158 101 100 230 01 228 27 00 15

2 142 115 63 57 01 56 05 00 06

3 165 128 57 32 01 31 03 00 05

4 173 118 65 35 01 34 03 00 05

High 160 102 70 46 04 44 03 00 05

t Highfrac14Low 08 06 141 738 87 746 277 24 131

Smallfirm

s Low 195 95 132 354 01 350 56 00 24

2 185 96 131 303 01 301 44 00 18

3 160 96 104 179 00 176 18 00 15

4 134 99 78 99 01 98 08 00 12

High 133 95 73 82 01 80 06 00 10

t Highfrac14Low 80 00 95 345 12 344 149 12 52

Largefirm

s Low 145 123 42 14 01 12 03 00 01

2 137 114 32 07 01 06 01 00 01

3 142 104 35 06 01 05 00 00 01

4 130 91 36 08 03 07 01 01 01

High 134 81 48 09 06 07 01 01 01

t Highfrac14Low 19 85 17 26 54 33 34 17 11

1450 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

Table AIV Debt and equity changesmdashwithout fixed effects

The sample contains nonfinancial firms listed on the annual Compustat files for the period from 1971

to 2009 The table presents estimates from regressions of change in debt (D) change in the book

value of equity (BVE) change in the market value of equity (MVE) and net equity issuances

(EqyIss) The explanatory variables include current and lagged change in operating income before

depreciation (Profits and Profitst1) and lagged change in book value of assets (Assetst1) The

regressions include year indicator variables The reported t-statistics are corrected for clustering at the

firm level

aMean significant at the 1 level

bMean significant at the 5 level

cMean significant at the 10 level

D BVE MVE EqyIss

(1) (2) (3) (4) (5) (6) (7) (8)

Profitst1 025a 024a 023a 038a 007 066a 007a 009a

(40) (32) (36) (39) (06) (29) (42) (51)Profits 051b 082a 149a 004a

(25) (33) (49) (33)Assetst1 002 005c 024a 001b

(08) (17) (34) (23)Constant 93 176a 45 71 1552a 1704a 130a 140a

(12) (28) (07) (09) (49) (49) (45) (45)Year FE Yes Yes Yes Yes Yes Yes Yes YesR2 001 005 001 011 001 005 002 002N 162056 162056 162130 162130 157550 157550 162154 162154

Table AIII Median profitability sorts

The sample contains nonfinancial firms listed on the annual Compustat files for the period from 1971

to 2009 The table presents median debt and equity median changes in debt and equity and median

equity issuances (EqyIss) in millions of dollars for all firms sorted on profitability and for profitability

sorts within the smallest and largest firms The sorts are done annually

D D BVE BVE MVE MVE EqyIss Assets(1) (2) (3) (4) (5) (6) (7) (8)

Allfirm

s

Low 21 00 75 17 256 20 01 1922 186 02 449 18 598 19 00 11033 816 06 1274 03 1637 05 00 32244 787 02 1541 18 2488 03 00 3556High 273 00 1361 66 3300 14 00 2578

Smallfirm

s Low 05 00 13 07 140 14 06 402 06 00 25 08 114 12 01 563 09 00 32 05 85 07 00 674 12 00 39 01 68 03 00 79High 10 00 51 04 98 01 00 90

Largefirm

s Low 11485 381 10269 371 13227 202 03 333902 13051 393 12786 140 17571 169 05 392463 11554 287 13429 24 21623 305 00 370314 9963 209 15052 194 29424 308 00 36705High 7361 00 16380 613 49142 439 00 34976

PROFITSndashLEVERAGE PUZZLEREVISITED 1451

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

References

Auerbach A J (1985) Real determinants of corporate leverage in B M Friedman (ed)

Corporate Capital Structures in the United States University of Chicago Press Chicago

IllinoisBertrand M and Mullainathan S (2003) Enjoying the quiet life corporate governance and

managerial preferences Journal of Political Economy 111 1043ndash1075Booth L Varouj A Demirguc-Kunt A and Maksimovic V (2001) Capital structures in

developing countries Journal of Finance 56 87ndash130Brennan M J and Schwartz E S (1984) Optimal financial policy and firm valuation

Journal of Finance 39 593ndash607Cameron A C and Trivedi P K (2010) Microeconometrics using Stata Stata Press TexasConstantinides G M (1979) Multiperiod consumption and investment behavior with

convex transactions costs Management Science 25 1127ndash1137Constantinides G M and Richard S F (1978) Existence of optimal simple policies for

discounted-cost inventory and cash management in continuous time Operations Research

26 620ndash636DeAngelo H DeAngelo L and Whited T M (2011) Capital structure dynamics and

transitory debt Journal of Financial Economics 99 235ndash261Dixit A K (1993) The Art of Smooth Pasting Routledge London and New YorkFama E and French K R (2002) Testing trade-off and pecking order predictions about

dividends and debt Review of Financial Studies 15 1ndash33Faulkender M and Petersen M A (2006) Does the source of capital affect capital struc-

ture Review of Financial Studies 19 45ndash79Faulkender M Flannery M J Hankins K W and Smith J M (2012) Cash Flows and

Leverage Adjustments Journal of Financial Economics 103 632ndash646Fischer E Heinkel R and Zechner J (1989) Dynamic capital structure choice theory

and tests Journal of Finance 44 19ndash40Frank M Z and Goyal V K (2008) Trade-off and pecking order theories of

debt in B E Eckbo (ed) Handbook of Corporate Finance Empirical Corporate Finance

Vol 2 Handbook of Finance Series Chapter 12 ElsevierNorth-Holland Amsterdam pp

135ndash202

Frank M Z and Goyal V K (2009) Capital structure decisions which factors are reliably

important Financial Management 38 1ndash37

Garvey G T and Hanka G (1999) Capital structure and corporate control the effect of

antitakeover statutes on firm leverage Journal of Finance 54 519ndash546

Goldstein R Ju N and Leland H (2001) An EBIT-based model of dynamic capital

structure Journal of Business 74 483ndash512

Graham J R and Tucker A L (2006) Tax shelters and corporate debt policy Journal of

Financial Economics 81 563ndash594

Hackbarth D and Mauer D C (2012) Optimal priority structure capital structure and

investment Review of Financial Studies 25 747ndash796

Halling M Yu J and Zechner J (2012) Leverage dynamics over the business cycle

Working paper University of Utah University of New South Wales and Vienna

University of Economics and Business AdministrationHennessy C A and Whited T A (2005) Debt dynamics Journal of Finance 60 1129ndash1165

Huang R and Ritter J R (2009) Testing theories of capital structure and estimating the

speed of adjustment Journal of Financial and Qualitative Analysis 44 237ndash271

1452 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

Hovakimian A Opler T and Titman S (2001) The debt-equity choice Journal of

Financial and Quantitative Analysis 36 1ndash24Jung K Kim Y and Stulz R M (1996) Timing investment opportunities managerial

discretion and the security issue decisions Journal of Financial Economics 42 159ndash185Kane A Marcus A J and McDonald R L (1984) How big is the tax advantage to

debt Journal of Finance 39 841ndash853Kayhan A and Titman S (2007) Firmsrsquo histories and their capital structures Journal of

Financial Economics 83 1ndash32Leary M T (2009) Bank loan supply lender choice and corporate capital structure

Journal of Finance 64 1143ndash1185Leary M T and Roberts M R (2005) Do firms rebalance their capital structures

Journal of Finance 60 2575ndash2619Leland H E (1994) Corporate debt value bond covenants and optimal capital structure

Journal of finance 49 1213ndash1252Lemmon M L Roberts M R and Zender J F (2008) Back to the beginning persist-

ence and the cross-section of corporate capital structure Journal of Finance 63 1575ndash1608Long M S and Malitz I B (1985) The investment-financing nexus some empirical

evidence Midland Corporate Finance Journal 3 53ndash59Mackie-Mason J K (1990) Do taxes affect corporate financing decisions Journal of

Finance 45 1471ndash1493Morellec E Nikolov B and Schurhoff N (2012) Corporate governance and capital

structure dynamics Journal of Finance 67 803ndash848Myers S C (1984) The capital structure puzzle Journal of Finance 39 575ndash592Myers S C (1993) Still searching for optimal capital structure Journal of Applied

Corporate Finance 61 4ndash14Rajan R and Zingales L (1995) What do we know about capital structure some evidence

from international data Journal of Finance 50 1421ndash1460Strebulaev I A (2007) Do tests of capital structure theory mean what they say Journal of

Finance 62 1747ndash1787

Strebulaev I A and Whited T M (2012) Dynamic Models and Structural Estimation in

Corporate Finance Now Publisher Boston and Delft

Titman S and Wessels R (1988) The determinants of capital structure choice Journal of

Finance 43 1ndash21

Welch I (2004) Capital structure and stock returns Journal of Political Economy 112

106ndash131

Welch I (2011) Two common problems in capital structure research the financial-debt-

to-asset ratio and issuing activity versus leverage changes International Review of Finance

11 1ndash17

PROFITSndashLEVERAGE PUZZLEREVISITED 1453

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

standard dynamic optimization conditions For a reader who wants a moreexplicit treatment of the mathematics we recommend Dixit (1993) The ob-servation that this mathematics can be interpreted in terms of corporateleverage in which firm size matters is also not novel It can be found inLeary and Roberts (2005) who provide a simulation with both fixed costsand a weakly convex adjustment costs What is new here is the analysis of theimplication for the impact of profits on leverageSuppose that without active rebalancing leverage would just wander

about due to changes in the stock market value of equity According tothe trade-off theory excessively high leverage is costly (bankruptcy andagency costs) as is excessively low leverage (lost tax benefits and free cashflow problems) If there are no transactions costs then under conventionaltrade-off theory assumptions there will be a unique leverage target that willbe maintained at all timesDue to the fixed costs such frequent rebalancing would be too costly As a

result the firm permits the leverage to drift However if the drift is tooextreme in either direction then the firm will actively rebalance in the direc-tion of the static optimum If the rebalanced firm (inclusive of rebalancingcosts) is worth more then rebalancing is worth it If the rebalanced firmwould be worth less then rebalancing is not worth it Rebalancing takesplace at the point at which the values just matchWhen the firm is rebalancing it also needs to worry about how far to

move in the desired direction When each step has costs the firm needs toensure that the marginal cost of the last step is just equal to the marginalbenefit As a result both from excessively high leverage and from excessivelylow leverage the movements are only part wayTo fix ideas consider an off-the-shelf model of the leverage ratio The

focus is on the impact of the adjustment cost structure The formal structureis essentially the cash management problem studied by Constantinides andRichard (1978) and Dixit (1993) We follow the presentation and notation inDixit (1993) very closely to show that this is just an application of well-established ideas to the leverage problemLet xt denote the leverage ratio at date t is the interest rate Ku is a fixed

cost of actively increasing leverage Kd is the fixed cost of actively reducingleverage mu is the marginal cost of each unit of active leverage increase md isthe marginal cost of each unit of active leverage decrease f(x) is the flowpayoff if leverage is at x F(x) is the expected present value of the payoff ifleverage is at x x is the lower adjustment threshold x is the lower targetleverage (when adjusting from below) x is the upper adjustment thresholdand x is the upper target leverage (when adjusting from above) Accordingto the trade-off theory F 0 gt 0 F 00 lt 0

PROFITSndashLEVERAGE PUZZLEREVISITED 1421

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

In the absence of active adjustments the leverage ratio xt follows astandard process

dx frac14 ethx tTHORNdtthorn ethx tTHORNdw

where is the trend coefficient is the volatility coefficient and w is anormal shock process with zero mean and unit variance If x lt x lt x thenfinancial inaction is optimal If x drops to x then it is reset to x and if xrises to x then it is reset to xThe trade-off theory is frequently thought of as a class of theories in which

firm value first increases in leverage and ultimately drops if leverage gets toohigh This can be motivated by taxes and bankruptcy or by various agencyconcerns A range of ideas can be allowed for when we write that the flowpayoff (dividends andor profits) is fethxtTHORN and it depends on the leverage statext

5 If the firm starts at x0frac14 x then F(x) is the expected present value of theflow payoffs The discount rate is which allows us to write

FethxTHORN frac14 E

Z 10

fethxtTHORNetdtjx0 frac14 x

To maximize the firm value it must choose a leverage policy that sayswhen to rebalance and how much to rebalance These are given by thefamous value matching and smooth pasting conditionsThe value matching conditions are as follows

FethxTHORN FethxTHORN frac14 Ku thornmuethx xTHORN

FethxTHORN FethxTHORN frac14 Kd thornmdethx xTHORN

The smooth pasting conditions are as follows

F 0ethxTHORN frac14 F 0ethxTHORN frac14 mu

F 0ethxTHORN frac14 F 0ethxTHORN frac14 md

As explained by Constantinides and Richard (1978) with standard as-sumptions about the shape of F there are a pair of lines at heights mu andmd such that they intersect the function F 0ethxTHORN in such a way as to define thepoints x lt x lt x lt xTo sum up in this trade-off model with both fixed and variable

rebalancing costs there are four critical leverage points x gt x gt x gt x

5 For simplicity we suppose that the flow payoff is a continuous function of leverageThere is no discrete exit or bankruptcy point This allows us to simply interpret theConstantinides and Richard model much as in Leary and Roberts (2005)

1422 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

If leverage is between x (upper boundary) and x (lower boundary) the firmdoes nothing active Changes in leverage reflect various shocks to revenuesand costs that hit the firm If leverage drifts up to x (or above) then the firmrebalances to x If leverage drifts down to x (or below) the firm rebalancesto xWhat does this imply empirically A firm that does nothing can be

anywhere within the interval between x and x A firm that is rebalancingis a firm that has just hit one of the outer boundaries How does profitabilityaffect the likelihood of hitting a boundary High profits increase the value ofequity This increases the denominator of the leverage ratio and thus makesit more likely that the firm hits the lower boundary Low profits reduce thevalue of equity and thus make it more likely that the firm hits the upperboundaryTo be more specific suppose that the costs are symmetric and the leverage

changes are also following an ordinary Brownian motion Then the median

inert firm will have leverage ofxthornx

2 A firm that just actively reduced leverage

will be at x gtxthornx

2 A firm that just actively increased leverage will be at

x ltxthornx

2 A firm that just reduced leverage will generally be a firm that had

very high leverage To the extent that this reflects a lack of activerefinancing such firm will have low value of equity The low value ofequity will generally reflect low profits A firm that just increased leveragewill generally be a firm that had very low leverage In the same way a firmthat just increased its leverage will tend to have had high value of equity dueto high operating profits6

In this setting it is natural to focus on firms that are actively adjusting andto compare them to each other and to firms that are inactive It is alsohelpful to look directly at financing decisions and profitability rather thanfocusing on a leverage ratioOur approach shares the perspective of Welch (2004) Leary and Roberts

(2005) and Strebulaev (2007) that it is very important to focus on activeadjustments by firms Welch (2004) transforms the data to put everythingback into a ratio form His ldquoimplied debt ratiosrdquo still have both debt andequity in the denominator Similarly Leary and Roberts (2005) andStrebulaev (2007) pay a great deal of attention to leverage ratios whenstudying financing decisions Strebulaev (2007) does not have fixed costssince he makes use of a rescaling property Empirically we find differentadjusting behavior by large and by small firms So the rescaling property is

6 If we knew the values of the critical leverage points then it would be possible to derivemore refined tests Unfortunately those critical values are not observable

PROFITSndashLEVERAGE PUZZLEREVISITED 1423

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

an issue of concern His model does not account for the negative relationshipbetween current leverage and past market-to-book ratioWe depart from the earlier literature by arguing that the use of leverage

ratios can be particularly misleading for profitability since the use of aleverage ratio causes identification problems It is hard to be sure whethera leverage change reflects changes in the numerator or the denominator Thisis particularly problematic when as here the rebalancing is partialIt should be noted that much empirical capital structure literature debates

about whether to use a book leverage ratio or a market leverage ratio Bookleverage is used based on the argument that managers borrow against assetsin place which are better reflected by book assets Others such as Welch(2004) argue that book values are essentially an accounting fiction and thatonly market leverage should count Empirical papers regularly attempt tosidestep this debate by carrying out tests on both book leverage and marketleverage Our point is that both kinds of leverage ratios are potentially mis-leading when examining the impact of profits on capital structure

3 Data

We use conventional data sources starting with the merged Compustat-CRSP data The data are annual and are converted into constant year2000 dollars using the GDP deflator We exclude financial companies (SIC6000-6999) firms involved in major mergers (Compustat footnote code AB)firms reporting financial data in currencies other than the US dollars andfirms with missing data on our key variables7 The ratio variables aretrimmed at the 1 level in both tails of the distribution This serves toremove outliers and the most extremely misrecorded data The finalsample consists of 179021 firm-year observations from 1971 to 2009Table I provides definitions of financial variables and reports summary

statistics The average debt (in constant US$) is about $653 million while themedian is $24 million A significant fraction of firms have zero debt (the 10thpercentile is 0) Book equity is slightly larger than book debt Market equityis almost three times larger than book debt Book assets average $2191million although the medians are considerably smallerIf issuing or retiring securities incurs no fixed costs then we would expect

to see many small actions and very few large actions (Leary and Roberts2005) If there were significant fixed costs involved in issuing or retiringoutstanding securities then small issues might not be worthwhile Table I

7 These include debt book value of equity market value of equity assets book andmarket leverage profitability market-to-book assets ratio and tangibility

1424 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

shows that although most firms issue little debt or equity in a given year theaverages are large In other words when firms actually enter debt and equitymarkets they intervene massivelyThe mean constant dollar debt issue is $165 million (the median is $2

million) In unreported tables we find that the average debt issue is about81 of assets (the median is 18) About 38 of the firms issue no debt8 issue between 0 and 1 of the value of their assets as debt another

Table I Data description

Data sources The sample comes from the annual Compustat files The sample period is

1971ndash2009 We exclude financial companies (SIC 6000-6999) firms involved in majormergers (Compustat footnote code AB) firms reporting financial data in currencies otherthan the US dollar and firms with missing data on the key variables All financial variables

are deflated to year 2000 using the GDP deflator N is the number of observations SD isthe standard deviation D BVE MVE Assets Debt issuance Equity issuance Debt re-payment Equity repayment are all measured in $ millions The ratio variables are trimmed

at the 1 level in both tails of the distribution

Variable definitions Debt (D)frac14 long-term debt (dltt)thorn short-term debt (dlc) book equity(BVE)frac14 common shareholder equity (ceq) market equity (MVE)frac14 no of outstanding

shares (csho) closing price ( prcc f ) assetsfrac14 book assets (at) debt issuancefrac14 issuance oflong-term debt (Maxethdltis 0THORN)thorn increase in current debt (Maxethdlcch 0THORN) equity issu-ancefrac14 sale of common stock (Maxethsstk 0THORN) debt repaymentfrac14 reduction of long-termdebt (Maxethdltr 0THORN)thorn decrease in current debt (Minethdlcch 0THORN) equity repurchasefrac14 purchase

of common stock (Maxethprstkc 0THORN) book leveragefrac14D(DthornBVE) market leveragefrac14D(DthornMVE) profitabilityfrac14EBITDA (oibdp)assets market-to-book ratio M

B frac14Marketvalue of assets (MVA)assets where MVAfrac14DthornMVEthorn preferred-liq value (pstkl)

deferred taxes (txditc) tangibilityfrac14 net property plant and equipment (ppent)assets

Variable N Mean SD

Distribution

10th 50th 90th

Debt (D) 179021 653 4592 0 24 1048

Book equity (BVE) 179021 815 4122 3 70 1373

Market equity (MVE) 179021 1784 9589 9 118 2698

Assets 179021 2191 11530 10 153 3575

Debt issuance 179021 165 1335 0 2 245

Equity issuance 179021 26 186 0 0 40

Debt repayment 179021 132 1123 0 3 181

Equity repurchase 179021 24 257 0 0 9

Book leverage 179021 036 034 000 031 072

Market leverage 179021 027 025 000 020 066

Profitability 179021 005 027 018 011 023MB 179021 165 206 052 102 324

Tangibility 179021 031 024 005 026 070

PROFITSndashLEVERAGE PUZZLEREVISITED 1425

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

16 issue between 1 and 5 of the value of their assets as debt and theremaining 38 issue debt in excess of 5 of the value of their assetsThe mean constant dollar equity issue is about $26 million (the median is

about $04 million) As a fraction of assets the mean and median equityissues are about 7 and 02 respectively About 33 of the firms issue noequity 34 of the firms issue between 0 and 1 of the value of theirassets as equity another 14 issue between 1 and 5 of the value of theirassets as equity and the remaining 19 of the firms issue equity that is inexcess of 5 of the value of their assets Average debt repayments are largerthan equity repurchases This perhaps reflects the finite maturity of debt andits contractual repayment The median firm does not repurchase equityWe construct both book and market leverage ratios Book leverage is

defined as debt over debt plus book equity Market leverage is defined asdebt over debt plus market equity8 The median book leverage is 036 (theaverage is 031) The median market leverage is 027 (the average is 020)Profitability is defined as the ratio of operating income before depreci-

ation to assets While the average firm is profitable (the ratio of EBITDA toassets is 005) the median firm is even more profitable (with a profitabilityratio of 011) But the sample also includes a large number of unprofitablefirms as the 10th percentile is 018 The table also reports descriptive stat-istics on the market-to-book ratio and the tangibility ratio The market-to-book ratio (MB) defined as the ratio of the market value of assets to bookassets averages at about 165 Tangibility defined as the ratio of netproperty plant and equipment to assets averages at about 31

4 Leverage Ratio Regressions

The previous literature focuses on estimates obtained using leverage ratiosHence we start with a similar estimation to check whether our results matchthose of previous studies Table II presents the results for book leverage Weobtain similar results for market leverage and they are presented inAppendix Table AI The regressions include leverage factors following theprevious capital structure literature (see eg Lemmon Roberts and Zender(2008) and Frank and Goyal (2009)) The factors include (i) profitability (ii)industry median leverage (iii) market-to-book assets ratio (iv) tangibility ofassets and (v) firm size Industry median leverage is constructed as themedian leverage of all other firms in the industry excluding the firm under

8 Welch (2011) stresses the idea that nonfinancial liabilities should not be implicitly mis-treated as if they were equity by paying excessively narrow attention to financial liabilitiesin a leverage ratio We have adopted his approach in the empirical work reported here

1426 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

consideration Firm size is defined as the natural log of assets Other factorsare defined in Section 3 We cluster standard errors by firm and estimatethese regressions both with and without fixed effectsIn Table II we run conventional cross-section leverage regressions Since

we employ the usual data it is not surprising that our results match thosereported in the existing capital structure literature Profitability has anegative sign in both the book leverage regressions and in the marketleverage regressions The coefficients on other factors largely match thosereported in earlier studies Firms operating in industries with high leverage

Table II Leverage and profitability

The table presents parameter estimates for book leverage regressions The sample consists

of nonfinancial companies in the annual Compustat files during the period 1971ndash2009 Theleverage is estimated as the ratio of debt over debt plus book equity The explanatoryvariables Profitabilityt1

MB

t1

Tangibilityt1 and LnethAssetsTHORNt1 are described in

Table I IndLevt1 is estimated as the median book leverage of all other firms in thesame industry The industry is defined at the level of the four-digit SIC code All specifi-cations include the year fixed effects The specifications in column (2) additionally include

the firm fixed effects We report t-statistics where the standard errors are clustered at thefirm level in parentheses

aSignificant at the 1 level

Quantile regressions

(1) (2) (3) (4) (5)

Profitabilityt1 026a 020a 011a 017a 041a

(293) (192) (378) (518) (1015)

IndLevt1 042a 012a 030a 052a 052a

(369) (107) (926) (1252) (950)MB

t1

002a 001a 001a 002a 002a

(203) (102) (392) (426) (418)

Tangibilityt1 018a 025a 024a 019a 012a

(206) (154) (964) (580) (276)

LnethAssetsTHORNt1 001a 002a 002a 002a 001a

(114) (76) (899) (478) (96)

Constant 012a 010a 012a 002a 029a

(156) (60) (284) (27) (387)

Year FE Yes Yes Yes Yes Yes

Firm FE No Yes No No No

Regression type OLS FE 25th tile Median 75thtile

Clustered SE Yes Yes No No No

R2 013 054

Pseudo R2 016 015 009

N 158824 158824 158824 158824 158824

PROFITSndashLEVERAGE PUZZLEREVISITED 1427

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

tend to have high leverage A higher market-to-book ratio is associated withlower leverage Larger firms are typically more highly levered The coeffi-cients on tangibility are positive The conclusion from Table II is clear Theleverage regression results match those reported in previous studiesAt times one may be concerned that in a regression what is being

estimated is a conditional mean while some parts of the distributionmight behave differently from other parts If so then ldquoaccidentsrdquo likesample selection issues or asymmetries in the underlying distributionsmight play a greater role than is usually recognized In other words thesource of identification could be misinterpreted To guard against this po-tential problem we estimate the same model using quantile regressions9

Columns 3 4 and 5 report what happens at the 25th 50th and 75th per-centiles The basic model results are robust in terms of both the sign and thestatistical significanceThe impact of profits does appear to be somewhat stronger among the

high leverage firms For instance in a book leverage regression the coeffi-cient on profits for the first quartile is 011 with a t-ratio of 378 For thethird quartile the coefficient on profits is 041 with a t-ratio of 1015This difference is curious and might deserve further study However the factthat in each case we get the negative sign and statistical significance is suf-ficient for the purposes of the present article So the results on profits arerather robust across the distributionTable III sorts firms into those that are actively adjusting their leverage

and those that are passive In the presence of both fixed and variable trans-acts costs this distinction is important We expect the coefficient on profit-ability in leverage ratio regressions to be more negative in firms that arepassive and less negative in firms that are actively adjustingWe define active firms as those that are either issuing net debt or net

equity in excess of 5 of their assets Passive firms are those that neitherissue debt nor equity in nontrivial amounts (ie greater than 5 of theirassets) Consistent with our expectations we find in Table III that the coef-ficients on profits are significantly more negative for passive firms than theyare for active firms With firm and year fixed effects the coefficient onprofits is 008 for active firms and 017 for the passive firms The twocoefficients are significantly different from each other (plt 001) InTable III the standard control variables are used They have the usualsigns and for the most part do not differ much between the active and thepassive firms

9 Cameron and Trivedi (2010) provide an extensive discussion of quantile regressions inStata

1428 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

5 Financing Activity

51 FREQUENCY OF DEBT AND EQUITY ISSUES

Table IV reports the percentage of firms issuing or repurchasing debt orequity for annual sorts based on lagged profitability In Panel A weemploy the conventional 5 cutoff rule to exclude minor fluctuations anddefine firms ldquoissuing debtrdquo as those that issue debt (both short-term andlong-term) in excess of 5 of the value of their assets Other decisions aresimilarly defined using a 5 cutoff Appendix Table AII examines the dis-tributional properties across profitability (and size) quintiles for ldquolargerdquo is-suances by increasing the 5 cutoff to higher levels such as 10 and 15How does security issuance behavior vary with firm profitability To

answer this question Table IV sorts firms by profitability and reports the

Table III Leverage and profitability for active and passive firms

The table presents estimates of leverage on firm characteristics for the subsamples of firms

that are active in either debt or equity markets Active firms are defined as those that issueeither net debt in excess of 5 of the value of assets or issue net equity in excess of 5 ofthe value of their assets The passive firms do not issue significant (more than 5) of either

debt or equity The sample comes from the annual Compustat files during the period 1971ndash2009 Financial firms are excluded The variables are described in Tables I and II Wereport t-statistics where the standard errors are clustered at the firm level in parentheses

aSignificant at the 1 level

Active Passive

(1) (2) (3) (4)

Profitabilityt1 006a 008a 022a 017a

(126) (148) (254) (188)

IndLevt1 044a 019a 041a 013a

(424) (149) (326) (127)MB

t1

002a 001a 003a 001a

(362) (196) (327) (172)

Tangibilityt1 010a 014a 014a 019a

(146) (118) (145) (130)

LnethAssetsTHORNt1 001a 004a 001a 004a

(121) (179) (147) (169)

Constant 012a 008a 003a 007a

(120) (57) (45) (42)

Year FE Yes Yes Yes Yes

Firm FE No Yes No Yes

R2 034 064 033 075

N 43395 43395 58943 58943

PROFITSndashLEVERAGE PUZZLEREVISITED 1429

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

percentage of firms issuing or repurchasing debt or equity The table alsoreports test statistics from tests that examine whether the percentage differ-ences between the high- and low-profit firms are significantly different fromeach other The results in Column (1) show that the likelihood of issuing debt

Table IV Profitability sorts for debt and equity issuers

The sample contains nonfinancial firms listed on the annual Compustat files for the period

from 1971 to 2009 The table presents the percentage of firms issuing and retiring (orrepurchasing) debt and equity The firms are classified as ldquoissuing debtrdquo if they issuedebt in excess of 5 of the value of their assets as ldquoissuing equityrdquo if they issue equity

in excess of 5 of the value of their assets as ldquoretiring debtrdquo if they retire debt in excess of5 of the value of their assets and as ldquorepurchasing equityrdquo if they repurchase equity inexcess of 5 of the value of their assets In addition we also report net debt issuers which

are firms that issue net debt over 5 of the value of their assets and net equity issuerswhich are firms that issue net equity in excess of 5 of the value of their assets Weannually sort firms on lagged profitability and report the percentage of firms in each ofthese categories The bottom part of each panel reports the percentage of firms issuing or

retiring securities by profitability within the smallest and largest asset quintiles The tablealso reports results from t-test for the difference in percentage of firms issuing or retiringsecurities in the lowest and the highest profit quintile

Percentage of firms

Iss Ret Iss Iss Rep Iss Both IssD IssE

D D NetD E E NetE DampE RepE RetD

(1) (2) (3) (4) (5) (6) (7) (8) (9)

Allfirm

s

Low 276 245 186 301 13 297 71 01 38

2 319 299 181 117 21 112 27 02 20

3 385 344 194 86 29 80 21 05 15

4 386 330 213 81 43 75 20 07 14

High 329 273 198 92 82 81 17 13 14

t Highfrac14Low 150 84 39 723 414 763 358 176 200

Smallfirm

s Low 275 204 201 411 11 405 109 01 47

2 291 233 213 379 10 378 99 01 45

3 281 246 204 265 14 261 66 01 41

4 275 272 186 162 15 159 34 02 35

High 275 281 178 136 27 131 25 02 26

t Highfrac14Low 01 84 28 316 55 318 171 18 55

Largefirm

s

Low 414 363 177 87 22 79 28 03 11

2 449 379 180 62 30 57 17 07 09

3 439 360 186 56 41 48 16 09 07

4 403 320 194 46 67 39 13 12 06

High 374 264 207 50 129 35 10 28 05

t Highfrac14Low 50 130 45 89 244 116 77 120 47

1430 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

is mostly independent of firm profitability In the aggregate there is a smalldifference but it is sensitive to sorting on firm size Similarly in Column (2)there is only a weak relation between the profitability and the likelihood thatthe firm retires debt To a first approximation the fraction of firms issuing orrepurchasing debt is largely independent of profits Large firms are somewhatmore likely than small firms to issue debt and also to retire debtUnlike debt issues the probability of issuing equity is strongly related to

profitability as can be seen in Column (4) Low-profit firms are much morelikely to issue equity than are high-profit firms In the lowest profit quintile301 of the firms issued equity In the highest profit quintile it was only92 The lower panel reports separate sorts for small and large firms Similarpatterns are found for both groups of firms that is the likelihood of equityissuance declines as profitability increases The differences between the quin-tiles are quite stark for the group of small firms All quintiles of small firmsissue equity much more frequently than the corresponding large firm quintilesHigh-profit firms are much more likely to repurchase equity as shown in

Column (5) In the overall sample only 13 of low-profit firmsrepurchased equity compared with 82 of the high-profit firms Whenwe separate the small and large firms again a difference emerges Ingeneral small firms do not repurchase much equity while a greaterfraction of the large firms do soGiven these facts it is natural that low profitability firms are much more

likely to be net issuers of equity than are high profitability firms as shown inColumn (6) Among the firms in the lowest quintiles of profits roughly 30of firms issue net equity that exceeds 5 percent of their assets In contrastfor the most profitable quintile only about 8 issue net equityFirm size is an important variable in the recent literaturemdashit is often used

as a proxy for access to capital markets (as in Faulkender and Petersen(2006) and Leary (2009)) Small firms are bank-dependent risky and infor-mationally opaque They have restricted access to public debt markets andconsequently face more severe supply constraints in their ability to issuedebt Thus we expect small firms to be more sluggish in adjusting theirdebt and equity in response to shocks to profitability Large firms incontrast have much easier access to public debt markets and they facefewer obstacles in accessing securities marketsTo examine how size and profitability interact the bottom part of Table IV

examines profitability sorts for the smallest and largest firms We first sortfirms annually by firm size and then within each size quintile we sort them onprofitability We do these two-way sorts to ensure that we have similarnumber of firms in profitability subgroups for both small and large firmsamples Unconditionally small firms are generally less profitable while

PROFITSndashLEVERAGE PUZZLEREVISITED 1431

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

large firms are relatively more profitable The two-way sorts allow us to havea uniform distribution of firms in various profitability quintiles for both smalland large firm samples The bottom part of Table IV reports results for thesmallest and the largest size quintiles Among the small firms there is littlerelation between profitability and the likelihood of issuing debt Howeverlarger firms exhibit a small increase in the likelihood of net issuance of debtwith increasing profitabilityIn contrast to debt the effects of profitability on equity issuance and

repurchases are much more consistent across size quintiles As profitabilityincreases firms are generally less likely to issue equity and more likely torepurchase it Across the two extreme size groups we note a striking differ-ence between the proportion of firms issuing equity and those issuing debtAmong low-profit small firms almost 41 are equity issuers whereas onlyabout 9 of low-profit large firms are equity issuers However regardless ofsize we note a monotonic reduction in the likelihood of issuing equity asprofitability increasesColumn (7) reports the fraction of firms that issue both debt and equity in

excess of 5 of assets both as a function of profits in the upper panel andas a function of firm size and profits in the lower panel In the upper panelwe see that low-profit firms are much more likely to issue both debt andequity (71) while the high-profit firms are less likely to do so (17) Thelower panel shows that small firms are much more likely to be issuers of bothin contrast to large firms In both firm size categories the profit pattern is thesame with low profits being more frequent issuers of both simultaneouslyImportantly Column (8) shows that the likelihood of issuing debt and

simultaneously repurchasing equity increases with profitability Converselyin Column (9) we find that the likelihood of doing the reverse that is issuingequity and retiring debt declines with profitability The effects of sorting onfirm size and firm profitability mirror those reported for all firms Overallthe results show that low-profit firms are less likely to issue debt and repur-chase equity they are instead more likely to issue equity and retire debtFirms with high profitability exhibit the reverse patternIn Appendix Table AII we present results from profitability (and size)

quintiles for large issuances defined using the 10 cutoff (in Panel A) andthe 15 cutoff (in Panel B) The results from these panels are largely con-sistent with those described above

52 MAGNITUDES OF ISSUANCES AND EQUITY CHANGES

In the previous section we considered the probability of having a nontriviallevel of debt or equity activity The next question is how large are the dollar

1432 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

values involved In Table V we sort firms according to profits and thentabulate the levels and changes in both debt and equity We do this firstfor all firms and then for small versus large firmsFor the analysis on all firms we observe in Column (1) that debt peaks at

the middle of the distribution This is because firms with medium profitabil-ity are also the largest as seen in Column (8) which reports average assetvalues for different profitability quintiles When we sort first by firm size andthen examine firms with different levels of profitability within size quintileswe find that for small firms debt is roughly independent of profits while forlarge firms less profitable firms have a higher level of debtAs expected Columns (3) and (5) show that more profitable firms have

higher equity values Columns (2) and (7) consider the relationship betweenissuances of debt and equity and firm profitability Debt issuances are sig-nificantly larger for more profitable firms In contrast equity issuances arelarger for low-profit firms Small low-profit firms issue substantially lessequity compared with small high-profit firms Among large firms the low-profit ones tend to issue equity while the high-profit ones tend to repurchaseequity The fact that more profitable firms issue debt and repurchase equitywhile the least profitable firms retire debt and issue equity is consistent withthe predicted relation between profitability and financing decisions under thetrade-off theoryWe also examine two-way sorts by size and profitability and report results

for the smallest and largest quintile of firms For the smallest quintile offirms the change in debt is largely unrelated to profits But for large firmsthere is a positive relation between profits and debt issuances High-profitfirms have a big positive change in debt Low-profit large firms have anegative change in debtColumns (4) and (6) provide an explanation of why the leverage ratio

regression results contradict those from the basic profitability sorts pre-sented here As we can see profitability indirectly affects leverage ratiosby increasing equity values Changes in both the book value of equity andthe market value of equity are positive and large for highly profitable firmsIn contrast these changes are negative for less profitable firms This evidenceillustrates an important issue concerning the use of leverage ratios Suchratios are often interpreted as essentially reflecting the use of debt by thefirm This interpretation while common is empirically misleadingFor the typical firm the change in the value of equity is larger than the

change in debt For example in the third profit quintile for large firmscolumn (7) the mean equity issue is just $9 million but the change in themarket value of equity is $242 million At the same time the mean change indebt is $27 million This suggests that a fair bit of the observed variation in

PROFITSndashLEVERAGE PUZZLEREVISITED 1433

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

the leverage ratios is primarily driven by the changes in the market value ofequity in the denominator rather than by the changes in debt in the numer-ator Since equity issues are often small this implies that the variation in theleverage ratio is primarily driven by internal operations rather than byexternal financing actions This again points to the fact that leverageratios can provide a misleading account of actual patterns in the dataThere is always a potential concern that the averages may be misleading

due to the impact of outliers To address this concern in AppendixTable AIII we present median values of the profitability sorts Theseresults generally reinforce the findings in Table VTable VI again considers the magnitudes of financing activity but this

time the issues are scaled by a measure of firm size That way it is easier to

Table V Magnitude of financing activity

The sample contains nonfinancial firms listed on the annual Compustat files for the period

from 1971 to 2009 The table presents average debt and equity changes in debt and equityand equity issuances (in millions of dollars) for firms sorted on profitability within sizeclasses The table reports information for all firms sorted on profitability and for profit-

ability sorts within the smallest and largest firms The sorts are done annually The tablealso reports results from t-test for the difference in percentage of firms issuing or retiringsecurities in the lowest and the highest profit quintile

D D BVE BVE MVE MVE EqyIss Assets

(1) (2) (3) (4) (5) (6) (7) (8)

Allfirm

s

Low 83 5 78 10 120 5 7 234

2 605 13 509 24 721 8 17 1691

3 1227 9 1123 22 1834 3 9 3551

4 833 23 1166 7 2408 74 9 2994

High 570 53 1171 71 3817 68 36 2581

t Highfrac14Low 31 9 36 9 38 2 16 36

Smallfirm

s Low 25 04 42 01 332 28 33 97

2 29 06 55 02 278 16 21 116

3 31 07 63 01 212 05 12 127

4 30 06 70 05 179 10 08 136

High 24 06 84 14 234 17 07 142

t Highfrac14Low 04 15 153 85 73 11 201 107

Largefirm

s Low 3929 55 2990 152 4671 11 63 10846

2 3826 19 3178 95 5122 66 11 10721

3 3074 27 3498 41 6173 242 9 9697

4 2623 62 3915 45 8473 94 57 9855

High 2180 193 4642 261 14924 202 148 10171

t Highfrac14Low 94 64 108 87 246 12 169 16

1434 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

see whether large firms behave disproportionately when compared withsmall firms Column (1) shows that low-profit firms have more debtas reported in many previous studies In Columns (4) and (6) we seethat low-profit firms have a major decline in both the book value ofequity and the market value of equity The high-profit firms have thereverse experienceMore interesting are the active decisions of the firms In Column (2) we see

that in contrast to the usual impression the low-profit firms are reducingdebt while the high-profit firms are increasing it In Column (7) we see thatthe low-profit firms are also reducing leverage by issuing equity The high-profit firms issue little equity In the lower panel similar results are foundwhen we separate small and large firms Equity issuing is particularly strong

Table VI Magnitude of financing activitymdashscaled levels and changes

The sample contains nonfinancial firms listed on the annual Compustat files for the period

from 1971 to 2009 The table presents average debt and equity changes in debt and equityand equity issuances scaled by assets for firms sorted on profitability within size classes Thetable reports information for all firms sorted on profitability and for profitability sorts

within the smallest and largest firms The sorts are done annually The table also reportsresults from t-test for the difference in percentage of firms issuing or retiring securities inthe lowest and the highest profit quintile

D D BVE BVE MVE MVE EqyIss

(1) (2) (3) (4) (5) (6) (7)

Allfirm

s

Low 386 122 12 871 3427 1868 217

2 285 58 407 119 940 157 28

3 293 12 425 14 742 03 12

4 269 26 461 02 955 27 10

High 192 15 551 47 1675 74 07

t Highfrac14Low 35 49 50 47 127 42 157

Smallfirm

s Low 545 154 841 2145 9688 9128 696

2 684 49 712 1915 5014 2318 288

3 331 253 141 969 3063 954 129

4 267 181 357 175 1673 221 55

High 194 05 543 26 1758 147 41

t Highfrac14Low 49 17 51 36 88 27 99

Largefirm

s Low 350 30 306 58 511 55 09

2 346 14 335 13 523 02 04

3 353 165 332 71 711 29 01

4 280 01 414 04 909 09 05

High 227 15 477 19 1643 27 14

t Highfrac14Low 371 112 453 181 570 57 221

PROFITSndashLEVERAGE PUZZLEREVISITED 1435

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

among the small low-profit firms that are also experiencing sharp reductionsin the value of equity In column (2) we see that both for the large and for thesmall firms net debt increases are primarily found in the highest quintile offirm profitability

6 Debt and Equity Regressions

So far we have demonstrated that in our data the conventional leverageregressions have the usual signs We then demonstrated that in simple sortsof the data strikingly different active financing patterns emerge The nexttask is to reconcile these differencesTable VII presents simple regressions predicting changes in debt book

equity and market equity as well as equity issuances Consistent with thesorts we control for changes in firm size and we also include firm and yeardummies There may be a concern about the speed of a firmrsquos responsesAccordingly we consider the lagged change in profits as the main case butwe also include the contemporaneous change in profits Whether the contem-poraneous change in profits is included or not does not in fact matter Wereport panel-robust standard errors adjusted for clustering at the firm levelThe results in Table VII support the findings in the sorts Firms that have

an increase in profits in 1 year increase debt both in the same year and in thenext year Firms that have an increase in profits experience an increase inboth the book value of equity and the market value of equity Importantlyhowever firms that experience and increase in profits reduce their equityissues All these effects are statistically significant (except for the marketvalue of equity in the column 5 specification) and seem intuitively reason-able In Appendix Table AIV similar regressions to those presented aboveare reported but without firm-fixed effects We find that empirically verylittle changes whether we include firm fixed effects or leave them out Thusthe targeting behavior to the extent that it happens does not account forwhat we are reportingThe next step is to bring the sorts and the simple regressions together into

a common setting We therefore include factors in addition to profitability(i) median industry leverage (ii) the market-to-book assets ratio (iii) tangi-bility of assets and (iv) firm size (measured by the log of assets) Rajan andZingales (1995) show that these factors are related to leverage in G7countries A number of studies have used these factors to estimateleverage targets Frank and Goyal (2009) show that these factors arerobustly related to leverage in the USA In these regressions we use indica-tors for the quintile that the firm is in for each factor

1436 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

Table VIII reports these results In Columns (1) and (2) we examine debtissues and again show that profits positively affect debt issuances The effectsare large and statistically significant at the 1 level Columns (3) and (4)which examine book equity show a strong effect of profits This effect is onlyslightly reduced by the inclusion of other factors Columns (5) and (6) illus-trate that the impact of profits on the change in market equity is also robustto the inclusion of the conventional factors Columns (7) and (8) examineequity issuances Here we again find that profits have a significant negativeimpact on equity issuances As before the results are robust to the inclusionof the conventional factorsIn unreported tables we estimate debt and equity changes using quantile

regressions In terms of the signs and significance of the profitability variable

Table VII Debt and equity changes

The sample contains nonfinancial firms listed on the annual Compustat files for the period

from 1971 to 2009 The table presents fixed effect estimates from regressions of changes indebt (D) changes in the book value of equity (BVE) changes in the market value ofequity (MVE) and net equity issuances (EqyIss) The explanatory variables include

current and lagged change in operating income before depreciation (Profits andProfitst1) and lagged change in the book value of assets (Assetst1) In addition tofixed firm effects the regressions include year indicator variables The reported t-statistics

are corrected for clustering at the firm level

aMean significant at the 1 level

bMean significant at the 5 level

cMean significant at the 10 level

D BVE MVE EqyIss

(1) (2) (3) (4) (5) (6) (7) (8)

Profitst1 022a 025a 015b 035a 004 060b 005b 007a

(33) (31) (25) (35) (03) (25) (25) (53)

Profits 050b 078a 140a 003a

(24) (32) (48) (34)

Assetst1 000 007b 025a 001b

(00) (21) (35) (25)

Constant 167 283a 34 51 1145a 1182a 605a 595a

(14) (30) (04) (05) (32) (29) (69) (74)

Year FE Yes Yes Yes Yes Yes Yes Yes Yes

Firm FE Yes Yes Yes Yes Yes Yes Yes Yes

R2 001 004 001 010 001 005 002 003

N 162056 162056 162130 162130 157550 157550 162154 162154

PROFITSndashLEVERAGE PUZZLEREVISITED 1437

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

the results are qualitatively similar to the OLS results reported in Table VIIIProfitability positively affects debt issuances and negatively affects equityissuances across various quantiles Overall the results in Table VIII showthat there is nothing that is special about using sorts or regressions to explaindebt and equity issues In either case we find that more profitable firms tendto increase their debt experience an increase in the value of equity andrepurchase shares Thus the control factors are not responsible for theusual rejection of the trade-off theory It would appear that what mattersis the size of the issuing activity relative to the organic increases in equityvalue due to profits In other words what matters in the leverage regressionseems to be coming from the presence of E in D

DthornE and not from D

Table VIII Debt and equity issuances profitability and leverage factor quintiles

The sample contains nonfinancial firms listed on the annual Compustat files for the period

from 1971 to 2009 The table presents estimates from regressions of change in debt (D)change in the book value of equity (BVE) change in the market value of equity (MVE)and net equity issuances (EquityIss) The explanatory variables include lagged profitability

quintiles and lagged leverage factor quintiles In addition to fixed firm effects the regres-sions include year indicator variables The reported t-statistics are corrected for clustering atthe firm level

aMean significant at the 1 level

bMean significant at the 5 level

cMean significant at the 10 level

D BVE MVE EqyIss

(1) (2) (3) (4) (5) (6) (7) (8)

ProfitabilityQint 214a 167a 283a 175a 80 429a 53a 55a

(70) (54) (66) (59) (08) (48) (55) (62)

IndLevQint 104a 43 99 20

(29) (15) (10) (10)

Ln(Assets)Qint52 17 1026a 53a

(16) (05) (77) (35)MB

Qint158a 372a 1155a 09

(45) (56) (106) (06)

TangibilityQint 05 177 436c 64a

(01) (15) (18) (31)

Constant 167 96 656b 235 4613a 8610a 113 570a

(08) (04) (26) (04) (61) (58) (11) (31)

Year FE Yes Yes Yes Yes Yes Yes Yes Yes

Firm FE Yes Yes Yes Yes Yes Yes Yes Yes

R2 0002 0002 0003 0004 0013 0014 0013 0014

N 155587 155037 155587 155037 155587 155037 155587 155037

1438 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

7 Are Financial Market Conditions Important

According to the trade-off theory capital structure is determined by con-sidering the impact of costs and benefits of debt The time variation inthe costs and benefits of debt imply that issuance decisions would alsovary over time Furthermore market conditions affect the transactioncosts of adjusting Adverse selection is a greater problem in a cold marketthan in a hot one Accordingly it may be easier to adjust in good marketconditions Related ideas have been developed by Huang and Ritter(2009) Faulkender et al (2012) and Halling Yu and Zechner (2012)Faulkender Flannery Hankins and Smithrsquos paper examines cross-sectionalvariation in adjustment speeds including market valuations andhow they may affect adjustment speeds It is natural to think that ingood market conditions it will be less costly to issue both debt and equityand adjusting leverage to respond to profitability shocks would be mucheasierTo test the importance of market conditions we require a definition of

good times and bad times Our empirical strategy is to estimate good timesversus bad times at the four-digit industry level We define an industry ashaving ldquogood timesrdquo if the median firm in that industry has a market-to-book ratio that is higher than the 67th percentile of the time-series distribu-tion of the industry median market-to-book ratios Conversely an industryis defined as having ldquobad timesrdquo if the median firm in the industry has amarket-to-book ratio that is lower than the 33rd percentile of the time-seriesdistribution of the industry median market-to-book ratiosPanel A of Table IX tabulates issuance activity for profitability sorts in

both good times and bad times As might be expected active debt and equityissuances are larger during good times There is also somewhat more activeswaps between debt and equity during good timesPanel B reports the issuance activity for the smallest and the largest firms

during the good times and bad times Within each size quintile firms aresorted on profitability This is a three-way sort As before we find that thereis much more active use of external markets during good times Small low-profit firms are more likely to issue equity in good times than in bad timesDebt issuances are significantly higher in good times Large high-profitfirms are significantly more likely to issue debt and repurchase equity ingood times than in bad timesPanel C reports the magnitudes of the financial variables rather than the

frequencies During bad times less profitable large firms retire substantialamounts of debt and they show a tendency to issue equity Small firms donot seem to engage in similar debt reduction activities However like the

PROFITSndashLEVERAGE PUZZLEREVISITED 1439

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

Table IX Debt and equity issues in good and bad times

The table reports the frequency and magnitude of financing activity for sorts on profits for subsamples

of firms in good and bad times An industry is defined as having ldquogood timesrdquo if the median firm in

that industry has a market-to-book ratio that is higher than the 67th percentile of the time-series

distribution of industry median market-to-book ratios Conversely an industry is defined as having

bad times if the median firm in that industry has a market-to-book ratio that is lower than the 33rd

percentile of the time-series distribution of industry median market-to-book ratios Panels A and B

report the percentage of firms (i) issuing net debt in excess of 5 of the value of their assets (ii) issuing

net equity in excess of 5 of the value of their assets (iii) issuing debt and repurchasing equity both in

excess of 5 of the value of their assets and (iv) issuing equity and retiring debt both in excess of 5

of the value of their assets Panel C tabulates the mean debt and equity levels and the mean changes in

debt and equity issuances The sample contains nonfinancial firms listed on the annual Compustat files

for the period from 1971 to 2009

Iss Iss Iss D Iss ENet D Net E Rep E Ret D(1) (2) (3) (4)

Panel A sorts on profitability

Badtimes

Low 145 131 01 202 170 49 02 083 173 46 04 074 182 47 03 10

High 177 51 06 10t Highfrac14Low 41 136 33 01

Goodtimes

Low 206 381 01 462 197 148 02 233 212 103 04 184 234 92 10 18

High 216 92 16 14t Highfrac14Low 23 658 138 23

Panel B sorts on firm size and profitability

Badtimes

Smallfirm

s Low 151 318 00 312 182 276 02 253 151 140 02 294 183 73 00 19

High 177 80 00 11t Highfrac14Low 10 90 NA 20

Largefirm

s Low 172 54 01 042 148 39 04 053 150 43 04 044 156 33 03 06

High 147 28 05 03t Highfrac14Low 17 32 20 02

Goodtimes Smallfirm

s Low 213 448 01 542 230 443 01 473 222 325 00 474 203 197 02 38

High 190 153 02 31t Highfrac14Low 19 230 10 39

Largefirm

s Low 203 96 03 122 208 66 09 133 232 54 16 064 215 39 19 05

High 234 36 38 04t Highfrac14Low 29 99 92 38

(continued)

1440 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

larger firms there is a tendency for small firms to issue more equitymdashat leastto the extent possibleDuring good times large profitable firms raise significant amounts of debt

as they experience an increase in the value of their equity Such firms alsoengage in share repurchases Debt issues by smaller firms are much lessaffected by their own profitability during good times and their change indebt is more or less independent of their profits However equity issuanceis primarily found among the low-profit small firms during good timesTable IX shows that empirically profitability has a time-varying impacton leverage choicesIn Table X the basic regressions for changes in debt and for equity

issuance are presented for both good times and bad times Comparinggood times with bad times we see that the effects of profits are muchstronger in good times In bad times the impacts are rather weak statisticallyThis difference is interesting and deserves further study However for ourpurposes recall that the static trade-off theory implies time-varying capital

Table IX Continued

D D BVE BVE MVE MVE EqyIss Assets

(1) (2) (3) (4) (5) (6) (7) (8)

Panel C magnitude of financing activity

Badtimes Smallfirm

s Low 40 03 58 14 305 06 22 140

2 44 05 82 07 240 67 17 180

3 45 04 94 02 176 18 05 187

4 43 09 98 02 137 04 03 188

High 36 10 114 13 209 16 04 199

Largefirm

s Low 3365 73 2523 169 2304 163 94 8972

2 3031 95 2983 35 3284 278 56 9068

3 2930 5 3360 14 3947 308 57 9225

4 2281 6 3817 19 5213 442 7 8902

High 1790 62 4767 13 8799 58 31 9573

Goodtimes

Smallfirm

s Low 24 06 38 05 339 23 39 90

2 26 08 48 01 301 04 25 104

3 30 09 60 04 246 03 17 125

4 28 07 66 08 206 18 12 130

High 23 06 85 16 261 13 08 141

Largefirm

s Low 4532 36 3477 103 6557 260 53 12465

2 4580 58 3527 139 6911 421 23 12522

3 2952 111 3485 2 7586 279 32 9535

4 2752 124 3939 72 10686 33 105 10226

High 2429 228 4819 377 18565 268 222 10830

PROFITSndashLEVERAGE PUZZLEREVISITED 1441

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

structure choices even when the target is time invariant Empirically we doobserve time-varying choices

8 Is Scaling Important

In the preceding section we provided results for unscaled debt and equityissues This seems appropriate to us However it involves making twochanges to the standard regression We would like to determine if both

Table X Debt and equity changes in good and bad timesmdashregressions

The sample contains nonfinancial firms listed on the annual Compustat files for the period

from 1971 to 2009 The table presents estimates from regressions of change in debt (D)and net equity issuances (EqyIss) in good and bad times The explanatory variables includechange in profits lagged change in profits and lagged change in assets The regressions

include both firm and year fixed effects An industry is defined as having ldquogood timesrdquo ifthe median firm in that industry has a market-to-book ratio that is higher than the 67thpercentile of the time-series distribution of industry median market-to-book ratios

Conversely an industry is defined as having bad times if the median firm in thatindustry has a market-to-book ratio that is lower than the 33rd percentile of the time-series distribution of industry median market-to-book ratios The reported t-statistics arecorrected for clustering at the firm level

aMean significant at the 1 level

bMean significant at the 5 level

cMean significant at the 10 level

Good times Bad times

D EqyIss D EqyIss

(1) (2) (3) (4) (5) (6) (7) (8)

Profits 087a 091a 006b 009a 060 065 003 003b

(30) (29) (20) (28) (15) (16) (16) (20)

Profitst1 029c 015a 034c 001

(19) (37) (17) (04)

Assetst1 000 002b 000 001

(02) (22) (00) (07)

Constant 1337 3989 8691a 10001a 1351 1851 4099a 4055a

(03) (09) (55) (37) (09) (13) (80) (78)

Year FE Yes Yes Yes Yes Yes Yes Yes Yes

Firm FE Yes Yes Yes Yes Yes Yes Yes Yes

R2 0071 0080 0018 0045 0178 0212 0034 0041

N 74827 68729 74891 68787 26965 26013 26971 26018

1442 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

changes are crucial Accordingly in this section we take a step back toconsider prior literature in which variables are scaled The questions arehow does scaling affect the results and whyTable XI examines alternative scalings Columns (1) and (2) scale debt and

equity issuances by the total firm net issuance Columns (3) and (4) scaledebt and equity issuances by the sum of debt and book equity Columns (5)and (6) consider the change in debt scaled by the sum of debt and bookequity Columns (7) and (8) consider the change in leverage due to externalfinancing Since we are examining changes we use the changes in thestandard regression factors as controls The primary interest is in the coef-ficient on the change in profitsWe find that a change in profits is associated with an increase in debt

issues But the sign reverses when we scale the debt issues by the sum of debtplus book equity This is telling What is really driving the results is thechange in equity in the denominator not the change in debt in the numer-ator This result is further substantiated in Columns (5) and (6) where thechange in debt to capital ratio is used instead of debt issuanceTo further explore the impact of active issuance relative to passive

experiencing of changes we decompose leverage by separating out theeffect of debt and equity issues from those due to changes in retainedearnings Thus we define leverage change due to external financing as

LevEF frac14Dt1 thorn d

Dt1 thorn Et1 thorn dthorn e

Dt1

Dt1 thorn Et1 eth1THORN

where D is debt E is book value of equity d is debt issues net of retirementsand e is equity issues net of repurchases Garvey and Hanka (1999) use asimilar measure in their study This measure directly captures the netleverage effect of debt and equity issuancesColumns (7) and (8) report results from regressions of the net leverage

effect of debt and equity issues on changes in profitability after controllingfor other leverage factors in differences and year indicator variables Theeffect of profitability on leverage change due to external financing is positiveand statistically significant This comparison suggests that changes in totaldebt and equity are only partly a result of debt and equity issuance decisionsOther balance sheet adjustments complicate the inferences from leverageratio regressionsFor most of our analysis we follow the literature in treating debt as homo-

genous It is of course potentially interesting to ask what happens when wedistinguish between short-term debt and long-term debt Which part of thefirmrsquos debt maturity structure drives the under-adjustment Table XII

PROFITSndashLEVERAGE PUZZLEREVISITED 1443

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

provides results from such decompositions Alternative scalings areconsidered The main results of interest are that the distinction betweenshort- and long-term debt is important Short-term debt issuance is largelyindependent of changes in profits The negative sign on profit change stemsfrom the action of long-term debt This seems intuitively reasonable

Table XI Scaled issuances

The sample contains nonfinancial firms listed on the annual Compustat files for the period

from 1971 to 2009 The table presents estimates from regressions of debt issuance to totalissuance (in columns (1) and (2)) debt issuance to total capitalization (in columns (3) and(4)) change in debt to capitalization ratio (in columns (5) and (6)) and change in leverage

ratio considering only external financing (in columns (7) and (8)) The change in leverageratio considering only external financing is defined as

LevEF frac14Dt1 thorn d

Dt1 thorn Et1 thorn dthorn e

Dt1

Dt1 thorn Et1

where D is debt E is book value of equity d is debt issues net of retirements and e is equity

issues net of repurchases The regressions include lagged leverage factors and year indicatorvariables Some of the specifications include firm fixed effects The reported t-statistics arecorrected for clustering at the firm level

aMean significant at the 1 level

bMean significant at the 5 level

cMean significant at the 10 level

DebtIssTotIss

DebtIssethDthornETHORN D

DthornE LevEF

(1) (2) (3) (4) (5) (6) (7) (8)

Profitt1 004a 000 002a 001c 003a 002 001b 002a

(35) (03) (34) (18) (35) (14) (23) (26)

IndLevt1 004c 001 001 001c 000 000 000 000

(19) (06) (15) (19) (02) (04) (03) (05)

Market=bookt1 001a 001a 000a 000a 000a 000a 000 000c

(47) (85) (68) (60) (55) (36) (05) (18)

Tangt1 004 005 013a 011a 011a 006a 003a 003a

(12) (15) (144) (112) (79) (38) (44) (42)

Assetst1 008a 004a 006a 004a 002a 001a 003a 002a

(97) (45) (283) (161) (53) (35) (147) (101)

Constant 079a 074a 002a 003a 000 000 000b 002a

(497) (442) (91) (111) (00) (05) (22) (93)

Year FE Yes Yes Yes Yes Yes Yes Yes Yes

Firm FE No Yes No Yes No Yes No Yes

R2 0015 0098 0022 0076 0009 0070 0012 0086

N 133905 133905 140552 140552 138229 138229 140572 140572

1444 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

inasmuch as short-term debt is probably more associated with day-to-dayoperations rather than longer term planning considerationsIn Table XIII we examine change in leverage ( D

DthornE) and change inleverage due to external financing (LEVEF) during good and bad timesseparately The traditional negative effect of a change in profits on achange in leverage is found both in good times and in bad timesHowever when we decompose the impact in order to isolate the externalcomponent as before the result disappears In good times the externalfinancing component of a change in profits is weak but still positiveDuring bad times it is not statistically significant This implies that distin-guishing between good and bad times does not fundamentally alter the basic

Table XII Short- and long-term debt issuances

The sample contains nonfinancial firms listed on the annual Compustat files for the period

from 1971 to 2009 The table presents estimates from regressions of short-term debtissuance to total issuance (in columns (1) and (2)) long-term debt issuance to totalissuance (in columns (3) and (4)) short-term debt issuance to total capitalization (in

columns (5) and (6)) and long-term debt to total capitalization (in columns (7) and (8))Some of the specifications include firm fixed effects The reported t-statistics are correctedfor clustering at the firm level

aMean significant at the 1 level

bMean significant at the 5 level

cMean significant at the 10 level

STDebtIssTotIss

LTDebtIssTotIss

STDebtIssethDthornETHORN

LTDebtIssethDthornETHORN

(1) (2) (3) (4) (5) (6) (7) (8)

Profitabilityt1 0002 0003 0042a 0008 0001 0001 0019a 0012a

(02) (04) (30) (05) (14) (10) (45) (28)

IndLevt1 0009 0007 0070b 0033 0001 0002c 0014a 0005

(06) (05) (26) (11) (08) (17) (26) (09)

M=Bt1 0003a 0003a 0003b 0010a 0000 0000 0003a 0002a

(48) (35) (22) (65) (10) (01) (84) (42)

Tangibilityt1 0020 0007 0011 0067 0005a 0006a 0097a 0081a

(12) (03) (03) (16) (34) (33) (108) (89)

Assetst1 0039a 0011b 0034a 0024b 0001a 0001 0046a 0020a

(89) (22) (38) (23) (36) (13) (190) (82)

Constant 0001b 0153a 0798a 0591a 0000c 0006a 0057a 0096a

(25) (144) (470) (288) (17) (124) (292) (234)

Year FE Yes Yes Yes Yes Yes Yes Yes Yes

Firm FE No Yes No Yes No Yes No Yes

R2 0025 0089 0018 0073 0030 0147 0016 0298

N 136146 136146 136140 136140 143031 143031 142847 142847

PROFITSndashLEVERAGE PUZZLEREVISITED 1445

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

Table XIII Externally financed leverage changes in good and bad times

The sample contains nonfinancial firms listed on the annual Compustat files for the period

from 1971 to 2009 The table presents estimates from regressions of change in debt tocapitalization ratio (in columns (1) (2) (5) and (6)) and change in leverage ratio consider-ing only external financing (in columns (3) (4) (7) and (8)) The change in leverage ratio

considering only external financing is defined as

LevEF frac14Dt1 thorn d

Dt1 thorn Et1 thorn dthorn e

Dt1

Dt1 thorn Et1

where D is debt E is book value of equity d is debt issues net of retirements and e is equityissues net of repurchases The regressions include lagged leverage factors and year indicator

variables The reported t-statistics are corrected for clustering at the firm level An industryis defined as having ldquogood timesrdquo if the median firm in that industry has a market-to-bookratio that is higher than the 67th percentile of the time-series distribution of industry

median market-to-book ratios Conversely an industry is defined as having ldquobad timesrdquoif the median firm in that industry has a market-to-book ratio that is lower than the 33rdpercentile of the time-series distribution of industry median market-to-book ratios

aMean significant at the 1 level

bMean significant at the 5 level

cMean significant at the 10 level

Good times Bad times

DDthornE LevEF D

DthornE LevEF

(1) (2) (3) (4) (5) (6) (7) (8)

Profitabiltyt1 0022c 0005 0016c 0025b 0073b 0041 0007 0007

(17) (03) (18) (24) (25) (11) (04) (03)

IndLevt1 0000 0011 0001 0009 0010 0018 0001 0000

(00) (09) (01) (13) (07) (10) (02) (00)

MB

t1

0004a 0003b 0000 0001 0008a 0001 0000 0000

(40) (24) (05) (08) (28) (04) (03) (03)

Tangt1 0097a 0041 0033a 0038b 0127a 0048 0059a 0033

(45) (16) (27) (25) (49) (14) (43) (16)

Assetst1 0014a 0013b 0024a 0019a 0037a 0023c 0036a 0025a

(30) (20) (84) (48) (45) (17) (88) (43)

Constant 0030b 0036 0000 0003 0028c 0014 0021 0001

(20) (16) (00) (07) (17) (06) (11) (01)

Year FE Yes Yes Yes Yes Yes Yes Yes Yes

Firm FE No Yes No Yes No Yes No Yes

R2 0006 0088 0009 0117 0026 0251 0023 0202

N 56820 56820 58200 58200 23462 23462 23511 23511

1446 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

conclusions The traditional leverage result on profits is driven by increasesin equity that are experienced and then partially but not completely offsetby firmsTo the extent that firms are engaging in active financing they are following

the prescriptions of the trade-off theory Thus we have a coherent picture ofhow firms are responding to profits Basically more profitable firms experiencean increase in the value of their equity They respond to this by issuing debtand repurchasing equity But the magnitude of the increase in equity is largerthan the magnitudes of the active steps taken in the debt and equity marketsThis means that the firmsrsquo actual actions are as predicted by the trade-offtheory But it leaves open the question of why they do not act more stronglyThe capital structure literature has long been interested in dynamics and

capital structure adjustment paths In recent years this interest has increas-ingly resulted in explicit dynamic optimizing models such as those ofGoldstein Ju and Leland (2001) Hennessy and Whited (2005) andStrebulaev (2007) These models commonly make specialized assumptionsleading some to question their generalityWe argue that both fixed transactions costs and variable transactions costs

are plausible and realistic in models that pin down firm size The simultaneouspresence of both is important Due to fixed costs firms only rebalance inter-mittently Due to variable adjustment costs the extent of the adjustment isreduced to the point at which an extra unit to adjustment is balanced againstthe extra unit transactions costs This means that both leverage increases andleverage reductions will only go part way toward a static optimumThe key point is that with both fixed and variable transactions costs we

should expect to only see intermittent leverage adjustments When leverage isadjusted it should ldquounder adjustrdquo when compared to a frictionless modelFurther discussion of transactions costs in a capital structure setting isprovided by Leary and Roberts (2005)

9 Conclusion

The connection between corporate profits and capital structure has beenvery influential in the assessment of the trade-off theory The conventionalinterpretation of the evidence has pushed the literature away from the trade-off theory and toward much more complex dynamic models and ideas As aresult it is important to make sure that the evidence has been correctlyinterpreted We show that the problem is with the interpretation of thetest results and a failure to take into account the costs associated with re-sponding to a profit shock

PROFITSndashLEVERAGE PUZZLEREVISITED 1447

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

We document that more profitable firms really do borrow more (not less)They tend to repurchase equity They experience an increase in both thebook value of equity and the market value of equity Less profitable firmsreally do tend to reduce their debt and to issue equityFirm size and market conditions also matter Larger firms tend to be more

active in the debt markets while smaller firms tend to be relatively moreactive in the equity markets During good times there is more use ofexternal financingThe usual profitsndashleverage puzzle result is primarily driven by the increase

in equity that is experienced by the more profitable firms Accordingly thepuzzle should be restated as asking why do firms not take sufficiently largeoffsetting actions to fully undo the change in equity What limits the mag-nitudes of the typical leverage response to profit shocksIn a frictionless model the partial response appears puzzling But there is

good empirical reason to believe that rebalancing entails both fixed andvariable costs and that firm size matters The rebalancing costs can befully avoided by doing nothing Accordingly the firm must decide whetherany given shock is big enough to be worth responding to If it is then thefirm must decide how big a response is called for These technical conditionsare known as ldquovalue matchingrdquo and ldquosmooth pastingrdquo Thus optimizationimplies that some shocks will be ignored Even if the shock is not ignoredthe optimal response will only partially undo the shock The magnitude ofthe leverage response must balance the marginal cost and the marginalbenefit of an extra unit of leverage Since the marginal cost of adjustingleverage is strictly positive the adjustment toward that static leverageoptimum will only go part way This is true both for leverage increasesand for leverage reductionsOne might conjecture that the variable costs would be time varying and

that such costs might have declined over time Since we do not have a fullaccounting of all of the variable costs it is of course impossible to be sureHowever there is some evidence suggesting that at least the directunderwriting fees are remarkably stable over time10

Surely nobody ought to be surprised that more profitable firms are moreinclined to repurchase equity while unprofitable firms tend not to do so Thefact that more profitable firms typically experience an increase in the value of

10 For IPOs Jay Ritter provides important summary evidence as httpbearwarringtonufleduritterIPOs2013Underwritingpdf He finds that for medium-sized deals a remark-

ably large fraction of the deals have a 7 fee year after year Whether this is due tobehavioral factors or some sort of a coordination device is not clear The extent towhich similar intertemporal stability carries over to other forms of external financemerits further study

1448 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

equity is equally natural Were it not for the contrary claims in the literature(ie the profitsndashleverage puzzle) we would argue that nobody ought to findit surprising that more profitable firms are more inclined to issue debt Thefact that leverage adjustments typically only partially offset profit shocksshould also sound quite natural These facts all fit together easily within thetrade-off theory once we take account of both fixed and variable costs ofactively adjusting capital structure

Appendix

Table AI Market leverage regressions

The table presents estimates of the regressions of market leverage on firm characteristics

The sample comes from the annual Compustat files during the period 1971ndash2009 Financialfirms are excluded The leverage is estimated as the ratio of debt over debt plus marketequity The explanatory variables Profitabilityt1

MB

t1

Tangibilityt1 and LnethAssetsTHORNt1are described in Table 1 IndMedianLevt1 is estimated as the median market leverage of allother firms in the same industry The industry is defined at the level of the four-digit SICcode All specifications include the year fixed effects The specifications in column (2)

additionally include the firm fixed effects We report t-statistics where the standard errorsare clustered at the firm level in parentheses

aSignificant at the 1 level

Quantile regressions

(1) (2) (3) (4) (5)

Profitabilityt1 013a 013a 007a 011a 020a

(275) (262) (349) (425) (538)

IndLevt1 046a 020a 032a 058a 063a

(506) (239) (1299) (1630) (1042)MB

t1

003a 001a 001a 002a 004a

(509) (303) (743) (709) (534)

Tangibilityt1 013a 018a 013a 014a 010a

(191) (181) (762) (571) (232)

LnethAssetsTHORNt1 001a 005a 001a 001a lt001a

(112) (258) (682) (342) (29)

Constant 010a 011a 005a 005a 029a

(171) (98) (173) (118) (398)

Year FE Yes Yes Yes Yes Yes

Firm FE No Yes No No No

Regression type OLS FE 25th tile Median 75thtile

Clustered SE Yes Yes No No No

R2 030 067

Pseudo R2 016 023 0201

N 158578 158578 158578 158578 158578

PROFITSndashLEVERAGE PUZZLEREVISITED 1449

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

Table AII Profitability sorts for firms issuing large amounts

Percentage of firms

Iss Ret Iss Iss Rep Iss Iss IssD IssE

D D NetD E E NetE DampE RepE RetD

(1) (2) (3) (4) (5) (6) (7) (8) (9)

Panel A 10 cutoff to define issuing or retiring activity

Allfirm

s

Low 205 150 134 260 05 257 45 00 23

2 207 175 103 77 06 75 11 01 10

3 243 196 101 50 08 47 08 01 09

4 252 186 113 50 13 48 08 02 08

High 227 161 118 60 27 57 07 04 09

t Highfrac14Low 70 38 62 755 226 768 320 90 152

Smallfirm

s Low 230 132 162 380 05 375 82 00 32

2 228 143 165 334 04 333 65 00 26

3 211 146 147 216 03 214 38 00 23

4 190 154 120 126 04 124 15 01 20

High 194 162 114 99 09 97 12 01 15

t Highfrac14Low 42 40 68 342 20 340 169 07 131

Largefirm

s Low 231 200 82 31 06 29 08 01 04

2 231 187 67 17 06 14 03 01 02

3 234 177 77 17 09 14 03 02 03

4 218 158 75 16 17 14 02 03 02

High 218 135 99 19 41 15 03 07 02

t Highfrac14Low 19 106 36 46 135 57 39 58 23

Percentage of firms

Iss Ret Iss Iss Rep Iss Both Iss D Iss E

D D Net D E E Net E DampE Rep E Ret D

(1) (2) (3) (4) (5) (6) (7) (8) (9)

Panel B 15 cutoff to define issuing or retiring activity

Allfirm

s

Low 158 101 100 230 01 228 27 00 15

2 142 115 63 57 01 56 05 00 06

3 165 128 57 32 01 31 03 00 05

4 173 118 65 35 01 34 03 00 05

High 160 102 70 46 04 44 03 00 05

t Highfrac14Low 08 06 141 738 87 746 277 24 131

Smallfirm

s Low 195 95 132 354 01 350 56 00 24

2 185 96 131 303 01 301 44 00 18

3 160 96 104 179 00 176 18 00 15

4 134 99 78 99 01 98 08 00 12

High 133 95 73 82 01 80 06 00 10

t Highfrac14Low 80 00 95 345 12 344 149 12 52

Largefirm

s Low 145 123 42 14 01 12 03 00 01

2 137 114 32 07 01 06 01 00 01

3 142 104 35 06 01 05 00 00 01

4 130 91 36 08 03 07 01 01 01

High 134 81 48 09 06 07 01 01 01

t Highfrac14Low 19 85 17 26 54 33 34 17 11

1450 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

Table AIV Debt and equity changesmdashwithout fixed effects

The sample contains nonfinancial firms listed on the annual Compustat files for the period from 1971

to 2009 The table presents estimates from regressions of change in debt (D) change in the book

value of equity (BVE) change in the market value of equity (MVE) and net equity issuances

(EqyIss) The explanatory variables include current and lagged change in operating income before

depreciation (Profits and Profitst1) and lagged change in book value of assets (Assetst1) The

regressions include year indicator variables The reported t-statistics are corrected for clustering at the

firm level

aMean significant at the 1 level

bMean significant at the 5 level

cMean significant at the 10 level

D BVE MVE EqyIss

(1) (2) (3) (4) (5) (6) (7) (8)

Profitst1 025a 024a 023a 038a 007 066a 007a 009a

(40) (32) (36) (39) (06) (29) (42) (51)Profits 051b 082a 149a 004a

(25) (33) (49) (33)Assetst1 002 005c 024a 001b

(08) (17) (34) (23)Constant 93 176a 45 71 1552a 1704a 130a 140a

(12) (28) (07) (09) (49) (49) (45) (45)Year FE Yes Yes Yes Yes Yes Yes Yes YesR2 001 005 001 011 001 005 002 002N 162056 162056 162130 162130 157550 157550 162154 162154

Table AIII Median profitability sorts

The sample contains nonfinancial firms listed on the annual Compustat files for the period from 1971

to 2009 The table presents median debt and equity median changes in debt and equity and median

equity issuances (EqyIss) in millions of dollars for all firms sorted on profitability and for profitability

sorts within the smallest and largest firms The sorts are done annually

D D BVE BVE MVE MVE EqyIss Assets(1) (2) (3) (4) (5) (6) (7) (8)

Allfirm

s

Low 21 00 75 17 256 20 01 1922 186 02 449 18 598 19 00 11033 816 06 1274 03 1637 05 00 32244 787 02 1541 18 2488 03 00 3556High 273 00 1361 66 3300 14 00 2578

Smallfirm

s Low 05 00 13 07 140 14 06 402 06 00 25 08 114 12 01 563 09 00 32 05 85 07 00 674 12 00 39 01 68 03 00 79High 10 00 51 04 98 01 00 90

Largefirm

s Low 11485 381 10269 371 13227 202 03 333902 13051 393 12786 140 17571 169 05 392463 11554 287 13429 24 21623 305 00 370314 9963 209 15052 194 29424 308 00 36705High 7361 00 16380 613 49142 439 00 34976

PROFITSndashLEVERAGE PUZZLEREVISITED 1451

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

References

Auerbach A J (1985) Real determinants of corporate leverage in B M Friedman (ed)

Corporate Capital Structures in the United States University of Chicago Press Chicago

IllinoisBertrand M and Mullainathan S (2003) Enjoying the quiet life corporate governance and

managerial preferences Journal of Political Economy 111 1043ndash1075Booth L Varouj A Demirguc-Kunt A and Maksimovic V (2001) Capital structures in

developing countries Journal of Finance 56 87ndash130Brennan M J and Schwartz E S (1984) Optimal financial policy and firm valuation

Journal of Finance 39 593ndash607Cameron A C and Trivedi P K (2010) Microeconometrics using Stata Stata Press TexasConstantinides G M (1979) Multiperiod consumption and investment behavior with

convex transactions costs Management Science 25 1127ndash1137Constantinides G M and Richard S F (1978) Existence of optimal simple policies for

discounted-cost inventory and cash management in continuous time Operations Research

26 620ndash636DeAngelo H DeAngelo L and Whited T M (2011) Capital structure dynamics and

transitory debt Journal of Financial Economics 99 235ndash261Dixit A K (1993) The Art of Smooth Pasting Routledge London and New YorkFama E and French K R (2002) Testing trade-off and pecking order predictions about

dividends and debt Review of Financial Studies 15 1ndash33Faulkender M and Petersen M A (2006) Does the source of capital affect capital struc-

ture Review of Financial Studies 19 45ndash79Faulkender M Flannery M J Hankins K W and Smith J M (2012) Cash Flows and

Leverage Adjustments Journal of Financial Economics 103 632ndash646Fischer E Heinkel R and Zechner J (1989) Dynamic capital structure choice theory

and tests Journal of Finance 44 19ndash40Frank M Z and Goyal V K (2008) Trade-off and pecking order theories of

debt in B E Eckbo (ed) Handbook of Corporate Finance Empirical Corporate Finance

Vol 2 Handbook of Finance Series Chapter 12 ElsevierNorth-Holland Amsterdam pp

135ndash202

Frank M Z and Goyal V K (2009) Capital structure decisions which factors are reliably

important Financial Management 38 1ndash37

Garvey G T and Hanka G (1999) Capital structure and corporate control the effect of

antitakeover statutes on firm leverage Journal of Finance 54 519ndash546

Goldstein R Ju N and Leland H (2001) An EBIT-based model of dynamic capital

structure Journal of Business 74 483ndash512

Graham J R and Tucker A L (2006) Tax shelters and corporate debt policy Journal of

Financial Economics 81 563ndash594

Hackbarth D and Mauer D C (2012) Optimal priority structure capital structure and

investment Review of Financial Studies 25 747ndash796

Halling M Yu J and Zechner J (2012) Leverage dynamics over the business cycle

Working paper University of Utah University of New South Wales and Vienna

University of Economics and Business AdministrationHennessy C A and Whited T A (2005) Debt dynamics Journal of Finance 60 1129ndash1165

Huang R and Ritter J R (2009) Testing theories of capital structure and estimating the

speed of adjustment Journal of Financial and Qualitative Analysis 44 237ndash271

1452 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

Hovakimian A Opler T and Titman S (2001) The debt-equity choice Journal of

Financial and Quantitative Analysis 36 1ndash24Jung K Kim Y and Stulz R M (1996) Timing investment opportunities managerial

discretion and the security issue decisions Journal of Financial Economics 42 159ndash185Kane A Marcus A J and McDonald R L (1984) How big is the tax advantage to

debt Journal of Finance 39 841ndash853Kayhan A and Titman S (2007) Firmsrsquo histories and their capital structures Journal of

Financial Economics 83 1ndash32Leary M T (2009) Bank loan supply lender choice and corporate capital structure

Journal of Finance 64 1143ndash1185Leary M T and Roberts M R (2005) Do firms rebalance their capital structures

Journal of Finance 60 2575ndash2619Leland H E (1994) Corporate debt value bond covenants and optimal capital structure

Journal of finance 49 1213ndash1252Lemmon M L Roberts M R and Zender J F (2008) Back to the beginning persist-

ence and the cross-section of corporate capital structure Journal of Finance 63 1575ndash1608Long M S and Malitz I B (1985) The investment-financing nexus some empirical

evidence Midland Corporate Finance Journal 3 53ndash59Mackie-Mason J K (1990) Do taxes affect corporate financing decisions Journal of

Finance 45 1471ndash1493Morellec E Nikolov B and Schurhoff N (2012) Corporate governance and capital

structure dynamics Journal of Finance 67 803ndash848Myers S C (1984) The capital structure puzzle Journal of Finance 39 575ndash592Myers S C (1993) Still searching for optimal capital structure Journal of Applied

Corporate Finance 61 4ndash14Rajan R and Zingales L (1995) What do we know about capital structure some evidence

from international data Journal of Finance 50 1421ndash1460Strebulaev I A (2007) Do tests of capital structure theory mean what they say Journal of

Finance 62 1747ndash1787

Strebulaev I A and Whited T M (2012) Dynamic Models and Structural Estimation in

Corporate Finance Now Publisher Boston and Delft

Titman S and Wessels R (1988) The determinants of capital structure choice Journal of

Finance 43 1ndash21

Welch I (2004) Capital structure and stock returns Journal of Political Economy 112

106ndash131

Welch I (2011) Two common problems in capital structure research the financial-debt-

to-asset ratio and issuing activity versus leverage changes International Review of Finance

11 1ndash17

PROFITSndashLEVERAGE PUZZLEREVISITED 1453

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

In the absence of active adjustments the leverage ratio xt follows astandard process

dx frac14 ethx tTHORNdtthorn ethx tTHORNdw

where is the trend coefficient is the volatility coefficient and w is anormal shock process with zero mean and unit variance If x lt x lt x thenfinancial inaction is optimal If x drops to x then it is reset to x and if xrises to x then it is reset to xThe trade-off theory is frequently thought of as a class of theories in which

firm value first increases in leverage and ultimately drops if leverage gets toohigh This can be motivated by taxes and bankruptcy or by various agencyconcerns A range of ideas can be allowed for when we write that the flowpayoff (dividends andor profits) is fethxtTHORN and it depends on the leverage statext

5 If the firm starts at x0frac14 x then F(x) is the expected present value of theflow payoffs The discount rate is which allows us to write

FethxTHORN frac14 E

Z 10

fethxtTHORNetdtjx0 frac14 x

To maximize the firm value it must choose a leverage policy that sayswhen to rebalance and how much to rebalance These are given by thefamous value matching and smooth pasting conditionsThe value matching conditions are as follows

FethxTHORN FethxTHORN frac14 Ku thornmuethx xTHORN

FethxTHORN FethxTHORN frac14 Kd thornmdethx xTHORN

The smooth pasting conditions are as follows

F 0ethxTHORN frac14 F 0ethxTHORN frac14 mu

F 0ethxTHORN frac14 F 0ethxTHORN frac14 md

As explained by Constantinides and Richard (1978) with standard as-sumptions about the shape of F there are a pair of lines at heights mu andmd such that they intersect the function F 0ethxTHORN in such a way as to define thepoints x lt x lt x lt xTo sum up in this trade-off model with both fixed and variable

rebalancing costs there are four critical leverage points x gt x gt x gt x

5 For simplicity we suppose that the flow payoff is a continuous function of leverageThere is no discrete exit or bankruptcy point This allows us to simply interpret theConstantinides and Richard model much as in Leary and Roberts (2005)

1422 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

If leverage is between x (upper boundary) and x (lower boundary) the firmdoes nothing active Changes in leverage reflect various shocks to revenuesand costs that hit the firm If leverage drifts up to x (or above) then the firmrebalances to x If leverage drifts down to x (or below) the firm rebalancesto xWhat does this imply empirically A firm that does nothing can be

anywhere within the interval between x and x A firm that is rebalancingis a firm that has just hit one of the outer boundaries How does profitabilityaffect the likelihood of hitting a boundary High profits increase the value ofequity This increases the denominator of the leverage ratio and thus makesit more likely that the firm hits the lower boundary Low profits reduce thevalue of equity and thus make it more likely that the firm hits the upperboundaryTo be more specific suppose that the costs are symmetric and the leverage

changes are also following an ordinary Brownian motion Then the median

inert firm will have leverage ofxthornx

2 A firm that just actively reduced leverage

will be at x gtxthornx

2 A firm that just actively increased leverage will be at

x ltxthornx

2 A firm that just reduced leverage will generally be a firm that had

very high leverage To the extent that this reflects a lack of activerefinancing such firm will have low value of equity The low value ofequity will generally reflect low profits A firm that just increased leveragewill generally be a firm that had very low leverage In the same way a firmthat just increased its leverage will tend to have had high value of equity dueto high operating profits6

In this setting it is natural to focus on firms that are actively adjusting andto compare them to each other and to firms that are inactive It is alsohelpful to look directly at financing decisions and profitability rather thanfocusing on a leverage ratioOur approach shares the perspective of Welch (2004) Leary and Roberts

(2005) and Strebulaev (2007) that it is very important to focus on activeadjustments by firms Welch (2004) transforms the data to put everythingback into a ratio form His ldquoimplied debt ratiosrdquo still have both debt andequity in the denominator Similarly Leary and Roberts (2005) andStrebulaev (2007) pay a great deal of attention to leverage ratios whenstudying financing decisions Strebulaev (2007) does not have fixed costssince he makes use of a rescaling property Empirically we find differentadjusting behavior by large and by small firms So the rescaling property is

6 If we knew the values of the critical leverage points then it would be possible to derivemore refined tests Unfortunately those critical values are not observable

PROFITSndashLEVERAGE PUZZLEREVISITED 1423

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

an issue of concern His model does not account for the negative relationshipbetween current leverage and past market-to-book ratioWe depart from the earlier literature by arguing that the use of leverage

ratios can be particularly misleading for profitability since the use of aleverage ratio causes identification problems It is hard to be sure whethera leverage change reflects changes in the numerator or the denominator Thisis particularly problematic when as here the rebalancing is partialIt should be noted that much empirical capital structure literature debates

about whether to use a book leverage ratio or a market leverage ratio Bookleverage is used based on the argument that managers borrow against assetsin place which are better reflected by book assets Others such as Welch(2004) argue that book values are essentially an accounting fiction and thatonly market leverage should count Empirical papers regularly attempt tosidestep this debate by carrying out tests on both book leverage and marketleverage Our point is that both kinds of leverage ratios are potentially mis-leading when examining the impact of profits on capital structure

3 Data

We use conventional data sources starting with the merged Compustat-CRSP data The data are annual and are converted into constant year2000 dollars using the GDP deflator We exclude financial companies (SIC6000-6999) firms involved in major mergers (Compustat footnote code AB)firms reporting financial data in currencies other than the US dollars andfirms with missing data on our key variables7 The ratio variables aretrimmed at the 1 level in both tails of the distribution This serves toremove outliers and the most extremely misrecorded data The finalsample consists of 179021 firm-year observations from 1971 to 2009Table I provides definitions of financial variables and reports summary

statistics The average debt (in constant US$) is about $653 million while themedian is $24 million A significant fraction of firms have zero debt (the 10thpercentile is 0) Book equity is slightly larger than book debt Market equityis almost three times larger than book debt Book assets average $2191million although the medians are considerably smallerIf issuing or retiring securities incurs no fixed costs then we would expect

to see many small actions and very few large actions (Leary and Roberts2005) If there were significant fixed costs involved in issuing or retiringoutstanding securities then small issues might not be worthwhile Table I

7 These include debt book value of equity market value of equity assets book andmarket leverage profitability market-to-book assets ratio and tangibility

1424 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

shows that although most firms issue little debt or equity in a given year theaverages are large In other words when firms actually enter debt and equitymarkets they intervene massivelyThe mean constant dollar debt issue is $165 million (the median is $2

million) In unreported tables we find that the average debt issue is about81 of assets (the median is 18) About 38 of the firms issue no debt8 issue between 0 and 1 of the value of their assets as debt another

Table I Data description

Data sources The sample comes from the annual Compustat files The sample period is

1971ndash2009 We exclude financial companies (SIC 6000-6999) firms involved in majormergers (Compustat footnote code AB) firms reporting financial data in currencies otherthan the US dollar and firms with missing data on the key variables All financial variables

are deflated to year 2000 using the GDP deflator N is the number of observations SD isthe standard deviation D BVE MVE Assets Debt issuance Equity issuance Debt re-payment Equity repayment are all measured in $ millions The ratio variables are trimmed

at the 1 level in both tails of the distribution

Variable definitions Debt (D)frac14 long-term debt (dltt)thorn short-term debt (dlc) book equity(BVE)frac14 common shareholder equity (ceq) market equity (MVE)frac14 no of outstanding

shares (csho) closing price ( prcc f ) assetsfrac14 book assets (at) debt issuancefrac14 issuance oflong-term debt (Maxethdltis 0THORN)thorn increase in current debt (Maxethdlcch 0THORN) equity issu-ancefrac14 sale of common stock (Maxethsstk 0THORN) debt repaymentfrac14 reduction of long-termdebt (Maxethdltr 0THORN)thorn decrease in current debt (Minethdlcch 0THORN) equity repurchasefrac14 purchase

of common stock (Maxethprstkc 0THORN) book leveragefrac14D(DthornBVE) market leveragefrac14D(DthornMVE) profitabilityfrac14EBITDA (oibdp)assets market-to-book ratio M

B frac14Marketvalue of assets (MVA)assets where MVAfrac14DthornMVEthorn preferred-liq value (pstkl)

deferred taxes (txditc) tangibilityfrac14 net property plant and equipment (ppent)assets

Variable N Mean SD

Distribution

10th 50th 90th

Debt (D) 179021 653 4592 0 24 1048

Book equity (BVE) 179021 815 4122 3 70 1373

Market equity (MVE) 179021 1784 9589 9 118 2698

Assets 179021 2191 11530 10 153 3575

Debt issuance 179021 165 1335 0 2 245

Equity issuance 179021 26 186 0 0 40

Debt repayment 179021 132 1123 0 3 181

Equity repurchase 179021 24 257 0 0 9

Book leverage 179021 036 034 000 031 072

Market leverage 179021 027 025 000 020 066

Profitability 179021 005 027 018 011 023MB 179021 165 206 052 102 324

Tangibility 179021 031 024 005 026 070

PROFITSndashLEVERAGE PUZZLEREVISITED 1425

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

16 issue between 1 and 5 of the value of their assets as debt and theremaining 38 issue debt in excess of 5 of the value of their assetsThe mean constant dollar equity issue is about $26 million (the median is

about $04 million) As a fraction of assets the mean and median equityissues are about 7 and 02 respectively About 33 of the firms issue noequity 34 of the firms issue between 0 and 1 of the value of theirassets as equity another 14 issue between 1 and 5 of the value of theirassets as equity and the remaining 19 of the firms issue equity that is inexcess of 5 of the value of their assets Average debt repayments are largerthan equity repurchases This perhaps reflects the finite maturity of debt andits contractual repayment The median firm does not repurchase equityWe construct both book and market leverage ratios Book leverage is

defined as debt over debt plus book equity Market leverage is defined asdebt over debt plus market equity8 The median book leverage is 036 (theaverage is 031) The median market leverage is 027 (the average is 020)Profitability is defined as the ratio of operating income before depreci-

ation to assets While the average firm is profitable (the ratio of EBITDA toassets is 005) the median firm is even more profitable (with a profitabilityratio of 011) But the sample also includes a large number of unprofitablefirms as the 10th percentile is 018 The table also reports descriptive stat-istics on the market-to-book ratio and the tangibility ratio The market-to-book ratio (MB) defined as the ratio of the market value of assets to bookassets averages at about 165 Tangibility defined as the ratio of netproperty plant and equipment to assets averages at about 31

4 Leverage Ratio Regressions

The previous literature focuses on estimates obtained using leverage ratiosHence we start with a similar estimation to check whether our results matchthose of previous studies Table II presents the results for book leverage Weobtain similar results for market leverage and they are presented inAppendix Table AI The regressions include leverage factors following theprevious capital structure literature (see eg Lemmon Roberts and Zender(2008) and Frank and Goyal (2009)) The factors include (i) profitability (ii)industry median leverage (iii) market-to-book assets ratio (iv) tangibility ofassets and (v) firm size Industry median leverage is constructed as themedian leverage of all other firms in the industry excluding the firm under

8 Welch (2011) stresses the idea that nonfinancial liabilities should not be implicitly mis-treated as if they were equity by paying excessively narrow attention to financial liabilitiesin a leverage ratio We have adopted his approach in the empirical work reported here

1426 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

consideration Firm size is defined as the natural log of assets Other factorsare defined in Section 3 We cluster standard errors by firm and estimatethese regressions both with and without fixed effectsIn Table II we run conventional cross-section leverage regressions Since

we employ the usual data it is not surprising that our results match thosereported in the existing capital structure literature Profitability has anegative sign in both the book leverage regressions and in the marketleverage regressions The coefficients on other factors largely match thosereported in earlier studies Firms operating in industries with high leverage

Table II Leverage and profitability

The table presents parameter estimates for book leverage regressions The sample consists

of nonfinancial companies in the annual Compustat files during the period 1971ndash2009 Theleverage is estimated as the ratio of debt over debt plus book equity The explanatoryvariables Profitabilityt1

MB

t1

Tangibilityt1 and LnethAssetsTHORNt1 are described in

Table I IndLevt1 is estimated as the median book leverage of all other firms in thesame industry The industry is defined at the level of the four-digit SIC code All specifi-cations include the year fixed effects The specifications in column (2) additionally include

the firm fixed effects We report t-statistics where the standard errors are clustered at thefirm level in parentheses

aSignificant at the 1 level

Quantile regressions

(1) (2) (3) (4) (5)

Profitabilityt1 026a 020a 011a 017a 041a

(293) (192) (378) (518) (1015)

IndLevt1 042a 012a 030a 052a 052a

(369) (107) (926) (1252) (950)MB

t1

002a 001a 001a 002a 002a

(203) (102) (392) (426) (418)

Tangibilityt1 018a 025a 024a 019a 012a

(206) (154) (964) (580) (276)

LnethAssetsTHORNt1 001a 002a 002a 002a 001a

(114) (76) (899) (478) (96)

Constant 012a 010a 012a 002a 029a

(156) (60) (284) (27) (387)

Year FE Yes Yes Yes Yes Yes

Firm FE No Yes No No No

Regression type OLS FE 25th tile Median 75thtile

Clustered SE Yes Yes No No No

R2 013 054

Pseudo R2 016 015 009

N 158824 158824 158824 158824 158824

PROFITSndashLEVERAGE PUZZLEREVISITED 1427

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

tend to have high leverage A higher market-to-book ratio is associated withlower leverage Larger firms are typically more highly levered The coeffi-cients on tangibility are positive The conclusion from Table II is clear Theleverage regression results match those reported in previous studiesAt times one may be concerned that in a regression what is being

estimated is a conditional mean while some parts of the distributionmight behave differently from other parts If so then ldquoaccidentsrdquo likesample selection issues or asymmetries in the underlying distributionsmight play a greater role than is usually recognized In other words thesource of identification could be misinterpreted To guard against this po-tential problem we estimate the same model using quantile regressions9

Columns 3 4 and 5 report what happens at the 25th 50th and 75th per-centiles The basic model results are robust in terms of both the sign and thestatistical significanceThe impact of profits does appear to be somewhat stronger among the

high leverage firms For instance in a book leverage regression the coeffi-cient on profits for the first quartile is 011 with a t-ratio of 378 For thethird quartile the coefficient on profits is 041 with a t-ratio of 1015This difference is curious and might deserve further study However the factthat in each case we get the negative sign and statistical significance is suf-ficient for the purposes of the present article So the results on profits arerather robust across the distributionTable III sorts firms into those that are actively adjusting their leverage

and those that are passive In the presence of both fixed and variable trans-acts costs this distinction is important We expect the coefficient on profit-ability in leverage ratio regressions to be more negative in firms that arepassive and less negative in firms that are actively adjustingWe define active firms as those that are either issuing net debt or net

equity in excess of 5 of their assets Passive firms are those that neitherissue debt nor equity in nontrivial amounts (ie greater than 5 of theirassets) Consistent with our expectations we find in Table III that the coef-ficients on profits are significantly more negative for passive firms than theyare for active firms With firm and year fixed effects the coefficient onprofits is 008 for active firms and 017 for the passive firms The twocoefficients are significantly different from each other (plt 001) InTable III the standard control variables are used They have the usualsigns and for the most part do not differ much between the active and thepassive firms

9 Cameron and Trivedi (2010) provide an extensive discussion of quantile regressions inStata

1428 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

5 Financing Activity

51 FREQUENCY OF DEBT AND EQUITY ISSUES

Table IV reports the percentage of firms issuing or repurchasing debt orequity for annual sorts based on lagged profitability In Panel A weemploy the conventional 5 cutoff rule to exclude minor fluctuations anddefine firms ldquoissuing debtrdquo as those that issue debt (both short-term andlong-term) in excess of 5 of the value of their assets Other decisions aresimilarly defined using a 5 cutoff Appendix Table AII examines the dis-tributional properties across profitability (and size) quintiles for ldquolargerdquo is-suances by increasing the 5 cutoff to higher levels such as 10 and 15How does security issuance behavior vary with firm profitability To

answer this question Table IV sorts firms by profitability and reports the

Table III Leverage and profitability for active and passive firms

The table presents estimates of leverage on firm characteristics for the subsamples of firms

that are active in either debt or equity markets Active firms are defined as those that issueeither net debt in excess of 5 of the value of assets or issue net equity in excess of 5 ofthe value of their assets The passive firms do not issue significant (more than 5) of either

debt or equity The sample comes from the annual Compustat files during the period 1971ndash2009 Financial firms are excluded The variables are described in Tables I and II Wereport t-statistics where the standard errors are clustered at the firm level in parentheses

aSignificant at the 1 level

Active Passive

(1) (2) (3) (4)

Profitabilityt1 006a 008a 022a 017a

(126) (148) (254) (188)

IndLevt1 044a 019a 041a 013a

(424) (149) (326) (127)MB

t1

002a 001a 003a 001a

(362) (196) (327) (172)

Tangibilityt1 010a 014a 014a 019a

(146) (118) (145) (130)

LnethAssetsTHORNt1 001a 004a 001a 004a

(121) (179) (147) (169)

Constant 012a 008a 003a 007a

(120) (57) (45) (42)

Year FE Yes Yes Yes Yes

Firm FE No Yes No Yes

R2 034 064 033 075

N 43395 43395 58943 58943

PROFITSndashLEVERAGE PUZZLEREVISITED 1429

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

percentage of firms issuing or repurchasing debt or equity The table alsoreports test statistics from tests that examine whether the percentage differ-ences between the high- and low-profit firms are significantly different fromeach other The results in Column (1) show that the likelihood of issuing debt

Table IV Profitability sorts for debt and equity issuers

The sample contains nonfinancial firms listed on the annual Compustat files for the period

from 1971 to 2009 The table presents the percentage of firms issuing and retiring (orrepurchasing) debt and equity The firms are classified as ldquoissuing debtrdquo if they issuedebt in excess of 5 of the value of their assets as ldquoissuing equityrdquo if they issue equity

in excess of 5 of the value of their assets as ldquoretiring debtrdquo if they retire debt in excess of5 of the value of their assets and as ldquorepurchasing equityrdquo if they repurchase equity inexcess of 5 of the value of their assets In addition we also report net debt issuers which

are firms that issue net debt over 5 of the value of their assets and net equity issuerswhich are firms that issue net equity in excess of 5 of the value of their assets Weannually sort firms on lagged profitability and report the percentage of firms in each ofthese categories The bottom part of each panel reports the percentage of firms issuing or

retiring securities by profitability within the smallest and largest asset quintiles The tablealso reports results from t-test for the difference in percentage of firms issuing or retiringsecurities in the lowest and the highest profit quintile

Percentage of firms

Iss Ret Iss Iss Rep Iss Both IssD IssE

D D NetD E E NetE DampE RepE RetD

(1) (2) (3) (4) (5) (6) (7) (8) (9)

Allfirm

s

Low 276 245 186 301 13 297 71 01 38

2 319 299 181 117 21 112 27 02 20

3 385 344 194 86 29 80 21 05 15

4 386 330 213 81 43 75 20 07 14

High 329 273 198 92 82 81 17 13 14

t Highfrac14Low 150 84 39 723 414 763 358 176 200

Smallfirm

s Low 275 204 201 411 11 405 109 01 47

2 291 233 213 379 10 378 99 01 45

3 281 246 204 265 14 261 66 01 41

4 275 272 186 162 15 159 34 02 35

High 275 281 178 136 27 131 25 02 26

t Highfrac14Low 01 84 28 316 55 318 171 18 55

Largefirm

s

Low 414 363 177 87 22 79 28 03 11

2 449 379 180 62 30 57 17 07 09

3 439 360 186 56 41 48 16 09 07

4 403 320 194 46 67 39 13 12 06

High 374 264 207 50 129 35 10 28 05

t Highfrac14Low 50 130 45 89 244 116 77 120 47

1430 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

is mostly independent of firm profitability In the aggregate there is a smalldifference but it is sensitive to sorting on firm size Similarly in Column (2)there is only a weak relation between the profitability and the likelihood thatthe firm retires debt To a first approximation the fraction of firms issuing orrepurchasing debt is largely independent of profits Large firms are somewhatmore likely than small firms to issue debt and also to retire debtUnlike debt issues the probability of issuing equity is strongly related to

profitability as can be seen in Column (4) Low-profit firms are much morelikely to issue equity than are high-profit firms In the lowest profit quintile301 of the firms issued equity In the highest profit quintile it was only92 The lower panel reports separate sorts for small and large firms Similarpatterns are found for both groups of firms that is the likelihood of equityissuance declines as profitability increases The differences between the quin-tiles are quite stark for the group of small firms All quintiles of small firmsissue equity much more frequently than the corresponding large firm quintilesHigh-profit firms are much more likely to repurchase equity as shown in

Column (5) In the overall sample only 13 of low-profit firmsrepurchased equity compared with 82 of the high-profit firms Whenwe separate the small and large firms again a difference emerges Ingeneral small firms do not repurchase much equity while a greaterfraction of the large firms do soGiven these facts it is natural that low profitability firms are much more

likely to be net issuers of equity than are high profitability firms as shown inColumn (6) Among the firms in the lowest quintiles of profits roughly 30of firms issue net equity that exceeds 5 percent of their assets In contrastfor the most profitable quintile only about 8 issue net equityFirm size is an important variable in the recent literaturemdashit is often used

as a proxy for access to capital markets (as in Faulkender and Petersen(2006) and Leary (2009)) Small firms are bank-dependent risky and infor-mationally opaque They have restricted access to public debt markets andconsequently face more severe supply constraints in their ability to issuedebt Thus we expect small firms to be more sluggish in adjusting theirdebt and equity in response to shocks to profitability Large firms incontrast have much easier access to public debt markets and they facefewer obstacles in accessing securities marketsTo examine how size and profitability interact the bottom part of Table IV

examines profitability sorts for the smallest and largest firms We first sortfirms annually by firm size and then within each size quintile we sort them onprofitability We do these two-way sorts to ensure that we have similarnumber of firms in profitability subgroups for both small and large firmsamples Unconditionally small firms are generally less profitable while

PROFITSndashLEVERAGE PUZZLEREVISITED 1431

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

large firms are relatively more profitable The two-way sorts allow us to havea uniform distribution of firms in various profitability quintiles for both smalland large firm samples The bottom part of Table IV reports results for thesmallest and the largest size quintiles Among the small firms there is littlerelation between profitability and the likelihood of issuing debt Howeverlarger firms exhibit a small increase in the likelihood of net issuance of debtwith increasing profitabilityIn contrast to debt the effects of profitability on equity issuance and

repurchases are much more consistent across size quintiles As profitabilityincreases firms are generally less likely to issue equity and more likely torepurchase it Across the two extreme size groups we note a striking differ-ence between the proportion of firms issuing equity and those issuing debtAmong low-profit small firms almost 41 are equity issuers whereas onlyabout 9 of low-profit large firms are equity issuers However regardless ofsize we note a monotonic reduction in the likelihood of issuing equity asprofitability increasesColumn (7) reports the fraction of firms that issue both debt and equity in

excess of 5 of assets both as a function of profits in the upper panel andas a function of firm size and profits in the lower panel In the upper panelwe see that low-profit firms are much more likely to issue both debt andequity (71) while the high-profit firms are less likely to do so (17) Thelower panel shows that small firms are much more likely to be issuers of bothin contrast to large firms In both firm size categories the profit pattern is thesame with low profits being more frequent issuers of both simultaneouslyImportantly Column (8) shows that the likelihood of issuing debt and

simultaneously repurchasing equity increases with profitability Converselyin Column (9) we find that the likelihood of doing the reverse that is issuingequity and retiring debt declines with profitability The effects of sorting onfirm size and firm profitability mirror those reported for all firms Overallthe results show that low-profit firms are less likely to issue debt and repur-chase equity they are instead more likely to issue equity and retire debtFirms with high profitability exhibit the reverse patternIn Appendix Table AII we present results from profitability (and size)

quintiles for large issuances defined using the 10 cutoff (in Panel A) andthe 15 cutoff (in Panel B) The results from these panels are largely con-sistent with those described above

52 MAGNITUDES OF ISSUANCES AND EQUITY CHANGES

In the previous section we considered the probability of having a nontriviallevel of debt or equity activity The next question is how large are the dollar

1432 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

values involved In Table V we sort firms according to profits and thentabulate the levels and changes in both debt and equity We do this firstfor all firms and then for small versus large firmsFor the analysis on all firms we observe in Column (1) that debt peaks at

the middle of the distribution This is because firms with medium profitabil-ity are also the largest as seen in Column (8) which reports average assetvalues for different profitability quintiles When we sort first by firm size andthen examine firms with different levels of profitability within size quintileswe find that for small firms debt is roughly independent of profits while forlarge firms less profitable firms have a higher level of debtAs expected Columns (3) and (5) show that more profitable firms have

higher equity values Columns (2) and (7) consider the relationship betweenissuances of debt and equity and firm profitability Debt issuances are sig-nificantly larger for more profitable firms In contrast equity issuances arelarger for low-profit firms Small low-profit firms issue substantially lessequity compared with small high-profit firms Among large firms the low-profit ones tend to issue equity while the high-profit ones tend to repurchaseequity The fact that more profitable firms issue debt and repurchase equitywhile the least profitable firms retire debt and issue equity is consistent withthe predicted relation between profitability and financing decisions under thetrade-off theoryWe also examine two-way sorts by size and profitability and report results

for the smallest and largest quintile of firms For the smallest quintile offirms the change in debt is largely unrelated to profits But for large firmsthere is a positive relation between profits and debt issuances High-profitfirms have a big positive change in debt Low-profit large firms have anegative change in debtColumns (4) and (6) provide an explanation of why the leverage ratio

regression results contradict those from the basic profitability sorts pre-sented here As we can see profitability indirectly affects leverage ratiosby increasing equity values Changes in both the book value of equity andthe market value of equity are positive and large for highly profitable firmsIn contrast these changes are negative for less profitable firms This evidenceillustrates an important issue concerning the use of leverage ratios Suchratios are often interpreted as essentially reflecting the use of debt by thefirm This interpretation while common is empirically misleadingFor the typical firm the change in the value of equity is larger than the

change in debt For example in the third profit quintile for large firmscolumn (7) the mean equity issue is just $9 million but the change in themarket value of equity is $242 million At the same time the mean change indebt is $27 million This suggests that a fair bit of the observed variation in

PROFITSndashLEVERAGE PUZZLEREVISITED 1433

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

the leverage ratios is primarily driven by the changes in the market value ofequity in the denominator rather than by the changes in debt in the numer-ator Since equity issues are often small this implies that the variation in theleverage ratio is primarily driven by internal operations rather than byexternal financing actions This again points to the fact that leverageratios can provide a misleading account of actual patterns in the dataThere is always a potential concern that the averages may be misleading

due to the impact of outliers To address this concern in AppendixTable AIII we present median values of the profitability sorts Theseresults generally reinforce the findings in Table VTable VI again considers the magnitudes of financing activity but this

time the issues are scaled by a measure of firm size That way it is easier to

Table V Magnitude of financing activity

The sample contains nonfinancial firms listed on the annual Compustat files for the period

from 1971 to 2009 The table presents average debt and equity changes in debt and equityand equity issuances (in millions of dollars) for firms sorted on profitability within sizeclasses The table reports information for all firms sorted on profitability and for profit-

ability sorts within the smallest and largest firms The sorts are done annually The tablealso reports results from t-test for the difference in percentage of firms issuing or retiringsecurities in the lowest and the highest profit quintile

D D BVE BVE MVE MVE EqyIss Assets

(1) (2) (3) (4) (5) (6) (7) (8)

Allfirm

s

Low 83 5 78 10 120 5 7 234

2 605 13 509 24 721 8 17 1691

3 1227 9 1123 22 1834 3 9 3551

4 833 23 1166 7 2408 74 9 2994

High 570 53 1171 71 3817 68 36 2581

t Highfrac14Low 31 9 36 9 38 2 16 36

Smallfirm

s Low 25 04 42 01 332 28 33 97

2 29 06 55 02 278 16 21 116

3 31 07 63 01 212 05 12 127

4 30 06 70 05 179 10 08 136

High 24 06 84 14 234 17 07 142

t Highfrac14Low 04 15 153 85 73 11 201 107

Largefirm

s Low 3929 55 2990 152 4671 11 63 10846

2 3826 19 3178 95 5122 66 11 10721

3 3074 27 3498 41 6173 242 9 9697

4 2623 62 3915 45 8473 94 57 9855

High 2180 193 4642 261 14924 202 148 10171

t Highfrac14Low 94 64 108 87 246 12 169 16

1434 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

see whether large firms behave disproportionately when compared withsmall firms Column (1) shows that low-profit firms have more debtas reported in many previous studies In Columns (4) and (6) we seethat low-profit firms have a major decline in both the book value ofequity and the market value of equity The high-profit firms have thereverse experienceMore interesting are the active decisions of the firms In Column (2) we see

that in contrast to the usual impression the low-profit firms are reducingdebt while the high-profit firms are increasing it In Column (7) we see thatthe low-profit firms are also reducing leverage by issuing equity The high-profit firms issue little equity In the lower panel similar results are foundwhen we separate small and large firms Equity issuing is particularly strong

Table VI Magnitude of financing activitymdashscaled levels and changes

The sample contains nonfinancial firms listed on the annual Compustat files for the period

from 1971 to 2009 The table presents average debt and equity changes in debt and equityand equity issuances scaled by assets for firms sorted on profitability within size classes Thetable reports information for all firms sorted on profitability and for profitability sorts

within the smallest and largest firms The sorts are done annually The table also reportsresults from t-test for the difference in percentage of firms issuing or retiring securities inthe lowest and the highest profit quintile

D D BVE BVE MVE MVE EqyIss

(1) (2) (3) (4) (5) (6) (7)

Allfirm

s

Low 386 122 12 871 3427 1868 217

2 285 58 407 119 940 157 28

3 293 12 425 14 742 03 12

4 269 26 461 02 955 27 10

High 192 15 551 47 1675 74 07

t Highfrac14Low 35 49 50 47 127 42 157

Smallfirm

s Low 545 154 841 2145 9688 9128 696

2 684 49 712 1915 5014 2318 288

3 331 253 141 969 3063 954 129

4 267 181 357 175 1673 221 55

High 194 05 543 26 1758 147 41

t Highfrac14Low 49 17 51 36 88 27 99

Largefirm

s Low 350 30 306 58 511 55 09

2 346 14 335 13 523 02 04

3 353 165 332 71 711 29 01

4 280 01 414 04 909 09 05

High 227 15 477 19 1643 27 14

t Highfrac14Low 371 112 453 181 570 57 221

PROFITSndashLEVERAGE PUZZLEREVISITED 1435

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

among the small low-profit firms that are also experiencing sharp reductionsin the value of equity In column (2) we see that both for the large and for thesmall firms net debt increases are primarily found in the highest quintile offirm profitability

6 Debt and Equity Regressions

So far we have demonstrated that in our data the conventional leverageregressions have the usual signs We then demonstrated that in simple sortsof the data strikingly different active financing patterns emerge The nexttask is to reconcile these differencesTable VII presents simple regressions predicting changes in debt book

equity and market equity as well as equity issuances Consistent with thesorts we control for changes in firm size and we also include firm and yeardummies There may be a concern about the speed of a firmrsquos responsesAccordingly we consider the lagged change in profits as the main case butwe also include the contemporaneous change in profits Whether the contem-poraneous change in profits is included or not does not in fact matter Wereport panel-robust standard errors adjusted for clustering at the firm levelThe results in Table VII support the findings in the sorts Firms that have

an increase in profits in 1 year increase debt both in the same year and in thenext year Firms that have an increase in profits experience an increase inboth the book value of equity and the market value of equity Importantlyhowever firms that experience and increase in profits reduce their equityissues All these effects are statistically significant (except for the marketvalue of equity in the column 5 specification) and seem intuitively reason-able In Appendix Table AIV similar regressions to those presented aboveare reported but without firm-fixed effects We find that empirically verylittle changes whether we include firm fixed effects or leave them out Thusthe targeting behavior to the extent that it happens does not account forwhat we are reportingThe next step is to bring the sorts and the simple regressions together into

a common setting We therefore include factors in addition to profitability(i) median industry leverage (ii) the market-to-book assets ratio (iii) tangi-bility of assets and (iv) firm size (measured by the log of assets) Rajan andZingales (1995) show that these factors are related to leverage in G7countries A number of studies have used these factors to estimateleverage targets Frank and Goyal (2009) show that these factors arerobustly related to leverage in the USA In these regressions we use indica-tors for the quintile that the firm is in for each factor

1436 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

Table VIII reports these results In Columns (1) and (2) we examine debtissues and again show that profits positively affect debt issuances The effectsare large and statistically significant at the 1 level Columns (3) and (4)which examine book equity show a strong effect of profits This effect is onlyslightly reduced by the inclusion of other factors Columns (5) and (6) illus-trate that the impact of profits on the change in market equity is also robustto the inclusion of the conventional factors Columns (7) and (8) examineequity issuances Here we again find that profits have a significant negativeimpact on equity issuances As before the results are robust to the inclusionof the conventional factorsIn unreported tables we estimate debt and equity changes using quantile

regressions In terms of the signs and significance of the profitability variable

Table VII Debt and equity changes

The sample contains nonfinancial firms listed on the annual Compustat files for the period

from 1971 to 2009 The table presents fixed effect estimates from regressions of changes indebt (D) changes in the book value of equity (BVE) changes in the market value ofequity (MVE) and net equity issuances (EqyIss) The explanatory variables include

current and lagged change in operating income before depreciation (Profits andProfitst1) and lagged change in the book value of assets (Assetst1) In addition tofixed firm effects the regressions include year indicator variables The reported t-statistics

are corrected for clustering at the firm level

aMean significant at the 1 level

bMean significant at the 5 level

cMean significant at the 10 level

D BVE MVE EqyIss

(1) (2) (3) (4) (5) (6) (7) (8)

Profitst1 022a 025a 015b 035a 004 060b 005b 007a

(33) (31) (25) (35) (03) (25) (25) (53)

Profits 050b 078a 140a 003a

(24) (32) (48) (34)

Assetst1 000 007b 025a 001b

(00) (21) (35) (25)

Constant 167 283a 34 51 1145a 1182a 605a 595a

(14) (30) (04) (05) (32) (29) (69) (74)

Year FE Yes Yes Yes Yes Yes Yes Yes Yes

Firm FE Yes Yes Yes Yes Yes Yes Yes Yes

R2 001 004 001 010 001 005 002 003

N 162056 162056 162130 162130 157550 157550 162154 162154

PROFITSndashLEVERAGE PUZZLEREVISITED 1437

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

the results are qualitatively similar to the OLS results reported in Table VIIIProfitability positively affects debt issuances and negatively affects equityissuances across various quantiles Overall the results in Table VIII showthat there is nothing that is special about using sorts or regressions to explaindebt and equity issues In either case we find that more profitable firms tendto increase their debt experience an increase in the value of equity andrepurchase shares Thus the control factors are not responsible for theusual rejection of the trade-off theory It would appear that what mattersis the size of the issuing activity relative to the organic increases in equityvalue due to profits In other words what matters in the leverage regressionseems to be coming from the presence of E in D

DthornE and not from D

Table VIII Debt and equity issuances profitability and leverage factor quintiles

The sample contains nonfinancial firms listed on the annual Compustat files for the period

from 1971 to 2009 The table presents estimates from regressions of change in debt (D)change in the book value of equity (BVE) change in the market value of equity (MVE)and net equity issuances (EquityIss) The explanatory variables include lagged profitability

quintiles and lagged leverage factor quintiles In addition to fixed firm effects the regres-sions include year indicator variables The reported t-statistics are corrected for clustering atthe firm level

aMean significant at the 1 level

bMean significant at the 5 level

cMean significant at the 10 level

D BVE MVE EqyIss

(1) (2) (3) (4) (5) (6) (7) (8)

ProfitabilityQint 214a 167a 283a 175a 80 429a 53a 55a

(70) (54) (66) (59) (08) (48) (55) (62)

IndLevQint 104a 43 99 20

(29) (15) (10) (10)

Ln(Assets)Qint52 17 1026a 53a

(16) (05) (77) (35)MB

Qint158a 372a 1155a 09

(45) (56) (106) (06)

TangibilityQint 05 177 436c 64a

(01) (15) (18) (31)

Constant 167 96 656b 235 4613a 8610a 113 570a

(08) (04) (26) (04) (61) (58) (11) (31)

Year FE Yes Yes Yes Yes Yes Yes Yes Yes

Firm FE Yes Yes Yes Yes Yes Yes Yes Yes

R2 0002 0002 0003 0004 0013 0014 0013 0014

N 155587 155037 155587 155037 155587 155037 155587 155037

1438 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

7 Are Financial Market Conditions Important

According to the trade-off theory capital structure is determined by con-sidering the impact of costs and benefits of debt The time variation inthe costs and benefits of debt imply that issuance decisions would alsovary over time Furthermore market conditions affect the transactioncosts of adjusting Adverse selection is a greater problem in a cold marketthan in a hot one Accordingly it may be easier to adjust in good marketconditions Related ideas have been developed by Huang and Ritter(2009) Faulkender et al (2012) and Halling Yu and Zechner (2012)Faulkender Flannery Hankins and Smithrsquos paper examines cross-sectionalvariation in adjustment speeds including market valuations andhow they may affect adjustment speeds It is natural to think that ingood market conditions it will be less costly to issue both debt and equityand adjusting leverage to respond to profitability shocks would be mucheasierTo test the importance of market conditions we require a definition of

good times and bad times Our empirical strategy is to estimate good timesversus bad times at the four-digit industry level We define an industry ashaving ldquogood timesrdquo if the median firm in that industry has a market-to-book ratio that is higher than the 67th percentile of the time-series distribu-tion of the industry median market-to-book ratios Conversely an industryis defined as having ldquobad timesrdquo if the median firm in the industry has amarket-to-book ratio that is lower than the 33rd percentile of the time-seriesdistribution of the industry median market-to-book ratiosPanel A of Table IX tabulates issuance activity for profitability sorts in

both good times and bad times As might be expected active debt and equityissuances are larger during good times There is also somewhat more activeswaps between debt and equity during good timesPanel B reports the issuance activity for the smallest and the largest firms

during the good times and bad times Within each size quintile firms aresorted on profitability This is a three-way sort As before we find that thereis much more active use of external markets during good times Small low-profit firms are more likely to issue equity in good times than in bad timesDebt issuances are significantly higher in good times Large high-profitfirms are significantly more likely to issue debt and repurchase equity ingood times than in bad timesPanel C reports the magnitudes of the financial variables rather than the

frequencies During bad times less profitable large firms retire substantialamounts of debt and they show a tendency to issue equity Small firms donot seem to engage in similar debt reduction activities However like the

PROFITSndashLEVERAGE PUZZLEREVISITED 1439

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

Table IX Debt and equity issues in good and bad times

The table reports the frequency and magnitude of financing activity for sorts on profits for subsamples

of firms in good and bad times An industry is defined as having ldquogood timesrdquo if the median firm in

that industry has a market-to-book ratio that is higher than the 67th percentile of the time-series

distribution of industry median market-to-book ratios Conversely an industry is defined as having

bad times if the median firm in that industry has a market-to-book ratio that is lower than the 33rd

percentile of the time-series distribution of industry median market-to-book ratios Panels A and B

report the percentage of firms (i) issuing net debt in excess of 5 of the value of their assets (ii) issuing

net equity in excess of 5 of the value of their assets (iii) issuing debt and repurchasing equity both in

excess of 5 of the value of their assets and (iv) issuing equity and retiring debt both in excess of 5

of the value of their assets Panel C tabulates the mean debt and equity levels and the mean changes in

debt and equity issuances The sample contains nonfinancial firms listed on the annual Compustat files

for the period from 1971 to 2009

Iss Iss Iss D Iss ENet D Net E Rep E Ret D(1) (2) (3) (4)

Panel A sorts on profitability

Badtimes

Low 145 131 01 202 170 49 02 083 173 46 04 074 182 47 03 10

High 177 51 06 10t Highfrac14Low 41 136 33 01

Goodtimes

Low 206 381 01 462 197 148 02 233 212 103 04 184 234 92 10 18

High 216 92 16 14t Highfrac14Low 23 658 138 23

Panel B sorts on firm size and profitability

Badtimes

Smallfirm

s Low 151 318 00 312 182 276 02 253 151 140 02 294 183 73 00 19

High 177 80 00 11t Highfrac14Low 10 90 NA 20

Largefirm

s Low 172 54 01 042 148 39 04 053 150 43 04 044 156 33 03 06

High 147 28 05 03t Highfrac14Low 17 32 20 02

Goodtimes Smallfirm

s Low 213 448 01 542 230 443 01 473 222 325 00 474 203 197 02 38

High 190 153 02 31t Highfrac14Low 19 230 10 39

Largefirm

s Low 203 96 03 122 208 66 09 133 232 54 16 064 215 39 19 05

High 234 36 38 04t Highfrac14Low 29 99 92 38

(continued)

1440 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

larger firms there is a tendency for small firms to issue more equitymdashat leastto the extent possibleDuring good times large profitable firms raise significant amounts of debt

as they experience an increase in the value of their equity Such firms alsoengage in share repurchases Debt issues by smaller firms are much lessaffected by their own profitability during good times and their change indebt is more or less independent of their profits However equity issuanceis primarily found among the low-profit small firms during good timesTable IX shows that empirically profitability has a time-varying impacton leverage choicesIn Table X the basic regressions for changes in debt and for equity

issuance are presented for both good times and bad times Comparinggood times with bad times we see that the effects of profits are muchstronger in good times In bad times the impacts are rather weak statisticallyThis difference is interesting and deserves further study However for ourpurposes recall that the static trade-off theory implies time-varying capital

Table IX Continued

D D BVE BVE MVE MVE EqyIss Assets

(1) (2) (3) (4) (5) (6) (7) (8)

Panel C magnitude of financing activity

Badtimes Smallfirm

s Low 40 03 58 14 305 06 22 140

2 44 05 82 07 240 67 17 180

3 45 04 94 02 176 18 05 187

4 43 09 98 02 137 04 03 188

High 36 10 114 13 209 16 04 199

Largefirm

s Low 3365 73 2523 169 2304 163 94 8972

2 3031 95 2983 35 3284 278 56 9068

3 2930 5 3360 14 3947 308 57 9225

4 2281 6 3817 19 5213 442 7 8902

High 1790 62 4767 13 8799 58 31 9573

Goodtimes

Smallfirm

s Low 24 06 38 05 339 23 39 90

2 26 08 48 01 301 04 25 104

3 30 09 60 04 246 03 17 125

4 28 07 66 08 206 18 12 130

High 23 06 85 16 261 13 08 141

Largefirm

s Low 4532 36 3477 103 6557 260 53 12465

2 4580 58 3527 139 6911 421 23 12522

3 2952 111 3485 2 7586 279 32 9535

4 2752 124 3939 72 10686 33 105 10226

High 2429 228 4819 377 18565 268 222 10830

PROFITSndashLEVERAGE PUZZLEREVISITED 1441

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

structure choices even when the target is time invariant Empirically we doobserve time-varying choices

8 Is Scaling Important

In the preceding section we provided results for unscaled debt and equityissues This seems appropriate to us However it involves making twochanges to the standard regression We would like to determine if both

Table X Debt and equity changes in good and bad timesmdashregressions

The sample contains nonfinancial firms listed on the annual Compustat files for the period

from 1971 to 2009 The table presents estimates from regressions of change in debt (D)and net equity issuances (EqyIss) in good and bad times The explanatory variables includechange in profits lagged change in profits and lagged change in assets The regressions

include both firm and year fixed effects An industry is defined as having ldquogood timesrdquo ifthe median firm in that industry has a market-to-book ratio that is higher than the 67thpercentile of the time-series distribution of industry median market-to-book ratios

Conversely an industry is defined as having bad times if the median firm in thatindustry has a market-to-book ratio that is lower than the 33rd percentile of the time-series distribution of industry median market-to-book ratios The reported t-statistics arecorrected for clustering at the firm level

aMean significant at the 1 level

bMean significant at the 5 level

cMean significant at the 10 level

Good times Bad times

D EqyIss D EqyIss

(1) (2) (3) (4) (5) (6) (7) (8)

Profits 087a 091a 006b 009a 060 065 003 003b

(30) (29) (20) (28) (15) (16) (16) (20)

Profitst1 029c 015a 034c 001

(19) (37) (17) (04)

Assetst1 000 002b 000 001

(02) (22) (00) (07)

Constant 1337 3989 8691a 10001a 1351 1851 4099a 4055a

(03) (09) (55) (37) (09) (13) (80) (78)

Year FE Yes Yes Yes Yes Yes Yes Yes Yes

Firm FE Yes Yes Yes Yes Yes Yes Yes Yes

R2 0071 0080 0018 0045 0178 0212 0034 0041

N 74827 68729 74891 68787 26965 26013 26971 26018

1442 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

changes are crucial Accordingly in this section we take a step back toconsider prior literature in which variables are scaled The questions arehow does scaling affect the results and whyTable XI examines alternative scalings Columns (1) and (2) scale debt and

equity issuances by the total firm net issuance Columns (3) and (4) scaledebt and equity issuances by the sum of debt and book equity Columns (5)and (6) consider the change in debt scaled by the sum of debt and bookequity Columns (7) and (8) consider the change in leverage due to externalfinancing Since we are examining changes we use the changes in thestandard regression factors as controls The primary interest is in the coef-ficient on the change in profitsWe find that a change in profits is associated with an increase in debt

issues But the sign reverses when we scale the debt issues by the sum of debtplus book equity This is telling What is really driving the results is thechange in equity in the denominator not the change in debt in the numer-ator This result is further substantiated in Columns (5) and (6) where thechange in debt to capital ratio is used instead of debt issuanceTo further explore the impact of active issuance relative to passive

experiencing of changes we decompose leverage by separating out theeffect of debt and equity issues from those due to changes in retainedearnings Thus we define leverage change due to external financing as

LevEF frac14Dt1 thorn d

Dt1 thorn Et1 thorn dthorn e

Dt1

Dt1 thorn Et1 eth1THORN

where D is debt E is book value of equity d is debt issues net of retirementsand e is equity issues net of repurchases Garvey and Hanka (1999) use asimilar measure in their study This measure directly captures the netleverage effect of debt and equity issuancesColumns (7) and (8) report results from regressions of the net leverage

effect of debt and equity issues on changes in profitability after controllingfor other leverage factors in differences and year indicator variables Theeffect of profitability on leverage change due to external financing is positiveand statistically significant This comparison suggests that changes in totaldebt and equity are only partly a result of debt and equity issuance decisionsOther balance sheet adjustments complicate the inferences from leverageratio regressionsFor most of our analysis we follow the literature in treating debt as homo-

genous It is of course potentially interesting to ask what happens when wedistinguish between short-term debt and long-term debt Which part of thefirmrsquos debt maturity structure drives the under-adjustment Table XII

PROFITSndashLEVERAGE PUZZLEREVISITED 1443

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

provides results from such decompositions Alternative scalings areconsidered The main results of interest are that the distinction betweenshort- and long-term debt is important Short-term debt issuance is largelyindependent of changes in profits The negative sign on profit change stemsfrom the action of long-term debt This seems intuitively reasonable

Table XI Scaled issuances

The sample contains nonfinancial firms listed on the annual Compustat files for the period

from 1971 to 2009 The table presents estimates from regressions of debt issuance to totalissuance (in columns (1) and (2)) debt issuance to total capitalization (in columns (3) and(4)) change in debt to capitalization ratio (in columns (5) and (6)) and change in leverage

ratio considering only external financing (in columns (7) and (8)) The change in leverageratio considering only external financing is defined as

LevEF frac14Dt1 thorn d

Dt1 thorn Et1 thorn dthorn e

Dt1

Dt1 thorn Et1

where D is debt E is book value of equity d is debt issues net of retirements and e is equity

issues net of repurchases The regressions include lagged leverage factors and year indicatorvariables Some of the specifications include firm fixed effects The reported t-statistics arecorrected for clustering at the firm level

aMean significant at the 1 level

bMean significant at the 5 level

cMean significant at the 10 level

DebtIssTotIss

DebtIssethDthornETHORN D

DthornE LevEF

(1) (2) (3) (4) (5) (6) (7) (8)

Profitt1 004a 000 002a 001c 003a 002 001b 002a

(35) (03) (34) (18) (35) (14) (23) (26)

IndLevt1 004c 001 001 001c 000 000 000 000

(19) (06) (15) (19) (02) (04) (03) (05)

Market=bookt1 001a 001a 000a 000a 000a 000a 000 000c

(47) (85) (68) (60) (55) (36) (05) (18)

Tangt1 004 005 013a 011a 011a 006a 003a 003a

(12) (15) (144) (112) (79) (38) (44) (42)

Assetst1 008a 004a 006a 004a 002a 001a 003a 002a

(97) (45) (283) (161) (53) (35) (147) (101)

Constant 079a 074a 002a 003a 000 000 000b 002a

(497) (442) (91) (111) (00) (05) (22) (93)

Year FE Yes Yes Yes Yes Yes Yes Yes Yes

Firm FE No Yes No Yes No Yes No Yes

R2 0015 0098 0022 0076 0009 0070 0012 0086

N 133905 133905 140552 140552 138229 138229 140572 140572

1444 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

inasmuch as short-term debt is probably more associated with day-to-dayoperations rather than longer term planning considerationsIn Table XIII we examine change in leverage ( D

DthornE) and change inleverage due to external financing (LEVEF) during good and bad timesseparately The traditional negative effect of a change in profits on achange in leverage is found both in good times and in bad timesHowever when we decompose the impact in order to isolate the externalcomponent as before the result disappears In good times the externalfinancing component of a change in profits is weak but still positiveDuring bad times it is not statistically significant This implies that distin-guishing between good and bad times does not fundamentally alter the basic

Table XII Short- and long-term debt issuances

The sample contains nonfinancial firms listed on the annual Compustat files for the period

from 1971 to 2009 The table presents estimates from regressions of short-term debtissuance to total issuance (in columns (1) and (2)) long-term debt issuance to totalissuance (in columns (3) and (4)) short-term debt issuance to total capitalization (in

columns (5) and (6)) and long-term debt to total capitalization (in columns (7) and (8))Some of the specifications include firm fixed effects The reported t-statistics are correctedfor clustering at the firm level

aMean significant at the 1 level

bMean significant at the 5 level

cMean significant at the 10 level

STDebtIssTotIss

LTDebtIssTotIss

STDebtIssethDthornETHORN

LTDebtIssethDthornETHORN

(1) (2) (3) (4) (5) (6) (7) (8)

Profitabilityt1 0002 0003 0042a 0008 0001 0001 0019a 0012a

(02) (04) (30) (05) (14) (10) (45) (28)

IndLevt1 0009 0007 0070b 0033 0001 0002c 0014a 0005

(06) (05) (26) (11) (08) (17) (26) (09)

M=Bt1 0003a 0003a 0003b 0010a 0000 0000 0003a 0002a

(48) (35) (22) (65) (10) (01) (84) (42)

Tangibilityt1 0020 0007 0011 0067 0005a 0006a 0097a 0081a

(12) (03) (03) (16) (34) (33) (108) (89)

Assetst1 0039a 0011b 0034a 0024b 0001a 0001 0046a 0020a

(89) (22) (38) (23) (36) (13) (190) (82)

Constant 0001b 0153a 0798a 0591a 0000c 0006a 0057a 0096a

(25) (144) (470) (288) (17) (124) (292) (234)

Year FE Yes Yes Yes Yes Yes Yes Yes Yes

Firm FE No Yes No Yes No Yes No Yes

R2 0025 0089 0018 0073 0030 0147 0016 0298

N 136146 136146 136140 136140 143031 143031 142847 142847

PROFITSndashLEVERAGE PUZZLEREVISITED 1445

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

Table XIII Externally financed leverage changes in good and bad times

The sample contains nonfinancial firms listed on the annual Compustat files for the period

from 1971 to 2009 The table presents estimates from regressions of change in debt tocapitalization ratio (in columns (1) (2) (5) and (6)) and change in leverage ratio consider-ing only external financing (in columns (3) (4) (7) and (8)) The change in leverage ratio

considering only external financing is defined as

LevEF frac14Dt1 thorn d

Dt1 thorn Et1 thorn dthorn e

Dt1

Dt1 thorn Et1

where D is debt E is book value of equity d is debt issues net of retirements and e is equityissues net of repurchases The regressions include lagged leverage factors and year indicator

variables The reported t-statistics are corrected for clustering at the firm level An industryis defined as having ldquogood timesrdquo if the median firm in that industry has a market-to-bookratio that is higher than the 67th percentile of the time-series distribution of industry

median market-to-book ratios Conversely an industry is defined as having ldquobad timesrdquoif the median firm in that industry has a market-to-book ratio that is lower than the 33rdpercentile of the time-series distribution of industry median market-to-book ratios

aMean significant at the 1 level

bMean significant at the 5 level

cMean significant at the 10 level

Good times Bad times

DDthornE LevEF D

DthornE LevEF

(1) (2) (3) (4) (5) (6) (7) (8)

Profitabiltyt1 0022c 0005 0016c 0025b 0073b 0041 0007 0007

(17) (03) (18) (24) (25) (11) (04) (03)

IndLevt1 0000 0011 0001 0009 0010 0018 0001 0000

(00) (09) (01) (13) (07) (10) (02) (00)

MB

t1

0004a 0003b 0000 0001 0008a 0001 0000 0000

(40) (24) (05) (08) (28) (04) (03) (03)

Tangt1 0097a 0041 0033a 0038b 0127a 0048 0059a 0033

(45) (16) (27) (25) (49) (14) (43) (16)

Assetst1 0014a 0013b 0024a 0019a 0037a 0023c 0036a 0025a

(30) (20) (84) (48) (45) (17) (88) (43)

Constant 0030b 0036 0000 0003 0028c 0014 0021 0001

(20) (16) (00) (07) (17) (06) (11) (01)

Year FE Yes Yes Yes Yes Yes Yes Yes Yes

Firm FE No Yes No Yes No Yes No Yes

R2 0006 0088 0009 0117 0026 0251 0023 0202

N 56820 56820 58200 58200 23462 23462 23511 23511

1446 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

conclusions The traditional leverage result on profits is driven by increasesin equity that are experienced and then partially but not completely offsetby firmsTo the extent that firms are engaging in active financing they are following

the prescriptions of the trade-off theory Thus we have a coherent picture ofhow firms are responding to profits Basically more profitable firms experiencean increase in the value of their equity They respond to this by issuing debtand repurchasing equity But the magnitude of the increase in equity is largerthan the magnitudes of the active steps taken in the debt and equity marketsThis means that the firmsrsquo actual actions are as predicted by the trade-offtheory But it leaves open the question of why they do not act more stronglyThe capital structure literature has long been interested in dynamics and

capital structure adjustment paths In recent years this interest has increas-ingly resulted in explicit dynamic optimizing models such as those ofGoldstein Ju and Leland (2001) Hennessy and Whited (2005) andStrebulaev (2007) These models commonly make specialized assumptionsleading some to question their generalityWe argue that both fixed transactions costs and variable transactions costs

are plausible and realistic in models that pin down firm size The simultaneouspresence of both is important Due to fixed costs firms only rebalance inter-mittently Due to variable adjustment costs the extent of the adjustment isreduced to the point at which an extra unit to adjustment is balanced againstthe extra unit transactions costs This means that both leverage increases andleverage reductions will only go part way toward a static optimumThe key point is that with both fixed and variable transactions costs we

should expect to only see intermittent leverage adjustments When leverage isadjusted it should ldquounder adjustrdquo when compared to a frictionless modelFurther discussion of transactions costs in a capital structure setting isprovided by Leary and Roberts (2005)

9 Conclusion

The connection between corporate profits and capital structure has beenvery influential in the assessment of the trade-off theory The conventionalinterpretation of the evidence has pushed the literature away from the trade-off theory and toward much more complex dynamic models and ideas As aresult it is important to make sure that the evidence has been correctlyinterpreted We show that the problem is with the interpretation of thetest results and a failure to take into account the costs associated with re-sponding to a profit shock

PROFITSndashLEVERAGE PUZZLEREVISITED 1447

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

We document that more profitable firms really do borrow more (not less)They tend to repurchase equity They experience an increase in both thebook value of equity and the market value of equity Less profitable firmsreally do tend to reduce their debt and to issue equityFirm size and market conditions also matter Larger firms tend to be more

active in the debt markets while smaller firms tend to be relatively moreactive in the equity markets During good times there is more use ofexternal financingThe usual profitsndashleverage puzzle result is primarily driven by the increase

in equity that is experienced by the more profitable firms Accordingly thepuzzle should be restated as asking why do firms not take sufficiently largeoffsetting actions to fully undo the change in equity What limits the mag-nitudes of the typical leverage response to profit shocksIn a frictionless model the partial response appears puzzling But there is

good empirical reason to believe that rebalancing entails both fixed andvariable costs and that firm size matters The rebalancing costs can befully avoided by doing nothing Accordingly the firm must decide whetherany given shock is big enough to be worth responding to If it is then thefirm must decide how big a response is called for These technical conditionsare known as ldquovalue matchingrdquo and ldquosmooth pastingrdquo Thus optimizationimplies that some shocks will be ignored Even if the shock is not ignoredthe optimal response will only partially undo the shock The magnitude ofthe leverage response must balance the marginal cost and the marginalbenefit of an extra unit of leverage Since the marginal cost of adjustingleverage is strictly positive the adjustment toward that static leverageoptimum will only go part way This is true both for leverage increasesand for leverage reductionsOne might conjecture that the variable costs would be time varying and

that such costs might have declined over time Since we do not have a fullaccounting of all of the variable costs it is of course impossible to be sureHowever there is some evidence suggesting that at least the directunderwriting fees are remarkably stable over time10

Surely nobody ought to be surprised that more profitable firms are moreinclined to repurchase equity while unprofitable firms tend not to do so Thefact that more profitable firms typically experience an increase in the value of

10 For IPOs Jay Ritter provides important summary evidence as httpbearwarringtonufleduritterIPOs2013Underwritingpdf He finds that for medium-sized deals a remark-

ably large fraction of the deals have a 7 fee year after year Whether this is due tobehavioral factors or some sort of a coordination device is not clear The extent towhich similar intertemporal stability carries over to other forms of external financemerits further study

1448 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

equity is equally natural Were it not for the contrary claims in the literature(ie the profitsndashleverage puzzle) we would argue that nobody ought to findit surprising that more profitable firms are more inclined to issue debt Thefact that leverage adjustments typically only partially offset profit shocksshould also sound quite natural These facts all fit together easily within thetrade-off theory once we take account of both fixed and variable costs ofactively adjusting capital structure

Appendix

Table AI Market leverage regressions

The table presents estimates of the regressions of market leverage on firm characteristics

The sample comes from the annual Compustat files during the period 1971ndash2009 Financialfirms are excluded The leverage is estimated as the ratio of debt over debt plus marketequity The explanatory variables Profitabilityt1

MB

t1

Tangibilityt1 and LnethAssetsTHORNt1are described in Table 1 IndMedianLevt1 is estimated as the median market leverage of allother firms in the same industry The industry is defined at the level of the four-digit SICcode All specifications include the year fixed effects The specifications in column (2)

additionally include the firm fixed effects We report t-statistics where the standard errorsare clustered at the firm level in parentheses

aSignificant at the 1 level

Quantile regressions

(1) (2) (3) (4) (5)

Profitabilityt1 013a 013a 007a 011a 020a

(275) (262) (349) (425) (538)

IndLevt1 046a 020a 032a 058a 063a

(506) (239) (1299) (1630) (1042)MB

t1

003a 001a 001a 002a 004a

(509) (303) (743) (709) (534)

Tangibilityt1 013a 018a 013a 014a 010a

(191) (181) (762) (571) (232)

LnethAssetsTHORNt1 001a 005a 001a 001a lt001a

(112) (258) (682) (342) (29)

Constant 010a 011a 005a 005a 029a

(171) (98) (173) (118) (398)

Year FE Yes Yes Yes Yes Yes

Firm FE No Yes No No No

Regression type OLS FE 25th tile Median 75thtile

Clustered SE Yes Yes No No No

R2 030 067

Pseudo R2 016 023 0201

N 158578 158578 158578 158578 158578

PROFITSndashLEVERAGE PUZZLEREVISITED 1449

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

Table AII Profitability sorts for firms issuing large amounts

Percentage of firms

Iss Ret Iss Iss Rep Iss Iss IssD IssE

D D NetD E E NetE DampE RepE RetD

(1) (2) (3) (4) (5) (6) (7) (8) (9)

Panel A 10 cutoff to define issuing or retiring activity

Allfirm

s

Low 205 150 134 260 05 257 45 00 23

2 207 175 103 77 06 75 11 01 10

3 243 196 101 50 08 47 08 01 09

4 252 186 113 50 13 48 08 02 08

High 227 161 118 60 27 57 07 04 09

t Highfrac14Low 70 38 62 755 226 768 320 90 152

Smallfirm

s Low 230 132 162 380 05 375 82 00 32

2 228 143 165 334 04 333 65 00 26

3 211 146 147 216 03 214 38 00 23

4 190 154 120 126 04 124 15 01 20

High 194 162 114 99 09 97 12 01 15

t Highfrac14Low 42 40 68 342 20 340 169 07 131

Largefirm

s Low 231 200 82 31 06 29 08 01 04

2 231 187 67 17 06 14 03 01 02

3 234 177 77 17 09 14 03 02 03

4 218 158 75 16 17 14 02 03 02

High 218 135 99 19 41 15 03 07 02

t Highfrac14Low 19 106 36 46 135 57 39 58 23

Percentage of firms

Iss Ret Iss Iss Rep Iss Both Iss D Iss E

D D Net D E E Net E DampE Rep E Ret D

(1) (2) (3) (4) (5) (6) (7) (8) (9)

Panel B 15 cutoff to define issuing or retiring activity

Allfirm

s

Low 158 101 100 230 01 228 27 00 15

2 142 115 63 57 01 56 05 00 06

3 165 128 57 32 01 31 03 00 05

4 173 118 65 35 01 34 03 00 05

High 160 102 70 46 04 44 03 00 05

t Highfrac14Low 08 06 141 738 87 746 277 24 131

Smallfirm

s Low 195 95 132 354 01 350 56 00 24

2 185 96 131 303 01 301 44 00 18

3 160 96 104 179 00 176 18 00 15

4 134 99 78 99 01 98 08 00 12

High 133 95 73 82 01 80 06 00 10

t Highfrac14Low 80 00 95 345 12 344 149 12 52

Largefirm

s Low 145 123 42 14 01 12 03 00 01

2 137 114 32 07 01 06 01 00 01

3 142 104 35 06 01 05 00 00 01

4 130 91 36 08 03 07 01 01 01

High 134 81 48 09 06 07 01 01 01

t Highfrac14Low 19 85 17 26 54 33 34 17 11

1450 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

Table AIV Debt and equity changesmdashwithout fixed effects

The sample contains nonfinancial firms listed on the annual Compustat files for the period from 1971

to 2009 The table presents estimates from regressions of change in debt (D) change in the book

value of equity (BVE) change in the market value of equity (MVE) and net equity issuances

(EqyIss) The explanatory variables include current and lagged change in operating income before

depreciation (Profits and Profitst1) and lagged change in book value of assets (Assetst1) The

regressions include year indicator variables The reported t-statistics are corrected for clustering at the

firm level

aMean significant at the 1 level

bMean significant at the 5 level

cMean significant at the 10 level

D BVE MVE EqyIss

(1) (2) (3) (4) (5) (6) (7) (8)

Profitst1 025a 024a 023a 038a 007 066a 007a 009a

(40) (32) (36) (39) (06) (29) (42) (51)Profits 051b 082a 149a 004a

(25) (33) (49) (33)Assetst1 002 005c 024a 001b

(08) (17) (34) (23)Constant 93 176a 45 71 1552a 1704a 130a 140a

(12) (28) (07) (09) (49) (49) (45) (45)Year FE Yes Yes Yes Yes Yes Yes Yes YesR2 001 005 001 011 001 005 002 002N 162056 162056 162130 162130 157550 157550 162154 162154

Table AIII Median profitability sorts

The sample contains nonfinancial firms listed on the annual Compustat files for the period from 1971

to 2009 The table presents median debt and equity median changes in debt and equity and median

equity issuances (EqyIss) in millions of dollars for all firms sorted on profitability and for profitability

sorts within the smallest and largest firms The sorts are done annually

D D BVE BVE MVE MVE EqyIss Assets(1) (2) (3) (4) (5) (6) (7) (8)

Allfirm

s

Low 21 00 75 17 256 20 01 1922 186 02 449 18 598 19 00 11033 816 06 1274 03 1637 05 00 32244 787 02 1541 18 2488 03 00 3556High 273 00 1361 66 3300 14 00 2578

Smallfirm

s Low 05 00 13 07 140 14 06 402 06 00 25 08 114 12 01 563 09 00 32 05 85 07 00 674 12 00 39 01 68 03 00 79High 10 00 51 04 98 01 00 90

Largefirm

s Low 11485 381 10269 371 13227 202 03 333902 13051 393 12786 140 17571 169 05 392463 11554 287 13429 24 21623 305 00 370314 9963 209 15052 194 29424 308 00 36705High 7361 00 16380 613 49142 439 00 34976

PROFITSndashLEVERAGE PUZZLEREVISITED 1451

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

References

Auerbach A J (1985) Real determinants of corporate leverage in B M Friedman (ed)

Corporate Capital Structures in the United States University of Chicago Press Chicago

IllinoisBertrand M and Mullainathan S (2003) Enjoying the quiet life corporate governance and

managerial preferences Journal of Political Economy 111 1043ndash1075Booth L Varouj A Demirguc-Kunt A and Maksimovic V (2001) Capital structures in

developing countries Journal of Finance 56 87ndash130Brennan M J and Schwartz E S (1984) Optimal financial policy and firm valuation

Journal of Finance 39 593ndash607Cameron A C and Trivedi P K (2010) Microeconometrics using Stata Stata Press TexasConstantinides G M (1979) Multiperiod consumption and investment behavior with

convex transactions costs Management Science 25 1127ndash1137Constantinides G M and Richard S F (1978) Existence of optimal simple policies for

discounted-cost inventory and cash management in continuous time Operations Research

26 620ndash636DeAngelo H DeAngelo L and Whited T M (2011) Capital structure dynamics and

transitory debt Journal of Financial Economics 99 235ndash261Dixit A K (1993) The Art of Smooth Pasting Routledge London and New YorkFama E and French K R (2002) Testing trade-off and pecking order predictions about

dividends and debt Review of Financial Studies 15 1ndash33Faulkender M and Petersen M A (2006) Does the source of capital affect capital struc-

ture Review of Financial Studies 19 45ndash79Faulkender M Flannery M J Hankins K W and Smith J M (2012) Cash Flows and

Leverage Adjustments Journal of Financial Economics 103 632ndash646Fischer E Heinkel R and Zechner J (1989) Dynamic capital structure choice theory

and tests Journal of Finance 44 19ndash40Frank M Z and Goyal V K (2008) Trade-off and pecking order theories of

debt in B E Eckbo (ed) Handbook of Corporate Finance Empirical Corporate Finance

Vol 2 Handbook of Finance Series Chapter 12 ElsevierNorth-Holland Amsterdam pp

135ndash202

Frank M Z and Goyal V K (2009) Capital structure decisions which factors are reliably

important Financial Management 38 1ndash37

Garvey G T and Hanka G (1999) Capital structure and corporate control the effect of

antitakeover statutes on firm leverage Journal of Finance 54 519ndash546

Goldstein R Ju N and Leland H (2001) An EBIT-based model of dynamic capital

structure Journal of Business 74 483ndash512

Graham J R and Tucker A L (2006) Tax shelters and corporate debt policy Journal of

Financial Economics 81 563ndash594

Hackbarth D and Mauer D C (2012) Optimal priority structure capital structure and

investment Review of Financial Studies 25 747ndash796

Halling M Yu J and Zechner J (2012) Leverage dynamics over the business cycle

Working paper University of Utah University of New South Wales and Vienna

University of Economics and Business AdministrationHennessy C A and Whited T A (2005) Debt dynamics Journal of Finance 60 1129ndash1165

Huang R and Ritter J R (2009) Testing theories of capital structure and estimating the

speed of adjustment Journal of Financial and Qualitative Analysis 44 237ndash271

1452 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

Hovakimian A Opler T and Titman S (2001) The debt-equity choice Journal of

Financial and Quantitative Analysis 36 1ndash24Jung K Kim Y and Stulz R M (1996) Timing investment opportunities managerial

discretion and the security issue decisions Journal of Financial Economics 42 159ndash185Kane A Marcus A J and McDonald R L (1984) How big is the tax advantage to

debt Journal of Finance 39 841ndash853Kayhan A and Titman S (2007) Firmsrsquo histories and their capital structures Journal of

Financial Economics 83 1ndash32Leary M T (2009) Bank loan supply lender choice and corporate capital structure

Journal of Finance 64 1143ndash1185Leary M T and Roberts M R (2005) Do firms rebalance their capital structures

Journal of Finance 60 2575ndash2619Leland H E (1994) Corporate debt value bond covenants and optimal capital structure

Journal of finance 49 1213ndash1252Lemmon M L Roberts M R and Zender J F (2008) Back to the beginning persist-

ence and the cross-section of corporate capital structure Journal of Finance 63 1575ndash1608Long M S and Malitz I B (1985) The investment-financing nexus some empirical

evidence Midland Corporate Finance Journal 3 53ndash59Mackie-Mason J K (1990) Do taxes affect corporate financing decisions Journal of

Finance 45 1471ndash1493Morellec E Nikolov B and Schurhoff N (2012) Corporate governance and capital

structure dynamics Journal of Finance 67 803ndash848Myers S C (1984) The capital structure puzzle Journal of Finance 39 575ndash592Myers S C (1993) Still searching for optimal capital structure Journal of Applied

Corporate Finance 61 4ndash14Rajan R and Zingales L (1995) What do we know about capital structure some evidence

from international data Journal of Finance 50 1421ndash1460Strebulaev I A (2007) Do tests of capital structure theory mean what they say Journal of

Finance 62 1747ndash1787

Strebulaev I A and Whited T M (2012) Dynamic Models and Structural Estimation in

Corporate Finance Now Publisher Boston and Delft

Titman S and Wessels R (1988) The determinants of capital structure choice Journal of

Finance 43 1ndash21

Welch I (2004) Capital structure and stock returns Journal of Political Economy 112

106ndash131

Welch I (2011) Two common problems in capital structure research the financial-debt-

to-asset ratio and issuing activity versus leverage changes International Review of Finance

11 1ndash17

PROFITSndashLEVERAGE PUZZLEREVISITED 1453

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

If leverage is between x (upper boundary) and x (lower boundary) the firmdoes nothing active Changes in leverage reflect various shocks to revenuesand costs that hit the firm If leverage drifts up to x (or above) then the firmrebalances to x If leverage drifts down to x (or below) the firm rebalancesto xWhat does this imply empirically A firm that does nothing can be

anywhere within the interval between x and x A firm that is rebalancingis a firm that has just hit one of the outer boundaries How does profitabilityaffect the likelihood of hitting a boundary High profits increase the value ofequity This increases the denominator of the leverage ratio and thus makesit more likely that the firm hits the lower boundary Low profits reduce thevalue of equity and thus make it more likely that the firm hits the upperboundaryTo be more specific suppose that the costs are symmetric and the leverage

changes are also following an ordinary Brownian motion Then the median

inert firm will have leverage ofxthornx

2 A firm that just actively reduced leverage

will be at x gtxthornx

2 A firm that just actively increased leverage will be at

x ltxthornx

2 A firm that just reduced leverage will generally be a firm that had

very high leverage To the extent that this reflects a lack of activerefinancing such firm will have low value of equity The low value ofequity will generally reflect low profits A firm that just increased leveragewill generally be a firm that had very low leverage In the same way a firmthat just increased its leverage will tend to have had high value of equity dueto high operating profits6

In this setting it is natural to focus on firms that are actively adjusting andto compare them to each other and to firms that are inactive It is alsohelpful to look directly at financing decisions and profitability rather thanfocusing on a leverage ratioOur approach shares the perspective of Welch (2004) Leary and Roberts

(2005) and Strebulaev (2007) that it is very important to focus on activeadjustments by firms Welch (2004) transforms the data to put everythingback into a ratio form His ldquoimplied debt ratiosrdquo still have both debt andequity in the denominator Similarly Leary and Roberts (2005) andStrebulaev (2007) pay a great deal of attention to leverage ratios whenstudying financing decisions Strebulaev (2007) does not have fixed costssince he makes use of a rescaling property Empirically we find differentadjusting behavior by large and by small firms So the rescaling property is

6 If we knew the values of the critical leverage points then it would be possible to derivemore refined tests Unfortunately those critical values are not observable

PROFITSndashLEVERAGE PUZZLEREVISITED 1423

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

an issue of concern His model does not account for the negative relationshipbetween current leverage and past market-to-book ratioWe depart from the earlier literature by arguing that the use of leverage

ratios can be particularly misleading for profitability since the use of aleverage ratio causes identification problems It is hard to be sure whethera leverage change reflects changes in the numerator or the denominator Thisis particularly problematic when as here the rebalancing is partialIt should be noted that much empirical capital structure literature debates

about whether to use a book leverage ratio or a market leverage ratio Bookleverage is used based on the argument that managers borrow against assetsin place which are better reflected by book assets Others such as Welch(2004) argue that book values are essentially an accounting fiction and thatonly market leverage should count Empirical papers regularly attempt tosidestep this debate by carrying out tests on both book leverage and marketleverage Our point is that both kinds of leverage ratios are potentially mis-leading when examining the impact of profits on capital structure

3 Data

We use conventional data sources starting with the merged Compustat-CRSP data The data are annual and are converted into constant year2000 dollars using the GDP deflator We exclude financial companies (SIC6000-6999) firms involved in major mergers (Compustat footnote code AB)firms reporting financial data in currencies other than the US dollars andfirms with missing data on our key variables7 The ratio variables aretrimmed at the 1 level in both tails of the distribution This serves toremove outliers and the most extremely misrecorded data The finalsample consists of 179021 firm-year observations from 1971 to 2009Table I provides definitions of financial variables and reports summary

statistics The average debt (in constant US$) is about $653 million while themedian is $24 million A significant fraction of firms have zero debt (the 10thpercentile is 0) Book equity is slightly larger than book debt Market equityis almost three times larger than book debt Book assets average $2191million although the medians are considerably smallerIf issuing or retiring securities incurs no fixed costs then we would expect

to see many small actions and very few large actions (Leary and Roberts2005) If there were significant fixed costs involved in issuing or retiringoutstanding securities then small issues might not be worthwhile Table I

7 These include debt book value of equity market value of equity assets book andmarket leverage profitability market-to-book assets ratio and tangibility

1424 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

shows that although most firms issue little debt or equity in a given year theaverages are large In other words when firms actually enter debt and equitymarkets they intervene massivelyThe mean constant dollar debt issue is $165 million (the median is $2

million) In unreported tables we find that the average debt issue is about81 of assets (the median is 18) About 38 of the firms issue no debt8 issue between 0 and 1 of the value of their assets as debt another

Table I Data description

Data sources The sample comes from the annual Compustat files The sample period is

1971ndash2009 We exclude financial companies (SIC 6000-6999) firms involved in majormergers (Compustat footnote code AB) firms reporting financial data in currencies otherthan the US dollar and firms with missing data on the key variables All financial variables

are deflated to year 2000 using the GDP deflator N is the number of observations SD isthe standard deviation D BVE MVE Assets Debt issuance Equity issuance Debt re-payment Equity repayment are all measured in $ millions The ratio variables are trimmed

at the 1 level in both tails of the distribution

Variable definitions Debt (D)frac14 long-term debt (dltt)thorn short-term debt (dlc) book equity(BVE)frac14 common shareholder equity (ceq) market equity (MVE)frac14 no of outstanding

shares (csho) closing price ( prcc f ) assetsfrac14 book assets (at) debt issuancefrac14 issuance oflong-term debt (Maxethdltis 0THORN)thorn increase in current debt (Maxethdlcch 0THORN) equity issu-ancefrac14 sale of common stock (Maxethsstk 0THORN) debt repaymentfrac14 reduction of long-termdebt (Maxethdltr 0THORN)thorn decrease in current debt (Minethdlcch 0THORN) equity repurchasefrac14 purchase

of common stock (Maxethprstkc 0THORN) book leveragefrac14D(DthornBVE) market leveragefrac14D(DthornMVE) profitabilityfrac14EBITDA (oibdp)assets market-to-book ratio M

B frac14Marketvalue of assets (MVA)assets where MVAfrac14DthornMVEthorn preferred-liq value (pstkl)

deferred taxes (txditc) tangibilityfrac14 net property plant and equipment (ppent)assets

Variable N Mean SD

Distribution

10th 50th 90th

Debt (D) 179021 653 4592 0 24 1048

Book equity (BVE) 179021 815 4122 3 70 1373

Market equity (MVE) 179021 1784 9589 9 118 2698

Assets 179021 2191 11530 10 153 3575

Debt issuance 179021 165 1335 0 2 245

Equity issuance 179021 26 186 0 0 40

Debt repayment 179021 132 1123 0 3 181

Equity repurchase 179021 24 257 0 0 9

Book leverage 179021 036 034 000 031 072

Market leverage 179021 027 025 000 020 066

Profitability 179021 005 027 018 011 023MB 179021 165 206 052 102 324

Tangibility 179021 031 024 005 026 070

PROFITSndashLEVERAGE PUZZLEREVISITED 1425

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

16 issue between 1 and 5 of the value of their assets as debt and theremaining 38 issue debt in excess of 5 of the value of their assetsThe mean constant dollar equity issue is about $26 million (the median is

about $04 million) As a fraction of assets the mean and median equityissues are about 7 and 02 respectively About 33 of the firms issue noequity 34 of the firms issue between 0 and 1 of the value of theirassets as equity another 14 issue between 1 and 5 of the value of theirassets as equity and the remaining 19 of the firms issue equity that is inexcess of 5 of the value of their assets Average debt repayments are largerthan equity repurchases This perhaps reflects the finite maturity of debt andits contractual repayment The median firm does not repurchase equityWe construct both book and market leverage ratios Book leverage is

defined as debt over debt plus book equity Market leverage is defined asdebt over debt plus market equity8 The median book leverage is 036 (theaverage is 031) The median market leverage is 027 (the average is 020)Profitability is defined as the ratio of operating income before depreci-

ation to assets While the average firm is profitable (the ratio of EBITDA toassets is 005) the median firm is even more profitable (with a profitabilityratio of 011) But the sample also includes a large number of unprofitablefirms as the 10th percentile is 018 The table also reports descriptive stat-istics on the market-to-book ratio and the tangibility ratio The market-to-book ratio (MB) defined as the ratio of the market value of assets to bookassets averages at about 165 Tangibility defined as the ratio of netproperty plant and equipment to assets averages at about 31

4 Leverage Ratio Regressions

The previous literature focuses on estimates obtained using leverage ratiosHence we start with a similar estimation to check whether our results matchthose of previous studies Table II presents the results for book leverage Weobtain similar results for market leverage and they are presented inAppendix Table AI The regressions include leverage factors following theprevious capital structure literature (see eg Lemmon Roberts and Zender(2008) and Frank and Goyal (2009)) The factors include (i) profitability (ii)industry median leverage (iii) market-to-book assets ratio (iv) tangibility ofassets and (v) firm size Industry median leverage is constructed as themedian leverage of all other firms in the industry excluding the firm under

8 Welch (2011) stresses the idea that nonfinancial liabilities should not be implicitly mis-treated as if they were equity by paying excessively narrow attention to financial liabilitiesin a leverage ratio We have adopted his approach in the empirical work reported here

1426 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

consideration Firm size is defined as the natural log of assets Other factorsare defined in Section 3 We cluster standard errors by firm and estimatethese regressions both with and without fixed effectsIn Table II we run conventional cross-section leverage regressions Since

we employ the usual data it is not surprising that our results match thosereported in the existing capital structure literature Profitability has anegative sign in both the book leverage regressions and in the marketleverage regressions The coefficients on other factors largely match thosereported in earlier studies Firms operating in industries with high leverage

Table II Leverage and profitability

The table presents parameter estimates for book leverage regressions The sample consists

of nonfinancial companies in the annual Compustat files during the period 1971ndash2009 Theleverage is estimated as the ratio of debt over debt plus book equity The explanatoryvariables Profitabilityt1

MB

t1

Tangibilityt1 and LnethAssetsTHORNt1 are described in

Table I IndLevt1 is estimated as the median book leverage of all other firms in thesame industry The industry is defined at the level of the four-digit SIC code All specifi-cations include the year fixed effects The specifications in column (2) additionally include

the firm fixed effects We report t-statistics where the standard errors are clustered at thefirm level in parentheses

aSignificant at the 1 level

Quantile regressions

(1) (2) (3) (4) (5)

Profitabilityt1 026a 020a 011a 017a 041a

(293) (192) (378) (518) (1015)

IndLevt1 042a 012a 030a 052a 052a

(369) (107) (926) (1252) (950)MB

t1

002a 001a 001a 002a 002a

(203) (102) (392) (426) (418)

Tangibilityt1 018a 025a 024a 019a 012a

(206) (154) (964) (580) (276)

LnethAssetsTHORNt1 001a 002a 002a 002a 001a

(114) (76) (899) (478) (96)

Constant 012a 010a 012a 002a 029a

(156) (60) (284) (27) (387)

Year FE Yes Yes Yes Yes Yes

Firm FE No Yes No No No

Regression type OLS FE 25th tile Median 75thtile

Clustered SE Yes Yes No No No

R2 013 054

Pseudo R2 016 015 009

N 158824 158824 158824 158824 158824

PROFITSndashLEVERAGE PUZZLEREVISITED 1427

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

tend to have high leverage A higher market-to-book ratio is associated withlower leverage Larger firms are typically more highly levered The coeffi-cients on tangibility are positive The conclusion from Table II is clear Theleverage regression results match those reported in previous studiesAt times one may be concerned that in a regression what is being

estimated is a conditional mean while some parts of the distributionmight behave differently from other parts If so then ldquoaccidentsrdquo likesample selection issues or asymmetries in the underlying distributionsmight play a greater role than is usually recognized In other words thesource of identification could be misinterpreted To guard against this po-tential problem we estimate the same model using quantile regressions9

Columns 3 4 and 5 report what happens at the 25th 50th and 75th per-centiles The basic model results are robust in terms of both the sign and thestatistical significanceThe impact of profits does appear to be somewhat stronger among the

high leverage firms For instance in a book leverage regression the coeffi-cient on profits for the first quartile is 011 with a t-ratio of 378 For thethird quartile the coefficient on profits is 041 with a t-ratio of 1015This difference is curious and might deserve further study However the factthat in each case we get the negative sign and statistical significance is suf-ficient for the purposes of the present article So the results on profits arerather robust across the distributionTable III sorts firms into those that are actively adjusting their leverage

and those that are passive In the presence of both fixed and variable trans-acts costs this distinction is important We expect the coefficient on profit-ability in leverage ratio regressions to be more negative in firms that arepassive and less negative in firms that are actively adjustingWe define active firms as those that are either issuing net debt or net

equity in excess of 5 of their assets Passive firms are those that neitherissue debt nor equity in nontrivial amounts (ie greater than 5 of theirassets) Consistent with our expectations we find in Table III that the coef-ficients on profits are significantly more negative for passive firms than theyare for active firms With firm and year fixed effects the coefficient onprofits is 008 for active firms and 017 for the passive firms The twocoefficients are significantly different from each other (plt 001) InTable III the standard control variables are used They have the usualsigns and for the most part do not differ much between the active and thepassive firms

9 Cameron and Trivedi (2010) provide an extensive discussion of quantile regressions inStata

1428 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

5 Financing Activity

51 FREQUENCY OF DEBT AND EQUITY ISSUES

Table IV reports the percentage of firms issuing or repurchasing debt orequity for annual sorts based on lagged profitability In Panel A weemploy the conventional 5 cutoff rule to exclude minor fluctuations anddefine firms ldquoissuing debtrdquo as those that issue debt (both short-term andlong-term) in excess of 5 of the value of their assets Other decisions aresimilarly defined using a 5 cutoff Appendix Table AII examines the dis-tributional properties across profitability (and size) quintiles for ldquolargerdquo is-suances by increasing the 5 cutoff to higher levels such as 10 and 15How does security issuance behavior vary with firm profitability To

answer this question Table IV sorts firms by profitability and reports the

Table III Leverage and profitability for active and passive firms

The table presents estimates of leverage on firm characteristics for the subsamples of firms

that are active in either debt or equity markets Active firms are defined as those that issueeither net debt in excess of 5 of the value of assets or issue net equity in excess of 5 ofthe value of their assets The passive firms do not issue significant (more than 5) of either

debt or equity The sample comes from the annual Compustat files during the period 1971ndash2009 Financial firms are excluded The variables are described in Tables I and II Wereport t-statistics where the standard errors are clustered at the firm level in parentheses

aSignificant at the 1 level

Active Passive

(1) (2) (3) (4)

Profitabilityt1 006a 008a 022a 017a

(126) (148) (254) (188)

IndLevt1 044a 019a 041a 013a

(424) (149) (326) (127)MB

t1

002a 001a 003a 001a

(362) (196) (327) (172)

Tangibilityt1 010a 014a 014a 019a

(146) (118) (145) (130)

LnethAssetsTHORNt1 001a 004a 001a 004a

(121) (179) (147) (169)

Constant 012a 008a 003a 007a

(120) (57) (45) (42)

Year FE Yes Yes Yes Yes

Firm FE No Yes No Yes

R2 034 064 033 075

N 43395 43395 58943 58943

PROFITSndashLEVERAGE PUZZLEREVISITED 1429

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

percentage of firms issuing or repurchasing debt or equity The table alsoreports test statistics from tests that examine whether the percentage differ-ences between the high- and low-profit firms are significantly different fromeach other The results in Column (1) show that the likelihood of issuing debt

Table IV Profitability sorts for debt and equity issuers

The sample contains nonfinancial firms listed on the annual Compustat files for the period

from 1971 to 2009 The table presents the percentage of firms issuing and retiring (orrepurchasing) debt and equity The firms are classified as ldquoissuing debtrdquo if they issuedebt in excess of 5 of the value of their assets as ldquoissuing equityrdquo if they issue equity

in excess of 5 of the value of their assets as ldquoretiring debtrdquo if they retire debt in excess of5 of the value of their assets and as ldquorepurchasing equityrdquo if they repurchase equity inexcess of 5 of the value of their assets In addition we also report net debt issuers which

are firms that issue net debt over 5 of the value of their assets and net equity issuerswhich are firms that issue net equity in excess of 5 of the value of their assets Weannually sort firms on lagged profitability and report the percentage of firms in each ofthese categories The bottom part of each panel reports the percentage of firms issuing or

retiring securities by profitability within the smallest and largest asset quintiles The tablealso reports results from t-test for the difference in percentage of firms issuing or retiringsecurities in the lowest and the highest profit quintile

Percentage of firms

Iss Ret Iss Iss Rep Iss Both IssD IssE

D D NetD E E NetE DampE RepE RetD

(1) (2) (3) (4) (5) (6) (7) (8) (9)

Allfirm

s

Low 276 245 186 301 13 297 71 01 38

2 319 299 181 117 21 112 27 02 20

3 385 344 194 86 29 80 21 05 15

4 386 330 213 81 43 75 20 07 14

High 329 273 198 92 82 81 17 13 14

t Highfrac14Low 150 84 39 723 414 763 358 176 200

Smallfirm

s Low 275 204 201 411 11 405 109 01 47

2 291 233 213 379 10 378 99 01 45

3 281 246 204 265 14 261 66 01 41

4 275 272 186 162 15 159 34 02 35

High 275 281 178 136 27 131 25 02 26

t Highfrac14Low 01 84 28 316 55 318 171 18 55

Largefirm

s

Low 414 363 177 87 22 79 28 03 11

2 449 379 180 62 30 57 17 07 09

3 439 360 186 56 41 48 16 09 07

4 403 320 194 46 67 39 13 12 06

High 374 264 207 50 129 35 10 28 05

t Highfrac14Low 50 130 45 89 244 116 77 120 47

1430 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

is mostly independent of firm profitability In the aggregate there is a smalldifference but it is sensitive to sorting on firm size Similarly in Column (2)there is only a weak relation between the profitability and the likelihood thatthe firm retires debt To a first approximation the fraction of firms issuing orrepurchasing debt is largely independent of profits Large firms are somewhatmore likely than small firms to issue debt and also to retire debtUnlike debt issues the probability of issuing equity is strongly related to

profitability as can be seen in Column (4) Low-profit firms are much morelikely to issue equity than are high-profit firms In the lowest profit quintile301 of the firms issued equity In the highest profit quintile it was only92 The lower panel reports separate sorts for small and large firms Similarpatterns are found for both groups of firms that is the likelihood of equityissuance declines as profitability increases The differences between the quin-tiles are quite stark for the group of small firms All quintiles of small firmsissue equity much more frequently than the corresponding large firm quintilesHigh-profit firms are much more likely to repurchase equity as shown in

Column (5) In the overall sample only 13 of low-profit firmsrepurchased equity compared with 82 of the high-profit firms Whenwe separate the small and large firms again a difference emerges Ingeneral small firms do not repurchase much equity while a greaterfraction of the large firms do soGiven these facts it is natural that low profitability firms are much more

likely to be net issuers of equity than are high profitability firms as shown inColumn (6) Among the firms in the lowest quintiles of profits roughly 30of firms issue net equity that exceeds 5 percent of their assets In contrastfor the most profitable quintile only about 8 issue net equityFirm size is an important variable in the recent literaturemdashit is often used

as a proxy for access to capital markets (as in Faulkender and Petersen(2006) and Leary (2009)) Small firms are bank-dependent risky and infor-mationally opaque They have restricted access to public debt markets andconsequently face more severe supply constraints in their ability to issuedebt Thus we expect small firms to be more sluggish in adjusting theirdebt and equity in response to shocks to profitability Large firms incontrast have much easier access to public debt markets and they facefewer obstacles in accessing securities marketsTo examine how size and profitability interact the bottom part of Table IV

examines profitability sorts for the smallest and largest firms We first sortfirms annually by firm size and then within each size quintile we sort them onprofitability We do these two-way sorts to ensure that we have similarnumber of firms in profitability subgroups for both small and large firmsamples Unconditionally small firms are generally less profitable while

PROFITSndashLEVERAGE PUZZLEREVISITED 1431

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

large firms are relatively more profitable The two-way sorts allow us to havea uniform distribution of firms in various profitability quintiles for both smalland large firm samples The bottom part of Table IV reports results for thesmallest and the largest size quintiles Among the small firms there is littlerelation between profitability and the likelihood of issuing debt Howeverlarger firms exhibit a small increase in the likelihood of net issuance of debtwith increasing profitabilityIn contrast to debt the effects of profitability on equity issuance and

repurchases are much more consistent across size quintiles As profitabilityincreases firms are generally less likely to issue equity and more likely torepurchase it Across the two extreme size groups we note a striking differ-ence between the proportion of firms issuing equity and those issuing debtAmong low-profit small firms almost 41 are equity issuers whereas onlyabout 9 of low-profit large firms are equity issuers However regardless ofsize we note a monotonic reduction in the likelihood of issuing equity asprofitability increasesColumn (7) reports the fraction of firms that issue both debt and equity in

excess of 5 of assets both as a function of profits in the upper panel andas a function of firm size and profits in the lower panel In the upper panelwe see that low-profit firms are much more likely to issue both debt andequity (71) while the high-profit firms are less likely to do so (17) Thelower panel shows that small firms are much more likely to be issuers of bothin contrast to large firms In both firm size categories the profit pattern is thesame with low profits being more frequent issuers of both simultaneouslyImportantly Column (8) shows that the likelihood of issuing debt and

simultaneously repurchasing equity increases with profitability Converselyin Column (9) we find that the likelihood of doing the reverse that is issuingequity and retiring debt declines with profitability The effects of sorting onfirm size and firm profitability mirror those reported for all firms Overallthe results show that low-profit firms are less likely to issue debt and repur-chase equity they are instead more likely to issue equity and retire debtFirms with high profitability exhibit the reverse patternIn Appendix Table AII we present results from profitability (and size)

quintiles for large issuances defined using the 10 cutoff (in Panel A) andthe 15 cutoff (in Panel B) The results from these panels are largely con-sistent with those described above

52 MAGNITUDES OF ISSUANCES AND EQUITY CHANGES

In the previous section we considered the probability of having a nontriviallevel of debt or equity activity The next question is how large are the dollar

1432 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

values involved In Table V we sort firms according to profits and thentabulate the levels and changes in both debt and equity We do this firstfor all firms and then for small versus large firmsFor the analysis on all firms we observe in Column (1) that debt peaks at

the middle of the distribution This is because firms with medium profitabil-ity are also the largest as seen in Column (8) which reports average assetvalues for different profitability quintiles When we sort first by firm size andthen examine firms with different levels of profitability within size quintileswe find that for small firms debt is roughly independent of profits while forlarge firms less profitable firms have a higher level of debtAs expected Columns (3) and (5) show that more profitable firms have

higher equity values Columns (2) and (7) consider the relationship betweenissuances of debt and equity and firm profitability Debt issuances are sig-nificantly larger for more profitable firms In contrast equity issuances arelarger for low-profit firms Small low-profit firms issue substantially lessequity compared with small high-profit firms Among large firms the low-profit ones tend to issue equity while the high-profit ones tend to repurchaseequity The fact that more profitable firms issue debt and repurchase equitywhile the least profitable firms retire debt and issue equity is consistent withthe predicted relation between profitability and financing decisions under thetrade-off theoryWe also examine two-way sorts by size and profitability and report results

for the smallest and largest quintile of firms For the smallest quintile offirms the change in debt is largely unrelated to profits But for large firmsthere is a positive relation between profits and debt issuances High-profitfirms have a big positive change in debt Low-profit large firms have anegative change in debtColumns (4) and (6) provide an explanation of why the leverage ratio

regression results contradict those from the basic profitability sorts pre-sented here As we can see profitability indirectly affects leverage ratiosby increasing equity values Changes in both the book value of equity andthe market value of equity are positive and large for highly profitable firmsIn contrast these changes are negative for less profitable firms This evidenceillustrates an important issue concerning the use of leverage ratios Suchratios are often interpreted as essentially reflecting the use of debt by thefirm This interpretation while common is empirically misleadingFor the typical firm the change in the value of equity is larger than the

change in debt For example in the third profit quintile for large firmscolumn (7) the mean equity issue is just $9 million but the change in themarket value of equity is $242 million At the same time the mean change indebt is $27 million This suggests that a fair bit of the observed variation in

PROFITSndashLEVERAGE PUZZLEREVISITED 1433

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

the leverage ratios is primarily driven by the changes in the market value ofequity in the denominator rather than by the changes in debt in the numer-ator Since equity issues are often small this implies that the variation in theleverage ratio is primarily driven by internal operations rather than byexternal financing actions This again points to the fact that leverageratios can provide a misleading account of actual patterns in the dataThere is always a potential concern that the averages may be misleading

due to the impact of outliers To address this concern in AppendixTable AIII we present median values of the profitability sorts Theseresults generally reinforce the findings in Table VTable VI again considers the magnitudes of financing activity but this

time the issues are scaled by a measure of firm size That way it is easier to

Table V Magnitude of financing activity

The sample contains nonfinancial firms listed on the annual Compustat files for the period

from 1971 to 2009 The table presents average debt and equity changes in debt and equityand equity issuances (in millions of dollars) for firms sorted on profitability within sizeclasses The table reports information for all firms sorted on profitability and for profit-

ability sorts within the smallest and largest firms The sorts are done annually The tablealso reports results from t-test for the difference in percentage of firms issuing or retiringsecurities in the lowest and the highest profit quintile

D D BVE BVE MVE MVE EqyIss Assets

(1) (2) (3) (4) (5) (6) (7) (8)

Allfirm

s

Low 83 5 78 10 120 5 7 234

2 605 13 509 24 721 8 17 1691

3 1227 9 1123 22 1834 3 9 3551

4 833 23 1166 7 2408 74 9 2994

High 570 53 1171 71 3817 68 36 2581

t Highfrac14Low 31 9 36 9 38 2 16 36

Smallfirm

s Low 25 04 42 01 332 28 33 97

2 29 06 55 02 278 16 21 116

3 31 07 63 01 212 05 12 127

4 30 06 70 05 179 10 08 136

High 24 06 84 14 234 17 07 142

t Highfrac14Low 04 15 153 85 73 11 201 107

Largefirm

s Low 3929 55 2990 152 4671 11 63 10846

2 3826 19 3178 95 5122 66 11 10721

3 3074 27 3498 41 6173 242 9 9697

4 2623 62 3915 45 8473 94 57 9855

High 2180 193 4642 261 14924 202 148 10171

t Highfrac14Low 94 64 108 87 246 12 169 16

1434 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

see whether large firms behave disproportionately when compared withsmall firms Column (1) shows that low-profit firms have more debtas reported in many previous studies In Columns (4) and (6) we seethat low-profit firms have a major decline in both the book value ofequity and the market value of equity The high-profit firms have thereverse experienceMore interesting are the active decisions of the firms In Column (2) we see

that in contrast to the usual impression the low-profit firms are reducingdebt while the high-profit firms are increasing it In Column (7) we see thatthe low-profit firms are also reducing leverage by issuing equity The high-profit firms issue little equity In the lower panel similar results are foundwhen we separate small and large firms Equity issuing is particularly strong

Table VI Magnitude of financing activitymdashscaled levels and changes

The sample contains nonfinancial firms listed on the annual Compustat files for the period

from 1971 to 2009 The table presents average debt and equity changes in debt and equityand equity issuances scaled by assets for firms sorted on profitability within size classes Thetable reports information for all firms sorted on profitability and for profitability sorts

within the smallest and largest firms The sorts are done annually The table also reportsresults from t-test for the difference in percentage of firms issuing or retiring securities inthe lowest and the highest profit quintile

D D BVE BVE MVE MVE EqyIss

(1) (2) (3) (4) (5) (6) (7)

Allfirm

s

Low 386 122 12 871 3427 1868 217

2 285 58 407 119 940 157 28

3 293 12 425 14 742 03 12

4 269 26 461 02 955 27 10

High 192 15 551 47 1675 74 07

t Highfrac14Low 35 49 50 47 127 42 157

Smallfirm

s Low 545 154 841 2145 9688 9128 696

2 684 49 712 1915 5014 2318 288

3 331 253 141 969 3063 954 129

4 267 181 357 175 1673 221 55

High 194 05 543 26 1758 147 41

t Highfrac14Low 49 17 51 36 88 27 99

Largefirm

s Low 350 30 306 58 511 55 09

2 346 14 335 13 523 02 04

3 353 165 332 71 711 29 01

4 280 01 414 04 909 09 05

High 227 15 477 19 1643 27 14

t Highfrac14Low 371 112 453 181 570 57 221

PROFITSndashLEVERAGE PUZZLEREVISITED 1435

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

among the small low-profit firms that are also experiencing sharp reductionsin the value of equity In column (2) we see that both for the large and for thesmall firms net debt increases are primarily found in the highest quintile offirm profitability

6 Debt and Equity Regressions

So far we have demonstrated that in our data the conventional leverageregressions have the usual signs We then demonstrated that in simple sortsof the data strikingly different active financing patterns emerge The nexttask is to reconcile these differencesTable VII presents simple regressions predicting changes in debt book

equity and market equity as well as equity issuances Consistent with thesorts we control for changes in firm size and we also include firm and yeardummies There may be a concern about the speed of a firmrsquos responsesAccordingly we consider the lagged change in profits as the main case butwe also include the contemporaneous change in profits Whether the contem-poraneous change in profits is included or not does not in fact matter Wereport panel-robust standard errors adjusted for clustering at the firm levelThe results in Table VII support the findings in the sorts Firms that have

an increase in profits in 1 year increase debt both in the same year and in thenext year Firms that have an increase in profits experience an increase inboth the book value of equity and the market value of equity Importantlyhowever firms that experience and increase in profits reduce their equityissues All these effects are statistically significant (except for the marketvalue of equity in the column 5 specification) and seem intuitively reason-able In Appendix Table AIV similar regressions to those presented aboveare reported but without firm-fixed effects We find that empirically verylittle changes whether we include firm fixed effects or leave them out Thusthe targeting behavior to the extent that it happens does not account forwhat we are reportingThe next step is to bring the sorts and the simple regressions together into

a common setting We therefore include factors in addition to profitability(i) median industry leverage (ii) the market-to-book assets ratio (iii) tangi-bility of assets and (iv) firm size (measured by the log of assets) Rajan andZingales (1995) show that these factors are related to leverage in G7countries A number of studies have used these factors to estimateleverage targets Frank and Goyal (2009) show that these factors arerobustly related to leverage in the USA In these regressions we use indica-tors for the quintile that the firm is in for each factor

1436 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

Table VIII reports these results In Columns (1) and (2) we examine debtissues and again show that profits positively affect debt issuances The effectsare large and statistically significant at the 1 level Columns (3) and (4)which examine book equity show a strong effect of profits This effect is onlyslightly reduced by the inclusion of other factors Columns (5) and (6) illus-trate that the impact of profits on the change in market equity is also robustto the inclusion of the conventional factors Columns (7) and (8) examineequity issuances Here we again find that profits have a significant negativeimpact on equity issuances As before the results are robust to the inclusionof the conventional factorsIn unreported tables we estimate debt and equity changes using quantile

regressions In terms of the signs and significance of the profitability variable

Table VII Debt and equity changes

The sample contains nonfinancial firms listed on the annual Compustat files for the period

from 1971 to 2009 The table presents fixed effect estimates from regressions of changes indebt (D) changes in the book value of equity (BVE) changes in the market value ofequity (MVE) and net equity issuances (EqyIss) The explanatory variables include

current and lagged change in operating income before depreciation (Profits andProfitst1) and lagged change in the book value of assets (Assetst1) In addition tofixed firm effects the regressions include year indicator variables The reported t-statistics

are corrected for clustering at the firm level

aMean significant at the 1 level

bMean significant at the 5 level

cMean significant at the 10 level

D BVE MVE EqyIss

(1) (2) (3) (4) (5) (6) (7) (8)

Profitst1 022a 025a 015b 035a 004 060b 005b 007a

(33) (31) (25) (35) (03) (25) (25) (53)

Profits 050b 078a 140a 003a

(24) (32) (48) (34)

Assetst1 000 007b 025a 001b

(00) (21) (35) (25)

Constant 167 283a 34 51 1145a 1182a 605a 595a

(14) (30) (04) (05) (32) (29) (69) (74)

Year FE Yes Yes Yes Yes Yes Yes Yes Yes

Firm FE Yes Yes Yes Yes Yes Yes Yes Yes

R2 001 004 001 010 001 005 002 003

N 162056 162056 162130 162130 157550 157550 162154 162154

PROFITSndashLEVERAGE PUZZLEREVISITED 1437

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

the results are qualitatively similar to the OLS results reported in Table VIIIProfitability positively affects debt issuances and negatively affects equityissuances across various quantiles Overall the results in Table VIII showthat there is nothing that is special about using sorts or regressions to explaindebt and equity issues In either case we find that more profitable firms tendto increase their debt experience an increase in the value of equity andrepurchase shares Thus the control factors are not responsible for theusual rejection of the trade-off theory It would appear that what mattersis the size of the issuing activity relative to the organic increases in equityvalue due to profits In other words what matters in the leverage regressionseems to be coming from the presence of E in D

DthornE and not from D

Table VIII Debt and equity issuances profitability and leverage factor quintiles

The sample contains nonfinancial firms listed on the annual Compustat files for the period

from 1971 to 2009 The table presents estimates from regressions of change in debt (D)change in the book value of equity (BVE) change in the market value of equity (MVE)and net equity issuances (EquityIss) The explanatory variables include lagged profitability

quintiles and lagged leverage factor quintiles In addition to fixed firm effects the regres-sions include year indicator variables The reported t-statistics are corrected for clustering atthe firm level

aMean significant at the 1 level

bMean significant at the 5 level

cMean significant at the 10 level

D BVE MVE EqyIss

(1) (2) (3) (4) (5) (6) (7) (8)

ProfitabilityQint 214a 167a 283a 175a 80 429a 53a 55a

(70) (54) (66) (59) (08) (48) (55) (62)

IndLevQint 104a 43 99 20

(29) (15) (10) (10)

Ln(Assets)Qint52 17 1026a 53a

(16) (05) (77) (35)MB

Qint158a 372a 1155a 09

(45) (56) (106) (06)

TangibilityQint 05 177 436c 64a

(01) (15) (18) (31)

Constant 167 96 656b 235 4613a 8610a 113 570a

(08) (04) (26) (04) (61) (58) (11) (31)

Year FE Yes Yes Yes Yes Yes Yes Yes Yes

Firm FE Yes Yes Yes Yes Yes Yes Yes Yes

R2 0002 0002 0003 0004 0013 0014 0013 0014

N 155587 155037 155587 155037 155587 155037 155587 155037

1438 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

7 Are Financial Market Conditions Important

According to the trade-off theory capital structure is determined by con-sidering the impact of costs and benefits of debt The time variation inthe costs and benefits of debt imply that issuance decisions would alsovary over time Furthermore market conditions affect the transactioncosts of adjusting Adverse selection is a greater problem in a cold marketthan in a hot one Accordingly it may be easier to adjust in good marketconditions Related ideas have been developed by Huang and Ritter(2009) Faulkender et al (2012) and Halling Yu and Zechner (2012)Faulkender Flannery Hankins and Smithrsquos paper examines cross-sectionalvariation in adjustment speeds including market valuations andhow they may affect adjustment speeds It is natural to think that ingood market conditions it will be less costly to issue both debt and equityand adjusting leverage to respond to profitability shocks would be mucheasierTo test the importance of market conditions we require a definition of

good times and bad times Our empirical strategy is to estimate good timesversus bad times at the four-digit industry level We define an industry ashaving ldquogood timesrdquo if the median firm in that industry has a market-to-book ratio that is higher than the 67th percentile of the time-series distribu-tion of the industry median market-to-book ratios Conversely an industryis defined as having ldquobad timesrdquo if the median firm in the industry has amarket-to-book ratio that is lower than the 33rd percentile of the time-seriesdistribution of the industry median market-to-book ratiosPanel A of Table IX tabulates issuance activity for profitability sorts in

both good times and bad times As might be expected active debt and equityissuances are larger during good times There is also somewhat more activeswaps between debt and equity during good timesPanel B reports the issuance activity for the smallest and the largest firms

during the good times and bad times Within each size quintile firms aresorted on profitability This is a three-way sort As before we find that thereis much more active use of external markets during good times Small low-profit firms are more likely to issue equity in good times than in bad timesDebt issuances are significantly higher in good times Large high-profitfirms are significantly more likely to issue debt and repurchase equity ingood times than in bad timesPanel C reports the magnitudes of the financial variables rather than the

frequencies During bad times less profitable large firms retire substantialamounts of debt and they show a tendency to issue equity Small firms donot seem to engage in similar debt reduction activities However like the

PROFITSndashLEVERAGE PUZZLEREVISITED 1439

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

Table IX Debt and equity issues in good and bad times

The table reports the frequency and magnitude of financing activity for sorts on profits for subsamples

of firms in good and bad times An industry is defined as having ldquogood timesrdquo if the median firm in

that industry has a market-to-book ratio that is higher than the 67th percentile of the time-series

distribution of industry median market-to-book ratios Conversely an industry is defined as having

bad times if the median firm in that industry has a market-to-book ratio that is lower than the 33rd

percentile of the time-series distribution of industry median market-to-book ratios Panels A and B

report the percentage of firms (i) issuing net debt in excess of 5 of the value of their assets (ii) issuing

net equity in excess of 5 of the value of their assets (iii) issuing debt and repurchasing equity both in

excess of 5 of the value of their assets and (iv) issuing equity and retiring debt both in excess of 5

of the value of their assets Panel C tabulates the mean debt and equity levels and the mean changes in

debt and equity issuances The sample contains nonfinancial firms listed on the annual Compustat files

for the period from 1971 to 2009

Iss Iss Iss D Iss ENet D Net E Rep E Ret D(1) (2) (3) (4)

Panel A sorts on profitability

Badtimes

Low 145 131 01 202 170 49 02 083 173 46 04 074 182 47 03 10

High 177 51 06 10t Highfrac14Low 41 136 33 01

Goodtimes

Low 206 381 01 462 197 148 02 233 212 103 04 184 234 92 10 18

High 216 92 16 14t Highfrac14Low 23 658 138 23

Panel B sorts on firm size and profitability

Badtimes

Smallfirm

s Low 151 318 00 312 182 276 02 253 151 140 02 294 183 73 00 19

High 177 80 00 11t Highfrac14Low 10 90 NA 20

Largefirm

s Low 172 54 01 042 148 39 04 053 150 43 04 044 156 33 03 06

High 147 28 05 03t Highfrac14Low 17 32 20 02

Goodtimes Smallfirm

s Low 213 448 01 542 230 443 01 473 222 325 00 474 203 197 02 38

High 190 153 02 31t Highfrac14Low 19 230 10 39

Largefirm

s Low 203 96 03 122 208 66 09 133 232 54 16 064 215 39 19 05

High 234 36 38 04t Highfrac14Low 29 99 92 38

(continued)

1440 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

larger firms there is a tendency for small firms to issue more equitymdashat leastto the extent possibleDuring good times large profitable firms raise significant amounts of debt

as they experience an increase in the value of their equity Such firms alsoengage in share repurchases Debt issues by smaller firms are much lessaffected by their own profitability during good times and their change indebt is more or less independent of their profits However equity issuanceis primarily found among the low-profit small firms during good timesTable IX shows that empirically profitability has a time-varying impacton leverage choicesIn Table X the basic regressions for changes in debt and for equity

issuance are presented for both good times and bad times Comparinggood times with bad times we see that the effects of profits are muchstronger in good times In bad times the impacts are rather weak statisticallyThis difference is interesting and deserves further study However for ourpurposes recall that the static trade-off theory implies time-varying capital

Table IX Continued

D D BVE BVE MVE MVE EqyIss Assets

(1) (2) (3) (4) (5) (6) (7) (8)

Panel C magnitude of financing activity

Badtimes Smallfirm

s Low 40 03 58 14 305 06 22 140

2 44 05 82 07 240 67 17 180

3 45 04 94 02 176 18 05 187

4 43 09 98 02 137 04 03 188

High 36 10 114 13 209 16 04 199

Largefirm

s Low 3365 73 2523 169 2304 163 94 8972

2 3031 95 2983 35 3284 278 56 9068

3 2930 5 3360 14 3947 308 57 9225

4 2281 6 3817 19 5213 442 7 8902

High 1790 62 4767 13 8799 58 31 9573

Goodtimes

Smallfirm

s Low 24 06 38 05 339 23 39 90

2 26 08 48 01 301 04 25 104

3 30 09 60 04 246 03 17 125

4 28 07 66 08 206 18 12 130

High 23 06 85 16 261 13 08 141

Largefirm

s Low 4532 36 3477 103 6557 260 53 12465

2 4580 58 3527 139 6911 421 23 12522

3 2952 111 3485 2 7586 279 32 9535

4 2752 124 3939 72 10686 33 105 10226

High 2429 228 4819 377 18565 268 222 10830

PROFITSndashLEVERAGE PUZZLEREVISITED 1441

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

structure choices even when the target is time invariant Empirically we doobserve time-varying choices

8 Is Scaling Important

In the preceding section we provided results for unscaled debt and equityissues This seems appropriate to us However it involves making twochanges to the standard regression We would like to determine if both

Table X Debt and equity changes in good and bad timesmdashregressions

The sample contains nonfinancial firms listed on the annual Compustat files for the period

from 1971 to 2009 The table presents estimates from regressions of change in debt (D)and net equity issuances (EqyIss) in good and bad times The explanatory variables includechange in profits lagged change in profits and lagged change in assets The regressions

include both firm and year fixed effects An industry is defined as having ldquogood timesrdquo ifthe median firm in that industry has a market-to-book ratio that is higher than the 67thpercentile of the time-series distribution of industry median market-to-book ratios

Conversely an industry is defined as having bad times if the median firm in thatindustry has a market-to-book ratio that is lower than the 33rd percentile of the time-series distribution of industry median market-to-book ratios The reported t-statistics arecorrected for clustering at the firm level

aMean significant at the 1 level

bMean significant at the 5 level

cMean significant at the 10 level

Good times Bad times

D EqyIss D EqyIss

(1) (2) (3) (4) (5) (6) (7) (8)

Profits 087a 091a 006b 009a 060 065 003 003b

(30) (29) (20) (28) (15) (16) (16) (20)

Profitst1 029c 015a 034c 001

(19) (37) (17) (04)

Assetst1 000 002b 000 001

(02) (22) (00) (07)

Constant 1337 3989 8691a 10001a 1351 1851 4099a 4055a

(03) (09) (55) (37) (09) (13) (80) (78)

Year FE Yes Yes Yes Yes Yes Yes Yes Yes

Firm FE Yes Yes Yes Yes Yes Yes Yes Yes

R2 0071 0080 0018 0045 0178 0212 0034 0041

N 74827 68729 74891 68787 26965 26013 26971 26018

1442 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

changes are crucial Accordingly in this section we take a step back toconsider prior literature in which variables are scaled The questions arehow does scaling affect the results and whyTable XI examines alternative scalings Columns (1) and (2) scale debt and

equity issuances by the total firm net issuance Columns (3) and (4) scaledebt and equity issuances by the sum of debt and book equity Columns (5)and (6) consider the change in debt scaled by the sum of debt and bookequity Columns (7) and (8) consider the change in leverage due to externalfinancing Since we are examining changes we use the changes in thestandard regression factors as controls The primary interest is in the coef-ficient on the change in profitsWe find that a change in profits is associated with an increase in debt

issues But the sign reverses when we scale the debt issues by the sum of debtplus book equity This is telling What is really driving the results is thechange in equity in the denominator not the change in debt in the numer-ator This result is further substantiated in Columns (5) and (6) where thechange in debt to capital ratio is used instead of debt issuanceTo further explore the impact of active issuance relative to passive

experiencing of changes we decompose leverage by separating out theeffect of debt and equity issues from those due to changes in retainedearnings Thus we define leverage change due to external financing as

LevEF frac14Dt1 thorn d

Dt1 thorn Et1 thorn dthorn e

Dt1

Dt1 thorn Et1 eth1THORN

where D is debt E is book value of equity d is debt issues net of retirementsand e is equity issues net of repurchases Garvey and Hanka (1999) use asimilar measure in their study This measure directly captures the netleverage effect of debt and equity issuancesColumns (7) and (8) report results from regressions of the net leverage

effect of debt and equity issues on changes in profitability after controllingfor other leverage factors in differences and year indicator variables Theeffect of profitability on leverage change due to external financing is positiveand statistically significant This comparison suggests that changes in totaldebt and equity are only partly a result of debt and equity issuance decisionsOther balance sheet adjustments complicate the inferences from leverageratio regressionsFor most of our analysis we follow the literature in treating debt as homo-

genous It is of course potentially interesting to ask what happens when wedistinguish between short-term debt and long-term debt Which part of thefirmrsquos debt maturity structure drives the under-adjustment Table XII

PROFITSndashLEVERAGE PUZZLEREVISITED 1443

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

provides results from such decompositions Alternative scalings areconsidered The main results of interest are that the distinction betweenshort- and long-term debt is important Short-term debt issuance is largelyindependent of changes in profits The negative sign on profit change stemsfrom the action of long-term debt This seems intuitively reasonable

Table XI Scaled issuances

The sample contains nonfinancial firms listed on the annual Compustat files for the period

from 1971 to 2009 The table presents estimates from regressions of debt issuance to totalissuance (in columns (1) and (2)) debt issuance to total capitalization (in columns (3) and(4)) change in debt to capitalization ratio (in columns (5) and (6)) and change in leverage

ratio considering only external financing (in columns (7) and (8)) The change in leverageratio considering only external financing is defined as

LevEF frac14Dt1 thorn d

Dt1 thorn Et1 thorn dthorn e

Dt1

Dt1 thorn Et1

where D is debt E is book value of equity d is debt issues net of retirements and e is equity

issues net of repurchases The regressions include lagged leverage factors and year indicatorvariables Some of the specifications include firm fixed effects The reported t-statistics arecorrected for clustering at the firm level

aMean significant at the 1 level

bMean significant at the 5 level

cMean significant at the 10 level

DebtIssTotIss

DebtIssethDthornETHORN D

DthornE LevEF

(1) (2) (3) (4) (5) (6) (7) (8)

Profitt1 004a 000 002a 001c 003a 002 001b 002a

(35) (03) (34) (18) (35) (14) (23) (26)

IndLevt1 004c 001 001 001c 000 000 000 000

(19) (06) (15) (19) (02) (04) (03) (05)

Market=bookt1 001a 001a 000a 000a 000a 000a 000 000c

(47) (85) (68) (60) (55) (36) (05) (18)

Tangt1 004 005 013a 011a 011a 006a 003a 003a

(12) (15) (144) (112) (79) (38) (44) (42)

Assetst1 008a 004a 006a 004a 002a 001a 003a 002a

(97) (45) (283) (161) (53) (35) (147) (101)

Constant 079a 074a 002a 003a 000 000 000b 002a

(497) (442) (91) (111) (00) (05) (22) (93)

Year FE Yes Yes Yes Yes Yes Yes Yes Yes

Firm FE No Yes No Yes No Yes No Yes

R2 0015 0098 0022 0076 0009 0070 0012 0086

N 133905 133905 140552 140552 138229 138229 140572 140572

1444 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

inasmuch as short-term debt is probably more associated with day-to-dayoperations rather than longer term planning considerationsIn Table XIII we examine change in leverage ( D

DthornE) and change inleverage due to external financing (LEVEF) during good and bad timesseparately The traditional negative effect of a change in profits on achange in leverage is found both in good times and in bad timesHowever when we decompose the impact in order to isolate the externalcomponent as before the result disappears In good times the externalfinancing component of a change in profits is weak but still positiveDuring bad times it is not statistically significant This implies that distin-guishing between good and bad times does not fundamentally alter the basic

Table XII Short- and long-term debt issuances

The sample contains nonfinancial firms listed on the annual Compustat files for the period

from 1971 to 2009 The table presents estimates from regressions of short-term debtissuance to total issuance (in columns (1) and (2)) long-term debt issuance to totalissuance (in columns (3) and (4)) short-term debt issuance to total capitalization (in

columns (5) and (6)) and long-term debt to total capitalization (in columns (7) and (8))Some of the specifications include firm fixed effects The reported t-statistics are correctedfor clustering at the firm level

aMean significant at the 1 level

bMean significant at the 5 level

cMean significant at the 10 level

STDebtIssTotIss

LTDebtIssTotIss

STDebtIssethDthornETHORN

LTDebtIssethDthornETHORN

(1) (2) (3) (4) (5) (6) (7) (8)

Profitabilityt1 0002 0003 0042a 0008 0001 0001 0019a 0012a

(02) (04) (30) (05) (14) (10) (45) (28)

IndLevt1 0009 0007 0070b 0033 0001 0002c 0014a 0005

(06) (05) (26) (11) (08) (17) (26) (09)

M=Bt1 0003a 0003a 0003b 0010a 0000 0000 0003a 0002a

(48) (35) (22) (65) (10) (01) (84) (42)

Tangibilityt1 0020 0007 0011 0067 0005a 0006a 0097a 0081a

(12) (03) (03) (16) (34) (33) (108) (89)

Assetst1 0039a 0011b 0034a 0024b 0001a 0001 0046a 0020a

(89) (22) (38) (23) (36) (13) (190) (82)

Constant 0001b 0153a 0798a 0591a 0000c 0006a 0057a 0096a

(25) (144) (470) (288) (17) (124) (292) (234)

Year FE Yes Yes Yes Yes Yes Yes Yes Yes

Firm FE No Yes No Yes No Yes No Yes

R2 0025 0089 0018 0073 0030 0147 0016 0298

N 136146 136146 136140 136140 143031 143031 142847 142847

PROFITSndashLEVERAGE PUZZLEREVISITED 1445

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

Table XIII Externally financed leverage changes in good and bad times

The sample contains nonfinancial firms listed on the annual Compustat files for the period

from 1971 to 2009 The table presents estimates from regressions of change in debt tocapitalization ratio (in columns (1) (2) (5) and (6)) and change in leverage ratio consider-ing only external financing (in columns (3) (4) (7) and (8)) The change in leverage ratio

considering only external financing is defined as

LevEF frac14Dt1 thorn d

Dt1 thorn Et1 thorn dthorn e

Dt1

Dt1 thorn Et1

where D is debt E is book value of equity d is debt issues net of retirements and e is equityissues net of repurchases The regressions include lagged leverage factors and year indicator

variables The reported t-statistics are corrected for clustering at the firm level An industryis defined as having ldquogood timesrdquo if the median firm in that industry has a market-to-bookratio that is higher than the 67th percentile of the time-series distribution of industry

median market-to-book ratios Conversely an industry is defined as having ldquobad timesrdquoif the median firm in that industry has a market-to-book ratio that is lower than the 33rdpercentile of the time-series distribution of industry median market-to-book ratios

aMean significant at the 1 level

bMean significant at the 5 level

cMean significant at the 10 level

Good times Bad times

DDthornE LevEF D

DthornE LevEF

(1) (2) (3) (4) (5) (6) (7) (8)

Profitabiltyt1 0022c 0005 0016c 0025b 0073b 0041 0007 0007

(17) (03) (18) (24) (25) (11) (04) (03)

IndLevt1 0000 0011 0001 0009 0010 0018 0001 0000

(00) (09) (01) (13) (07) (10) (02) (00)

MB

t1

0004a 0003b 0000 0001 0008a 0001 0000 0000

(40) (24) (05) (08) (28) (04) (03) (03)

Tangt1 0097a 0041 0033a 0038b 0127a 0048 0059a 0033

(45) (16) (27) (25) (49) (14) (43) (16)

Assetst1 0014a 0013b 0024a 0019a 0037a 0023c 0036a 0025a

(30) (20) (84) (48) (45) (17) (88) (43)

Constant 0030b 0036 0000 0003 0028c 0014 0021 0001

(20) (16) (00) (07) (17) (06) (11) (01)

Year FE Yes Yes Yes Yes Yes Yes Yes Yes

Firm FE No Yes No Yes No Yes No Yes

R2 0006 0088 0009 0117 0026 0251 0023 0202

N 56820 56820 58200 58200 23462 23462 23511 23511

1446 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

conclusions The traditional leverage result on profits is driven by increasesin equity that are experienced and then partially but not completely offsetby firmsTo the extent that firms are engaging in active financing they are following

the prescriptions of the trade-off theory Thus we have a coherent picture ofhow firms are responding to profits Basically more profitable firms experiencean increase in the value of their equity They respond to this by issuing debtand repurchasing equity But the magnitude of the increase in equity is largerthan the magnitudes of the active steps taken in the debt and equity marketsThis means that the firmsrsquo actual actions are as predicted by the trade-offtheory But it leaves open the question of why they do not act more stronglyThe capital structure literature has long been interested in dynamics and

capital structure adjustment paths In recent years this interest has increas-ingly resulted in explicit dynamic optimizing models such as those ofGoldstein Ju and Leland (2001) Hennessy and Whited (2005) andStrebulaev (2007) These models commonly make specialized assumptionsleading some to question their generalityWe argue that both fixed transactions costs and variable transactions costs

are plausible and realistic in models that pin down firm size The simultaneouspresence of both is important Due to fixed costs firms only rebalance inter-mittently Due to variable adjustment costs the extent of the adjustment isreduced to the point at which an extra unit to adjustment is balanced againstthe extra unit transactions costs This means that both leverage increases andleverage reductions will only go part way toward a static optimumThe key point is that with both fixed and variable transactions costs we

should expect to only see intermittent leverage adjustments When leverage isadjusted it should ldquounder adjustrdquo when compared to a frictionless modelFurther discussion of transactions costs in a capital structure setting isprovided by Leary and Roberts (2005)

9 Conclusion

The connection between corporate profits and capital structure has beenvery influential in the assessment of the trade-off theory The conventionalinterpretation of the evidence has pushed the literature away from the trade-off theory and toward much more complex dynamic models and ideas As aresult it is important to make sure that the evidence has been correctlyinterpreted We show that the problem is with the interpretation of thetest results and a failure to take into account the costs associated with re-sponding to a profit shock

PROFITSndashLEVERAGE PUZZLEREVISITED 1447

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

We document that more profitable firms really do borrow more (not less)They tend to repurchase equity They experience an increase in both thebook value of equity and the market value of equity Less profitable firmsreally do tend to reduce their debt and to issue equityFirm size and market conditions also matter Larger firms tend to be more

active in the debt markets while smaller firms tend to be relatively moreactive in the equity markets During good times there is more use ofexternal financingThe usual profitsndashleverage puzzle result is primarily driven by the increase

in equity that is experienced by the more profitable firms Accordingly thepuzzle should be restated as asking why do firms not take sufficiently largeoffsetting actions to fully undo the change in equity What limits the mag-nitudes of the typical leverage response to profit shocksIn a frictionless model the partial response appears puzzling But there is

good empirical reason to believe that rebalancing entails both fixed andvariable costs and that firm size matters The rebalancing costs can befully avoided by doing nothing Accordingly the firm must decide whetherany given shock is big enough to be worth responding to If it is then thefirm must decide how big a response is called for These technical conditionsare known as ldquovalue matchingrdquo and ldquosmooth pastingrdquo Thus optimizationimplies that some shocks will be ignored Even if the shock is not ignoredthe optimal response will only partially undo the shock The magnitude ofthe leverage response must balance the marginal cost and the marginalbenefit of an extra unit of leverage Since the marginal cost of adjustingleverage is strictly positive the adjustment toward that static leverageoptimum will only go part way This is true both for leverage increasesand for leverage reductionsOne might conjecture that the variable costs would be time varying and

that such costs might have declined over time Since we do not have a fullaccounting of all of the variable costs it is of course impossible to be sureHowever there is some evidence suggesting that at least the directunderwriting fees are remarkably stable over time10

Surely nobody ought to be surprised that more profitable firms are moreinclined to repurchase equity while unprofitable firms tend not to do so Thefact that more profitable firms typically experience an increase in the value of

10 For IPOs Jay Ritter provides important summary evidence as httpbearwarringtonufleduritterIPOs2013Underwritingpdf He finds that for medium-sized deals a remark-

ably large fraction of the deals have a 7 fee year after year Whether this is due tobehavioral factors or some sort of a coordination device is not clear The extent towhich similar intertemporal stability carries over to other forms of external financemerits further study

1448 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

equity is equally natural Were it not for the contrary claims in the literature(ie the profitsndashleverage puzzle) we would argue that nobody ought to findit surprising that more profitable firms are more inclined to issue debt Thefact that leverage adjustments typically only partially offset profit shocksshould also sound quite natural These facts all fit together easily within thetrade-off theory once we take account of both fixed and variable costs ofactively adjusting capital structure

Appendix

Table AI Market leverage regressions

The table presents estimates of the regressions of market leverage on firm characteristics

The sample comes from the annual Compustat files during the period 1971ndash2009 Financialfirms are excluded The leverage is estimated as the ratio of debt over debt plus marketequity The explanatory variables Profitabilityt1

MB

t1

Tangibilityt1 and LnethAssetsTHORNt1are described in Table 1 IndMedianLevt1 is estimated as the median market leverage of allother firms in the same industry The industry is defined at the level of the four-digit SICcode All specifications include the year fixed effects The specifications in column (2)

additionally include the firm fixed effects We report t-statistics where the standard errorsare clustered at the firm level in parentheses

aSignificant at the 1 level

Quantile regressions

(1) (2) (3) (4) (5)

Profitabilityt1 013a 013a 007a 011a 020a

(275) (262) (349) (425) (538)

IndLevt1 046a 020a 032a 058a 063a

(506) (239) (1299) (1630) (1042)MB

t1

003a 001a 001a 002a 004a

(509) (303) (743) (709) (534)

Tangibilityt1 013a 018a 013a 014a 010a

(191) (181) (762) (571) (232)

LnethAssetsTHORNt1 001a 005a 001a 001a lt001a

(112) (258) (682) (342) (29)

Constant 010a 011a 005a 005a 029a

(171) (98) (173) (118) (398)

Year FE Yes Yes Yes Yes Yes

Firm FE No Yes No No No

Regression type OLS FE 25th tile Median 75thtile

Clustered SE Yes Yes No No No

R2 030 067

Pseudo R2 016 023 0201

N 158578 158578 158578 158578 158578

PROFITSndashLEVERAGE PUZZLEREVISITED 1449

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

Table AII Profitability sorts for firms issuing large amounts

Percentage of firms

Iss Ret Iss Iss Rep Iss Iss IssD IssE

D D NetD E E NetE DampE RepE RetD

(1) (2) (3) (4) (5) (6) (7) (8) (9)

Panel A 10 cutoff to define issuing or retiring activity

Allfirm

s

Low 205 150 134 260 05 257 45 00 23

2 207 175 103 77 06 75 11 01 10

3 243 196 101 50 08 47 08 01 09

4 252 186 113 50 13 48 08 02 08

High 227 161 118 60 27 57 07 04 09

t Highfrac14Low 70 38 62 755 226 768 320 90 152

Smallfirm

s Low 230 132 162 380 05 375 82 00 32

2 228 143 165 334 04 333 65 00 26

3 211 146 147 216 03 214 38 00 23

4 190 154 120 126 04 124 15 01 20

High 194 162 114 99 09 97 12 01 15

t Highfrac14Low 42 40 68 342 20 340 169 07 131

Largefirm

s Low 231 200 82 31 06 29 08 01 04

2 231 187 67 17 06 14 03 01 02

3 234 177 77 17 09 14 03 02 03

4 218 158 75 16 17 14 02 03 02

High 218 135 99 19 41 15 03 07 02

t Highfrac14Low 19 106 36 46 135 57 39 58 23

Percentage of firms

Iss Ret Iss Iss Rep Iss Both Iss D Iss E

D D Net D E E Net E DampE Rep E Ret D

(1) (2) (3) (4) (5) (6) (7) (8) (9)

Panel B 15 cutoff to define issuing or retiring activity

Allfirm

s

Low 158 101 100 230 01 228 27 00 15

2 142 115 63 57 01 56 05 00 06

3 165 128 57 32 01 31 03 00 05

4 173 118 65 35 01 34 03 00 05

High 160 102 70 46 04 44 03 00 05

t Highfrac14Low 08 06 141 738 87 746 277 24 131

Smallfirm

s Low 195 95 132 354 01 350 56 00 24

2 185 96 131 303 01 301 44 00 18

3 160 96 104 179 00 176 18 00 15

4 134 99 78 99 01 98 08 00 12

High 133 95 73 82 01 80 06 00 10

t Highfrac14Low 80 00 95 345 12 344 149 12 52

Largefirm

s Low 145 123 42 14 01 12 03 00 01

2 137 114 32 07 01 06 01 00 01

3 142 104 35 06 01 05 00 00 01

4 130 91 36 08 03 07 01 01 01

High 134 81 48 09 06 07 01 01 01

t Highfrac14Low 19 85 17 26 54 33 34 17 11

1450 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

Table AIV Debt and equity changesmdashwithout fixed effects

The sample contains nonfinancial firms listed on the annual Compustat files for the period from 1971

to 2009 The table presents estimates from regressions of change in debt (D) change in the book

value of equity (BVE) change in the market value of equity (MVE) and net equity issuances

(EqyIss) The explanatory variables include current and lagged change in operating income before

depreciation (Profits and Profitst1) and lagged change in book value of assets (Assetst1) The

regressions include year indicator variables The reported t-statistics are corrected for clustering at the

firm level

aMean significant at the 1 level

bMean significant at the 5 level

cMean significant at the 10 level

D BVE MVE EqyIss

(1) (2) (3) (4) (5) (6) (7) (8)

Profitst1 025a 024a 023a 038a 007 066a 007a 009a

(40) (32) (36) (39) (06) (29) (42) (51)Profits 051b 082a 149a 004a

(25) (33) (49) (33)Assetst1 002 005c 024a 001b

(08) (17) (34) (23)Constant 93 176a 45 71 1552a 1704a 130a 140a

(12) (28) (07) (09) (49) (49) (45) (45)Year FE Yes Yes Yes Yes Yes Yes Yes YesR2 001 005 001 011 001 005 002 002N 162056 162056 162130 162130 157550 157550 162154 162154

Table AIII Median profitability sorts

The sample contains nonfinancial firms listed on the annual Compustat files for the period from 1971

to 2009 The table presents median debt and equity median changes in debt and equity and median

equity issuances (EqyIss) in millions of dollars for all firms sorted on profitability and for profitability

sorts within the smallest and largest firms The sorts are done annually

D D BVE BVE MVE MVE EqyIss Assets(1) (2) (3) (4) (5) (6) (7) (8)

Allfirm

s

Low 21 00 75 17 256 20 01 1922 186 02 449 18 598 19 00 11033 816 06 1274 03 1637 05 00 32244 787 02 1541 18 2488 03 00 3556High 273 00 1361 66 3300 14 00 2578

Smallfirm

s Low 05 00 13 07 140 14 06 402 06 00 25 08 114 12 01 563 09 00 32 05 85 07 00 674 12 00 39 01 68 03 00 79High 10 00 51 04 98 01 00 90

Largefirm

s Low 11485 381 10269 371 13227 202 03 333902 13051 393 12786 140 17571 169 05 392463 11554 287 13429 24 21623 305 00 370314 9963 209 15052 194 29424 308 00 36705High 7361 00 16380 613 49142 439 00 34976

PROFITSndashLEVERAGE PUZZLEREVISITED 1451

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

References

Auerbach A J (1985) Real determinants of corporate leverage in B M Friedman (ed)

Corporate Capital Structures in the United States University of Chicago Press Chicago

IllinoisBertrand M and Mullainathan S (2003) Enjoying the quiet life corporate governance and

managerial preferences Journal of Political Economy 111 1043ndash1075Booth L Varouj A Demirguc-Kunt A and Maksimovic V (2001) Capital structures in

developing countries Journal of Finance 56 87ndash130Brennan M J and Schwartz E S (1984) Optimal financial policy and firm valuation

Journal of Finance 39 593ndash607Cameron A C and Trivedi P K (2010) Microeconometrics using Stata Stata Press TexasConstantinides G M (1979) Multiperiod consumption and investment behavior with

convex transactions costs Management Science 25 1127ndash1137Constantinides G M and Richard S F (1978) Existence of optimal simple policies for

discounted-cost inventory and cash management in continuous time Operations Research

26 620ndash636DeAngelo H DeAngelo L and Whited T M (2011) Capital structure dynamics and

transitory debt Journal of Financial Economics 99 235ndash261Dixit A K (1993) The Art of Smooth Pasting Routledge London and New YorkFama E and French K R (2002) Testing trade-off and pecking order predictions about

dividends and debt Review of Financial Studies 15 1ndash33Faulkender M and Petersen M A (2006) Does the source of capital affect capital struc-

ture Review of Financial Studies 19 45ndash79Faulkender M Flannery M J Hankins K W and Smith J M (2012) Cash Flows and

Leverage Adjustments Journal of Financial Economics 103 632ndash646Fischer E Heinkel R and Zechner J (1989) Dynamic capital structure choice theory

and tests Journal of Finance 44 19ndash40Frank M Z and Goyal V K (2008) Trade-off and pecking order theories of

debt in B E Eckbo (ed) Handbook of Corporate Finance Empirical Corporate Finance

Vol 2 Handbook of Finance Series Chapter 12 ElsevierNorth-Holland Amsterdam pp

135ndash202

Frank M Z and Goyal V K (2009) Capital structure decisions which factors are reliably

important Financial Management 38 1ndash37

Garvey G T and Hanka G (1999) Capital structure and corporate control the effect of

antitakeover statutes on firm leverage Journal of Finance 54 519ndash546

Goldstein R Ju N and Leland H (2001) An EBIT-based model of dynamic capital

structure Journal of Business 74 483ndash512

Graham J R and Tucker A L (2006) Tax shelters and corporate debt policy Journal of

Financial Economics 81 563ndash594

Hackbarth D and Mauer D C (2012) Optimal priority structure capital structure and

investment Review of Financial Studies 25 747ndash796

Halling M Yu J and Zechner J (2012) Leverage dynamics over the business cycle

Working paper University of Utah University of New South Wales and Vienna

University of Economics and Business AdministrationHennessy C A and Whited T A (2005) Debt dynamics Journal of Finance 60 1129ndash1165

Huang R and Ritter J R (2009) Testing theories of capital structure and estimating the

speed of adjustment Journal of Financial and Qualitative Analysis 44 237ndash271

1452 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

Hovakimian A Opler T and Titman S (2001) The debt-equity choice Journal of

Financial and Quantitative Analysis 36 1ndash24Jung K Kim Y and Stulz R M (1996) Timing investment opportunities managerial

discretion and the security issue decisions Journal of Financial Economics 42 159ndash185Kane A Marcus A J and McDonald R L (1984) How big is the tax advantage to

debt Journal of Finance 39 841ndash853Kayhan A and Titman S (2007) Firmsrsquo histories and their capital structures Journal of

Financial Economics 83 1ndash32Leary M T (2009) Bank loan supply lender choice and corporate capital structure

Journal of Finance 64 1143ndash1185Leary M T and Roberts M R (2005) Do firms rebalance their capital structures

Journal of Finance 60 2575ndash2619Leland H E (1994) Corporate debt value bond covenants and optimal capital structure

Journal of finance 49 1213ndash1252Lemmon M L Roberts M R and Zender J F (2008) Back to the beginning persist-

ence and the cross-section of corporate capital structure Journal of Finance 63 1575ndash1608Long M S and Malitz I B (1985) The investment-financing nexus some empirical

evidence Midland Corporate Finance Journal 3 53ndash59Mackie-Mason J K (1990) Do taxes affect corporate financing decisions Journal of

Finance 45 1471ndash1493Morellec E Nikolov B and Schurhoff N (2012) Corporate governance and capital

structure dynamics Journal of Finance 67 803ndash848Myers S C (1984) The capital structure puzzle Journal of Finance 39 575ndash592Myers S C (1993) Still searching for optimal capital structure Journal of Applied

Corporate Finance 61 4ndash14Rajan R and Zingales L (1995) What do we know about capital structure some evidence

from international data Journal of Finance 50 1421ndash1460Strebulaev I A (2007) Do tests of capital structure theory mean what they say Journal of

Finance 62 1747ndash1787

Strebulaev I A and Whited T M (2012) Dynamic Models and Structural Estimation in

Corporate Finance Now Publisher Boston and Delft

Titman S and Wessels R (1988) The determinants of capital structure choice Journal of

Finance 43 1ndash21

Welch I (2004) Capital structure and stock returns Journal of Political Economy 112

106ndash131

Welch I (2011) Two common problems in capital structure research the financial-debt-

to-asset ratio and issuing activity versus leverage changes International Review of Finance

11 1ndash17

PROFITSndashLEVERAGE PUZZLEREVISITED 1453

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

an issue of concern His model does not account for the negative relationshipbetween current leverage and past market-to-book ratioWe depart from the earlier literature by arguing that the use of leverage

ratios can be particularly misleading for profitability since the use of aleverage ratio causes identification problems It is hard to be sure whethera leverage change reflects changes in the numerator or the denominator Thisis particularly problematic when as here the rebalancing is partialIt should be noted that much empirical capital structure literature debates

about whether to use a book leverage ratio or a market leverage ratio Bookleverage is used based on the argument that managers borrow against assetsin place which are better reflected by book assets Others such as Welch(2004) argue that book values are essentially an accounting fiction and thatonly market leverage should count Empirical papers regularly attempt tosidestep this debate by carrying out tests on both book leverage and marketleverage Our point is that both kinds of leverage ratios are potentially mis-leading when examining the impact of profits on capital structure

3 Data

We use conventional data sources starting with the merged Compustat-CRSP data The data are annual and are converted into constant year2000 dollars using the GDP deflator We exclude financial companies (SIC6000-6999) firms involved in major mergers (Compustat footnote code AB)firms reporting financial data in currencies other than the US dollars andfirms with missing data on our key variables7 The ratio variables aretrimmed at the 1 level in both tails of the distribution This serves toremove outliers and the most extremely misrecorded data The finalsample consists of 179021 firm-year observations from 1971 to 2009Table I provides definitions of financial variables and reports summary

statistics The average debt (in constant US$) is about $653 million while themedian is $24 million A significant fraction of firms have zero debt (the 10thpercentile is 0) Book equity is slightly larger than book debt Market equityis almost three times larger than book debt Book assets average $2191million although the medians are considerably smallerIf issuing or retiring securities incurs no fixed costs then we would expect

to see many small actions and very few large actions (Leary and Roberts2005) If there were significant fixed costs involved in issuing or retiringoutstanding securities then small issues might not be worthwhile Table I

7 These include debt book value of equity market value of equity assets book andmarket leverage profitability market-to-book assets ratio and tangibility

1424 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

shows that although most firms issue little debt or equity in a given year theaverages are large In other words when firms actually enter debt and equitymarkets they intervene massivelyThe mean constant dollar debt issue is $165 million (the median is $2

million) In unreported tables we find that the average debt issue is about81 of assets (the median is 18) About 38 of the firms issue no debt8 issue between 0 and 1 of the value of their assets as debt another

Table I Data description

Data sources The sample comes from the annual Compustat files The sample period is

1971ndash2009 We exclude financial companies (SIC 6000-6999) firms involved in majormergers (Compustat footnote code AB) firms reporting financial data in currencies otherthan the US dollar and firms with missing data on the key variables All financial variables

are deflated to year 2000 using the GDP deflator N is the number of observations SD isthe standard deviation D BVE MVE Assets Debt issuance Equity issuance Debt re-payment Equity repayment are all measured in $ millions The ratio variables are trimmed

at the 1 level in both tails of the distribution

Variable definitions Debt (D)frac14 long-term debt (dltt)thorn short-term debt (dlc) book equity(BVE)frac14 common shareholder equity (ceq) market equity (MVE)frac14 no of outstanding

shares (csho) closing price ( prcc f ) assetsfrac14 book assets (at) debt issuancefrac14 issuance oflong-term debt (Maxethdltis 0THORN)thorn increase in current debt (Maxethdlcch 0THORN) equity issu-ancefrac14 sale of common stock (Maxethsstk 0THORN) debt repaymentfrac14 reduction of long-termdebt (Maxethdltr 0THORN)thorn decrease in current debt (Minethdlcch 0THORN) equity repurchasefrac14 purchase

of common stock (Maxethprstkc 0THORN) book leveragefrac14D(DthornBVE) market leveragefrac14D(DthornMVE) profitabilityfrac14EBITDA (oibdp)assets market-to-book ratio M

B frac14Marketvalue of assets (MVA)assets where MVAfrac14DthornMVEthorn preferred-liq value (pstkl)

deferred taxes (txditc) tangibilityfrac14 net property plant and equipment (ppent)assets

Variable N Mean SD

Distribution

10th 50th 90th

Debt (D) 179021 653 4592 0 24 1048

Book equity (BVE) 179021 815 4122 3 70 1373

Market equity (MVE) 179021 1784 9589 9 118 2698

Assets 179021 2191 11530 10 153 3575

Debt issuance 179021 165 1335 0 2 245

Equity issuance 179021 26 186 0 0 40

Debt repayment 179021 132 1123 0 3 181

Equity repurchase 179021 24 257 0 0 9

Book leverage 179021 036 034 000 031 072

Market leverage 179021 027 025 000 020 066

Profitability 179021 005 027 018 011 023MB 179021 165 206 052 102 324

Tangibility 179021 031 024 005 026 070

PROFITSndashLEVERAGE PUZZLEREVISITED 1425

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

16 issue between 1 and 5 of the value of their assets as debt and theremaining 38 issue debt in excess of 5 of the value of their assetsThe mean constant dollar equity issue is about $26 million (the median is

about $04 million) As a fraction of assets the mean and median equityissues are about 7 and 02 respectively About 33 of the firms issue noequity 34 of the firms issue between 0 and 1 of the value of theirassets as equity another 14 issue between 1 and 5 of the value of theirassets as equity and the remaining 19 of the firms issue equity that is inexcess of 5 of the value of their assets Average debt repayments are largerthan equity repurchases This perhaps reflects the finite maturity of debt andits contractual repayment The median firm does not repurchase equityWe construct both book and market leverage ratios Book leverage is

defined as debt over debt plus book equity Market leverage is defined asdebt over debt plus market equity8 The median book leverage is 036 (theaverage is 031) The median market leverage is 027 (the average is 020)Profitability is defined as the ratio of operating income before depreci-

ation to assets While the average firm is profitable (the ratio of EBITDA toassets is 005) the median firm is even more profitable (with a profitabilityratio of 011) But the sample also includes a large number of unprofitablefirms as the 10th percentile is 018 The table also reports descriptive stat-istics on the market-to-book ratio and the tangibility ratio The market-to-book ratio (MB) defined as the ratio of the market value of assets to bookassets averages at about 165 Tangibility defined as the ratio of netproperty plant and equipment to assets averages at about 31

4 Leverage Ratio Regressions

The previous literature focuses on estimates obtained using leverage ratiosHence we start with a similar estimation to check whether our results matchthose of previous studies Table II presents the results for book leverage Weobtain similar results for market leverage and they are presented inAppendix Table AI The regressions include leverage factors following theprevious capital structure literature (see eg Lemmon Roberts and Zender(2008) and Frank and Goyal (2009)) The factors include (i) profitability (ii)industry median leverage (iii) market-to-book assets ratio (iv) tangibility ofassets and (v) firm size Industry median leverage is constructed as themedian leverage of all other firms in the industry excluding the firm under

8 Welch (2011) stresses the idea that nonfinancial liabilities should not be implicitly mis-treated as if they were equity by paying excessively narrow attention to financial liabilitiesin a leverage ratio We have adopted his approach in the empirical work reported here

1426 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

consideration Firm size is defined as the natural log of assets Other factorsare defined in Section 3 We cluster standard errors by firm and estimatethese regressions both with and without fixed effectsIn Table II we run conventional cross-section leverage regressions Since

we employ the usual data it is not surprising that our results match thosereported in the existing capital structure literature Profitability has anegative sign in both the book leverage regressions and in the marketleverage regressions The coefficients on other factors largely match thosereported in earlier studies Firms operating in industries with high leverage

Table II Leverage and profitability

The table presents parameter estimates for book leverage regressions The sample consists

of nonfinancial companies in the annual Compustat files during the period 1971ndash2009 Theleverage is estimated as the ratio of debt over debt plus book equity The explanatoryvariables Profitabilityt1

MB

t1

Tangibilityt1 and LnethAssetsTHORNt1 are described in

Table I IndLevt1 is estimated as the median book leverage of all other firms in thesame industry The industry is defined at the level of the four-digit SIC code All specifi-cations include the year fixed effects The specifications in column (2) additionally include

the firm fixed effects We report t-statistics where the standard errors are clustered at thefirm level in parentheses

aSignificant at the 1 level

Quantile regressions

(1) (2) (3) (4) (5)

Profitabilityt1 026a 020a 011a 017a 041a

(293) (192) (378) (518) (1015)

IndLevt1 042a 012a 030a 052a 052a

(369) (107) (926) (1252) (950)MB

t1

002a 001a 001a 002a 002a

(203) (102) (392) (426) (418)

Tangibilityt1 018a 025a 024a 019a 012a

(206) (154) (964) (580) (276)

LnethAssetsTHORNt1 001a 002a 002a 002a 001a

(114) (76) (899) (478) (96)

Constant 012a 010a 012a 002a 029a

(156) (60) (284) (27) (387)

Year FE Yes Yes Yes Yes Yes

Firm FE No Yes No No No

Regression type OLS FE 25th tile Median 75thtile

Clustered SE Yes Yes No No No

R2 013 054

Pseudo R2 016 015 009

N 158824 158824 158824 158824 158824

PROFITSndashLEVERAGE PUZZLEREVISITED 1427

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

tend to have high leverage A higher market-to-book ratio is associated withlower leverage Larger firms are typically more highly levered The coeffi-cients on tangibility are positive The conclusion from Table II is clear Theleverage regression results match those reported in previous studiesAt times one may be concerned that in a regression what is being

estimated is a conditional mean while some parts of the distributionmight behave differently from other parts If so then ldquoaccidentsrdquo likesample selection issues or asymmetries in the underlying distributionsmight play a greater role than is usually recognized In other words thesource of identification could be misinterpreted To guard against this po-tential problem we estimate the same model using quantile regressions9

Columns 3 4 and 5 report what happens at the 25th 50th and 75th per-centiles The basic model results are robust in terms of both the sign and thestatistical significanceThe impact of profits does appear to be somewhat stronger among the

high leverage firms For instance in a book leverage regression the coeffi-cient on profits for the first quartile is 011 with a t-ratio of 378 For thethird quartile the coefficient on profits is 041 with a t-ratio of 1015This difference is curious and might deserve further study However the factthat in each case we get the negative sign and statistical significance is suf-ficient for the purposes of the present article So the results on profits arerather robust across the distributionTable III sorts firms into those that are actively adjusting their leverage

and those that are passive In the presence of both fixed and variable trans-acts costs this distinction is important We expect the coefficient on profit-ability in leverage ratio regressions to be more negative in firms that arepassive and less negative in firms that are actively adjustingWe define active firms as those that are either issuing net debt or net

equity in excess of 5 of their assets Passive firms are those that neitherissue debt nor equity in nontrivial amounts (ie greater than 5 of theirassets) Consistent with our expectations we find in Table III that the coef-ficients on profits are significantly more negative for passive firms than theyare for active firms With firm and year fixed effects the coefficient onprofits is 008 for active firms and 017 for the passive firms The twocoefficients are significantly different from each other (plt 001) InTable III the standard control variables are used They have the usualsigns and for the most part do not differ much between the active and thepassive firms

9 Cameron and Trivedi (2010) provide an extensive discussion of quantile regressions inStata

1428 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

5 Financing Activity

51 FREQUENCY OF DEBT AND EQUITY ISSUES

Table IV reports the percentage of firms issuing or repurchasing debt orequity for annual sorts based on lagged profitability In Panel A weemploy the conventional 5 cutoff rule to exclude minor fluctuations anddefine firms ldquoissuing debtrdquo as those that issue debt (both short-term andlong-term) in excess of 5 of the value of their assets Other decisions aresimilarly defined using a 5 cutoff Appendix Table AII examines the dis-tributional properties across profitability (and size) quintiles for ldquolargerdquo is-suances by increasing the 5 cutoff to higher levels such as 10 and 15How does security issuance behavior vary with firm profitability To

answer this question Table IV sorts firms by profitability and reports the

Table III Leverage and profitability for active and passive firms

The table presents estimates of leverage on firm characteristics for the subsamples of firms

that are active in either debt or equity markets Active firms are defined as those that issueeither net debt in excess of 5 of the value of assets or issue net equity in excess of 5 ofthe value of their assets The passive firms do not issue significant (more than 5) of either

debt or equity The sample comes from the annual Compustat files during the period 1971ndash2009 Financial firms are excluded The variables are described in Tables I and II Wereport t-statistics where the standard errors are clustered at the firm level in parentheses

aSignificant at the 1 level

Active Passive

(1) (2) (3) (4)

Profitabilityt1 006a 008a 022a 017a

(126) (148) (254) (188)

IndLevt1 044a 019a 041a 013a

(424) (149) (326) (127)MB

t1

002a 001a 003a 001a

(362) (196) (327) (172)

Tangibilityt1 010a 014a 014a 019a

(146) (118) (145) (130)

LnethAssetsTHORNt1 001a 004a 001a 004a

(121) (179) (147) (169)

Constant 012a 008a 003a 007a

(120) (57) (45) (42)

Year FE Yes Yes Yes Yes

Firm FE No Yes No Yes

R2 034 064 033 075

N 43395 43395 58943 58943

PROFITSndashLEVERAGE PUZZLEREVISITED 1429

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

percentage of firms issuing or repurchasing debt or equity The table alsoreports test statistics from tests that examine whether the percentage differ-ences between the high- and low-profit firms are significantly different fromeach other The results in Column (1) show that the likelihood of issuing debt

Table IV Profitability sorts for debt and equity issuers

The sample contains nonfinancial firms listed on the annual Compustat files for the period

from 1971 to 2009 The table presents the percentage of firms issuing and retiring (orrepurchasing) debt and equity The firms are classified as ldquoissuing debtrdquo if they issuedebt in excess of 5 of the value of their assets as ldquoissuing equityrdquo if they issue equity

in excess of 5 of the value of their assets as ldquoretiring debtrdquo if they retire debt in excess of5 of the value of their assets and as ldquorepurchasing equityrdquo if they repurchase equity inexcess of 5 of the value of their assets In addition we also report net debt issuers which

are firms that issue net debt over 5 of the value of their assets and net equity issuerswhich are firms that issue net equity in excess of 5 of the value of their assets Weannually sort firms on lagged profitability and report the percentage of firms in each ofthese categories The bottom part of each panel reports the percentage of firms issuing or

retiring securities by profitability within the smallest and largest asset quintiles The tablealso reports results from t-test for the difference in percentage of firms issuing or retiringsecurities in the lowest and the highest profit quintile

Percentage of firms

Iss Ret Iss Iss Rep Iss Both IssD IssE

D D NetD E E NetE DampE RepE RetD

(1) (2) (3) (4) (5) (6) (7) (8) (9)

Allfirm

s

Low 276 245 186 301 13 297 71 01 38

2 319 299 181 117 21 112 27 02 20

3 385 344 194 86 29 80 21 05 15

4 386 330 213 81 43 75 20 07 14

High 329 273 198 92 82 81 17 13 14

t Highfrac14Low 150 84 39 723 414 763 358 176 200

Smallfirm

s Low 275 204 201 411 11 405 109 01 47

2 291 233 213 379 10 378 99 01 45

3 281 246 204 265 14 261 66 01 41

4 275 272 186 162 15 159 34 02 35

High 275 281 178 136 27 131 25 02 26

t Highfrac14Low 01 84 28 316 55 318 171 18 55

Largefirm

s

Low 414 363 177 87 22 79 28 03 11

2 449 379 180 62 30 57 17 07 09

3 439 360 186 56 41 48 16 09 07

4 403 320 194 46 67 39 13 12 06

High 374 264 207 50 129 35 10 28 05

t Highfrac14Low 50 130 45 89 244 116 77 120 47

1430 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

is mostly independent of firm profitability In the aggregate there is a smalldifference but it is sensitive to sorting on firm size Similarly in Column (2)there is only a weak relation between the profitability and the likelihood thatthe firm retires debt To a first approximation the fraction of firms issuing orrepurchasing debt is largely independent of profits Large firms are somewhatmore likely than small firms to issue debt and also to retire debtUnlike debt issues the probability of issuing equity is strongly related to

profitability as can be seen in Column (4) Low-profit firms are much morelikely to issue equity than are high-profit firms In the lowest profit quintile301 of the firms issued equity In the highest profit quintile it was only92 The lower panel reports separate sorts for small and large firms Similarpatterns are found for both groups of firms that is the likelihood of equityissuance declines as profitability increases The differences between the quin-tiles are quite stark for the group of small firms All quintiles of small firmsissue equity much more frequently than the corresponding large firm quintilesHigh-profit firms are much more likely to repurchase equity as shown in

Column (5) In the overall sample only 13 of low-profit firmsrepurchased equity compared with 82 of the high-profit firms Whenwe separate the small and large firms again a difference emerges Ingeneral small firms do not repurchase much equity while a greaterfraction of the large firms do soGiven these facts it is natural that low profitability firms are much more

likely to be net issuers of equity than are high profitability firms as shown inColumn (6) Among the firms in the lowest quintiles of profits roughly 30of firms issue net equity that exceeds 5 percent of their assets In contrastfor the most profitable quintile only about 8 issue net equityFirm size is an important variable in the recent literaturemdashit is often used

as a proxy for access to capital markets (as in Faulkender and Petersen(2006) and Leary (2009)) Small firms are bank-dependent risky and infor-mationally opaque They have restricted access to public debt markets andconsequently face more severe supply constraints in their ability to issuedebt Thus we expect small firms to be more sluggish in adjusting theirdebt and equity in response to shocks to profitability Large firms incontrast have much easier access to public debt markets and they facefewer obstacles in accessing securities marketsTo examine how size and profitability interact the bottom part of Table IV

examines profitability sorts for the smallest and largest firms We first sortfirms annually by firm size and then within each size quintile we sort them onprofitability We do these two-way sorts to ensure that we have similarnumber of firms in profitability subgroups for both small and large firmsamples Unconditionally small firms are generally less profitable while

PROFITSndashLEVERAGE PUZZLEREVISITED 1431

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

large firms are relatively more profitable The two-way sorts allow us to havea uniform distribution of firms in various profitability quintiles for both smalland large firm samples The bottom part of Table IV reports results for thesmallest and the largest size quintiles Among the small firms there is littlerelation between profitability and the likelihood of issuing debt Howeverlarger firms exhibit a small increase in the likelihood of net issuance of debtwith increasing profitabilityIn contrast to debt the effects of profitability on equity issuance and

repurchases are much more consistent across size quintiles As profitabilityincreases firms are generally less likely to issue equity and more likely torepurchase it Across the two extreme size groups we note a striking differ-ence between the proportion of firms issuing equity and those issuing debtAmong low-profit small firms almost 41 are equity issuers whereas onlyabout 9 of low-profit large firms are equity issuers However regardless ofsize we note a monotonic reduction in the likelihood of issuing equity asprofitability increasesColumn (7) reports the fraction of firms that issue both debt and equity in

excess of 5 of assets both as a function of profits in the upper panel andas a function of firm size and profits in the lower panel In the upper panelwe see that low-profit firms are much more likely to issue both debt andequity (71) while the high-profit firms are less likely to do so (17) Thelower panel shows that small firms are much more likely to be issuers of bothin contrast to large firms In both firm size categories the profit pattern is thesame with low profits being more frequent issuers of both simultaneouslyImportantly Column (8) shows that the likelihood of issuing debt and

simultaneously repurchasing equity increases with profitability Converselyin Column (9) we find that the likelihood of doing the reverse that is issuingequity and retiring debt declines with profitability The effects of sorting onfirm size and firm profitability mirror those reported for all firms Overallthe results show that low-profit firms are less likely to issue debt and repur-chase equity they are instead more likely to issue equity and retire debtFirms with high profitability exhibit the reverse patternIn Appendix Table AII we present results from profitability (and size)

quintiles for large issuances defined using the 10 cutoff (in Panel A) andthe 15 cutoff (in Panel B) The results from these panels are largely con-sistent with those described above

52 MAGNITUDES OF ISSUANCES AND EQUITY CHANGES

In the previous section we considered the probability of having a nontriviallevel of debt or equity activity The next question is how large are the dollar

1432 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

values involved In Table V we sort firms according to profits and thentabulate the levels and changes in both debt and equity We do this firstfor all firms and then for small versus large firmsFor the analysis on all firms we observe in Column (1) that debt peaks at

the middle of the distribution This is because firms with medium profitabil-ity are also the largest as seen in Column (8) which reports average assetvalues for different profitability quintiles When we sort first by firm size andthen examine firms with different levels of profitability within size quintileswe find that for small firms debt is roughly independent of profits while forlarge firms less profitable firms have a higher level of debtAs expected Columns (3) and (5) show that more profitable firms have

higher equity values Columns (2) and (7) consider the relationship betweenissuances of debt and equity and firm profitability Debt issuances are sig-nificantly larger for more profitable firms In contrast equity issuances arelarger for low-profit firms Small low-profit firms issue substantially lessequity compared with small high-profit firms Among large firms the low-profit ones tend to issue equity while the high-profit ones tend to repurchaseequity The fact that more profitable firms issue debt and repurchase equitywhile the least profitable firms retire debt and issue equity is consistent withthe predicted relation between profitability and financing decisions under thetrade-off theoryWe also examine two-way sorts by size and profitability and report results

for the smallest and largest quintile of firms For the smallest quintile offirms the change in debt is largely unrelated to profits But for large firmsthere is a positive relation between profits and debt issuances High-profitfirms have a big positive change in debt Low-profit large firms have anegative change in debtColumns (4) and (6) provide an explanation of why the leverage ratio

regression results contradict those from the basic profitability sorts pre-sented here As we can see profitability indirectly affects leverage ratiosby increasing equity values Changes in both the book value of equity andthe market value of equity are positive and large for highly profitable firmsIn contrast these changes are negative for less profitable firms This evidenceillustrates an important issue concerning the use of leverage ratios Suchratios are often interpreted as essentially reflecting the use of debt by thefirm This interpretation while common is empirically misleadingFor the typical firm the change in the value of equity is larger than the

change in debt For example in the third profit quintile for large firmscolumn (7) the mean equity issue is just $9 million but the change in themarket value of equity is $242 million At the same time the mean change indebt is $27 million This suggests that a fair bit of the observed variation in

PROFITSndashLEVERAGE PUZZLEREVISITED 1433

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

the leverage ratios is primarily driven by the changes in the market value ofequity in the denominator rather than by the changes in debt in the numer-ator Since equity issues are often small this implies that the variation in theleverage ratio is primarily driven by internal operations rather than byexternal financing actions This again points to the fact that leverageratios can provide a misleading account of actual patterns in the dataThere is always a potential concern that the averages may be misleading

due to the impact of outliers To address this concern in AppendixTable AIII we present median values of the profitability sorts Theseresults generally reinforce the findings in Table VTable VI again considers the magnitudes of financing activity but this

time the issues are scaled by a measure of firm size That way it is easier to

Table V Magnitude of financing activity

The sample contains nonfinancial firms listed on the annual Compustat files for the period

from 1971 to 2009 The table presents average debt and equity changes in debt and equityand equity issuances (in millions of dollars) for firms sorted on profitability within sizeclasses The table reports information for all firms sorted on profitability and for profit-

ability sorts within the smallest and largest firms The sorts are done annually The tablealso reports results from t-test for the difference in percentage of firms issuing or retiringsecurities in the lowest and the highest profit quintile

D D BVE BVE MVE MVE EqyIss Assets

(1) (2) (3) (4) (5) (6) (7) (8)

Allfirm

s

Low 83 5 78 10 120 5 7 234

2 605 13 509 24 721 8 17 1691

3 1227 9 1123 22 1834 3 9 3551

4 833 23 1166 7 2408 74 9 2994

High 570 53 1171 71 3817 68 36 2581

t Highfrac14Low 31 9 36 9 38 2 16 36

Smallfirm

s Low 25 04 42 01 332 28 33 97

2 29 06 55 02 278 16 21 116

3 31 07 63 01 212 05 12 127

4 30 06 70 05 179 10 08 136

High 24 06 84 14 234 17 07 142

t Highfrac14Low 04 15 153 85 73 11 201 107

Largefirm

s Low 3929 55 2990 152 4671 11 63 10846

2 3826 19 3178 95 5122 66 11 10721

3 3074 27 3498 41 6173 242 9 9697

4 2623 62 3915 45 8473 94 57 9855

High 2180 193 4642 261 14924 202 148 10171

t Highfrac14Low 94 64 108 87 246 12 169 16

1434 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

see whether large firms behave disproportionately when compared withsmall firms Column (1) shows that low-profit firms have more debtas reported in many previous studies In Columns (4) and (6) we seethat low-profit firms have a major decline in both the book value ofequity and the market value of equity The high-profit firms have thereverse experienceMore interesting are the active decisions of the firms In Column (2) we see

that in contrast to the usual impression the low-profit firms are reducingdebt while the high-profit firms are increasing it In Column (7) we see thatthe low-profit firms are also reducing leverage by issuing equity The high-profit firms issue little equity In the lower panel similar results are foundwhen we separate small and large firms Equity issuing is particularly strong

Table VI Magnitude of financing activitymdashscaled levels and changes

The sample contains nonfinancial firms listed on the annual Compustat files for the period

from 1971 to 2009 The table presents average debt and equity changes in debt and equityand equity issuances scaled by assets for firms sorted on profitability within size classes Thetable reports information for all firms sorted on profitability and for profitability sorts

within the smallest and largest firms The sorts are done annually The table also reportsresults from t-test for the difference in percentage of firms issuing or retiring securities inthe lowest and the highest profit quintile

D D BVE BVE MVE MVE EqyIss

(1) (2) (3) (4) (5) (6) (7)

Allfirm

s

Low 386 122 12 871 3427 1868 217

2 285 58 407 119 940 157 28

3 293 12 425 14 742 03 12

4 269 26 461 02 955 27 10

High 192 15 551 47 1675 74 07

t Highfrac14Low 35 49 50 47 127 42 157

Smallfirm

s Low 545 154 841 2145 9688 9128 696

2 684 49 712 1915 5014 2318 288

3 331 253 141 969 3063 954 129

4 267 181 357 175 1673 221 55

High 194 05 543 26 1758 147 41

t Highfrac14Low 49 17 51 36 88 27 99

Largefirm

s Low 350 30 306 58 511 55 09

2 346 14 335 13 523 02 04

3 353 165 332 71 711 29 01

4 280 01 414 04 909 09 05

High 227 15 477 19 1643 27 14

t Highfrac14Low 371 112 453 181 570 57 221

PROFITSndashLEVERAGE PUZZLEREVISITED 1435

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

among the small low-profit firms that are also experiencing sharp reductionsin the value of equity In column (2) we see that both for the large and for thesmall firms net debt increases are primarily found in the highest quintile offirm profitability

6 Debt and Equity Regressions

So far we have demonstrated that in our data the conventional leverageregressions have the usual signs We then demonstrated that in simple sortsof the data strikingly different active financing patterns emerge The nexttask is to reconcile these differencesTable VII presents simple regressions predicting changes in debt book

equity and market equity as well as equity issuances Consistent with thesorts we control for changes in firm size and we also include firm and yeardummies There may be a concern about the speed of a firmrsquos responsesAccordingly we consider the lagged change in profits as the main case butwe also include the contemporaneous change in profits Whether the contem-poraneous change in profits is included or not does not in fact matter Wereport panel-robust standard errors adjusted for clustering at the firm levelThe results in Table VII support the findings in the sorts Firms that have

an increase in profits in 1 year increase debt both in the same year and in thenext year Firms that have an increase in profits experience an increase inboth the book value of equity and the market value of equity Importantlyhowever firms that experience and increase in profits reduce their equityissues All these effects are statistically significant (except for the marketvalue of equity in the column 5 specification) and seem intuitively reason-able In Appendix Table AIV similar regressions to those presented aboveare reported but without firm-fixed effects We find that empirically verylittle changes whether we include firm fixed effects or leave them out Thusthe targeting behavior to the extent that it happens does not account forwhat we are reportingThe next step is to bring the sorts and the simple regressions together into

a common setting We therefore include factors in addition to profitability(i) median industry leverage (ii) the market-to-book assets ratio (iii) tangi-bility of assets and (iv) firm size (measured by the log of assets) Rajan andZingales (1995) show that these factors are related to leverage in G7countries A number of studies have used these factors to estimateleverage targets Frank and Goyal (2009) show that these factors arerobustly related to leverage in the USA In these regressions we use indica-tors for the quintile that the firm is in for each factor

1436 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

Table VIII reports these results In Columns (1) and (2) we examine debtissues and again show that profits positively affect debt issuances The effectsare large and statistically significant at the 1 level Columns (3) and (4)which examine book equity show a strong effect of profits This effect is onlyslightly reduced by the inclusion of other factors Columns (5) and (6) illus-trate that the impact of profits on the change in market equity is also robustto the inclusion of the conventional factors Columns (7) and (8) examineequity issuances Here we again find that profits have a significant negativeimpact on equity issuances As before the results are robust to the inclusionof the conventional factorsIn unreported tables we estimate debt and equity changes using quantile

regressions In terms of the signs and significance of the profitability variable

Table VII Debt and equity changes

The sample contains nonfinancial firms listed on the annual Compustat files for the period

from 1971 to 2009 The table presents fixed effect estimates from regressions of changes indebt (D) changes in the book value of equity (BVE) changes in the market value ofequity (MVE) and net equity issuances (EqyIss) The explanatory variables include

current and lagged change in operating income before depreciation (Profits andProfitst1) and lagged change in the book value of assets (Assetst1) In addition tofixed firm effects the regressions include year indicator variables The reported t-statistics

are corrected for clustering at the firm level

aMean significant at the 1 level

bMean significant at the 5 level

cMean significant at the 10 level

D BVE MVE EqyIss

(1) (2) (3) (4) (5) (6) (7) (8)

Profitst1 022a 025a 015b 035a 004 060b 005b 007a

(33) (31) (25) (35) (03) (25) (25) (53)

Profits 050b 078a 140a 003a

(24) (32) (48) (34)

Assetst1 000 007b 025a 001b

(00) (21) (35) (25)

Constant 167 283a 34 51 1145a 1182a 605a 595a

(14) (30) (04) (05) (32) (29) (69) (74)

Year FE Yes Yes Yes Yes Yes Yes Yes Yes

Firm FE Yes Yes Yes Yes Yes Yes Yes Yes

R2 001 004 001 010 001 005 002 003

N 162056 162056 162130 162130 157550 157550 162154 162154

PROFITSndashLEVERAGE PUZZLEREVISITED 1437

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

the results are qualitatively similar to the OLS results reported in Table VIIIProfitability positively affects debt issuances and negatively affects equityissuances across various quantiles Overall the results in Table VIII showthat there is nothing that is special about using sorts or regressions to explaindebt and equity issues In either case we find that more profitable firms tendto increase their debt experience an increase in the value of equity andrepurchase shares Thus the control factors are not responsible for theusual rejection of the trade-off theory It would appear that what mattersis the size of the issuing activity relative to the organic increases in equityvalue due to profits In other words what matters in the leverage regressionseems to be coming from the presence of E in D

DthornE and not from D

Table VIII Debt and equity issuances profitability and leverage factor quintiles

The sample contains nonfinancial firms listed on the annual Compustat files for the period

from 1971 to 2009 The table presents estimates from regressions of change in debt (D)change in the book value of equity (BVE) change in the market value of equity (MVE)and net equity issuances (EquityIss) The explanatory variables include lagged profitability

quintiles and lagged leverage factor quintiles In addition to fixed firm effects the regres-sions include year indicator variables The reported t-statistics are corrected for clustering atthe firm level

aMean significant at the 1 level

bMean significant at the 5 level

cMean significant at the 10 level

D BVE MVE EqyIss

(1) (2) (3) (4) (5) (6) (7) (8)

ProfitabilityQint 214a 167a 283a 175a 80 429a 53a 55a

(70) (54) (66) (59) (08) (48) (55) (62)

IndLevQint 104a 43 99 20

(29) (15) (10) (10)

Ln(Assets)Qint52 17 1026a 53a

(16) (05) (77) (35)MB

Qint158a 372a 1155a 09

(45) (56) (106) (06)

TangibilityQint 05 177 436c 64a

(01) (15) (18) (31)

Constant 167 96 656b 235 4613a 8610a 113 570a

(08) (04) (26) (04) (61) (58) (11) (31)

Year FE Yes Yes Yes Yes Yes Yes Yes Yes

Firm FE Yes Yes Yes Yes Yes Yes Yes Yes

R2 0002 0002 0003 0004 0013 0014 0013 0014

N 155587 155037 155587 155037 155587 155037 155587 155037

1438 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

7 Are Financial Market Conditions Important

According to the trade-off theory capital structure is determined by con-sidering the impact of costs and benefits of debt The time variation inthe costs and benefits of debt imply that issuance decisions would alsovary over time Furthermore market conditions affect the transactioncosts of adjusting Adverse selection is a greater problem in a cold marketthan in a hot one Accordingly it may be easier to adjust in good marketconditions Related ideas have been developed by Huang and Ritter(2009) Faulkender et al (2012) and Halling Yu and Zechner (2012)Faulkender Flannery Hankins and Smithrsquos paper examines cross-sectionalvariation in adjustment speeds including market valuations andhow they may affect adjustment speeds It is natural to think that ingood market conditions it will be less costly to issue both debt and equityand adjusting leverage to respond to profitability shocks would be mucheasierTo test the importance of market conditions we require a definition of

good times and bad times Our empirical strategy is to estimate good timesversus bad times at the four-digit industry level We define an industry ashaving ldquogood timesrdquo if the median firm in that industry has a market-to-book ratio that is higher than the 67th percentile of the time-series distribu-tion of the industry median market-to-book ratios Conversely an industryis defined as having ldquobad timesrdquo if the median firm in the industry has amarket-to-book ratio that is lower than the 33rd percentile of the time-seriesdistribution of the industry median market-to-book ratiosPanel A of Table IX tabulates issuance activity for profitability sorts in

both good times and bad times As might be expected active debt and equityissuances are larger during good times There is also somewhat more activeswaps between debt and equity during good timesPanel B reports the issuance activity for the smallest and the largest firms

during the good times and bad times Within each size quintile firms aresorted on profitability This is a three-way sort As before we find that thereis much more active use of external markets during good times Small low-profit firms are more likely to issue equity in good times than in bad timesDebt issuances are significantly higher in good times Large high-profitfirms are significantly more likely to issue debt and repurchase equity ingood times than in bad timesPanel C reports the magnitudes of the financial variables rather than the

frequencies During bad times less profitable large firms retire substantialamounts of debt and they show a tendency to issue equity Small firms donot seem to engage in similar debt reduction activities However like the

PROFITSndashLEVERAGE PUZZLEREVISITED 1439

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

Table IX Debt and equity issues in good and bad times

The table reports the frequency and magnitude of financing activity for sorts on profits for subsamples

of firms in good and bad times An industry is defined as having ldquogood timesrdquo if the median firm in

that industry has a market-to-book ratio that is higher than the 67th percentile of the time-series

distribution of industry median market-to-book ratios Conversely an industry is defined as having

bad times if the median firm in that industry has a market-to-book ratio that is lower than the 33rd

percentile of the time-series distribution of industry median market-to-book ratios Panels A and B

report the percentage of firms (i) issuing net debt in excess of 5 of the value of their assets (ii) issuing

net equity in excess of 5 of the value of their assets (iii) issuing debt and repurchasing equity both in

excess of 5 of the value of their assets and (iv) issuing equity and retiring debt both in excess of 5

of the value of their assets Panel C tabulates the mean debt and equity levels and the mean changes in

debt and equity issuances The sample contains nonfinancial firms listed on the annual Compustat files

for the period from 1971 to 2009

Iss Iss Iss D Iss ENet D Net E Rep E Ret D(1) (2) (3) (4)

Panel A sorts on profitability

Badtimes

Low 145 131 01 202 170 49 02 083 173 46 04 074 182 47 03 10

High 177 51 06 10t Highfrac14Low 41 136 33 01

Goodtimes

Low 206 381 01 462 197 148 02 233 212 103 04 184 234 92 10 18

High 216 92 16 14t Highfrac14Low 23 658 138 23

Panel B sorts on firm size and profitability

Badtimes

Smallfirm

s Low 151 318 00 312 182 276 02 253 151 140 02 294 183 73 00 19

High 177 80 00 11t Highfrac14Low 10 90 NA 20

Largefirm

s Low 172 54 01 042 148 39 04 053 150 43 04 044 156 33 03 06

High 147 28 05 03t Highfrac14Low 17 32 20 02

Goodtimes Smallfirm

s Low 213 448 01 542 230 443 01 473 222 325 00 474 203 197 02 38

High 190 153 02 31t Highfrac14Low 19 230 10 39

Largefirm

s Low 203 96 03 122 208 66 09 133 232 54 16 064 215 39 19 05

High 234 36 38 04t Highfrac14Low 29 99 92 38

(continued)

1440 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

larger firms there is a tendency for small firms to issue more equitymdashat leastto the extent possibleDuring good times large profitable firms raise significant amounts of debt

as they experience an increase in the value of their equity Such firms alsoengage in share repurchases Debt issues by smaller firms are much lessaffected by their own profitability during good times and their change indebt is more or less independent of their profits However equity issuanceis primarily found among the low-profit small firms during good timesTable IX shows that empirically profitability has a time-varying impacton leverage choicesIn Table X the basic regressions for changes in debt and for equity

issuance are presented for both good times and bad times Comparinggood times with bad times we see that the effects of profits are muchstronger in good times In bad times the impacts are rather weak statisticallyThis difference is interesting and deserves further study However for ourpurposes recall that the static trade-off theory implies time-varying capital

Table IX Continued

D D BVE BVE MVE MVE EqyIss Assets

(1) (2) (3) (4) (5) (6) (7) (8)

Panel C magnitude of financing activity

Badtimes Smallfirm

s Low 40 03 58 14 305 06 22 140

2 44 05 82 07 240 67 17 180

3 45 04 94 02 176 18 05 187

4 43 09 98 02 137 04 03 188

High 36 10 114 13 209 16 04 199

Largefirm

s Low 3365 73 2523 169 2304 163 94 8972

2 3031 95 2983 35 3284 278 56 9068

3 2930 5 3360 14 3947 308 57 9225

4 2281 6 3817 19 5213 442 7 8902

High 1790 62 4767 13 8799 58 31 9573

Goodtimes

Smallfirm

s Low 24 06 38 05 339 23 39 90

2 26 08 48 01 301 04 25 104

3 30 09 60 04 246 03 17 125

4 28 07 66 08 206 18 12 130

High 23 06 85 16 261 13 08 141

Largefirm

s Low 4532 36 3477 103 6557 260 53 12465

2 4580 58 3527 139 6911 421 23 12522

3 2952 111 3485 2 7586 279 32 9535

4 2752 124 3939 72 10686 33 105 10226

High 2429 228 4819 377 18565 268 222 10830

PROFITSndashLEVERAGE PUZZLEREVISITED 1441

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

structure choices even when the target is time invariant Empirically we doobserve time-varying choices

8 Is Scaling Important

In the preceding section we provided results for unscaled debt and equityissues This seems appropriate to us However it involves making twochanges to the standard regression We would like to determine if both

Table X Debt and equity changes in good and bad timesmdashregressions

The sample contains nonfinancial firms listed on the annual Compustat files for the period

from 1971 to 2009 The table presents estimates from regressions of change in debt (D)and net equity issuances (EqyIss) in good and bad times The explanatory variables includechange in profits lagged change in profits and lagged change in assets The regressions

include both firm and year fixed effects An industry is defined as having ldquogood timesrdquo ifthe median firm in that industry has a market-to-book ratio that is higher than the 67thpercentile of the time-series distribution of industry median market-to-book ratios

Conversely an industry is defined as having bad times if the median firm in thatindustry has a market-to-book ratio that is lower than the 33rd percentile of the time-series distribution of industry median market-to-book ratios The reported t-statistics arecorrected for clustering at the firm level

aMean significant at the 1 level

bMean significant at the 5 level

cMean significant at the 10 level

Good times Bad times

D EqyIss D EqyIss

(1) (2) (3) (4) (5) (6) (7) (8)

Profits 087a 091a 006b 009a 060 065 003 003b

(30) (29) (20) (28) (15) (16) (16) (20)

Profitst1 029c 015a 034c 001

(19) (37) (17) (04)

Assetst1 000 002b 000 001

(02) (22) (00) (07)

Constant 1337 3989 8691a 10001a 1351 1851 4099a 4055a

(03) (09) (55) (37) (09) (13) (80) (78)

Year FE Yes Yes Yes Yes Yes Yes Yes Yes

Firm FE Yes Yes Yes Yes Yes Yes Yes Yes

R2 0071 0080 0018 0045 0178 0212 0034 0041

N 74827 68729 74891 68787 26965 26013 26971 26018

1442 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

changes are crucial Accordingly in this section we take a step back toconsider prior literature in which variables are scaled The questions arehow does scaling affect the results and whyTable XI examines alternative scalings Columns (1) and (2) scale debt and

equity issuances by the total firm net issuance Columns (3) and (4) scaledebt and equity issuances by the sum of debt and book equity Columns (5)and (6) consider the change in debt scaled by the sum of debt and bookequity Columns (7) and (8) consider the change in leverage due to externalfinancing Since we are examining changes we use the changes in thestandard regression factors as controls The primary interest is in the coef-ficient on the change in profitsWe find that a change in profits is associated with an increase in debt

issues But the sign reverses when we scale the debt issues by the sum of debtplus book equity This is telling What is really driving the results is thechange in equity in the denominator not the change in debt in the numer-ator This result is further substantiated in Columns (5) and (6) where thechange in debt to capital ratio is used instead of debt issuanceTo further explore the impact of active issuance relative to passive

experiencing of changes we decompose leverage by separating out theeffect of debt and equity issues from those due to changes in retainedearnings Thus we define leverage change due to external financing as

LevEF frac14Dt1 thorn d

Dt1 thorn Et1 thorn dthorn e

Dt1

Dt1 thorn Et1 eth1THORN

where D is debt E is book value of equity d is debt issues net of retirementsand e is equity issues net of repurchases Garvey and Hanka (1999) use asimilar measure in their study This measure directly captures the netleverage effect of debt and equity issuancesColumns (7) and (8) report results from regressions of the net leverage

effect of debt and equity issues on changes in profitability after controllingfor other leverage factors in differences and year indicator variables Theeffect of profitability on leverage change due to external financing is positiveand statistically significant This comparison suggests that changes in totaldebt and equity are only partly a result of debt and equity issuance decisionsOther balance sheet adjustments complicate the inferences from leverageratio regressionsFor most of our analysis we follow the literature in treating debt as homo-

genous It is of course potentially interesting to ask what happens when wedistinguish between short-term debt and long-term debt Which part of thefirmrsquos debt maturity structure drives the under-adjustment Table XII

PROFITSndashLEVERAGE PUZZLEREVISITED 1443

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

provides results from such decompositions Alternative scalings areconsidered The main results of interest are that the distinction betweenshort- and long-term debt is important Short-term debt issuance is largelyindependent of changes in profits The negative sign on profit change stemsfrom the action of long-term debt This seems intuitively reasonable

Table XI Scaled issuances

The sample contains nonfinancial firms listed on the annual Compustat files for the period

from 1971 to 2009 The table presents estimates from regressions of debt issuance to totalissuance (in columns (1) and (2)) debt issuance to total capitalization (in columns (3) and(4)) change in debt to capitalization ratio (in columns (5) and (6)) and change in leverage

ratio considering only external financing (in columns (7) and (8)) The change in leverageratio considering only external financing is defined as

LevEF frac14Dt1 thorn d

Dt1 thorn Et1 thorn dthorn e

Dt1

Dt1 thorn Et1

where D is debt E is book value of equity d is debt issues net of retirements and e is equity

issues net of repurchases The regressions include lagged leverage factors and year indicatorvariables Some of the specifications include firm fixed effects The reported t-statistics arecorrected for clustering at the firm level

aMean significant at the 1 level

bMean significant at the 5 level

cMean significant at the 10 level

DebtIssTotIss

DebtIssethDthornETHORN D

DthornE LevEF

(1) (2) (3) (4) (5) (6) (7) (8)

Profitt1 004a 000 002a 001c 003a 002 001b 002a

(35) (03) (34) (18) (35) (14) (23) (26)

IndLevt1 004c 001 001 001c 000 000 000 000

(19) (06) (15) (19) (02) (04) (03) (05)

Market=bookt1 001a 001a 000a 000a 000a 000a 000 000c

(47) (85) (68) (60) (55) (36) (05) (18)

Tangt1 004 005 013a 011a 011a 006a 003a 003a

(12) (15) (144) (112) (79) (38) (44) (42)

Assetst1 008a 004a 006a 004a 002a 001a 003a 002a

(97) (45) (283) (161) (53) (35) (147) (101)

Constant 079a 074a 002a 003a 000 000 000b 002a

(497) (442) (91) (111) (00) (05) (22) (93)

Year FE Yes Yes Yes Yes Yes Yes Yes Yes

Firm FE No Yes No Yes No Yes No Yes

R2 0015 0098 0022 0076 0009 0070 0012 0086

N 133905 133905 140552 140552 138229 138229 140572 140572

1444 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

inasmuch as short-term debt is probably more associated with day-to-dayoperations rather than longer term planning considerationsIn Table XIII we examine change in leverage ( D

DthornE) and change inleverage due to external financing (LEVEF) during good and bad timesseparately The traditional negative effect of a change in profits on achange in leverage is found both in good times and in bad timesHowever when we decompose the impact in order to isolate the externalcomponent as before the result disappears In good times the externalfinancing component of a change in profits is weak but still positiveDuring bad times it is not statistically significant This implies that distin-guishing between good and bad times does not fundamentally alter the basic

Table XII Short- and long-term debt issuances

The sample contains nonfinancial firms listed on the annual Compustat files for the period

from 1971 to 2009 The table presents estimates from regressions of short-term debtissuance to total issuance (in columns (1) and (2)) long-term debt issuance to totalissuance (in columns (3) and (4)) short-term debt issuance to total capitalization (in

columns (5) and (6)) and long-term debt to total capitalization (in columns (7) and (8))Some of the specifications include firm fixed effects The reported t-statistics are correctedfor clustering at the firm level

aMean significant at the 1 level

bMean significant at the 5 level

cMean significant at the 10 level

STDebtIssTotIss

LTDebtIssTotIss

STDebtIssethDthornETHORN

LTDebtIssethDthornETHORN

(1) (2) (3) (4) (5) (6) (7) (8)

Profitabilityt1 0002 0003 0042a 0008 0001 0001 0019a 0012a

(02) (04) (30) (05) (14) (10) (45) (28)

IndLevt1 0009 0007 0070b 0033 0001 0002c 0014a 0005

(06) (05) (26) (11) (08) (17) (26) (09)

M=Bt1 0003a 0003a 0003b 0010a 0000 0000 0003a 0002a

(48) (35) (22) (65) (10) (01) (84) (42)

Tangibilityt1 0020 0007 0011 0067 0005a 0006a 0097a 0081a

(12) (03) (03) (16) (34) (33) (108) (89)

Assetst1 0039a 0011b 0034a 0024b 0001a 0001 0046a 0020a

(89) (22) (38) (23) (36) (13) (190) (82)

Constant 0001b 0153a 0798a 0591a 0000c 0006a 0057a 0096a

(25) (144) (470) (288) (17) (124) (292) (234)

Year FE Yes Yes Yes Yes Yes Yes Yes Yes

Firm FE No Yes No Yes No Yes No Yes

R2 0025 0089 0018 0073 0030 0147 0016 0298

N 136146 136146 136140 136140 143031 143031 142847 142847

PROFITSndashLEVERAGE PUZZLEREVISITED 1445

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

Table XIII Externally financed leverage changes in good and bad times

The sample contains nonfinancial firms listed on the annual Compustat files for the period

from 1971 to 2009 The table presents estimates from regressions of change in debt tocapitalization ratio (in columns (1) (2) (5) and (6)) and change in leverage ratio consider-ing only external financing (in columns (3) (4) (7) and (8)) The change in leverage ratio

considering only external financing is defined as

LevEF frac14Dt1 thorn d

Dt1 thorn Et1 thorn dthorn e

Dt1

Dt1 thorn Et1

where D is debt E is book value of equity d is debt issues net of retirements and e is equityissues net of repurchases The regressions include lagged leverage factors and year indicator

variables The reported t-statistics are corrected for clustering at the firm level An industryis defined as having ldquogood timesrdquo if the median firm in that industry has a market-to-bookratio that is higher than the 67th percentile of the time-series distribution of industry

median market-to-book ratios Conversely an industry is defined as having ldquobad timesrdquoif the median firm in that industry has a market-to-book ratio that is lower than the 33rdpercentile of the time-series distribution of industry median market-to-book ratios

aMean significant at the 1 level

bMean significant at the 5 level

cMean significant at the 10 level

Good times Bad times

DDthornE LevEF D

DthornE LevEF

(1) (2) (3) (4) (5) (6) (7) (8)

Profitabiltyt1 0022c 0005 0016c 0025b 0073b 0041 0007 0007

(17) (03) (18) (24) (25) (11) (04) (03)

IndLevt1 0000 0011 0001 0009 0010 0018 0001 0000

(00) (09) (01) (13) (07) (10) (02) (00)

MB

t1

0004a 0003b 0000 0001 0008a 0001 0000 0000

(40) (24) (05) (08) (28) (04) (03) (03)

Tangt1 0097a 0041 0033a 0038b 0127a 0048 0059a 0033

(45) (16) (27) (25) (49) (14) (43) (16)

Assetst1 0014a 0013b 0024a 0019a 0037a 0023c 0036a 0025a

(30) (20) (84) (48) (45) (17) (88) (43)

Constant 0030b 0036 0000 0003 0028c 0014 0021 0001

(20) (16) (00) (07) (17) (06) (11) (01)

Year FE Yes Yes Yes Yes Yes Yes Yes Yes

Firm FE No Yes No Yes No Yes No Yes

R2 0006 0088 0009 0117 0026 0251 0023 0202

N 56820 56820 58200 58200 23462 23462 23511 23511

1446 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

conclusions The traditional leverage result on profits is driven by increasesin equity that are experienced and then partially but not completely offsetby firmsTo the extent that firms are engaging in active financing they are following

the prescriptions of the trade-off theory Thus we have a coherent picture ofhow firms are responding to profits Basically more profitable firms experiencean increase in the value of their equity They respond to this by issuing debtand repurchasing equity But the magnitude of the increase in equity is largerthan the magnitudes of the active steps taken in the debt and equity marketsThis means that the firmsrsquo actual actions are as predicted by the trade-offtheory But it leaves open the question of why they do not act more stronglyThe capital structure literature has long been interested in dynamics and

capital structure adjustment paths In recent years this interest has increas-ingly resulted in explicit dynamic optimizing models such as those ofGoldstein Ju and Leland (2001) Hennessy and Whited (2005) andStrebulaev (2007) These models commonly make specialized assumptionsleading some to question their generalityWe argue that both fixed transactions costs and variable transactions costs

are plausible and realistic in models that pin down firm size The simultaneouspresence of both is important Due to fixed costs firms only rebalance inter-mittently Due to variable adjustment costs the extent of the adjustment isreduced to the point at which an extra unit to adjustment is balanced againstthe extra unit transactions costs This means that both leverage increases andleverage reductions will only go part way toward a static optimumThe key point is that with both fixed and variable transactions costs we

should expect to only see intermittent leverage adjustments When leverage isadjusted it should ldquounder adjustrdquo when compared to a frictionless modelFurther discussion of transactions costs in a capital structure setting isprovided by Leary and Roberts (2005)

9 Conclusion

The connection between corporate profits and capital structure has beenvery influential in the assessment of the trade-off theory The conventionalinterpretation of the evidence has pushed the literature away from the trade-off theory and toward much more complex dynamic models and ideas As aresult it is important to make sure that the evidence has been correctlyinterpreted We show that the problem is with the interpretation of thetest results and a failure to take into account the costs associated with re-sponding to a profit shock

PROFITSndashLEVERAGE PUZZLEREVISITED 1447

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

We document that more profitable firms really do borrow more (not less)They tend to repurchase equity They experience an increase in both thebook value of equity and the market value of equity Less profitable firmsreally do tend to reduce their debt and to issue equityFirm size and market conditions also matter Larger firms tend to be more

active in the debt markets while smaller firms tend to be relatively moreactive in the equity markets During good times there is more use ofexternal financingThe usual profitsndashleverage puzzle result is primarily driven by the increase

in equity that is experienced by the more profitable firms Accordingly thepuzzle should be restated as asking why do firms not take sufficiently largeoffsetting actions to fully undo the change in equity What limits the mag-nitudes of the typical leverage response to profit shocksIn a frictionless model the partial response appears puzzling But there is

good empirical reason to believe that rebalancing entails both fixed andvariable costs and that firm size matters The rebalancing costs can befully avoided by doing nothing Accordingly the firm must decide whetherany given shock is big enough to be worth responding to If it is then thefirm must decide how big a response is called for These technical conditionsare known as ldquovalue matchingrdquo and ldquosmooth pastingrdquo Thus optimizationimplies that some shocks will be ignored Even if the shock is not ignoredthe optimal response will only partially undo the shock The magnitude ofthe leverage response must balance the marginal cost and the marginalbenefit of an extra unit of leverage Since the marginal cost of adjustingleverage is strictly positive the adjustment toward that static leverageoptimum will only go part way This is true both for leverage increasesand for leverage reductionsOne might conjecture that the variable costs would be time varying and

that such costs might have declined over time Since we do not have a fullaccounting of all of the variable costs it is of course impossible to be sureHowever there is some evidence suggesting that at least the directunderwriting fees are remarkably stable over time10

Surely nobody ought to be surprised that more profitable firms are moreinclined to repurchase equity while unprofitable firms tend not to do so Thefact that more profitable firms typically experience an increase in the value of

10 For IPOs Jay Ritter provides important summary evidence as httpbearwarringtonufleduritterIPOs2013Underwritingpdf He finds that for medium-sized deals a remark-

ably large fraction of the deals have a 7 fee year after year Whether this is due tobehavioral factors or some sort of a coordination device is not clear The extent towhich similar intertemporal stability carries over to other forms of external financemerits further study

1448 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

equity is equally natural Were it not for the contrary claims in the literature(ie the profitsndashleverage puzzle) we would argue that nobody ought to findit surprising that more profitable firms are more inclined to issue debt Thefact that leverage adjustments typically only partially offset profit shocksshould also sound quite natural These facts all fit together easily within thetrade-off theory once we take account of both fixed and variable costs ofactively adjusting capital structure

Appendix

Table AI Market leverage regressions

The table presents estimates of the regressions of market leverage on firm characteristics

The sample comes from the annual Compustat files during the period 1971ndash2009 Financialfirms are excluded The leverage is estimated as the ratio of debt over debt plus marketequity The explanatory variables Profitabilityt1

MB

t1

Tangibilityt1 and LnethAssetsTHORNt1are described in Table 1 IndMedianLevt1 is estimated as the median market leverage of allother firms in the same industry The industry is defined at the level of the four-digit SICcode All specifications include the year fixed effects The specifications in column (2)

additionally include the firm fixed effects We report t-statistics where the standard errorsare clustered at the firm level in parentheses

aSignificant at the 1 level

Quantile regressions

(1) (2) (3) (4) (5)

Profitabilityt1 013a 013a 007a 011a 020a

(275) (262) (349) (425) (538)

IndLevt1 046a 020a 032a 058a 063a

(506) (239) (1299) (1630) (1042)MB

t1

003a 001a 001a 002a 004a

(509) (303) (743) (709) (534)

Tangibilityt1 013a 018a 013a 014a 010a

(191) (181) (762) (571) (232)

LnethAssetsTHORNt1 001a 005a 001a 001a lt001a

(112) (258) (682) (342) (29)

Constant 010a 011a 005a 005a 029a

(171) (98) (173) (118) (398)

Year FE Yes Yes Yes Yes Yes

Firm FE No Yes No No No

Regression type OLS FE 25th tile Median 75thtile

Clustered SE Yes Yes No No No

R2 030 067

Pseudo R2 016 023 0201

N 158578 158578 158578 158578 158578

PROFITSndashLEVERAGE PUZZLEREVISITED 1449

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

Table AII Profitability sorts for firms issuing large amounts

Percentage of firms

Iss Ret Iss Iss Rep Iss Iss IssD IssE

D D NetD E E NetE DampE RepE RetD

(1) (2) (3) (4) (5) (6) (7) (8) (9)

Panel A 10 cutoff to define issuing or retiring activity

Allfirm

s

Low 205 150 134 260 05 257 45 00 23

2 207 175 103 77 06 75 11 01 10

3 243 196 101 50 08 47 08 01 09

4 252 186 113 50 13 48 08 02 08

High 227 161 118 60 27 57 07 04 09

t Highfrac14Low 70 38 62 755 226 768 320 90 152

Smallfirm

s Low 230 132 162 380 05 375 82 00 32

2 228 143 165 334 04 333 65 00 26

3 211 146 147 216 03 214 38 00 23

4 190 154 120 126 04 124 15 01 20

High 194 162 114 99 09 97 12 01 15

t Highfrac14Low 42 40 68 342 20 340 169 07 131

Largefirm

s Low 231 200 82 31 06 29 08 01 04

2 231 187 67 17 06 14 03 01 02

3 234 177 77 17 09 14 03 02 03

4 218 158 75 16 17 14 02 03 02

High 218 135 99 19 41 15 03 07 02

t Highfrac14Low 19 106 36 46 135 57 39 58 23

Percentage of firms

Iss Ret Iss Iss Rep Iss Both Iss D Iss E

D D Net D E E Net E DampE Rep E Ret D

(1) (2) (3) (4) (5) (6) (7) (8) (9)

Panel B 15 cutoff to define issuing or retiring activity

Allfirm

s

Low 158 101 100 230 01 228 27 00 15

2 142 115 63 57 01 56 05 00 06

3 165 128 57 32 01 31 03 00 05

4 173 118 65 35 01 34 03 00 05

High 160 102 70 46 04 44 03 00 05

t Highfrac14Low 08 06 141 738 87 746 277 24 131

Smallfirm

s Low 195 95 132 354 01 350 56 00 24

2 185 96 131 303 01 301 44 00 18

3 160 96 104 179 00 176 18 00 15

4 134 99 78 99 01 98 08 00 12

High 133 95 73 82 01 80 06 00 10

t Highfrac14Low 80 00 95 345 12 344 149 12 52

Largefirm

s Low 145 123 42 14 01 12 03 00 01

2 137 114 32 07 01 06 01 00 01

3 142 104 35 06 01 05 00 00 01

4 130 91 36 08 03 07 01 01 01

High 134 81 48 09 06 07 01 01 01

t Highfrac14Low 19 85 17 26 54 33 34 17 11

1450 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

Table AIV Debt and equity changesmdashwithout fixed effects

The sample contains nonfinancial firms listed on the annual Compustat files for the period from 1971

to 2009 The table presents estimates from regressions of change in debt (D) change in the book

value of equity (BVE) change in the market value of equity (MVE) and net equity issuances

(EqyIss) The explanatory variables include current and lagged change in operating income before

depreciation (Profits and Profitst1) and lagged change in book value of assets (Assetst1) The

regressions include year indicator variables The reported t-statistics are corrected for clustering at the

firm level

aMean significant at the 1 level

bMean significant at the 5 level

cMean significant at the 10 level

D BVE MVE EqyIss

(1) (2) (3) (4) (5) (6) (7) (8)

Profitst1 025a 024a 023a 038a 007 066a 007a 009a

(40) (32) (36) (39) (06) (29) (42) (51)Profits 051b 082a 149a 004a

(25) (33) (49) (33)Assetst1 002 005c 024a 001b

(08) (17) (34) (23)Constant 93 176a 45 71 1552a 1704a 130a 140a

(12) (28) (07) (09) (49) (49) (45) (45)Year FE Yes Yes Yes Yes Yes Yes Yes YesR2 001 005 001 011 001 005 002 002N 162056 162056 162130 162130 157550 157550 162154 162154

Table AIII Median profitability sorts

The sample contains nonfinancial firms listed on the annual Compustat files for the period from 1971

to 2009 The table presents median debt and equity median changes in debt and equity and median

equity issuances (EqyIss) in millions of dollars for all firms sorted on profitability and for profitability

sorts within the smallest and largest firms The sorts are done annually

D D BVE BVE MVE MVE EqyIss Assets(1) (2) (3) (4) (5) (6) (7) (8)

Allfirm

s

Low 21 00 75 17 256 20 01 1922 186 02 449 18 598 19 00 11033 816 06 1274 03 1637 05 00 32244 787 02 1541 18 2488 03 00 3556High 273 00 1361 66 3300 14 00 2578

Smallfirm

s Low 05 00 13 07 140 14 06 402 06 00 25 08 114 12 01 563 09 00 32 05 85 07 00 674 12 00 39 01 68 03 00 79High 10 00 51 04 98 01 00 90

Largefirm

s Low 11485 381 10269 371 13227 202 03 333902 13051 393 12786 140 17571 169 05 392463 11554 287 13429 24 21623 305 00 370314 9963 209 15052 194 29424 308 00 36705High 7361 00 16380 613 49142 439 00 34976

PROFITSndashLEVERAGE PUZZLEREVISITED 1451

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

References

Auerbach A J (1985) Real determinants of corporate leverage in B M Friedman (ed)

Corporate Capital Structures in the United States University of Chicago Press Chicago

IllinoisBertrand M and Mullainathan S (2003) Enjoying the quiet life corporate governance and

managerial preferences Journal of Political Economy 111 1043ndash1075Booth L Varouj A Demirguc-Kunt A and Maksimovic V (2001) Capital structures in

developing countries Journal of Finance 56 87ndash130Brennan M J and Schwartz E S (1984) Optimal financial policy and firm valuation

Journal of Finance 39 593ndash607Cameron A C and Trivedi P K (2010) Microeconometrics using Stata Stata Press TexasConstantinides G M (1979) Multiperiod consumption and investment behavior with

convex transactions costs Management Science 25 1127ndash1137Constantinides G M and Richard S F (1978) Existence of optimal simple policies for

discounted-cost inventory and cash management in continuous time Operations Research

26 620ndash636DeAngelo H DeAngelo L and Whited T M (2011) Capital structure dynamics and

transitory debt Journal of Financial Economics 99 235ndash261Dixit A K (1993) The Art of Smooth Pasting Routledge London and New YorkFama E and French K R (2002) Testing trade-off and pecking order predictions about

dividends and debt Review of Financial Studies 15 1ndash33Faulkender M and Petersen M A (2006) Does the source of capital affect capital struc-

ture Review of Financial Studies 19 45ndash79Faulkender M Flannery M J Hankins K W and Smith J M (2012) Cash Flows and

Leverage Adjustments Journal of Financial Economics 103 632ndash646Fischer E Heinkel R and Zechner J (1989) Dynamic capital structure choice theory

and tests Journal of Finance 44 19ndash40Frank M Z and Goyal V K (2008) Trade-off and pecking order theories of

debt in B E Eckbo (ed) Handbook of Corporate Finance Empirical Corporate Finance

Vol 2 Handbook of Finance Series Chapter 12 ElsevierNorth-Holland Amsterdam pp

135ndash202

Frank M Z and Goyal V K (2009) Capital structure decisions which factors are reliably

important Financial Management 38 1ndash37

Garvey G T and Hanka G (1999) Capital structure and corporate control the effect of

antitakeover statutes on firm leverage Journal of Finance 54 519ndash546

Goldstein R Ju N and Leland H (2001) An EBIT-based model of dynamic capital

structure Journal of Business 74 483ndash512

Graham J R and Tucker A L (2006) Tax shelters and corporate debt policy Journal of

Financial Economics 81 563ndash594

Hackbarth D and Mauer D C (2012) Optimal priority structure capital structure and

investment Review of Financial Studies 25 747ndash796

Halling M Yu J and Zechner J (2012) Leverage dynamics over the business cycle

Working paper University of Utah University of New South Wales and Vienna

University of Economics and Business AdministrationHennessy C A and Whited T A (2005) Debt dynamics Journal of Finance 60 1129ndash1165

Huang R and Ritter J R (2009) Testing theories of capital structure and estimating the

speed of adjustment Journal of Financial and Qualitative Analysis 44 237ndash271

1452 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

Hovakimian A Opler T and Titman S (2001) The debt-equity choice Journal of

Financial and Quantitative Analysis 36 1ndash24Jung K Kim Y and Stulz R M (1996) Timing investment opportunities managerial

discretion and the security issue decisions Journal of Financial Economics 42 159ndash185Kane A Marcus A J and McDonald R L (1984) How big is the tax advantage to

debt Journal of Finance 39 841ndash853Kayhan A and Titman S (2007) Firmsrsquo histories and their capital structures Journal of

Financial Economics 83 1ndash32Leary M T (2009) Bank loan supply lender choice and corporate capital structure

Journal of Finance 64 1143ndash1185Leary M T and Roberts M R (2005) Do firms rebalance their capital structures

Journal of Finance 60 2575ndash2619Leland H E (1994) Corporate debt value bond covenants and optimal capital structure

Journal of finance 49 1213ndash1252Lemmon M L Roberts M R and Zender J F (2008) Back to the beginning persist-

ence and the cross-section of corporate capital structure Journal of Finance 63 1575ndash1608Long M S and Malitz I B (1985) The investment-financing nexus some empirical

evidence Midland Corporate Finance Journal 3 53ndash59Mackie-Mason J K (1990) Do taxes affect corporate financing decisions Journal of

Finance 45 1471ndash1493Morellec E Nikolov B and Schurhoff N (2012) Corporate governance and capital

structure dynamics Journal of Finance 67 803ndash848Myers S C (1984) The capital structure puzzle Journal of Finance 39 575ndash592Myers S C (1993) Still searching for optimal capital structure Journal of Applied

Corporate Finance 61 4ndash14Rajan R and Zingales L (1995) What do we know about capital structure some evidence

from international data Journal of Finance 50 1421ndash1460Strebulaev I A (2007) Do tests of capital structure theory mean what they say Journal of

Finance 62 1747ndash1787

Strebulaev I A and Whited T M (2012) Dynamic Models and Structural Estimation in

Corporate Finance Now Publisher Boston and Delft

Titman S and Wessels R (1988) The determinants of capital structure choice Journal of

Finance 43 1ndash21

Welch I (2004) Capital structure and stock returns Journal of Political Economy 112

106ndash131

Welch I (2011) Two common problems in capital structure research the financial-debt-

to-asset ratio and issuing activity versus leverage changes International Review of Finance

11 1ndash17

PROFITSndashLEVERAGE PUZZLEREVISITED 1453

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

shows that although most firms issue little debt or equity in a given year theaverages are large In other words when firms actually enter debt and equitymarkets they intervene massivelyThe mean constant dollar debt issue is $165 million (the median is $2

million) In unreported tables we find that the average debt issue is about81 of assets (the median is 18) About 38 of the firms issue no debt8 issue between 0 and 1 of the value of their assets as debt another

Table I Data description

Data sources The sample comes from the annual Compustat files The sample period is

1971ndash2009 We exclude financial companies (SIC 6000-6999) firms involved in majormergers (Compustat footnote code AB) firms reporting financial data in currencies otherthan the US dollar and firms with missing data on the key variables All financial variables

are deflated to year 2000 using the GDP deflator N is the number of observations SD isthe standard deviation D BVE MVE Assets Debt issuance Equity issuance Debt re-payment Equity repayment are all measured in $ millions The ratio variables are trimmed

at the 1 level in both tails of the distribution

Variable definitions Debt (D)frac14 long-term debt (dltt)thorn short-term debt (dlc) book equity(BVE)frac14 common shareholder equity (ceq) market equity (MVE)frac14 no of outstanding

shares (csho) closing price ( prcc f ) assetsfrac14 book assets (at) debt issuancefrac14 issuance oflong-term debt (Maxethdltis 0THORN)thorn increase in current debt (Maxethdlcch 0THORN) equity issu-ancefrac14 sale of common stock (Maxethsstk 0THORN) debt repaymentfrac14 reduction of long-termdebt (Maxethdltr 0THORN)thorn decrease in current debt (Minethdlcch 0THORN) equity repurchasefrac14 purchase

of common stock (Maxethprstkc 0THORN) book leveragefrac14D(DthornBVE) market leveragefrac14D(DthornMVE) profitabilityfrac14EBITDA (oibdp)assets market-to-book ratio M

B frac14Marketvalue of assets (MVA)assets where MVAfrac14DthornMVEthorn preferred-liq value (pstkl)

deferred taxes (txditc) tangibilityfrac14 net property plant and equipment (ppent)assets

Variable N Mean SD

Distribution

10th 50th 90th

Debt (D) 179021 653 4592 0 24 1048

Book equity (BVE) 179021 815 4122 3 70 1373

Market equity (MVE) 179021 1784 9589 9 118 2698

Assets 179021 2191 11530 10 153 3575

Debt issuance 179021 165 1335 0 2 245

Equity issuance 179021 26 186 0 0 40

Debt repayment 179021 132 1123 0 3 181

Equity repurchase 179021 24 257 0 0 9

Book leverage 179021 036 034 000 031 072

Market leverage 179021 027 025 000 020 066

Profitability 179021 005 027 018 011 023MB 179021 165 206 052 102 324

Tangibility 179021 031 024 005 026 070

PROFITSndashLEVERAGE PUZZLEREVISITED 1425

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

16 issue between 1 and 5 of the value of their assets as debt and theremaining 38 issue debt in excess of 5 of the value of their assetsThe mean constant dollar equity issue is about $26 million (the median is

about $04 million) As a fraction of assets the mean and median equityissues are about 7 and 02 respectively About 33 of the firms issue noequity 34 of the firms issue between 0 and 1 of the value of theirassets as equity another 14 issue between 1 and 5 of the value of theirassets as equity and the remaining 19 of the firms issue equity that is inexcess of 5 of the value of their assets Average debt repayments are largerthan equity repurchases This perhaps reflects the finite maturity of debt andits contractual repayment The median firm does not repurchase equityWe construct both book and market leverage ratios Book leverage is

defined as debt over debt plus book equity Market leverage is defined asdebt over debt plus market equity8 The median book leverage is 036 (theaverage is 031) The median market leverage is 027 (the average is 020)Profitability is defined as the ratio of operating income before depreci-

ation to assets While the average firm is profitable (the ratio of EBITDA toassets is 005) the median firm is even more profitable (with a profitabilityratio of 011) But the sample also includes a large number of unprofitablefirms as the 10th percentile is 018 The table also reports descriptive stat-istics on the market-to-book ratio and the tangibility ratio The market-to-book ratio (MB) defined as the ratio of the market value of assets to bookassets averages at about 165 Tangibility defined as the ratio of netproperty plant and equipment to assets averages at about 31

4 Leverage Ratio Regressions

The previous literature focuses on estimates obtained using leverage ratiosHence we start with a similar estimation to check whether our results matchthose of previous studies Table II presents the results for book leverage Weobtain similar results for market leverage and they are presented inAppendix Table AI The regressions include leverage factors following theprevious capital structure literature (see eg Lemmon Roberts and Zender(2008) and Frank and Goyal (2009)) The factors include (i) profitability (ii)industry median leverage (iii) market-to-book assets ratio (iv) tangibility ofassets and (v) firm size Industry median leverage is constructed as themedian leverage of all other firms in the industry excluding the firm under

8 Welch (2011) stresses the idea that nonfinancial liabilities should not be implicitly mis-treated as if they were equity by paying excessively narrow attention to financial liabilitiesin a leverage ratio We have adopted his approach in the empirical work reported here

1426 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

consideration Firm size is defined as the natural log of assets Other factorsare defined in Section 3 We cluster standard errors by firm and estimatethese regressions both with and without fixed effectsIn Table II we run conventional cross-section leverage regressions Since

we employ the usual data it is not surprising that our results match thosereported in the existing capital structure literature Profitability has anegative sign in both the book leverage regressions and in the marketleverage regressions The coefficients on other factors largely match thosereported in earlier studies Firms operating in industries with high leverage

Table II Leverage and profitability

The table presents parameter estimates for book leverage regressions The sample consists

of nonfinancial companies in the annual Compustat files during the period 1971ndash2009 Theleverage is estimated as the ratio of debt over debt plus book equity The explanatoryvariables Profitabilityt1

MB

t1

Tangibilityt1 and LnethAssetsTHORNt1 are described in

Table I IndLevt1 is estimated as the median book leverage of all other firms in thesame industry The industry is defined at the level of the four-digit SIC code All specifi-cations include the year fixed effects The specifications in column (2) additionally include

the firm fixed effects We report t-statistics where the standard errors are clustered at thefirm level in parentheses

aSignificant at the 1 level

Quantile regressions

(1) (2) (3) (4) (5)

Profitabilityt1 026a 020a 011a 017a 041a

(293) (192) (378) (518) (1015)

IndLevt1 042a 012a 030a 052a 052a

(369) (107) (926) (1252) (950)MB

t1

002a 001a 001a 002a 002a

(203) (102) (392) (426) (418)

Tangibilityt1 018a 025a 024a 019a 012a

(206) (154) (964) (580) (276)

LnethAssetsTHORNt1 001a 002a 002a 002a 001a

(114) (76) (899) (478) (96)

Constant 012a 010a 012a 002a 029a

(156) (60) (284) (27) (387)

Year FE Yes Yes Yes Yes Yes

Firm FE No Yes No No No

Regression type OLS FE 25th tile Median 75thtile

Clustered SE Yes Yes No No No

R2 013 054

Pseudo R2 016 015 009

N 158824 158824 158824 158824 158824

PROFITSndashLEVERAGE PUZZLEREVISITED 1427

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

tend to have high leverage A higher market-to-book ratio is associated withlower leverage Larger firms are typically more highly levered The coeffi-cients on tangibility are positive The conclusion from Table II is clear Theleverage regression results match those reported in previous studiesAt times one may be concerned that in a regression what is being

estimated is a conditional mean while some parts of the distributionmight behave differently from other parts If so then ldquoaccidentsrdquo likesample selection issues or asymmetries in the underlying distributionsmight play a greater role than is usually recognized In other words thesource of identification could be misinterpreted To guard against this po-tential problem we estimate the same model using quantile regressions9

Columns 3 4 and 5 report what happens at the 25th 50th and 75th per-centiles The basic model results are robust in terms of both the sign and thestatistical significanceThe impact of profits does appear to be somewhat stronger among the

high leverage firms For instance in a book leverage regression the coeffi-cient on profits for the first quartile is 011 with a t-ratio of 378 For thethird quartile the coefficient on profits is 041 with a t-ratio of 1015This difference is curious and might deserve further study However the factthat in each case we get the negative sign and statistical significance is suf-ficient for the purposes of the present article So the results on profits arerather robust across the distributionTable III sorts firms into those that are actively adjusting their leverage

and those that are passive In the presence of both fixed and variable trans-acts costs this distinction is important We expect the coefficient on profit-ability in leverage ratio regressions to be more negative in firms that arepassive and less negative in firms that are actively adjustingWe define active firms as those that are either issuing net debt or net

equity in excess of 5 of their assets Passive firms are those that neitherissue debt nor equity in nontrivial amounts (ie greater than 5 of theirassets) Consistent with our expectations we find in Table III that the coef-ficients on profits are significantly more negative for passive firms than theyare for active firms With firm and year fixed effects the coefficient onprofits is 008 for active firms and 017 for the passive firms The twocoefficients are significantly different from each other (plt 001) InTable III the standard control variables are used They have the usualsigns and for the most part do not differ much between the active and thepassive firms

9 Cameron and Trivedi (2010) provide an extensive discussion of quantile regressions inStata

1428 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

5 Financing Activity

51 FREQUENCY OF DEBT AND EQUITY ISSUES

Table IV reports the percentage of firms issuing or repurchasing debt orequity for annual sorts based on lagged profitability In Panel A weemploy the conventional 5 cutoff rule to exclude minor fluctuations anddefine firms ldquoissuing debtrdquo as those that issue debt (both short-term andlong-term) in excess of 5 of the value of their assets Other decisions aresimilarly defined using a 5 cutoff Appendix Table AII examines the dis-tributional properties across profitability (and size) quintiles for ldquolargerdquo is-suances by increasing the 5 cutoff to higher levels such as 10 and 15How does security issuance behavior vary with firm profitability To

answer this question Table IV sorts firms by profitability and reports the

Table III Leverage and profitability for active and passive firms

The table presents estimates of leverage on firm characteristics for the subsamples of firms

that are active in either debt or equity markets Active firms are defined as those that issueeither net debt in excess of 5 of the value of assets or issue net equity in excess of 5 ofthe value of their assets The passive firms do not issue significant (more than 5) of either

debt or equity The sample comes from the annual Compustat files during the period 1971ndash2009 Financial firms are excluded The variables are described in Tables I and II Wereport t-statistics where the standard errors are clustered at the firm level in parentheses

aSignificant at the 1 level

Active Passive

(1) (2) (3) (4)

Profitabilityt1 006a 008a 022a 017a

(126) (148) (254) (188)

IndLevt1 044a 019a 041a 013a

(424) (149) (326) (127)MB

t1

002a 001a 003a 001a

(362) (196) (327) (172)

Tangibilityt1 010a 014a 014a 019a

(146) (118) (145) (130)

LnethAssetsTHORNt1 001a 004a 001a 004a

(121) (179) (147) (169)

Constant 012a 008a 003a 007a

(120) (57) (45) (42)

Year FE Yes Yes Yes Yes

Firm FE No Yes No Yes

R2 034 064 033 075

N 43395 43395 58943 58943

PROFITSndashLEVERAGE PUZZLEREVISITED 1429

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

percentage of firms issuing or repurchasing debt or equity The table alsoreports test statistics from tests that examine whether the percentage differ-ences between the high- and low-profit firms are significantly different fromeach other The results in Column (1) show that the likelihood of issuing debt

Table IV Profitability sorts for debt and equity issuers

The sample contains nonfinancial firms listed on the annual Compustat files for the period

from 1971 to 2009 The table presents the percentage of firms issuing and retiring (orrepurchasing) debt and equity The firms are classified as ldquoissuing debtrdquo if they issuedebt in excess of 5 of the value of their assets as ldquoissuing equityrdquo if they issue equity

in excess of 5 of the value of their assets as ldquoretiring debtrdquo if they retire debt in excess of5 of the value of their assets and as ldquorepurchasing equityrdquo if they repurchase equity inexcess of 5 of the value of their assets In addition we also report net debt issuers which

are firms that issue net debt over 5 of the value of their assets and net equity issuerswhich are firms that issue net equity in excess of 5 of the value of their assets Weannually sort firms on lagged profitability and report the percentage of firms in each ofthese categories The bottom part of each panel reports the percentage of firms issuing or

retiring securities by profitability within the smallest and largest asset quintiles The tablealso reports results from t-test for the difference in percentage of firms issuing or retiringsecurities in the lowest and the highest profit quintile

Percentage of firms

Iss Ret Iss Iss Rep Iss Both IssD IssE

D D NetD E E NetE DampE RepE RetD

(1) (2) (3) (4) (5) (6) (7) (8) (9)

Allfirm

s

Low 276 245 186 301 13 297 71 01 38

2 319 299 181 117 21 112 27 02 20

3 385 344 194 86 29 80 21 05 15

4 386 330 213 81 43 75 20 07 14

High 329 273 198 92 82 81 17 13 14

t Highfrac14Low 150 84 39 723 414 763 358 176 200

Smallfirm

s Low 275 204 201 411 11 405 109 01 47

2 291 233 213 379 10 378 99 01 45

3 281 246 204 265 14 261 66 01 41

4 275 272 186 162 15 159 34 02 35

High 275 281 178 136 27 131 25 02 26

t Highfrac14Low 01 84 28 316 55 318 171 18 55

Largefirm

s

Low 414 363 177 87 22 79 28 03 11

2 449 379 180 62 30 57 17 07 09

3 439 360 186 56 41 48 16 09 07

4 403 320 194 46 67 39 13 12 06

High 374 264 207 50 129 35 10 28 05

t Highfrac14Low 50 130 45 89 244 116 77 120 47

1430 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

is mostly independent of firm profitability In the aggregate there is a smalldifference but it is sensitive to sorting on firm size Similarly in Column (2)there is only a weak relation between the profitability and the likelihood thatthe firm retires debt To a first approximation the fraction of firms issuing orrepurchasing debt is largely independent of profits Large firms are somewhatmore likely than small firms to issue debt and also to retire debtUnlike debt issues the probability of issuing equity is strongly related to

profitability as can be seen in Column (4) Low-profit firms are much morelikely to issue equity than are high-profit firms In the lowest profit quintile301 of the firms issued equity In the highest profit quintile it was only92 The lower panel reports separate sorts for small and large firms Similarpatterns are found for both groups of firms that is the likelihood of equityissuance declines as profitability increases The differences between the quin-tiles are quite stark for the group of small firms All quintiles of small firmsissue equity much more frequently than the corresponding large firm quintilesHigh-profit firms are much more likely to repurchase equity as shown in

Column (5) In the overall sample only 13 of low-profit firmsrepurchased equity compared with 82 of the high-profit firms Whenwe separate the small and large firms again a difference emerges Ingeneral small firms do not repurchase much equity while a greaterfraction of the large firms do soGiven these facts it is natural that low profitability firms are much more

likely to be net issuers of equity than are high profitability firms as shown inColumn (6) Among the firms in the lowest quintiles of profits roughly 30of firms issue net equity that exceeds 5 percent of their assets In contrastfor the most profitable quintile only about 8 issue net equityFirm size is an important variable in the recent literaturemdashit is often used

as a proxy for access to capital markets (as in Faulkender and Petersen(2006) and Leary (2009)) Small firms are bank-dependent risky and infor-mationally opaque They have restricted access to public debt markets andconsequently face more severe supply constraints in their ability to issuedebt Thus we expect small firms to be more sluggish in adjusting theirdebt and equity in response to shocks to profitability Large firms incontrast have much easier access to public debt markets and they facefewer obstacles in accessing securities marketsTo examine how size and profitability interact the bottom part of Table IV

examines profitability sorts for the smallest and largest firms We first sortfirms annually by firm size and then within each size quintile we sort them onprofitability We do these two-way sorts to ensure that we have similarnumber of firms in profitability subgroups for both small and large firmsamples Unconditionally small firms are generally less profitable while

PROFITSndashLEVERAGE PUZZLEREVISITED 1431

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

large firms are relatively more profitable The two-way sorts allow us to havea uniform distribution of firms in various profitability quintiles for both smalland large firm samples The bottom part of Table IV reports results for thesmallest and the largest size quintiles Among the small firms there is littlerelation between profitability and the likelihood of issuing debt Howeverlarger firms exhibit a small increase in the likelihood of net issuance of debtwith increasing profitabilityIn contrast to debt the effects of profitability on equity issuance and

repurchases are much more consistent across size quintiles As profitabilityincreases firms are generally less likely to issue equity and more likely torepurchase it Across the two extreme size groups we note a striking differ-ence between the proportion of firms issuing equity and those issuing debtAmong low-profit small firms almost 41 are equity issuers whereas onlyabout 9 of low-profit large firms are equity issuers However regardless ofsize we note a monotonic reduction in the likelihood of issuing equity asprofitability increasesColumn (7) reports the fraction of firms that issue both debt and equity in

excess of 5 of assets both as a function of profits in the upper panel andas a function of firm size and profits in the lower panel In the upper panelwe see that low-profit firms are much more likely to issue both debt andequity (71) while the high-profit firms are less likely to do so (17) Thelower panel shows that small firms are much more likely to be issuers of bothin contrast to large firms In both firm size categories the profit pattern is thesame with low profits being more frequent issuers of both simultaneouslyImportantly Column (8) shows that the likelihood of issuing debt and

simultaneously repurchasing equity increases with profitability Converselyin Column (9) we find that the likelihood of doing the reverse that is issuingequity and retiring debt declines with profitability The effects of sorting onfirm size and firm profitability mirror those reported for all firms Overallthe results show that low-profit firms are less likely to issue debt and repur-chase equity they are instead more likely to issue equity and retire debtFirms with high profitability exhibit the reverse patternIn Appendix Table AII we present results from profitability (and size)

quintiles for large issuances defined using the 10 cutoff (in Panel A) andthe 15 cutoff (in Panel B) The results from these panels are largely con-sistent with those described above

52 MAGNITUDES OF ISSUANCES AND EQUITY CHANGES

In the previous section we considered the probability of having a nontriviallevel of debt or equity activity The next question is how large are the dollar

1432 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

values involved In Table V we sort firms according to profits and thentabulate the levels and changes in both debt and equity We do this firstfor all firms and then for small versus large firmsFor the analysis on all firms we observe in Column (1) that debt peaks at

the middle of the distribution This is because firms with medium profitabil-ity are also the largest as seen in Column (8) which reports average assetvalues for different profitability quintiles When we sort first by firm size andthen examine firms with different levels of profitability within size quintileswe find that for small firms debt is roughly independent of profits while forlarge firms less profitable firms have a higher level of debtAs expected Columns (3) and (5) show that more profitable firms have

higher equity values Columns (2) and (7) consider the relationship betweenissuances of debt and equity and firm profitability Debt issuances are sig-nificantly larger for more profitable firms In contrast equity issuances arelarger for low-profit firms Small low-profit firms issue substantially lessequity compared with small high-profit firms Among large firms the low-profit ones tend to issue equity while the high-profit ones tend to repurchaseequity The fact that more profitable firms issue debt and repurchase equitywhile the least profitable firms retire debt and issue equity is consistent withthe predicted relation between profitability and financing decisions under thetrade-off theoryWe also examine two-way sorts by size and profitability and report results

for the smallest and largest quintile of firms For the smallest quintile offirms the change in debt is largely unrelated to profits But for large firmsthere is a positive relation between profits and debt issuances High-profitfirms have a big positive change in debt Low-profit large firms have anegative change in debtColumns (4) and (6) provide an explanation of why the leverage ratio

regression results contradict those from the basic profitability sorts pre-sented here As we can see profitability indirectly affects leverage ratiosby increasing equity values Changes in both the book value of equity andthe market value of equity are positive and large for highly profitable firmsIn contrast these changes are negative for less profitable firms This evidenceillustrates an important issue concerning the use of leverage ratios Suchratios are often interpreted as essentially reflecting the use of debt by thefirm This interpretation while common is empirically misleadingFor the typical firm the change in the value of equity is larger than the

change in debt For example in the third profit quintile for large firmscolumn (7) the mean equity issue is just $9 million but the change in themarket value of equity is $242 million At the same time the mean change indebt is $27 million This suggests that a fair bit of the observed variation in

PROFITSndashLEVERAGE PUZZLEREVISITED 1433

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

the leverage ratios is primarily driven by the changes in the market value ofequity in the denominator rather than by the changes in debt in the numer-ator Since equity issues are often small this implies that the variation in theleverage ratio is primarily driven by internal operations rather than byexternal financing actions This again points to the fact that leverageratios can provide a misleading account of actual patterns in the dataThere is always a potential concern that the averages may be misleading

due to the impact of outliers To address this concern in AppendixTable AIII we present median values of the profitability sorts Theseresults generally reinforce the findings in Table VTable VI again considers the magnitudes of financing activity but this

time the issues are scaled by a measure of firm size That way it is easier to

Table V Magnitude of financing activity

The sample contains nonfinancial firms listed on the annual Compustat files for the period

from 1971 to 2009 The table presents average debt and equity changes in debt and equityand equity issuances (in millions of dollars) for firms sorted on profitability within sizeclasses The table reports information for all firms sorted on profitability and for profit-

ability sorts within the smallest and largest firms The sorts are done annually The tablealso reports results from t-test for the difference in percentage of firms issuing or retiringsecurities in the lowest and the highest profit quintile

D D BVE BVE MVE MVE EqyIss Assets

(1) (2) (3) (4) (5) (6) (7) (8)

Allfirm

s

Low 83 5 78 10 120 5 7 234

2 605 13 509 24 721 8 17 1691

3 1227 9 1123 22 1834 3 9 3551

4 833 23 1166 7 2408 74 9 2994

High 570 53 1171 71 3817 68 36 2581

t Highfrac14Low 31 9 36 9 38 2 16 36

Smallfirm

s Low 25 04 42 01 332 28 33 97

2 29 06 55 02 278 16 21 116

3 31 07 63 01 212 05 12 127

4 30 06 70 05 179 10 08 136

High 24 06 84 14 234 17 07 142

t Highfrac14Low 04 15 153 85 73 11 201 107

Largefirm

s Low 3929 55 2990 152 4671 11 63 10846

2 3826 19 3178 95 5122 66 11 10721

3 3074 27 3498 41 6173 242 9 9697

4 2623 62 3915 45 8473 94 57 9855

High 2180 193 4642 261 14924 202 148 10171

t Highfrac14Low 94 64 108 87 246 12 169 16

1434 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

see whether large firms behave disproportionately when compared withsmall firms Column (1) shows that low-profit firms have more debtas reported in many previous studies In Columns (4) and (6) we seethat low-profit firms have a major decline in both the book value ofequity and the market value of equity The high-profit firms have thereverse experienceMore interesting are the active decisions of the firms In Column (2) we see

that in contrast to the usual impression the low-profit firms are reducingdebt while the high-profit firms are increasing it In Column (7) we see thatthe low-profit firms are also reducing leverage by issuing equity The high-profit firms issue little equity In the lower panel similar results are foundwhen we separate small and large firms Equity issuing is particularly strong

Table VI Magnitude of financing activitymdashscaled levels and changes

The sample contains nonfinancial firms listed on the annual Compustat files for the period

from 1971 to 2009 The table presents average debt and equity changes in debt and equityand equity issuances scaled by assets for firms sorted on profitability within size classes Thetable reports information for all firms sorted on profitability and for profitability sorts

within the smallest and largest firms The sorts are done annually The table also reportsresults from t-test for the difference in percentage of firms issuing or retiring securities inthe lowest and the highest profit quintile

D D BVE BVE MVE MVE EqyIss

(1) (2) (3) (4) (5) (6) (7)

Allfirm

s

Low 386 122 12 871 3427 1868 217

2 285 58 407 119 940 157 28

3 293 12 425 14 742 03 12

4 269 26 461 02 955 27 10

High 192 15 551 47 1675 74 07

t Highfrac14Low 35 49 50 47 127 42 157

Smallfirm

s Low 545 154 841 2145 9688 9128 696

2 684 49 712 1915 5014 2318 288

3 331 253 141 969 3063 954 129

4 267 181 357 175 1673 221 55

High 194 05 543 26 1758 147 41

t Highfrac14Low 49 17 51 36 88 27 99

Largefirm

s Low 350 30 306 58 511 55 09

2 346 14 335 13 523 02 04

3 353 165 332 71 711 29 01

4 280 01 414 04 909 09 05

High 227 15 477 19 1643 27 14

t Highfrac14Low 371 112 453 181 570 57 221

PROFITSndashLEVERAGE PUZZLEREVISITED 1435

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

among the small low-profit firms that are also experiencing sharp reductionsin the value of equity In column (2) we see that both for the large and for thesmall firms net debt increases are primarily found in the highest quintile offirm profitability

6 Debt and Equity Regressions

So far we have demonstrated that in our data the conventional leverageregressions have the usual signs We then demonstrated that in simple sortsof the data strikingly different active financing patterns emerge The nexttask is to reconcile these differencesTable VII presents simple regressions predicting changes in debt book

equity and market equity as well as equity issuances Consistent with thesorts we control for changes in firm size and we also include firm and yeardummies There may be a concern about the speed of a firmrsquos responsesAccordingly we consider the lagged change in profits as the main case butwe also include the contemporaneous change in profits Whether the contem-poraneous change in profits is included or not does not in fact matter Wereport panel-robust standard errors adjusted for clustering at the firm levelThe results in Table VII support the findings in the sorts Firms that have

an increase in profits in 1 year increase debt both in the same year and in thenext year Firms that have an increase in profits experience an increase inboth the book value of equity and the market value of equity Importantlyhowever firms that experience and increase in profits reduce their equityissues All these effects are statistically significant (except for the marketvalue of equity in the column 5 specification) and seem intuitively reason-able In Appendix Table AIV similar regressions to those presented aboveare reported but without firm-fixed effects We find that empirically verylittle changes whether we include firm fixed effects or leave them out Thusthe targeting behavior to the extent that it happens does not account forwhat we are reportingThe next step is to bring the sorts and the simple regressions together into

a common setting We therefore include factors in addition to profitability(i) median industry leverage (ii) the market-to-book assets ratio (iii) tangi-bility of assets and (iv) firm size (measured by the log of assets) Rajan andZingales (1995) show that these factors are related to leverage in G7countries A number of studies have used these factors to estimateleverage targets Frank and Goyal (2009) show that these factors arerobustly related to leverage in the USA In these regressions we use indica-tors for the quintile that the firm is in for each factor

1436 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

Table VIII reports these results In Columns (1) and (2) we examine debtissues and again show that profits positively affect debt issuances The effectsare large and statistically significant at the 1 level Columns (3) and (4)which examine book equity show a strong effect of profits This effect is onlyslightly reduced by the inclusion of other factors Columns (5) and (6) illus-trate that the impact of profits on the change in market equity is also robustto the inclusion of the conventional factors Columns (7) and (8) examineequity issuances Here we again find that profits have a significant negativeimpact on equity issuances As before the results are robust to the inclusionof the conventional factorsIn unreported tables we estimate debt and equity changes using quantile

regressions In terms of the signs and significance of the profitability variable

Table VII Debt and equity changes

The sample contains nonfinancial firms listed on the annual Compustat files for the period

from 1971 to 2009 The table presents fixed effect estimates from regressions of changes indebt (D) changes in the book value of equity (BVE) changes in the market value ofequity (MVE) and net equity issuances (EqyIss) The explanatory variables include

current and lagged change in operating income before depreciation (Profits andProfitst1) and lagged change in the book value of assets (Assetst1) In addition tofixed firm effects the regressions include year indicator variables The reported t-statistics

are corrected for clustering at the firm level

aMean significant at the 1 level

bMean significant at the 5 level

cMean significant at the 10 level

D BVE MVE EqyIss

(1) (2) (3) (4) (5) (6) (7) (8)

Profitst1 022a 025a 015b 035a 004 060b 005b 007a

(33) (31) (25) (35) (03) (25) (25) (53)

Profits 050b 078a 140a 003a

(24) (32) (48) (34)

Assetst1 000 007b 025a 001b

(00) (21) (35) (25)

Constant 167 283a 34 51 1145a 1182a 605a 595a

(14) (30) (04) (05) (32) (29) (69) (74)

Year FE Yes Yes Yes Yes Yes Yes Yes Yes

Firm FE Yes Yes Yes Yes Yes Yes Yes Yes

R2 001 004 001 010 001 005 002 003

N 162056 162056 162130 162130 157550 157550 162154 162154

PROFITSndashLEVERAGE PUZZLEREVISITED 1437

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

the results are qualitatively similar to the OLS results reported in Table VIIIProfitability positively affects debt issuances and negatively affects equityissuances across various quantiles Overall the results in Table VIII showthat there is nothing that is special about using sorts or regressions to explaindebt and equity issues In either case we find that more profitable firms tendto increase their debt experience an increase in the value of equity andrepurchase shares Thus the control factors are not responsible for theusual rejection of the trade-off theory It would appear that what mattersis the size of the issuing activity relative to the organic increases in equityvalue due to profits In other words what matters in the leverage regressionseems to be coming from the presence of E in D

DthornE and not from D

Table VIII Debt and equity issuances profitability and leverage factor quintiles

The sample contains nonfinancial firms listed on the annual Compustat files for the period

from 1971 to 2009 The table presents estimates from regressions of change in debt (D)change in the book value of equity (BVE) change in the market value of equity (MVE)and net equity issuances (EquityIss) The explanatory variables include lagged profitability

quintiles and lagged leverage factor quintiles In addition to fixed firm effects the regres-sions include year indicator variables The reported t-statistics are corrected for clustering atthe firm level

aMean significant at the 1 level

bMean significant at the 5 level

cMean significant at the 10 level

D BVE MVE EqyIss

(1) (2) (3) (4) (5) (6) (7) (8)

ProfitabilityQint 214a 167a 283a 175a 80 429a 53a 55a

(70) (54) (66) (59) (08) (48) (55) (62)

IndLevQint 104a 43 99 20

(29) (15) (10) (10)

Ln(Assets)Qint52 17 1026a 53a

(16) (05) (77) (35)MB

Qint158a 372a 1155a 09

(45) (56) (106) (06)

TangibilityQint 05 177 436c 64a

(01) (15) (18) (31)

Constant 167 96 656b 235 4613a 8610a 113 570a

(08) (04) (26) (04) (61) (58) (11) (31)

Year FE Yes Yes Yes Yes Yes Yes Yes Yes

Firm FE Yes Yes Yes Yes Yes Yes Yes Yes

R2 0002 0002 0003 0004 0013 0014 0013 0014

N 155587 155037 155587 155037 155587 155037 155587 155037

1438 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

7 Are Financial Market Conditions Important

According to the trade-off theory capital structure is determined by con-sidering the impact of costs and benefits of debt The time variation inthe costs and benefits of debt imply that issuance decisions would alsovary over time Furthermore market conditions affect the transactioncosts of adjusting Adverse selection is a greater problem in a cold marketthan in a hot one Accordingly it may be easier to adjust in good marketconditions Related ideas have been developed by Huang and Ritter(2009) Faulkender et al (2012) and Halling Yu and Zechner (2012)Faulkender Flannery Hankins and Smithrsquos paper examines cross-sectionalvariation in adjustment speeds including market valuations andhow they may affect adjustment speeds It is natural to think that ingood market conditions it will be less costly to issue both debt and equityand adjusting leverage to respond to profitability shocks would be mucheasierTo test the importance of market conditions we require a definition of

good times and bad times Our empirical strategy is to estimate good timesversus bad times at the four-digit industry level We define an industry ashaving ldquogood timesrdquo if the median firm in that industry has a market-to-book ratio that is higher than the 67th percentile of the time-series distribu-tion of the industry median market-to-book ratios Conversely an industryis defined as having ldquobad timesrdquo if the median firm in the industry has amarket-to-book ratio that is lower than the 33rd percentile of the time-seriesdistribution of the industry median market-to-book ratiosPanel A of Table IX tabulates issuance activity for profitability sorts in

both good times and bad times As might be expected active debt and equityissuances are larger during good times There is also somewhat more activeswaps between debt and equity during good timesPanel B reports the issuance activity for the smallest and the largest firms

during the good times and bad times Within each size quintile firms aresorted on profitability This is a three-way sort As before we find that thereis much more active use of external markets during good times Small low-profit firms are more likely to issue equity in good times than in bad timesDebt issuances are significantly higher in good times Large high-profitfirms are significantly more likely to issue debt and repurchase equity ingood times than in bad timesPanel C reports the magnitudes of the financial variables rather than the

frequencies During bad times less profitable large firms retire substantialamounts of debt and they show a tendency to issue equity Small firms donot seem to engage in similar debt reduction activities However like the

PROFITSndashLEVERAGE PUZZLEREVISITED 1439

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

Table IX Debt and equity issues in good and bad times

The table reports the frequency and magnitude of financing activity for sorts on profits for subsamples

of firms in good and bad times An industry is defined as having ldquogood timesrdquo if the median firm in

that industry has a market-to-book ratio that is higher than the 67th percentile of the time-series

distribution of industry median market-to-book ratios Conversely an industry is defined as having

bad times if the median firm in that industry has a market-to-book ratio that is lower than the 33rd

percentile of the time-series distribution of industry median market-to-book ratios Panels A and B

report the percentage of firms (i) issuing net debt in excess of 5 of the value of their assets (ii) issuing

net equity in excess of 5 of the value of their assets (iii) issuing debt and repurchasing equity both in

excess of 5 of the value of their assets and (iv) issuing equity and retiring debt both in excess of 5

of the value of their assets Panel C tabulates the mean debt and equity levels and the mean changes in

debt and equity issuances The sample contains nonfinancial firms listed on the annual Compustat files

for the period from 1971 to 2009

Iss Iss Iss D Iss ENet D Net E Rep E Ret D(1) (2) (3) (4)

Panel A sorts on profitability

Badtimes

Low 145 131 01 202 170 49 02 083 173 46 04 074 182 47 03 10

High 177 51 06 10t Highfrac14Low 41 136 33 01

Goodtimes

Low 206 381 01 462 197 148 02 233 212 103 04 184 234 92 10 18

High 216 92 16 14t Highfrac14Low 23 658 138 23

Panel B sorts on firm size and profitability

Badtimes

Smallfirm

s Low 151 318 00 312 182 276 02 253 151 140 02 294 183 73 00 19

High 177 80 00 11t Highfrac14Low 10 90 NA 20

Largefirm

s Low 172 54 01 042 148 39 04 053 150 43 04 044 156 33 03 06

High 147 28 05 03t Highfrac14Low 17 32 20 02

Goodtimes Smallfirm

s Low 213 448 01 542 230 443 01 473 222 325 00 474 203 197 02 38

High 190 153 02 31t Highfrac14Low 19 230 10 39

Largefirm

s Low 203 96 03 122 208 66 09 133 232 54 16 064 215 39 19 05

High 234 36 38 04t Highfrac14Low 29 99 92 38

(continued)

1440 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

larger firms there is a tendency for small firms to issue more equitymdashat leastto the extent possibleDuring good times large profitable firms raise significant amounts of debt

as they experience an increase in the value of their equity Such firms alsoengage in share repurchases Debt issues by smaller firms are much lessaffected by their own profitability during good times and their change indebt is more or less independent of their profits However equity issuanceis primarily found among the low-profit small firms during good timesTable IX shows that empirically profitability has a time-varying impacton leverage choicesIn Table X the basic regressions for changes in debt and for equity

issuance are presented for both good times and bad times Comparinggood times with bad times we see that the effects of profits are muchstronger in good times In bad times the impacts are rather weak statisticallyThis difference is interesting and deserves further study However for ourpurposes recall that the static trade-off theory implies time-varying capital

Table IX Continued

D D BVE BVE MVE MVE EqyIss Assets

(1) (2) (3) (4) (5) (6) (7) (8)

Panel C magnitude of financing activity

Badtimes Smallfirm

s Low 40 03 58 14 305 06 22 140

2 44 05 82 07 240 67 17 180

3 45 04 94 02 176 18 05 187

4 43 09 98 02 137 04 03 188

High 36 10 114 13 209 16 04 199

Largefirm

s Low 3365 73 2523 169 2304 163 94 8972

2 3031 95 2983 35 3284 278 56 9068

3 2930 5 3360 14 3947 308 57 9225

4 2281 6 3817 19 5213 442 7 8902

High 1790 62 4767 13 8799 58 31 9573

Goodtimes

Smallfirm

s Low 24 06 38 05 339 23 39 90

2 26 08 48 01 301 04 25 104

3 30 09 60 04 246 03 17 125

4 28 07 66 08 206 18 12 130

High 23 06 85 16 261 13 08 141

Largefirm

s Low 4532 36 3477 103 6557 260 53 12465

2 4580 58 3527 139 6911 421 23 12522

3 2952 111 3485 2 7586 279 32 9535

4 2752 124 3939 72 10686 33 105 10226

High 2429 228 4819 377 18565 268 222 10830

PROFITSndashLEVERAGE PUZZLEREVISITED 1441

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

structure choices even when the target is time invariant Empirically we doobserve time-varying choices

8 Is Scaling Important

In the preceding section we provided results for unscaled debt and equityissues This seems appropriate to us However it involves making twochanges to the standard regression We would like to determine if both

Table X Debt and equity changes in good and bad timesmdashregressions

The sample contains nonfinancial firms listed on the annual Compustat files for the period

from 1971 to 2009 The table presents estimates from regressions of change in debt (D)and net equity issuances (EqyIss) in good and bad times The explanatory variables includechange in profits lagged change in profits and lagged change in assets The regressions

include both firm and year fixed effects An industry is defined as having ldquogood timesrdquo ifthe median firm in that industry has a market-to-book ratio that is higher than the 67thpercentile of the time-series distribution of industry median market-to-book ratios

Conversely an industry is defined as having bad times if the median firm in thatindustry has a market-to-book ratio that is lower than the 33rd percentile of the time-series distribution of industry median market-to-book ratios The reported t-statistics arecorrected for clustering at the firm level

aMean significant at the 1 level

bMean significant at the 5 level

cMean significant at the 10 level

Good times Bad times

D EqyIss D EqyIss

(1) (2) (3) (4) (5) (6) (7) (8)

Profits 087a 091a 006b 009a 060 065 003 003b

(30) (29) (20) (28) (15) (16) (16) (20)

Profitst1 029c 015a 034c 001

(19) (37) (17) (04)

Assetst1 000 002b 000 001

(02) (22) (00) (07)

Constant 1337 3989 8691a 10001a 1351 1851 4099a 4055a

(03) (09) (55) (37) (09) (13) (80) (78)

Year FE Yes Yes Yes Yes Yes Yes Yes Yes

Firm FE Yes Yes Yes Yes Yes Yes Yes Yes

R2 0071 0080 0018 0045 0178 0212 0034 0041

N 74827 68729 74891 68787 26965 26013 26971 26018

1442 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

changes are crucial Accordingly in this section we take a step back toconsider prior literature in which variables are scaled The questions arehow does scaling affect the results and whyTable XI examines alternative scalings Columns (1) and (2) scale debt and

equity issuances by the total firm net issuance Columns (3) and (4) scaledebt and equity issuances by the sum of debt and book equity Columns (5)and (6) consider the change in debt scaled by the sum of debt and bookequity Columns (7) and (8) consider the change in leverage due to externalfinancing Since we are examining changes we use the changes in thestandard regression factors as controls The primary interest is in the coef-ficient on the change in profitsWe find that a change in profits is associated with an increase in debt

issues But the sign reverses when we scale the debt issues by the sum of debtplus book equity This is telling What is really driving the results is thechange in equity in the denominator not the change in debt in the numer-ator This result is further substantiated in Columns (5) and (6) where thechange in debt to capital ratio is used instead of debt issuanceTo further explore the impact of active issuance relative to passive

experiencing of changes we decompose leverage by separating out theeffect of debt and equity issues from those due to changes in retainedearnings Thus we define leverage change due to external financing as

LevEF frac14Dt1 thorn d

Dt1 thorn Et1 thorn dthorn e

Dt1

Dt1 thorn Et1 eth1THORN

where D is debt E is book value of equity d is debt issues net of retirementsand e is equity issues net of repurchases Garvey and Hanka (1999) use asimilar measure in their study This measure directly captures the netleverage effect of debt and equity issuancesColumns (7) and (8) report results from regressions of the net leverage

effect of debt and equity issues on changes in profitability after controllingfor other leverage factors in differences and year indicator variables Theeffect of profitability on leverage change due to external financing is positiveand statistically significant This comparison suggests that changes in totaldebt and equity are only partly a result of debt and equity issuance decisionsOther balance sheet adjustments complicate the inferences from leverageratio regressionsFor most of our analysis we follow the literature in treating debt as homo-

genous It is of course potentially interesting to ask what happens when wedistinguish between short-term debt and long-term debt Which part of thefirmrsquos debt maturity structure drives the under-adjustment Table XII

PROFITSndashLEVERAGE PUZZLEREVISITED 1443

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

provides results from such decompositions Alternative scalings areconsidered The main results of interest are that the distinction betweenshort- and long-term debt is important Short-term debt issuance is largelyindependent of changes in profits The negative sign on profit change stemsfrom the action of long-term debt This seems intuitively reasonable

Table XI Scaled issuances

The sample contains nonfinancial firms listed on the annual Compustat files for the period

from 1971 to 2009 The table presents estimates from regressions of debt issuance to totalissuance (in columns (1) and (2)) debt issuance to total capitalization (in columns (3) and(4)) change in debt to capitalization ratio (in columns (5) and (6)) and change in leverage

ratio considering only external financing (in columns (7) and (8)) The change in leverageratio considering only external financing is defined as

LevEF frac14Dt1 thorn d

Dt1 thorn Et1 thorn dthorn e

Dt1

Dt1 thorn Et1

where D is debt E is book value of equity d is debt issues net of retirements and e is equity

issues net of repurchases The regressions include lagged leverage factors and year indicatorvariables Some of the specifications include firm fixed effects The reported t-statistics arecorrected for clustering at the firm level

aMean significant at the 1 level

bMean significant at the 5 level

cMean significant at the 10 level

DebtIssTotIss

DebtIssethDthornETHORN D

DthornE LevEF

(1) (2) (3) (4) (5) (6) (7) (8)

Profitt1 004a 000 002a 001c 003a 002 001b 002a

(35) (03) (34) (18) (35) (14) (23) (26)

IndLevt1 004c 001 001 001c 000 000 000 000

(19) (06) (15) (19) (02) (04) (03) (05)

Market=bookt1 001a 001a 000a 000a 000a 000a 000 000c

(47) (85) (68) (60) (55) (36) (05) (18)

Tangt1 004 005 013a 011a 011a 006a 003a 003a

(12) (15) (144) (112) (79) (38) (44) (42)

Assetst1 008a 004a 006a 004a 002a 001a 003a 002a

(97) (45) (283) (161) (53) (35) (147) (101)

Constant 079a 074a 002a 003a 000 000 000b 002a

(497) (442) (91) (111) (00) (05) (22) (93)

Year FE Yes Yes Yes Yes Yes Yes Yes Yes

Firm FE No Yes No Yes No Yes No Yes

R2 0015 0098 0022 0076 0009 0070 0012 0086

N 133905 133905 140552 140552 138229 138229 140572 140572

1444 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

inasmuch as short-term debt is probably more associated with day-to-dayoperations rather than longer term planning considerationsIn Table XIII we examine change in leverage ( D

DthornE) and change inleverage due to external financing (LEVEF) during good and bad timesseparately The traditional negative effect of a change in profits on achange in leverage is found both in good times and in bad timesHowever when we decompose the impact in order to isolate the externalcomponent as before the result disappears In good times the externalfinancing component of a change in profits is weak but still positiveDuring bad times it is not statistically significant This implies that distin-guishing between good and bad times does not fundamentally alter the basic

Table XII Short- and long-term debt issuances

The sample contains nonfinancial firms listed on the annual Compustat files for the period

from 1971 to 2009 The table presents estimates from regressions of short-term debtissuance to total issuance (in columns (1) and (2)) long-term debt issuance to totalissuance (in columns (3) and (4)) short-term debt issuance to total capitalization (in

columns (5) and (6)) and long-term debt to total capitalization (in columns (7) and (8))Some of the specifications include firm fixed effects The reported t-statistics are correctedfor clustering at the firm level

aMean significant at the 1 level

bMean significant at the 5 level

cMean significant at the 10 level

STDebtIssTotIss

LTDebtIssTotIss

STDebtIssethDthornETHORN

LTDebtIssethDthornETHORN

(1) (2) (3) (4) (5) (6) (7) (8)

Profitabilityt1 0002 0003 0042a 0008 0001 0001 0019a 0012a

(02) (04) (30) (05) (14) (10) (45) (28)

IndLevt1 0009 0007 0070b 0033 0001 0002c 0014a 0005

(06) (05) (26) (11) (08) (17) (26) (09)

M=Bt1 0003a 0003a 0003b 0010a 0000 0000 0003a 0002a

(48) (35) (22) (65) (10) (01) (84) (42)

Tangibilityt1 0020 0007 0011 0067 0005a 0006a 0097a 0081a

(12) (03) (03) (16) (34) (33) (108) (89)

Assetst1 0039a 0011b 0034a 0024b 0001a 0001 0046a 0020a

(89) (22) (38) (23) (36) (13) (190) (82)

Constant 0001b 0153a 0798a 0591a 0000c 0006a 0057a 0096a

(25) (144) (470) (288) (17) (124) (292) (234)

Year FE Yes Yes Yes Yes Yes Yes Yes Yes

Firm FE No Yes No Yes No Yes No Yes

R2 0025 0089 0018 0073 0030 0147 0016 0298

N 136146 136146 136140 136140 143031 143031 142847 142847

PROFITSndashLEVERAGE PUZZLEREVISITED 1445

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

Table XIII Externally financed leverage changes in good and bad times

The sample contains nonfinancial firms listed on the annual Compustat files for the period

from 1971 to 2009 The table presents estimates from regressions of change in debt tocapitalization ratio (in columns (1) (2) (5) and (6)) and change in leverage ratio consider-ing only external financing (in columns (3) (4) (7) and (8)) The change in leverage ratio

considering only external financing is defined as

LevEF frac14Dt1 thorn d

Dt1 thorn Et1 thorn dthorn e

Dt1

Dt1 thorn Et1

where D is debt E is book value of equity d is debt issues net of retirements and e is equityissues net of repurchases The regressions include lagged leverage factors and year indicator

variables The reported t-statistics are corrected for clustering at the firm level An industryis defined as having ldquogood timesrdquo if the median firm in that industry has a market-to-bookratio that is higher than the 67th percentile of the time-series distribution of industry

median market-to-book ratios Conversely an industry is defined as having ldquobad timesrdquoif the median firm in that industry has a market-to-book ratio that is lower than the 33rdpercentile of the time-series distribution of industry median market-to-book ratios

aMean significant at the 1 level

bMean significant at the 5 level

cMean significant at the 10 level

Good times Bad times

DDthornE LevEF D

DthornE LevEF

(1) (2) (3) (4) (5) (6) (7) (8)

Profitabiltyt1 0022c 0005 0016c 0025b 0073b 0041 0007 0007

(17) (03) (18) (24) (25) (11) (04) (03)

IndLevt1 0000 0011 0001 0009 0010 0018 0001 0000

(00) (09) (01) (13) (07) (10) (02) (00)

MB

t1

0004a 0003b 0000 0001 0008a 0001 0000 0000

(40) (24) (05) (08) (28) (04) (03) (03)

Tangt1 0097a 0041 0033a 0038b 0127a 0048 0059a 0033

(45) (16) (27) (25) (49) (14) (43) (16)

Assetst1 0014a 0013b 0024a 0019a 0037a 0023c 0036a 0025a

(30) (20) (84) (48) (45) (17) (88) (43)

Constant 0030b 0036 0000 0003 0028c 0014 0021 0001

(20) (16) (00) (07) (17) (06) (11) (01)

Year FE Yes Yes Yes Yes Yes Yes Yes Yes

Firm FE No Yes No Yes No Yes No Yes

R2 0006 0088 0009 0117 0026 0251 0023 0202

N 56820 56820 58200 58200 23462 23462 23511 23511

1446 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

conclusions The traditional leverage result on profits is driven by increasesin equity that are experienced and then partially but not completely offsetby firmsTo the extent that firms are engaging in active financing they are following

the prescriptions of the trade-off theory Thus we have a coherent picture ofhow firms are responding to profits Basically more profitable firms experiencean increase in the value of their equity They respond to this by issuing debtand repurchasing equity But the magnitude of the increase in equity is largerthan the magnitudes of the active steps taken in the debt and equity marketsThis means that the firmsrsquo actual actions are as predicted by the trade-offtheory But it leaves open the question of why they do not act more stronglyThe capital structure literature has long been interested in dynamics and

capital structure adjustment paths In recent years this interest has increas-ingly resulted in explicit dynamic optimizing models such as those ofGoldstein Ju and Leland (2001) Hennessy and Whited (2005) andStrebulaev (2007) These models commonly make specialized assumptionsleading some to question their generalityWe argue that both fixed transactions costs and variable transactions costs

are plausible and realistic in models that pin down firm size The simultaneouspresence of both is important Due to fixed costs firms only rebalance inter-mittently Due to variable adjustment costs the extent of the adjustment isreduced to the point at which an extra unit to adjustment is balanced againstthe extra unit transactions costs This means that both leverage increases andleverage reductions will only go part way toward a static optimumThe key point is that with both fixed and variable transactions costs we

should expect to only see intermittent leverage adjustments When leverage isadjusted it should ldquounder adjustrdquo when compared to a frictionless modelFurther discussion of transactions costs in a capital structure setting isprovided by Leary and Roberts (2005)

9 Conclusion

The connection between corporate profits and capital structure has beenvery influential in the assessment of the trade-off theory The conventionalinterpretation of the evidence has pushed the literature away from the trade-off theory and toward much more complex dynamic models and ideas As aresult it is important to make sure that the evidence has been correctlyinterpreted We show that the problem is with the interpretation of thetest results and a failure to take into account the costs associated with re-sponding to a profit shock

PROFITSndashLEVERAGE PUZZLEREVISITED 1447

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

We document that more profitable firms really do borrow more (not less)They tend to repurchase equity They experience an increase in both thebook value of equity and the market value of equity Less profitable firmsreally do tend to reduce their debt and to issue equityFirm size and market conditions also matter Larger firms tend to be more

active in the debt markets while smaller firms tend to be relatively moreactive in the equity markets During good times there is more use ofexternal financingThe usual profitsndashleverage puzzle result is primarily driven by the increase

in equity that is experienced by the more profitable firms Accordingly thepuzzle should be restated as asking why do firms not take sufficiently largeoffsetting actions to fully undo the change in equity What limits the mag-nitudes of the typical leverage response to profit shocksIn a frictionless model the partial response appears puzzling But there is

good empirical reason to believe that rebalancing entails both fixed andvariable costs and that firm size matters The rebalancing costs can befully avoided by doing nothing Accordingly the firm must decide whetherany given shock is big enough to be worth responding to If it is then thefirm must decide how big a response is called for These technical conditionsare known as ldquovalue matchingrdquo and ldquosmooth pastingrdquo Thus optimizationimplies that some shocks will be ignored Even if the shock is not ignoredthe optimal response will only partially undo the shock The magnitude ofthe leverage response must balance the marginal cost and the marginalbenefit of an extra unit of leverage Since the marginal cost of adjustingleverage is strictly positive the adjustment toward that static leverageoptimum will only go part way This is true both for leverage increasesand for leverage reductionsOne might conjecture that the variable costs would be time varying and

that such costs might have declined over time Since we do not have a fullaccounting of all of the variable costs it is of course impossible to be sureHowever there is some evidence suggesting that at least the directunderwriting fees are remarkably stable over time10

Surely nobody ought to be surprised that more profitable firms are moreinclined to repurchase equity while unprofitable firms tend not to do so Thefact that more profitable firms typically experience an increase in the value of

10 For IPOs Jay Ritter provides important summary evidence as httpbearwarringtonufleduritterIPOs2013Underwritingpdf He finds that for medium-sized deals a remark-

ably large fraction of the deals have a 7 fee year after year Whether this is due tobehavioral factors or some sort of a coordination device is not clear The extent towhich similar intertemporal stability carries over to other forms of external financemerits further study

1448 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

equity is equally natural Were it not for the contrary claims in the literature(ie the profitsndashleverage puzzle) we would argue that nobody ought to findit surprising that more profitable firms are more inclined to issue debt Thefact that leverage adjustments typically only partially offset profit shocksshould also sound quite natural These facts all fit together easily within thetrade-off theory once we take account of both fixed and variable costs ofactively adjusting capital structure

Appendix

Table AI Market leverage regressions

The table presents estimates of the regressions of market leverage on firm characteristics

The sample comes from the annual Compustat files during the period 1971ndash2009 Financialfirms are excluded The leverage is estimated as the ratio of debt over debt plus marketequity The explanatory variables Profitabilityt1

MB

t1

Tangibilityt1 and LnethAssetsTHORNt1are described in Table 1 IndMedianLevt1 is estimated as the median market leverage of allother firms in the same industry The industry is defined at the level of the four-digit SICcode All specifications include the year fixed effects The specifications in column (2)

additionally include the firm fixed effects We report t-statistics where the standard errorsare clustered at the firm level in parentheses

aSignificant at the 1 level

Quantile regressions

(1) (2) (3) (4) (5)

Profitabilityt1 013a 013a 007a 011a 020a

(275) (262) (349) (425) (538)

IndLevt1 046a 020a 032a 058a 063a

(506) (239) (1299) (1630) (1042)MB

t1

003a 001a 001a 002a 004a

(509) (303) (743) (709) (534)

Tangibilityt1 013a 018a 013a 014a 010a

(191) (181) (762) (571) (232)

LnethAssetsTHORNt1 001a 005a 001a 001a lt001a

(112) (258) (682) (342) (29)

Constant 010a 011a 005a 005a 029a

(171) (98) (173) (118) (398)

Year FE Yes Yes Yes Yes Yes

Firm FE No Yes No No No

Regression type OLS FE 25th tile Median 75thtile

Clustered SE Yes Yes No No No

R2 030 067

Pseudo R2 016 023 0201

N 158578 158578 158578 158578 158578

PROFITSndashLEVERAGE PUZZLEREVISITED 1449

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

Table AII Profitability sorts for firms issuing large amounts

Percentage of firms

Iss Ret Iss Iss Rep Iss Iss IssD IssE

D D NetD E E NetE DampE RepE RetD

(1) (2) (3) (4) (5) (6) (7) (8) (9)

Panel A 10 cutoff to define issuing or retiring activity

Allfirm

s

Low 205 150 134 260 05 257 45 00 23

2 207 175 103 77 06 75 11 01 10

3 243 196 101 50 08 47 08 01 09

4 252 186 113 50 13 48 08 02 08

High 227 161 118 60 27 57 07 04 09

t Highfrac14Low 70 38 62 755 226 768 320 90 152

Smallfirm

s Low 230 132 162 380 05 375 82 00 32

2 228 143 165 334 04 333 65 00 26

3 211 146 147 216 03 214 38 00 23

4 190 154 120 126 04 124 15 01 20

High 194 162 114 99 09 97 12 01 15

t Highfrac14Low 42 40 68 342 20 340 169 07 131

Largefirm

s Low 231 200 82 31 06 29 08 01 04

2 231 187 67 17 06 14 03 01 02

3 234 177 77 17 09 14 03 02 03

4 218 158 75 16 17 14 02 03 02

High 218 135 99 19 41 15 03 07 02

t Highfrac14Low 19 106 36 46 135 57 39 58 23

Percentage of firms

Iss Ret Iss Iss Rep Iss Both Iss D Iss E

D D Net D E E Net E DampE Rep E Ret D

(1) (2) (3) (4) (5) (6) (7) (8) (9)

Panel B 15 cutoff to define issuing or retiring activity

Allfirm

s

Low 158 101 100 230 01 228 27 00 15

2 142 115 63 57 01 56 05 00 06

3 165 128 57 32 01 31 03 00 05

4 173 118 65 35 01 34 03 00 05

High 160 102 70 46 04 44 03 00 05

t Highfrac14Low 08 06 141 738 87 746 277 24 131

Smallfirm

s Low 195 95 132 354 01 350 56 00 24

2 185 96 131 303 01 301 44 00 18

3 160 96 104 179 00 176 18 00 15

4 134 99 78 99 01 98 08 00 12

High 133 95 73 82 01 80 06 00 10

t Highfrac14Low 80 00 95 345 12 344 149 12 52

Largefirm

s Low 145 123 42 14 01 12 03 00 01

2 137 114 32 07 01 06 01 00 01

3 142 104 35 06 01 05 00 00 01

4 130 91 36 08 03 07 01 01 01

High 134 81 48 09 06 07 01 01 01

t Highfrac14Low 19 85 17 26 54 33 34 17 11

1450 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

Table AIV Debt and equity changesmdashwithout fixed effects

The sample contains nonfinancial firms listed on the annual Compustat files for the period from 1971

to 2009 The table presents estimates from regressions of change in debt (D) change in the book

value of equity (BVE) change in the market value of equity (MVE) and net equity issuances

(EqyIss) The explanatory variables include current and lagged change in operating income before

depreciation (Profits and Profitst1) and lagged change in book value of assets (Assetst1) The

regressions include year indicator variables The reported t-statistics are corrected for clustering at the

firm level

aMean significant at the 1 level

bMean significant at the 5 level

cMean significant at the 10 level

D BVE MVE EqyIss

(1) (2) (3) (4) (5) (6) (7) (8)

Profitst1 025a 024a 023a 038a 007 066a 007a 009a

(40) (32) (36) (39) (06) (29) (42) (51)Profits 051b 082a 149a 004a

(25) (33) (49) (33)Assetst1 002 005c 024a 001b

(08) (17) (34) (23)Constant 93 176a 45 71 1552a 1704a 130a 140a

(12) (28) (07) (09) (49) (49) (45) (45)Year FE Yes Yes Yes Yes Yes Yes Yes YesR2 001 005 001 011 001 005 002 002N 162056 162056 162130 162130 157550 157550 162154 162154

Table AIII Median profitability sorts

The sample contains nonfinancial firms listed on the annual Compustat files for the period from 1971

to 2009 The table presents median debt and equity median changes in debt and equity and median

equity issuances (EqyIss) in millions of dollars for all firms sorted on profitability and for profitability

sorts within the smallest and largest firms The sorts are done annually

D D BVE BVE MVE MVE EqyIss Assets(1) (2) (3) (4) (5) (6) (7) (8)

Allfirm

s

Low 21 00 75 17 256 20 01 1922 186 02 449 18 598 19 00 11033 816 06 1274 03 1637 05 00 32244 787 02 1541 18 2488 03 00 3556High 273 00 1361 66 3300 14 00 2578

Smallfirm

s Low 05 00 13 07 140 14 06 402 06 00 25 08 114 12 01 563 09 00 32 05 85 07 00 674 12 00 39 01 68 03 00 79High 10 00 51 04 98 01 00 90

Largefirm

s Low 11485 381 10269 371 13227 202 03 333902 13051 393 12786 140 17571 169 05 392463 11554 287 13429 24 21623 305 00 370314 9963 209 15052 194 29424 308 00 36705High 7361 00 16380 613 49142 439 00 34976

PROFITSndashLEVERAGE PUZZLEREVISITED 1451

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

References

Auerbach A J (1985) Real determinants of corporate leverage in B M Friedman (ed)

Corporate Capital Structures in the United States University of Chicago Press Chicago

IllinoisBertrand M and Mullainathan S (2003) Enjoying the quiet life corporate governance and

managerial preferences Journal of Political Economy 111 1043ndash1075Booth L Varouj A Demirguc-Kunt A and Maksimovic V (2001) Capital structures in

developing countries Journal of Finance 56 87ndash130Brennan M J and Schwartz E S (1984) Optimal financial policy and firm valuation

Journal of Finance 39 593ndash607Cameron A C and Trivedi P K (2010) Microeconometrics using Stata Stata Press TexasConstantinides G M (1979) Multiperiod consumption and investment behavior with

convex transactions costs Management Science 25 1127ndash1137Constantinides G M and Richard S F (1978) Existence of optimal simple policies for

discounted-cost inventory and cash management in continuous time Operations Research

26 620ndash636DeAngelo H DeAngelo L and Whited T M (2011) Capital structure dynamics and

transitory debt Journal of Financial Economics 99 235ndash261Dixit A K (1993) The Art of Smooth Pasting Routledge London and New YorkFama E and French K R (2002) Testing trade-off and pecking order predictions about

dividends and debt Review of Financial Studies 15 1ndash33Faulkender M and Petersen M A (2006) Does the source of capital affect capital struc-

ture Review of Financial Studies 19 45ndash79Faulkender M Flannery M J Hankins K W and Smith J M (2012) Cash Flows and

Leverage Adjustments Journal of Financial Economics 103 632ndash646Fischer E Heinkel R and Zechner J (1989) Dynamic capital structure choice theory

and tests Journal of Finance 44 19ndash40Frank M Z and Goyal V K (2008) Trade-off and pecking order theories of

debt in B E Eckbo (ed) Handbook of Corporate Finance Empirical Corporate Finance

Vol 2 Handbook of Finance Series Chapter 12 ElsevierNorth-Holland Amsterdam pp

135ndash202

Frank M Z and Goyal V K (2009) Capital structure decisions which factors are reliably

important Financial Management 38 1ndash37

Garvey G T and Hanka G (1999) Capital structure and corporate control the effect of

antitakeover statutes on firm leverage Journal of Finance 54 519ndash546

Goldstein R Ju N and Leland H (2001) An EBIT-based model of dynamic capital

structure Journal of Business 74 483ndash512

Graham J R and Tucker A L (2006) Tax shelters and corporate debt policy Journal of

Financial Economics 81 563ndash594

Hackbarth D and Mauer D C (2012) Optimal priority structure capital structure and

investment Review of Financial Studies 25 747ndash796

Halling M Yu J and Zechner J (2012) Leverage dynamics over the business cycle

Working paper University of Utah University of New South Wales and Vienna

University of Economics and Business AdministrationHennessy C A and Whited T A (2005) Debt dynamics Journal of Finance 60 1129ndash1165

Huang R and Ritter J R (2009) Testing theories of capital structure and estimating the

speed of adjustment Journal of Financial and Qualitative Analysis 44 237ndash271

1452 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

Hovakimian A Opler T and Titman S (2001) The debt-equity choice Journal of

Financial and Quantitative Analysis 36 1ndash24Jung K Kim Y and Stulz R M (1996) Timing investment opportunities managerial

discretion and the security issue decisions Journal of Financial Economics 42 159ndash185Kane A Marcus A J and McDonald R L (1984) How big is the tax advantage to

debt Journal of Finance 39 841ndash853Kayhan A and Titman S (2007) Firmsrsquo histories and their capital structures Journal of

Financial Economics 83 1ndash32Leary M T (2009) Bank loan supply lender choice and corporate capital structure

Journal of Finance 64 1143ndash1185Leary M T and Roberts M R (2005) Do firms rebalance their capital structures

Journal of Finance 60 2575ndash2619Leland H E (1994) Corporate debt value bond covenants and optimal capital structure

Journal of finance 49 1213ndash1252Lemmon M L Roberts M R and Zender J F (2008) Back to the beginning persist-

ence and the cross-section of corporate capital structure Journal of Finance 63 1575ndash1608Long M S and Malitz I B (1985) The investment-financing nexus some empirical

evidence Midland Corporate Finance Journal 3 53ndash59Mackie-Mason J K (1990) Do taxes affect corporate financing decisions Journal of

Finance 45 1471ndash1493Morellec E Nikolov B and Schurhoff N (2012) Corporate governance and capital

structure dynamics Journal of Finance 67 803ndash848Myers S C (1984) The capital structure puzzle Journal of Finance 39 575ndash592Myers S C (1993) Still searching for optimal capital structure Journal of Applied

Corporate Finance 61 4ndash14Rajan R and Zingales L (1995) What do we know about capital structure some evidence

from international data Journal of Finance 50 1421ndash1460Strebulaev I A (2007) Do tests of capital structure theory mean what they say Journal of

Finance 62 1747ndash1787

Strebulaev I A and Whited T M (2012) Dynamic Models and Structural Estimation in

Corporate Finance Now Publisher Boston and Delft

Titman S and Wessels R (1988) The determinants of capital structure choice Journal of

Finance 43 1ndash21

Welch I (2004) Capital structure and stock returns Journal of Political Economy 112

106ndash131

Welch I (2011) Two common problems in capital structure research the financial-debt-

to-asset ratio and issuing activity versus leverage changes International Review of Finance

11 1ndash17

PROFITSndashLEVERAGE PUZZLEREVISITED 1453

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

16 issue between 1 and 5 of the value of their assets as debt and theremaining 38 issue debt in excess of 5 of the value of their assetsThe mean constant dollar equity issue is about $26 million (the median is

about $04 million) As a fraction of assets the mean and median equityissues are about 7 and 02 respectively About 33 of the firms issue noequity 34 of the firms issue between 0 and 1 of the value of theirassets as equity another 14 issue between 1 and 5 of the value of theirassets as equity and the remaining 19 of the firms issue equity that is inexcess of 5 of the value of their assets Average debt repayments are largerthan equity repurchases This perhaps reflects the finite maturity of debt andits contractual repayment The median firm does not repurchase equityWe construct both book and market leverage ratios Book leverage is

defined as debt over debt plus book equity Market leverage is defined asdebt over debt plus market equity8 The median book leverage is 036 (theaverage is 031) The median market leverage is 027 (the average is 020)Profitability is defined as the ratio of operating income before depreci-

ation to assets While the average firm is profitable (the ratio of EBITDA toassets is 005) the median firm is even more profitable (with a profitabilityratio of 011) But the sample also includes a large number of unprofitablefirms as the 10th percentile is 018 The table also reports descriptive stat-istics on the market-to-book ratio and the tangibility ratio The market-to-book ratio (MB) defined as the ratio of the market value of assets to bookassets averages at about 165 Tangibility defined as the ratio of netproperty plant and equipment to assets averages at about 31

4 Leverage Ratio Regressions

The previous literature focuses on estimates obtained using leverage ratiosHence we start with a similar estimation to check whether our results matchthose of previous studies Table II presents the results for book leverage Weobtain similar results for market leverage and they are presented inAppendix Table AI The regressions include leverage factors following theprevious capital structure literature (see eg Lemmon Roberts and Zender(2008) and Frank and Goyal (2009)) The factors include (i) profitability (ii)industry median leverage (iii) market-to-book assets ratio (iv) tangibility ofassets and (v) firm size Industry median leverage is constructed as themedian leverage of all other firms in the industry excluding the firm under

8 Welch (2011) stresses the idea that nonfinancial liabilities should not be implicitly mis-treated as if they were equity by paying excessively narrow attention to financial liabilitiesin a leverage ratio We have adopted his approach in the empirical work reported here

1426 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

consideration Firm size is defined as the natural log of assets Other factorsare defined in Section 3 We cluster standard errors by firm and estimatethese regressions both with and without fixed effectsIn Table II we run conventional cross-section leverage regressions Since

we employ the usual data it is not surprising that our results match thosereported in the existing capital structure literature Profitability has anegative sign in both the book leverage regressions and in the marketleverage regressions The coefficients on other factors largely match thosereported in earlier studies Firms operating in industries with high leverage

Table II Leverage and profitability

The table presents parameter estimates for book leverage regressions The sample consists

of nonfinancial companies in the annual Compustat files during the period 1971ndash2009 Theleverage is estimated as the ratio of debt over debt plus book equity The explanatoryvariables Profitabilityt1

MB

t1

Tangibilityt1 and LnethAssetsTHORNt1 are described in

Table I IndLevt1 is estimated as the median book leverage of all other firms in thesame industry The industry is defined at the level of the four-digit SIC code All specifi-cations include the year fixed effects The specifications in column (2) additionally include

the firm fixed effects We report t-statistics where the standard errors are clustered at thefirm level in parentheses

aSignificant at the 1 level

Quantile regressions

(1) (2) (3) (4) (5)

Profitabilityt1 026a 020a 011a 017a 041a

(293) (192) (378) (518) (1015)

IndLevt1 042a 012a 030a 052a 052a

(369) (107) (926) (1252) (950)MB

t1

002a 001a 001a 002a 002a

(203) (102) (392) (426) (418)

Tangibilityt1 018a 025a 024a 019a 012a

(206) (154) (964) (580) (276)

LnethAssetsTHORNt1 001a 002a 002a 002a 001a

(114) (76) (899) (478) (96)

Constant 012a 010a 012a 002a 029a

(156) (60) (284) (27) (387)

Year FE Yes Yes Yes Yes Yes

Firm FE No Yes No No No

Regression type OLS FE 25th tile Median 75thtile

Clustered SE Yes Yes No No No

R2 013 054

Pseudo R2 016 015 009

N 158824 158824 158824 158824 158824

PROFITSndashLEVERAGE PUZZLEREVISITED 1427

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

tend to have high leverage A higher market-to-book ratio is associated withlower leverage Larger firms are typically more highly levered The coeffi-cients on tangibility are positive The conclusion from Table II is clear Theleverage regression results match those reported in previous studiesAt times one may be concerned that in a regression what is being

estimated is a conditional mean while some parts of the distributionmight behave differently from other parts If so then ldquoaccidentsrdquo likesample selection issues or asymmetries in the underlying distributionsmight play a greater role than is usually recognized In other words thesource of identification could be misinterpreted To guard against this po-tential problem we estimate the same model using quantile regressions9

Columns 3 4 and 5 report what happens at the 25th 50th and 75th per-centiles The basic model results are robust in terms of both the sign and thestatistical significanceThe impact of profits does appear to be somewhat stronger among the

high leverage firms For instance in a book leverage regression the coeffi-cient on profits for the first quartile is 011 with a t-ratio of 378 For thethird quartile the coefficient on profits is 041 with a t-ratio of 1015This difference is curious and might deserve further study However the factthat in each case we get the negative sign and statistical significance is suf-ficient for the purposes of the present article So the results on profits arerather robust across the distributionTable III sorts firms into those that are actively adjusting their leverage

and those that are passive In the presence of both fixed and variable trans-acts costs this distinction is important We expect the coefficient on profit-ability in leverage ratio regressions to be more negative in firms that arepassive and less negative in firms that are actively adjustingWe define active firms as those that are either issuing net debt or net

equity in excess of 5 of their assets Passive firms are those that neitherissue debt nor equity in nontrivial amounts (ie greater than 5 of theirassets) Consistent with our expectations we find in Table III that the coef-ficients on profits are significantly more negative for passive firms than theyare for active firms With firm and year fixed effects the coefficient onprofits is 008 for active firms and 017 for the passive firms The twocoefficients are significantly different from each other (plt 001) InTable III the standard control variables are used They have the usualsigns and for the most part do not differ much between the active and thepassive firms

9 Cameron and Trivedi (2010) provide an extensive discussion of quantile regressions inStata

1428 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

5 Financing Activity

51 FREQUENCY OF DEBT AND EQUITY ISSUES

Table IV reports the percentage of firms issuing or repurchasing debt orequity for annual sorts based on lagged profitability In Panel A weemploy the conventional 5 cutoff rule to exclude minor fluctuations anddefine firms ldquoissuing debtrdquo as those that issue debt (both short-term andlong-term) in excess of 5 of the value of their assets Other decisions aresimilarly defined using a 5 cutoff Appendix Table AII examines the dis-tributional properties across profitability (and size) quintiles for ldquolargerdquo is-suances by increasing the 5 cutoff to higher levels such as 10 and 15How does security issuance behavior vary with firm profitability To

answer this question Table IV sorts firms by profitability and reports the

Table III Leverage and profitability for active and passive firms

The table presents estimates of leverage on firm characteristics for the subsamples of firms

that are active in either debt or equity markets Active firms are defined as those that issueeither net debt in excess of 5 of the value of assets or issue net equity in excess of 5 ofthe value of their assets The passive firms do not issue significant (more than 5) of either

debt or equity The sample comes from the annual Compustat files during the period 1971ndash2009 Financial firms are excluded The variables are described in Tables I and II Wereport t-statistics where the standard errors are clustered at the firm level in parentheses

aSignificant at the 1 level

Active Passive

(1) (2) (3) (4)

Profitabilityt1 006a 008a 022a 017a

(126) (148) (254) (188)

IndLevt1 044a 019a 041a 013a

(424) (149) (326) (127)MB

t1

002a 001a 003a 001a

(362) (196) (327) (172)

Tangibilityt1 010a 014a 014a 019a

(146) (118) (145) (130)

LnethAssetsTHORNt1 001a 004a 001a 004a

(121) (179) (147) (169)

Constant 012a 008a 003a 007a

(120) (57) (45) (42)

Year FE Yes Yes Yes Yes

Firm FE No Yes No Yes

R2 034 064 033 075

N 43395 43395 58943 58943

PROFITSndashLEVERAGE PUZZLEREVISITED 1429

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

percentage of firms issuing or repurchasing debt or equity The table alsoreports test statistics from tests that examine whether the percentage differ-ences between the high- and low-profit firms are significantly different fromeach other The results in Column (1) show that the likelihood of issuing debt

Table IV Profitability sorts for debt and equity issuers

The sample contains nonfinancial firms listed on the annual Compustat files for the period

from 1971 to 2009 The table presents the percentage of firms issuing and retiring (orrepurchasing) debt and equity The firms are classified as ldquoissuing debtrdquo if they issuedebt in excess of 5 of the value of their assets as ldquoissuing equityrdquo if they issue equity

in excess of 5 of the value of their assets as ldquoretiring debtrdquo if they retire debt in excess of5 of the value of their assets and as ldquorepurchasing equityrdquo if they repurchase equity inexcess of 5 of the value of their assets In addition we also report net debt issuers which

are firms that issue net debt over 5 of the value of their assets and net equity issuerswhich are firms that issue net equity in excess of 5 of the value of their assets Weannually sort firms on lagged profitability and report the percentage of firms in each ofthese categories The bottom part of each panel reports the percentage of firms issuing or

retiring securities by profitability within the smallest and largest asset quintiles The tablealso reports results from t-test for the difference in percentage of firms issuing or retiringsecurities in the lowest and the highest profit quintile

Percentage of firms

Iss Ret Iss Iss Rep Iss Both IssD IssE

D D NetD E E NetE DampE RepE RetD

(1) (2) (3) (4) (5) (6) (7) (8) (9)

Allfirm

s

Low 276 245 186 301 13 297 71 01 38

2 319 299 181 117 21 112 27 02 20

3 385 344 194 86 29 80 21 05 15

4 386 330 213 81 43 75 20 07 14

High 329 273 198 92 82 81 17 13 14

t Highfrac14Low 150 84 39 723 414 763 358 176 200

Smallfirm

s Low 275 204 201 411 11 405 109 01 47

2 291 233 213 379 10 378 99 01 45

3 281 246 204 265 14 261 66 01 41

4 275 272 186 162 15 159 34 02 35

High 275 281 178 136 27 131 25 02 26

t Highfrac14Low 01 84 28 316 55 318 171 18 55

Largefirm

s

Low 414 363 177 87 22 79 28 03 11

2 449 379 180 62 30 57 17 07 09

3 439 360 186 56 41 48 16 09 07

4 403 320 194 46 67 39 13 12 06

High 374 264 207 50 129 35 10 28 05

t Highfrac14Low 50 130 45 89 244 116 77 120 47

1430 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

is mostly independent of firm profitability In the aggregate there is a smalldifference but it is sensitive to sorting on firm size Similarly in Column (2)there is only a weak relation between the profitability and the likelihood thatthe firm retires debt To a first approximation the fraction of firms issuing orrepurchasing debt is largely independent of profits Large firms are somewhatmore likely than small firms to issue debt and also to retire debtUnlike debt issues the probability of issuing equity is strongly related to

profitability as can be seen in Column (4) Low-profit firms are much morelikely to issue equity than are high-profit firms In the lowest profit quintile301 of the firms issued equity In the highest profit quintile it was only92 The lower panel reports separate sorts for small and large firms Similarpatterns are found for both groups of firms that is the likelihood of equityissuance declines as profitability increases The differences between the quin-tiles are quite stark for the group of small firms All quintiles of small firmsissue equity much more frequently than the corresponding large firm quintilesHigh-profit firms are much more likely to repurchase equity as shown in

Column (5) In the overall sample only 13 of low-profit firmsrepurchased equity compared with 82 of the high-profit firms Whenwe separate the small and large firms again a difference emerges Ingeneral small firms do not repurchase much equity while a greaterfraction of the large firms do soGiven these facts it is natural that low profitability firms are much more

likely to be net issuers of equity than are high profitability firms as shown inColumn (6) Among the firms in the lowest quintiles of profits roughly 30of firms issue net equity that exceeds 5 percent of their assets In contrastfor the most profitable quintile only about 8 issue net equityFirm size is an important variable in the recent literaturemdashit is often used

as a proxy for access to capital markets (as in Faulkender and Petersen(2006) and Leary (2009)) Small firms are bank-dependent risky and infor-mationally opaque They have restricted access to public debt markets andconsequently face more severe supply constraints in their ability to issuedebt Thus we expect small firms to be more sluggish in adjusting theirdebt and equity in response to shocks to profitability Large firms incontrast have much easier access to public debt markets and they facefewer obstacles in accessing securities marketsTo examine how size and profitability interact the bottom part of Table IV

examines profitability sorts for the smallest and largest firms We first sortfirms annually by firm size and then within each size quintile we sort them onprofitability We do these two-way sorts to ensure that we have similarnumber of firms in profitability subgroups for both small and large firmsamples Unconditionally small firms are generally less profitable while

PROFITSndashLEVERAGE PUZZLEREVISITED 1431

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

large firms are relatively more profitable The two-way sorts allow us to havea uniform distribution of firms in various profitability quintiles for both smalland large firm samples The bottom part of Table IV reports results for thesmallest and the largest size quintiles Among the small firms there is littlerelation between profitability and the likelihood of issuing debt Howeverlarger firms exhibit a small increase in the likelihood of net issuance of debtwith increasing profitabilityIn contrast to debt the effects of profitability on equity issuance and

repurchases are much more consistent across size quintiles As profitabilityincreases firms are generally less likely to issue equity and more likely torepurchase it Across the two extreme size groups we note a striking differ-ence between the proportion of firms issuing equity and those issuing debtAmong low-profit small firms almost 41 are equity issuers whereas onlyabout 9 of low-profit large firms are equity issuers However regardless ofsize we note a monotonic reduction in the likelihood of issuing equity asprofitability increasesColumn (7) reports the fraction of firms that issue both debt and equity in

excess of 5 of assets both as a function of profits in the upper panel andas a function of firm size and profits in the lower panel In the upper panelwe see that low-profit firms are much more likely to issue both debt andequity (71) while the high-profit firms are less likely to do so (17) Thelower panel shows that small firms are much more likely to be issuers of bothin contrast to large firms In both firm size categories the profit pattern is thesame with low profits being more frequent issuers of both simultaneouslyImportantly Column (8) shows that the likelihood of issuing debt and

simultaneously repurchasing equity increases with profitability Converselyin Column (9) we find that the likelihood of doing the reverse that is issuingequity and retiring debt declines with profitability The effects of sorting onfirm size and firm profitability mirror those reported for all firms Overallthe results show that low-profit firms are less likely to issue debt and repur-chase equity they are instead more likely to issue equity and retire debtFirms with high profitability exhibit the reverse patternIn Appendix Table AII we present results from profitability (and size)

quintiles for large issuances defined using the 10 cutoff (in Panel A) andthe 15 cutoff (in Panel B) The results from these panels are largely con-sistent with those described above

52 MAGNITUDES OF ISSUANCES AND EQUITY CHANGES

In the previous section we considered the probability of having a nontriviallevel of debt or equity activity The next question is how large are the dollar

1432 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

values involved In Table V we sort firms according to profits and thentabulate the levels and changes in both debt and equity We do this firstfor all firms and then for small versus large firmsFor the analysis on all firms we observe in Column (1) that debt peaks at

the middle of the distribution This is because firms with medium profitabil-ity are also the largest as seen in Column (8) which reports average assetvalues for different profitability quintiles When we sort first by firm size andthen examine firms with different levels of profitability within size quintileswe find that for small firms debt is roughly independent of profits while forlarge firms less profitable firms have a higher level of debtAs expected Columns (3) and (5) show that more profitable firms have

higher equity values Columns (2) and (7) consider the relationship betweenissuances of debt and equity and firm profitability Debt issuances are sig-nificantly larger for more profitable firms In contrast equity issuances arelarger for low-profit firms Small low-profit firms issue substantially lessequity compared with small high-profit firms Among large firms the low-profit ones tend to issue equity while the high-profit ones tend to repurchaseequity The fact that more profitable firms issue debt and repurchase equitywhile the least profitable firms retire debt and issue equity is consistent withthe predicted relation between profitability and financing decisions under thetrade-off theoryWe also examine two-way sorts by size and profitability and report results

for the smallest and largest quintile of firms For the smallest quintile offirms the change in debt is largely unrelated to profits But for large firmsthere is a positive relation between profits and debt issuances High-profitfirms have a big positive change in debt Low-profit large firms have anegative change in debtColumns (4) and (6) provide an explanation of why the leverage ratio

regression results contradict those from the basic profitability sorts pre-sented here As we can see profitability indirectly affects leverage ratiosby increasing equity values Changes in both the book value of equity andthe market value of equity are positive and large for highly profitable firmsIn contrast these changes are negative for less profitable firms This evidenceillustrates an important issue concerning the use of leverage ratios Suchratios are often interpreted as essentially reflecting the use of debt by thefirm This interpretation while common is empirically misleadingFor the typical firm the change in the value of equity is larger than the

change in debt For example in the third profit quintile for large firmscolumn (7) the mean equity issue is just $9 million but the change in themarket value of equity is $242 million At the same time the mean change indebt is $27 million This suggests that a fair bit of the observed variation in

PROFITSndashLEVERAGE PUZZLEREVISITED 1433

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

the leverage ratios is primarily driven by the changes in the market value ofequity in the denominator rather than by the changes in debt in the numer-ator Since equity issues are often small this implies that the variation in theleverage ratio is primarily driven by internal operations rather than byexternal financing actions This again points to the fact that leverageratios can provide a misleading account of actual patterns in the dataThere is always a potential concern that the averages may be misleading

due to the impact of outliers To address this concern in AppendixTable AIII we present median values of the profitability sorts Theseresults generally reinforce the findings in Table VTable VI again considers the magnitudes of financing activity but this

time the issues are scaled by a measure of firm size That way it is easier to

Table V Magnitude of financing activity

The sample contains nonfinancial firms listed on the annual Compustat files for the period

from 1971 to 2009 The table presents average debt and equity changes in debt and equityand equity issuances (in millions of dollars) for firms sorted on profitability within sizeclasses The table reports information for all firms sorted on profitability and for profit-

ability sorts within the smallest and largest firms The sorts are done annually The tablealso reports results from t-test for the difference in percentage of firms issuing or retiringsecurities in the lowest and the highest profit quintile

D D BVE BVE MVE MVE EqyIss Assets

(1) (2) (3) (4) (5) (6) (7) (8)

Allfirm

s

Low 83 5 78 10 120 5 7 234

2 605 13 509 24 721 8 17 1691

3 1227 9 1123 22 1834 3 9 3551

4 833 23 1166 7 2408 74 9 2994

High 570 53 1171 71 3817 68 36 2581

t Highfrac14Low 31 9 36 9 38 2 16 36

Smallfirm

s Low 25 04 42 01 332 28 33 97

2 29 06 55 02 278 16 21 116

3 31 07 63 01 212 05 12 127

4 30 06 70 05 179 10 08 136

High 24 06 84 14 234 17 07 142

t Highfrac14Low 04 15 153 85 73 11 201 107

Largefirm

s Low 3929 55 2990 152 4671 11 63 10846

2 3826 19 3178 95 5122 66 11 10721

3 3074 27 3498 41 6173 242 9 9697

4 2623 62 3915 45 8473 94 57 9855

High 2180 193 4642 261 14924 202 148 10171

t Highfrac14Low 94 64 108 87 246 12 169 16

1434 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

see whether large firms behave disproportionately when compared withsmall firms Column (1) shows that low-profit firms have more debtas reported in many previous studies In Columns (4) and (6) we seethat low-profit firms have a major decline in both the book value ofequity and the market value of equity The high-profit firms have thereverse experienceMore interesting are the active decisions of the firms In Column (2) we see

that in contrast to the usual impression the low-profit firms are reducingdebt while the high-profit firms are increasing it In Column (7) we see thatthe low-profit firms are also reducing leverage by issuing equity The high-profit firms issue little equity In the lower panel similar results are foundwhen we separate small and large firms Equity issuing is particularly strong

Table VI Magnitude of financing activitymdashscaled levels and changes

The sample contains nonfinancial firms listed on the annual Compustat files for the period

from 1971 to 2009 The table presents average debt and equity changes in debt and equityand equity issuances scaled by assets for firms sorted on profitability within size classes Thetable reports information for all firms sorted on profitability and for profitability sorts

within the smallest and largest firms The sorts are done annually The table also reportsresults from t-test for the difference in percentage of firms issuing or retiring securities inthe lowest and the highest profit quintile

D D BVE BVE MVE MVE EqyIss

(1) (2) (3) (4) (5) (6) (7)

Allfirm

s

Low 386 122 12 871 3427 1868 217

2 285 58 407 119 940 157 28

3 293 12 425 14 742 03 12

4 269 26 461 02 955 27 10

High 192 15 551 47 1675 74 07

t Highfrac14Low 35 49 50 47 127 42 157

Smallfirm

s Low 545 154 841 2145 9688 9128 696

2 684 49 712 1915 5014 2318 288

3 331 253 141 969 3063 954 129

4 267 181 357 175 1673 221 55

High 194 05 543 26 1758 147 41

t Highfrac14Low 49 17 51 36 88 27 99

Largefirm

s Low 350 30 306 58 511 55 09

2 346 14 335 13 523 02 04

3 353 165 332 71 711 29 01

4 280 01 414 04 909 09 05

High 227 15 477 19 1643 27 14

t Highfrac14Low 371 112 453 181 570 57 221

PROFITSndashLEVERAGE PUZZLEREVISITED 1435

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

among the small low-profit firms that are also experiencing sharp reductionsin the value of equity In column (2) we see that both for the large and for thesmall firms net debt increases are primarily found in the highest quintile offirm profitability

6 Debt and Equity Regressions

So far we have demonstrated that in our data the conventional leverageregressions have the usual signs We then demonstrated that in simple sortsof the data strikingly different active financing patterns emerge The nexttask is to reconcile these differencesTable VII presents simple regressions predicting changes in debt book

equity and market equity as well as equity issuances Consistent with thesorts we control for changes in firm size and we also include firm and yeardummies There may be a concern about the speed of a firmrsquos responsesAccordingly we consider the lagged change in profits as the main case butwe also include the contemporaneous change in profits Whether the contem-poraneous change in profits is included or not does not in fact matter Wereport panel-robust standard errors adjusted for clustering at the firm levelThe results in Table VII support the findings in the sorts Firms that have

an increase in profits in 1 year increase debt both in the same year and in thenext year Firms that have an increase in profits experience an increase inboth the book value of equity and the market value of equity Importantlyhowever firms that experience and increase in profits reduce their equityissues All these effects are statistically significant (except for the marketvalue of equity in the column 5 specification) and seem intuitively reason-able In Appendix Table AIV similar regressions to those presented aboveare reported but without firm-fixed effects We find that empirically verylittle changes whether we include firm fixed effects or leave them out Thusthe targeting behavior to the extent that it happens does not account forwhat we are reportingThe next step is to bring the sorts and the simple regressions together into

a common setting We therefore include factors in addition to profitability(i) median industry leverage (ii) the market-to-book assets ratio (iii) tangi-bility of assets and (iv) firm size (measured by the log of assets) Rajan andZingales (1995) show that these factors are related to leverage in G7countries A number of studies have used these factors to estimateleverage targets Frank and Goyal (2009) show that these factors arerobustly related to leverage in the USA In these regressions we use indica-tors for the quintile that the firm is in for each factor

1436 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

Table VIII reports these results In Columns (1) and (2) we examine debtissues and again show that profits positively affect debt issuances The effectsare large and statistically significant at the 1 level Columns (3) and (4)which examine book equity show a strong effect of profits This effect is onlyslightly reduced by the inclusion of other factors Columns (5) and (6) illus-trate that the impact of profits on the change in market equity is also robustto the inclusion of the conventional factors Columns (7) and (8) examineequity issuances Here we again find that profits have a significant negativeimpact on equity issuances As before the results are robust to the inclusionof the conventional factorsIn unreported tables we estimate debt and equity changes using quantile

regressions In terms of the signs and significance of the profitability variable

Table VII Debt and equity changes

The sample contains nonfinancial firms listed on the annual Compustat files for the period

from 1971 to 2009 The table presents fixed effect estimates from regressions of changes indebt (D) changes in the book value of equity (BVE) changes in the market value ofequity (MVE) and net equity issuances (EqyIss) The explanatory variables include

current and lagged change in operating income before depreciation (Profits andProfitst1) and lagged change in the book value of assets (Assetst1) In addition tofixed firm effects the regressions include year indicator variables The reported t-statistics

are corrected for clustering at the firm level

aMean significant at the 1 level

bMean significant at the 5 level

cMean significant at the 10 level

D BVE MVE EqyIss

(1) (2) (3) (4) (5) (6) (7) (8)

Profitst1 022a 025a 015b 035a 004 060b 005b 007a

(33) (31) (25) (35) (03) (25) (25) (53)

Profits 050b 078a 140a 003a

(24) (32) (48) (34)

Assetst1 000 007b 025a 001b

(00) (21) (35) (25)

Constant 167 283a 34 51 1145a 1182a 605a 595a

(14) (30) (04) (05) (32) (29) (69) (74)

Year FE Yes Yes Yes Yes Yes Yes Yes Yes

Firm FE Yes Yes Yes Yes Yes Yes Yes Yes

R2 001 004 001 010 001 005 002 003

N 162056 162056 162130 162130 157550 157550 162154 162154

PROFITSndashLEVERAGE PUZZLEREVISITED 1437

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

the results are qualitatively similar to the OLS results reported in Table VIIIProfitability positively affects debt issuances and negatively affects equityissuances across various quantiles Overall the results in Table VIII showthat there is nothing that is special about using sorts or regressions to explaindebt and equity issues In either case we find that more profitable firms tendto increase their debt experience an increase in the value of equity andrepurchase shares Thus the control factors are not responsible for theusual rejection of the trade-off theory It would appear that what mattersis the size of the issuing activity relative to the organic increases in equityvalue due to profits In other words what matters in the leverage regressionseems to be coming from the presence of E in D

DthornE and not from D

Table VIII Debt and equity issuances profitability and leverage factor quintiles

The sample contains nonfinancial firms listed on the annual Compustat files for the period

from 1971 to 2009 The table presents estimates from regressions of change in debt (D)change in the book value of equity (BVE) change in the market value of equity (MVE)and net equity issuances (EquityIss) The explanatory variables include lagged profitability

quintiles and lagged leverage factor quintiles In addition to fixed firm effects the regres-sions include year indicator variables The reported t-statistics are corrected for clustering atthe firm level

aMean significant at the 1 level

bMean significant at the 5 level

cMean significant at the 10 level

D BVE MVE EqyIss

(1) (2) (3) (4) (5) (6) (7) (8)

ProfitabilityQint 214a 167a 283a 175a 80 429a 53a 55a

(70) (54) (66) (59) (08) (48) (55) (62)

IndLevQint 104a 43 99 20

(29) (15) (10) (10)

Ln(Assets)Qint52 17 1026a 53a

(16) (05) (77) (35)MB

Qint158a 372a 1155a 09

(45) (56) (106) (06)

TangibilityQint 05 177 436c 64a

(01) (15) (18) (31)

Constant 167 96 656b 235 4613a 8610a 113 570a

(08) (04) (26) (04) (61) (58) (11) (31)

Year FE Yes Yes Yes Yes Yes Yes Yes Yes

Firm FE Yes Yes Yes Yes Yes Yes Yes Yes

R2 0002 0002 0003 0004 0013 0014 0013 0014

N 155587 155037 155587 155037 155587 155037 155587 155037

1438 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

7 Are Financial Market Conditions Important

According to the trade-off theory capital structure is determined by con-sidering the impact of costs and benefits of debt The time variation inthe costs and benefits of debt imply that issuance decisions would alsovary over time Furthermore market conditions affect the transactioncosts of adjusting Adverse selection is a greater problem in a cold marketthan in a hot one Accordingly it may be easier to adjust in good marketconditions Related ideas have been developed by Huang and Ritter(2009) Faulkender et al (2012) and Halling Yu and Zechner (2012)Faulkender Flannery Hankins and Smithrsquos paper examines cross-sectionalvariation in adjustment speeds including market valuations andhow they may affect adjustment speeds It is natural to think that ingood market conditions it will be less costly to issue both debt and equityand adjusting leverage to respond to profitability shocks would be mucheasierTo test the importance of market conditions we require a definition of

good times and bad times Our empirical strategy is to estimate good timesversus bad times at the four-digit industry level We define an industry ashaving ldquogood timesrdquo if the median firm in that industry has a market-to-book ratio that is higher than the 67th percentile of the time-series distribu-tion of the industry median market-to-book ratios Conversely an industryis defined as having ldquobad timesrdquo if the median firm in the industry has amarket-to-book ratio that is lower than the 33rd percentile of the time-seriesdistribution of the industry median market-to-book ratiosPanel A of Table IX tabulates issuance activity for profitability sorts in

both good times and bad times As might be expected active debt and equityissuances are larger during good times There is also somewhat more activeswaps between debt and equity during good timesPanel B reports the issuance activity for the smallest and the largest firms

during the good times and bad times Within each size quintile firms aresorted on profitability This is a three-way sort As before we find that thereis much more active use of external markets during good times Small low-profit firms are more likely to issue equity in good times than in bad timesDebt issuances are significantly higher in good times Large high-profitfirms are significantly more likely to issue debt and repurchase equity ingood times than in bad timesPanel C reports the magnitudes of the financial variables rather than the

frequencies During bad times less profitable large firms retire substantialamounts of debt and they show a tendency to issue equity Small firms donot seem to engage in similar debt reduction activities However like the

PROFITSndashLEVERAGE PUZZLEREVISITED 1439

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

Table IX Debt and equity issues in good and bad times

The table reports the frequency and magnitude of financing activity for sorts on profits for subsamples

of firms in good and bad times An industry is defined as having ldquogood timesrdquo if the median firm in

that industry has a market-to-book ratio that is higher than the 67th percentile of the time-series

distribution of industry median market-to-book ratios Conversely an industry is defined as having

bad times if the median firm in that industry has a market-to-book ratio that is lower than the 33rd

percentile of the time-series distribution of industry median market-to-book ratios Panels A and B

report the percentage of firms (i) issuing net debt in excess of 5 of the value of their assets (ii) issuing

net equity in excess of 5 of the value of their assets (iii) issuing debt and repurchasing equity both in

excess of 5 of the value of their assets and (iv) issuing equity and retiring debt both in excess of 5

of the value of their assets Panel C tabulates the mean debt and equity levels and the mean changes in

debt and equity issuances The sample contains nonfinancial firms listed on the annual Compustat files

for the period from 1971 to 2009

Iss Iss Iss D Iss ENet D Net E Rep E Ret D(1) (2) (3) (4)

Panel A sorts on profitability

Badtimes

Low 145 131 01 202 170 49 02 083 173 46 04 074 182 47 03 10

High 177 51 06 10t Highfrac14Low 41 136 33 01

Goodtimes

Low 206 381 01 462 197 148 02 233 212 103 04 184 234 92 10 18

High 216 92 16 14t Highfrac14Low 23 658 138 23

Panel B sorts on firm size and profitability

Badtimes

Smallfirm

s Low 151 318 00 312 182 276 02 253 151 140 02 294 183 73 00 19

High 177 80 00 11t Highfrac14Low 10 90 NA 20

Largefirm

s Low 172 54 01 042 148 39 04 053 150 43 04 044 156 33 03 06

High 147 28 05 03t Highfrac14Low 17 32 20 02

Goodtimes Smallfirm

s Low 213 448 01 542 230 443 01 473 222 325 00 474 203 197 02 38

High 190 153 02 31t Highfrac14Low 19 230 10 39

Largefirm

s Low 203 96 03 122 208 66 09 133 232 54 16 064 215 39 19 05

High 234 36 38 04t Highfrac14Low 29 99 92 38

(continued)

1440 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

larger firms there is a tendency for small firms to issue more equitymdashat leastto the extent possibleDuring good times large profitable firms raise significant amounts of debt

as they experience an increase in the value of their equity Such firms alsoengage in share repurchases Debt issues by smaller firms are much lessaffected by their own profitability during good times and their change indebt is more or less independent of their profits However equity issuanceis primarily found among the low-profit small firms during good timesTable IX shows that empirically profitability has a time-varying impacton leverage choicesIn Table X the basic regressions for changes in debt and for equity

issuance are presented for both good times and bad times Comparinggood times with bad times we see that the effects of profits are muchstronger in good times In bad times the impacts are rather weak statisticallyThis difference is interesting and deserves further study However for ourpurposes recall that the static trade-off theory implies time-varying capital

Table IX Continued

D D BVE BVE MVE MVE EqyIss Assets

(1) (2) (3) (4) (5) (6) (7) (8)

Panel C magnitude of financing activity

Badtimes Smallfirm

s Low 40 03 58 14 305 06 22 140

2 44 05 82 07 240 67 17 180

3 45 04 94 02 176 18 05 187

4 43 09 98 02 137 04 03 188

High 36 10 114 13 209 16 04 199

Largefirm

s Low 3365 73 2523 169 2304 163 94 8972

2 3031 95 2983 35 3284 278 56 9068

3 2930 5 3360 14 3947 308 57 9225

4 2281 6 3817 19 5213 442 7 8902

High 1790 62 4767 13 8799 58 31 9573

Goodtimes

Smallfirm

s Low 24 06 38 05 339 23 39 90

2 26 08 48 01 301 04 25 104

3 30 09 60 04 246 03 17 125

4 28 07 66 08 206 18 12 130

High 23 06 85 16 261 13 08 141

Largefirm

s Low 4532 36 3477 103 6557 260 53 12465

2 4580 58 3527 139 6911 421 23 12522

3 2952 111 3485 2 7586 279 32 9535

4 2752 124 3939 72 10686 33 105 10226

High 2429 228 4819 377 18565 268 222 10830

PROFITSndashLEVERAGE PUZZLEREVISITED 1441

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

structure choices even when the target is time invariant Empirically we doobserve time-varying choices

8 Is Scaling Important

In the preceding section we provided results for unscaled debt and equityissues This seems appropriate to us However it involves making twochanges to the standard regression We would like to determine if both

Table X Debt and equity changes in good and bad timesmdashregressions

The sample contains nonfinancial firms listed on the annual Compustat files for the period

from 1971 to 2009 The table presents estimates from regressions of change in debt (D)and net equity issuances (EqyIss) in good and bad times The explanatory variables includechange in profits lagged change in profits and lagged change in assets The regressions

include both firm and year fixed effects An industry is defined as having ldquogood timesrdquo ifthe median firm in that industry has a market-to-book ratio that is higher than the 67thpercentile of the time-series distribution of industry median market-to-book ratios

Conversely an industry is defined as having bad times if the median firm in thatindustry has a market-to-book ratio that is lower than the 33rd percentile of the time-series distribution of industry median market-to-book ratios The reported t-statistics arecorrected for clustering at the firm level

aMean significant at the 1 level

bMean significant at the 5 level

cMean significant at the 10 level

Good times Bad times

D EqyIss D EqyIss

(1) (2) (3) (4) (5) (6) (7) (8)

Profits 087a 091a 006b 009a 060 065 003 003b

(30) (29) (20) (28) (15) (16) (16) (20)

Profitst1 029c 015a 034c 001

(19) (37) (17) (04)

Assetst1 000 002b 000 001

(02) (22) (00) (07)

Constant 1337 3989 8691a 10001a 1351 1851 4099a 4055a

(03) (09) (55) (37) (09) (13) (80) (78)

Year FE Yes Yes Yes Yes Yes Yes Yes Yes

Firm FE Yes Yes Yes Yes Yes Yes Yes Yes

R2 0071 0080 0018 0045 0178 0212 0034 0041

N 74827 68729 74891 68787 26965 26013 26971 26018

1442 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

changes are crucial Accordingly in this section we take a step back toconsider prior literature in which variables are scaled The questions arehow does scaling affect the results and whyTable XI examines alternative scalings Columns (1) and (2) scale debt and

equity issuances by the total firm net issuance Columns (3) and (4) scaledebt and equity issuances by the sum of debt and book equity Columns (5)and (6) consider the change in debt scaled by the sum of debt and bookequity Columns (7) and (8) consider the change in leverage due to externalfinancing Since we are examining changes we use the changes in thestandard regression factors as controls The primary interest is in the coef-ficient on the change in profitsWe find that a change in profits is associated with an increase in debt

issues But the sign reverses when we scale the debt issues by the sum of debtplus book equity This is telling What is really driving the results is thechange in equity in the denominator not the change in debt in the numer-ator This result is further substantiated in Columns (5) and (6) where thechange in debt to capital ratio is used instead of debt issuanceTo further explore the impact of active issuance relative to passive

experiencing of changes we decompose leverage by separating out theeffect of debt and equity issues from those due to changes in retainedearnings Thus we define leverage change due to external financing as

LevEF frac14Dt1 thorn d

Dt1 thorn Et1 thorn dthorn e

Dt1

Dt1 thorn Et1 eth1THORN

where D is debt E is book value of equity d is debt issues net of retirementsand e is equity issues net of repurchases Garvey and Hanka (1999) use asimilar measure in their study This measure directly captures the netleverage effect of debt and equity issuancesColumns (7) and (8) report results from regressions of the net leverage

effect of debt and equity issues on changes in profitability after controllingfor other leverage factors in differences and year indicator variables Theeffect of profitability on leverage change due to external financing is positiveand statistically significant This comparison suggests that changes in totaldebt and equity are only partly a result of debt and equity issuance decisionsOther balance sheet adjustments complicate the inferences from leverageratio regressionsFor most of our analysis we follow the literature in treating debt as homo-

genous It is of course potentially interesting to ask what happens when wedistinguish between short-term debt and long-term debt Which part of thefirmrsquos debt maturity structure drives the under-adjustment Table XII

PROFITSndashLEVERAGE PUZZLEREVISITED 1443

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

provides results from such decompositions Alternative scalings areconsidered The main results of interest are that the distinction betweenshort- and long-term debt is important Short-term debt issuance is largelyindependent of changes in profits The negative sign on profit change stemsfrom the action of long-term debt This seems intuitively reasonable

Table XI Scaled issuances

The sample contains nonfinancial firms listed on the annual Compustat files for the period

from 1971 to 2009 The table presents estimates from regressions of debt issuance to totalissuance (in columns (1) and (2)) debt issuance to total capitalization (in columns (3) and(4)) change in debt to capitalization ratio (in columns (5) and (6)) and change in leverage

ratio considering only external financing (in columns (7) and (8)) The change in leverageratio considering only external financing is defined as

LevEF frac14Dt1 thorn d

Dt1 thorn Et1 thorn dthorn e

Dt1

Dt1 thorn Et1

where D is debt E is book value of equity d is debt issues net of retirements and e is equity

issues net of repurchases The regressions include lagged leverage factors and year indicatorvariables Some of the specifications include firm fixed effects The reported t-statistics arecorrected for clustering at the firm level

aMean significant at the 1 level

bMean significant at the 5 level

cMean significant at the 10 level

DebtIssTotIss

DebtIssethDthornETHORN D

DthornE LevEF

(1) (2) (3) (4) (5) (6) (7) (8)

Profitt1 004a 000 002a 001c 003a 002 001b 002a

(35) (03) (34) (18) (35) (14) (23) (26)

IndLevt1 004c 001 001 001c 000 000 000 000

(19) (06) (15) (19) (02) (04) (03) (05)

Market=bookt1 001a 001a 000a 000a 000a 000a 000 000c

(47) (85) (68) (60) (55) (36) (05) (18)

Tangt1 004 005 013a 011a 011a 006a 003a 003a

(12) (15) (144) (112) (79) (38) (44) (42)

Assetst1 008a 004a 006a 004a 002a 001a 003a 002a

(97) (45) (283) (161) (53) (35) (147) (101)

Constant 079a 074a 002a 003a 000 000 000b 002a

(497) (442) (91) (111) (00) (05) (22) (93)

Year FE Yes Yes Yes Yes Yes Yes Yes Yes

Firm FE No Yes No Yes No Yes No Yes

R2 0015 0098 0022 0076 0009 0070 0012 0086

N 133905 133905 140552 140552 138229 138229 140572 140572

1444 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

inasmuch as short-term debt is probably more associated with day-to-dayoperations rather than longer term planning considerationsIn Table XIII we examine change in leverage ( D

DthornE) and change inleverage due to external financing (LEVEF) during good and bad timesseparately The traditional negative effect of a change in profits on achange in leverage is found both in good times and in bad timesHowever when we decompose the impact in order to isolate the externalcomponent as before the result disappears In good times the externalfinancing component of a change in profits is weak but still positiveDuring bad times it is not statistically significant This implies that distin-guishing between good and bad times does not fundamentally alter the basic

Table XII Short- and long-term debt issuances

The sample contains nonfinancial firms listed on the annual Compustat files for the period

from 1971 to 2009 The table presents estimates from regressions of short-term debtissuance to total issuance (in columns (1) and (2)) long-term debt issuance to totalissuance (in columns (3) and (4)) short-term debt issuance to total capitalization (in

columns (5) and (6)) and long-term debt to total capitalization (in columns (7) and (8))Some of the specifications include firm fixed effects The reported t-statistics are correctedfor clustering at the firm level

aMean significant at the 1 level

bMean significant at the 5 level

cMean significant at the 10 level

STDebtIssTotIss

LTDebtIssTotIss

STDebtIssethDthornETHORN

LTDebtIssethDthornETHORN

(1) (2) (3) (4) (5) (6) (7) (8)

Profitabilityt1 0002 0003 0042a 0008 0001 0001 0019a 0012a

(02) (04) (30) (05) (14) (10) (45) (28)

IndLevt1 0009 0007 0070b 0033 0001 0002c 0014a 0005

(06) (05) (26) (11) (08) (17) (26) (09)

M=Bt1 0003a 0003a 0003b 0010a 0000 0000 0003a 0002a

(48) (35) (22) (65) (10) (01) (84) (42)

Tangibilityt1 0020 0007 0011 0067 0005a 0006a 0097a 0081a

(12) (03) (03) (16) (34) (33) (108) (89)

Assetst1 0039a 0011b 0034a 0024b 0001a 0001 0046a 0020a

(89) (22) (38) (23) (36) (13) (190) (82)

Constant 0001b 0153a 0798a 0591a 0000c 0006a 0057a 0096a

(25) (144) (470) (288) (17) (124) (292) (234)

Year FE Yes Yes Yes Yes Yes Yes Yes Yes

Firm FE No Yes No Yes No Yes No Yes

R2 0025 0089 0018 0073 0030 0147 0016 0298

N 136146 136146 136140 136140 143031 143031 142847 142847

PROFITSndashLEVERAGE PUZZLEREVISITED 1445

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

Table XIII Externally financed leverage changes in good and bad times

The sample contains nonfinancial firms listed on the annual Compustat files for the period

from 1971 to 2009 The table presents estimates from regressions of change in debt tocapitalization ratio (in columns (1) (2) (5) and (6)) and change in leverage ratio consider-ing only external financing (in columns (3) (4) (7) and (8)) The change in leverage ratio

considering only external financing is defined as

LevEF frac14Dt1 thorn d

Dt1 thorn Et1 thorn dthorn e

Dt1

Dt1 thorn Et1

where D is debt E is book value of equity d is debt issues net of retirements and e is equityissues net of repurchases The regressions include lagged leverage factors and year indicator

variables The reported t-statistics are corrected for clustering at the firm level An industryis defined as having ldquogood timesrdquo if the median firm in that industry has a market-to-bookratio that is higher than the 67th percentile of the time-series distribution of industry

median market-to-book ratios Conversely an industry is defined as having ldquobad timesrdquoif the median firm in that industry has a market-to-book ratio that is lower than the 33rdpercentile of the time-series distribution of industry median market-to-book ratios

aMean significant at the 1 level

bMean significant at the 5 level

cMean significant at the 10 level

Good times Bad times

DDthornE LevEF D

DthornE LevEF

(1) (2) (3) (4) (5) (6) (7) (8)

Profitabiltyt1 0022c 0005 0016c 0025b 0073b 0041 0007 0007

(17) (03) (18) (24) (25) (11) (04) (03)

IndLevt1 0000 0011 0001 0009 0010 0018 0001 0000

(00) (09) (01) (13) (07) (10) (02) (00)

MB

t1

0004a 0003b 0000 0001 0008a 0001 0000 0000

(40) (24) (05) (08) (28) (04) (03) (03)

Tangt1 0097a 0041 0033a 0038b 0127a 0048 0059a 0033

(45) (16) (27) (25) (49) (14) (43) (16)

Assetst1 0014a 0013b 0024a 0019a 0037a 0023c 0036a 0025a

(30) (20) (84) (48) (45) (17) (88) (43)

Constant 0030b 0036 0000 0003 0028c 0014 0021 0001

(20) (16) (00) (07) (17) (06) (11) (01)

Year FE Yes Yes Yes Yes Yes Yes Yes Yes

Firm FE No Yes No Yes No Yes No Yes

R2 0006 0088 0009 0117 0026 0251 0023 0202

N 56820 56820 58200 58200 23462 23462 23511 23511

1446 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

conclusions The traditional leverage result on profits is driven by increasesin equity that are experienced and then partially but not completely offsetby firmsTo the extent that firms are engaging in active financing they are following

the prescriptions of the trade-off theory Thus we have a coherent picture ofhow firms are responding to profits Basically more profitable firms experiencean increase in the value of their equity They respond to this by issuing debtand repurchasing equity But the magnitude of the increase in equity is largerthan the magnitudes of the active steps taken in the debt and equity marketsThis means that the firmsrsquo actual actions are as predicted by the trade-offtheory But it leaves open the question of why they do not act more stronglyThe capital structure literature has long been interested in dynamics and

capital structure adjustment paths In recent years this interest has increas-ingly resulted in explicit dynamic optimizing models such as those ofGoldstein Ju and Leland (2001) Hennessy and Whited (2005) andStrebulaev (2007) These models commonly make specialized assumptionsleading some to question their generalityWe argue that both fixed transactions costs and variable transactions costs

are plausible and realistic in models that pin down firm size The simultaneouspresence of both is important Due to fixed costs firms only rebalance inter-mittently Due to variable adjustment costs the extent of the adjustment isreduced to the point at which an extra unit to adjustment is balanced againstthe extra unit transactions costs This means that both leverage increases andleverage reductions will only go part way toward a static optimumThe key point is that with both fixed and variable transactions costs we

should expect to only see intermittent leverage adjustments When leverage isadjusted it should ldquounder adjustrdquo when compared to a frictionless modelFurther discussion of transactions costs in a capital structure setting isprovided by Leary and Roberts (2005)

9 Conclusion

The connection between corporate profits and capital structure has beenvery influential in the assessment of the trade-off theory The conventionalinterpretation of the evidence has pushed the literature away from the trade-off theory and toward much more complex dynamic models and ideas As aresult it is important to make sure that the evidence has been correctlyinterpreted We show that the problem is with the interpretation of thetest results and a failure to take into account the costs associated with re-sponding to a profit shock

PROFITSndashLEVERAGE PUZZLEREVISITED 1447

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

We document that more profitable firms really do borrow more (not less)They tend to repurchase equity They experience an increase in both thebook value of equity and the market value of equity Less profitable firmsreally do tend to reduce their debt and to issue equityFirm size and market conditions also matter Larger firms tend to be more

active in the debt markets while smaller firms tend to be relatively moreactive in the equity markets During good times there is more use ofexternal financingThe usual profitsndashleverage puzzle result is primarily driven by the increase

in equity that is experienced by the more profitable firms Accordingly thepuzzle should be restated as asking why do firms not take sufficiently largeoffsetting actions to fully undo the change in equity What limits the mag-nitudes of the typical leverage response to profit shocksIn a frictionless model the partial response appears puzzling But there is

good empirical reason to believe that rebalancing entails both fixed andvariable costs and that firm size matters The rebalancing costs can befully avoided by doing nothing Accordingly the firm must decide whetherany given shock is big enough to be worth responding to If it is then thefirm must decide how big a response is called for These technical conditionsare known as ldquovalue matchingrdquo and ldquosmooth pastingrdquo Thus optimizationimplies that some shocks will be ignored Even if the shock is not ignoredthe optimal response will only partially undo the shock The magnitude ofthe leverage response must balance the marginal cost and the marginalbenefit of an extra unit of leverage Since the marginal cost of adjustingleverage is strictly positive the adjustment toward that static leverageoptimum will only go part way This is true both for leverage increasesand for leverage reductionsOne might conjecture that the variable costs would be time varying and

that such costs might have declined over time Since we do not have a fullaccounting of all of the variable costs it is of course impossible to be sureHowever there is some evidence suggesting that at least the directunderwriting fees are remarkably stable over time10

Surely nobody ought to be surprised that more profitable firms are moreinclined to repurchase equity while unprofitable firms tend not to do so Thefact that more profitable firms typically experience an increase in the value of

10 For IPOs Jay Ritter provides important summary evidence as httpbearwarringtonufleduritterIPOs2013Underwritingpdf He finds that for medium-sized deals a remark-

ably large fraction of the deals have a 7 fee year after year Whether this is due tobehavioral factors or some sort of a coordination device is not clear The extent towhich similar intertemporal stability carries over to other forms of external financemerits further study

1448 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

equity is equally natural Were it not for the contrary claims in the literature(ie the profitsndashleverage puzzle) we would argue that nobody ought to findit surprising that more profitable firms are more inclined to issue debt Thefact that leverage adjustments typically only partially offset profit shocksshould also sound quite natural These facts all fit together easily within thetrade-off theory once we take account of both fixed and variable costs ofactively adjusting capital structure

Appendix

Table AI Market leverage regressions

The table presents estimates of the regressions of market leverage on firm characteristics

The sample comes from the annual Compustat files during the period 1971ndash2009 Financialfirms are excluded The leverage is estimated as the ratio of debt over debt plus marketequity The explanatory variables Profitabilityt1

MB

t1

Tangibilityt1 and LnethAssetsTHORNt1are described in Table 1 IndMedianLevt1 is estimated as the median market leverage of allother firms in the same industry The industry is defined at the level of the four-digit SICcode All specifications include the year fixed effects The specifications in column (2)

additionally include the firm fixed effects We report t-statistics where the standard errorsare clustered at the firm level in parentheses

aSignificant at the 1 level

Quantile regressions

(1) (2) (3) (4) (5)

Profitabilityt1 013a 013a 007a 011a 020a

(275) (262) (349) (425) (538)

IndLevt1 046a 020a 032a 058a 063a

(506) (239) (1299) (1630) (1042)MB

t1

003a 001a 001a 002a 004a

(509) (303) (743) (709) (534)

Tangibilityt1 013a 018a 013a 014a 010a

(191) (181) (762) (571) (232)

LnethAssetsTHORNt1 001a 005a 001a 001a lt001a

(112) (258) (682) (342) (29)

Constant 010a 011a 005a 005a 029a

(171) (98) (173) (118) (398)

Year FE Yes Yes Yes Yes Yes

Firm FE No Yes No No No

Regression type OLS FE 25th tile Median 75thtile

Clustered SE Yes Yes No No No

R2 030 067

Pseudo R2 016 023 0201

N 158578 158578 158578 158578 158578

PROFITSndashLEVERAGE PUZZLEREVISITED 1449

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

Table AII Profitability sorts for firms issuing large amounts

Percentage of firms

Iss Ret Iss Iss Rep Iss Iss IssD IssE

D D NetD E E NetE DampE RepE RetD

(1) (2) (3) (4) (5) (6) (7) (8) (9)

Panel A 10 cutoff to define issuing or retiring activity

Allfirm

s

Low 205 150 134 260 05 257 45 00 23

2 207 175 103 77 06 75 11 01 10

3 243 196 101 50 08 47 08 01 09

4 252 186 113 50 13 48 08 02 08

High 227 161 118 60 27 57 07 04 09

t Highfrac14Low 70 38 62 755 226 768 320 90 152

Smallfirm

s Low 230 132 162 380 05 375 82 00 32

2 228 143 165 334 04 333 65 00 26

3 211 146 147 216 03 214 38 00 23

4 190 154 120 126 04 124 15 01 20

High 194 162 114 99 09 97 12 01 15

t Highfrac14Low 42 40 68 342 20 340 169 07 131

Largefirm

s Low 231 200 82 31 06 29 08 01 04

2 231 187 67 17 06 14 03 01 02

3 234 177 77 17 09 14 03 02 03

4 218 158 75 16 17 14 02 03 02

High 218 135 99 19 41 15 03 07 02

t Highfrac14Low 19 106 36 46 135 57 39 58 23

Percentage of firms

Iss Ret Iss Iss Rep Iss Both Iss D Iss E

D D Net D E E Net E DampE Rep E Ret D

(1) (2) (3) (4) (5) (6) (7) (8) (9)

Panel B 15 cutoff to define issuing or retiring activity

Allfirm

s

Low 158 101 100 230 01 228 27 00 15

2 142 115 63 57 01 56 05 00 06

3 165 128 57 32 01 31 03 00 05

4 173 118 65 35 01 34 03 00 05

High 160 102 70 46 04 44 03 00 05

t Highfrac14Low 08 06 141 738 87 746 277 24 131

Smallfirm

s Low 195 95 132 354 01 350 56 00 24

2 185 96 131 303 01 301 44 00 18

3 160 96 104 179 00 176 18 00 15

4 134 99 78 99 01 98 08 00 12

High 133 95 73 82 01 80 06 00 10

t Highfrac14Low 80 00 95 345 12 344 149 12 52

Largefirm

s Low 145 123 42 14 01 12 03 00 01

2 137 114 32 07 01 06 01 00 01

3 142 104 35 06 01 05 00 00 01

4 130 91 36 08 03 07 01 01 01

High 134 81 48 09 06 07 01 01 01

t Highfrac14Low 19 85 17 26 54 33 34 17 11

1450 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

Table AIV Debt and equity changesmdashwithout fixed effects

The sample contains nonfinancial firms listed on the annual Compustat files for the period from 1971

to 2009 The table presents estimates from regressions of change in debt (D) change in the book

value of equity (BVE) change in the market value of equity (MVE) and net equity issuances

(EqyIss) The explanatory variables include current and lagged change in operating income before

depreciation (Profits and Profitst1) and lagged change in book value of assets (Assetst1) The

regressions include year indicator variables The reported t-statistics are corrected for clustering at the

firm level

aMean significant at the 1 level

bMean significant at the 5 level

cMean significant at the 10 level

D BVE MVE EqyIss

(1) (2) (3) (4) (5) (6) (7) (8)

Profitst1 025a 024a 023a 038a 007 066a 007a 009a

(40) (32) (36) (39) (06) (29) (42) (51)Profits 051b 082a 149a 004a

(25) (33) (49) (33)Assetst1 002 005c 024a 001b

(08) (17) (34) (23)Constant 93 176a 45 71 1552a 1704a 130a 140a

(12) (28) (07) (09) (49) (49) (45) (45)Year FE Yes Yes Yes Yes Yes Yes Yes YesR2 001 005 001 011 001 005 002 002N 162056 162056 162130 162130 157550 157550 162154 162154

Table AIII Median profitability sorts

The sample contains nonfinancial firms listed on the annual Compustat files for the period from 1971

to 2009 The table presents median debt and equity median changes in debt and equity and median

equity issuances (EqyIss) in millions of dollars for all firms sorted on profitability and for profitability

sorts within the smallest and largest firms The sorts are done annually

D D BVE BVE MVE MVE EqyIss Assets(1) (2) (3) (4) (5) (6) (7) (8)

Allfirm

s

Low 21 00 75 17 256 20 01 1922 186 02 449 18 598 19 00 11033 816 06 1274 03 1637 05 00 32244 787 02 1541 18 2488 03 00 3556High 273 00 1361 66 3300 14 00 2578

Smallfirm

s Low 05 00 13 07 140 14 06 402 06 00 25 08 114 12 01 563 09 00 32 05 85 07 00 674 12 00 39 01 68 03 00 79High 10 00 51 04 98 01 00 90

Largefirm

s Low 11485 381 10269 371 13227 202 03 333902 13051 393 12786 140 17571 169 05 392463 11554 287 13429 24 21623 305 00 370314 9963 209 15052 194 29424 308 00 36705High 7361 00 16380 613 49142 439 00 34976

PROFITSndashLEVERAGE PUZZLEREVISITED 1451

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

References

Auerbach A J (1985) Real determinants of corporate leverage in B M Friedman (ed)

Corporate Capital Structures in the United States University of Chicago Press Chicago

IllinoisBertrand M and Mullainathan S (2003) Enjoying the quiet life corporate governance and

managerial preferences Journal of Political Economy 111 1043ndash1075Booth L Varouj A Demirguc-Kunt A and Maksimovic V (2001) Capital structures in

developing countries Journal of Finance 56 87ndash130Brennan M J and Schwartz E S (1984) Optimal financial policy and firm valuation

Journal of Finance 39 593ndash607Cameron A C and Trivedi P K (2010) Microeconometrics using Stata Stata Press TexasConstantinides G M (1979) Multiperiod consumption and investment behavior with

convex transactions costs Management Science 25 1127ndash1137Constantinides G M and Richard S F (1978) Existence of optimal simple policies for

discounted-cost inventory and cash management in continuous time Operations Research

26 620ndash636DeAngelo H DeAngelo L and Whited T M (2011) Capital structure dynamics and

transitory debt Journal of Financial Economics 99 235ndash261Dixit A K (1993) The Art of Smooth Pasting Routledge London and New YorkFama E and French K R (2002) Testing trade-off and pecking order predictions about

dividends and debt Review of Financial Studies 15 1ndash33Faulkender M and Petersen M A (2006) Does the source of capital affect capital struc-

ture Review of Financial Studies 19 45ndash79Faulkender M Flannery M J Hankins K W and Smith J M (2012) Cash Flows and

Leverage Adjustments Journal of Financial Economics 103 632ndash646Fischer E Heinkel R and Zechner J (1989) Dynamic capital structure choice theory

and tests Journal of Finance 44 19ndash40Frank M Z and Goyal V K (2008) Trade-off and pecking order theories of

debt in B E Eckbo (ed) Handbook of Corporate Finance Empirical Corporate Finance

Vol 2 Handbook of Finance Series Chapter 12 ElsevierNorth-Holland Amsterdam pp

135ndash202

Frank M Z and Goyal V K (2009) Capital structure decisions which factors are reliably

important Financial Management 38 1ndash37

Garvey G T and Hanka G (1999) Capital structure and corporate control the effect of

antitakeover statutes on firm leverage Journal of Finance 54 519ndash546

Goldstein R Ju N and Leland H (2001) An EBIT-based model of dynamic capital

structure Journal of Business 74 483ndash512

Graham J R and Tucker A L (2006) Tax shelters and corporate debt policy Journal of

Financial Economics 81 563ndash594

Hackbarth D and Mauer D C (2012) Optimal priority structure capital structure and

investment Review of Financial Studies 25 747ndash796

Halling M Yu J and Zechner J (2012) Leverage dynamics over the business cycle

Working paper University of Utah University of New South Wales and Vienna

University of Economics and Business AdministrationHennessy C A and Whited T A (2005) Debt dynamics Journal of Finance 60 1129ndash1165

Huang R and Ritter J R (2009) Testing theories of capital structure and estimating the

speed of adjustment Journal of Financial and Qualitative Analysis 44 237ndash271

1452 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

Hovakimian A Opler T and Titman S (2001) The debt-equity choice Journal of

Financial and Quantitative Analysis 36 1ndash24Jung K Kim Y and Stulz R M (1996) Timing investment opportunities managerial

discretion and the security issue decisions Journal of Financial Economics 42 159ndash185Kane A Marcus A J and McDonald R L (1984) How big is the tax advantage to

debt Journal of Finance 39 841ndash853Kayhan A and Titman S (2007) Firmsrsquo histories and their capital structures Journal of

Financial Economics 83 1ndash32Leary M T (2009) Bank loan supply lender choice and corporate capital structure

Journal of Finance 64 1143ndash1185Leary M T and Roberts M R (2005) Do firms rebalance their capital structures

Journal of Finance 60 2575ndash2619Leland H E (1994) Corporate debt value bond covenants and optimal capital structure

Journal of finance 49 1213ndash1252Lemmon M L Roberts M R and Zender J F (2008) Back to the beginning persist-

ence and the cross-section of corporate capital structure Journal of Finance 63 1575ndash1608Long M S and Malitz I B (1985) The investment-financing nexus some empirical

evidence Midland Corporate Finance Journal 3 53ndash59Mackie-Mason J K (1990) Do taxes affect corporate financing decisions Journal of

Finance 45 1471ndash1493Morellec E Nikolov B and Schurhoff N (2012) Corporate governance and capital

structure dynamics Journal of Finance 67 803ndash848Myers S C (1984) The capital structure puzzle Journal of Finance 39 575ndash592Myers S C (1993) Still searching for optimal capital structure Journal of Applied

Corporate Finance 61 4ndash14Rajan R and Zingales L (1995) What do we know about capital structure some evidence

from international data Journal of Finance 50 1421ndash1460Strebulaev I A (2007) Do tests of capital structure theory mean what they say Journal of

Finance 62 1747ndash1787

Strebulaev I A and Whited T M (2012) Dynamic Models and Structural Estimation in

Corporate Finance Now Publisher Boston and Delft

Titman S and Wessels R (1988) The determinants of capital structure choice Journal of

Finance 43 1ndash21

Welch I (2004) Capital structure and stock returns Journal of Political Economy 112

106ndash131

Welch I (2011) Two common problems in capital structure research the financial-debt-

to-asset ratio and issuing activity versus leverage changes International Review of Finance

11 1ndash17

PROFITSndashLEVERAGE PUZZLEREVISITED 1453

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

consideration Firm size is defined as the natural log of assets Other factorsare defined in Section 3 We cluster standard errors by firm and estimatethese regressions both with and without fixed effectsIn Table II we run conventional cross-section leverage regressions Since

we employ the usual data it is not surprising that our results match thosereported in the existing capital structure literature Profitability has anegative sign in both the book leverage regressions and in the marketleverage regressions The coefficients on other factors largely match thosereported in earlier studies Firms operating in industries with high leverage

Table II Leverage and profitability

The table presents parameter estimates for book leverage regressions The sample consists

of nonfinancial companies in the annual Compustat files during the period 1971ndash2009 Theleverage is estimated as the ratio of debt over debt plus book equity The explanatoryvariables Profitabilityt1

MB

t1

Tangibilityt1 and LnethAssetsTHORNt1 are described in

Table I IndLevt1 is estimated as the median book leverage of all other firms in thesame industry The industry is defined at the level of the four-digit SIC code All specifi-cations include the year fixed effects The specifications in column (2) additionally include

the firm fixed effects We report t-statistics where the standard errors are clustered at thefirm level in parentheses

aSignificant at the 1 level

Quantile regressions

(1) (2) (3) (4) (5)

Profitabilityt1 026a 020a 011a 017a 041a

(293) (192) (378) (518) (1015)

IndLevt1 042a 012a 030a 052a 052a

(369) (107) (926) (1252) (950)MB

t1

002a 001a 001a 002a 002a

(203) (102) (392) (426) (418)

Tangibilityt1 018a 025a 024a 019a 012a

(206) (154) (964) (580) (276)

LnethAssetsTHORNt1 001a 002a 002a 002a 001a

(114) (76) (899) (478) (96)

Constant 012a 010a 012a 002a 029a

(156) (60) (284) (27) (387)

Year FE Yes Yes Yes Yes Yes

Firm FE No Yes No No No

Regression type OLS FE 25th tile Median 75thtile

Clustered SE Yes Yes No No No

R2 013 054

Pseudo R2 016 015 009

N 158824 158824 158824 158824 158824

PROFITSndashLEVERAGE PUZZLEREVISITED 1427

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

tend to have high leverage A higher market-to-book ratio is associated withlower leverage Larger firms are typically more highly levered The coeffi-cients on tangibility are positive The conclusion from Table II is clear Theleverage regression results match those reported in previous studiesAt times one may be concerned that in a regression what is being

estimated is a conditional mean while some parts of the distributionmight behave differently from other parts If so then ldquoaccidentsrdquo likesample selection issues or asymmetries in the underlying distributionsmight play a greater role than is usually recognized In other words thesource of identification could be misinterpreted To guard against this po-tential problem we estimate the same model using quantile regressions9

Columns 3 4 and 5 report what happens at the 25th 50th and 75th per-centiles The basic model results are robust in terms of both the sign and thestatistical significanceThe impact of profits does appear to be somewhat stronger among the

high leverage firms For instance in a book leverage regression the coeffi-cient on profits for the first quartile is 011 with a t-ratio of 378 For thethird quartile the coefficient on profits is 041 with a t-ratio of 1015This difference is curious and might deserve further study However the factthat in each case we get the negative sign and statistical significance is suf-ficient for the purposes of the present article So the results on profits arerather robust across the distributionTable III sorts firms into those that are actively adjusting their leverage

and those that are passive In the presence of both fixed and variable trans-acts costs this distinction is important We expect the coefficient on profit-ability in leverage ratio regressions to be more negative in firms that arepassive and less negative in firms that are actively adjustingWe define active firms as those that are either issuing net debt or net

equity in excess of 5 of their assets Passive firms are those that neitherissue debt nor equity in nontrivial amounts (ie greater than 5 of theirassets) Consistent with our expectations we find in Table III that the coef-ficients on profits are significantly more negative for passive firms than theyare for active firms With firm and year fixed effects the coefficient onprofits is 008 for active firms and 017 for the passive firms The twocoefficients are significantly different from each other (plt 001) InTable III the standard control variables are used They have the usualsigns and for the most part do not differ much between the active and thepassive firms

9 Cameron and Trivedi (2010) provide an extensive discussion of quantile regressions inStata

1428 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

5 Financing Activity

51 FREQUENCY OF DEBT AND EQUITY ISSUES

Table IV reports the percentage of firms issuing or repurchasing debt orequity for annual sorts based on lagged profitability In Panel A weemploy the conventional 5 cutoff rule to exclude minor fluctuations anddefine firms ldquoissuing debtrdquo as those that issue debt (both short-term andlong-term) in excess of 5 of the value of their assets Other decisions aresimilarly defined using a 5 cutoff Appendix Table AII examines the dis-tributional properties across profitability (and size) quintiles for ldquolargerdquo is-suances by increasing the 5 cutoff to higher levels such as 10 and 15How does security issuance behavior vary with firm profitability To

answer this question Table IV sorts firms by profitability and reports the

Table III Leverage and profitability for active and passive firms

The table presents estimates of leverage on firm characteristics for the subsamples of firms

that are active in either debt or equity markets Active firms are defined as those that issueeither net debt in excess of 5 of the value of assets or issue net equity in excess of 5 ofthe value of their assets The passive firms do not issue significant (more than 5) of either

debt or equity The sample comes from the annual Compustat files during the period 1971ndash2009 Financial firms are excluded The variables are described in Tables I and II Wereport t-statistics where the standard errors are clustered at the firm level in parentheses

aSignificant at the 1 level

Active Passive

(1) (2) (3) (4)

Profitabilityt1 006a 008a 022a 017a

(126) (148) (254) (188)

IndLevt1 044a 019a 041a 013a

(424) (149) (326) (127)MB

t1

002a 001a 003a 001a

(362) (196) (327) (172)

Tangibilityt1 010a 014a 014a 019a

(146) (118) (145) (130)

LnethAssetsTHORNt1 001a 004a 001a 004a

(121) (179) (147) (169)

Constant 012a 008a 003a 007a

(120) (57) (45) (42)

Year FE Yes Yes Yes Yes

Firm FE No Yes No Yes

R2 034 064 033 075

N 43395 43395 58943 58943

PROFITSndashLEVERAGE PUZZLEREVISITED 1429

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

percentage of firms issuing or repurchasing debt or equity The table alsoreports test statistics from tests that examine whether the percentage differ-ences between the high- and low-profit firms are significantly different fromeach other The results in Column (1) show that the likelihood of issuing debt

Table IV Profitability sorts for debt and equity issuers

The sample contains nonfinancial firms listed on the annual Compustat files for the period

from 1971 to 2009 The table presents the percentage of firms issuing and retiring (orrepurchasing) debt and equity The firms are classified as ldquoissuing debtrdquo if they issuedebt in excess of 5 of the value of their assets as ldquoissuing equityrdquo if they issue equity

in excess of 5 of the value of their assets as ldquoretiring debtrdquo if they retire debt in excess of5 of the value of their assets and as ldquorepurchasing equityrdquo if they repurchase equity inexcess of 5 of the value of their assets In addition we also report net debt issuers which

are firms that issue net debt over 5 of the value of their assets and net equity issuerswhich are firms that issue net equity in excess of 5 of the value of their assets Weannually sort firms on lagged profitability and report the percentage of firms in each ofthese categories The bottom part of each panel reports the percentage of firms issuing or

retiring securities by profitability within the smallest and largest asset quintiles The tablealso reports results from t-test for the difference in percentage of firms issuing or retiringsecurities in the lowest and the highest profit quintile

Percentage of firms

Iss Ret Iss Iss Rep Iss Both IssD IssE

D D NetD E E NetE DampE RepE RetD

(1) (2) (3) (4) (5) (6) (7) (8) (9)

Allfirm

s

Low 276 245 186 301 13 297 71 01 38

2 319 299 181 117 21 112 27 02 20

3 385 344 194 86 29 80 21 05 15

4 386 330 213 81 43 75 20 07 14

High 329 273 198 92 82 81 17 13 14

t Highfrac14Low 150 84 39 723 414 763 358 176 200

Smallfirm

s Low 275 204 201 411 11 405 109 01 47

2 291 233 213 379 10 378 99 01 45

3 281 246 204 265 14 261 66 01 41

4 275 272 186 162 15 159 34 02 35

High 275 281 178 136 27 131 25 02 26

t Highfrac14Low 01 84 28 316 55 318 171 18 55

Largefirm

s

Low 414 363 177 87 22 79 28 03 11

2 449 379 180 62 30 57 17 07 09

3 439 360 186 56 41 48 16 09 07

4 403 320 194 46 67 39 13 12 06

High 374 264 207 50 129 35 10 28 05

t Highfrac14Low 50 130 45 89 244 116 77 120 47

1430 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

is mostly independent of firm profitability In the aggregate there is a smalldifference but it is sensitive to sorting on firm size Similarly in Column (2)there is only a weak relation between the profitability and the likelihood thatthe firm retires debt To a first approximation the fraction of firms issuing orrepurchasing debt is largely independent of profits Large firms are somewhatmore likely than small firms to issue debt and also to retire debtUnlike debt issues the probability of issuing equity is strongly related to

profitability as can be seen in Column (4) Low-profit firms are much morelikely to issue equity than are high-profit firms In the lowest profit quintile301 of the firms issued equity In the highest profit quintile it was only92 The lower panel reports separate sorts for small and large firms Similarpatterns are found for both groups of firms that is the likelihood of equityissuance declines as profitability increases The differences between the quin-tiles are quite stark for the group of small firms All quintiles of small firmsissue equity much more frequently than the corresponding large firm quintilesHigh-profit firms are much more likely to repurchase equity as shown in

Column (5) In the overall sample only 13 of low-profit firmsrepurchased equity compared with 82 of the high-profit firms Whenwe separate the small and large firms again a difference emerges Ingeneral small firms do not repurchase much equity while a greaterfraction of the large firms do soGiven these facts it is natural that low profitability firms are much more

likely to be net issuers of equity than are high profitability firms as shown inColumn (6) Among the firms in the lowest quintiles of profits roughly 30of firms issue net equity that exceeds 5 percent of their assets In contrastfor the most profitable quintile only about 8 issue net equityFirm size is an important variable in the recent literaturemdashit is often used

as a proxy for access to capital markets (as in Faulkender and Petersen(2006) and Leary (2009)) Small firms are bank-dependent risky and infor-mationally opaque They have restricted access to public debt markets andconsequently face more severe supply constraints in their ability to issuedebt Thus we expect small firms to be more sluggish in adjusting theirdebt and equity in response to shocks to profitability Large firms incontrast have much easier access to public debt markets and they facefewer obstacles in accessing securities marketsTo examine how size and profitability interact the bottom part of Table IV

examines profitability sorts for the smallest and largest firms We first sortfirms annually by firm size and then within each size quintile we sort them onprofitability We do these two-way sorts to ensure that we have similarnumber of firms in profitability subgroups for both small and large firmsamples Unconditionally small firms are generally less profitable while

PROFITSndashLEVERAGE PUZZLEREVISITED 1431

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

large firms are relatively more profitable The two-way sorts allow us to havea uniform distribution of firms in various profitability quintiles for both smalland large firm samples The bottom part of Table IV reports results for thesmallest and the largest size quintiles Among the small firms there is littlerelation between profitability and the likelihood of issuing debt Howeverlarger firms exhibit a small increase in the likelihood of net issuance of debtwith increasing profitabilityIn contrast to debt the effects of profitability on equity issuance and

repurchases are much more consistent across size quintiles As profitabilityincreases firms are generally less likely to issue equity and more likely torepurchase it Across the two extreme size groups we note a striking differ-ence between the proportion of firms issuing equity and those issuing debtAmong low-profit small firms almost 41 are equity issuers whereas onlyabout 9 of low-profit large firms are equity issuers However regardless ofsize we note a monotonic reduction in the likelihood of issuing equity asprofitability increasesColumn (7) reports the fraction of firms that issue both debt and equity in

excess of 5 of assets both as a function of profits in the upper panel andas a function of firm size and profits in the lower panel In the upper panelwe see that low-profit firms are much more likely to issue both debt andequity (71) while the high-profit firms are less likely to do so (17) Thelower panel shows that small firms are much more likely to be issuers of bothin contrast to large firms In both firm size categories the profit pattern is thesame with low profits being more frequent issuers of both simultaneouslyImportantly Column (8) shows that the likelihood of issuing debt and

simultaneously repurchasing equity increases with profitability Converselyin Column (9) we find that the likelihood of doing the reverse that is issuingequity and retiring debt declines with profitability The effects of sorting onfirm size and firm profitability mirror those reported for all firms Overallthe results show that low-profit firms are less likely to issue debt and repur-chase equity they are instead more likely to issue equity and retire debtFirms with high profitability exhibit the reverse patternIn Appendix Table AII we present results from profitability (and size)

quintiles for large issuances defined using the 10 cutoff (in Panel A) andthe 15 cutoff (in Panel B) The results from these panels are largely con-sistent with those described above

52 MAGNITUDES OF ISSUANCES AND EQUITY CHANGES

In the previous section we considered the probability of having a nontriviallevel of debt or equity activity The next question is how large are the dollar

1432 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

values involved In Table V we sort firms according to profits and thentabulate the levels and changes in both debt and equity We do this firstfor all firms and then for small versus large firmsFor the analysis on all firms we observe in Column (1) that debt peaks at

the middle of the distribution This is because firms with medium profitabil-ity are also the largest as seen in Column (8) which reports average assetvalues for different profitability quintiles When we sort first by firm size andthen examine firms with different levels of profitability within size quintileswe find that for small firms debt is roughly independent of profits while forlarge firms less profitable firms have a higher level of debtAs expected Columns (3) and (5) show that more profitable firms have

higher equity values Columns (2) and (7) consider the relationship betweenissuances of debt and equity and firm profitability Debt issuances are sig-nificantly larger for more profitable firms In contrast equity issuances arelarger for low-profit firms Small low-profit firms issue substantially lessequity compared with small high-profit firms Among large firms the low-profit ones tend to issue equity while the high-profit ones tend to repurchaseequity The fact that more profitable firms issue debt and repurchase equitywhile the least profitable firms retire debt and issue equity is consistent withthe predicted relation between profitability and financing decisions under thetrade-off theoryWe also examine two-way sorts by size and profitability and report results

for the smallest and largest quintile of firms For the smallest quintile offirms the change in debt is largely unrelated to profits But for large firmsthere is a positive relation between profits and debt issuances High-profitfirms have a big positive change in debt Low-profit large firms have anegative change in debtColumns (4) and (6) provide an explanation of why the leverage ratio

regression results contradict those from the basic profitability sorts pre-sented here As we can see profitability indirectly affects leverage ratiosby increasing equity values Changes in both the book value of equity andthe market value of equity are positive and large for highly profitable firmsIn contrast these changes are negative for less profitable firms This evidenceillustrates an important issue concerning the use of leverage ratios Suchratios are often interpreted as essentially reflecting the use of debt by thefirm This interpretation while common is empirically misleadingFor the typical firm the change in the value of equity is larger than the

change in debt For example in the third profit quintile for large firmscolumn (7) the mean equity issue is just $9 million but the change in themarket value of equity is $242 million At the same time the mean change indebt is $27 million This suggests that a fair bit of the observed variation in

PROFITSndashLEVERAGE PUZZLEREVISITED 1433

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

the leverage ratios is primarily driven by the changes in the market value ofequity in the denominator rather than by the changes in debt in the numer-ator Since equity issues are often small this implies that the variation in theleverage ratio is primarily driven by internal operations rather than byexternal financing actions This again points to the fact that leverageratios can provide a misleading account of actual patterns in the dataThere is always a potential concern that the averages may be misleading

due to the impact of outliers To address this concern in AppendixTable AIII we present median values of the profitability sorts Theseresults generally reinforce the findings in Table VTable VI again considers the magnitudes of financing activity but this

time the issues are scaled by a measure of firm size That way it is easier to

Table V Magnitude of financing activity

The sample contains nonfinancial firms listed on the annual Compustat files for the period

from 1971 to 2009 The table presents average debt and equity changes in debt and equityand equity issuances (in millions of dollars) for firms sorted on profitability within sizeclasses The table reports information for all firms sorted on profitability and for profit-

ability sorts within the smallest and largest firms The sorts are done annually The tablealso reports results from t-test for the difference in percentage of firms issuing or retiringsecurities in the lowest and the highest profit quintile

D D BVE BVE MVE MVE EqyIss Assets

(1) (2) (3) (4) (5) (6) (7) (8)

Allfirm

s

Low 83 5 78 10 120 5 7 234

2 605 13 509 24 721 8 17 1691

3 1227 9 1123 22 1834 3 9 3551

4 833 23 1166 7 2408 74 9 2994

High 570 53 1171 71 3817 68 36 2581

t Highfrac14Low 31 9 36 9 38 2 16 36

Smallfirm

s Low 25 04 42 01 332 28 33 97

2 29 06 55 02 278 16 21 116

3 31 07 63 01 212 05 12 127

4 30 06 70 05 179 10 08 136

High 24 06 84 14 234 17 07 142

t Highfrac14Low 04 15 153 85 73 11 201 107

Largefirm

s Low 3929 55 2990 152 4671 11 63 10846

2 3826 19 3178 95 5122 66 11 10721

3 3074 27 3498 41 6173 242 9 9697

4 2623 62 3915 45 8473 94 57 9855

High 2180 193 4642 261 14924 202 148 10171

t Highfrac14Low 94 64 108 87 246 12 169 16

1434 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

see whether large firms behave disproportionately when compared withsmall firms Column (1) shows that low-profit firms have more debtas reported in many previous studies In Columns (4) and (6) we seethat low-profit firms have a major decline in both the book value ofequity and the market value of equity The high-profit firms have thereverse experienceMore interesting are the active decisions of the firms In Column (2) we see

that in contrast to the usual impression the low-profit firms are reducingdebt while the high-profit firms are increasing it In Column (7) we see thatthe low-profit firms are also reducing leverage by issuing equity The high-profit firms issue little equity In the lower panel similar results are foundwhen we separate small and large firms Equity issuing is particularly strong

Table VI Magnitude of financing activitymdashscaled levels and changes

The sample contains nonfinancial firms listed on the annual Compustat files for the period

from 1971 to 2009 The table presents average debt and equity changes in debt and equityand equity issuances scaled by assets for firms sorted on profitability within size classes Thetable reports information for all firms sorted on profitability and for profitability sorts

within the smallest and largest firms The sorts are done annually The table also reportsresults from t-test for the difference in percentage of firms issuing or retiring securities inthe lowest and the highest profit quintile

D D BVE BVE MVE MVE EqyIss

(1) (2) (3) (4) (5) (6) (7)

Allfirm

s

Low 386 122 12 871 3427 1868 217

2 285 58 407 119 940 157 28

3 293 12 425 14 742 03 12

4 269 26 461 02 955 27 10

High 192 15 551 47 1675 74 07

t Highfrac14Low 35 49 50 47 127 42 157

Smallfirm

s Low 545 154 841 2145 9688 9128 696

2 684 49 712 1915 5014 2318 288

3 331 253 141 969 3063 954 129

4 267 181 357 175 1673 221 55

High 194 05 543 26 1758 147 41

t Highfrac14Low 49 17 51 36 88 27 99

Largefirm

s Low 350 30 306 58 511 55 09

2 346 14 335 13 523 02 04

3 353 165 332 71 711 29 01

4 280 01 414 04 909 09 05

High 227 15 477 19 1643 27 14

t Highfrac14Low 371 112 453 181 570 57 221

PROFITSndashLEVERAGE PUZZLEREVISITED 1435

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

among the small low-profit firms that are also experiencing sharp reductionsin the value of equity In column (2) we see that both for the large and for thesmall firms net debt increases are primarily found in the highest quintile offirm profitability

6 Debt and Equity Regressions

So far we have demonstrated that in our data the conventional leverageregressions have the usual signs We then demonstrated that in simple sortsof the data strikingly different active financing patterns emerge The nexttask is to reconcile these differencesTable VII presents simple regressions predicting changes in debt book

equity and market equity as well as equity issuances Consistent with thesorts we control for changes in firm size and we also include firm and yeardummies There may be a concern about the speed of a firmrsquos responsesAccordingly we consider the lagged change in profits as the main case butwe also include the contemporaneous change in profits Whether the contem-poraneous change in profits is included or not does not in fact matter Wereport panel-robust standard errors adjusted for clustering at the firm levelThe results in Table VII support the findings in the sorts Firms that have

an increase in profits in 1 year increase debt both in the same year and in thenext year Firms that have an increase in profits experience an increase inboth the book value of equity and the market value of equity Importantlyhowever firms that experience and increase in profits reduce their equityissues All these effects are statistically significant (except for the marketvalue of equity in the column 5 specification) and seem intuitively reason-able In Appendix Table AIV similar regressions to those presented aboveare reported but without firm-fixed effects We find that empirically verylittle changes whether we include firm fixed effects or leave them out Thusthe targeting behavior to the extent that it happens does not account forwhat we are reportingThe next step is to bring the sorts and the simple regressions together into

a common setting We therefore include factors in addition to profitability(i) median industry leverage (ii) the market-to-book assets ratio (iii) tangi-bility of assets and (iv) firm size (measured by the log of assets) Rajan andZingales (1995) show that these factors are related to leverage in G7countries A number of studies have used these factors to estimateleverage targets Frank and Goyal (2009) show that these factors arerobustly related to leverage in the USA In these regressions we use indica-tors for the quintile that the firm is in for each factor

1436 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

Table VIII reports these results In Columns (1) and (2) we examine debtissues and again show that profits positively affect debt issuances The effectsare large and statistically significant at the 1 level Columns (3) and (4)which examine book equity show a strong effect of profits This effect is onlyslightly reduced by the inclusion of other factors Columns (5) and (6) illus-trate that the impact of profits on the change in market equity is also robustto the inclusion of the conventional factors Columns (7) and (8) examineequity issuances Here we again find that profits have a significant negativeimpact on equity issuances As before the results are robust to the inclusionof the conventional factorsIn unreported tables we estimate debt and equity changes using quantile

regressions In terms of the signs and significance of the profitability variable

Table VII Debt and equity changes

The sample contains nonfinancial firms listed on the annual Compustat files for the period

from 1971 to 2009 The table presents fixed effect estimates from regressions of changes indebt (D) changes in the book value of equity (BVE) changes in the market value ofequity (MVE) and net equity issuances (EqyIss) The explanatory variables include

current and lagged change in operating income before depreciation (Profits andProfitst1) and lagged change in the book value of assets (Assetst1) In addition tofixed firm effects the regressions include year indicator variables The reported t-statistics

are corrected for clustering at the firm level

aMean significant at the 1 level

bMean significant at the 5 level

cMean significant at the 10 level

D BVE MVE EqyIss

(1) (2) (3) (4) (5) (6) (7) (8)

Profitst1 022a 025a 015b 035a 004 060b 005b 007a

(33) (31) (25) (35) (03) (25) (25) (53)

Profits 050b 078a 140a 003a

(24) (32) (48) (34)

Assetst1 000 007b 025a 001b

(00) (21) (35) (25)

Constant 167 283a 34 51 1145a 1182a 605a 595a

(14) (30) (04) (05) (32) (29) (69) (74)

Year FE Yes Yes Yes Yes Yes Yes Yes Yes

Firm FE Yes Yes Yes Yes Yes Yes Yes Yes

R2 001 004 001 010 001 005 002 003

N 162056 162056 162130 162130 157550 157550 162154 162154

PROFITSndashLEVERAGE PUZZLEREVISITED 1437

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

the results are qualitatively similar to the OLS results reported in Table VIIIProfitability positively affects debt issuances and negatively affects equityissuances across various quantiles Overall the results in Table VIII showthat there is nothing that is special about using sorts or regressions to explaindebt and equity issues In either case we find that more profitable firms tendto increase their debt experience an increase in the value of equity andrepurchase shares Thus the control factors are not responsible for theusual rejection of the trade-off theory It would appear that what mattersis the size of the issuing activity relative to the organic increases in equityvalue due to profits In other words what matters in the leverage regressionseems to be coming from the presence of E in D

DthornE and not from D

Table VIII Debt and equity issuances profitability and leverage factor quintiles

The sample contains nonfinancial firms listed on the annual Compustat files for the period

from 1971 to 2009 The table presents estimates from regressions of change in debt (D)change in the book value of equity (BVE) change in the market value of equity (MVE)and net equity issuances (EquityIss) The explanatory variables include lagged profitability

quintiles and lagged leverage factor quintiles In addition to fixed firm effects the regres-sions include year indicator variables The reported t-statistics are corrected for clustering atthe firm level

aMean significant at the 1 level

bMean significant at the 5 level

cMean significant at the 10 level

D BVE MVE EqyIss

(1) (2) (3) (4) (5) (6) (7) (8)

ProfitabilityQint 214a 167a 283a 175a 80 429a 53a 55a

(70) (54) (66) (59) (08) (48) (55) (62)

IndLevQint 104a 43 99 20

(29) (15) (10) (10)

Ln(Assets)Qint52 17 1026a 53a

(16) (05) (77) (35)MB

Qint158a 372a 1155a 09

(45) (56) (106) (06)

TangibilityQint 05 177 436c 64a

(01) (15) (18) (31)

Constant 167 96 656b 235 4613a 8610a 113 570a

(08) (04) (26) (04) (61) (58) (11) (31)

Year FE Yes Yes Yes Yes Yes Yes Yes Yes

Firm FE Yes Yes Yes Yes Yes Yes Yes Yes

R2 0002 0002 0003 0004 0013 0014 0013 0014

N 155587 155037 155587 155037 155587 155037 155587 155037

1438 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

7 Are Financial Market Conditions Important

According to the trade-off theory capital structure is determined by con-sidering the impact of costs and benefits of debt The time variation inthe costs and benefits of debt imply that issuance decisions would alsovary over time Furthermore market conditions affect the transactioncosts of adjusting Adverse selection is a greater problem in a cold marketthan in a hot one Accordingly it may be easier to adjust in good marketconditions Related ideas have been developed by Huang and Ritter(2009) Faulkender et al (2012) and Halling Yu and Zechner (2012)Faulkender Flannery Hankins and Smithrsquos paper examines cross-sectionalvariation in adjustment speeds including market valuations andhow they may affect adjustment speeds It is natural to think that ingood market conditions it will be less costly to issue both debt and equityand adjusting leverage to respond to profitability shocks would be mucheasierTo test the importance of market conditions we require a definition of

good times and bad times Our empirical strategy is to estimate good timesversus bad times at the four-digit industry level We define an industry ashaving ldquogood timesrdquo if the median firm in that industry has a market-to-book ratio that is higher than the 67th percentile of the time-series distribu-tion of the industry median market-to-book ratios Conversely an industryis defined as having ldquobad timesrdquo if the median firm in the industry has amarket-to-book ratio that is lower than the 33rd percentile of the time-seriesdistribution of the industry median market-to-book ratiosPanel A of Table IX tabulates issuance activity for profitability sorts in

both good times and bad times As might be expected active debt and equityissuances are larger during good times There is also somewhat more activeswaps between debt and equity during good timesPanel B reports the issuance activity for the smallest and the largest firms

during the good times and bad times Within each size quintile firms aresorted on profitability This is a three-way sort As before we find that thereis much more active use of external markets during good times Small low-profit firms are more likely to issue equity in good times than in bad timesDebt issuances are significantly higher in good times Large high-profitfirms are significantly more likely to issue debt and repurchase equity ingood times than in bad timesPanel C reports the magnitudes of the financial variables rather than the

frequencies During bad times less profitable large firms retire substantialamounts of debt and they show a tendency to issue equity Small firms donot seem to engage in similar debt reduction activities However like the

PROFITSndashLEVERAGE PUZZLEREVISITED 1439

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

Table IX Debt and equity issues in good and bad times

The table reports the frequency and magnitude of financing activity for sorts on profits for subsamples

of firms in good and bad times An industry is defined as having ldquogood timesrdquo if the median firm in

that industry has a market-to-book ratio that is higher than the 67th percentile of the time-series

distribution of industry median market-to-book ratios Conversely an industry is defined as having

bad times if the median firm in that industry has a market-to-book ratio that is lower than the 33rd

percentile of the time-series distribution of industry median market-to-book ratios Panels A and B

report the percentage of firms (i) issuing net debt in excess of 5 of the value of their assets (ii) issuing

net equity in excess of 5 of the value of their assets (iii) issuing debt and repurchasing equity both in

excess of 5 of the value of their assets and (iv) issuing equity and retiring debt both in excess of 5

of the value of their assets Panel C tabulates the mean debt and equity levels and the mean changes in

debt and equity issuances The sample contains nonfinancial firms listed on the annual Compustat files

for the period from 1971 to 2009

Iss Iss Iss D Iss ENet D Net E Rep E Ret D(1) (2) (3) (4)

Panel A sorts on profitability

Badtimes

Low 145 131 01 202 170 49 02 083 173 46 04 074 182 47 03 10

High 177 51 06 10t Highfrac14Low 41 136 33 01

Goodtimes

Low 206 381 01 462 197 148 02 233 212 103 04 184 234 92 10 18

High 216 92 16 14t Highfrac14Low 23 658 138 23

Panel B sorts on firm size and profitability

Badtimes

Smallfirm

s Low 151 318 00 312 182 276 02 253 151 140 02 294 183 73 00 19

High 177 80 00 11t Highfrac14Low 10 90 NA 20

Largefirm

s Low 172 54 01 042 148 39 04 053 150 43 04 044 156 33 03 06

High 147 28 05 03t Highfrac14Low 17 32 20 02

Goodtimes Smallfirm

s Low 213 448 01 542 230 443 01 473 222 325 00 474 203 197 02 38

High 190 153 02 31t Highfrac14Low 19 230 10 39

Largefirm

s Low 203 96 03 122 208 66 09 133 232 54 16 064 215 39 19 05

High 234 36 38 04t Highfrac14Low 29 99 92 38

(continued)

1440 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

larger firms there is a tendency for small firms to issue more equitymdashat leastto the extent possibleDuring good times large profitable firms raise significant amounts of debt

as they experience an increase in the value of their equity Such firms alsoengage in share repurchases Debt issues by smaller firms are much lessaffected by their own profitability during good times and their change indebt is more or less independent of their profits However equity issuanceis primarily found among the low-profit small firms during good timesTable IX shows that empirically profitability has a time-varying impacton leverage choicesIn Table X the basic regressions for changes in debt and for equity

issuance are presented for both good times and bad times Comparinggood times with bad times we see that the effects of profits are muchstronger in good times In bad times the impacts are rather weak statisticallyThis difference is interesting and deserves further study However for ourpurposes recall that the static trade-off theory implies time-varying capital

Table IX Continued

D D BVE BVE MVE MVE EqyIss Assets

(1) (2) (3) (4) (5) (6) (7) (8)

Panel C magnitude of financing activity

Badtimes Smallfirm

s Low 40 03 58 14 305 06 22 140

2 44 05 82 07 240 67 17 180

3 45 04 94 02 176 18 05 187

4 43 09 98 02 137 04 03 188

High 36 10 114 13 209 16 04 199

Largefirm

s Low 3365 73 2523 169 2304 163 94 8972

2 3031 95 2983 35 3284 278 56 9068

3 2930 5 3360 14 3947 308 57 9225

4 2281 6 3817 19 5213 442 7 8902

High 1790 62 4767 13 8799 58 31 9573

Goodtimes

Smallfirm

s Low 24 06 38 05 339 23 39 90

2 26 08 48 01 301 04 25 104

3 30 09 60 04 246 03 17 125

4 28 07 66 08 206 18 12 130

High 23 06 85 16 261 13 08 141

Largefirm

s Low 4532 36 3477 103 6557 260 53 12465

2 4580 58 3527 139 6911 421 23 12522

3 2952 111 3485 2 7586 279 32 9535

4 2752 124 3939 72 10686 33 105 10226

High 2429 228 4819 377 18565 268 222 10830

PROFITSndashLEVERAGE PUZZLEREVISITED 1441

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

structure choices even when the target is time invariant Empirically we doobserve time-varying choices

8 Is Scaling Important

In the preceding section we provided results for unscaled debt and equityissues This seems appropriate to us However it involves making twochanges to the standard regression We would like to determine if both

Table X Debt and equity changes in good and bad timesmdashregressions

The sample contains nonfinancial firms listed on the annual Compustat files for the period

from 1971 to 2009 The table presents estimates from regressions of change in debt (D)and net equity issuances (EqyIss) in good and bad times The explanatory variables includechange in profits lagged change in profits and lagged change in assets The regressions

include both firm and year fixed effects An industry is defined as having ldquogood timesrdquo ifthe median firm in that industry has a market-to-book ratio that is higher than the 67thpercentile of the time-series distribution of industry median market-to-book ratios

Conversely an industry is defined as having bad times if the median firm in thatindustry has a market-to-book ratio that is lower than the 33rd percentile of the time-series distribution of industry median market-to-book ratios The reported t-statistics arecorrected for clustering at the firm level

aMean significant at the 1 level

bMean significant at the 5 level

cMean significant at the 10 level

Good times Bad times

D EqyIss D EqyIss

(1) (2) (3) (4) (5) (6) (7) (8)

Profits 087a 091a 006b 009a 060 065 003 003b

(30) (29) (20) (28) (15) (16) (16) (20)

Profitst1 029c 015a 034c 001

(19) (37) (17) (04)

Assetst1 000 002b 000 001

(02) (22) (00) (07)

Constant 1337 3989 8691a 10001a 1351 1851 4099a 4055a

(03) (09) (55) (37) (09) (13) (80) (78)

Year FE Yes Yes Yes Yes Yes Yes Yes Yes

Firm FE Yes Yes Yes Yes Yes Yes Yes Yes

R2 0071 0080 0018 0045 0178 0212 0034 0041

N 74827 68729 74891 68787 26965 26013 26971 26018

1442 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

changes are crucial Accordingly in this section we take a step back toconsider prior literature in which variables are scaled The questions arehow does scaling affect the results and whyTable XI examines alternative scalings Columns (1) and (2) scale debt and

equity issuances by the total firm net issuance Columns (3) and (4) scaledebt and equity issuances by the sum of debt and book equity Columns (5)and (6) consider the change in debt scaled by the sum of debt and bookequity Columns (7) and (8) consider the change in leverage due to externalfinancing Since we are examining changes we use the changes in thestandard regression factors as controls The primary interest is in the coef-ficient on the change in profitsWe find that a change in profits is associated with an increase in debt

issues But the sign reverses when we scale the debt issues by the sum of debtplus book equity This is telling What is really driving the results is thechange in equity in the denominator not the change in debt in the numer-ator This result is further substantiated in Columns (5) and (6) where thechange in debt to capital ratio is used instead of debt issuanceTo further explore the impact of active issuance relative to passive

experiencing of changes we decompose leverage by separating out theeffect of debt and equity issues from those due to changes in retainedearnings Thus we define leverage change due to external financing as

LevEF frac14Dt1 thorn d

Dt1 thorn Et1 thorn dthorn e

Dt1

Dt1 thorn Et1 eth1THORN

where D is debt E is book value of equity d is debt issues net of retirementsand e is equity issues net of repurchases Garvey and Hanka (1999) use asimilar measure in their study This measure directly captures the netleverage effect of debt and equity issuancesColumns (7) and (8) report results from regressions of the net leverage

effect of debt and equity issues on changes in profitability after controllingfor other leverage factors in differences and year indicator variables Theeffect of profitability on leverage change due to external financing is positiveand statistically significant This comparison suggests that changes in totaldebt and equity are only partly a result of debt and equity issuance decisionsOther balance sheet adjustments complicate the inferences from leverageratio regressionsFor most of our analysis we follow the literature in treating debt as homo-

genous It is of course potentially interesting to ask what happens when wedistinguish between short-term debt and long-term debt Which part of thefirmrsquos debt maturity structure drives the under-adjustment Table XII

PROFITSndashLEVERAGE PUZZLEREVISITED 1443

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

provides results from such decompositions Alternative scalings areconsidered The main results of interest are that the distinction betweenshort- and long-term debt is important Short-term debt issuance is largelyindependent of changes in profits The negative sign on profit change stemsfrom the action of long-term debt This seems intuitively reasonable

Table XI Scaled issuances

The sample contains nonfinancial firms listed on the annual Compustat files for the period

from 1971 to 2009 The table presents estimates from regressions of debt issuance to totalissuance (in columns (1) and (2)) debt issuance to total capitalization (in columns (3) and(4)) change in debt to capitalization ratio (in columns (5) and (6)) and change in leverage

ratio considering only external financing (in columns (7) and (8)) The change in leverageratio considering only external financing is defined as

LevEF frac14Dt1 thorn d

Dt1 thorn Et1 thorn dthorn e

Dt1

Dt1 thorn Et1

where D is debt E is book value of equity d is debt issues net of retirements and e is equity

issues net of repurchases The regressions include lagged leverage factors and year indicatorvariables Some of the specifications include firm fixed effects The reported t-statistics arecorrected for clustering at the firm level

aMean significant at the 1 level

bMean significant at the 5 level

cMean significant at the 10 level

DebtIssTotIss

DebtIssethDthornETHORN D

DthornE LevEF

(1) (2) (3) (4) (5) (6) (7) (8)

Profitt1 004a 000 002a 001c 003a 002 001b 002a

(35) (03) (34) (18) (35) (14) (23) (26)

IndLevt1 004c 001 001 001c 000 000 000 000

(19) (06) (15) (19) (02) (04) (03) (05)

Market=bookt1 001a 001a 000a 000a 000a 000a 000 000c

(47) (85) (68) (60) (55) (36) (05) (18)

Tangt1 004 005 013a 011a 011a 006a 003a 003a

(12) (15) (144) (112) (79) (38) (44) (42)

Assetst1 008a 004a 006a 004a 002a 001a 003a 002a

(97) (45) (283) (161) (53) (35) (147) (101)

Constant 079a 074a 002a 003a 000 000 000b 002a

(497) (442) (91) (111) (00) (05) (22) (93)

Year FE Yes Yes Yes Yes Yes Yes Yes Yes

Firm FE No Yes No Yes No Yes No Yes

R2 0015 0098 0022 0076 0009 0070 0012 0086

N 133905 133905 140552 140552 138229 138229 140572 140572

1444 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

inasmuch as short-term debt is probably more associated with day-to-dayoperations rather than longer term planning considerationsIn Table XIII we examine change in leverage ( D

DthornE) and change inleverage due to external financing (LEVEF) during good and bad timesseparately The traditional negative effect of a change in profits on achange in leverage is found both in good times and in bad timesHowever when we decompose the impact in order to isolate the externalcomponent as before the result disappears In good times the externalfinancing component of a change in profits is weak but still positiveDuring bad times it is not statistically significant This implies that distin-guishing between good and bad times does not fundamentally alter the basic

Table XII Short- and long-term debt issuances

The sample contains nonfinancial firms listed on the annual Compustat files for the period

from 1971 to 2009 The table presents estimates from regressions of short-term debtissuance to total issuance (in columns (1) and (2)) long-term debt issuance to totalissuance (in columns (3) and (4)) short-term debt issuance to total capitalization (in

columns (5) and (6)) and long-term debt to total capitalization (in columns (7) and (8))Some of the specifications include firm fixed effects The reported t-statistics are correctedfor clustering at the firm level

aMean significant at the 1 level

bMean significant at the 5 level

cMean significant at the 10 level

STDebtIssTotIss

LTDebtIssTotIss

STDebtIssethDthornETHORN

LTDebtIssethDthornETHORN

(1) (2) (3) (4) (5) (6) (7) (8)

Profitabilityt1 0002 0003 0042a 0008 0001 0001 0019a 0012a

(02) (04) (30) (05) (14) (10) (45) (28)

IndLevt1 0009 0007 0070b 0033 0001 0002c 0014a 0005

(06) (05) (26) (11) (08) (17) (26) (09)

M=Bt1 0003a 0003a 0003b 0010a 0000 0000 0003a 0002a

(48) (35) (22) (65) (10) (01) (84) (42)

Tangibilityt1 0020 0007 0011 0067 0005a 0006a 0097a 0081a

(12) (03) (03) (16) (34) (33) (108) (89)

Assetst1 0039a 0011b 0034a 0024b 0001a 0001 0046a 0020a

(89) (22) (38) (23) (36) (13) (190) (82)

Constant 0001b 0153a 0798a 0591a 0000c 0006a 0057a 0096a

(25) (144) (470) (288) (17) (124) (292) (234)

Year FE Yes Yes Yes Yes Yes Yes Yes Yes

Firm FE No Yes No Yes No Yes No Yes

R2 0025 0089 0018 0073 0030 0147 0016 0298

N 136146 136146 136140 136140 143031 143031 142847 142847

PROFITSndashLEVERAGE PUZZLEREVISITED 1445

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

Table XIII Externally financed leverage changes in good and bad times

The sample contains nonfinancial firms listed on the annual Compustat files for the period

from 1971 to 2009 The table presents estimates from regressions of change in debt tocapitalization ratio (in columns (1) (2) (5) and (6)) and change in leverage ratio consider-ing only external financing (in columns (3) (4) (7) and (8)) The change in leverage ratio

considering only external financing is defined as

LevEF frac14Dt1 thorn d

Dt1 thorn Et1 thorn dthorn e

Dt1

Dt1 thorn Et1

where D is debt E is book value of equity d is debt issues net of retirements and e is equityissues net of repurchases The regressions include lagged leverage factors and year indicator

variables The reported t-statistics are corrected for clustering at the firm level An industryis defined as having ldquogood timesrdquo if the median firm in that industry has a market-to-bookratio that is higher than the 67th percentile of the time-series distribution of industry

median market-to-book ratios Conversely an industry is defined as having ldquobad timesrdquoif the median firm in that industry has a market-to-book ratio that is lower than the 33rdpercentile of the time-series distribution of industry median market-to-book ratios

aMean significant at the 1 level

bMean significant at the 5 level

cMean significant at the 10 level

Good times Bad times

DDthornE LevEF D

DthornE LevEF

(1) (2) (3) (4) (5) (6) (7) (8)

Profitabiltyt1 0022c 0005 0016c 0025b 0073b 0041 0007 0007

(17) (03) (18) (24) (25) (11) (04) (03)

IndLevt1 0000 0011 0001 0009 0010 0018 0001 0000

(00) (09) (01) (13) (07) (10) (02) (00)

MB

t1

0004a 0003b 0000 0001 0008a 0001 0000 0000

(40) (24) (05) (08) (28) (04) (03) (03)

Tangt1 0097a 0041 0033a 0038b 0127a 0048 0059a 0033

(45) (16) (27) (25) (49) (14) (43) (16)

Assetst1 0014a 0013b 0024a 0019a 0037a 0023c 0036a 0025a

(30) (20) (84) (48) (45) (17) (88) (43)

Constant 0030b 0036 0000 0003 0028c 0014 0021 0001

(20) (16) (00) (07) (17) (06) (11) (01)

Year FE Yes Yes Yes Yes Yes Yes Yes Yes

Firm FE No Yes No Yes No Yes No Yes

R2 0006 0088 0009 0117 0026 0251 0023 0202

N 56820 56820 58200 58200 23462 23462 23511 23511

1446 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

conclusions The traditional leverage result on profits is driven by increasesin equity that are experienced and then partially but not completely offsetby firmsTo the extent that firms are engaging in active financing they are following

the prescriptions of the trade-off theory Thus we have a coherent picture ofhow firms are responding to profits Basically more profitable firms experiencean increase in the value of their equity They respond to this by issuing debtand repurchasing equity But the magnitude of the increase in equity is largerthan the magnitudes of the active steps taken in the debt and equity marketsThis means that the firmsrsquo actual actions are as predicted by the trade-offtheory But it leaves open the question of why they do not act more stronglyThe capital structure literature has long been interested in dynamics and

capital structure adjustment paths In recent years this interest has increas-ingly resulted in explicit dynamic optimizing models such as those ofGoldstein Ju and Leland (2001) Hennessy and Whited (2005) andStrebulaev (2007) These models commonly make specialized assumptionsleading some to question their generalityWe argue that both fixed transactions costs and variable transactions costs

are plausible and realistic in models that pin down firm size The simultaneouspresence of both is important Due to fixed costs firms only rebalance inter-mittently Due to variable adjustment costs the extent of the adjustment isreduced to the point at which an extra unit to adjustment is balanced againstthe extra unit transactions costs This means that both leverage increases andleverage reductions will only go part way toward a static optimumThe key point is that with both fixed and variable transactions costs we

should expect to only see intermittent leverage adjustments When leverage isadjusted it should ldquounder adjustrdquo when compared to a frictionless modelFurther discussion of transactions costs in a capital structure setting isprovided by Leary and Roberts (2005)

9 Conclusion

The connection between corporate profits and capital structure has beenvery influential in the assessment of the trade-off theory The conventionalinterpretation of the evidence has pushed the literature away from the trade-off theory and toward much more complex dynamic models and ideas As aresult it is important to make sure that the evidence has been correctlyinterpreted We show that the problem is with the interpretation of thetest results and a failure to take into account the costs associated with re-sponding to a profit shock

PROFITSndashLEVERAGE PUZZLEREVISITED 1447

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

We document that more profitable firms really do borrow more (not less)They tend to repurchase equity They experience an increase in both thebook value of equity and the market value of equity Less profitable firmsreally do tend to reduce their debt and to issue equityFirm size and market conditions also matter Larger firms tend to be more

active in the debt markets while smaller firms tend to be relatively moreactive in the equity markets During good times there is more use ofexternal financingThe usual profitsndashleverage puzzle result is primarily driven by the increase

in equity that is experienced by the more profitable firms Accordingly thepuzzle should be restated as asking why do firms not take sufficiently largeoffsetting actions to fully undo the change in equity What limits the mag-nitudes of the typical leverage response to profit shocksIn a frictionless model the partial response appears puzzling But there is

good empirical reason to believe that rebalancing entails both fixed andvariable costs and that firm size matters The rebalancing costs can befully avoided by doing nothing Accordingly the firm must decide whetherany given shock is big enough to be worth responding to If it is then thefirm must decide how big a response is called for These technical conditionsare known as ldquovalue matchingrdquo and ldquosmooth pastingrdquo Thus optimizationimplies that some shocks will be ignored Even if the shock is not ignoredthe optimal response will only partially undo the shock The magnitude ofthe leverage response must balance the marginal cost and the marginalbenefit of an extra unit of leverage Since the marginal cost of adjustingleverage is strictly positive the adjustment toward that static leverageoptimum will only go part way This is true both for leverage increasesand for leverage reductionsOne might conjecture that the variable costs would be time varying and

that such costs might have declined over time Since we do not have a fullaccounting of all of the variable costs it is of course impossible to be sureHowever there is some evidence suggesting that at least the directunderwriting fees are remarkably stable over time10

Surely nobody ought to be surprised that more profitable firms are moreinclined to repurchase equity while unprofitable firms tend not to do so Thefact that more profitable firms typically experience an increase in the value of

10 For IPOs Jay Ritter provides important summary evidence as httpbearwarringtonufleduritterIPOs2013Underwritingpdf He finds that for medium-sized deals a remark-

ably large fraction of the deals have a 7 fee year after year Whether this is due tobehavioral factors or some sort of a coordination device is not clear The extent towhich similar intertemporal stability carries over to other forms of external financemerits further study

1448 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

equity is equally natural Were it not for the contrary claims in the literature(ie the profitsndashleverage puzzle) we would argue that nobody ought to findit surprising that more profitable firms are more inclined to issue debt Thefact that leverage adjustments typically only partially offset profit shocksshould also sound quite natural These facts all fit together easily within thetrade-off theory once we take account of both fixed and variable costs ofactively adjusting capital structure

Appendix

Table AI Market leverage regressions

The table presents estimates of the regressions of market leverage on firm characteristics

The sample comes from the annual Compustat files during the period 1971ndash2009 Financialfirms are excluded The leverage is estimated as the ratio of debt over debt plus marketequity The explanatory variables Profitabilityt1

MB

t1

Tangibilityt1 and LnethAssetsTHORNt1are described in Table 1 IndMedianLevt1 is estimated as the median market leverage of allother firms in the same industry The industry is defined at the level of the four-digit SICcode All specifications include the year fixed effects The specifications in column (2)

additionally include the firm fixed effects We report t-statistics where the standard errorsare clustered at the firm level in parentheses

aSignificant at the 1 level

Quantile regressions

(1) (2) (3) (4) (5)

Profitabilityt1 013a 013a 007a 011a 020a

(275) (262) (349) (425) (538)

IndLevt1 046a 020a 032a 058a 063a

(506) (239) (1299) (1630) (1042)MB

t1

003a 001a 001a 002a 004a

(509) (303) (743) (709) (534)

Tangibilityt1 013a 018a 013a 014a 010a

(191) (181) (762) (571) (232)

LnethAssetsTHORNt1 001a 005a 001a 001a lt001a

(112) (258) (682) (342) (29)

Constant 010a 011a 005a 005a 029a

(171) (98) (173) (118) (398)

Year FE Yes Yes Yes Yes Yes

Firm FE No Yes No No No

Regression type OLS FE 25th tile Median 75thtile

Clustered SE Yes Yes No No No

R2 030 067

Pseudo R2 016 023 0201

N 158578 158578 158578 158578 158578

PROFITSndashLEVERAGE PUZZLEREVISITED 1449

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

Table AII Profitability sorts for firms issuing large amounts

Percentage of firms

Iss Ret Iss Iss Rep Iss Iss IssD IssE

D D NetD E E NetE DampE RepE RetD

(1) (2) (3) (4) (5) (6) (7) (8) (9)

Panel A 10 cutoff to define issuing or retiring activity

Allfirm

s

Low 205 150 134 260 05 257 45 00 23

2 207 175 103 77 06 75 11 01 10

3 243 196 101 50 08 47 08 01 09

4 252 186 113 50 13 48 08 02 08

High 227 161 118 60 27 57 07 04 09

t Highfrac14Low 70 38 62 755 226 768 320 90 152

Smallfirm

s Low 230 132 162 380 05 375 82 00 32

2 228 143 165 334 04 333 65 00 26

3 211 146 147 216 03 214 38 00 23

4 190 154 120 126 04 124 15 01 20

High 194 162 114 99 09 97 12 01 15

t Highfrac14Low 42 40 68 342 20 340 169 07 131

Largefirm

s Low 231 200 82 31 06 29 08 01 04

2 231 187 67 17 06 14 03 01 02

3 234 177 77 17 09 14 03 02 03

4 218 158 75 16 17 14 02 03 02

High 218 135 99 19 41 15 03 07 02

t Highfrac14Low 19 106 36 46 135 57 39 58 23

Percentage of firms

Iss Ret Iss Iss Rep Iss Both Iss D Iss E

D D Net D E E Net E DampE Rep E Ret D

(1) (2) (3) (4) (5) (6) (7) (8) (9)

Panel B 15 cutoff to define issuing or retiring activity

Allfirm

s

Low 158 101 100 230 01 228 27 00 15

2 142 115 63 57 01 56 05 00 06

3 165 128 57 32 01 31 03 00 05

4 173 118 65 35 01 34 03 00 05

High 160 102 70 46 04 44 03 00 05

t Highfrac14Low 08 06 141 738 87 746 277 24 131

Smallfirm

s Low 195 95 132 354 01 350 56 00 24

2 185 96 131 303 01 301 44 00 18

3 160 96 104 179 00 176 18 00 15

4 134 99 78 99 01 98 08 00 12

High 133 95 73 82 01 80 06 00 10

t Highfrac14Low 80 00 95 345 12 344 149 12 52

Largefirm

s Low 145 123 42 14 01 12 03 00 01

2 137 114 32 07 01 06 01 00 01

3 142 104 35 06 01 05 00 00 01

4 130 91 36 08 03 07 01 01 01

High 134 81 48 09 06 07 01 01 01

t Highfrac14Low 19 85 17 26 54 33 34 17 11

1450 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

Table AIV Debt and equity changesmdashwithout fixed effects

The sample contains nonfinancial firms listed on the annual Compustat files for the period from 1971

to 2009 The table presents estimates from regressions of change in debt (D) change in the book

value of equity (BVE) change in the market value of equity (MVE) and net equity issuances

(EqyIss) The explanatory variables include current and lagged change in operating income before

depreciation (Profits and Profitst1) and lagged change in book value of assets (Assetst1) The

regressions include year indicator variables The reported t-statistics are corrected for clustering at the

firm level

aMean significant at the 1 level

bMean significant at the 5 level

cMean significant at the 10 level

D BVE MVE EqyIss

(1) (2) (3) (4) (5) (6) (7) (8)

Profitst1 025a 024a 023a 038a 007 066a 007a 009a

(40) (32) (36) (39) (06) (29) (42) (51)Profits 051b 082a 149a 004a

(25) (33) (49) (33)Assetst1 002 005c 024a 001b

(08) (17) (34) (23)Constant 93 176a 45 71 1552a 1704a 130a 140a

(12) (28) (07) (09) (49) (49) (45) (45)Year FE Yes Yes Yes Yes Yes Yes Yes YesR2 001 005 001 011 001 005 002 002N 162056 162056 162130 162130 157550 157550 162154 162154

Table AIII Median profitability sorts

The sample contains nonfinancial firms listed on the annual Compustat files for the period from 1971

to 2009 The table presents median debt and equity median changes in debt and equity and median

equity issuances (EqyIss) in millions of dollars for all firms sorted on profitability and for profitability

sorts within the smallest and largest firms The sorts are done annually

D D BVE BVE MVE MVE EqyIss Assets(1) (2) (3) (4) (5) (6) (7) (8)

Allfirm

s

Low 21 00 75 17 256 20 01 1922 186 02 449 18 598 19 00 11033 816 06 1274 03 1637 05 00 32244 787 02 1541 18 2488 03 00 3556High 273 00 1361 66 3300 14 00 2578

Smallfirm

s Low 05 00 13 07 140 14 06 402 06 00 25 08 114 12 01 563 09 00 32 05 85 07 00 674 12 00 39 01 68 03 00 79High 10 00 51 04 98 01 00 90

Largefirm

s Low 11485 381 10269 371 13227 202 03 333902 13051 393 12786 140 17571 169 05 392463 11554 287 13429 24 21623 305 00 370314 9963 209 15052 194 29424 308 00 36705High 7361 00 16380 613 49142 439 00 34976

PROFITSndashLEVERAGE PUZZLEREVISITED 1451

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

References

Auerbach A J (1985) Real determinants of corporate leverage in B M Friedman (ed)

Corporate Capital Structures in the United States University of Chicago Press Chicago

IllinoisBertrand M and Mullainathan S (2003) Enjoying the quiet life corporate governance and

managerial preferences Journal of Political Economy 111 1043ndash1075Booth L Varouj A Demirguc-Kunt A and Maksimovic V (2001) Capital structures in

developing countries Journal of Finance 56 87ndash130Brennan M J and Schwartz E S (1984) Optimal financial policy and firm valuation

Journal of Finance 39 593ndash607Cameron A C and Trivedi P K (2010) Microeconometrics using Stata Stata Press TexasConstantinides G M (1979) Multiperiod consumption and investment behavior with

convex transactions costs Management Science 25 1127ndash1137Constantinides G M and Richard S F (1978) Existence of optimal simple policies for

discounted-cost inventory and cash management in continuous time Operations Research

26 620ndash636DeAngelo H DeAngelo L and Whited T M (2011) Capital structure dynamics and

transitory debt Journal of Financial Economics 99 235ndash261Dixit A K (1993) The Art of Smooth Pasting Routledge London and New YorkFama E and French K R (2002) Testing trade-off and pecking order predictions about

dividends and debt Review of Financial Studies 15 1ndash33Faulkender M and Petersen M A (2006) Does the source of capital affect capital struc-

ture Review of Financial Studies 19 45ndash79Faulkender M Flannery M J Hankins K W and Smith J M (2012) Cash Flows and

Leverage Adjustments Journal of Financial Economics 103 632ndash646Fischer E Heinkel R and Zechner J (1989) Dynamic capital structure choice theory

and tests Journal of Finance 44 19ndash40Frank M Z and Goyal V K (2008) Trade-off and pecking order theories of

debt in B E Eckbo (ed) Handbook of Corporate Finance Empirical Corporate Finance

Vol 2 Handbook of Finance Series Chapter 12 ElsevierNorth-Holland Amsterdam pp

135ndash202

Frank M Z and Goyal V K (2009) Capital structure decisions which factors are reliably

important Financial Management 38 1ndash37

Garvey G T and Hanka G (1999) Capital structure and corporate control the effect of

antitakeover statutes on firm leverage Journal of Finance 54 519ndash546

Goldstein R Ju N and Leland H (2001) An EBIT-based model of dynamic capital

structure Journal of Business 74 483ndash512

Graham J R and Tucker A L (2006) Tax shelters and corporate debt policy Journal of

Financial Economics 81 563ndash594

Hackbarth D and Mauer D C (2012) Optimal priority structure capital structure and

investment Review of Financial Studies 25 747ndash796

Halling M Yu J and Zechner J (2012) Leverage dynamics over the business cycle

Working paper University of Utah University of New South Wales and Vienna

University of Economics and Business AdministrationHennessy C A and Whited T A (2005) Debt dynamics Journal of Finance 60 1129ndash1165

Huang R and Ritter J R (2009) Testing theories of capital structure and estimating the

speed of adjustment Journal of Financial and Qualitative Analysis 44 237ndash271

1452 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

Hovakimian A Opler T and Titman S (2001) The debt-equity choice Journal of

Financial and Quantitative Analysis 36 1ndash24Jung K Kim Y and Stulz R M (1996) Timing investment opportunities managerial

discretion and the security issue decisions Journal of Financial Economics 42 159ndash185Kane A Marcus A J and McDonald R L (1984) How big is the tax advantage to

debt Journal of Finance 39 841ndash853Kayhan A and Titman S (2007) Firmsrsquo histories and their capital structures Journal of

Financial Economics 83 1ndash32Leary M T (2009) Bank loan supply lender choice and corporate capital structure

Journal of Finance 64 1143ndash1185Leary M T and Roberts M R (2005) Do firms rebalance their capital structures

Journal of Finance 60 2575ndash2619Leland H E (1994) Corporate debt value bond covenants and optimal capital structure

Journal of finance 49 1213ndash1252Lemmon M L Roberts M R and Zender J F (2008) Back to the beginning persist-

ence and the cross-section of corporate capital structure Journal of Finance 63 1575ndash1608Long M S and Malitz I B (1985) The investment-financing nexus some empirical

evidence Midland Corporate Finance Journal 3 53ndash59Mackie-Mason J K (1990) Do taxes affect corporate financing decisions Journal of

Finance 45 1471ndash1493Morellec E Nikolov B and Schurhoff N (2012) Corporate governance and capital

structure dynamics Journal of Finance 67 803ndash848Myers S C (1984) The capital structure puzzle Journal of Finance 39 575ndash592Myers S C (1993) Still searching for optimal capital structure Journal of Applied

Corporate Finance 61 4ndash14Rajan R and Zingales L (1995) What do we know about capital structure some evidence

from international data Journal of Finance 50 1421ndash1460Strebulaev I A (2007) Do tests of capital structure theory mean what they say Journal of

Finance 62 1747ndash1787

Strebulaev I A and Whited T M (2012) Dynamic Models and Structural Estimation in

Corporate Finance Now Publisher Boston and Delft

Titman S and Wessels R (1988) The determinants of capital structure choice Journal of

Finance 43 1ndash21

Welch I (2004) Capital structure and stock returns Journal of Political Economy 112

106ndash131

Welch I (2011) Two common problems in capital structure research the financial-debt-

to-asset ratio and issuing activity versus leverage changes International Review of Finance

11 1ndash17

PROFITSndashLEVERAGE PUZZLEREVISITED 1453

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

tend to have high leverage A higher market-to-book ratio is associated withlower leverage Larger firms are typically more highly levered The coeffi-cients on tangibility are positive The conclusion from Table II is clear Theleverage regression results match those reported in previous studiesAt times one may be concerned that in a regression what is being

estimated is a conditional mean while some parts of the distributionmight behave differently from other parts If so then ldquoaccidentsrdquo likesample selection issues or asymmetries in the underlying distributionsmight play a greater role than is usually recognized In other words thesource of identification could be misinterpreted To guard against this po-tential problem we estimate the same model using quantile regressions9

Columns 3 4 and 5 report what happens at the 25th 50th and 75th per-centiles The basic model results are robust in terms of both the sign and thestatistical significanceThe impact of profits does appear to be somewhat stronger among the

high leverage firms For instance in a book leverage regression the coeffi-cient on profits for the first quartile is 011 with a t-ratio of 378 For thethird quartile the coefficient on profits is 041 with a t-ratio of 1015This difference is curious and might deserve further study However the factthat in each case we get the negative sign and statistical significance is suf-ficient for the purposes of the present article So the results on profits arerather robust across the distributionTable III sorts firms into those that are actively adjusting their leverage

and those that are passive In the presence of both fixed and variable trans-acts costs this distinction is important We expect the coefficient on profit-ability in leverage ratio regressions to be more negative in firms that arepassive and less negative in firms that are actively adjustingWe define active firms as those that are either issuing net debt or net

equity in excess of 5 of their assets Passive firms are those that neitherissue debt nor equity in nontrivial amounts (ie greater than 5 of theirassets) Consistent with our expectations we find in Table III that the coef-ficients on profits are significantly more negative for passive firms than theyare for active firms With firm and year fixed effects the coefficient onprofits is 008 for active firms and 017 for the passive firms The twocoefficients are significantly different from each other (plt 001) InTable III the standard control variables are used They have the usualsigns and for the most part do not differ much between the active and thepassive firms

9 Cameron and Trivedi (2010) provide an extensive discussion of quantile regressions inStata

1428 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

5 Financing Activity

51 FREQUENCY OF DEBT AND EQUITY ISSUES

Table IV reports the percentage of firms issuing or repurchasing debt orequity for annual sorts based on lagged profitability In Panel A weemploy the conventional 5 cutoff rule to exclude minor fluctuations anddefine firms ldquoissuing debtrdquo as those that issue debt (both short-term andlong-term) in excess of 5 of the value of their assets Other decisions aresimilarly defined using a 5 cutoff Appendix Table AII examines the dis-tributional properties across profitability (and size) quintiles for ldquolargerdquo is-suances by increasing the 5 cutoff to higher levels such as 10 and 15How does security issuance behavior vary with firm profitability To

answer this question Table IV sorts firms by profitability and reports the

Table III Leverage and profitability for active and passive firms

The table presents estimates of leverage on firm characteristics for the subsamples of firms

that are active in either debt or equity markets Active firms are defined as those that issueeither net debt in excess of 5 of the value of assets or issue net equity in excess of 5 ofthe value of their assets The passive firms do not issue significant (more than 5) of either

debt or equity The sample comes from the annual Compustat files during the period 1971ndash2009 Financial firms are excluded The variables are described in Tables I and II Wereport t-statistics where the standard errors are clustered at the firm level in parentheses

aSignificant at the 1 level

Active Passive

(1) (2) (3) (4)

Profitabilityt1 006a 008a 022a 017a

(126) (148) (254) (188)

IndLevt1 044a 019a 041a 013a

(424) (149) (326) (127)MB

t1

002a 001a 003a 001a

(362) (196) (327) (172)

Tangibilityt1 010a 014a 014a 019a

(146) (118) (145) (130)

LnethAssetsTHORNt1 001a 004a 001a 004a

(121) (179) (147) (169)

Constant 012a 008a 003a 007a

(120) (57) (45) (42)

Year FE Yes Yes Yes Yes

Firm FE No Yes No Yes

R2 034 064 033 075

N 43395 43395 58943 58943

PROFITSndashLEVERAGE PUZZLEREVISITED 1429

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

percentage of firms issuing or repurchasing debt or equity The table alsoreports test statistics from tests that examine whether the percentage differ-ences between the high- and low-profit firms are significantly different fromeach other The results in Column (1) show that the likelihood of issuing debt

Table IV Profitability sorts for debt and equity issuers

The sample contains nonfinancial firms listed on the annual Compustat files for the period

from 1971 to 2009 The table presents the percentage of firms issuing and retiring (orrepurchasing) debt and equity The firms are classified as ldquoissuing debtrdquo if they issuedebt in excess of 5 of the value of their assets as ldquoissuing equityrdquo if they issue equity

in excess of 5 of the value of their assets as ldquoretiring debtrdquo if they retire debt in excess of5 of the value of their assets and as ldquorepurchasing equityrdquo if they repurchase equity inexcess of 5 of the value of their assets In addition we also report net debt issuers which

are firms that issue net debt over 5 of the value of their assets and net equity issuerswhich are firms that issue net equity in excess of 5 of the value of their assets Weannually sort firms on lagged profitability and report the percentage of firms in each ofthese categories The bottom part of each panel reports the percentage of firms issuing or

retiring securities by profitability within the smallest and largest asset quintiles The tablealso reports results from t-test for the difference in percentage of firms issuing or retiringsecurities in the lowest and the highest profit quintile

Percentage of firms

Iss Ret Iss Iss Rep Iss Both IssD IssE

D D NetD E E NetE DampE RepE RetD

(1) (2) (3) (4) (5) (6) (7) (8) (9)

Allfirm

s

Low 276 245 186 301 13 297 71 01 38

2 319 299 181 117 21 112 27 02 20

3 385 344 194 86 29 80 21 05 15

4 386 330 213 81 43 75 20 07 14

High 329 273 198 92 82 81 17 13 14

t Highfrac14Low 150 84 39 723 414 763 358 176 200

Smallfirm

s Low 275 204 201 411 11 405 109 01 47

2 291 233 213 379 10 378 99 01 45

3 281 246 204 265 14 261 66 01 41

4 275 272 186 162 15 159 34 02 35

High 275 281 178 136 27 131 25 02 26

t Highfrac14Low 01 84 28 316 55 318 171 18 55

Largefirm

s

Low 414 363 177 87 22 79 28 03 11

2 449 379 180 62 30 57 17 07 09

3 439 360 186 56 41 48 16 09 07

4 403 320 194 46 67 39 13 12 06

High 374 264 207 50 129 35 10 28 05

t Highfrac14Low 50 130 45 89 244 116 77 120 47

1430 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

is mostly independent of firm profitability In the aggregate there is a smalldifference but it is sensitive to sorting on firm size Similarly in Column (2)there is only a weak relation between the profitability and the likelihood thatthe firm retires debt To a first approximation the fraction of firms issuing orrepurchasing debt is largely independent of profits Large firms are somewhatmore likely than small firms to issue debt and also to retire debtUnlike debt issues the probability of issuing equity is strongly related to

profitability as can be seen in Column (4) Low-profit firms are much morelikely to issue equity than are high-profit firms In the lowest profit quintile301 of the firms issued equity In the highest profit quintile it was only92 The lower panel reports separate sorts for small and large firms Similarpatterns are found for both groups of firms that is the likelihood of equityissuance declines as profitability increases The differences between the quin-tiles are quite stark for the group of small firms All quintiles of small firmsissue equity much more frequently than the corresponding large firm quintilesHigh-profit firms are much more likely to repurchase equity as shown in

Column (5) In the overall sample only 13 of low-profit firmsrepurchased equity compared with 82 of the high-profit firms Whenwe separate the small and large firms again a difference emerges Ingeneral small firms do not repurchase much equity while a greaterfraction of the large firms do soGiven these facts it is natural that low profitability firms are much more

likely to be net issuers of equity than are high profitability firms as shown inColumn (6) Among the firms in the lowest quintiles of profits roughly 30of firms issue net equity that exceeds 5 percent of their assets In contrastfor the most profitable quintile only about 8 issue net equityFirm size is an important variable in the recent literaturemdashit is often used

as a proxy for access to capital markets (as in Faulkender and Petersen(2006) and Leary (2009)) Small firms are bank-dependent risky and infor-mationally opaque They have restricted access to public debt markets andconsequently face more severe supply constraints in their ability to issuedebt Thus we expect small firms to be more sluggish in adjusting theirdebt and equity in response to shocks to profitability Large firms incontrast have much easier access to public debt markets and they facefewer obstacles in accessing securities marketsTo examine how size and profitability interact the bottom part of Table IV

examines profitability sorts for the smallest and largest firms We first sortfirms annually by firm size and then within each size quintile we sort them onprofitability We do these two-way sorts to ensure that we have similarnumber of firms in profitability subgroups for both small and large firmsamples Unconditionally small firms are generally less profitable while

PROFITSndashLEVERAGE PUZZLEREVISITED 1431

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

large firms are relatively more profitable The two-way sorts allow us to havea uniform distribution of firms in various profitability quintiles for both smalland large firm samples The bottom part of Table IV reports results for thesmallest and the largest size quintiles Among the small firms there is littlerelation between profitability and the likelihood of issuing debt Howeverlarger firms exhibit a small increase in the likelihood of net issuance of debtwith increasing profitabilityIn contrast to debt the effects of profitability on equity issuance and

repurchases are much more consistent across size quintiles As profitabilityincreases firms are generally less likely to issue equity and more likely torepurchase it Across the two extreme size groups we note a striking differ-ence between the proportion of firms issuing equity and those issuing debtAmong low-profit small firms almost 41 are equity issuers whereas onlyabout 9 of low-profit large firms are equity issuers However regardless ofsize we note a monotonic reduction in the likelihood of issuing equity asprofitability increasesColumn (7) reports the fraction of firms that issue both debt and equity in

excess of 5 of assets both as a function of profits in the upper panel andas a function of firm size and profits in the lower panel In the upper panelwe see that low-profit firms are much more likely to issue both debt andequity (71) while the high-profit firms are less likely to do so (17) Thelower panel shows that small firms are much more likely to be issuers of bothin contrast to large firms In both firm size categories the profit pattern is thesame with low profits being more frequent issuers of both simultaneouslyImportantly Column (8) shows that the likelihood of issuing debt and

simultaneously repurchasing equity increases with profitability Converselyin Column (9) we find that the likelihood of doing the reverse that is issuingequity and retiring debt declines with profitability The effects of sorting onfirm size and firm profitability mirror those reported for all firms Overallthe results show that low-profit firms are less likely to issue debt and repur-chase equity they are instead more likely to issue equity and retire debtFirms with high profitability exhibit the reverse patternIn Appendix Table AII we present results from profitability (and size)

quintiles for large issuances defined using the 10 cutoff (in Panel A) andthe 15 cutoff (in Panel B) The results from these panels are largely con-sistent with those described above

52 MAGNITUDES OF ISSUANCES AND EQUITY CHANGES

In the previous section we considered the probability of having a nontriviallevel of debt or equity activity The next question is how large are the dollar

1432 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

values involved In Table V we sort firms according to profits and thentabulate the levels and changes in both debt and equity We do this firstfor all firms and then for small versus large firmsFor the analysis on all firms we observe in Column (1) that debt peaks at

the middle of the distribution This is because firms with medium profitabil-ity are also the largest as seen in Column (8) which reports average assetvalues for different profitability quintiles When we sort first by firm size andthen examine firms with different levels of profitability within size quintileswe find that for small firms debt is roughly independent of profits while forlarge firms less profitable firms have a higher level of debtAs expected Columns (3) and (5) show that more profitable firms have

higher equity values Columns (2) and (7) consider the relationship betweenissuances of debt and equity and firm profitability Debt issuances are sig-nificantly larger for more profitable firms In contrast equity issuances arelarger for low-profit firms Small low-profit firms issue substantially lessequity compared with small high-profit firms Among large firms the low-profit ones tend to issue equity while the high-profit ones tend to repurchaseequity The fact that more profitable firms issue debt and repurchase equitywhile the least profitable firms retire debt and issue equity is consistent withthe predicted relation between profitability and financing decisions under thetrade-off theoryWe also examine two-way sorts by size and profitability and report results

for the smallest and largest quintile of firms For the smallest quintile offirms the change in debt is largely unrelated to profits But for large firmsthere is a positive relation between profits and debt issuances High-profitfirms have a big positive change in debt Low-profit large firms have anegative change in debtColumns (4) and (6) provide an explanation of why the leverage ratio

regression results contradict those from the basic profitability sorts pre-sented here As we can see profitability indirectly affects leverage ratiosby increasing equity values Changes in both the book value of equity andthe market value of equity are positive and large for highly profitable firmsIn contrast these changes are negative for less profitable firms This evidenceillustrates an important issue concerning the use of leverage ratios Suchratios are often interpreted as essentially reflecting the use of debt by thefirm This interpretation while common is empirically misleadingFor the typical firm the change in the value of equity is larger than the

change in debt For example in the third profit quintile for large firmscolumn (7) the mean equity issue is just $9 million but the change in themarket value of equity is $242 million At the same time the mean change indebt is $27 million This suggests that a fair bit of the observed variation in

PROFITSndashLEVERAGE PUZZLEREVISITED 1433

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

the leverage ratios is primarily driven by the changes in the market value ofequity in the denominator rather than by the changes in debt in the numer-ator Since equity issues are often small this implies that the variation in theleverage ratio is primarily driven by internal operations rather than byexternal financing actions This again points to the fact that leverageratios can provide a misleading account of actual patterns in the dataThere is always a potential concern that the averages may be misleading

due to the impact of outliers To address this concern in AppendixTable AIII we present median values of the profitability sorts Theseresults generally reinforce the findings in Table VTable VI again considers the magnitudes of financing activity but this

time the issues are scaled by a measure of firm size That way it is easier to

Table V Magnitude of financing activity

The sample contains nonfinancial firms listed on the annual Compustat files for the period

from 1971 to 2009 The table presents average debt and equity changes in debt and equityand equity issuances (in millions of dollars) for firms sorted on profitability within sizeclasses The table reports information for all firms sorted on profitability and for profit-

ability sorts within the smallest and largest firms The sorts are done annually The tablealso reports results from t-test for the difference in percentage of firms issuing or retiringsecurities in the lowest and the highest profit quintile

D D BVE BVE MVE MVE EqyIss Assets

(1) (2) (3) (4) (5) (6) (7) (8)

Allfirm

s

Low 83 5 78 10 120 5 7 234

2 605 13 509 24 721 8 17 1691

3 1227 9 1123 22 1834 3 9 3551

4 833 23 1166 7 2408 74 9 2994

High 570 53 1171 71 3817 68 36 2581

t Highfrac14Low 31 9 36 9 38 2 16 36

Smallfirm

s Low 25 04 42 01 332 28 33 97

2 29 06 55 02 278 16 21 116

3 31 07 63 01 212 05 12 127

4 30 06 70 05 179 10 08 136

High 24 06 84 14 234 17 07 142

t Highfrac14Low 04 15 153 85 73 11 201 107

Largefirm

s Low 3929 55 2990 152 4671 11 63 10846

2 3826 19 3178 95 5122 66 11 10721

3 3074 27 3498 41 6173 242 9 9697

4 2623 62 3915 45 8473 94 57 9855

High 2180 193 4642 261 14924 202 148 10171

t Highfrac14Low 94 64 108 87 246 12 169 16

1434 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

see whether large firms behave disproportionately when compared withsmall firms Column (1) shows that low-profit firms have more debtas reported in many previous studies In Columns (4) and (6) we seethat low-profit firms have a major decline in both the book value ofequity and the market value of equity The high-profit firms have thereverse experienceMore interesting are the active decisions of the firms In Column (2) we see

that in contrast to the usual impression the low-profit firms are reducingdebt while the high-profit firms are increasing it In Column (7) we see thatthe low-profit firms are also reducing leverage by issuing equity The high-profit firms issue little equity In the lower panel similar results are foundwhen we separate small and large firms Equity issuing is particularly strong

Table VI Magnitude of financing activitymdashscaled levels and changes

The sample contains nonfinancial firms listed on the annual Compustat files for the period

from 1971 to 2009 The table presents average debt and equity changes in debt and equityand equity issuances scaled by assets for firms sorted on profitability within size classes Thetable reports information for all firms sorted on profitability and for profitability sorts

within the smallest and largest firms The sorts are done annually The table also reportsresults from t-test for the difference in percentage of firms issuing or retiring securities inthe lowest and the highest profit quintile

D D BVE BVE MVE MVE EqyIss

(1) (2) (3) (4) (5) (6) (7)

Allfirm

s

Low 386 122 12 871 3427 1868 217

2 285 58 407 119 940 157 28

3 293 12 425 14 742 03 12

4 269 26 461 02 955 27 10

High 192 15 551 47 1675 74 07

t Highfrac14Low 35 49 50 47 127 42 157

Smallfirm

s Low 545 154 841 2145 9688 9128 696

2 684 49 712 1915 5014 2318 288

3 331 253 141 969 3063 954 129

4 267 181 357 175 1673 221 55

High 194 05 543 26 1758 147 41

t Highfrac14Low 49 17 51 36 88 27 99

Largefirm

s Low 350 30 306 58 511 55 09

2 346 14 335 13 523 02 04

3 353 165 332 71 711 29 01

4 280 01 414 04 909 09 05

High 227 15 477 19 1643 27 14

t Highfrac14Low 371 112 453 181 570 57 221

PROFITSndashLEVERAGE PUZZLEREVISITED 1435

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

among the small low-profit firms that are also experiencing sharp reductionsin the value of equity In column (2) we see that both for the large and for thesmall firms net debt increases are primarily found in the highest quintile offirm profitability

6 Debt and Equity Regressions

So far we have demonstrated that in our data the conventional leverageregressions have the usual signs We then demonstrated that in simple sortsof the data strikingly different active financing patterns emerge The nexttask is to reconcile these differencesTable VII presents simple regressions predicting changes in debt book

equity and market equity as well as equity issuances Consistent with thesorts we control for changes in firm size and we also include firm and yeardummies There may be a concern about the speed of a firmrsquos responsesAccordingly we consider the lagged change in profits as the main case butwe also include the contemporaneous change in profits Whether the contem-poraneous change in profits is included or not does not in fact matter Wereport panel-robust standard errors adjusted for clustering at the firm levelThe results in Table VII support the findings in the sorts Firms that have

an increase in profits in 1 year increase debt both in the same year and in thenext year Firms that have an increase in profits experience an increase inboth the book value of equity and the market value of equity Importantlyhowever firms that experience and increase in profits reduce their equityissues All these effects are statistically significant (except for the marketvalue of equity in the column 5 specification) and seem intuitively reason-able In Appendix Table AIV similar regressions to those presented aboveare reported but without firm-fixed effects We find that empirically verylittle changes whether we include firm fixed effects or leave them out Thusthe targeting behavior to the extent that it happens does not account forwhat we are reportingThe next step is to bring the sorts and the simple regressions together into

a common setting We therefore include factors in addition to profitability(i) median industry leverage (ii) the market-to-book assets ratio (iii) tangi-bility of assets and (iv) firm size (measured by the log of assets) Rajan andZingales (1995) show that these factors are related to leverage in G7countries A number of studies have used these factors to estimateleverage targets Frank and Goyal (2009) show that these factors arerobustly related to leverage in the USA In these regressions we use indica-tors for the quintile that the firm is in for each factor

1436 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

Table VIII reports these results In Columns (1) and (2) we examine debtissues and again show that profits positively affect debt issuances The effectsare large and statistically significant at the 1 level Columns (3) and (4)which examine book equity show a strong effect of profits This effect is onlyslightly reduced by the inclusion of other factors Columns (5) and (6) illus-trate that the impact of profits on the change in market equity is also robustto the inclusion of the conventional factors Columns (7) and (8) examineequity issuances Here we again find that profits have a significant negativeimpact on equity issuances As before the results are robust to the inclusionof the conventional factorsIn unreported tables we estimate debt and equity changes using quantile

regressions In terms of the signs and significance of the profitability variable

Table VII Debt and equity changes

The sample contains nonfinancial firms listed on the annual Compustat files for the period

from 1971 to 2009 The table presents fixed effect estimates from regressions of changes indebt (D) changes in the book value of equity (BVE) changes in the market value ofequity (MVE) and net equity issuances (EqyIss) The explanatory variables include

current and lagged change in operating income before depreciation (Profits andProfitst1) and lagged change in the book value of assets (Assetst1) In addition tofixed firm effects the regressions include year indicator variables The reported t-statistics

are corrected for clustering at the firm level

aMean significant at the 1 level

bMean significant at the 5 level

cMean significant at the 10 level

D BVE MVE EqyIss

(1) (2) (3) (4) (5) (6) (7) (8)

Profitst1 022a 025a 015b 035a 004 060b 005b 007a

(33) (31) (25) (35) (03) (25) (25) (53)

Profits 050b 078a 140a 003a

(24) (32) (48) (34)

Assetst1 000 007b 025a 001b

(00) (21) (35) (25)

Constant 167 283a 34 51 1145a 1182a 605a 595a

(14) (30) (04) (05) (32) (29) (69) (74)

Year FE Yes Yes Yes Yes Yes Yes Yes Yes

Firm FE Yes Yes Yes Yes Yes Yes Yes Yes

R2 001 004 001 010 001 005 002 003

N 162056 162056 162130 162130 157550 157550 162154 162154

PROFITSndashLEVERAGE PUZZLEREVISITED 1437

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

the results are qualitatively similar to the OLS results reported in Table VIIIProfitability positively affects debt issuances and negatively affects equityissuances across various quantiles Overall the results in Table VIII showthat there is nothing that is special about using sorts or regressions to explaindebt and equity issues In either case we find that more profitable firms tendto increase their debt experience an increase in the value of equity andrepurchase shares Thus the control factors are not responsible for theusual rejection of the trade-off theory It would appear that what mattersis the size of the issuing activity relative to the organic increases in equityvalue due to profits In other words what matters in the leverage regressionseems to be coming from the presence of E in D

DthornE and not from D

Table VIII Debt and equity issuances profitability and leverage factor quintiles

The sample contains nonfinancial firms listed on the annual Compustat files for the period

from 1971 to 2009 The table presents estimates from regressions of change in debt (D)change in the book value of equity (BVE) change in the market value of equity (MVE)and net equity issuances (EquityIss) The explanatory variables include lagged profitability

quintiles and lagged leverage factor quintiles In addition to fixed firm effects the regres-sions include year indicator variables The reported t-statistics are corrected for clustering atthe firm level

aMean significant at the 1 level

bMean significant at the 5 level

cMean significant at the 10 level

D BVE MVE EqyIss

(1) (2) (3) (4) (5) (6) (7) (8)

ProfitabilityQint 214a 167a 283a 175a 80 429a 53a 55a

(70) (54) (66) (59) (08) (48) (55) (62)

IndLevQint 104a 43 99 20

(29) (15) (10) (10)

Ln(Assets)Qint52 17 1026a 53a

(16) (05) (77) (35)MB

Qint158a 372a 1155a 09

(45) (56) (106) (06)

TangibilityQint 05 177 436c 64a

(01) (15) (18) (31)

Constant 167 96 656b 235 4613a 8610a 113 570a

(08) (04) (26) (04) (61) (58) (11) (31)

Year FE Yes Yes Yes Yes Yes Yes Yes Yes

Firm FE Yes Yes Yes Yes Yes Yes Yes Yes

R2 0002 0002 0003 0004 0013 0014 0013 0014

N 155587 155037 155587 155037 155587 155037 155587 155037

1438 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

7 Are Financial Market Conditions Important

According to the trade-off theory capital structure is determined by con-sidering the impact of costs and benefits of debt The time variation inthe costs and benefits of debt imply that issuance decisions would alsovary over time Furthermore market conditions affect the transactioncosts of adjusting Adverse selection is a greater problem in a cold marketthan in a hot one Accordingly it may be easier to adjust in good marketconditions Related ideas have been developed by Huang and Ritter(2009) Faulkender et al (2012) and Halling Yu and Zechner (2012)Faulkender Flannery Hankins and Smithrsquos paper examines cross-sectionalvariation in adjustment speeds including market valuations andhow they may affect adjustment speeds It is natural to think that ingood market conditions it will be less costly to issue both debt and equityand adjusting leverage to respond to profitability shocks would be mucheasierTo test the importance of market conditions we require a definition of

good times and bad times Our empirical strategy is to estimate good timesversus bad times at the four-digit industry level We define an industry ashaving ldquogood timesrdquo if the median firm in that industry has a market-to-book ratio that is higher than the 67th percentile of the time-series distribu-tion of the industry median market-to-book ratios Conversely an industryis defined as having ldquobad timesrdquo if the median firm in the industry has amarket-to-book ratio that is lower than the 33rd percentile of the time-seriesdistribution of the industry median market-to-book ratiosPanel A of Table IX tabulates issuance activity for profitability sorts in

both good times and bad times As might be expected active debt and equityissuances are larger during good times There is also somewhat more activeswaps between debt and equity during good timesPanel B reports the issuance activity for the smallest and the largest firms

during the good times and bad times Within each size quintile firms aresorted on profitability This is a three-way sort As before we find that thereis much more active use of external markets during good times Small low-profit firms are more likely to issue equity in good times than in bad timesDebt issuances are significantly higher in good times Large high-profitfirms are significantly more likely to issue debt and repurchase equity ingood times than in bad timesPanel C reports the magnitudes of the financial variables rather than the

frequencies During bad times less profitable large firms retire substantialamounts of debt and they show a tendency to issue equity Small firms donot seem to engage in similar debt reduction activities However like the

PROFITSndashLEVERAGE PUZZLEREVISITED 1439

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

Table IX Debt and equity issues in good and bad times

The table reports the frequency and magnitude of financing activity for sorts on profits for subsamples

of firms in good and bad times An industry is defined as having ldquogood timesrdquo if the median firm in

that industry has a market-to-book ratio that is higher than the 67th percentile of the time-series

distribution of industry median market-to-book ratios Conversely an industry is defined as having

bad times if the median firm in that industry has a market-to-book ratio that is lower than the 33rd

percentile of the time-series distribution of industry median market-to-book ratios Panels A and B

report the percentage of firms (i) issuing net debt in excess of 5 of the value of their assets (ii) issuing

net equity in excess of 5 of the value of their assets (iii) issuing debt and repurchasing equity both in

excess of 5 of the value of their assets and (iv) issuing equity and retiring debt both in excess of 5

of the value of their assets Panel C tabulates the mean debt and equity levels and the mean changes in

debt and equity issuances The sample contains nonfinancial firms listed on the annual Compustat files

for the period from 1971 to 2009

Iss Iss Iss D Iss ENet D Net E Rep E Ret D(1) (2) (3) (4)

Panel A sorts on profitability

Badtimes

Low 145 131 01 202 170 49 02 083 173 46 04 074 182 47 03 10

High 177 51 06 10t Highfrac14Low 41 136 33 01

Goodtimes

Low 206 381 01 462 197 148 02 233 212 103 04 184 234 92 10 18

High 216 92 16 14t Highfrac14Low 23 658 138 23

Panel B sorts on firm size and profitability

Badtimes

Smallfirm

s Low 151 318 00 312 182 276 02 253 151 140 02 294 183 73 00 19

High 177 80 00 11t Highfrac14Low 10 90 NA 20

Largefirm

s Low 172 54 01 042 148 39 04 053 150 43 04 044 156 33 03 06

High 147 28 05 03t Highfrac14Low 17 32 20 02

Goodtimes Smallfirm

s Low 213 448 01 542 230 443 01 473 222 325 00 474 203 197 02 38

High 190 153 02 31t Highfrac14Low 19 230 10 39

Largefirm

s Low 203 96 03 122 208 66 09 133 232 54 16 064 215 39 19 05

High 234 36 38 04t Highfrac14Low 29 99 92 38

(continued)

1440 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

larger firms there is a tendency for small firms to issue more equitymdashat leastto the extent possibleDuring good times large profitable firms raise significant amounts of debt

as they experience an increase in the value of their equity Such firms alsoengage in share repurchases Debt issues by smaller firms are much lessaffected by their own profitability during good times and their change indebt is more or less independent of their profits However equity issuanceis primarily found among the low-profit small firms during good timesTable IX shows that empirically profitability has a time-varying impacton leverage choicesIn Table X the basic regressions for changes in debt and for equity

issuance are presented for both good times and bad times Comparinggood times with bad times we see that the effects of profits are muchstronger in good times In bad times the impacts are rather weak statisticallyThis difference is interesting and deserves further study However for ourpurposes recall that the static trade-off theory implies time-varying capital

Table IX Continued

D D BVE BVE MVE MVE EqyIss Assets

(1) (2) (3) (4) (5) (6) (7) (8)

Panel C magnitude of financing activity

Badtimes Smallfirm

s Low 40 03 58 14 305 06 22 140

2 44 05 82 07 240 67 17 180

3 45 04 94 02 176 18 05 187

4 43 09 98 02 137 04 03 188

High 36 10 114 13 209 16 04 199

Largefirm

s Low 3365 73 2523 169 2304 163 94 8972

2 3031 95 2983 35 3284 278 56 9068

3 2930 5 3360 14 3947 308 57 9225

4 2281 6 3817 19 5213 442 7 8902

High 1790 62 4767 13 8799 58 31 9573

Goodtimes

Smallfirm

s Low 24 06 38 05 339 23 39 90

2 26 08 48 01 301 04 25 104

3 30 09 60 04 246 03 17 125

4 28 07 66 08 206 18 12 130

High 23 06 85 16 261 13 08 141

Largefirm

s Low 4532 36 3477 103 6557 260 53 12465

2 4580 58 3527 139 6911 421 23 12522

3 2952 111 3485 2 7586 279 32 9535

4 2752 124 3939 72 10686 33 105 10226

High 2429 228 4819 377 18565 268 222 10830

PROFITSndashLEVERAGE PUZZLEREVISITED 1441

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

structure choices even when the target is time invariant Empirically we doobserve time-varying choices

8 Is Scaling Important

In the preceding section we provided results for unscaled debt and equityissues This seems appropriate to us However it involves making twochanges to the standard regression We would like to determine if both

Table X Debt and equity changes in good and bad timesmdashregressions

The sample contains nonfinancial firms listed on the annual Compustat files for the period

from 1971 to 2009 The table presents estimates from regressions of change in debt (D)and net equity issuances (EqyIss) in good and bad times The explanatory variables includechange in profits lagged change in profits and lagged change in assets The regressions

include both firm and year fixed effects An industry is defined as having ldquogood timesrdquo ifthe median firm in that industry has a market-to-book ratio that is higher than the 67thpercentile of the time-series distribution of industry median market-to-book ratios

Conversely an industry is defined as having bad times if the median firm in thatindustry has a market-to-book ratio that is lower than the 33rd percentile of the time-series distribution of industry median market-to-book ratios The reported t-statistics arecorrected for clustering at the firm level

aMean significant at the 1 level

bMean significant at the 5 level

cMean significant at the 10 level

Good times Bad times

D EqyIss D EqyIss

(1) (2) (3) (4) (5) (6) (7) (8)

Profits 087a 091a 006b 009a 060 065 003 003b

(30) (29) (20) (28) (15) (16) (16) (20)

Profitst1 029c 015a 034c 001

(19) (37) (17) (04)

Assetst1 000 002b 000 001

(02) (22) (00) (07)

Constant 1337 3989 8691a 10001a 1351 1851 4099a 4055a

(03) (09) (55) (37) (09) (13) (80) (78)

Year FE Yes Yes Yes Yes Yes Yes Yes Yes

Firm FE Yes Yes Yes Yes Yes Yes Yes Yes

R2 0071 0080 0018 0045 0178 0212 0034 0041

N 74827 68729 74891 68787 26965 26013 26971 26018

1442 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

changes are crucial Accordingly in this section we take a step back toconsider prior literature in which variables are scaled The questions arehow does scaling affect the results and whyTable XI examines alternative scalings Columns (1) and (2) scale debt and

equity issuances by the total firm net issuance Columns (3) and (4) scaledebt and equity issuances by the sum of debt and book equity Columns (5)and (6) consider the change in debt scaled by the sum of debt and bookequity Columns (7) and (8) consider the change in leverage due to externalfinancing Since we are examining changes we use the changes in thestandard regression factors as controls The primary interest is in the coef-ficient on the change in profitsWe find that a change in profits is associated with an increase in debt

issues But the sign reverses when we scale the debt issues by the sum of debtplus book equity This is telling What is really driving the results is thechange in equity in the denominator not the change in debt in the numer-ator This result is further substantiated in Columns (5) and (6) where thechange in debt to capital ratio is used instead of debt issuanceTo further explore the impact of active issuance relative to passive

experiencing of changes we decompose leverage by separating out theeffect of debt and equity issues from those due to changes in retainedearnings Thus we define leverage change due to external financing as

LevEF frac14Dt1 thorn d

Dt1 thorn Et1 thorn dthorn e

Dt1

Dt1 thorn Et1 eth1THORN

where D is debt E is book value of equity d is debt issues net of retirementsand e is equity issues net of repurchases Garvey and Hanka (1999) use asimilar measure in their study This measure directly captures the netleverage effect of debt and equity issuancesColumns (7) and (8) report results from regressions of the net leverage

effect of debt and equity issues on changes in profitability after controllingfor other leverage factors in differences and year indicator variables Theeffect of profitability on leverage change due to external financing is positiveand statistically significant This comparison suggests that changes in totaldebt and equity are only partly a result of debt and equity issuance decisionsOther balance sheet adjustments complicate the inferences from leverageratio regressionsFor most of our analysis we follow the literature in treating debt as homo-

genous It is of course potentially interesting to ask what happens when wedistinguish between short-term debt and long-term debt Which part of thefirmrsquos debt maturity structure drives the under-adjustment Table XII

PROFITSndashLEVERAGE PUZZLEREVISITED 1443

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

provides results from such decompositions Alternative scalings areconsidered The main results of interest are that the distinction betweenshort- and long-term debt is important Short-term debt issuance is largelyindependent of changes in profits The negative sign on profit change stemsfrom the action of long-term debt This seems intuitively reasonable

Table XI Scaled issuances

The sample contains nonfinancial firms listed on the annual Compustat files for the period

from 1971 to 2009 The table presents estimates from regressions of debt issuance to totalissuance (in columns (1) and (2)) debt issuance to total capitalization (in columns (3) and(4)) change in debt to capitalization ratio (in columns (5) and (6)) and change in leverage

ratio considering only external financing (in columns (7) and (8)) The change in leverageratio considering only external financing is defined as

LevEF frac14Dt1 thorn d

Dt1 thorn Et1 thorn dthorn e

Dt1

Dt1 thorn Et1

where D is debt E is book value of equity d is debt issues net of retirements and e is equity

issues net of repurchases The regressions include lagged leverage factors and year indicatorvariables Some of the specifications include firm fixed effects The reported t-statistics arecorrected for clustering at the firm level

aMean significant at the 1 level

bMean significant at the 5 level

cMean significant at the 10 level

DebtIssTotIss

DebtIssethDthornETHORN D

DthornE LevEF

(1) (2) (3) (4) (5) (6) (7) (8)

Profitt1 004a 000 002a 001c 003a 002 001b 002a

(35) (03) (34) (18) (35) (14) (23) (26)

IndLevt1 004c 001 001 001c 000 000 000 000

(19) (06) (15) (19) (02) (04) (03) (05)

Market=bookt1 001a 001a 000a 000a 000a 000a 000 000c

(47) (85) (68) (60) (55) (36) (05) (18)

Tangt1 004 005 013a 011a 011a 006a 003a 003a

(12) (15) (144) (112) (79) (38) (44) (42)

Assetst1 008a 004a 006a 004a 002a 001a 003a 002a

(97) (45) (283) (161) (53) (35) (147) (101)

Constant 079a 074a 002a 003a 000 000 000b 002a

(497) (442) (91) (111) (00) (05) (22) (93)

Year FE Yes Yes Yes Yes Yes Yes Yes Yes

Firm FE No Yes No Yes No Yes No Yes

R2 0015 0098 0022 0076 0009 0070 0012 0086

N 133905 133905 140552 140552 138229 138229 140572 140572

1444 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

inasmuch as short-term debt is probably more associated with day-to-dayoperations rather than longer term planning considerationsIn Table XIII we examine change in leverage ( D

DthornE) and change inleverage due to external financing (LEVEF) during good and bad timesseparately The traditional negative effect of a change in profits on achange in leverage is found both in good times and in bad timesHowever when we decompose the impact in order to isolate the externalcomponent as before the result disappears In good times the externalfinancing component of a change in profits is weak but still positiveDuring bad times it is not statistically significant This implies that distin-guishing between good and bad times does not fundamentally alter the basic

Table XII Short- and long-term debt issuances

The sample contains nonfinancial firms listed on the annual Compustat files for the period

from 1971 to 2009 The table presents estimates from regressions of short-term debtissuance to total issuance (in columns (1) and (2)) long-term debt issuance to totalissuance (in columns (3) and (4)) short-term debt issuance to total capitalization (in

columns (5) and (6)) and long-term debt to total capitalization (in columns (7) and (8))Some of the specifications include firm fixed effects The reported t-statistics are correctedfor clustering at the firm level

aMean significant at the 1 level

bMean significant at the 5 level

cMean significant at the 10 level

STDebtIssTotIss

LTDebtIssTotIss

STDebtIssethDthornETHORN

LTDebtIssethDthornETHORN

(1) (2) (3) (4) (5) (6) (7) (8)

Profitabilityt1 0002 0003 0042a 0008 0001 0001 0019a 0012a

(02) (04) (30) (05) (14) (10) (45) (28)

IndLevt1 0009 0007 0070b 0033 0001 0002c 0014a 0005

(06) (05) (26) (11) (08) (17) (26) (09)

M=Bt1 0003a 0003a 0003b 0010a 0000 0000 0003a 0002a

(48) (35) (22) (65) (10) (01) (84) (42)

Tangibilityt1 0020 0007 0011 0067 0005a 0006a 0097a 0081a

(12) (03) (03) (16) (34) (33) (108) (89)

Assetst1 0039a 0011b 0034a 0024b 0001a 0001 0046a 0020a

(89) (22) (38) (23) (36) (13) (190) (82)

Constant 0001b 0153a 0798a 0591a 0000c 0006a 0057a 0096a

(25) (144) (470) (288) (17) (124) (292) (234)

Year FE Yes Yes Yes Yes Yes Yes Yes Yes

Firm FE No Yes No Yes No Yes No Yes

R2 0025 0089 0018 0073 0030 0147 0016 0298

N 136146 136146 136140 136140 143031 143031 142847 142847

PROFITSndashLEVERAGE PUZZLEREVISITED 1445

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

Table XIII Externally financed leverage changes in good and bad times

The sample contains nonfinancial firms listed on the annual Compustat files for the period

from 1971 to 2009 The table presents estimates from regressions of change in debt tocapitalization ratio (in columns (1) (2) (5) and (6)) and change in leverage ratio consider-ing only external financing (in columns (3) (4) (7) and (8)) The change in leverage ratio

considering only external financing is defined as

LevEF frac14Dt1 thorn d

Dt1 thorn Et1 thorn dthorn e

Dt1

Dt1 thorn Et1

where D is debt E is book value of equity d is debt issues net of retirements and e is equityissues net of repurchases The regressions include lagged leverage factors and year indicator

variables The reported t-statistics are corrected for clustering at the firm level An industryis defined as having ldquogood timesrdquo if the median firm in that industry has a market-to-bookratio that is higher than the 67th percentile of the time-series distribution of industry

median market-to-book ratios Conversely an industry is defined as having ldquobad timesrdquoif the median firm in that industry has a market-to-book ratio that is lower than the 33rdpercentile of the time-series distribution of industry median market-to-book ratios

aMean significant at the 1 level

bMean significant at the 5 level

cMean significant at the 10 level

Good times Bad times

DDthornE LevEF D

DthornE LevEF

(1) (2) (3) (4) (5) (6) (7) (8)

Profitabiltyt1 0022c 0005 0016c 0025b 0073b 0041 0007 0007

(17) (03) (18) (24) (25) (11) (04) (03)

IndLevt1 0000 0011 0001 0009 0010 0018 0001 0000

(00) (09) (01) (13) (07) (10) (02) (00)

MB

t1

0004a 0003b 0000 0001 0008a 0001 0000 0000

(40) (24) (05) (08) (28) (04) (03) (03)

Tangt1 0097a 0041 0033a 0038b 0127a 0048 0059a 0033

(45) (16) (27) (25) (49) (14) (43) (16)

Assetst1 0014a 0013b 0024a 0019a 0037a 0023c 0036a 0025a

(30) (20) (84) (48) (45) (17) (88) (43)

Constant 0030b 0036 0000 0003 0028c 0014 0021 0001

(20) (16) (00) (07) (17) (06) (11) (01)

Year FE Yes Yes Yes Yes Yes Yes Yes Yes

Firm FE No Yes No Yes No Yes No Yes

R2 0006 0088 0009 0117 0026 0251 0023 0202

N 56820 56820 58200 58200 23462 23462 23511 23511

1446 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

conclusions The traditional leverage result on profits is driven by increasesin equity that are experienced and then partially but not completely offsetby firmsTo the extent that firms are engaging in active financing they are following

the prescriptions of the trade-off theory Thus we have a coherent picture ofhow firms are responding to profits Basically more profitable firms experiencean increase in the value of their equity They respond to this by issuing debtand repurchasing equity But the magnitude of the increase in equity is largerthan the magnitudes of the active steps taken in the debt and equity marketsThis means that the firmsrsquo actual actions are as predicted by the trade-offtheory But it leaves open the question of why they do not act more stronglyThe capital structure literature has long been interested in dynamics and

capital structure adjustment paths In recent years this interest has increas-ingly resulted in explicit dynamic optimizing models such as those ofGoldstein Ju and Leland (2001) Hennessy and Whited (2005) andStrebulaev (2007) These models commonly make specialized assumptionsleading some to question their generalityWe argue that both fixed transactions costs and variable transactions costs

are plausible and realistic in models that pin down firm size The simultaneouspresence of both is important Due to fixed costs firms only rebalance inter-mittently Due to variable adjustment costs the extent of the adjustment isreduced to the point at which an extra unit to adjustment is balanced againstthe extra unit transactions costs This means that both leverage increases andleverage reductions will only go part way toward a static optimumThe key point is that with both fixed and variable transactions costs we

should expect to only see intermittent leverage adjustments When leverage isadjusted it should ldquounder adjustrdquo when compared to a frictionless modelFurther discussion of transactions costs in a capital structure setting isprovided by Leary and Roberts (2005)

9 Conclusion

The connection between corporate profits and capital structure has beenvery influential in the assessment of the trade-off theory The conventionalinterpretation of the evidence has pushed the literature away from the trade-off theory and toward much more complex dynamic models and ideas As aresult it is important to make sure that the evidence has been correctlyinterpreted We show that the problem is with the interpretation of thetest results and a failure to take into account the costs associated with re-sponding to a profit shock

PROFITSndashLEVERAGE PUZZLEREVISITED 1447

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

We document that more profitable firms really do borrow more (not less)They tend to repurchase equity They experience an increase in both thebook value of equity and the market value of equity Less profitable firmsreally do tend to reduce their debt and to issue equityFirm size and market conditions also matter Larger firms tend to be more

active in the debt markets while smaller firms tend to be relatively moreactive in the equity markets During good times there is more use ofexternal financingThe usual profitsndashleverage puzzle result is primarily driven by the increase

in equity that is experienced by the more profitable firms Accordingly thepuzzle should be restated as asking why do firms not take sufficiently largeoffsetting actions to fully undo the change in equity What limits the mag-nitudes of the typical leverage response to profit shocksIn a frictionless model the partial response appears puzzling But there is

good empirical reason to believe that rebalancing entails both fixed andvariable costs and that firm size matters The rebalancing costs can befully avoided by doing nothing Accordingly the firm must decide whetherany given shock is big enough to be worth responding to If it is then thefirm must decide how big a response is called for These technical conditionsare known as ldquovalue matchingrdquo and ldquosmooth pastingrdquo Thus optimizationimplies that some shocks will be ignored Even if the shock is not ignoredthe optimal response will only partially undo the shock The magnitude ofthe leverage response must balance the marginal cost and the marginalbenefit of an extra unit of leverage Since the marginal cost of adjustingleverage is strictly positive the adjustment toward that static leverageoptimum will only go part way This is true both for leverage increasesand for leverage reductionsOne might conjecture that the variable costs would be time varying and

that such costs might have declined over time Since we do not have a fullaccounting of all of the variable costs it is of course impossible to be sureHowever there is some evidence suggesting that at least the directunderwriting fees are remarkably stable over time10

Surely nobody ought to be surprised that more profitable firms are moreinclined to repurchase equity while unprofitable firms tend not to do so Thefact that more profitable firms typically experience an increase in the value of

10 For IPOs Jay Ritter provides important summary evidence as httpbearwarringtonufleduritterIPOs2013Underwritingpdf He finds that for medium-sized deals a remark-

ably large fraction of the deals have a 7 fee year after year Whether this is due tobehavioral factors or some sort of a coordination device is not clear The extent towhich similar intertemporal stability carries over to other forms of external financemerits further study

1448 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

equity is equally natural Were it not for the contrary claims in the literature(ie the profitsndashleverage puzzle) we would argue that nobody ought to findit surprising that more profitable firms are more inclined to issue debt Thefact that leverage adjustments typically only partially offset profit shocksshould also sound quite natural These facts all fit together easily within thetrade-off theory once we take account of both fixed and variable costs ofactively adjusting capital structure

Appendix

Table AI Market leverage regressions

The table presents estimates of the regressions of market leverage on firm characteristics

The sample comes from the annual Compustat files during the period 1971ndash2009 Financialfirms are excluded The leverage is estimated as the ratio of debt over debt plus marketequity The explanatory variables Profitabilityt1

MB

t1

Tangibilityt1 and LnethAssetsTHORNt1are described in Table 1 IndMedianLevt1 is estimated as the median market leverage of allother firms in the same industry The industry is defined at the level of the four-digit SICcode All specifications include the year fixed effects The specifications in column (2)

additionally include the firm fixed effects We report t-statistics where the standard errorsare clustered at the firm level in parentheses

aSignificant at the 1 level

Quantile regressions

(1) (2) (3) (4) (5)

Profitabilityt1 013a 013a 007a 011a 020a

(275) (262) (349) (425) (538)

IndLevt1 046a 020a 032a 058a 063a

(506) (239) (1299) (1630) (1042)MB

t1

003a 001a 001a 002a 004a

(509) (303) (743) (709) (534)

Tangibilityt1 013a 018a 013a 014a 010a

(191) (181) (762) (571) (232)

LnethAssetsTHORNt1 001a 005a 001a 001a lt001a

(112) (258) (682) (342) (29)

Constant 010a 011a 005a 005a 029a

(171) (98) (173) (118) (398)

Year FE Yes Yes Yes Yes Yes

Firm FE No Yes No No No

Regression type OLS FE 25th tile Median 75thtile

Clustered SE Yes Yes No No No

R2 030 067

Pseudo R2 016 023 0201

N 158578 158578 158578 158578 158578

PROFITSndashLEVERAGE PUZZLEREVISITED 1449

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

Table AII Profitability sorts for firms issuing large amounts

Percentage of firms

Iss Ret Iss Iss Rep Iss Iss IssD IssE

D D NetD E E NetE DampE RepE RetD

(1) (2) (3) (4) (5) (6) (7) (8) (9)

Panel A 10 cutoff to define issuing or retiring activity

Allfirm

s

Low 205 150 134 260 05 257 45 00 23

2 207 175 103 77 06 75 11 01 10

3 243 196 101 50 08 47 08 01 09

4 252 186 113 50 13 48 08 02 08

High 227 161 118 60 27 57 07 04 09

t Highfrac14Low 70 38 62 755 226 768 320 90 152

Smallfirm

s Low 230 132 162 380 05 375 82 00 32

2 228 143 165 334 04 333 65 00 26

3 211 146 147 216 03 214 38 00 23

4 190 154 120 126 04 124 15 01 20

High 194 162 114 99 09 97 12 01 15

t Highfrac14Low 42 40 68 342 20 340 169 07 131

Largefirm

s Low 231 200 82 31 06 29 08 01 04

2 231 187 67 17 06 14 03 01 02

3 234 177 77 17 09 14 03 02 03

4 218 158 75 16 17 14 02 03 02

High 218 135 99 19 41 15 03 07 02

t Highfrac14Low 19 106 36 46 135 57 39 58 23

Percentage of firms

Iss Ret Iss Iss Rep Iss Both Iss D Iss E

D D Net D E E Net E DampE Rep E Ret D

(1) (2) (3) (4) (5) (6) (7) (8) (9)

Panel B 15 cutoff to define issuing or retiring activity

Allfirm

s

Low 158 101 100 230 01 228 27 00 15

2 142 115 63 57 01 56 05 00 06

3 165 128 57 32 01 31 03 00 05

4 173 118 65 35 01 34 03 00 05

High 160 102 70 46 04 44 03 00 05

t Highfrac14Low 08 06 141 738 87 746 277 24 131

Smallfirm

s Low 195 95 132 354 01 350 56 00 24

2 185 96 131 303 01 301 44 00 18

3 160 96 104 179 00 176 18 00 15

4 134 99 78 99 01 98 08 00 12

High 133 95 73 82 01 80 06 00 10

t Highfrac14Low 80 00 95 345 12 344 149 12 52

Largefirm

s Low 145 123 42 14 01 12 03 00 01

2 137 114 32 07 01 06 01 00 01

3 142 104 35 06 01 05 00 00 01

4 130 91 36 08 03 07 01 01 01

High 134 81 48 09 06 07 01 01 01

t Highfrac14Low 19 85 17 26 54 33 34 17 11

1450 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

Table AIV Debt and equity changesmdashwithout fixed effects

The sample contains nonfinancial firms listed on the annual Compustat files for the period from 1971

to 2009 The table presents estimates from regressions of change in debt (D) change in the book

value of equity (BVE) change in the market value of equity (MVE) and net equity issuances

(EqyIss) The explanatory variables include current and lagged change in operating income before

depreciation (Profits and Profitst1) and lagged change in book value of assets (Assetst1) The

regressions include year indicator variables The reported t-statistics are corrected for clustering at the

firm level

aMean significant at the 1 level

bMean significant at the 5 level

cMean significant at the 10 level

D BVE MVE EqyIss

(1) (2) (3) (4) (5) (6) (7) (8)

Profitst1 025a 024a 023a 038a 007 066a 007a 009a

(40) (32) (36) (39) (06) (29) (42) (51)Profits 051b 082a 149a 004a

(25) (33) (49) (33)Assetst1 002 005c 024a 001b

(08) (17) (34) (23)Constant 93 176a 45 71 1552a 1704a 130a 140a

(12) (28) (07) (09) (49) (49) (45) (45)Year FE Yes Yes Yes Yes Yes Yes Yes YesR2 001 005 001 011 001 005 002 002N 162056 162056 162130 162130 157550 157550 162154 162154

Table AIII Median profitability sorts

The sample contains nonfinancial firms listed on the annual Compustat files for the period from 1971

to 2009 The table presents median debt and equity median changes in debt and equity and median

equity issuances (EqyIss) in millions of dollars for all firms sorted on profitability and for profitability

sorts within the smallest and largest firms The sorts are done annually

D D BVE BVE MVE MVE EqyIss Assets(1) (2) (3) (4) (5) (6) (7) (8)

Allfirm

s

Low 21 00 75 17 256 20 01 1922 186 02 449 18 598 19 00 11033 816 06 1274 03 1637 05 00 32244 787 02 1541 18 2488 03 00 3556High 273 00 1361 66 3300 14 00 2578

Smallfirm

s Low 05 00 13 07 140 14 06 402 06 00 25 08 114 12 01 563 09 00 32 05 85 07 00 674 12 00 39 01 68 03 00 79High 10 00 51 04 98 01 00 90

Largefirm

s Low 11485 381 10269 371 13227 202 03 333902 13051 393 12786 140 17571 169 05 392463 11554 287 13429 24 21623 305 00 370314 9963 209 15052 194 29424 308 00 36705High 7361 00 16380 613 49142 439 00 34976

PROFITSndashLEVERAGE PUZZLEREVISITED 1451

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

References

Auerbach A J (1985) Real determinants of corporate leverage in B M Friedman (ed)

Corporate Capital Structures in the United States University of Chicago Press Chicago

IllinoisBertrand M and Mullainathan S (2003) Enjoying the quiet life corporate governance and

managerial preferences Journal of Political Economy 111 1043ndash1075Booth L Varouj A Demirguc-Kunt A and Maksimovic V (2001) Capital structures in

developing countries Journal of Finance 56 87ndash130Brennan M J and Schwartz E S (1984) Optimal financial policy and firm valuation

Journal of Finance 39 593ndash607Cameron A C and Trivedi P K (2010) Microeconometrics using Stata Stata Press TexasConstantinides G M (1979) Multiperiod consumption and investment behavior with

convex transactions costs Management Science 25 1127ndash1137Constantinides G M and Richard S F (1978) Existence of optimal simple policies for

discounted-cost inventory and cash management in continuous time Operations Research

26 620ndash636DeAngelo H DeAngelo L and Whited T M (2011) Capital structure dynamics and

transitory debt Journal of Financial Economics 99 235ndash261Dixit A K (1993) The Art of Smooth Pasting Routledge London and New YorkFama E and French K R (2002) Testing trade-off and pecking order predictions about

dividends and debt Review of Financial Studies 15 1ndash33Faulkender M and Petersen M A (2006) Does the source of capital affect capital struc-

ture Review of Financial Studies 19 45ndash79Faulkender M Flannery M J Hankins K W and Smith J M (2012) Cash Flows and

Leverage Adjustments Journal of Financial Economics 103 632ndash646Fischer E Heinkel R and Zechner J (1989) Dynamic capital structure choice theory

and tests Journal of Finance 44 19ndash40Frank M Z and Goyal V K (2008) Trade-off and pecking order theories of

debt in B E Eckbo (ed) Handbook of Corporate Finance Empirical Corporate Finance

Vol 2 Handbook of Finance Series Chapter 12 ElsevierNorth-Holland Amsterdam pp

135ndash202

Frank M Z and Goyal V K (2009) Capital structure decisions which factors are reliably

important Financial Management 38 1ndash37

Garvey G T and Hanka G (1999) Capital structure and corporate control the effect of

antitakeover statutes on firm leverage Journal of Finance 54 519ndash546

Goldstein R Ju N and Leland H (2001) An EBIT-based model of dynamic capital

structure Journal of Business 74 483ndash512

Graham J R and Tucker A L (2006) Tax shelters and corporate debt policy Journal of

Financial Economics 81 563ndash594

Hackbarth D and Mauer D C (2012) Optimal priority structure capital structure and

investment Review of Financial Studies 25 747ndash796

Halling M Yu J and Zechner J (2012) Leverage dynamics over the business cycle

Working paper University of Utah University of New South Wales and Vienna

University of Economics and Business AdministrationHennessy C A and Whited T A (2005) Debt dynamics Journal of Finance 60 1129ndash1165

Huang R and Ritter J R (2009) Testing theories of capital structure and estimating the

speed of adjustment Journal of Financial and Qualitative Analysis 44 237ndash271

1452 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

Hovakimian A Opler T and Titman S (2001) The debt-equity choice Journal of

Financial and Quantitative Analysis 36 1ndash24Jung K Kim Y and Stulz R M (1996) Timing investment opportunities managerial

discretion and the security issue decisions Journal of Financial Economics 42 159ndash185Kane A Marcus A J and McDonald R L (1984) How big is the tax advantage to

debt Journal of Finance 39 841ndash853Kayhan A and Titman S (2007) Firmsrsquo histories and their capital structures Journal of

Financial Economics 83 1ndash32Leary M T (2009) Bank loan supply lender choice and corporate capital structure

Journal of Finance 64 1143ndash1185Leary M T and Roberts M R (2005) Do firms rebalance their capital structures

Journal of Finance 60 2575ndash2619Leland H E (1994) Corporate debt value bond covenants and optimal capital structure

Journal of finance 49 1213ndash1252Lemmon M L Roberts M R and Zender J F (2008) Back to the beginning persist-

ence and the cross-section of corporate capital structure Journal of Finance 63 1575ndash1608Long M S and Malitz I B (1985) The investment-financing nexus some empirical

evidence Midland Corporate Finance Journal 3 53ndash59Mackie-Mason J K (1990) Do taxes affect corporate financing decisions Journal of

Finance 45 1471ndash1493Morellec E Nikolov B and Schurhoff N (2012) Corporate governance and capital

structure dynamics Journal of Finance 67 803ndash848Myers S C (1984) The capital structure puzzle Journal of Finance 39 575ndash592Myers S C (1993) Still searching for optimal capital structure Journal of Applied

Corporate Finance 61 4ndash14Rajan R and Zingales L (1995) What do we know about capital structure some evidence

from international data Journal of Finance 50 1421ndash1460Strebulaev I A (2007) Do tests of capital structure theory mean what they say Journal of

Finance 62 1747ndash1787

Strebulaev I A and Whited T M (2012) Dynamic Models and Structural Estimation in

Corporate Finance Now Publisher Boston and Delft

Titman S and Wessels R (1988) The determinants of capital structure choice Journal of

Finance 43 1ndash21

Welch I (2004) Capital structure and stock returns Journal of Political Economy 112

106ndash131

Welch I (2011) Two common problems in capital structure research the financial-debt-

to-asset ratio and issuing activity versus leverage changes International Review of Finance

11 1ndash17

PROFITSndashLEVERAGE PUZZLEREVISITED 1453

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

5 Financing Activity

51 FREQUENCY OF DEBT AND EQUITY ISSUES

Table IV reports the percentage of firms issuing or repurchasing debt orequity for annual sorts based on lagged profitability In Panel A weemploy the conventional 5 cutoff rule to exclude minor fluctuations anddefine firms ldquoissuing debtrdquo as those that issue debt (both short-term andlong-term) in excess of 5 of the value of their assets Other decisions aresimilarly defined using a 5 cutoff Appendix Table AII examines the dis-tributional properties across profitability (and size) quintiles for ldquolargerdquo is-suances by increasing the 5 cutoff to higher levels such as 10 and 15How does security issuance behavior vary with firm profitability To

answer this question Table IV sorts firms by profitability and reports the

Table III Leverage and profitability for active and passive firms

The table presents estimates of leverage on firm characteristics for the subsamples of firms

that are active in either debt or equity markets Active firms are defined as those that issueeither net debt in excess of 5 of the value of assets or issue net equity in excess of 5 ofthe value of their assets The passive firms do not issue significant (more than 5) of either

debt or equity The sample comes from the annual Compustat files during the period 1971ndash2009 Financial firms are excluded The variables are described in Tables I and II Wereport t-statistics where the standard errors are clustered at the firm level in parentheses

aSignificant at the 1 level

Active Passive

(1) (2) (3) (4)

Profitabilityt1 006a 008a 022a 017a

(126) (148) (254) (188)

IndLevt1 044a 019a 041a 013a

(424) (149) (326) (127)MB

t1

002a 001a 003a 001a

(362) (196) (327) (172)

Tangibilityt1 010a 014a 014a 019a

(146) (118) (145) (130)

LnethAssetsTHORNt1 001a 004a 001a 004a

(121) (179) (147) (169)

Constant 012a 008a 003a 007a

(120) (57) (45) (42)

Year FE Yes Yes Yes Yes

Firm FE No Yes No Yes

R2 034 064 033 075

N 43395 43395 58943 58943

PROFITSndashLEVERAGE PUZZLEREVISITED 1429

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

percentage of firms issuing or repurchasing debt or equity The table alsoreports test statistics from tests that examine whether the percentage differ-ences between the high- and low-profit firms are significantly different fromeach other The results in Column (1) show that the likelihood of issuing debt

Table IV Profitability sorts for debt and equity issuers

The sample contains nonfinancial firms listed on the annual Compustat files for the period

from 1971 to 2009 The table presents the percentage of firms issuing and retiring (orrepurchasing) debt and equity The firms are classified as ldquoissuing debtrdquo if they issuedebt in excess of 5 of the value of their assets as ldquoissuing equityrdquo if they issue equity

in excess of 5 of the value of their assets as ldquoretiring debtrdquo if they retire debt in excess of5 of the value of their assets and as ldquorepurchasing equityrdquo if they repurchase equity inexcess of 5 of the value of their assets In addition we also report net debt issuers which

are firms that issue net debt over 5 of the value of their assets and net equity issuerswhich are firms that issue net equity in excess of 5 of the value of their assets Weannually sort firms on lagged profitability and report the percentage of firms in each ofthese categories The bottom part of each panel reports the percentage of firms issuing or

retiring securities by profitability within the smallest and largest asset quintiles The tablealso reports results from t-test for the difference in percentage of firms issuing or retiringsecurities in the lowest and the highest profit quintile

Percentage of firms

Iss Ret Iss Iss Rep Iss Both IssD IssE

D D NetD E E NetE DampE RepE RetD

(1) (2) (3) (4) (5) (6) (7) (8) (9)

Allfirm

s

Low 276 245 186 301 13 297 71 01 38

2 319 299 181 117 21 112 27 02 20

3 385 344 194 86 29 80 21 05 15

4 386 330 213 81 43 75 20 07 14

High 329 273 198 92 82 81 17 13 14

t Highfrac14Low 150 84 39 723 414 763 358 176 200

Smallfirm

s Low 275 204 201 411 11 405 109 01 47

2 291 233 213 379 10 378 99 01 45

3 281 246 204 265 14 261 66 01 41

4 275 272 186 162 15 159 34 02 35

High 275 281 178 136 27 131 25 02 26

t Highfrac14Low 01 84 28 316 55 318 171 18 55

Largefirm

s

Low 414 363 177 87 22 79 28 03 11

2 449 379 180 62 30 57 17 07 09

3 439 360 186 56 41 48 16 09 07

4 403 320 194 46 67 39 13 12 06

High 374 264 207 50 129 35 10 28 05

t Highfrac14Low 50 130 45 89 244 116 77 120 47

1430 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

is mostly independent of firm profitability In the aggregate there is a smalldifference but it is sensitive to sorting on firm size Similarly in Column (2)there is only a weak relation between the profitability and the likelihood thatthe firm retires debt To a first approximation the fraction of firms issuing orrepurchasing debt is largely independent of profits Large firms are somewhatmore likely than small firms to issue debt and also to retire debtUnlike debt issues the probability of issuing equity is strongly related to

profitability as can be seen in Column (4) Low-profit firms are much morelikely to issue equity than are high-profit firms In the lowest profit quintile301 of the firms issued equity In the highest profit quintile it was only92 The lower panel reports separate sorts for small and large firms Similarpatterns are found for both groups of firms that is the likelihood of equityissuance declines as profitability increases The differences between the quin-tiles are quite stark for the group of small firms All quintiles of small firmsissue equity much more frequently than the corresponding large firm quintilesHigh-profit firms are much more likely to repurchase equity as shown in

Column (5) In the overall sample only 13 of low-profit firmsrepurchased equity compared with 82 of the high-profit firms Whenwe separate the small and large firms again a difference emerges Ingeneral small firms do not repurchase much equity while a greaterfraction of the large firms do soGiven these facts it is natural that low profitability firms are much more

likely to be net issuers of equity than are high profitability firms as shown inColumn (6) Among the firms in the lowest quintiles of profits roughly 30of firms issue net equity that exceeds 5 percent of their assets In contrastfor the most profitable quintile only about 8 issue net equityFirm size is an important variable in the recent literaturemdashit is often used

as a proxy for access to capital markets (as in Faulkender and Petersen(2006) and Leary (2009)) Small firms are bank-dependent risky and infor-mationally opaque They have restricted access to public debt markets andconsequently face more severe supply constraints in their ability to issuedebt Thus we expect small firms to be more sluggish in adjusting theirdebt and equity in response to shocks to profitability Large firms incontrast have much easier access to public debt markets and they facefewer obstacles in accessing securities marketsTo examine how size and profitability interact the bottom part of Table IV

examines profitability sorts for the smallest and largest firms We first sortfirms annually by firm size and then within each size quintile we sort them onprofitability We do these two-way sorts to ensure that we have similarnumber of firms in profitability subgroups for both small and large firmsamples Unconditionally small firms are generally less profitable while

PROFITSndashLEVERAGE PUZZLEREVISITED 1431

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

large firms are relatively more profitable The two-way sorts allow us to havea uniform distribution of firms in various profitability quintiles for both smalland large firm samples The bottom part of Table IV reports results for thesmallest and the largest size quintiles Among the small firms there is littlerelation between profitability and the likelihood of issuing debt Howeverlarger firms exhibit a small increase in the likelihood of net issuance of debtwith increasing profitabilityIn contrast to debt the effects of profitability on equity issuance and

repurchases are much more consistent across size quintiles As profitabilityincreases firms are generally less likely to issue equity and more likely torepurchase it Across the two extreme size groups we note a striking differ-ence between the proportion of firms issuing equity and those issuing debtAmong low-profit small firms almost 41 are equity issuers whereas onlyabout 9 of low-profit large firms are equity issuers However regardless ofsize we note a monotonic reduction in the likelihood of issuing equity asprofitability increasesColumn (7) reports the fraction of firms that issue both debt and equity in

excess of 5 of assets both as a function of profits in the upper panel andas a function of firm size and profits in the lower panel In the upper panelwe see that low-profit firms are much more likely to issue both debt andequity (71) while the high-profit firms are less likely to do so (17) Thelower panel shows that small firms are much more likely to be issuers of bothin contrast to large firms In both firm size categories the profit pattern is thesame with low profits being more frequent issuers of both simultaneouslyImportantly Column (8) shows that the likelihood of issuing debt and

simultaneously repurchasing equity increases with profitability Converselyin Column (9) we find that the likelihood of doing the reverse that is issuingequity and retiring debt declines with profitability The effects of sorting onfirm size and firm profitability mirror those reported for all firms Overallthe results show that low-profit firms are less likely to issue debt and repur-chase equity they are instead more likely to issue equity and retire debtFirms with high profitability exhibit the reverse patternIn Appendix Table AII we present results from profitability (and size)

quintiles for large issuances defined using the 10 cutoff (in Panel A) andthe 15 cutoff (in Panel B) The results from these panels are largely con-sistent with those described above

52 MAGNITUDES OF ISSUANCES AND EQUITY CHANGES

In the previous section we considered the probability of having a nontriviallevel of debt or equity activity The next question is how large are the dollar

1432 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

values involved In Table V we sort firms according to profits and thentabulate the levels and changes in both debt and equity We do this firstfor all firms and then for small versus large firmsFor the analysis on all firms we observe in Column (1) that debt peaks at

the middle of the distribution This is because firms with medium profitabil-ity are also the largest as seen in Column (8) which reports average assetvalues for different profitability quintiles When we sort first by firm size andthen examine firms with different levels of profitability within size quintileswe find that for small firms debt is roughly independent of profits while forlarge firms less profitable firms have a higher level of debtAs expected Columns (3) and (5) show that more profitable firms have

higher equity values Columns (2) and (7) consider the relationship betweenissuances of debt and equity and firm profitability Debt issuances are sig-nificantly larger for more profitable firms In contrast equity issuances arelarger for low-profit firms Small low-profit firms issue substantially lessequity compared with small high-profit firms Among large firms the low-profit ones tend to issue equity while the high-profit ones tend to repurchaseequity The fact that more profitable firms issue debt and repurchase equitywhile the least profitable firms retire debt and issue equity is consistent withthe predicted relation between profitability and financing decisions under thetrade-off theoryWe also examine two-way sorts by size and profitability and report results

for the smallest and largest quintile of firms For the smallest quintile offirms the change in debt is largely unrelated to profits But for large firmsthere is a positive relation between profits and debt issuances High-profitfirms have a big positive change in debt Low-profit large firms have anegative change in debtColumns (4) and (6) provide an explanation of why the leverage ratio

regression results contradict those from the basic profitability sorts pre-sented here As we can see profitability indirectly affects leverage ratiosby increasing equity values Changes in both the book value of equity andthe market value of equity are positive and large for highly profitable firmsIn contrast these changes are negative for less profitable firms This evidenceillustrates an important issue concerning the use of leverage ratios Suchratios are often interpreted as essentially reflecting the use of debt by thefirm This interpretation while common is empirically misleadingFor the typical firm the change in the value of equity is larger than the

change in debt For example in the third profit quintile for large firmscolumn (7) the mean equity issue is just $9 million but the change in themarket value of equity is $242 million At the same time the mean change indebt is $27 million This suggests that a fair bit of the observed variation in

PROFITSndashLEVERAGE PUZZLEREVISITED 1433

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

the leverage ratios is primarily driven by the changes in the market value ofequity in the denominator rather than by the changes in debt in the numer-ator Since equity issues are often small this implies that the variation in theleverage ratio is primarily driven by internal operations rather than byexternal financing actions This again points to the fact that leverageratios can provide a misleading account of actual patterns in the dataThere is always a potential concern that the averages may be misleading

due to the impact of outliers To address this concern in AppendixTable AIII we present median values of the profitability sorts Theseresults generally reinforce the findings in Table VTable VI again considers the magnitudes of financing activity but this

time the issues are scaled by a measure of firm size That way it is easier to

Table V Magnitude of financing activity

The sample contains nonfinancial firms listed on the annual Compustat files for the period

from 1971 to 2009 The table presents average debt and equity changes in debt and equityand equity issuances (in millions of dollars) for firms sorted on profitability within sizeclasses The table reports information for all firms sorted on profitability and for profit-

ability sorts within the smallest and largest firms The sorts are done annually The tablealso reports results from t-test for the difference in percentage of firms issuing or retiringsecurities in the lowest and the highest profit quintile

D D BVE BVE MVE MVE EqyIss Assets

(1) (2) (3) (4) (5) (6) (7) (8)

Allfirm

s

Low 83 5 78 10 120 5 7 234

2 605 13 509 24 721 8 17 1691

3 1227 9 1123 22 1834 3 9 3551

4 833 23 1166 7 2408 74 9 2994

High 570 53 1171 71 3817 68 36 2581

t Highfrac14Low 31 9 36 9 38 2 16 36

Smallfirm

s Low 25 04 42 01 332 28 33 97

2 29 06 55 02 278 16 21 116

3 31 07 63 01 212 05 12 127

4 30 06 70 05 179 10 08 136

High 24 06 84 14 234 17 07 142

t Highfrac14Low 04 15 153 85 73 11 201 107

Largefirm

s Low 3929 55 2990 152 4671 11 63 10846

2 3826 19 3178 95 5122 66 11 10721

3 3074 27 3498 41 6173 242 9 9697

4 2623 62 3915 45 8473 94 57 9855

High 2180 193 4642 261 14924 202 148 10171

t Highfrac14Low 94 64 108 87 246 12 169 16

1434 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

see whether large firms behave disproportionately when compared withsmall firms Column (1) shows that low-profit firms have more debtas reported in many previous studies In Columns (4) and (6) we seethat low-profit firms have a major decline in both the book value ofequity and the market value of equity The high-profit firms have thereverse experienceMore interesting are the active decisions of the firms In Column (2) we see

that in contrast to the usual impression the low-profit firms are reducingdebt while the high-profit firms are increasing it In Column (7) we see thatthe low-profit firms are also reducing leverage by issuing equity The high-profit firms issue little equity In the lower panel similar results are foundwhen we separate small and large firms Equity issuing is particularly strong

Table VI Magnitude of financing activitymdashscaled levels and changes

The sample contains nonfinancial firms listed on the annual Compustat files for the period

from 1971 to 2009 The table presents average debt and equity changes in debt and equityand equity issuances scaled by assets for firms sorted on profitability within size classes Thetable reports information for all firms sorted on profitability and for profitability sorts

within the smallest and largest firms The sorts are done annually The table also reportsresults from t-test for the difference in percentage of firms issuing or retiring securities inthe lowest and the highest profit quintile

D D BVE BVE MVE MVE EqyIss

(1) (2) (3) (4) (5) (6) (7)

Allfirm

s

Low 386 122 12 871 3427 1868 217

2 285 58 407 119 940 157 28

3 293 12 425 14 742 03 12

4 269 26 461 02 955 27 10

High 192 15 551 47 1675 74 07

t Highfrac14Low 35 49 50 47 127 42 157

Smallfirm

s Low 545 154 841 2145 9688 9128 696

2 684 49 712 1915 5014 2318 288

3 331 253 141 969 3063 954 129

4 267 181 357 175 1673 221 55

High 194 05 543 26 1758 147 41

t Highfrac14Low 49 17 51 36 88 27 99

Largefirm

s Low 350 30 306 58 511 55 09

2 346 14 335 13 523 02 04

3 353 165 332 71 711 29 01

4 280 01 414 04 909 09 05

High 227 15 477 19 1643 27 14

t Highfrac14Low 371 112 453 181 570 57 221

PROFITSndashLEVERAGE PUZZLEREVISITED 1435

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

among the small low-profit firms that are also experiencing sharp reductionsin the value of equity In column (2) we see that both for the large and for thesmall firms net debt increases are primarily found in the highest quintile offirm profitability

6 Debt and Equity Regressions

So far we have demonstrated that in our data the conventional leverageregressions have the usual signs We then demonstrated that in simple sortsof the data strikingly different active financing patterns emerge The nexttask is to reconcile these differencesTable VII presents simple regressions predicting changes in debt book

equity and market equity as well as equity issuances Consistent with thesorts we control for changes in firm size and we also include firm and yeardummies There may be a concern about the speed of a firmrsquos responsesAccordingly we consider the lagged change in profits as the main case butwe also include the contemporaneous change in profits Whether the contem-poraneous change in profits is included or not does not in fact matter Wereport panel-robust standard errors adjusted for clustering at the firm levelThe results in Table VII support the findings in the sorts Firms that have

an increase in profits in 1 year increase debt both in the same year and in thenext year Firms that have an increase in profits experience an increase inboth the book value of equity and the market value of equity Importantlyhowever firms that experience and increase in profits reduce their equityissues All these effects are statistically significant (except for the marketvalue of equity in the column 5 specification) and seem intuitively reason-able In Appendix Table AIV similar regressions to those presented aboveare reported but without firm-fixed effects We find that empirically verylittle changes whether we include firm fixed effects or leave them out Thusthe targeting behavior to the extent that it happens does not account forwhat we are reportingThe next step is to bring the sorts and the simple regressions together into

a common setting We therefore include factors in addition to profitability(i) median industry leverage (ii) the market-to-book assets ratio (iii) tangi-bility of assets and (iv) firm size (measured by the log of assets) Rajan andZingales (1995) show that these factors are related to leverage in G7countries A number of studies have used these factors to estimateleverage targets Frank and Goyal (2009) show that these factors arerobustly related to leverage in the USA In these regressions we use indica-tors for the quintile that the firm is in for each factor

1436 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

Table VIII reports these results In Columns (1) and (2) we examine debtissues and again show that profits positively affect debt issuances The effectsare large and statistically significant at the 1 level Columns (3) and (4)which examine book equity show a strong effect of profits This effect is onlyslightly reduced by the inclusion of other factors Columns (5) and (6) illus-trate that the impact of profits on the change in market equity is also robustto the inclusion of the conventional factors Columns (7) and (8) examineequity issuances Here we again find that profits have a significant negativeimpact on equity issuances As before the results are robust to the inclusionof the conventional factorsIn unreported tables we estimate debt and equity changes using quantile

regressions In terms of the signs and significance of the profitability variable

Table VII Debt and equity changes

The sample contains nonfinancial firms listed on the annual Compustat files for the period

from 1971 to 2009 The table presents fixed effect estimates from regressions of changes indebt (D) changes in the book value of equity (BVE) changes in the market value ofequity (MVE) and net equity issuances (EqyIss) The explanatory variables include

current and lagged change in operating income before depreciation (Profits andProfitst1) and lagged change in the book value of assets (Assetst1) In addition tofixed firm effects the regressions include year indicator variables The reported t-statistics

are corrected for clustering at the firm level

aMean significant at the 1 level

bMean significant at the 5 level

cMean significant at the 10 level

D BVE MVE EqyIss

(1) (2) (3) (4) (5) (6) (7) (8)

Profitst1 022a 025a 015b 035a 004 060b 005b 007a

(33) (31) (25) (35) (03) (25) (25) (53)

Profits 050b 078a 140a 003a

(24) (32) (48) (34)

Assetst1 000 007b 025a 001b

(00) (21) (35) (25)

Constant 167 283a 34 51 1145a 1182a 605a 595a

(14) (30) (04) (05) (32) (29) (69) (74)

Year FE Yes Yes Yes Yes Yes Yes Yes Yes

Firm FE Yes Yes Yes Yes Yes Yes Yes Yes

R2 001 004 001 010 001 005 002 003

N 162056 162056 162130 162130 157550 157550 162154 162154

PROFITSndashLEVERAGE PUZZLEREVISITED 1437

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

the results are qualitatively similar to the OLS results reported in Table VIIIProfitability positively affects debt issuances and negatively affects equityissuances across various quantiles Overall the results in Table VIII showthat there is nothing that is special about using sorts or regressions to explaindebt and equity issues In either case we find that more profitable firms tendto increase their debt experience an increase in the value of equity andrepurchase shares Thus the control factors are not responsible for theusual rejection of the trade-off theory It would appear that what mattersis the size of the issuing activity relative to the organic increases in equityvalue due to profits In other words what matters in the leverage regressionseems to be coming from the presence of E in D

DthornE and not from D

Table VIII Debt and equity issuances profitability and leverage factor quintiles

The sample contains nonfinancial firms listed on the annual Compustat files for the period

from 1971 to 2009 The table presents estimates from regressions of change in debt (D)change in the book value of equity (BVE) change in the market value of equity (MVE)and net equity issuances (EquityIss) The explanatory variables include lagged profitability

quintiles and lagged leverage factor quintiles In addition to fixed firm effects the regres-sions include year indicator variables The reported t-statistics are corrected for clustering atthe firm level

aMean significant at the 1 level

bMean significant at the 5 level

cMean significant at the 10 level

D BVE MVE EqyIss

(1) (2) (3) (4) (5) (6) (7) (8)

ProfitabilityQint 214a 167a 283a 175a 80 429a 53a 55a

(70) (54) (66) (59) (08) (48) (55) (62)

IndLevQint 104a 43 99 20

(29) (15) (10) (10)

Ln(Assets)Qint52 17 1026a 53a

(16) (05) (77) (35)MB

Qint158a 372a 1155a 09

(45) (56) (106) (06)

TangibilityQint 05 177 436c 64a

(01) (15) (18) (31)

Constant 167 96 656b 235 4613a 8610a 113 570a

(08) (04) (26) (04) (61) (58) (11) (31)

Year FE Yes Yes Yes Yes Yes Yes Yes Yes

Firm FE Yes Yes Yes Yes Yes Yes Yes Yes

R2 0002 0002 0003 0004 0013 0014 0013 0014

N 155587 155037 155587 155037 155587 155037 155587 155037

1438 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

7 Are Financial Market Conditions Important

According to the trade-off theory capital structure is determined by con-sidering the impact of costs and benefits of debt The time variation inthe costs and benefits of debt imply that issuance decisions would alsovary over time Furthermore market conditions affect the transactioncosts of adjusting Adverse selection is a greater problem in a cold marketthan in a hot one Accordingly it may be easier to adjust in good marketconditions Related ideas have been developed by Huang and Ritter(2009) Faulkender et al (2012) and Halling Yu and Zechner (2012)Faulkender Flannery Hankins and Smithrsquos paper examines cross-sectionalvariation in adjustment speeds including market valuations andhow they may affect adjustment speeds It is natural to think that ingood market conditions it will be less costly to issue both debt and equityand adjusting leverage to respond to profitability shocks would be mucheasierTo test the importance of market conditions we require a definition of

good times and bad times Our empirical strategy is to estimate good timesversus bad times at the four-digit industry level We define an industry ashaving ldquogood timesrdquo if the median firm in that industry has a market-to-book ratio that is higher than the 67th percentile of the time-series distribu-tion of the industry median market-to-book ratios Conversely an industryis defined as having ldquobad timesrdquo if the median firm in the industry has amarket-to-book ratio that is lower than the 33rd percentile of the time-seriesdistribution of the industry median market-to-book ratiosPanel A of Table IX tabulates issuance activity for profitability sorts in

both good times and bad times As might be expected active debt and equityissuances are larger during good times There is also somewhat more activeswaps between debt and equity during good timesPanel B reports the issuance activity for the smallest and the largest firms

during the good times and bad times Within each size quintile firms aresorted on profitability This is a three-way sort As before we find that thereis much more active use of external markets during good times Small low-profit firms are more likely to issue equity in good times than in bad timesDebt issuances are significantly higher in good times Large high-profitfirms are significantly more likely to issue debt and repurchase equity ingood times than in bad timesPanel C reports the magnitudes of the financial variables rather than the

frequencies During bad times less profitable large firms retire substantialamounts of debt and they show a tendency to issue equity Small firms donot seem to engage in similar debt reduction activities However like the

PROFITSndashLEVERAGE PUZZLEREVISITED 1439

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

Table IX Debt and equity issues in good and bad times

The table reports the frequency and magnitude of financing activity for sorts on profits for subsamples

of firms in good and bad times An industry is defined as having ldquogood timesrdquo if the median firm in

that industry has a market-to-book ratio that is higher than the 67th percentile of the time-series

distribution of industry median market-to-book ratios Conversely an industry is defined as having

bad times if the median firm in that industry has a market-to-book ratio that is lower than the 33rd

percentile of the time-series distribution of industry median market-to-book ratios Panels A and B

report the percentage of firms (i) issuing net debt in excess of 5 of the value of their assets (ii) issuing

net equity in excess of 5 of the value of their assets (iii) issuing debt and repurchasing equity both in

excess of 5 of the value of their assets and (iv) issuing equity and retiring debt both in excess of 5

of the value of their assets Panel C tabulates the mean debt and equity levels and the mean changes in

debt and equity issuances The sample contains nonfinancial firms listed on the annual Compustat files

for the period from 1971 to 2009

Iss Iss Iss D Iss ENet D Net E Rep E Ret D(1) (2) (3) (4)

Panel A sorts on profitability

Badtimes

Low 145 131 01 202 170 49 02 083 173 46 04 074 182 47 03 10

High 177 51 06 10t Highfrac14Low 41 136 33 01

Goodtimes

Low 206 381 01 462 197 148 02 233 212 103 04 184 234 92 10 18

High 216 92 16 14t Highfrac14Low 23 658 138 23

Panel B sorts on firm size and profitability

Badtimes

Smallfirm

s Low 151 318 00 312 182 276 02 253 151 140 02 294 183 73 00 19

High 177 80 00 11t Highfrac14Low 10 90 NA 20

Largefirm

s Low 172 54 01 042 148 39 04 053 150 43 04 044 156 33 03 06

High 147 28 05 03t Highfrac14Low 17 32 20 02

Goodtimes Smallfirm

s Low 213 448 01 542 230 443 01 473 222 325 00 474 203 197 02 38

High 190 153 02 31t Highfrac14Low 19 230 10 39

Largefirm

s Low 203 96 03 122 208 66 09 133 232 54 16 064 215 39 19 05

High 234 36 38 04t Highfrac14Low 29 99 92 38

(continued)

1440 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

larger firms there is a tendency for small firms to issue more equitymdashat leastto the extent possibleDuring good times large profitable firms raise significant amounts of debt

as they experience an increase in the value of their equity Such firms alsoengage in share repurchases Debt issues by smaller firms are much lessaffected by their own profitability during good times and their change indebt is more or less independent of their profits However equity issuanceis primarily found among the low-profit small firms during good timesTable IX shows that empirically profitability has a time-varying impacton leverage choicesIn Table X the basic regressions for changes in debt and for equity

issuance are presented for both good times and bad times Comparinggood times with bad times we see that the effects of profits are muchstronger in good times In bad times the impacts are rather weak statisticallyThis difference is interesting and deserves further study However for ourpurposes recall that the static trade-off theory implies time-varying capital

Table IX Continued

D D BVE BVE MVE MVE EqyIss Assets

(1) (2) (3) (4) (5) (6) (7) (8)

Panel C magnitude of financing activity

Badtimes Smallfirm

s Low 40 03 58 14 305 06 22 140

2 44 05 82 07 240 67 17 180

3 45 04 94 02 176 18 05 187

4 43 09 98 02 137 04 03 188

High 36 10 114 13 209 16 04 199

Largefirm

s Low 3365 73 2523 169 2304 163 94 8972

2 3031 95 2983 35 3284 278 56 9068

3 2930 5 3360 14 3947 308 57 9225

4 2281 6 3817 19 5213 442 7 8902

High 1790 62 4767 13 8799 58 31 9573

Goodtimes

Smallfirm

s Low 24 06 38 05 339 23 39 90

2 26 08 48 01 301 04 25 104

3 30 09 60 04 246 03 17 125

4 28 07 66 08 206 18 12 130

High 23 06 85 16 261 13 08 141

Largefirm

s Low 4532 36 3477 103 6557 260 53 12465

2 4580 58 3527 139 6911 421 23 12522

3 2952 111 3485 2 7586 279 32 9535

4 2752 124 3939 72 10686 33 105 10226

High 2429 228 4819 377 18565 268 222 10830

PROFITSndashLEVERAGE PUZZLEREVISITED 1441

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

structure choices even when the target is time invariant Empirically we doobserve time-varying choices

8 Is Scaling Important

In the preceding section we provided results for unscaled debt and equityissues This seems appropriate to us However it involves making twochanges to the standard regression We would like to determine if both

Table X Debt and equity changes in good and bad timesmdashregressions

The sample contains nonfinancial firms listed on the annual Compustat files for the period

from 1971 to 2009 The table presents estimates from regressions of change in debt (D)and net equity issuances (EqyIss) in good and bad times The explanatory variables includechange in profits lagged change in profits and lagged change in assets The regressions

include both firm and year fixed effects An industry is defined as having ldquogood timesrdquo ifthe median firm in that industry has a market-to-book ratio that is higher than the 67thpercentile of the time-series distribution of industry median market-to-book ratios

Conversely an industry is defined as having bad times if the median firm in thatindustry has a market-to-book ratio that is lower than the 33rd percentile of the time-series distribution of industry median market-to-book ratios The reported t-statistics arecorrected for clustering at the firm level

aMean significant at the 1 level

bMean significant at the 5 level

cMean significant at the 10 level

Good times Bad times

D EqyIss D EqyIss

(1) (2) (3) (4) (5) (6) (7) (8)

Profits 087a 091a 006b 009a 060 065 003 003b

(30) (29) (20) (28) (15) (16) (16) (20)

Profitst1 029c 015a 034c 001

(19) (37) (17) (04)

Assetst1 000 002b 000 001

(02) (22) (00) (07)

Constant 1337 3989 8691a 10001a 1351 1851 4099a 4055a

(03) (09) (55) (37) (09) (13) (80) (78)

Year FE Yes Yes Yes Yes Yes Yes Yes Yes

Firm FE Yes Yes Yes Yes Yes Yes Yes Yes

R2 0071 0080 0018 0045 0178 0212 0034 0041

N 74827 68729 74891 68787 26965 26013 26971 26018

1442 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

changes are crucial Accordingly in this section we take a step back toconsider prior literature in which variables are scaled The questions arehow does scaling affect the results and whyTable XI examines alternative scalings Columns (1) and (2) scale debt and

equity issuances by the total firm net issuance Columns (3) and (4) scaledebt and equity issuances by the sum of debt and book equity Columns (5)and (6) consider the change in debt scaled by the sum of debt and bookequity Columns (7) and (8) consider the change in leverage due to externalfinancing Since we are examining changes we use the changes in thestandard regression factors as controls The primary interest is in the coef-ficient on the change in profitsWe find that a change in profits is associated with an increase in debt

issues But the sign reverses when we scale the debt issues by the sum of debtplus book equity This is telling What is really driving the results is thechange in equity in the denominator not the change in debt in the numer-ator This result is further substantiated in Columns (5) and (6) where thechange in debt to capital ratio is used instead of debt issuanceTo further explore the impact of active issuance relative to passive

experiencing of changes we decompose leverage by separating out theeffect of debt and equity issues from those due to changes in retainedearnings Thus we define leverage change due to external financing as

LevEF frac14Dt1 thorn d

Dt1 thorn Et1 thorn dthorn e

Dt1

Dt1 thorn Et1 eth1THORN

where D is debt E is book value of equity d is debt issues net of retirementsand e is equity issues net of repurchases Garvey and Hanka (1999) use asimilar measure in their study This measure directly captures the netleverage effect of debt and equity issuancesColumns (7) and (8) report results from regressions of the net leverage

effect of debt and equity issues on changes in profitability after controllingfor other leverage factors in differences and year indicator variables Theeffect of profitability on leverage change due to external financing is positiveand statistically significant This comparison suggests that changes in totaldebt and equity are only partly a result of debt and equity issuance decisionsOther balance sheet adjustments complicate the inferences from leverageratio regressionsFor most of our analysis we follow the literature in treating debt as homo-

genous It is of course potentially interesting to ask what happens when wedistinguish between short-term debt and long-term debt Which part of thefirmrsquos debt maturity structure drives the under-adjustment Table XII

PROFITSndashLEVERAGE PUZZLEREVISITED 1443

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

provides results from such decompositions Alternative scalings areconsidered The main results of interest are that the distinction betweenshort- and long-term debt is important Short-term debt issuance is largelyindependent of changes in profits The negative sign on profit change stemsfrom the action of long-term debt This seems intuitively reasonable

Table XI Scaled issuances

The sample contains nonfinancial firms listed on the annual Compustat files for the period

from 1971 to 2009 The table presents estimates from regressions of debt issuance to totalissuance (in columns (1) and (2)) debt issuance to total capitalization (in columns (3) and(4)) change in debt to capitalization ratio (in columns (5) and (6)) and change in leverage

ratio considering only external financing (in columns (7) and (8)) The change in leverageratio considering only external financing is defined as

LevEF frac14Dt1 thorn d

Dt1 thorn Et1 thorn dthorn e

Dt1

Dt1 thorn Et1

where D is debt E is book value of equity d is debt issues net of retirements and e is equity

issues net of repurchases The regressions include lagged leverage factors and year indicatorvariables Some of the specifications include firm fixed effects The reported t-statistics arecorrected for clustering at the firm level

aMean significant at the 1 level

bMean significant at the 5 level

cMean significant at the 10 level

DebtIssTotIss

DebtIssethDthornETHORN D

DthornE LevEF

(1) (2) (3) (4) (5) (6) (7) (8)

Profitt1 004a 000 002a 001c 003a 002 001b 002a

(35) (03) (34) (18) (35) (14) (23) (26)

IndLevt1 004c 001 001 001c 000 000 000 000

(19) (06) (15) (19) (02) (04) (03) (05)

Market=bookt1 001a 001a 000a 000a 000a 000a 000 000c

(47) (85) (68) (60) (55) (36) (05) (18)

Tangt1 004 005 013a 011a 011a 006a 003a 003a

(12) (15) (144) (112) (79) (38) (44) (42)

Assetst1 008a 004a 006a 004a 002a 001a 003a 002a

(97) (45) (283) (161) (53) (35) (147) (101)

Constant 079a 074a 002a 003a 000 000 000b 002a

(497) (442) (91) (111) (00) (05) (22) (93)

Year FE Yes Yes Yes Yes Yes Yes Yes Yes

Firm FE No Yes No Yes No Yes No Yes

R2 0015 0098 0022 0076 0009 0070 0012 0086

N 133905 133905 140552 140552 138229 138229 140572 140572

1444 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

inasmuch as short-term debt is probably more associated with day-to-dayoperations rather than longer term planning considerationsIn Table XIII we examine change in leverage ( D

DthornE) and change inleverage due to external financing (LEVEF) during good and bad timesseparately The traditional negative effect of a change in profits on achange in leverage is found both in good times and in bad timesHowever when we decompose the impact in order to isolate the externalcomponent as before the result disappears In good times the externalfinancing component of a change in profits is weak but still positiveDuring bad times it is not statistically significant This implies that distin-guishing between good and bad times does not fundamentally alter the basic

Table XII Short- and long-term debt issuances

The sample contains nonfinancial firms listed on the annual Compustat files for the period

from 1971 to 2009 The table presents estimates from regressions of short-term debtissuance to total issuance (in columns (1) and (2)) long-term debt issuance to totalissuance (in columns (3) and (4)) short-term debt issuance to total capitalization (in

columns (5) and (6)) and long-term debt to total capitalization (in columns (7) and (8))Some of the specifications include firm fixed effects The reported t-statistics are correctedfor clustering at the firm level

aMean significant at the 1 level

bMean significant at the 5 level

cMean significant at the 10 level

STDebtIssTotIss

LTDebtIssTotIss

STDebtIssethDthornETHORN

LTDebtIssethDthornETHORN

(1) (2) (3) (4) (5) (6) (7) (8)

Profitabilityt1 0002 0003 0042a 0008 0001 0001 0019a 0012a

(02) (04) (30) (05) (14) (10) (45) (28)

IndLevt1 0009 0007 0070b 0033 0001 0002c 0014a 0005

(06) (05) (26) (11) (08) (17) (26) (09)

M=Bt1 0003a 0003a 0003b 0010a 0000 0000 0003a 0002a

(48) (35) (22) (65) (10) (01) (84) (42)

Tangibilityt1 0020 0007 0011 0067 0005a 0006a 0097a 0081a

(12) (03) (03) (16) (34) (33) (108) (89)

Assetst1 0039a 0011b 0034a 0024b 0001a 0001 0046a 0020a

(89) (22) (38) (23) (36) (13) (190) (82)

Constant 0001b 0153a 0798a 0591a 0000c 0006a 0057a 0096a

(25) (144) (470) (288) (17) (124) (292) (234)

Year FE Yes Yes Yes Yes Yes Yes Yes Yes

Firm FE No Yes No Yes No Yes No Yes

R2 0025 0089 0018 0073 0030 0147 0016 0298

N 136146 136146 136140 136140 143031 143031 142847 142847

PROFITSndashLEVERAGE PUZZLEREVISITED 1445

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

Table XIII Externally financed leverage changes in good and bad times

The sample contains nonfinancial firms listed on the annual Compustat files for the period

from 1971 to 2009 The table presents estimates from regressions of change in debt tocapitalization ratio (in columns (1) (2) (5) and (6)) and change in leverage ratio consider-ing only external financing (in columns (3) (4) (7) and (8)) The change in leverage ratio

considering only external financing is defined as

LevEF frac14Dt1 thorn d

Dt1 thorn Et1 thorn dthorn e

Dt1

Dt1 thorn Et1

where D is debt E is book value of equity d is debt issues net of retirements and e is equityissues net of repurchases The regressions include lagged leverage factors and year indicator

variables The reported t-statistics are corrected for clustering at the firm level An industryis defined as having ldquogood timesrdquo if the median firm in that industry has a market-to-bookratio that is higher than the 67th percentile of the time-series distribution of industry

median market-to-book ratios Conversely an industry is defined as having ldquobad timesrdquoif the median firm in that industry has a market-to-book ratio that is lower than the 33rdpercentile of the time-series distribution of industry median market-to-book ratios

aMean significant at the 1 level

bMean significant at the 5 level

cMean significant at the 10 level

Good times Bad times

DDthornE LevEF D

DthornE LevEF

(1) (2) (3) (4) (5) (6) (7) (8)

Profitabiltyt1 0022c 0005 0016c 0025b 0073b 0041 0007 0007

(17) (03) (18) (24) (25) (11) (04) (03)

IndLevt1 0000 0011 0001 0009 0010 0018 0001 0000

(00) (09) (01) (13) (07) (10) (02) (00)

MB

t1

0004a 0003b 0000 0001 0008a 0001 0000 0000

(40) (24) (05) (08) (28) (04) (03) (03)

Tangt1 0097a 0041 0033a 0038b 0127a 0048 0059a 0033

(45) (16) (27) (25) (49) (14) (43) (16)

Assetst1 0014a 0013b 0024a 0019a 0037a 0023c 0036a 0025a

(30) (20) (84) (48) (45) (17) (88) (43)

Constant 0030b 0036 0000 0003 0028c 0014 0021 0001

(20) (16) (00) (07) (17) (06) (11) (01)

Year FE Yes Yes Yes Yes Yes Yes Yes Yes

Firm FE No Yes No Yes No Yes No Yes

R2 0006 0088 0009 0117 0026 0251 0023 0202

N 56820 56820 58200 58200 23462 23462 23511 23511

1446 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

conclusions The traditional leverage result on profits is driven by increasesin equity that are experienced and then partially but not completely offsetby firmsTo the extent that firms are engaging in active financing they are following

the prescriptions of the trade-off theory Thus we have a coherent picture ofhow firms are responding to profits Basically more profitable firms experiencean increase in the value of their equity They respond to this by issuing debtand repurchasing equity But the magnitude of the increase in equity is largerthan the magnitudes of the active steps taken in the debt and equity marketsThis means that the firmsrsquo actual actions are as predicted by the trade-offtheory But it leaves open the question of why they do not act more stronglyThe capital structure literature has long been interested in dynamics and

capital structure adjustment paths In recent years this interest has increas-ingly resulted in explicit dynamic optimizing models such as those ofGoldstein Ju and Leland (2001) Hennessy and Whited (2005) andStrebulaev (2007) These models commonly make specialized assumptionsleading some to question their generalityWe argue that both fixed transactions costs and variable transactions costs

are plausible and realistic in models that pin down firm size The simultaneouspresence of both is important Due to fixed costs firms only rebalance inter-mittently Due to variable adjustment costs the extent of the adjustment isreduced to the point at which an extra unit to adjustment is balanced againstthe extra unit transactions costs This means that both leverage increases andleverage reductions will only go part way toward a static optimumThe key point is that with both fixed and variable transactions costs we

should expect to only see intermittent leverage adjustments When leverage isadjusted it should ldquounder adjustrdquo when compared to a frictionless modelFurther discussion of transactions costs in a capital structure setting isprovided by Leary and Roberts (2005)

9 Conclusion

The connection between corporate profits and capital structure has beenvery influential in the assessment of the trade-off theory The conventionalinterpretation of the evidence has pushed the literature away from the trade-off theory and toward much more complex dynamic models and ideas As aresult it is important to make sure that the evidence has been correctlyinterpreted We show that the problem is with the interpretation of thetest results and a failure to take into account the costs associated with re-sponding to a profit shock

PROFITSndashLEVERAGE PUZZLEREVISITED 1447

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

We document that more profitable firms really do borrow more (not less)They tend to repurchase equity They experience an increase in both thebook value of equity and the market value of equity Less profitable firmsreally do tend to reduce their debt and to issue equityFirm size and market conditions also matter Larger firms tend to be more

active in the debt markets while smaller firms tend to be relatively moreactive in the equity markets During good times there is more use ofexternal financingThe usual profitsndashleverage puzzle result is primarily driven by the increase

in equity that is experienced by the more profitable firms Accordingly thepuzzle should be restated as asking why do firms not take sufficiently largeoffsetting actions to fully undo the change in equity What limits the mag-nitudes of the typical leverage response to profit shocksIn a frictionless model the partial response appears puzzling But there is

good empirical reason to believe that rebalancing entails both fixed andvariable costs and that firm size matters The rebalancing costs can befully avoided by doing nothing Accordingly the firm must decide whetherany given shock is big enough to be worth responding to If it is then thefirm must decide how big a response is called for These technical conditionsare known as ldquovalue matchingrdquo and ldquosmooth pastingrdquo Thus optimizationimplies that some shocks will be ignored Even if the shock is not ignoredthe optimal response will only partially undo the shock The magnitude ofthe leverage response must balance the marginal cost and the marginalbenefit of an extra unit of leverage Since the marginal cost of adjustingleverage is strictly positive the adjustment toward that static leverageoptimum will only go part way This is true both for leverage increasesand for leverage reductionsOne might conjecture that the variable costs would be time varying and

that such costs might have declined over time Since we do not have a fullaccounting of all of the variable costs it is of course impossible to be sureHowever there is some evidence suggesting that at least the directunderwriting fees are remarkably stable over time10

Surely nobody ought to be surprised that more profitable firms are moreinclined to repurchase equity while unprofitable firms tend not to do so Thefact that more profitable firms typically experience an increase in the value of

10 For IPOs Jay Ritter provides important summary evidence as httpbearwarringtonufleduritterIPOs2013Underwritingpdf He finds that for medium-sized deals a remark-

ably large fraction of the deals have a 7 fee year after year Whether this is due tobehavioral factors or some sort of a coordination device is not clear The extent towhich similar intertemporal stability carries over to other forms of external financemerits further study

1448 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

equity is equally natural Were it not for the contrary claims in the literature(ie the profitsndashleverage puzzle) we would argue that nobody ought to findit surprising that more profitable firms are more inclined to issue debt Thefact that leverage adjustments typically only partially offset profit shocksshould also sound quite natural These facts all fit together easily within thetrade-off theory once we take account of both fixed and variable costs ofactively adjusting capital structure

Appendix

Table AI Market leverage regressions

The table presents estimates of the regressions of market leverage on firm characteristics

The sample comes from the annual Compustat files during the period 1971ndash2009 Financialfirms are excluded The leverage is estimated as the ratio of debt over debt plus marketequity The explanatory variables Profitabilityt1

MB

t1

Tangibilityt1 and LnethAssetsTHORNt1are described in Table 1 IndMedianLevt1 is estimated as the median market leverage of allother firms in the same industry The industry is defined at the level of the four-digit SICcode All specifications include the year fixed effects The specifications in column (2)

additionally include the firm fixed effects We report t-statistics where the standard errorsare clustered at the firm level in parentheses

aSignificant at the 1 level

Quantile regressions

(1) (2) (3) (4) (5)

Profitabilityt1 013a 013a 007a 011a 020a

(275) (262) (349) (425) (538)

IndLevt1 046a 020a 032a 058a 063a

(506) (239) (1299) (1630) (1042)MB

t1

003a 001a 001a 002a 004a

(509) (303) (743) (709) (534)

Tangibilityt1 013a 018a 013a 014a 010a

(191) (181) (762) (571) (232)

LnethAssetsTHORNt1 001a 005a 001a 001a lt001a

(112) (258) (682) (342) (29)

Constant 010a 011a 005a 005a 029a

(171) (98) (173) (118) (398)

Year FE Yes Yes Yes Yes Yes

Firm FE No Yes No No No

Regression type OLS FE 25th tile Median 75thtile

Clustered SE Yes Yes No No No

R2 030 067

Pseudo R2 016 023 0201

N 158578 158578 158578 158578 158578

PROFITSndashLEVERAGE PUZZLEREVISITED 1449

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

Table AII Profitability sorts for firms issuing large amounts

Percentage of firms

Iss Ret Iss Iss Rep Iss Iss IssD IssE

D D NetD E E NetE DampE RepE RetD

(1) (2) (3) (4) (5) (6) (7) (8) (9)

Panel A 10 cutoff to define issuing or retiring activity

Allfirm

s

Low 205 150 134 260 05 257 45 00 23

2 207 175 103 77 06 75 11 01 10

3 243 196 101 50 08 47 08 01 09

4 252 186 113 50 13 48 08 02 08

High 227 161 118 60 27 57 07 04 09

t Highfrac14Low 70 38 62 755 226 768 320 90 152

Smallfirm

s Low 230 132 162 380 05 375 82 00 32

2 228 143 165 334 04 333 65 00 26

3 211 146 147 216 03 214 38 00 23

4 190 154 120 126 04 124 15 01 20

High 194 162 114 99 09 97 12 01 15

t Highfrac14Low 42 40 68 342 20 340 169 07 131

Largefirm

s Low 231 200 82 31 06 29 08 01 04

2 231 187 67 17 06 14 03 01 02

3 234 177 77 17 09 14 03 02 03

4 218 158 75 16 17 14 02 03 02

High 218 135 99 19 41 15 03 07 02

t Highfrac14Low 19 106 36 46 135 57 39 58 23

Percentage of firms

Iss Ret Iss Iss Rep Iss Both Iss D Iss E

D D Net D E E Net E DampE Rep E Ret D

(1) (2) (3) (4) (5) (6) (7) (8) (9)

Panel B 15 cutoff to define issuing or retiring activity

Allfirm

s

Low 158 101 100 230 01 228 27 00 15

2 142 115 63 57 01 56 05 00 06

3 165 128 57 32 01 31 03 00 05

4 173 118 65 35 01 34 03 00 05

High 160 102 70 46 04 44 03 00 05

t Highfrac14Low 08 06 141 738 87 746 277 24 131

Smallfirm

s Low 195 95 132 354 01 350 56 00 24

2 185 96 131 303 01 301 44 00 18

3 160 96 104 179 00 176 18 00 15

4 134 99 78 99 01 98 08 00 12

High 133 95 73 82 01 80 06 00 10

t Highfrac14Low 80 00 95 345 12 344 149 12 52

Largefirm

s Low 145 123 42 14 01 12 03 00 01

2 137 114 32 07 01 06 01 00 01

3 142 104 35 06 01 05 00 00 01

4 130 91 36 08 03 07 01 01 01

High 134 81 48 09 06 07 01 01 01

t Highfrac14Low 19 85 17 26 54 33 34 17 11

1450 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

Table AIV Debt and equity changesmdashwithout fixed effects

The sample contains nonfinancial firms listed on the annual Compustat files for the period from 1971

to 2009 The table presents estimates from regressions of change in debt (D) change in the book

value of equity (BVE) change in the market value of equity (MVE) and net equity issuances

(EqyIss) The explanatory variables include current and lagged change in operating income before

depreciation (Profits and Profitst1) and lagged change in book value of assets (Assetst1) The

regressions include year indicator variables The reported t-statistics are corrected for clustering at the

firm level

aMean significant at the 1 level

bMean significant at the 5 level

cMean significant at the 10 level

D BVE MVE EqyIss

(1) (2) (3) (4) (5) (6) (7) (8)

Profitst1 025a 024a 023a 038a 007 066a 007a 009a

(40) (32) (36) (39) (06) (29) (42) (51)Profits 051b 082a 149a 004a

(25) (33) (49) (33)Assetst1 002 005c 024a 001b

(08) (17) (34) (23)Constant 93 176a 45 71 1552a 1704a 130a 140a

(12) (28) (07) (09) (49) (49) (45) (45)Year FE Yes Yes Yes Yes Yes Yes Yes YesR2 001 005 001 011 001 005 002 002N 162056 162056 162130 162130 157550 157550 162154 162154

Table AIII Median profitability sorts

The sample contains nonfinancial firms listed on the annual Compustat files for the period from 1971

to 2009 The table presents median debt and equity median changes in debt and equity and median

equity issuances (EqyIss) in millions of dollars for all firms sorted on profitability and for profitability

sorts within the smallest and largest firms The sorts are done annually

D D BVE BVE MVE MVE EqyIss Assets(1) (2) (3) (4) (5) (6) (7) (8)

Allfirm

s

Low 21 00 75 17 256 20 01 1922 186 02 449 18 598 19 00 11033 816 06 1274 03 1637 05 00 32244 787 02 1541 18 2488 03 00 3556High 273 00 1361 66 3300 14 00 2578

Smallfirm

s Low 05 00 13 07 140 14 06 402 06 00 25 08 114 12 01 563 09 00 32 05 85 07 00 674 12 00 39 01 68 03 00 79High 10 00 51 04 98 01 00 90

Largefirm

s Low 11485 381 10269 371 13227 202 03 333902 13051 393 12786 140 17571 169 05 392463 11554 287 13429 24 21623 305 00 370314 9963 209 15052 194 29424 308 00 36705High 7361 00 16380 613 49142 439 00 34976

PROFITSndashLEVERAGE PUZZLEREVISITED 1451

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

References

Auerbach A J (1985) Real determinants of corporate leverage in B M Friedman (ed)

Corporate Capital Structures in the United States University of Chicago Press Chicago

IllinoisBertrand M and Mullainathan S (2003) Enjoying the quiet life corporate governance and

managerial preferences Journal of Political Economy 111 1043ndash1075Booth L Varouj A Demirguc-Kunt A and Maksimovic V (2001) Capital structures in

developing countries Journal of Finance 56 87ndash130Brennan M J and Schwartz E S (1984) Optimal financial policy and firm valuation

Journal of Finance 39 593ndash607Cameron A C and Trivedi P K (2010) Microeconometrics using Stata Stata Press TexasConstantinides G M (1979) Multiperiod consumption and investment behavior with

convex transactions costs Management Science 25 1127ndash1137Constantinides G M and Richard S F (1978) Existence of optimal simple policies for

discounted-cost inventory and cash management in continuous time Operations Research

26 620ndash636DeAngelo H DeAngelo L and Whited T M (2011) Capital structure dynamics and

transitory debt Journal of Financial Economics 99 235ndash261Dixit A K (1993) The Art of Smooth Pasting Routledge London and New YorkFama E and French K R (2002) Testing trade-off and pecking order predictions about

dividends and debt Review of Financial Studies 15 1ndash33Faulkender M and Petersen M A (2006) Does the source of capital affect capital struc-

ture Review of Financial Studies 19 45ndash79Faulkender M Flannery M J Hankins K W and Smith J M (2012) Cash Flows and

Leverage Adjustments Journal of Financial Economics 103 632ndash646Fischer E Heinkel R and Zechner J (1989) Dynamic capital structure choice theory

and tests Journal of Finance 44 19ndash40Frank M Z and Goyal V K (2008) Trade-off and pecking order theories of

debt in B E Eckbo (ed) Handbook of Corporate Finance Empirical Corporate Finance

Vol 2 Handbook of Finance Series Chapter 12 ElsevierNorth-Holland Amsterdam pp

135ndash202

Frank M Z and Goyal V K (2009) Capital structure decisions which factors are reliably

important Financial Management 38 1ndash37

Garvey G T and Hanka G (1999) Capital structure and corporate control the effect of

antitakeover statutes on firm leverage Journal of Finance 54 519ndash546

Goldstein R Ju N and Leland H (2001) An EBIT-based model of dynamic capital

structure Journal of Business 74 483ndash512

Graham J R and Tucker A L (2006) Tax shelters and corporate debt policy Journal of

Financial Economics 81 563ndash594

Hackbarth D and Mauer D C (2012) Optimal priority structure capital structure and

investment Review of Financial Studies 25 747ndash796

Halling M Yu J and Zechner J (2012) Leverage dynamics over the business cycle

Working paper University of Utah University of New South Wales and Vienna

University of Economics and Business AdministrationHennessy C A and Whited T A (2005) Debt dynamics Journal of Finance 60 1129ndash1165

Huang R and Ritter J R (2009) Testing theories of capital structure and estimating the

speed of adjustment Journal of Financial and Qualitative Analysis 44 237ndash271

1452 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

Hovakimian A Opler T and Titman S (2001) The debt-equity choice Journal of

Financial and Quantitative Analysis 36 1ndash24Jung K Kim Y and Stulz R M (1996) Timing investment opportunities managerial

discretion and the security issue decisions Journal of Financial Economics 42 159ndash185Kane A Marcus A J and McDonald R L (1984) How big is the tax advantage to

debt Journal of Finance 39 841ndash853Kayhan A and Titman S (2007) Firmsrsquo histories and their capital structures Journal of

Financial Economics 83 1ndash32Leary M T (2009) Bank loan supply lender choice and corporate capital structure

Journal of Finance 64 1143ndash1185Leary M T and Roberts M R (2005) Do firms rebalance their capital structures

Journal of Finance 60 2575ndash2619Leland H E (1994) Corporate debt value bond covenants and optimal capital structure

Journal of finance 49 1213ndash1252Lemmon M L Roberts M R and Zender J F (2008) Back to the beginning persist-

ence and the cross-section of corporate capital structure Journal of Finance 63 1575ndash1608Long M S and Malitz I B (1985) The investment-financing nexus some empirical

evidence Midland Corporate Finance Journal 3 53ndash59Mackie-Mason J K (1990) Do taxes affect corporate financing decisions Journal of

Finance 45 1471ndash1493Morellec E Nikolov B and Schurhoff N (2012) Corporate governance and capital

structure dynamics Journal of Finance 67 803ndash848Myers S C (1984) The capital structure puzzle Journal of Finance 39 575ndash592Myers S C (1993) Still searching for optimal capital structure Journal of Applied

Corporate Finance 61 4ndash14Rajan R and Zingales L (1995) What do we know about capital structure some evidence

from international data Journal of Finance 50 1421ndash1460Strebulaev I A (2007) Do tests of capital structure theory mean what they say Journal of

Finance 62 1747ndash1787

Strebulaev I A and Whited T M (2012) Dynamic Models and Structural Estimation in

Corporate Finance Now Publisher Boston and Delft

Titman S and Wessels R (1988) The determinants of capital structure choice Journal of

Finance 43 1ndash21

Welch I (2004) Capital structure and stock returns Journal of Political Economy 112

106ndash131

Welch I (2011) Two common problems in capital structure research the financial-debt-

to-asset ratio and issuing activity versus leverage changes International Review of Finance

11 1ndash17

PROFITSndashLEVERAGE PUZZLEREVISITED 1453

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

percentage of firms issuing or repurchasing debt or equity The table alsoreports test statistics from tests that examine whether the percentage differ-ences between the high- and low-profit firms are significantly different fromeach other The results in Column (1) show that the likelihood of issuing debt

Table IV Profitability sorts for debt and equity issuers

The sample contains nonfinancial firms listed on the annual Compustat files for the period

from 1971 to 2009 The table presents the percentage of firms issuing and retiring (orrepurchasing) debt and equity The firms are classified as ldquoissuing debtrdquo if they issuedebt in excess of 5 of the value of their assets as ldquoissuing equityrdquo if they issue equity

in excess of 5 of the value of their assets as ldquoretiring debtrdquo if they retire debt in excess of5 of the value of their assets and as ldquorepurchasing equityrdquo if they repurchase equity inexcess of 5 of the value of their assets In addition we also report net debt issuers which

are firms that issue net debt over 5 of the value of their assets and net equity issuerswhich are firms that issue net equity in excess of 5 of the value of their assets Weannually sort firms on lagged profitability and report the percentage of firms in each ofthese categories The bottom part of each panel reports the percentage of firms issuing or

retiring securities by profitability within the smallest and largest asset quintiles The tablealso reports results from t-test for the difference in percentage of firms issuing or retiringsecurities in the lowest and the highest profit quintile

Percentage of firms

Iss Ret Iss Iss Rep Iss Both IssD IssE

D D NetD E E NetE DampE RepE RetD

(1) (2) (3) (4) (5) (6) (7) (8) (9)

Allfirm

s

Low 276 245 186 301 13 297 71 01 38

2 319 299 181 117 21 112 27 02 20

3 385 344 194 86 29 80 21 05 15

4 386 330 213 81 43 75 20 07 14

High 329 273 198 92 82 81 17 13 14

t Highfrac14Low 150 84 39 723 414 763 358 176 200

Smallfirm

s Low 275 204 201 411 11 405 109 01 47

2 291 233 213 379 10 378 99 01 45

3 281 246 204 265 14 261 66 01 41

4 275 272 186 162 15 159 34 02 35

High 275 281 178 136 27 131 25 02 26

t Highfrac14Low 01 84 28 316 55 318 171 18 55

Largefirm

s

Low 414 363 177 87 22 79 28 03 11

2 449 379 180 62 30 57 17 07 09

3 439 360 186 56 41 48 16 09 07

4 403 320 194 46 67 39 13 12 06

High 374 264 207 50 129 35 10 28 05

t Highfrac14Low 50 130 45 89 244 116 77 120 47

1430 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

is mostly independent of firm profitability In the aggregate there is a smalldifference but it is sensitive to sorting on firm size Similarly in Column (2)there is only a weak relation between the profitability and the likelihood thatthe firm retires debt To a first approximation the fraction of firms issuing orrepurchasing debt is largely independent of profits Large firms are somewhatmore likely than small firms to issue debt and also to retire debtUnlike debt issues the probability of issuing equity is strongly related to

profitability as can be seen in Column (4) Low-profit firms are much morelikely to issue equity than are high-profit firms In the lowest profit quintile301 of the firms issued equity In the highest profit quintile it was only92 The lower panel reports separate sorts for small and large firms Similarpatterns are found for both groups of firms that is the likelihood of equityissuance declines as profitability increases The differences between the quin-tiles are quite stark for the group of small firms All quintiles of small firmsissue equity much more frequently than the corresponding large firm quintilesHigh-profit firms are much more likely to repurchase equity as shown in

Column (5) In the overall sample only 13 of low-profit firmsrepurchased equity compared with 82 of the high-profit firms Whenwe separate the small and large firms again a difference emerges Ingeneral small firms do not repurchase much equity while a greaterfraction of the large firms do soGiven these facts it is natural that low profitability firms are much more

likely to be net issuers of equity than are high profitability firms as shown inColumn (6) Among the firms in the lowest quintiles of profits roughly 30of firms issue net equity that exceeds 5 percent of their assets In contrastfor the most profitable quintile only about 8 issue net equityFirm size is an important variable in the recent literaturemdashit is often used

as a proxy for access to capital markets (as in Faulkender and Petersen(2006) and Leary (2009)) Small firms are bank-dependent risky and infor-mationally opaque They have restricted access to public debt markets andconsequently face more severe supply constraints in their ability to issuedebt Thus we expect small firms to be more sluggish in adjusting theirdebt and equity in response to shocks to profitability Large firms incontrast have much easier access to public debt markets and they facefewer obstacles in accessing securities marketsTo examine how size and profitability interact the bottom part of Table IV

examines profitability sorts for the smallest and largest firms We first sortfirms annually by firm size and then within each size quintile we sort them onprofitability We do these two-way sorts to ensure that we have similarnumber of firms in profitability subgroups for both small and large firmsamples Unconditionally small firms are generally less profitable while

PROFITSndashLEVERAGE PUZZLEREVISITED 1431

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

large firms are relatively more profitable The two-way sorts allow us to havea uniform distribution of firms in various profitability quintiles for both smalland large firm samples The bottom part of Table IV reports results for thesmallest and the largest size quintiles Among the small firms there is littlerelation between profitability and the likelihood of issuing debt Howeverlarger firms exhibit a small increase in the likelihood of net issuance of debtwith increasing profitabilityIn contrast to debt the effects of profitability on equity issuance and

repurchases are much more consistent across size quintiles As profitabilityincreases firms are generally less likely to issue equity and more likely torepurchase it Across the two extreme size groups we note a striking differ-ence between the proportion of firms issuing equity and those issuing debtAmong low-profit small firms almost 41 are equity issuers whereas onlyabout 9 of low-profit large firms are equity issuers However regardless ofsize we note a monotonic reduction in the likelihood of issuing equity asprofitability increasesColumn (7) reports the fraction of firms that issue both debt and equity in

excess of 5 of assets both as a function of profits in the upper panel andas a function of firm size and profits in the lower panel In the upper panelwe see that low-profit firms are much more likely to issue both debt andequity (71) while the high-profit firms are less likely to do so (17) Thelower panel shows that small firms are much more likely to be issuers of bothin contrast to large firms In both firm size categories the profit pattern is thesame with low profits being more frequent issuers of both simultaneouslyImportantly Column (8) shows that the likelihood of issuing debt and

simultaneously repurchasing equity increases with profitability Converselyin Column (9) we find that the likelihood of doing the reverse that is issuingequity and retiring debt declines with profitability The effects of sorting onfirm size and firm profitability mirror those reported for all firms Overallthe results show that low-profit firms are less likely to issue debt and repur-chase equity they are instead more likely to issue equity and retire debtFirms with high profitability exhibit the reverse patternIn Appendix Table AII we present results from profitability (and size)

quintiles for large issuances defined using the 10 cutoff (in Panel A) andthe 15 cutoff (in Panel B) The results from these panels are largely con-sistent with those described above

52 MAGNITUDES OF ISSUANCES AND EQUITY CHANGES

In the previous section we considered the probability of having a nontriviallevel of debt or equity activity The next question is how large are the dollar

1432 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

values involved In Table V we sort firms according to profits and thentabulate the levels and changes in both debt and equity We do this firstfor all firms and then for small versus large firmsFor the analysis on all firms we observe in Column (1) that debt peaks at

the middle of the distribution This is because firms with medium profitabil-ity are also the largest as seen in Column (8) which reports average assetvalues for different profitability quintiles When we sort first by firm size andthen examine firms with different levels of profitability within size quintileswe find that for small firms debt is roughly independent of profits while forlarge firms less profitable firms have a higher level of debtAs expected Columns (3) and (5) show that more profitable firms have

higher equity values Columns (2) and (7) consider the relationship betweenissuances of debt and equity and firm profitability Debt issuances are sig-nificantly larger for more profitable firms In contrast equity issuances arelarger for low-profit firms Small low-profit firms issue substantially lessequity compared with small high-profit firms Among large firms the low-profit ones tend to issue equity while the high-profit ones tend to repurchaseequity The fact that more profitable firms issue debt and repurchase equitywhile the least profitable firms retire debt and issue equity is consistent withthe predicted relation between profitability and financing decisions under thetrade-off theoryWe also examine two-way sorts by size and profitability and report results

for the smallest and largest quintile of firms For the smallest quintile offirms the change in debt is largely unrelated to profits But for large firmsthere is a positive relation between profits and debt issuances High-profitfirms have a big positive change in debt Low-profit large firms have anegative change in debtColumns (4) and (6) provide an explanation of why the leverage ratio

regression results contradict those from the basic profitability sorts pre-sented here As we can see profitability indirectly affects leverage ratiosby increasing equity values Changes in both the book value of equity andthe market value of equity are positive and large for highly profitable firmsIn contrast these changes are negative for less profitable firms This evidenceillustrates an important issue concerning the use of leverage ratios Suchratios are often interpreted as essentially reflecting the use of debt by thefirm This interpretation while common is empirically misleadingFor the typical firm the change in the value of equity is larger than the

change in debt For example in the third profit quintile for large firmscolumn (7) the mean equity issue is just $9 million but the change in themarket value of equity is $242 million At the same time the mean change indebt is $27 million This suggests that a fair bit of the observed variation in

PROFITSndashLEVERAGE PUZZLEREVISITED 1433

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

the leverage ratios is primarily driven by the changes in the market value ofequity in the denominator rather than by the changes in debt in the numer-ator Since equity issues are often small this implies that the variation in theleverage ratio is primarily driven by internal operations rather than byexternal financing actions This again points to the fact that leverageratios can provide a misleading account of actual patterns in the dataThere is always a potential concern that the averages may be misleading

due to the impact of outliers To address this concern in AppendixTable AIII we present median values of the profitability sorts Theseresults generally reinforce the findings in Table VTable VI again considers the magnitudes of financing activity but this

time the issues are scaled by a measure of firm size That way it is easier to

Table V Magnitude of financing activity

The sample contains nonfinancial firms listed on the annual Compustat files for the period

from 1971 to 2009 The table presents average debt and equity changes in debt and equityand equity issuances (in millions of dollars) for firms sorted on profitability within sizeclasses The table reports information for all firms sorted on profitability and for profit-

ability sorts within the smallest and largest firms The sorts are done annually The tablealso reports results from t-test for the difference in percentage of firms issuing or retiringsecurities in the lowest and the highest profit quintile

D D BVE BVE MVE MVE EqyIss Assets

(1) (2) (3) (4) (5) (6) (7) (8)

Allfirm

s

Low 83 5 78 10 120 5 7 234

2 605 13 509 24 721 8 17 1691

3 1227 9 1123 22 1834 3 9 3551

4 833 23 1166 7 2408 74 9 2994

High 570 53 1171 71 3817 68 36 2581

t Highfrac14Low 31 9 36 9 38 2 16 36

Smallfirm

s Low 25 04 42 01 332 28 33 97

2 29 06 55 02 278 16 21 116

3 31 07 63 01 212 05 12 127

4 30 06 70 05 179 10 08 136

High 24 06 84 14 234 17 07 142

t Highfrac14Low 04 15 153 85 73 11 201 107

Largefirm

s Low 3929 55 2990 152 4671 11 63 10846

2 3826 19 3178 95 5122 66 11 10721

3 3074 27 3498 41 6173 242 9 9697

4 2623 62 3915 45 8473 94 57 9855

High 2180 193 4642 261 14924 202 148 10171

t Highfrac14Low 94 64 108 87 246 12 169 16

1434 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

see whether large firms behave disproportionately when compared withsmall firms Column (1) shows that low-profit firms have more debtas reported in many previous studies In Columns (4) and (6) we seethat low-profit firms have a major decline in both the book value ofequity and the market value of equity The high-profit firms have thereverse experienceMore interesting are the active decisions of the firms In Column (2) we see

that in contrast to the usual impression the low-profit firms are reducingdebt while the high-profit firms are increasing it In Column (7) we see thatthe low-profit firms are also reducing leverage by issuing equity The high-profit firms issue little equity In the lower panel similar results are foundwhen we separate small and large firms Equity issuing is particularly strong

Table VI Magnitude of financing activitymdashscaled levels and changes

The sample contains nonfinancial firms listed on the annual Compustat files for the period

from 1971 to 2009 The table presents average debt and equity changes in debt and equityand equity issuances scaled by assets for firms sorted on profitability within size classes Thetable reports information for all firms sorted on profitability and for profitability sorts

within the smallest and largest firms The sorts are done annually The table also reportsresults from t-test for the difference in percentage of firms issuing or retiring securities inthe lowest and the highest profit quintile

D D BVE BVE MVE MVE EqyIss

(1) (2) (3) (4) (5) (6) (7)

Allfirm

s

Low 386 122 12 871 3427 1868 217

2 285 58 407 119 940 157 28

3 293 12 425 14 742 03 12

4 269 26 461 02 955 27 10

High 192 15 551 47 1675 74 07

t Highfrac14Low 35 49 50 47 127 42 157

Smallfirm

s Low 545 154 841 2145 9688 9128 696

2 684 49 712 1915 5014 2318 288

3 331 253 141 969 3063 954 129

4 267 181 357 175 1673 221 55

High 194 05 543 26 1758 147 41

t Highfrac14Low 49 17 51 36 88 27 99

Largefirm

s Low 350 30 306 58 511 55 09

2 346 14 335 13 523 02 04

3 353 165 332 71 711 29 01

4 280 01 414 04 909 09 05

High 227 15 477 19 1643 27 14

t Highfrac14Low 371 112 453 181 570 57 221

PROFITSndashLEVERAGE PUZZLEREVISITED 1435

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

among the small low-profit firms that are also experiencing sharp reductionsin the value of equity In column (2) we see that both for the large and for thesmall firms net debt increases are primarily found in the highest quintile offirm profitability

6 Debt and Equity Regressions

So far we have demonstrated that in our data the conventional leverageregressions have the usual signs We then demonstrated that in simple sortsof the data strikingly different active financing patterns emerge The nexttask is to reconcile these differencesTable VII presents simple regressions predicting changes in debt book

equity and market equity as well as equity issuances Consistent with thesorts we control for changes in firm size and we also include firm and yeardummies There may be a concern about the speed of a firmrsquos responsesAccordingly we consider the lagged change in profits as the main case butwe also include the contemporaneous change in profits Whether the contem-poraneous change in profits is included or not does not in fact matter Wereport panel-robust standard errors adjusted for clustering at the firm levelThe results in Table VII support the findings in the sorts Firms that have

an increase in profits in 1 year increase debt both in the same year and in thenext year Firms that have an increase in profits experience an increase inboth the book value of equity and the market value of equity Importantlyhowever firms that experience and increase in profits reduce their equityissues All these effects are statistically significant (except for the marketvalue of equity in the column 5 specification) and seem intuitively reason-able In Appendix Table AIV similar regressions to those presented aboveare reported but without firm-fixed effects We find that empirically verylittle changes whether we include firm fixed effects or leave them out Thusthe targeting behavior to the extent that it happens does not account forwhat we are reportingThe next step is to bring the sorts and the simple regressions together into

a common setting We therefore include factors in addition to profitability(i) median industry leverage (ii) the market-to-book assets ratio (iii) tangi-bility of assets and (iv) firm size (measured by the log of assets) Rajan andZingales (1995) show that these factors are related to leverage in G7countries A number of studies have used these factors to estimateleverage targets Frank and Goyal (2009) show that these factors arerobustly related to leverage in the USA In these regressions we use indica-tors for the quintile that the firm is in for each factor

1436 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

Table VIII reports these results In Columns (1) and (2) we examine debtissues and again show that profits positively affect debt issuances The effectsare large and statistically significant at the 1 level Columns (3) and (4)which examine book equity show a strong effect of profits This effect is onlyslightly reduced by the inclusion of other factors Columns (5) and (6) illus-trate that the impact of profits on the change in market equity is also robustto the inclusion of the conventional factors Columns (7) and (8) examineequity issuances Here we again find that profits have a significant negativeimpact on equity issuances As before the results are robust to the inclusionof the conventional factorsIn unreported tables we estimate debt and equity changes using quantile

regressions In terms of the signs and significance of the profitability variable

Table VII Debt and equity changes

The sample contains nonfinancial firms listed on the annual Compustat files for the period

from 1971 to 2009 The table presents fixed effect estimates from regressions of changes indebt (D) changes in the book value of equity (BVE) changes in the market value ofequity (MVE) and net equity issuances (EqyIss) The explanatory variables include

current and lagged change in operating income before depreciation (Profits andProfitst1) and lagged change in the book value of assets (Assetst1) In addition tofixed firm effects the regressions include year indicator variables The reported t-statistics

are corrected for clustering at the firm level

aMean significant at the 1 level

bMean significant at the 5 level

cMean significant at the 10 level

D BVE MVE EqyIss

(1) (2) (3) (4) (5) (6) (7) (8)

Profitst1 022a 025a 015b 035a 004 060b 005b 007a

(33) (31) (25) (35) (03) (25) (25) (53)

Profits 050b 078a 140a 003a

(24) (32) (48) (34)

Assetst1 000 007b 025a 001b

(00) (21) (35) (25)

Constant 167 283a 34 51 1145a 1182a 605a 595a

(14) (30) (04) (05) (32) (29) (69) (74)

Year FE Yes Yes Yes Yes Yes Yes Yes Yes

Firm FE Yes Yes Yes Yes Yes Yes Yes Yes

R2 001 004 001 010 001 005 002 003

N 162056 162056 162130 162130 157550 157550 162154 162154

PROFITSndashLEVERAGE PUZZLEREVISITED 1437

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

the results are qualitatively similar to the OLS results reported in Table VIIIProfitability positively affects debt issuances and negatively affects equityissuances across various quantiles Overall the results in Table VIII showthat there is nothing that is special about using sorts or regressions to explaindebt and equity issues In either case we find that more profitable firms tendto increase their debt experience an increase in the value of equity andrepurchase shares Thus the control factors are not responsible for theusual rejection of the trade-off theory It would appear that what mattersis the size of the issuing activity relative to the organic increases in equityvalue due to profits In other words what matters in the leverage regressionseems to be coming from the presence of E in D

DthornE and not from D

Table VIII Debt and equity issuances profitability and leverage factor quintiles

The sample contains nonfinancial firms listed on the annual Compustat files for the period

from 1971 to 2009 The table presents estimates from regressions of change in debt (D)change in the book value of equity (BVE) change in the market value of equity (MVE)and net equity issuances (EquityIss) The explanatory variables include lagged profitability

quintiles and lagged leverage factor quintiles In addition to fixed firm effects the regres-sions include year indicator variables The reported t-statistics are corrected for clustering atthe firm level

aMean significant at the 1 level

bMean significant at the 5 level

cMean significant at the 10 level

D BVE MVE EqyIss

(1) (2) (3) (4) (5) (6) (7) (8)

ProfitabilityQint 214a 167a 283a 175a 80 429a 53a 55a

(70) (54) (66) (59) (08) (48) (55) (62)

IndLevQint 104a 43 99 20

(29) (15) (10) (10)

Ln(Assets)Qint52 17 1026a 53a

(16) (05) (77) (35)MB

Qint158a 372a 1155a 09

(45) (56) (106) (06)

TangibilityQint 05 177 436c 64a

(01) (15) (18) (31)

Constant 167 96 656b 235 4613a 8610a 113 570a

(08) (04) (26) (04) (61) (58) (11) (31)

Year FE Yes Yes Yes Yes Yes Yes Yes Yes

Firm FE Yes Yes Yes Yes Yes Yes Yes Yes

R2 0002 0002 0003 0004 0013 0014 0013 0014

N 155587 155037 155587 155037 155587 155037 155587 155037

1438 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

7 Are Financial Market Conditions Important

According to the trade-off theory capital structure is determined by con-sidering the impact of costs and benefits of debt The time variation inthe costs and benefits of debt imply that issuance decisions would alsovary over time Furthermore market conditions affect the transactioncosts of adjusting Adverse selection is a greater problem in a cold marketthan in a hot one Accordingly it may be easier to adjust in good marketconditions Related ideas have been developed by Huang and Ritter(2009) Faulkender et al (2012) and Halling Yu and Zechner (2012)Faulkender Flannery Hankins and Smithrsquos paper examines cross-sectionalvariation in adjustment speeds including market valuations andhow they may affect adjustment speeds It is natural to think that ingood market conditions it will be less costly to issue both debt and equityand adjusting leverage to respond to profitability shocks would be mucheasierTo test the importance of market conditions we require a definition of

good times and bad times Our empirical strategy is to estimate good timesversus bad times at the four-digit industry level We define an industry ashaving ldquogood timesrdquo if the median firm in that industry has a market-to-book ratio that is higher than the 67th percentile of the time-series distribu-tion of the industry median market-to-book ratios Conversely an industryis defined as having ldquobad timesrdquo if the median firm in the industry has amarket-to-book ratio that is lower than the 33rd percentile of the time-seriesdistribution of the industry median market-to-book ratiosPanel A of Table IX tabulates issuance activity for profitability sorts in

both good times and bad times As might be expected active debt and equityissuances are larger during good times There is also somewhat more activeswaps between debt and equity during good timesPanel B reports the issuance activity for the smallest and the largest firms

during the good times and bad times Within each size quintile firms aresorted on profitability This is a three-way sort As before we find that thereis much more active use of external markets during good times Small low-profit firms are more likely to issue equity in good times than in bad timesDebt issuances are significantly higher in good times Large high-profitfirms are significantly more likely to issue debt and repurchase equity ingood times than in bad timesPanel C reports the magnitudes of the financial variables rather than the

frequencies During bad times less profitable large firms retire substantialamounts of debt and they show a tendency to issue equity Small firms donot seem to engage in similar debt reduction activities However like the

PROFITSndashLEVERAGE PUZZLEREVISITED 1439

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

Table IX Debt and equity issues in good and bad times

The table reports the frequency and magnitude of financing activity for sorts on profits for subsamples

of firms in good and bad times An industry is defined as having ldquogood timesrdquo if the median firm in

that industry has a market-to-book ratio that is higher than the 67th percentile of the time-series

distribution of industry median market-to-book ratios Conversely an industry is defined as having

bad times if the median firm in that industry has a market-to-book ratio that is lower than the 33rd

percentile of the time-series distribution of industry median market-to-book ratios Panels A and B

report the percentage of firms (i) issuing net debt in excess of 5 of the value of their assets (ii) issuing

net equity in excess of 5 of the value of their assets (iii) issuing debt and repurchasing equity both in

excess of 5 of the value of their assets and (iv) issuing equity and retiring debt both in excess of 5

of the value of their assets Panel C tabulates the mean debt and equity levels and the mean changes in

debt and equity issuances The sample contains nonfinancial firms listed on the annual Compustat files

for the period from 1971 to 2009

Iss Iss Iss D Iss ENet D Net E Rep E Ret D(1) (2) (3) (4)

Panel A sorts on profitability

Badtimes

Low 145 131 01 202 170 49 02 083 173 46 04 074 182 47 03 10

High 177 51 06 10t Highfrac14Low 41 136 33 01

Goodtimes

Low 206 381 01 462 197 148 02 233 212 103 04 184 234 92 10 18

High 216 92 16 14t Highfrac14Low 23 658 138 23

Panel B sorts on firm size and profitability

Badtimes

Smallfirm

s Low 151 318 00 312 182 276 02 253 151 140 02 294 183 73 00 19

High 177 80 00 11t Highfrac14Low 10 90 NA 20

Largefirm

s Low 172 54 01 042 148 39 04 053 150 43 04 044 156 33 03 06

High 147 28 05 03t Highfrac14Low 17 32 20 02

Goodtimes Smallfirm

s Low 213 448 01 542 230 443 01 473 222 325 00 474 203 197 02 38

High 190 153 02 31t Highfrac14Low 19 230 10 39

Largefirm

s Low 203 96 03 122 208 66 09 133 232 54 16 064 215 39 19 05

High 234 36 38 04t Highfrac14Low 29 99 92 38

(continued)

1440 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

larger firms there is a tendency for small firms to issue more equitymdashat leastto the extent possibleDuring good times large profitable firms raise significant amounts of debt

as they experience an increase in the value of their equity Such firms alsoengage in share repurchases Debt issues by smaller firms are much lessaffected by their own profitability during good times and their change indebt is more or less independent of their profits However equity issuanceis primarily found among the low-profit small firms during good timesTable IX shows that empirically profitability has a time-varying impacton leverage choicesIn Table X the basic regressions for changes in debt and for equity

issuance are presented for both good times and bad times Comparinggood times with bad times we see that the effects of profits are muchstronger in good times In bad times the impacts are rather weak statisticallyThis difference is interesting and deserves further study However for ourpurposes recall that the static trade-off theory implies time-varying capital

Table IX Continued

D D BVE BVE MVE MVE EqyIss Assets

(1) (2) (3) (4) (5) (6) (7) (8)

Panel C magnitude of financing activity

Badtimes Smallfirm

s Low 40 03 58 14 305 06 22 140

2 44 05 82 07 240 67 17 180

3 45 04 94 02 176 18 05 187

4 43 09 98 02 137 04 03 188

High 36 10 114 13 209 16 04 199

Largefirm

s Low 3365 73 2523 169 2304 163 94 8972

2 3031 95 2983 35 3284 278 56 9068

3 2930 5 3360 14 3947 308 57 9225

4 2281 6 3817 19 5213 442 7 8902

High 1790 62 4767 13 8799 58 31 9573

Goodtimes

Smallfirm

s Low 24 06 38 05 339 23 39 90

2 26 08 48 01 301 04 25 104

3 30 09 60 04 246 03 17 125

4 28 07 66 08 206 18 12 130

High 23 06 85 16 261 13 08 141

Largefirm

s Low 4532 36 3477 103 6557 260 53 12465

2 4580 58 3527 139 6911 421 23 12522

3 2952 111 3485 2 7586 279 32 9535

4 2752 124 3939 72 10686 33 105 10226

High 2429 228 4819 377 18565 268 222 10830

PROFITSndashLEVERAGE PUZZLEREVISITED 1441

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

structure choices even when the target is time invariant Empirically we doobserve time-varying choices

8 Is Scaling Important

In the preceding section we provided results for unscaled debt and equityissues This seems appropriate to us However it involves making twochanges to the standard regression We would like to determine if both

Table X Debt and equity changes in good and bad timesmdashregressions

The sample contains nonfinancial firms listed on the annual Compustat files for the period

from 1971 to 2009 The table presents estimates from regressions of change in debt (D)and net equity issuances (EqyIss) in good and bad times The explanatory variables includechange in profits lagged change in profits and lagged change in assets The regressions

include both firm and year fixed effects An industry is defined as having ldquogood timesrdquo ifthe median firm in that industry has a market-to-book ratio that is higher than the 67thpercentile of the time-series distribution of industry median market-to-book ratios

Conversely an industry is defined as having bad times if the median firm in thatindustry has a market-to-book ratio that is lower than the 33rd percentile of the time-series distribution of industry median market-to-book ratios The reported t-statistics arecorrected for clustering at the firm level

aMean significant at the 1 level

bMean significant at the 5 level

cMean significant at the 10 level

Good times Bad times

D EqyIss D EqyIss

(1) (2) (3) (4) (5) (6) (7) (8)

Profits 087a 091a 006b 009a 060 065 003 003b

(30) (29) (20) (28) (15) (16) (16) (20)

Profitst1 029c 015a 034c 001

(19) (37) (17) (04)

Assetst1 000 002b 000 001

(02) (22) (00) (07)

Constant 1337 3989 8691a 10001a 1351 1851 4099a 4055a

(03) (09) (55) (37) (09) (13) (80) (78)

Year FE Yes Yes Yes Yes Yes Yes Yes Yes

Firm FE Yes Yes Yes Yes Yes Yes Yes Yes

R2 0071 0080 0018 0045 0178 0212 0034 0041

N 74827 68729 74891 68787 26965 26013 26971 26018

1442 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

changes are crucial Accordingly in this section we take a step back toconsider prior literature in which variables are scaled The questions arehow does scaling affect the results and whyTable XI examines alternative scalings Columns (1) and (2) scale debt and

equity issuances by the total firm net issuance Columns (3) and (4) scaledebt and equity issuances by the sum of debt and book equity Columns (5)and (6) consider the change in debt scaled by the sum of debt and bookequity Columns (7) and (8) consider the change in leverage due to externalfinancing Since we are examining changes we use the changes in thestandard regression factors as controls The primary interest is in the coef-ficient on the change in profitsWe find that a change in profits is associated with an increase in debt

issues But the sign reverses when we scale the debt issues by the sum of debtplus book equity This is telling What is really driving the results is thechange in equity in the denominator not the change in debt in the numer-ator This result is further substantiated in Columns (5) and (6) where thechange in debt to capital ratio is used instead of debt issuanceTo further explore the impact of active issuance relative to passive

experiencing of changes we decompose leverage by separating out theeffect of debt and equity issues from those due to changes in retainedearnings Thus we define leverage change due to external financing as

LevEF frac14Dt1 thorn d

Dt1 thorn Et1 thorn dthorn e

Dt1

Dt1 thorn Et1 eth1THORN

where D is debt E is book value of equity d is debt issues net of retirementsand e is equity issues net of repurchases Garvey and Hanka (1999) use asimilar measure in their study This measure directly captures the netleverage effect of debt and equity issuancesColumns (7) and (8) report results from regressions of the net leverage

effect of debt and equity issues on changes in profitability after controllingfor other leverage factors in differences and year indicator variables Theeffect of profitability on leverage change due to external financing is positiveand statistically significant This comparison suggests that changes in totaldebt and equity are only partly a result of debt and equity issuance decisionsOther balance sheet adjustments complicate the inferences from leverageratio regressionsFor most of our analysis we follow the literature in treating debt as homo-

genous It is of course potentially interesting to ask what happens when wedistinguish between short-term debt and long-term debt Which part of thefirmrsquos debt maturity structure drives the under-adjustment Table XII

PROFITSndashLEVERAGE PUZZLEREVISITED 1443

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

provides results from such decompositions Alternative scalings areconsidered The main results of interest are that the distinction betweenshort- and long-term debt is important Short-term debt issuance is largelyindependent of changes in profits The negative sign on profit change stemsfrom the action of long-term debt This seems intuitively reasonable

Table XI Scaled issuances

The sample contains nonfinancial firms listed on the annual Compustat files for the period

from 1971 to 2009 The table presents estimates from regressions of debt issuance to totalissuance (in columns (1) and (2)) debt issuance to total capitalization (in columns (3) and(4)) change in debt to capitalization ratio (in columns (5) and (6)) and change in leverage

ratio considering only external financing (in columns (7) and (8)) The change in leverageratio considering only external financing is defined as

LevEF frac14Dt1 thorn d

Dt1 thorn Et1 thorn dthorn e

Dt1

Dt1 thorn Et1

where D is debt E is book value of equity d is debt issues net of retirements and e is equity

issues net of repurchases The regressions include lagged leverage factors and year indicatorvariables Some of the specifications include firm fixed effects The reported t-statistics arecorrected for clustering at the firm level

aMean significant at the 1 level

bMean significant at the 5 level

cMean significant at the 10 level

DebtIssTotIss

DebtIssethDthornETHORN D

DthornE LevEF

(1) (2) (3) (4) (5) (6) (7) (8)

Profitt1 004a 000 002a 001c 003a 002 001b 002a

(35) (03) (34) (18) (35) (14) (23) (26)

IndLevt1 004c 001 001 001c 000 000 000 000

(19) (06) (15) (19) (02) (04) (03) (05)

Market=bookt1 001a 001a 000a 000a 000a 000a 000 000c

(47) (85) (68) (60) (55) (36) (05) (18)

Tangt1 004 005 013a 011a 011a 006a 003a 003a

(12) (15) (144) (112) (79) (38) (44) (42)

Assetst1 008a 004a 006a 004a 002a 001a 003a 002a

(97) (45) (283) (161) (53) (35) (147) (101)

Constant 079a 074a 002a 003a 000 000 000b 002a

(497) (442) (91) (111) (00) (05) (22) (93)

Year FE Yes Yes Yes Yes Yes Yes Yes Yes

Firm FE No Yes No Yes No Yes No Yes

R2 0015 0098 0022 0076 0009 0070 0012 0086

N 133905 133905 140552 140552 138229 138229 140572 140572

1444 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

inasmuch as short-term debt is probably more associated with day-to-dayoperations rather than longer term planning considerationsIn Table XIII we examine change in leverage ( D

DthornE) and change inleverage due to external financing (LEVEF) during good and bad timesseparately The traditional negative effect of a change in profits on achange in leverage is found both in good times and in bad timesHowever when we decompose the impact in order to isolate the externalcomponent as before the result disappears In good times the externalfinancing component of a change in profits is weak but still positiveDuring bad times it is not statistically significant This implies that distin-guishing between good and bad times does not fundamentally alter the basic

Table XII Short- and long-term debt issuances

The sample contains nonfinancial firms listed on the annual Compustat files for the period

from 1971 to 2009 The table presents estimates from regressions of short-term debtissuance to total issuance (in columns (1) and (2)) long-term debt issuance to totalissuance (in columns (3) and (4)) short-term debt issuance to total capitalization (in

columns (5) and (6)) and long-term debt to total capitalization (in columns (7) and (8))Some of the specifications include firm fixed effects The reported t-statistics are correctedfor clustering at the firm level

aMean significant at the 1 level

bMean significant at the 5 level

cMean significant at the 10 level

STDebtIssTotIss

LTDebtIssTotIss

STDebtIssethDthornETHORN

LTDebtIssethDthornETHORN

(1) (2) (3) (4) (5) (6) (7) (8)

Profitabilityt1 0002 0003 0042a 0008 0001 0001 0019a 0012a

(02) (04) (30) (05) (14) (10) (45) (28)

IndLevt1 0009 0007 0070b 0033 0001 0002c 0014a 0005

(06) (05) (26) (11) (08) (17) (26) (09)

M=Bt1 0003a 0003a 0003b 0010a 0000 0000 0003a 0002a

(48) (35) (22) (65) (10) (01) (84) (42)

Tangibilityt1 0020 0007 0011 0067 0005a 0006a 0097a 0081a

(12) (03) (03) (16) (34) (33) (108) (89)

Assetst1 0039a 0011b 0034a 0024b 0001a 0001 0046a 0020a

(89) (22) (38) (23) (36) (13) (190) (82)

Constant 0001b 0153a 0798a 0591a 0000c 0006a 0057a 0096a

(25) (144) (470) (288) (17) (124) (292) (234)

Year FE Yes Yes Yes Yes Yes Yes Yes Yes

Firm FE No Yes No Yes No Yes No Yes

R2 0025 0089 0018 0073 0030 0147 0016 0298

N 136146 136146 136140 136140 143031 143031 142847 142847

PROFITSndashLEVERAGE PUZZLEREVISITED 1445

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

Table XIII Externally financed leverage changes in good and bad times

The sample contains nonfinancial firms listed on the annual Compustat files for the period

from 1971 to 2009 The table presents estimates from regressions of change in debt tocapitalization ratio (in columns (1) (2) (5) and (6)) and change in leverage ratio consider-ing only external financing (in columns (3) (4) (7) and (8)) The change in leverage ratio

considering only external financing is defined as

LevEF frac14Dt1 thorn d

Dt1 thorn Et1 thorn dthorn e

Dt1

Dt1 thorn Et1

where D is debt E is book value of equity d is debt issues net of retirements and e is equityissues net of repurchases The regressions include lagged leverage factors and year indicator

variables The reported t-statistics are corrected for clustering at the firm level An industryis defined as having ldquogood timesrdquo if the median firm in that industry has a market-to-bookratio that is higher than the 67th percentile of the time-series distribution of industry

median market-to-book ratios Conversely an industry is defined as having ldquobad timesrdquoif the median firm in that industry has a market-to-book ratio that is lower than the 33rdpercentile of the time-series distribution of industry median market-to-book ratios

aMean significant at the 1 level

bMean significant at the 5 level

cMean significant at the 10 level

Good times Bad times

DDthornE LevEF D

DthornE LevEF

(1) (2) (3) (4) (5) (6) (7) (8)

Profitabiltyt1 0022c 0005 0016c 0025b 0073b 0041 0007 0007

(17) (03) (18) (24) (25) (11) (04) (03)

IndLevt1 0000 0011 0001 0009 0010 0018 0001 0000

(00) (09) (01) (13) (07) (10) (02) (00)

MB

t1

0004a 0003b 0000 0001 0008a 0001 0000 0000

(40) (24) (05) (08) (28) (04) (03) (03)

Tangt1 0097a 0041 0033a 0038b 0127a 0048 0059a 0033

(45) (16) (27) (25) (49) (14) (43) (16)

Assetst1 0014a 0013b 0024a 0019a 0037a 0023c 0036a 0025a

(30) (20) (84) (48) (45) (17) (88) (43)

Constant 0030b 0036 0000 0003 0028c 0014 0021 0001

(20) (16) (00) (07) (17) (06) (11) (01)

Year FE Yes Yes Yes Yes Yes Yes Yes Yes

Firm FE No Yes No Yes No Yes No Yes

R2 0006 0088 0009 0117 0026 0251 0023 0202

N 56820 56820 58200 58200 23462 23462 23511 23511

1446 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

conclusions The traditional leverage result on profits is driven by increasesin equity that are experienced and then partially but not completely offsetby firmsTo the extent that firms are engaging in active financing they are following

the prescriptions of the trade-off theory Thus we have a coherent picture ofhow firms are responding to profits Basically more profitable firms experiencean increase in the value of their equity They respond to this by issuing debtand repurchasing equity But the magnitude of the increase in equity is largerthan the magnitudes of the active steps taken in the debt and equity marketsThis means that the firmsrsquo actual actions are as predicted by the trade-offtheory But it leaves open the question of why they do not act more stronglyThe capital structure literature has long been interested in dynamics and

capital structure adjustment paths In recent years this interest has increas-ingly resulted in explicit dynamic optimizing models such as those ofGoldstein Ju and Leland (2001) Hennessy and Whited (2005) andStrebulaev (2007) These models commonly make specialized assumptionsleading some to question their generalityWe argue that both fixed transactions costs and variable transactions costs

are plausible and realistic in models that pin down firm size The simultaneouspresence of both is important Due to fixed costs firms only rebalance inter-mittently Due to variable adjustment costs the extent of the adjustment isreduced to the point at which an extra unit to adjustment is balanced againstthe extra unit transactions costs This means that both leverage increases andleverage reductions will only go part way toward a static optimumThe key point is that with both fixed and variable transactions costs we

should expect to only see intermittent leverage adjustments When leverage isadjusted it should ldquounder adjustrdquo when compared to a frictionless modelFurther discussion of transactions costs in a capital structure setting isprovided by Leary and Roberts (2005)

9 Conclusion

The connection between corporate profits and capital structure has beenvery influential in the assessment of the trade-off theory The conventionalinterpretation of the evidence has pushed the literature away from the trade-off theory and toward much more complex dynamic models and ideas As aresult it is important to make sure that the evidence has been correctlyinterpreted We show that the problem is with the interpretation of thetest results and a failure to take into account the costs associated with re-sponding to a profit shock

PROFITSndashLEVERAGE PUZZLEREVISITED 1447

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

We document that more profitable firms really do borrow more (not less)They tend to repurchase equity They experience an increase in both thebook value of equity and the market value of equity Less profitable firmsreally do tend to reduce their debt and to issue equityFirm size and market conditions also matter Larger firms tend to be more

active in the debt markets while smaller firms tend to be relatively moreactive in the equity markets During good times there is more use ofexternal financingThe usual profitsndashleverage puzzle result is primarily driven by the increase

in equity that is experienced by the more profitable firms Accordingly thepuzzle should be restated as asking why do firms not take sufficiently largeoffsetting actions to fully undo the change in equity What limits the mag-nitudes of the typical leverage response to profit shocksIn a frictionless model the partial response appears puzzling But there is

good empirical reason to believe that rebalancing entails both fixed andvariable costs and that firm size matters The rebalancing costs can befully avoided by doing nothing Accordingly the firm must decide whetherany given shock is big enough to be worth responding to If it is then thefirm must decide how big a response is called for These technical conditionsare known as ldquovalue matchingrdquo and ldquosmooth pastingrdquo Thus optimizationimplies that some shocks will be ignored Even if the shock is not ignoredthe optimal response will only partially undo the shock The magnitude ofthe leverage response must balance the marginal cost and the marginalbenefit of an extra unit of leverage Since the marginal cost of adjustingleverage is strictly positive the adjustment toward that static leverageoptimum will only go part way This is true both for leverage increasesand for leverage reductionsOne might conjecture that the variable costs would be time varying and

that such costs might have declined over time Since we do not have a fullaccounting of all of the variable costs it is of course impossible to be sureHowever there is some evidence suggesting that at least the directunderwriting fees are remarkably stable over time10

Surely nobody ought to be surprised that more profitable firms are moreinclined to repurchase equity while unprofitable firms tend not to do so Thefact that more profitable firms typically experience an increase in the value of

10 For IPOs Jay Ritter provides important summary evidence as httpbearwarringtonufleduritterIPOs2013Underwritingpdf He finds that for medium-sized deals a remark-

ably large fraction of the deals have a 7 fee year after year Whether this is due tobehavioral factors or some sort of a coordination device is not clear The extent towhich similar intertemporal stability carries over to other forms of external financemerits further study

1448 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

equity is equally natural Were it not for the contrary claims in the literature(ie the profitsndashleverage puzzle) we would argue that nobody ought to findit surprising that more profitable firms are more inclined to issue debt Thefact that leverage adjustments typically only partially offset profit shocksshould also sound quite natural These facts all fit together easily within thetrade-off theory once we take account of both fixed and variable costs ofactively adjusting capital structure

Appendix

Table AI Market leverage regressions

The table presents estimates of the regressions of market leverage on firm characteristics

The sample comes from the annual Compustat files during the period 1971ndash2009 Financialfirms are excluded The leverage is estimated as the ratio of debt over debt plus marketequity The explanatory variables Profitabilityt1

MB

t1

Tangibilityt1 and LnethAssetsTHORNt1are described in Table 1 IndMedianLevt1 is estimated as the median market leverage of allother firms in the same industry The industry is defined at the level of the four-digit SICcode All specifications include the year fixed effects The specifications in column (2)

additionally include the firm fixed effects We report t-statistics where the standard errorsare clustered at the firm level in parentheses

aSignificant at the 1 level

Quantile regressions

(1) (2) (3) (4) (5)

Profitabilityt1 013a 013a 007a 011a 020a

(275) (262) (349) (425) (538)

IndLevt1 046a 020a 032a 058a 063a

(506) (239) (1299) (1630) (1042)MB

t1

003a 001a 001a 002a 004a

(509) (303) (743) (709) (534)

Tangibilityt1 013a 018a 013a 014a 010a

(191) (181) (762) (571) (232)

LnethAssetsTHORNt1 001a 005a 001a 001a lt001a

(112) (258) (682) (342) (29)

Constant 010a 011a 005a 005a 029a

(171) (98) (173) (118) (398)

Year FE Yes Yes Yes Yes Yes

Firm FE No Yes No No No

Regression type OLS FE 25th tile Median 75thtile

Clustered SE Yes Yes No No No

R2 030 067

Pseudo R2 016 023 0201

N 158578 158578 158578 158578 158578

PROFITSndashLEVERAGE PUZZLEREVISITED 1449

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

Table AII Profitability sorts for firms issuing large amounts

Percentage of firms

Iss Ret Iss Iss Rep Iss Iss IssD IssE

D D NetD E E NetE DampE RepE RetD

(1) (2) (3) (4) (5) (6) (7) (8) (9)

Panel A 10 cutoff to define issuing or retiring activity

Allfirm

s

Low 205 150 134 260 05 257 45 00 23

2 207 175 103 77 06 75 11 01 10

3 243 196 101 50 08 47 08 01 09

4 252 186 113 50 13 48 08 02 08

High 227 161 118 60 27 57 07 04 09

t Highfrac14Low 70 38 62 755 226 768 320 90 152

Smallfirm

s Low 230 132 162 380 05 375 82 00 32

2 228 143 165 334 04 333 65 00 26

3 211 146 147 216 03 214 38 00 23

4 190 154 120 126 04 124 15 01 20

High 194 162 114 99 09 97 12 01 15

t Highfrac14Low 42 40 68 342 20 340 169 07 131

Largefirm

s Low 231 200 82 31 06 29 08 01 04

2 231 187 67 17 06 14 03 01 02

3 234 177 77 17 09 14 03 02 03

4 218 158 75 16 17 14 02 03 02

High 218 135 99 19 41 15 03 07 02

t Highfrac14Low 19 106 36 46 135 57 39 58 23

Percentage of firms

Iss Ret Iss Iss Rep Iss Both Iss D Iss E

D D Net D E E Net E DampE Rep E Ret D

(1) (2) (3) (4) (5) (6) (7) (8) (9)

Panel B 15 cutoff to define issuing or retiring activity

Allfirm

s

Low 158 101 100 230 01 228 27 00 15

2 142 115 63 57 01 56 05 00 06

3 165 128 57 32 01 31 03 00 05

4 173 118 65 35 01 34 03 00 05

High 160 102 70 46 04 44 03 00 05

t Highfrac14Low 08 06 141 738 87 746 277 24 131

Smallfirm

s Low 195 95 132 354 01 350 56 00 24

2 185 96 131 303 01 301 44 00 18

3 160 96 104 179 00 176 18 00 15

4 134 99 78 99 01 98 08 00 12

High 133 95 73 82 01 80 06 00 10

t Highfrac14Low 80 00 95 345 12 344 149 12 52

Largefirm

s Low 145 123 42 14 01 12 03 00 01

2 137 114 32 07 01 06 01 00 01

3 142 104 35 06 01 05 00 00 01

4 130 91 36 08 03 07 01 01 01

High 134 81 48 09 06 07 01 01 01

t Highfrac14Low 19 85 17 26 54 33 34 17 11

1450 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

Table AIV Debt and equity changesmdashwithout fixed effects

The sample contains nonfinancial firms listed on the annual Compustat files for the period from 1971

to 2009 The table presents estimates from regressions of change in debt (D) change in the book

value of equity (BVE) change in the market value of equity (MVE) and net equity issuances

(EqyIss) The explanatory variables include current and lagged change in operating income before

depreciation (Profits and Profitst1) and lagged change in book value of assets (Assetst1) The

regressions include year indicator variables The reported t-statistics are corrected for clustering at the

firm level

aMean significant at the 1 level

bMean significant at the 5 level

cMean significant at the 10 level

D BVE MVE EqyIss

(1) (2) (3) (4) (5) (6) (7) (8)

Profitst1 025a 024a 023a 038a 007 066a 007a 009a

(40) (32) (36) (39) (06) (29) (42) (51)Profits 051b 082a 149a 004a

(25) (33) (49) (33)Assetst1 002 005c 024a 001b

(08) (17) (34) (23)Constant 93 176a 45 71 1552a 1704a 130a 140a

(12) (28) (07) (09) (49) (49) (45) (45)Year FE Yes Yes Yes Yes Yes Yes Yes YesR2 001 005 001 011 001 005 002 002N 162056 162056 162130 162130 157550 157550 162154 162154

Table AIII Median profitability sorts

The sample contains nonfinancial firms listed on the annual Compustat files for the period from 1971

to 2009 The table presents median debt and equity median changes in debt and equity and median

equity issuances (EqyIss) in millions of dollars for all firms sorted on profitability and for profitability

sorts within the smallest and largest firms The sorts are done annually

D D BVE BVE MVE MVE EqyIss Assets(1) (2) (3) (4) (5) (6) (7) (8)

Allfirm

s

Low 21 00 75 17 256 20 01 1922 186 02 449 18 598 19 00 11033 816 06 1274 03 1637 05 00 32244 787 02 1541 18 2488 03 00 3556High 273 00 1361 66 3300 14 00 2578

Smallfirm

s Low 05 00 13 07 140 14 06 402 06 00 25 08 114 12 01 563 09 00 32 05 85 07 00 674 12 00 39 01 68 03 00 79High 10 00 51 04 98 01 00 90

Largefirm

s Low 11485 381 10269 371 13227 202 03 333902 13051 393 12786 140 17571 169 05 392463 11554 287 13429 24 21623 305 00 370314 9963 209 15052 194 29424 308 00 36705High 7361 00 16380 613 49142 439 00 34976

PROFITSndashLEVERAGE PUZZLEREVISITED 1451

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

References

Auerbach A J (1985) Real determinants of corporate leverage in B M Friedman (ed)

Corporate Capital Structures in the United States University of Chicago Press Chicago

IllinoisBertrand M and Mullainathan S (2003) Enjoying the quiet life corporate governance and

managerial preferences Journal of Political Economy 111 1043ndash1075Booth L Varouj A Demirguc-Kunt A and Maksimovic V (2001) Capital structures in

developing countries Journal of Finance 56 87ndash130Brennan M J and Schwartz E S (1984) Optimal financial policy and firm valuation

Journal of Finance 39 593ndash607Cameron A C and Trivedi P K (2010) Microeconometrics using Stata Stata Press TexasConstantinides G M (1979) Multiperiod consumption and investment behavior with

convex transactions costs Management Science 25 1127ndash1137Constantinides G M and Richard S F (1978) Existence of optimal simple policies for

discounted-cost inventory and cash management in continuous time Operations Research

26 620ndash636DeAngelo H DeAngelo L and Whited T M (2011) Capital structure dynamics and

transitory debt Journal of Financial Economics 99 235ndash261Dixit A K (1993) The Art of Smooth Pasting Routledge London and New YorkFama E and French K R (2002) Testing trade-off and pecking order predictions about

dividends and debt Review of Financial Studies 15 1ndash33Faulkender M and Petersen M A (2006) Does the source of capital affect capital struc-

ture Review of Financial Studies 19 45ndash79Faulkender M Flannery M J Hankins K W and Smith J M (2012) Cash Flows and

Leverage Adjustments Journal of Financial Economics 103 632ndash646Fischer E Heinkel R and Zechner J (1989) Dynamic capital structure choice theory

and tests Journal of Finance 44 19ndash40Frank M Z and Goyal V K (2008) Trade-off and pecking order theories of

debt in B E Eckbo (ed) Handbook of Corporate Finance Empirical Corporate Finance

Vol 2 Handbook of Finance Series Chapter 12 ElsevierNorth-Holland Amsterdam pp

135ndash202

Frank M Z and Goyal V K (2009) Capital structure decisions which factors are reliably

important Financial Management 38 1ndash37

Garvey G T and Hanka G (1999) Capital structure and corporate control the effect of

antitakeover statutes on firm leverage Journal of Finance 54 519ndash546

Goldstein R Ju N and Leland H (2001) An EBIT-based model of dynamic capital

structure Journal of Business 74 483ndash512

Graham J R and Tucker A L (2006) Tax shelters and corporate debt policy Journal of

Financial Economics 81 563ndash594

Hackbarth D and Mauer D C (2012) Optimal priority structure capital structure and

investment Review of Financial Studies 25 747ndash796

Halling M Yu J and Zechner J (2012) Leverage dynamics over the business cycle

Working paper University of Utah University of New South Wales and Vienna

University of Economics and Business AdministrationHennessy C A and Whited T A (2005) Debt dynamics Journal of Finance 60 1129ndash1165

Huang R and Ritter J R (2009) Testing theories of capital structure and estimating the

speed of adjustment Journal of Financial and Qualitative Analysis 44 237ndash271

1452 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

Hovakimian A Opler T and Titman S (2001) The debt-equity choice Journal of

Financial and Quantitative Analysis 36 1ndash24Jung K Kim Y and Stulz R M (1996) Timing investment opportunities managerial

discretion and the security issue decisions Journal of Financial Economics 42 159ndash185Kane A Marcus A J and McDonald R L (1984) How big is the tax advantage to

debt Journal of Finance 39 841ndash853Kayhan A and Titman S (2007) Firmsrsquo histories and their capital structures Journal of

Financial Economics 83 1ndash32Leary M T (2009) Bank loan supply lender choice and corporate capital structure

Journal of Finance 64 1143ndash1185Leary M T and Roberts M R (2005) Do firms rebalance their capital structures

Journal of Finance 60 2575ndash2619Leland H E (1994) Corporate debt value bond covenants and optimal capital structure

Journal of finance 49 1213ndash1252Lemmon M L Roberts M R and Zender J F (2008) Back to the beginning persist-

ence and the cross-section of corporate capital structure Journal of Finance 63 1575ndash1608Long M S and Malitz I B (1985) The investment-financing nexus some empirical

evidence Midland Corporate Finance Journal 3 53ndash59Mackie-Mason J K (1990) Do taxes affect corporate financing decisions Journal of

Finance 45 1471ndash1493Morellec E Nikolov B and Schurhoff N (2012) Corporate governance and capital

structure dynamics Journal of Finance 67 803ndash848Myers S C (1984) The capital structure puzzle Journal of Finance 39 575ndash592Myers S C (1993) Still searching for optimal capital structure Journal of Applied

Corporate Finance 61 4ndash14Rajan R and Zingales L (1995) What do we know about capital structure some evidence

from international data Journal of Finance 50 1421ndash1460Strebulaev I A (2007) Do tests of capital structure theory mean what they say Journal of

Finance 62 1747ndash1787

Strebulaev I A and Whited T M (2012) Dynamic Models and Structural Estimation in

Corporate Finance Now Publisher Boston and Delft

Titman S and Wessels R (1988) The determinants of capital structure choice Journal of

Finance 43 1ndash21

Welch I (2004) Capital structure and stock returns Journal of Political Economy 112

106ndash131

Welch I (2011) Two common problems in capital structure research the financial-debt-

to-asset ratio and issuing activity versus leverage changes International Review of Finance

11 1ndash17

PROFITSndashLEVERAGE PUZZLEREVISITED 1453

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

is mostly independent of firm profitability In the aggregate there is a smalldifference but it is sensitive to sorting on firm size Similarly in Column (2)there is only a weak relation between the profitability and the likelihood thatthe firm retires debt To a first approximation the fraction of firms issuing orrepurchasing debt is largely independent of profits Large firms are somewhatmore likely than small firms to issue debt and also to retire debtUnlike debt issues the probability of issuing equity is strongly related to

profitability as can be seen in Column (4) Low-profit firms are much morelikely to issue equity than are high-profit firms In the lowest profit quintile301 of the firms issued equity In the highest profit quintile it was only92 The lower panel reports separate sorts for small and large firms Similarpatterns are found for both groups of firms that is the likelihood of equityissuance declines as profitability increases The differences between the quin-tiles are quite stark for the group of small firms All quintiles of small firmsissue equity much more frequently than the corresponding large firm quintilesHigh-profit firms are much more likely to repurchase equity as shown in

Column (5) In the overall sample only 13 of low-profit firmsrepurchased equity compared with 82 of the high-profit firms Whenwe separate the small and large firms again a difference emerges Ingeneral small firms do not repurchase much equity while a greaterfraction of the large firms do soGiven these facts it is natural that low profitability firms are much more

likely to be net issuers of equity than are high profitability firms as shown inColumn (6) Among the firms in the lowest quintiles of profits roughly 30of firms issue net equity that exceeds 5 percent of their assets In contrastfor the most profitable quintile only about 8 issue net equityFirm size is an important variable in the recent literaturemdashit is often used

as a proxy for access to capital markets (as in Faulkender and Petersen(2006) and Leary (2009)) Small firms are bank-dependent risky and infor-mationally opaque They have restricted access to public debt markets andconsequently face more severe supply constraints in their ability to issuedebt Thus we expect small firms to be more sluggish in adjusting theirdebt and equity in response to shocks to profitability Large firms incontrast have much easier access to public debt markets and they facefewer obstacles in accessing securities marketsTo examine how size and profitability interact the bottom part of Table IV

examines profitability sorts for the smallest and largest firms We first sortfirms annually by firm size and then within each size quintile we sort them onprofitability We do these two-way sorts to ensure that we have similarnumber of firms in profitability subgroups for both small and large firmsamples Unconditionally small firms are generally less profitable while

PROFITSndashLEVERAGE PUZZLEREVISITED 1431

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

large firms are relatively more profitable The two-way sorts allow us to havea uniform distribution of firms in various profitability quintiles for both smalland large firm samples The bottom part of Table IV reports results for thesmallest and the largest size quintiles Among the small firms there is littlerelation between profitability and the likelihood of issuing debt Howeverlarger firms exhibit a small increase in the likelihood of net issuance of debtwith increasing profitabilityIn contrast to debt the effects of profitability on equity issuance and

repurchases are much more consistent across size quintiles As profitabilityincreases firms are generally less likely to issue equity and more likely torepurchase it Across the two extreme size groups we note a striking differ-ence between the proportion of firms issuing equity and those issuing debtAmong low-profit small firms almost 41 are equity issuers whereas onlyabout 9 of low-profit large firms are equity issuers However regardless ofsize we note a monotonic reduction in the likelihood of issuing equity asprofitability increasesColumn (7) reports the fraction of firms that issue both debt and equity in

excess of 5 of assets both as a function of profits in the upper panel andas a function of firm size and profits in the lower panel In the upper panelwe see that low-profit firms are much more likely to issue both debt andequity (71) while the high-profit firms are less likely to do so (17) Thelower panel shows that small firms are much more likely to be issuers of bothin contrast to large firms In both firm size categories the profit pattern is thesame with low profits being more frequent issuers of both simultaneouslyImportantly Column (8) shows that the likelihood of issuing debt and

simultaneously repurchasing equity increases with profitability Converselyin Column (9) we find that the likelihood of doing the reverse that is issuingequity and retiring debt declines with profitability The effects of sorting onfirm size and firm profitability mirror those reported for all firms Overallthe results show that low-profit firms are less likely to issue debt and repur-chase equity they are instead more likely to issue equity and retire debtFirms with high profitability exhibit the reverse patternIn Appendix Table AII we present results from profitability (and size)

quintiles for large issuances defined using the 10 cutoff (in Panel A) andthe 15 cutoff (in Panel B) The results from these panels are largely con-sistent with those described above

52 MAGNITUDES OF ISSUANCES AND EQUITY CHANGES

In the previous section we considered the probability of having a nontriviallevel of debt or equity activity The next question is how large are the dollar

1432 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

values involved In Table V we sort firms according to profits and thentabulate the levels and changes in both debt and equity We do this firstfor all firms and then for small versus large firmsFor the analysis on all firms we observe in Column (1) that debt peaks at

the middle of the distribution This is because firms with medium profitabil-ity are also the largest as seen in Column (8) which reports average assetvalues for different profitability quintiles When we sort first by firm size andthen examine firms with different levels of profitability within size quintileswe find that for small firms debt is roughly independent of profits while forlarge firms less profitable firms have a higher level of debtAs expected Columns (3) and (5) show that more profitable firms have

higher equity values Columns (2) and (7) consider the relationship betweenissuances of debt and equity and firm profitability Debt issuances are sig-nificantly larger for more profitable firms In contrast equity issuances arelarger for low-profit firms Small low-profit firms issue substantially lessequity compared with small high-profit firms Among large firms the low-profit ones tend to issue equity while the high-profit ones tend to repurchaseequity The fact that more profitable firms issue debt and repurchase equitywhile the least profitable firms retire debt and issue equity is consistent withthe predicted relation between profitability and financing decisions under thetrade-off theoryWe also examine two-way sorts by size and profitability and report results

for the smallest and largest quintile of firms For the smallest quintile offirms the change in debt is largely unrelated to profits But for large firmsthere is a positive relation between profits and debt issuances High-profitfirms have a big positive change in debt Low-profit large firms have anegative change in debtColumns (4) and (6) provide an explanation of why the leverage ratio

regression results contradict those from the basic profitability sorts pre-sented here As we can see profitability indirectly affects leverage ratiosby increasing equity values Changes in both the book value of equity andthe market value of equity are positive and large for highly profitable firmsIn contrast these changes are negative for less profitable firms This evidenceillustrates an important issue concerning the use of leverage ratios Suchratios are often interpreted as essentially reflecting the use of debt by thefirm This interpretation while common is empirically misleadingFor the typical firm the change in the value of equity is larger than the

change in debt For example in the third profit quintile for large firmscolumn (7) the mean equity issue is just $9 million but the change in themarket value of equity is $242 million At the same time the mean change indebt is $27 million This suggests that a fair bit of the observed variation in

PROFITSndashLEVERAGE PUZZLEREVISITED 1433

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

the leverage ratios is primarily driven by the changes in the market value ofequity in the denominator rather than by the changes in debt in the numer-ator Since equity issues are often small this implies that the variation in theleverage ratio is primarily driven by internal operations rather than byexternal financing actions This again points to the fact that leverageratios can provide a misleading account of actual patterns in the dataThere is always a potential concern that the averages may be misleading

due to the impact of outliers To address this concern in AppendixTable AIII we present median values of the profitability sorts Theseresults generally reinforce the findings in Table VTable VI again considers the magnitudes of financing activity but this

time the issues are scaled by a measure of firm size That way it is easier to

Table V Magnitude of financing activity

The sample contains nonfinancial firms listed on the annual Compustat files for the period

from 1971 to 2009 The table presents average debt and equity changes in debt and equityand equity issuances (in millions of dollars) for firms sorted on profitability within sizeclasses The table reports information for all firms sorted on profitability and for profit-

ability sorts within the smallest and largest firms The sorts are done annually The tablealso reports results from t-test for the difference in percentage of firms issuing or retiringsecurities in the lowest and the highest profit quintile

D D BVE BVE MVE MVE EqyIss Assets

(1) (2) (3) (4) (5) (6) (7) (8)

Allfirm

s

Low 83 5 78 10 120 5 7 234

2 605 13 509 24 721 8 17 1691

3 1227 9 1123 22 1834 3 9 3551

4 833 23 1166 7 2408 74 9 2994

High 570 53 1171 71 3817 68 36 2581

t Highfrac14Low 31 9 36 9 38 2 16 36

Smallfirm

s Low 25 04 42 01 332 28 33 97

2 29 06 55 02 278 16 21 116

3 31 07 63 01 212 05 12 127

4 30 06 70 05 179 10 08 136

High 24 06 84 14 234 17 07 142

t Highfrac14Low 04 15 153 85 73 11 201 107

Largefirm

s Low 3929 55 2990 152 4671 11 63 10846

2 3826 19 3178 95 5122 66 11 10721

3 3074 27 3498 41 6173 242 9 9697

4 2623 62 3915 45 8473 94 57 9855

High 2180 193 4642 261 14924 202 148 10171

t Highfrac14Low 94 64 108 87 246 12 169 16

1434 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

see whether large firms behave disproportionately when compared withsmall firms Column (1) shows that low-profit firms have more debtas reported in many previous studies In Columns (4) and (6) we seethat low-profit firms have a major decline in both the book value ofequity and the market value of equity The high-profit firms have thereverse experienceMore interesting are the active decisions of the firms In Column (2) we see

that in contrast to the usual impression the low-profit firms are reducingdebt while the high-profit firms are increasing it In Column (7) we see thatthe low-profit firms are also reducing leverage by issuing equity The high-profit firms issue little equity In the lower panel similar results are foundwhen we separate small and large firms Equity issuing is particularly strong

Table VI Magnitude of financing activitymdashscaled levels and changes

The sample contains nonfinancial firms listed on the annual Compustat files for the period

from 1971 to 2009 The table presents average debt and equity changes in debt and equityand equity issuances scaled by assets for firms sorted on profitability within size classes Thetable reports information for all firms sorted on profitability and for profitability sorts

within the smallest and largest firms The sorts are done annually The table also reportsresults from t-test for the difference in percentage of firms issuing or retiring securities inthe lowest and the highest profit quintile

D D BVE BVE MVE MVE EqyIss

(1) (2) (3) (4) (5) (6) (7)

Allfirm

s

Low 386 122 12 871 3427 1868 217

2 285 58 407 119 940 157 28

3 293 12 425 14 742 03 12

4 269 26 461 02 955 27 10

High 192 15 551 47 1675 74 07

t Highfrac14Low 35 49 50 47 127 42 157

Smallfirm

s Low 545 154 841 2145 9688 9128 696

2 684 49 712 1915 5014 2318 288

3 331 253 141 969 3063 954 129

4 267 181 357 175 1673 221 55

High 194 05 543 26 1758 147 41

t Highfrac14Low 49 17 51 36 88 27 99

Largefirm

s Low 350 30 306 58 511 55 09

2 346 14 335 13 523 02 04

3 353 165 332 71 711 29 01

4 280 01 414 04 909 09 05

High 227 15 477 19 1643 27 14

t Highfrac14Low 371 112 453 181 570 57 221

PROFITSndashLEVERAGE PUZZLEREVISITED 1435

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

among the small low-profit firms that are also experiencing sharp reductionsin the value of equity In column (2) we see that both for the large and for thesmall firms net debt increases are primarily found in the highest quintile offirm profitability

6 Debt and Equity Regressions

So far we have demonstrated that in our data the conventional leverageregressions have the usual signs We then demonstrated that in simple sortsof the data strikingly different active financing patterns emerge The nexttask is to reconcile these differencesTable VII presents simple regressions predicting changes in debt book

equity and market equity as well as equity issuances Consistent with thesorts we control for changes in firm size and we also include firm and yeardummies There may be a concern about the speed of a firmrsquos responsesAccordingly we consider the lagged change in profits as the main case butwe also include the contemporaneous change in profits Whether the contem-poraneous change in profits is included or not does not in fact matter Wereport panel-robust standard errors adjusted for clustering at the firm levelThe results in Table VII support the findings in the sorts Firms that have

an increase in profits in 1 year increase debt both in the same year and in thenext year Firms that have an increase in profits experience an increase inboth the book value of equity and the market value of equity Importantlyhowever firms that experience and increase in profits reduce their equityissues All these effects are statistically significant (except for the marketvalue of equity in the column 5 specification) and seem intuitively reason-able In Appendix Table AIV similar regressions to those presented aboveare reported but without firm-fixed effects We find that empirically verylittle changes whether we include firm fixed effects or leave them out Thusthe targeting behavior to the extent that it happens does not account forwhat we are reportingThe next step is to bring the sorts and the simple regressions together into

a common setting We therefore include factors in addition to profitability(i) median industry leverage (ii) the market-to-book assets ratio (iii) tangi-bility of assets and (iv) firm size (measured by the log of assets) Rajan andZingales (1995) show that these factors are related to leverage in G7countries A number of studies have used these factors to estimateleverage targets Frank and Goyal (2009) show that these factors arerobustly related to leverage in the USA In these regressions we use indica-tors for the quintile that the firm is in for each factor

1436 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

Table VIII reports these results In Columns (1) and (2) we examine debtissues and again show that profits positively affect debt issuances The effectsare large and statistically significant at the 1 level Columns (3) and (4)which examine book equity show a strong effect of profits This effect is onlyslightly reduced by the inclusion of other factors Columns (5) and (6) illus-trate that the impact of profits on the change in market equity is also robustto the inclusion of the conventional factors Columns (7) and (8) examineequity issuances Here we again find that profits have a significant negativeimpact on equity issuances As before the results are robust to the inclusionof the conventional factorsIn unreported tables we estimate debt and equity changes using quantile

regressions In terms of the signs and significance of the profitability variable

Table VII Debt and equity changes

The sample contains nonfinancial firms listed on the annual Compustat files for the period

from 1971 to 2009 The table presents fixed effect estimates from regressions of changes indebt (D) changes in the book value of equity (BVE) changes in the market value ofequity (MVE) and net equity issuances (EqyIss) The explanatory variables include

current and lagged change in operating income before depreciation (Profits andProfitst1) and lagged change in the book value of assets (Assetst1) In addition tofixed firm effects the regressions include year indicator variables The reported t-statistics

are corrected for clustering at the firm level

aMean significant at the 1 level

bMean significant at the 5 level

cMean significant at the 10 level

D BVE MVE EqyIss

(1) (2) (3) (4) (5) (6) (7) (8)

Profitst1 022a 025a 015b 035a 004 060b 005b 007a

(33) (31) (25) (35) (03) (25) (25) (53)

Profits 050b 078a 140a 003a

(24) (32) (48) (34)

Assetst1 000 007b 025a 001b

(00) (21) (35) (25)

Constant 167 283a 34 51 1145a 1182a 605a 595a

(14) (30) (04) (05) (32) (29) (69) (74)

Year FE Yes Yes Yes Yes Yes Yes Yes Yes

Firm FE Yes Yes Yes Yes Yes Yes Yes Yes

R2 001 004 001 010 001 005 002 003

N 162056 162056 162130 162130 157550 157550 162154 162154

PROFITSndashLEVERAGE PUZZLEREVISITED 1437

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

the results are qualitatively similar to the OLS results reported in Table VIIIProfitability positively affects debt issuances and negatively affects equityissuances across various quantiles Overall the results in Table VIII showthat there is nothing that is special about using sorts or regressions to explaindebt and equity issues In either case we find that more profitable firms tendto increase their debt experience an increase in the value of equity andrepurchase shares Thus the control factors are not responsible for theusual rejection of the trade-off theory It would appear that what mattersis the size of the issuing activity relative to the organic increases in equityvalue due to profits In other words what matters in the leverage regressionseems to be coming from the presence of E in D

DthornE and not from D

Table VIII Debt and equity issuances profitability and leverage factor quintiles

The sample contains nonfinancial firms listed on the annual Compustat files for the period

from 1971 to 2009 The table presents estimates from regressions of change in debt (D)change in the book value of equity (BVE) change in the market value of equity (MVE)and net equity issuances (EquityIss) The explanatory variables include lagged profitability

quintiles and lagged leverage factor quintiles In addition to fixed firm effects the regres-sions include year indicator variables The reported t-statistics are corrected for clustering atthe firm level

aMean significant at the 1 level

bMean significant at the 5 level

cMean significant at the 10 level

D BVE MVE EqyIss

(1) (2) (3) (4) (5) (6) (7) (8)

ProfitabilityQint 214a 167a 283a 175a 80 429a 53a 55a

(70) (54) (66) (59) (08) (48) (55) (62)

IndLevQint 104a 43 99 20

(29) (15) (10) (10)

Ln(Assets)Qint52 17 1026a 53a

(16) (05) (77) (35)MB

Qint158a 372a 1155a 09

(45) (56) (106) (06)

TangibilityQint 05 177 436c 64a

(01) (15) (18) (31)

Constant 167 96 656b 235 4613a 8610a 113 570a

(08) (04) (26) (04) (61) (58) (11) (31)

Year FE Yes Yes Yes Yes Yes Yes Yes Yes

Firm FE Yes Yes Yes Yes Yes Yes Yes Yes

R2 0002 0002 0003 0004 0013 0014 0013 0014

N 155587 155037 155587 155037 155587 155037 155587 155037

1438 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

7 Are Financial Market Conditions Important

According to the trade-off theory capital structure is determined by con-sidering the impact of costs and benefits of debt The time variation inthe costs and benefits of debt imply that issuance decisions would alsovary over time Furthermore market conditions affect the transactioncosts of adjusting Adverse selection is a greater problem in a cold marketthan in a hot one Accordingly it may be easier to adjust in good marketconditions Related ideas have been developed by Huang and Ritter(2009) Faulkender et al (2012) and Halling Yu and Zechner (2012)Faulkender Flannery Hankins and Smithrsquos paper examines cross-sectionalvariation in adjustment speeds including market valuations andhow they may affect adjustment speeds It is natural to think that ingood market conditions it will be less costly to issue both debt and equityand adjusting leverage to respond to profitability shocks would be mucheasierTo test the importance of market conditions we require a definition of

good times and bad times Our empirical strategy is to estimate good timesversus bad times at the four-digit industry level We define an industry ashaving ldquogood timesrdquo if the median firm in that industry has a market-to-book ratio that is higher than the 67th percentile of the time-series distribu-tion of the industry median market-to-book ratios Conversely an industryis defined as having ldquobad timesrdquo if the median firm in the industry has amarket-to-book ratio that is lower than the 33rd percentile of the time-seriesdistribution of the industry median market-to-book ratiosPanel A of Table IX tabulates issuance activity for profitability sorts in

both good times and bad times As might be expected active debt and equityissuances are larger during good times There is also somewhat more activeswaps between debt and equity during good timesPanel B reports the issuance activity for the smallest and the largest firms

during the good times and bad times Within each size quintile firms aresorted on profitability This is a three-way sort As before we find that thereis much more active use of external markets during good times Small low-profit firms are more likely to issue equity in good times than in bad timesDebt issuances are significantly higher in good times Large high-profitfirms are significantly more likely to issue debt and repurchase equity ingood times than in bad timesPanel C reports the magnitudes of the financial variables rather than the

frequencies During bad times less profitable large firms retire substantialamounts of debt and they show a tendency to issue equity Small firms donot seem to engage in similar debt reduction activities However like the

PROFITSndashLEVERAGE PUZZLEREVISITED 1439

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

Table IX Debt and equity issues in good and bad times

The table reports the frequency and magnitude of financing activity for sorts on profits for subsamples

of firms in good and bad times An industry is defined as having ldquogood timesrdquo if the median firm in

that industry has a market-to-book ratio that is higher than the 67th percentile of the time-series

distribution of industry median market-to-book ratios Conversely an industry is defined as having

bad times if the median firm in that industry has a market-to-book ratio that is lower than the 33rd

percentile of the time-series distribution of industry median market-to-book ratios Panels A and B

report the percentage of firms (i) issuing net debt in excess of 5 of the value of their assets (ii) issuing

net equity in excess of 5 of the value of their assets (iii) issuing debt and repurchasing equity both in

excess of 5 of the value of their assets and (iv) issuing equity and retiring debt both in excess of 5

of the value of their assets Panel C tabulates the mean debt and equity levels and the mean changes in

debt and equity issuances The sample contains nonfinancial firms listed on the annual Compustat files

for the period from 1971 to 2009

Iss Iss Iss D Iss ENet D Net E Rep E Ret D(1) (2) (3) (4)

Panel A sorts on profitability

Badtimes

Low 145 131 01 202 170 49 02 083 173 46 04 074 182 47 03 10

High 177 51 06 10t Highfrac14Low 41 136 33 01

Goodtimes

Low 206 381 01 462 197 148 02 233 212 103 04 184 234 92 10 18

High 216 92 16 14t Highfrac14Low 23 658 138 23

Panel B sorts on firm size and profitability

Badtimes

Smallfirm

s Low 151 318 00 312 182 276 02 253 151 140 02 294 183 73 00 19

High 177 80 00 11t Highfrac14Low 10 90 NA 20

Largefirm

s Low 172 54 01 042 148 39 04 053 150 43 04 044 156 33 03 06

High 147 28 05 03t Highfrac14Low 17 32 20 02

Goodtimes Smallfirm

s Low 213 448 01 542 230 443 01 473 222 325 00 474 203 197 02 38

High 190 153 02 31t Highfrac14Low 19 230 10 39

Largefirm

s Low 203 96 03 122 208 66 09 133 232 54 16 064 215 39 19 05

High 234 36 38 04t Highfrac14Low 29 99 92 38

(continued)

1440 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

larger firms there is a tendency for small firms to issue more equitymdashat leastto the extent possibleDuring good times large profitable firms raise significant amounts of debt

as they experience an increase in the value of their equity Such firms alsoengage in share repurchases Debt issues by smaller firms are much lessaffected by their own profitability during good times and their change indebt is more or less independent of their profits However equity issuanceis primarily found among the low-profit small firms during good timesTable IX shows that empirically profitability has a time-varying impacton leverage choicesIn Table X the basic regressions for changes in debt and for equity

issuance are presented for both good times and bad times Comparinggood times with bad times we see that the effects of profits are muchstronger in good times In bad times the impacts are rather weak statisticallyThis difference is interesting and deserves further study However for ourpurposes recall that the static trade-off theory implies time-varying capital

Table IX Continued

D D BVE BVE MVE MVE EqyIss Assets

(1) (2) (3) (4) (5) (6) (7) (8)

Panel C magnitude of financing activity

Badtimes Smallfirm

s Low 40 03 58 14 305 06 22 140

2 44 05 82 07 240 67 17 180

3 45 04 94 02 176 18 05 187

4 43 09 98 02 137 04 03 188

High 36 10 114 13 209 16 04 199

Largefirm

s Low 3365 73 2523 169 2304 163 94 8972

2 3031 95 2983 35 3284 278 56 9068

3 2930 5 3360 14 3947 308 57 9225

4 2281 6 3817 19 5213 442 7 8902

High 1790 62 4767 13 8799 58 31 9573

Goodtimes

Smallfirm

s Low 24 06 38 05 339 23 39 90

2 26 08 48 01 301 04 25 104

3 30 09 60 04 246 03 17 125

4 28 07 66 08 206 18 12 130

High 23 06 85 16 261 13 08 141

Largefirm

s Low 4532 36 3477 103 6557 260 53 12465

2 4580 58 3527 139 6911 421 23 12522

3 2952 111 3485 2 7586 279 32 9535

4 2752 124 3939 72 10686 33 105 10226

High 2429 228 4819 377 18565 268 222 10830

PROFITSndashLEVERAGE PUZZLEREVISITED 1441

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

structure choices even when the target is time invariant Empirically we doobserve time-varying choices

8 Is Scaling Important

In the preceding section we provided results for unscaled debt and equityissues This seems appropriate to us However it involves making twochanges to the standard regression We would like to determine if both

Table X Debt and equity changes in good and bad timesmdashregressions

The sample contains nonfinancial firms listed on the annual Compustat files for the period

from 1971 to 2009 The table presents estimates from regressions of change in debt (D)and net equity issuances (EqyIss) in good and bad times The explanatory variables includechange in profits lagged change in profits and lagged change in assets The regressions

include both firm and year fixed effects An industry is defined as having ldquogood timesrdquo ifthe median firm in that industry has a market-to-book ratio that is higher than the 67thpercentile of the time-series distribution of industry median market-to-book ratios

Conversely an industry is defined as having bad times if the median firm in thatindustry has a market-to-book ratio that is lower than the 33rd percentile of the time-series distribution of industry median market-to-book ratios The reported t-statistics arecorrected for clustering at the firm level

aMean significant at the 1 level

bMean significant at the 5 level

cMean significant at the 10 level

Good times Bad times

D EqyIss D EqyIss

(1) (2) (3) (4) (5) (6) (7) (8)

Profits 087a 091a 006b 009a 060 065 003 003b

(30) (29) (20) (28) (15) (16) (16) (20)

Profitst1 029c 015a 034c 001

(19) (37) (17) (04)

Assetst1 000 002b 000 001

(02) (22) (00) (07)

Constant 1337 3989 8691a 10001a 1351 1851 4099a 4055a

(03) (09) (55) (37) (09) (13) (80) (78)

Year FE Yes Yes Yes Yes Yes Yes Yes Yes

Firm FE Yes Yes Yes Yes Yes Yes Yes Yes

R2 0071 0080 0018 0045 0178 0212 0034 0041

N 74827 68729 74891 68787 26965 26013 26971 26018

1442 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

changes are crucial Accordingly in this section we take a step back toconsider prior literature in which variables are scaled The questions arehow does scaling affect the results and whyTable XI examines alternative scalings Columns (1) and (2) scale debt and

equity issuances by the total firm net issuance Columns (3) and (4) scaledebt and equity issuances by the sum of debt and book equity Columns (5)and (6) consider the change in debt scaled by the sum of debt and bookequity Columns (7) and (8) consider the change in leverage due to externalfinancing Since we are examining changes we use the changes in thestandard regression factors as controls The primary interest is in the coef-ficient on the change in profitsWe find that a change in profits is associated with an increase in debt

issues But the sign reverses when we scale the debt issues by the sum of debtplus book equity This is telling What is really driving the results is thechange in equity in the denominator not the change in debt in the numer-ator This result is further substantiated in Columns (5) and (6) where thechange in debt to capital ratio is used instead of debt issuanceTo further explore the impact of active issuance relative to passive

experiencing of changes we decompose leverage by separating out theeffect of debt and equity issues from those due to changes in retainedearnings Thus we define leverage change due to external financing as

LevEF frac14Dt1 thorn d

Dt1 thorn Et1 thorn dthorn e

Dt1

Dt1 thorn Et1 eth1THORN

where D is debt E is book value of equity d is debt issues net of retirementsand e is equity issues net of repurchases Garvey and Hanka (1999) use asimilar measure in their study This measure directly captures the netleverage effect of debt and equity issuancesColumns (7) and (8) report results from regressions of the net leverage

effect of debt and equity issues on changes in profitability after controllingfor other leverage factors in differences and year indicator variables Theeffect of profitability on leverage change due to external financing is positiveand statistically significant This comparison suggests that changes in totaldebt and equity are only partly a result of debt and equity issuance decisionsOther balance sheet adjustments complicate the inferences from leverageratio regressionsFor most of our analysis we follow the literature in treating debt as homo-

genous It is of course potentially interesting to ask what happens when wedistinguish between short-term debt and long-term debt Which part of thefirmrsquos debt maturity structure drives the under-adjustment Table XII

PROFITSndashLEVERAGE PUZZLEREVISITED 1443

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

provides results from such decompositions Alternative scalings areconsidered The main results of interest are that the distinction betweenshort- and long-term debt is important Short-term debt issuance is largelyindependent of changes in profits The negative sign on profit change stemsfrom the action of long-term debt This seems intuitively reasonable

Table XI Scaled issuances

The sample contains nonfinancial firms listed on the annual Compustat files for the period

from 1971 to 2009 The table presents estimates from regressions of debt issuance to totalissuance (in columns (1) and (2)) debt issuance to total capitalization (in columns (3) and(4)) change in debt to capitalization ratio (in columns (5) and (6)) and change in leverage

ratio considering only external financing (in columns (7) and (8)) The change in leverageratio considering only external financing is defined as

LevEF frac14Dt1 thorn d

Dt1 thorn Et1 thorn dthorn e

Dt1

Dt1 thorn Et1

where D is debt E is book value of equity d is debt issues net of retirements and e is equity

issues net of repurchases The regressions include lagged leverage factors and year indicatorvariables Some of the specifications include firm fixed effects The reported t-statistics arecorrected for clustering at the firm level

aMean significant at the 1 level

bMean significant at the 5 level

cMean significant at the 10 level

DebtIssTotIss

DebtIssethDthornETHORN D

DthornE LevEF

(1) (2) (3) (4) (5) (6) (7) (8)

Profitt1 004a 000 002a 001c 003a 002 001b 002a

(35) (03) (34) (18) (35) (14) (23) (26)

IndLevt1 004c 001 001 001c 000 000 000 000

(19) (06) (15) (19) (02) (04) (03) (05)

Market=bookt1 001a 001a 000a 000a 000a 000a 000 000c

(47) (85) (68) (60) (55) (36) (05) (18)

Tangt1 004 005 013a 011a 011a 006a 003a 003a

(12) (15) (144) (112) (79) (38) (44) (42)

Assetst1 008a 004a 006a 004a 002a 001a 003a 002a

(97) (45) (283) (161) (53) (35) (147) (101)

Constant 079a 074a 002a 003a 000 000 000b 002a

(497) (442) (91) (111) (00) (05) (22) (93)

Year FE Yes Yes Yes Yes Yes Yes Yes Yes

Firm FE No Yes No Yes No Yes No Yes

R2 0015 0098 0022 0076 0009 0070 0012 0086

N 133905 133905 140552 140552 138229 138229 140572 140572

1444 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

inasmuch as short-term debt is probably more associated with day-to-dayoperations rather than longer term planning considerationsIn Table XIII we examine change in leverage ( D

DthornE) and change inleverage due to external financing (LEVEF) during good and bad timesseparately The traditional negative effect of a change in profits on achange in leverage is found both in good times and in bad timesHowever when we decompose the impact in order to isolate the externalcomponent as before the result disappears In good times the externalfinancing component of a change in profits is weak but still positiveDuring bad times it is not statistically significant This implies that distin-guishing between good and bad times does not fundamentally alter the basic

Table XII Short- and long-term debt issuances

The sample contains nonfinancial firms listed on the annual Compustat files for the period

from 1971 to 2009 The table presents estimates from regressions of short-term debtissuance to total issuance (in columns (1) and (2)) long-term debt issuance to totalissuance (in columns (3) and (4)) short-term debt issuance to total capitalization (in

columns (5) and (6)) and long-term debt to total capitalization (in columns (7) and (8))Some of the specifications include firm fixed effects The reported t-statistics are correctedfor clustering at the firm level

aMean significant at the 1 level

bMean significant at the 5 level

cMean significant at the 10 level

STDebtIssTotIss

LTDebtIssTotIss

STDebtIssethDthornETHORN

LTDebtIssethDthornETHORN

(1) (2) (3) (4) (5) (6) (7) (8)

Profitabilityt1 0002 0003 0042a 0008 0001 0001 0019a 0012a

(02) (04) (30) (05) (14) (10) (45) (28)

IndLevt1 0009 0007 0070b 0033 0001 0002c 0014a 0005

(06) (05) (26) (11) (08) (17) (26) (09)

M=Bt1 0003a 0003a 0003b 0010a 0000 0000 0003a 0002a

(48) (35) (22) (65) (10) (01) (84) (42)

Tangibilityt1 0020 0007 0011 0067 0005a 0006a 0097a 0081a

(12) (03) (03) (16) (34) (33) (108) (89)

Assetst1 0039a 0011b 0034a 0024b 0001a 0001 0046a 0020a

(89) (22) (38) (23) (36) (13) (190) (82)

Constant 0001b 0153a 0798a 0591a 0000c 0006a 0057a 0096a

(25) (144) (470) (288) (17) (124) (292) (234)

Year FE Yes Yes Yes Yes Yes Yes Yes Yes

Firm FE No Yes No Yes No Yes No Yes

R2 0025 0089 0018 0073 0030 0147 0016 0298

N 136146 136146 136140 136140 143031 143031 142847 142847

PROFITSndashLEVERAGE PUZZLEREVISITED 1445

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

Table XIII Externally financed leverage changes in good and bad times

The sample contains nonfinancial firms listed on the annual Compustat files for the period

from 1971 to 2009 The table presents estimates from regressions of change in debt tocapitalization ratio (in columns (1) (2) (5) and (6)) and change in leverage ratio consider-ing only external financing (in columns (3) (4) (7) and (8)) The change in leverage ratio

considering only external financing is defined as

LevEF frac14Dt1 thorn d

Dt1 thorn Et1 thorn dthorn e

Dt1

Dt1 thorn Et1

where D is debt E is book value of equity d is debt issues net of retirements and e is equityissues net of repurchases The regressions include lagged leverage factors and year indicator

variables The reported t-statistics are corrected for clustering at the firm level An industryis defined as having ldquogood timesrdquo if the median firm in that industry has a market-to-bookratio that is higher than the 67th percentile of the time-series distribution of industry

median market-to-book ratios Conversely an industry is defined as having ldquobad timesrdquoif the median firm in that industry has a market-to-book ratio that is lower than the 33rdpercentile of the time-series distribution of industry median market-to-book ratios

aMean significant at the 1 level

bMean significant at the 5 level

cMean significant at the 10 level

Good times Bad times

DDthornE LevEF D

DthornE LevEF

(1) (2) (3) (4) (5) (6) (7) (8)

Profitabiltyt1 0022c 0005 0016c 0025b 0073b 0041 0007 0007

(17) (03) (18) (24) (25) (11) (04) (03)

IndLevt1 0000 0011 0001 0009 0010 0018 0001 0000

(00) (09) (01) (13) (07) (10) (02) (00)

MB

t1

0004a 0003b 0000 0001 0008a 0001 0000 0000

(40) (24) (05) (08) (28) (04) (03) (03)

Tangt1 0097a 0041 0033a 0038b 0127a 0048 0059a 0033

(45) (16) (27) (25) (49) (14) (43) (16)

Assetst1 0014a 0013b 0024a 0019a 0037a 0023c 0036a 0025a

(30) (20) (84) (48) (45) (17) (88) (43)

Constant 0030b 0036 0000 0003 0028c 0014 0021 0001

(20) (16) (00) (07) (17) (06) (11) (01)

Year FE Yes Yes Yes Yes Yes Yes Yes Yes

Firm FE No Yes No Yes No Yes No Yes

R2 0006 0088 0009 0117 0026 0251 0023 0202

N 56820 56820 58200 58200 23462 23462 23511 23511

1446 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

conclusions The traditional leverage result on profits is driven by increasesin equity that are experienced and then partially but not completely offsetby firmsTo the extent that firms are engaging in active financing they are following

the prescriptions of the trade-off theory Thus we have a coherent picture ofhow firms are responding to profits Basically more profitable firms experiencean increase in the value of their equity They respond to this by issuing debtand repurchasing equity But the magnitude of the increase in equity is largerthan the magnitudes of the active steps taken in the debt and equity marketsThis means that the firmsrsquo actual actions are as predicted by the trade-offtheory But it leaves open the question of why they do not act more stronglyThe capital structure literature has long been interested in dynamics and

capital structure adjustment paths In recent years this interest has increas-ingly resulted in explicit dynamic optimizing models such as those ofGoldstein Ju and Leland (2001) Hennessy and Whited (2005) andStrebulaev (2007) These models commonly make specialized assumptionsleading some to question their generalityWe argue that both fixed transactions costs and variable transactions costs

are plausible and realistic in models that pin down firm size The simultaneouspresence of both is important Due to fixed costs firms only rebalance inter-mittently Due to variable adjustment costs the extent of the adjustment isreduced to the point at which an extra unit to adjustment is balanced againstthe extra unit transactions costs This means that both leverage increases andleverage reductions will only go part way toward a static optimumThe key point is that with both fixed and variable transactions costs we

should expect to only see intermittent leverage adjustments When leverage isadjusted it should ldquounder adjustrdquo when compared to a frictionless modelFurther discussion of transactions costs in a capital structure setting isprovided by Leary and Roberts (2005)

9 Conclusion

The connection between corporate profits and capital structure has beenvery influential in the assessment of the trade-off theory The conventionalinterpretation of the evidence has pushed the literature away from the trade-off theory and toward much more complex dynamic models and ideas As aresult it is important to make sure that the evidence has been correctlyinterpreted We show that the problem is with the interpretation of thetest results and a failure to take into account the costs associated with re-sponding to a profit shock

PROFITSndashLEVERAGE PUZZLEREVISITED 1447

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

We document that more profitable firms really do borrow more (not less)They tend to repurchase equity They experience an increase in both thebook value of equity and the market value of equity Less profitable firmsreally do tend to reduce their debt and to issue equityFirm size and market conditions also matter Larger firms tend to be more

active in the debt markets while smaller firms tend to be relatively moreactive in the equity markets During good times there is more use ofexternal financingThe usual profitsndashleverage puzzle result is primarily driven by the increase

in equity that is experienced by the more profitable firms Accordingly thepuzzle should be restated as asking why do firms not take sufficiently largeoffsetting actions to fully undo the change in equity What limits the mag-nitudes of the typical leverage response to profit shocksIn a frictionless model the partial response appears puzzling But there is

good empirical reason to believe that rebalancing entails both fixed andvariable costs and that firm size matters The rebalancing costs can befully avoided by doing nothing Accordingly the firm must decide whetherany given shock is big enough to be worth responding to If it is then thefirm must decide how big a response is called for These technical conditionsare known as ldquovalue matchingrdquo and ldquosmooth pastingrdquo Thus optimizationimplies that some shocks will be ignored Even if the shock is not ignoredthe optimal response will only partially undo the shock The magnitude ofthe leverage response must balance the marginal cost and the marginalbenefit of an extra unit of leverage Since the marginal cost of adjustingleverage is strictly positive the adjustment toward that static leverageoptimum will only go part way This is true both for leverage increasesand for leverage reductionsOne might conjecture that the variable costs would be time varying and

that such costs might have declined over time Since we do not have a fullaccounting of all of the variable costs it is of course impossible to be sureHowever there is some evidence suggesting that at least the directunderwriting fees are remarkably stable over time10

Surely nobody ought to be surprised that more profitable firms are moreinclined to repurchase equity while unprofitable firms tend not to do so Thefact that more profitable firms typically experience an increase in the value of

10 For IPOs Jay Ritter provides important summary evidence as httpbearwarringtonufleduritterIPOs2013Underwritingpdf He finds that for medium-sized deals a remark-

ably large fraction of the deals have a 7 fee year after year Whether this is due tobehavioral factors or some sort of a coordination device is not clear The extent towhich similar intertemporal stability carries over to other forms of external financemerits further study

1448 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

equity is equally natural Were it not for the contrary claims in the literature(ie the profitsndashleverage puzzle) we would argue that nobody ought to findit surprising that more profitable firms are more inclined to issue debt Thefact that leverage adjustments typically only partially offset profit shocksshould also sound quite natural These facts all fit together easily within thetrade-off theory once we take account of both fixed and variable costs ofactively adjusting capital structure

Appendix

Table AI Market leverage regressions

The table presents estimates of the regressions of market leverage on firm characteristics

The sample comes from the annual Compustat files during the period 1971ndash2009 Financialfirms are excluded The leverage is estimated as the ratio of debt over debt plus marketequity The explanatory variables Profitabilityt1

MB

t1

Tangibilityt1 and LnethAssetsTHORNt1are described in Table 1 IndMedianLevt1 is estimated as the median market leverage of allother firms in the same industry The industry is defined at the level of the four-digit SICcode All specifications include the year fixed effects The specifications in column (2)

additionally include the firm fixed effects We report t-statistics where the standard errorsare clustered at the firm level in parentheses

aSignificant at the 1 level

Quantile regressions

(1) (2) (3) (4) (5)

Profitabilityt1 013a 013a 007a 011a 020a

(275) (262) (349) (425) (538)

IndLevt1 046a 020a 032a 058a 063a

(506) (239) (1299) (1630) (1042)MB

t1

003a 001a 001a 002a 004a

(509) (303) (743) (709) (534)

Tangibilityt1 013a 018a 013a 014a 010a

(191) (181) (762) (571) (232)

LnethAssetsTHORNt1 001a 005a 001a 001a lt001a

(112) (258) (682) (342) (29)

Constant 010a 011a 005a 005a 029a

(171) (98) (173) (118) (398)

Year FE Yes Yes Yes Yes Yes

Firm FE No Yes No No No

Regression type OLS FE 25th tile Median 75thtile

Clustered SE Yes Yes No No No

R2 030 067

Pseudo R2 016 023 0201

N 158578 158578 158578 158578 158578

PROFITSndashLEVERAGE PUZZLEREVISITED 1449

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

Table AII Profitability sorts for firms issuing large amounts

Percentage of firms

Iss Ret Iss Iss Rep Iss Iss IssD IssE

D D NetD E E NetE DampE RepE RetD

(1) (2) (3) (4) (5) (6) (7) (8) (9)

Panel A 10 cutoff to define issuing or retiring activity

Allfirm

s

Low 205 150 134 260 05 257 45 00 23

2 207 175 103 77 06 75 11 01 10

3 243 196 101 50 08 47 08 01 09

4 252 186 113 50 13 48 08 02 08

High 227 161 118 60 27 57 07 04 09

t Highfrac14Low 70 38 62 755 226 768 320 90 152

Smallfirm

s Low 230 132 162 380 05 375 82 00 32

2 228 143 165 334 04 333 65 00 26

3 211 146 147 216 03 214 38 00 23

4 190 154 120 126 04 124 15 01 20

High 194 162 114 99 09 97 12 01 15

t Highfrac14Low 42 40 68 342 20 340 169 07 131

Largefirm

s Low 231 200 82 31 06 29 08 01 04

2 231 187 67 17 06 14 03 01 02

3 234 177 77 17 09 14 03 02 03

4 218 158 75 16 17 14 02 03 02

High 218 135 99 19 41 15 03 07 02

t Highfrac14Low 19 106 36 46 135 57 39 58 23

Percentage of firms

Iss Ret Iss Iss Rep Iss Both Iss D Iss E

D D Net D E E Net E DampE Rep E Ret D

(1) (2) (3) (4) (5) (6) (7) (8) (9)

Panel B 15 cutoff to define issuing or retiring activity

Allfirm

s

Low 158 101 100 230 01 228 27 00 15

2 142 115 63 57 01 56 05 00 06

3 165 128 57 32 01 31 03 00 05

4 173 118 65 35 01 34 03 00 05

High 160 102 70 46 04 44 03 00 05

t Highfrac14Low 08 06 141 738 87 746 277 24 131

Smallfirm

s Low 195 95 132 354 01 350 56 00 24

2 185 96 131 303 01 301 44 00 18

3 160 96 104 179 00 176 18 00 15

4 134 99 78 99 01 98 08 00 12

High 133 95 73 82 01 80 06 00 10

t Highfrac14Low 80 00 95 345 12 344 149 12 52

Largefirm

s Low 145 123 42 14 01 12 03 00 01

2 137 114 32 07 01 06 01 00 01

3 142 104 35 06 01 05 00 00 01

4 130 91 36 08 03 07 01 01 01

High 134 81 48 09 06 07 01 01 01

t Highfrac14Low 19 85 17 26 54 33 34 17 11

1450 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

Table AIV Debt and equity changesmdashwithout fixed effects

The sample contains nonfinancial firms listed on the annual Compustat files for the period from 1971

to 2009 The table presents estimates from regressions of change in debt (D) change in the book

value of equity (BVE) change in the market value of equity (MVE) and net equity issuances

(EqyIss) The explanatory variables include current and lagged change in operating income before

depreciation (Profits and Profitst1) and lagged change in book value of assets (Assetst1) The

regressions include year indicator variables The reported t-statistics are corrected for clustering at the

firm level

aMean significant at the 1 level

bMean significant at the 5 level

cMean significant at the 10 level

D BVE MVE EqyIss

(1) (2) (3) (4) (5) (6) (7) (8)

Profitst1 025a 024a 023a 038a 007 066a 007a 009a

(40) (32) (36) (39) (06) (29) (42) (51)Profits 051b 082a 149a 004a

(25) (33) (49) (33)Assetst1 002 005c 024a 001b

(08) (17) (34) (23)Constant 93 176a 45 71 1552a 1704a 130a 140a

(12) (28) (07) (09) (49) (49) (45) (45)Year FE Yes Yes Yes Yes Yes Yes Yes YesR2 001 005 001 011 001 005 002 002N 162056 162056 162130 162130 157550 157550 162154 162154

Table AIII Median profitability sorts

The sample contains nonfinancial firms listed on the annual Compustat files for the period from 1971

to 2009 The table presents median debt and equity median changes in debt and equity and median

equity issuances (EqyIss) in millions of dollars for all firms sorted on profitability and for profitability

sorts within the smallest and largest firms The sorts are done annually

D D BVE BVE MVE MVE EqyIss Assets(1) (2) (3) (4) (5) (6) (7) (8)

Allfirm

s

Low 21 00 75 17 256 20 01 1922 186 02 449 18 598 19 00 11033 816 06 1274 03 1637 05 00 32244 787 02 1541 18 2488 03 00 3556High 273 00 1361 66 3300 14 00 2578

Smallfirm

s Low 05 00 13 07 140 14 06 402 06 00 25 08 114 12 01 563 09 00 32 05 85 07 00 674 12 00 39 01 68 03 00 79High 10 00 51 04 98 01 00 90

Largefirm

s Low 11485 381 10269 371 13227 202 03 333902 13051 393 12786 140 17571 169 05 392463 11554 287 13429 24 21623 305 00 370314 9963 209 15052 194 29424 308 00 36705High 7361 00 16380 613 49142 439 00 34976

PROFITSndashLEVERAGE PUZZLEREVISITED 1451

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

References

Auerbach A J (1985) Real determinants of corporate leverage in B M Friedman (ed)

Corporate Capital Structures in the United States University of Chicago Press Chicago

IllinoisBertrand M and Mullainathan S (2003) Enjoying the quiet life corporate governance and

managerial preferences Journal of Political Economy 111 1043ndash1075Booth L Varouj A Demirguc-Kunt A and Maksimovic V (2001) Capital structures in

developing countries Journal of Finance 56 87ndash130Brennan M J and Schwartz E S (1984) Optimal financial policy and firm valuation

Journal of Finance 39 593ndash607Cameron A C and Trivedi P K (2010) Microeconometrics using Stata Stata Press TexasConstantinides G M (1979) Multiperiod consumption and investment behavior with

convex transactions costs Management Science 25 1127ndash1137Constantinides G M and Richard S F (1978) Existence of optimal simple policies for

discounted-cost inventory and cash management in continuous time Operations Research

26 620ndash636DeAngelo H DeAngelo L and Whited T M (2011) Capital structure dynamics and

transitory debt Journal of Financial Economics 99 235ndash261Dixit A K (1993) The Art of Smooth Pasting Routledge London and New YorkFama E and French K R (2002) Testing trade-off and pecking order predictions about

dividends and debt Review of Financial Studies 15 1ndash33Faulkender M and Petersen M A (2006) Does the source of capital affect capital struc-

ture Review of Financial Studies 19 45ndash79Faulkender M Flannery M J Hankins K W and Smith J M (2012) Cash Flows and

Leverage Adjustments Journal of Financial Economics 103 632ndash646Fischer E Heinkel R and Zechner J (1989) Dynamic capital structure choice theory

and tests Journal of Finance 44 19ndash40Frank M Z and Goyal V K (2008) Trade-off and pecking order theories of

debt in B E Eckbo (ed) Handbook of Corporate Finance Empirical Corporate Finance

Vol 2 Handbook of Finance Series Chapter 12 ElsevierNorth-Holland Amsterdam pp

135ndash202

Frank M Z and Goyal V K (2009) Capital structure decisions which factors are reliably

important Financial Management 38 1ndash37

Garvey G T and Hanka G (1999) Capital structure and corporate control the effect of

antitakeover statutes on firm leverage Journal of Finance 54 519ndash546

Goldstein R Ju N and Leland H (2001) An EBIT-based model of dynamic capital

structure Journal of Business 74 483ndash512

Graham J R and Tucker A L (2006) Tax shelters and corporate debt policy Journal of

Financial Economics 81 563ndash594

Hackbarth D and Mauer D C (2012) Optimal priority structure capital structure and

investment Review of Financial Studies 25 747ndash796

Halling M Yu J and Zechner J (2012) Leverage dynamics over the business cycle

Working paper University of Utah University of New South Wales and Vienna

University of Economics and Business AdministrationHennessy C A and Whited T A (2005) Debt dynamics Journal of Finance 60 1129ndash1165

Huang R and Ritter J R (2009) Testing theories of capital structure and estimating the

speed of adjustment Journal of Financial and Qualitative Analysis 44 237ndash271

1452 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

Hovakimian A Opler T and Titman S (2001) The debt-equity choice Journal of

Financial and Quantitative Analysis 36 1ndash24Jung K Kim Y and Stulz R M (1996) Timing investment opportunities managerial

discretion and the security issue decisions Journal of Financial Economics 42 159ndash185Kane A Marcus A J and McDonald R L (1984) How big is the tax advantage to

debt Journal of Finance 39 841ndash853Kayhan A and Titman S (2007) Firmsrsquo histories and their capital structures Journal of

Financial Economics 83 1ndash32Leary M T (2009) Bank loan supply lender choice and corporate capital structure

Journal of Finance 64 1143ndash1185Leary M T and Roberts M R (2005) Do firms rebalance their capital structures

Journal of Finance 60 2575ndash2619Leland H E (1994) Corporate debt value bond covenants and optimal capital structure

Journal of finance 49 1213ndash1252Lemmon M L Roberts M R and Zender J F (2008) Back to the beginning persist-

ence and the cross-section of corporate capital structure Journal of Finance 63 1575ndash1608Long M S and Malitz I B (1985) The investment-financing nexus some empirical

evidence Midland Corporate Finance Journal 3 53ndash59Mackie-Mason J K (1990) Do taxes affect corporate financing decisions Journal of

Finance 45 1471ndash1493Morellec E Nikolov B and Schurhoff N (2012) Corporate governance and capital

structure dynamics Journal of Finance 67 803ndash848Myers S C (1984) The capital structure puzzle Journal of Finance 39 575ndash592Myers S C (1993) Still searching for optimal capital structure Journal of Applied

Corporate Finance 61 4ndash14Rajan R and Zingales L (1995) What do we know about capital structure some evidence

from international data Journal of Finance 50 1421ndash1460Strebulaev I A (2007) Do tests of capital structure theory mean what they say Journal of

Finance 62 1747ndash1787

Strebulaev I A and Whited T M (2012) Dynamic Models and Structural Estimation in

Corporate Finance Now Publisher Boston and Delft

Titman S and Wessels R (1988) The determinants of capital structure choice Journal of

Finance 43 1ndash21

Welch I (2004) Capital structure and stock returns Journal of Political Economy 112

106ndash131

Welch I (2011) Two common problems in capital structure research the financial-debt-

to-asset ratio and issuing activity versus leverage changes International Review of Finance

11 1ndash17

PROFITSndashLEVERAGE PUZZLEREVISITED 1453

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

large firms are relatively more profitable The two-way sorts allow us to havea uniform distribution of firms in various profitability quintiles for both smalland large firm samples The bottom part of Table IV reports results for thesmallest and the largest size quintiles Among the small firms there is littlerelation between profitability and the likelihood of issuing debt Howeverlarger firms exhibit a small increase in the likelihood of net issuance of debtwith increasing profitabilityIn contrast to debt the effects of profitability on equity issuance and

repurchases are much more consistent across size quintiles As profitabilityincreases firms are generally less likely to issue equity and more likely torepurchase it Across the two extreme size groups we note a striking differ-ence between the proportion of firms issuing equity and those issuing debtAmong low-profit small firms almost 41 are equity issuers whereas onlyabout 9 of low-profit large firms are equity issuers However regardless ofsize we note a monotonic reduction in the likelihood of issuing equity asprofitability increasesColumn (7) reports the fraction of firms that issue both debt and equity in

excess of 5 of assets both as a function of profits in the upper panel andas a function of firm size and profits in the lower panel In the upper panelwe see that low-profit firms are much more likely to issue both debt andequity (71) while the high-profit firms are less likely to do so (17) Thelower panel shows that small firms are much more likely to be issuers of bothin contrast to large firms In both firm size categories the profit pattern is thesame with low profits being more frequent issuers of both simultaneouslyImportantly Column (8) shows that the likelihood of issuing debt and

simultaneously repurchasing equity increases with profitability Converselyin Column (9) we find that the likelihood of doing the reverse that is issuingequity and retiring debt declines with profitability The effects of sorting onfirm size and firm profitability mirror those reported for all firms Overallthe results show that low-profit firms are less likely to issue debt and repur-chase equity they are instead more likely to issue equity and retire debtFirms with high profitability exhibit the reverse patternIn Appendix Table AII we present results from profitability (and size)

quintiles for large issuances defined using the 10 cutoff (in Panel A) andthe 15 cutoff (in Panel B) The results from these panels are largely con-sistent with those described above

52 MAGNITUDES OF ISSUANCES AND EQUITY CHANGES

In the previous section we considered the probability of having a nontriviallevel of debt or equity activity The next question is how large are the dollar

1432 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

values involved In Table V we sort firms according to profits and thentabulate the levels and changes in both debt and equity We do this firstfor all firms and then for small versus large firmsFor the analysis on all firms we observe in Column (1) that debt peaks at

the middle of the distribution This is because firms with medium profitabil-ity are also the largest as seen in Column (8) which reports average assetvalues for different profitability quintiles When we sort first by firm size andthen examine firms with different levels of profitability within size quintileswe find that for small firms debt is roughly independent of profits while forlarge firms less profitable firms have a higher level of debtAs expected Columns (3) and (5) show that more profitable firms have

higher equity values Columns (2) and (7) consider the relationship betweenissuances of debt and equity and firm profitability Debt issuances are sig-nificantly larger for more profitable firms In contrast equity issuances arelarger for low-profit firms Small low-profit firms issue substantially lessequity compared with small high-profit firms Among large firms the low-profit ones tend to issue equity while the high-profit ones tend to repurchaseequity The fact that more profitable firms issue debt and repurchase equitywhile the least profitable firms retire debt and issue equity is consistent withthe predicted relation between profitability and financing decisions under thetrade-off theoryWe also examine two-way sorts by size and profitability and report results

for the smallest and largest quintile of firms For the smallest quintile offirms the change in debt is largely unrelated to profits But for large firmsthere is a positive relation between profits and debt issuances High-profitfirms have a big positive change in debt Low-profit large firms have anegative change in debtColumns (4) and (6) provide an explanation of why the leverage ratio

regression results contradict those from the basic profitability sorts pre-sented here As we can see profitability indirectly affects leverage ratiosby increasing equity values Changes in both the book value of equity andthe market value of equity are positive and large for highly profitable firmsIn contrast these changes are negative for less profitable firms This evidenceillustrates an important issue concerning the use of leverage ratios Suchratios are often interpreted as essentially reflecting the use of debt by thefirm This interpretation while common is empirically misleadingFor the typical firm the change in the value of equity is larger than the

change in debt For example in the third profit quintile for large firmscolumn (7) the mean equity issue is just $9 million but the change in themarket value of equity is $242 million At the same time the mean change indebt is $27 million This suggests that a fair bit of the observed variation in

PROFITSndashLEVERAGE PUZZLEREVISITED 1433

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

the leverage ratios is primarily driven by the changes in the market value ofequity in the denominator rather than by the changes in debt in the numer-ator Since equity issues are often small this implies that the variation in theleverage ratio is primarily driven by internal operations rather than byexternal financing actions This again points to the fact that leverageratios can provide a misleading account of actual patterns in the dataThere is always a potential concern that the averages may be misleading

due to the impact of outliers To address this concern in AppendixTable AIII we present median values of the profitability sorts Theseresults generally reinforce the findings in Table VTable VI again considers the magnitudes of financing activity but this

time the issues are scaled by a measure of firm size That way it is easier to

Table V Magnitude of financing activity

The sample contains nonfinancial firms listed on the annual Compustat files for the period

from 1971 to 2009 The table presents average debt and equity changes in debt and equityand equity issuances (in millions of dollars) for firms sorted on profitability within sizeclasses The table reports information for all firms sorted on profitability and for profit-

ability sorts within the smallest and largest firms The sorts are done annually The tablealso reports results from t-test for the difference in percentage of firms issuing or retiringsecurities in the lowest and the highest profit quintile

D D BVE BVE MVE MVE EqyIss Assets

(1) (2) (3) (4) (5) (6) (7) (8)

Allfirm

s

Low 83 5 78 10 120 5 7 234

2 605 13 509 24 721 8 17 1691

3 1227 9 1123 22 1834 3 9 3551

4 833 23 1166 7 2408 74 9 2994

High 570 53 1171 71 3817 68 36 2581

t Highfrac14Low 31 9 36 9 38 2 16 36

Smallfirm

s Low 25 04 42 01 332 28 33 97

2 29 06 55 02 278 16 21 116

3 31 07 63 01 212 05 12 127

4 30 06 70 05 179 10 08 136

High 24 06 84 14 234 17 07 142

t Highfrac14Low 04 15 153 85 73 11 201 107

Largefirm

s Low 3929 55 2990 152 4671 11 63 10846

2 3826 19 3178 95 5122 66 11 10721

3 3074 27 3498 41 6173 242 9 9697

4 2623 62 3915 45 8473 94 57 9855

High 2180 193 4642 261 14924 202 148 10171

t Highfrac14Low 94 64 108 87 246 12 169 16

1434 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

see whether large firms behave disproportionately when compared withsmall firms Column (1) shows that low-profit firms have more debtas reported in many previous studies In Columns (4) and (6) we seethat low-profit firms have a major decline in both the book value ofequity and the market value of equity The high-profit firms have thereverse experienceMore interesting are the active decisions of the firms In Column (2) we see

that in contrast to the usual impression the low-profit firms are reducingdebt while the high-profit firms are increasing it In Column (7) we see thatthe low-profit firms are also reducing leverage by issuing equity The high-profit firms issue little equity In the lower panel similar results are foundwhen we separate small and large firms Equity issuing is particularly strong

Table VI Magnitude of financing activitymdashscaled levels and changes

The sample contains nonfinancial firms listed on the annual Compustat files for the period

from 1971 to 2009 The table presents average debt and equity changes in debt and equityand equity issuances scaled by assets for firms sorted on profitability within size classes Thetable reports information for all firms sorted on profitability and for profitability sorts

within the smallest and largest firms The sorts are done annually The table also reportsresults from t-test for the difference in percentage of firms issuing or retiring securities inthe lowest and the highest profit quintile

D D BVE BVE MVE MVE EqyIss

(1) (2) (3) (4) (5) (6) (7)

Allfirm

s

Low 386 122 12 871 3427 1868 217

2 285 58 407 119 940 157 28

3 293 12 425 14 742 03 12

4 269 26 461 02 955 27 10

High 192 15 551 47 1675 74 07

t Highfrac14Low 35 49 50 47 127 42 157

Smallfirm

s Low 545 154 841 2145 9688 9128 696

2 684 49 712 1915 5014 2318 288

3 331 253 141 969 3063 954 129

4 267 181 357 175 1673 221 55

High 194 05 543 26 1758 147 41

t Highfrac14Low 49 17 51 36 88 27 99

Largefirm

s Low 350 30 306 58 511 55 09

2 346 14 335 13 523 02 04

3 353 165 332 71 711 29 01

4 280 01 414 04 909 09 05

High 227 15 477 19 1643 27 14

t Highfrac14Low 371 112 453 181 570 57 221

PROFITSndashLEVERAGE PUZZLEREVISITED 1435

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

among the small low-profit firms that are also experiencing sharp reductionsin the value of equity In column (2) we see that both for the large and for thesmall firms net debt increases are primarily found in the highest quintile offirm profitability

6 Debt and Equity Regressions

So far we have demonstrated that in our data the conventional leverageregressions have the usual signs We then demonstrated that in simple sortsof the data strikingly different active financing patterns emerge The nexttask is to reconcile these differencesTable VII presents simple regressions predicting changes in debt book

equity and market equity as well as equity issuances Consistent with thesorts we control for changes in firm size and we also include firm and yeardummies There may be a concern about the speed of a firmrsquos responsesAccordingly we consider the lagged change in profits as the main case butwe also include the contemporaneous change in profits Whether the contem-poraneous change in profits is included or not does not in fact matter Wereport panel-robust standard errors adjusted for clustering at the firm levelThe results in Table VII support the findings in the sorts Firms that have

an increase in profits in 1 year increase debt both in the same year and in thenext year Firms that have an increase in profits experience an increase inboth the book value of equity and the market value of equity Importantlyhowever firms that experience and increase in profits reduce their equityissues All these effects are statistically significant (except for the marketvalue of equity in the column 5 specification) and seem intuitively reason-able In Appendix Table AIV similar regressions to those presented aboveare reported but without firm-fixed effects We find that empirically verylittle changes whether we include firm fixed effects or leave them out Thusthe targeting behavior to the extent that it happens does not account forwhat we are reportingThe next step is to bring the sorts and the simple regressions together into

a common setting We therefore include factors in addition to profitability(i) median industry leverage (ii) the market-to-book assets ratio (iii) tangi-bility of assets and (iv) firm size (measured by the log of assets) Rajan andZingales (1995) show that these factors are related to leverage in G7countries A number of studies have used these factors to estimateleverage targets Frank and Goyal (2009) show that these factors arerobustly related to leverage in the USA In these regressions we use indica-tors for the quintile that the firm is in for each factor

1436 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

Table VIII reports these results In Columns (1) and (2) we examine debtissues and again show that profits positively affect debt issuances The effectsare large and statistically significant at the 1 level Columns (3) and (4)which examine book equity show a strong effect of profits This effect is onlyslightly reduced by the inclusion of other factors Columns (5) and (6) illus-trate that the impact of profits on the change in market equity is also robustto the inclusion of the conventional factors Columns (7) and (8) examineequity issuances Here we again find that profits have a significant negativeimpact on equity issuances As before the results are robust to the inclusionof the conventional factorsIn unreported tables we estimate debt and equity changes using quantile

regressions In terms of the signs and significance of the profitability variable

Table VII Debt and equity changes

The sample contains nonfinancial firms listed on the annual Compustat files for the period

from 1971 to 2009 The table presents fixed effect estimates from regressions of changes indebt (D) changes in the book value of equity (BVE) changes in the market value ofequity (MVE) and net equity issuances (EqyIss) The explanatory variables include

current and lagged change in operating income before depreciation (Profits andProfitst1) and lagged change in the book value of assets (Assetst1) In addition tofixed firm effects the regressions include year indicator variables The reported t-statistics

are corrected for clustering at the firm level

aMean significant at the 1 level

bMean significant at the 5 level

cMean significant at the 10 level

D BVE MVE EqyIss

(1) (2) (3) (4) (5) (6) (7) (8)

Profitst1 022a 025a 015b 035a 004 060b 005b 007a

(33) (31) (25) (35) (03) (25) (25) (53)

Profits 050b 078a 140a 003a

(24) (32) (48) (34)

Assetst1 000 007b 025a 001b

(00) (21) (35) (25)

Constant 167 283a 34 51 1145a 1182a 605a 595a

(14) (30) (04) (05) (32) (29) (69) (74)

Year FE Yes Yes Yes Yes Yes Yes Yes Yes

Firm FE Yes Yes Yes Yes Yes Yes Yes Yes

R2 001 004 001 010 001 005 002 003

N 162056 162056 162130 162130 157550 157550 162154 162154

PROFITSndashLEVERAGE PUZZLEREVISITED 1437

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

the results are qualitatively similar to the OLS results reported in Table VIIIProfitability positively affects debt issuances and negatively affects equityissuances across various quantiles Overall the results in Table VIII showthat there is nothing that is special about using sorts or regressions to explaindebt and equity issues In either case we find that more profitable firms tendto increase their debt experience an increase in the value of equity andrepurchase shares Thus the control factors are not responsible for theusual rejection of the trade-off theory It would appear that what mattersis the size of the issuing activity relative to the organic increases in equityvalue due to profits In other words what matters in the leverage regressionseems to be coming from the presence of E in D

DthornE and not from D

Table VIII Debt and equity issuances profitability and leverage factor quintiles

The sample contains nonfinancial firms listed on the annual Compustat files for the period

from 1971 to 2009 The table presents estimates from regressions of change in debt (D)change in the book value of equity (BVE) change in the market value of equity (MVE)and net equity issuances (EquityIss) The explanatory variables include lagged profitability

quintiles and lagged leverage factor quintiles In addition to fixed firm effects the regres-sions include year indicator variables The reported t-statistics are corrected for clustering atthe firm level

aMean significant at the 1 level

bMean significant at the 5 level

cMean significant at the 10 level

D BVE MVE EqyIss

(1) (2) (3) (4) (5) (6) (7) (8)

ProfitabilityQint 214a 167a 283a 175a 80 429a 53a 55a

(70) (54) (66) (59) (08) (48) (55) (62)

IndLevQint 104a 43 99 20

(29) (15) (10) (10)

Ln(Assets)Qint52 17 1026a 53a

(16) (05) (77) (35)MB

Qint158a 372a 1155a 09

(45) (56) (106) (06)

TangibilityQint 05 177 436c 64a

(01) (15) (18) (31)

Constant 167 96 656b 235 4613a 8610a 113 570a

(08) (04) (26) (04) (61) (58) (11) (31)

Year FE Yes Yes Yes Yes Yes Yes Yes Yes

Firm FE Yes Yes Yes Yes Yes Yes Yes Yes

R2 0002 0002 0003 0004 0013 0014 0013 0014

N 155587 155037 155587 155037 155587 155037 155587 155037

1438 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

7 Are Financial Market Conditions Important

According to the trade-off theory capital structure is determined by con-sidering the impact of costs and benefits of debt The time variation inthe costs and benefits of debt imply that issuance decisions would alsovary over time Furthermore market conditions affect the transactioncosts of adjusting Adverse selection is a greater problem in a cold marketthan in a hot one Accordingly it may be easier to adjust in good marketconditions Related ideas have been developed by Huang and Ritter(2009) Faulkender et al (2012) and Halling Yu and Zechner (2012)Faulkender Flannery Hankins and Smithrsquos paper examines cross-sectionalvariation in adjustment speeds including market valuations andhow they may affect adjustment speeds It is natural to think that ingood market conditions it will be less costly to issue both debt and equityand adjusting leverage to respond to profitability shocks would be mucheasierTo test the importance of market conditions we require a definition of

good times and bad times Our empirical strategy is to estimate good timesversus bad times at the four-digit industry level We define an industry ashaving ldquogood timesrdquo if the median firm in that industry has a market-to-book ratio that is higher than the 67th percentile of the time-series distribu-tion of the industry median market-to-book ratios Conversely an industryis defined as having ldquobad timesrdquo if the median firm in the industry has amarket-to-book ratio that is lower than the 33rd percentile of the time-seriesdistribution of the industry median market-to-book ratiosPanel A of Table IX tabulates issuance activity for profitability sorts in

both good times and bad times As might be expected active debt and equityissuances are larger during good times There is also somewhat more activeswaps between debt and equity during good timesPanel B reports the issuance activity for the smallest and the largest firms

during the good times and bad times Within each size quintile firms aresorted on profitability This is a three-way sort As before we find that thereis much more active use of external markets during good times Small low-profit firms are more likely to issue equity in good times than in bad timesDebt issuances are significantly higher in good times Large high-profitfirms are significantly more likely to issue debt and repurchase equity ingood times than in bad timesPanel C reports the magnitudes of the financial variables rather than the

frequencies During bad times less profitable large firms retire substantialamounts of debt and they show a tendency to issue equity Small firms donot seem to engage in similar debt reduction activities However like the

PROFITSndashLEVERAGE PUZZLEREVISITED 1439

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

Table IX Debt and equity issues in good and bad times

The table reports the frequency and magnitude of financing activity for sorts on profits for subsamples

of firms in good and bad times An industry is defined as having ldquogood timesrdquo if the median firm in

that industry has a market-to-book ratio that is higher than the 67th percentile of the time-series

distribution of industry median market-to-book ratios Conversely an industry is defined as having

bad times if the median firm in that industry has a market-to-book ratio that is lower than the 33rd

percentile of the time-series distribution of industry median market-to-book ratios Panels A and B

report the percentage of firms (i) issuing net debt in excess of 5 of the value of their assets (ii) issuing

net equity in excess of 5 of the value of their assets (iii) issuing debt and repurchasing equity both in

excess of 5 of the value of their assets and (iv) issuing equity and retiring debt both in excess of 5

of the value of their assets Panel C tabulates the mean debt and equity levels and the mean changes in

debt and equity issuances The sample contains nonfinancial firms listed on the annual Compustat files

for the period from 1971 to 2009

Iss Iss Iss D Iss ENet D Net E Rep E Ret D(1) (2) (3) (4)

Panel A sorts on profitability

Badtimes

Low 145 131 01 202 170 49 02 083 173 46 04 074 182 47 03 10

High 177 51 06 10t Highfrac14Low 41 136 33 01

Goodtimes

Low 206 381 01 462 197 148 02 233 212 103 04 184 234 92 10 18

High 216 92 16 14t Highfrac14Low 23 658 138 23

Panel B sorts on firm size and profitability

Badtimes

Smallfirm

s Low 151 318 00 312 182 276 02 253 151 140 02 294 183 73 00 19

High 177 80 00 11t Highfrac14Low 10 90 NA 20

Largefirm

s Low 172 54 01 042 148 39 04 053 150 43 04 044 156 33 03 06

High 147 28 05 03t Highfrac14Low 17 32 20 02

Goodtimes Smallfirm

s Low 213 448 01 542 230 443 01 473 222 325 00 474 203 197 02 38

High 190 153 02 31t Highfrac14Low 19 230 10 39

Largefirm

s Low 203 96 03 122 208 66 09 133 232 54 16 064 215 39 19 05

High 234 36 38 04t Highfrac14Low 29 99 92 38

(continued)

1440 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

larger firms there is a tendency for small firms to issue more equitymdashat leastto the extent possibleDuring good times large profitable firms raise significant amounts of debt

as they experience an increase in the value of their equity Such firms alsoengage in share repurchases Debt issues by smaller firms are much lessaffected by their own profitability during good times and their change indebt is more or less independent of their profits However equity issuanceis primarily found among the low-profit small firms during good timesTable IX shows that empirically profitability has a time-varying impacton leverage choicesIn Table X the basic regressions for changes in debt and for equity

issuance are presented for both good times and bad times Comparinggood times with bad times we see that the effects of profits are muchstronger in good times In bad times the impacts are rather weak statisticallyThis difference is interesting and deserves further study However for ourpurposes recall that the static trade-off theory implies time-varying capital

Table IX Continued

D D BVE BVE MVE MVE EqyIss Assets

(1) (2) (3) (4) (5) (6) (7) (8)

Panel C magnitude of financing activity

Badtimes Smallfirm

s Low 40 03 58 14 305 06 22 140

2 44 05 82 07 240 67 17 180

3 45 04 94 02 176 18 05 187

4 43 09 98 02 137 04 03 188

High 36 10 114 13 209 16 04 199

Largefirm

s Low 3365 73 2523 169 2304 163 94 8972

2 3031 95 2983 35 3284 278 56 9068

3 2930 5 3360 14 3947 308 57 9225

4 2281 6 3817 19 5213 442 7 8902

High 1790 62 4767 13 8799 58 31 9573

Goodtimes

Smallfirm

s Low 24 06 38 05 339 23 39 90

2 26 08 48 01 301 04 25 104

3 30 09 60 04 246 03 17 125

4 28 07 66 08 206 18 12 130

High 23 06 85 16 261 13 08 141

Largefirm

s Low 4532 36 3477 103 6557 260 53 12465

2 4580 58 3527 139 6911 421 23 12522

3 2952 111 3485 2 7586 279 32 9535

4 2752 124 3939 72 10686 33 105 10226

High 2429 228 4819 377 18565 268 222 10830

PROFITSndashLEVERAGE PUZZLEREVISITED 1441

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

structure choices even when the target is time invariant Empirically we doobserve time-varying choices

8 Is Scaling Important

In the preceding section we provided results for unscaled debt and equityissues This seems appropriate to us However it involves making twochanges to the standard regression We would like to determine if both

Table X Debt and equity changes in good and bad timesmdashregressions

The sample contains nonfinancial firms listed on the annual Compustat files for the period

from 1971 to 2009 The table presents estimates from regressions of change in debt (D)and net equity issuances (EqyIss) in good and bad times The explanatory variables includechange in profits lagged change in profits and lagged change in assets The regressions

include both firm and year fixed effects An industry is defined as having ldquogood timesrdquo ifthe median firm in that industry has a market-to-book ratio that is higher than the 67thpercentile of the time-series distribution of industry median market-to-book ratios

Conversely an industry is defined as having bad times if the median firm in thatindustry has a market-to-book ratio that is lower than the 33rd percentile of the time-series distribution of industry median market-to-book ratios The reported t-statistics arecorrected for clustering at the firm level

aMean significant at the 1 level

bMean significant at the 5 level

cMean significant at the 10 level

Good times Bad times

D EqyIss D EqyIss

(1) (2) (3) (4) (5) (6) (7) (8)

Profits 087a 091a 006b 009a 060 065 003 003b

(30) (29) (20) (28) (15) (16) (16) (20)

Profitst1 029c 015a 034c 001

(19) (37) (17) (04)

Assetst1 000 002b 000 001

(02) (22) (00) (07)

Constant 1337 3989 8691a 10001a 1351 1851 4099a 4055a

(03) (09) (55) (37) (09) (13) (80) (78)

Year FE Yes Yes Yes Yes Yes Yes Yes Yes

Firm FE Yes Yes Yes Yes Yes Yes Yes Yes

R2 0071 0080 0018 0045 0178 0212 0034 0041

N 74827 68729 74891 68787 26965 26013 26971 26018

1442 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

changes are crucial Accordingly in this section we take a step back toconsider prior literature in which variables are scaled The questions arehow does scaling affect the results and whyTable XI examines alternative scalings Columns (1) and (2) scale debt and

equity issuances by the total firm net issuance Columns (3) and (4) scaledebt and equity issuances by the sum of debt and book equity Columns (5)and (6) consider the change in debt scaled by the sum of debt and bookequity Columns (7) and (8) consider the change in leverage due to externalfinancing Since we are examining changes we use the changes in thestandard regression factors as controls The primary interest is in the coef-ficient on the change in profitsWe find that a change in profits is associated with an increase in debt

issues But the sign reverses when we scale the debt issues by the sum of debtplus book equity This is telling What is really driving the results is thechange in equity in the denominator not the change in debt in the numer-ator This result is further substantiated in Columns (5) and (6) where thechange in debt to capital ratio is used instead of debt issuanceTo further explore the impact of active issuance relative to passive

experiencing of changes we decompose leverage by separating out theeffect of debt and equity issues from those due to changes in retainedearnings Thus we define leverage change due to external financing as

LevEF frac14Dt1 thorn d

Dt1 thorn Et1 thorn dthorn e

Dt1

Dt1 thorn Et1 eth1THORN

where D is debt E is book value of equity d is debt issues net of retirementsand e is equity issues net of repurchases Garvey and Hanka (1999) use asimilar measure in their study This measure directly captures the netleverage effect of debt and equity issuancesColumns (7) and (8) report results from regressions of the net leverage

effect of debt and equity issues on changes in profitability after controllingfor other leverage factors in differences and year indicator variables Theeffect of profitability on leverage change due to external financing is positiveand statistically significant This comparison suggests that changes in totaldebt and equity are only partly a result of debt and equity issuance decisionsOther balance sheet adjustments complicate the inferences from leverageratio regressionsFor most of our analysis we follow the literature in treating debt as homo-

genous It is of course potentially interesting to ask what happens when wedistinguish between short-term debt and long-term debt Which part of thefirmrsquos debt maturity structure drives the under-adjustment Table XII

PROFITSndashLEVERAGE PUZZLEREVISITED 1443

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

provides results from such decompositions Alternative scalings areconsidered The main results of interest are that the distinction betweenshort- and long-term debt is important Short-term debt issuance is largelyindependent of changes in profits The negative sign on profit change stemsfrom the action of long-term debt This seems intuitively reasonable

Table XI Scaled issuances

The sample contains nonfinancial firms listed on the annual Compustat files for the period

from 1971 to 2009 The table presents estimates from regressions of debt issuance to totalissuance (in columns (1) and (2)) debt issuance to total capitalization (in columns (3) and(4)) change in debt to capitalization ratio (in columns (5) and (6)) and change in leverage

ratio considering only external financing (in columns (7) and (8)) The change in leverageratio considering only external financing is defined as

LevEF frac14Dt1 thorn d

Dt1 thorn Et1 thorn dthorn e

Dt1

Dt1 thorn Et1

where D is debt E is book value of equity d is debt issues net of retirements and e is equity

issues net of repurchases The regressions include lagged leverage factors and year indicatorvariables Some of the specifications include firm fixed effects The reported t-statistics arecorrected for clustering at the firm level

aMean significant at the 1 level

bMean significant at the 5 level

cMean significant at the 10 level

DebtIssTotIss

DebtIssethDthornETHORN D

DthornE LevEF

(1) (2) (3) (4) (5) (6) (7) (8)

Profitt1 004a 000 002a 001c 003a 002 001b 002a

(35) (03) (34) (18) (35) (14) (23) (26)

IndLevt1 004c 001 001 001c 000 000 000 000

(19) (06) (15) (19) (02) (04) (03) (05)

Market=bookt1 001a 001a 000a 000a 000a 000a 000 000c

(47) (85) (68) (60) (55) (36) (05) (18)

Tangt1 004 005 013a 011a 011a 006a 003a 003a

(12) (15) (144) (112) (79) (38) (44) (42)

Assetst1 008a 004a 006a 004a 002a 001a 003a 002a

(97) (45) (283) (161) (53) (35) (147) (101)

Constant 079a 074a 002a 003a 000 000 000b 002a

(497) (442) (91) (111) (00) (05) (22) (93)

Year FE Yes Yes Yes Yes Yes Yes Yes Yes

Firm FE No Yes No Yes No Yes No Yes

R2 0015 0098 0022 0076 0009 0070 0012 0086

N 133905 133905 140552 140552 138229 138229 140572 140572

1444 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

inasmuch as short-term debt is probably more associated with day-to-dayoperations rather than longer term planning considerationsIn Table XIII we examine change in leverage ( D

DthornE) and change inleverage due to external financing (LEVEF) during good and bad timesseparately The traditional negative effect of a change in profits on achange in leverage is found both in good times and in bad timesHowever when we decompose the impact in order to isolate the externalcomponent as before the result disappears In good times the externalfinancing component of a change in profits is weak but still positiveDuring bad times it is not statistically significant This implies that distin-guishing between good and bad times does not fundamentally alter the basic

Table XII Short- and long-term debt issuances

The sample contains nonfinancial firms listed on the annual Compustat files for the period

from 1971 to 2009 The table presents estimates from regressions of short-term debtissuance to total issuance (in columns (1) and (2)) long-term debt issuance to totalissuance (in columns (3) and (4)) short-term debt issuance to total capitalization (in

columns (5) and (6)) and long-term debt to total capitalization (in columns (7) and (8))Some of the specifications include firm fixed effects The reported t-statistics are correctedfor clustering at the firm level

aMean significant at the 1 level

bMean significant at the 5 level

cMean significant at the 10 level

STDebtIssTotIss

LTDebtIssTotIss

STDebtIssethDthornETHORN

LTDebtIssethDthornETHORN

(1) (2) (3) (4) (5) (6) (7) (8)

Profitabilityt1 0002 0003 0042a 0008 0001 0001 0019a 0012a

(02) (04) (30) (05) (14) (10) (45) (28)

IndLevt1 0009 0007 0070b 0033 0001 0002c 0014a 0005

(06) (05) (26) (11) (08) (17) (26) (09)

M=Bt1 0003a 0003a 0003b 0010a 0000 0000 0003a 0002a

(48) (35) (22) (65) (10) (01) (84) (42)

Tangibilityt1 0020 0007 0011 0067 0005a 0006a 0097a 0081a

(12) (03) (03) (16) (34) (33) (108) (89)

Assetst1 0039a 0011b 0034a 0024b 0001a 0001 0046a 0020a

(89) (22) (38) (23) (36) (13) (190) (82)

Constant 0001b 0153a 0798a 0591a 0000c 0006a 0057a 0096a

(25) (144) (470) (288) (17) (124) (292) (234)

Year FE Yes Yes Yes Yes Yes Yes Yes Yes

Firm FE No Yes No Yes No Yes No Yes

R2 0025 0089 0018 0073 0030 0147 0016 0298

N 136146 136146 136140 136140 143031 143031 142847 142847

PROFITSndashLEVERAGE PUZZLEREVISITED 1445

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

Table XIII Externally financed leverage changes in good and bad times

The sample contains nonfinancial firms listed on the annual Compustat files for the period

from 1971 to 2009 The table presents estimates from regressions of change in debt tocapitalization ratio (in columns (1) (2) (5) and (6)) and change in leverage ratio consider-ing only external financing (in columns (3) (4) (7) and (8)) The change in leverage ratio

considering only external financing is defined as

LevEF frac14Dt1 thorn d

Dt1 thorn Et1 thorn dthorn e

Dt1

Dt1 thorn Et1

where D is debt E is book value of equity d is debt issues net of retirements and e is equityissues net of repurchases The regressions include lagged leverage factors and year indicator

variables The reported t-statistics are corrected for clustering at the firm level An industryis defined as having ldquogood timesrdquo if the median firm in that industry has a market-to-bookratio that is higher than the 67th percentile of the time-series distribution of industry

median market-to-book ratios Conversely an industry is defined as having ldquobad timesrdquoif the median firm in that industry has a market-to-book ratio that is lower than the 33rdpercentile of the time-series distribution of industry median market-to-book ratios

aMean significant at the 1 level

bMean significant at the 5 level

cMean significant at the 10 level

Good times Bad times

DDthornE LevEF D

DthornE LevEF

(1) (2) (3) (4) (5) (6) (7) (8)

Profitabiltyt1 0022c 0005 0016c 0025b 0073b 0041 0007 0007

(17) (03) (18) (24) (25) (11) (04) (03)

IndLevt1 0000 0011 0001 0009 0010 0018 0001 0000

(00) (09) (01) (13) (07) (10) (02) (00)

MB

t1

0004a 0003b 0000 0001 0008a 0001 0000 0000

(40) (24) (05) (08) (28) (04) (03) (03)

Tangt1 0097a 0041 0033a 0038b 0127a 0048 0059a 0033

(45) (16) (27) (25) (49) (14) (43) (16)

Assetst1 0014a 0013b 0024a 0019a 0037a 0023c 0036a 0025a

(30) (20) (84) (48) (45) (17) (88) (43)

Constant 0030b 0036 0000 0003 0028c 0014 0021 0001

(20) (16) (00) (07) (17) (06) (11) (01)

Year FE Yes Yes Yes Yes Yes Yes Yes Yes

Firm FE No Yes No Yes No Yes No Yes

R2 0006 0088 0009 0117 0026 0251 0023 0202

N 56820 56820 58200 58200 23462 23462 23511 23511

1446 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

conclusions The traditional leverage result on profits is driven by increasesin equity that are experienced and then partially but not completely offsetby firmsTo the extent that firms are engaging in active financing they are following

the prescriptions of the trade-off theory Thus we have a coherent picture ofhow firms are responding to profits Basically more profitable firms experiencean increase in the value of their equity They respond to this by issuing debtand repurchasing equity But the magnitude of the increase in equity is largerthan the magnitudes of the active steps taken in the debt and equity marketsThis means that the firmsrsquo actual actions are as predicted by the trade-offtheory But it leaves open the question of why they do not act more stronglyThe capital structure literature has long been interested in dynamics and

capital structure adjustment paths In recent years this interest has increas-ingly resulted in explicit dynamic optimizing models such as those ofGoldstein Ju and Leland (2001) Hennessy and Whited (2005) andStrebulaev (2007) These models commonly make specialized assumptionsleading some to question their generalityWe argue that both fixed transactions costs and variable transactions costs

are plausible and realistic in models that pin down firm size The simultaneouspresence of both is important Due to fixed costs firms only rebalance inter-mittently Due to variable adjustment costs the extent of the adjustment isreduced to the point at which an extra unit to adjustment is balanced againstthe extra unit transactions costs This means that both leverage increases andleverage reductions will only go part way toward a static optimumThe key point is that with both fixed and variable transactions costs we

should expect to only see intermittent leverage adjustments When leverage isadjusted it should ldquounder adjustrdquo when compared to a frictionless modelFurther discussion of transactions costs in a capital structure setting isprovided by Leary and Roberts (2005)

9 Conclusion

The connection between corporate profits and capital structure has beenvery influential in the assessment of the trade-off theory The conventionalinterpretation of the evidence has pushed the literature away from the trade-off theory and toward much more complex dynamic models and ideas As aresult it is important to make sure that the evidence has been correctlyinterpreted We show that the problem is with the interpretation of thetest results and a failure to take into account the costs associated with re-sponding to a profit shock

PROFITSndashLEVERAGE PUZZLEREVISITED 1447

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

We document that more profitable firms really do borrow more (not less)They tend to repurchase equity They experience an increase in both thebook value of equity and the market value of equity Less profitable firmsreally do tend to reduce their debt and to issue equityFirm size and market conditions also matter Larger firms tend to be more

active in the debt markets while smaller firms tend to be relatively moreactive in the equity markets During good times there is more use ofexternal financingThe usual profitsndashleverage puzzle result is primarily driven by the increase

in equity that is experienced by the more profitable firms Accordingly thepuzzle should be restated as asking why do firms not take sufficiently largeoffsetting actions to fully undo the change in equity What limits the mag-nitudes of the typical leverage response to profit shocksIn a frictionless model the partial response appears puzzling But there is

good empirical reason to believe that rebalancing entails both fixed andvariable costs and that firm size matters The rebalancing costs can befully avoided by doing nothing Accordingly the firm must decide whetherany given shock is big enough to be worth responding to If it is then thefirm must decide how big a response is called for These technical conditionsare known as ldquovalue matchingrdquo and ldquosmooth pastingrdquo Thus optimizationimplies that some shocks will be ignored Even if the shock is not ignoredthe optimal response will only partially undo the shock The magnitude ofthe leverage response must balance the marginal cost and the marginalbenefit of an extra unit of leverage Since the marginal cost of adjustingleverage is strictly positive the adjustment toward that static leverageoptimum will only go part way This is true both for leverage increasesand for leverage reductionsOne might conjecture that the variable costs would be time varying and

that such costs might have declined over time Since we do not have a fullaccounting of all of the variable costs it is of course impossible to be sureHowever there is some evidence suggesting that at least the directunderwriting fees are remarkably stable over time10

Surely nobody ought to be surprised that more profitable firms are moreinclined to repurchase equity while unprofitable firms tend not to do so Thefact that more profitable firms typically experience an increase in the value of

10 For IPOs Jay Ritter provides important summary evidence as httpbearwarringtonufleduritterIPOs2013Underwritingpdf He finds that for medium-sized deals a remark-

ably large fraction of the deals have a 7 fee year after year Whether this is due tobehavioral factors or some sort of a coordination device is not clear The extent towhich similar intertemporal stability carries over to other forms of external financemerits further study

1448 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

equity is equally natural Were it not for the contrary claims in the literature(ie the profitsndashleverage puzzle) we would argue that nobody ought to findit surprising that more profitable firms are more inclined to issue debt Thefact that leverage adjustments typically only partially offset profit shocksshould also sound quite natural These facts all fit together easily within thetrade-off theory once we take account of both fixed and variable costs ofactively adjusting capital structure

Appendix

Table AI Market leverage regressions

The table presents estimates of the regressions of market leverage on firm characteristics

The sample comes from the annual Compustat files during the period 1971ndash2009 Financialfirms are excluded The leverage is estimated as the ratio of debt over debt plus marketequity The explanatory variables Profitabilityt1

MB

t1

Tangibilityt1 and LnethAssetsTHORNt1are described in Table 1 IndMedianLevt1 is estimated as the median market leverage of allother firms in the same industry The industry is defined at the level of the four-digit SICcode All specifications include the year fixed effects The specifications in column (2)

additionally include the firm fixed effects We report t-statistics where the standard errorsare clustered at the firm level in parentheses

aSignificant at the 1 level

Quantile regressions

(1) (2) (3) (4) (5)

Profitabilityt1 013a 013a 007a 011a 020a

(275) (262) (349) (425) (538)

IndLevt1 046a 020a 032a 058a 063a

(506) (239) (1299) (1630) (1042)MB

t1

003a 001a 001a 002a 004a

(509) (303) (743) (709) (534)

Tangibilityt1 013a 018a 013a 014a 010a

(191) (181) (762) (571) (232)

LnethAssetsTHORNt1 001a 005a 001a 001a lt001a

(112) (258) (682) (342) (29)

Constant 010a 011a 005a 005a 029a

(171) (98) (173) (118) (398)

Year FE Yes Yes Yes Yes Yes

Firm FE No Yes No No No

Regression type OLS FE 25th tile Median 75thtile

Clustered SE Yes Yes No No No

R2 030 067

Pseudo R2 016 023 0201

N 158578 158578 158578 158578 158578

PROFITSndashLEVERAGE PUZZLEREVISITED 1449

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

Table AII Profitability sorts for firms issuing large amounts

Percentage of firms

Iss Ret Iss Iss Rep Iss Iss IssD IssE

D D NetD E E NetE DampE RepE RetD

(1) (2) (3) (4) (5) (6) (7) (8) (9)

Panel A 10 cutoff to define issuing or retiring activity

Allfirm

s

Low 205 150 134 260 05 257 45 00 23

2 207 175 103 77 06 75 11 01 10

3 243 196 101 50 08 47 08 01 09

4 252 186 113 50 13 48 08 02 08

High 227 161 118 60 27 57 07 04 09

t Highfrac14Low 70 38 62 755 226 768 320 90 152

Smallfirm

s Low 230 132 162 380 05 375 82 00 32

2 228 143 165 334 04 333 65 00 26

3 211 146 147 216 03 214 38 00 23

4 190 154 120 126 04 124 15 01 20

High 194 162 114 99 09 97 12 01 15

t Highfrac14Low 42 40 68 342 20 340 169 07 131

Largefirm

s Low 231 200 82 31 06 29 08 01 04

2 231 187 67 17 06 14 03 01 02

3 234 177 77 17 09 14 03 02 03

4 218 158 75 16 17 14 02 03 02

High 218 135 99 19 41 15 03 07 02

t Highfrac14Low 19 106 36 46 135 57 39 58 23

Percentage of firms

Iss Ret Iss Iss Rep Iss Both Iss D Iss E

D D Net D E E Net E DampE Rep E Ret D

(1) (2) (3) (4) (5) (6) (7) (8) (9)

Panel B 15 cutoff to define issuing or retiring activity

Allfirm

s

Low 158 101 100 230 01 228 27 00 15

2 142 115 63 57 01 56 05 00 06

3 165 128 57 32 01 31 03 00 05

4 173 118 65 35 01 34 03 00 05

High 160 102 70 46 04 44 03 00 05

t Highfrac14Low 08 06 141 738 87 746 277 24 131

Smallfirm

s Low 195 95 132 354 01 350 56 00 24

2 185 96 131 303 01 301 44 00 18

3 160 96 104 179 00 176 18 00 15

4 134 99 78 99 01 98 08 00 12

High 133 95 73 82 01 80 06 00 10

t Highfrac14Low 80 00 95 345 12 344 149 12 52

Largefirm

s Low 145 123 42 14 01 12 03 00 01

2 137 114 32 07 01 06 01 00 01

3 142 104 35 06 01 05 00 00 01

4 130 91 36 08 03 07 01 01 01

High 134 81 48 09 06 07 01 01 01

t Highfrac14Low 19 85 17 26 54 33 34 17 11

1450 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

Table AIV Debt and equity changesmdashwithout fixed effects

The sample contains nonfinancial firms listed on the annual Compustat files for the period from 1971

to 2009 The table presents estimates from regressions of change in debt (D) change in the book

value of equity (BVE) change in the market value of equity (MVE) and net equity issuances

(EqyIss) The explanatory variables include current and lagged change in operating income before

depreciation (Profits and Profitst1) and lagged change in book value of assets (Assetst1) The

regressions include year indicator variables The reported t-statistics are corrected for clustering at the

firm level

aMean significant at the 1 level

bMean significant at the 5 level

cMean significant at the 10 level

D BVE MVE EqyIss

(1) (2) (3) (4) (5) (6) (7) (8)

Profitst1 025a 024a 023a 038a 007 066a 007a 009a

(40) (32) (36) (39) (06) (29) (42) (51)Profits 051b 082a 149a 004a

(25) (33) (49) (33)Assetst1 002 005c 024a 001b

(08) (17) (34) (23)Constant 93 176a 45 71 1552a 1704a 130a 140a

(12) (28) (07) (09) (49) (49) (45) (45)Year FE Yes Yes Yes Yes Yes Yes Yes YesR2 001 005 001 011 001 005 002 002N 162056 162056 162130 162130 157550 157550 162154 162154

Table AIII Median profitability sorts

The sample contains nonfinancial firms listed on the annual Compustat files for the period from 1971

to 2009 The table presents median debt and equity median changes in debt and equity and median

equity issuances (EqyIss) in millions of dollars for all firms sorted on profitability and for profitability

sorts within the smallest and largest firms The sorts are done annually

D D BVE BVE MVE MVE EqyIss Assets(1) (2) (3) (4) (5) (6) (7) (8)

Allfirm

s

Low 21 00 75 17 256 20 01 1922 186 02 449 18 598 19 00 11033 816 06 1274 03 1637 05 00 32244 787 02 1541 18 2488 03 00 3556High 273 00 1361 66 3300 14 00 2578

Smallfirm

s Low 05 00 13 07 140 14 06 402 06 00 25 08 114 12 01 563 09 00 32 05 85 07 00 674 12 00 39 01 68 03 00 79High 10 00 51 04 98 01 00 90

Largefirm

s Low 11485 381 10269 371 13227 202 03 333902 13051 393 12786 140 17571 169 05 392463 11554 287 13429 24 21623 305 00 370314 9963 209 15052 194 29424 308 00 36705High 7361 00 16380 613 49142 439 00 34976

PROFITSndashLEVERAGE PUZZLEREVISITED 1451

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

References

Auerbach A J (1985) Real determinants of corporate leverage in B M Friedman (ed)

Corporate Capital Structures in the United States University of Chicago Press Chicago

IllinoisBertrand M and Mullainathan S (2003) Enjoying the quiet life corporate governance and

managerial preferences Journal of Political Economy 111 1043ndash1075Booth L Varouj A Demirguc-Kunt A and Maksimovic V (2001) Capital structures in

developing countries Journal of Finance 56 87ndash130Brennan M J and Schwartz E S (1984) Optimal financial policy and firm valuation

Journal of Finance 39 593ndash607Cameron A C and Trivedi P K (2010) Microeconometrics using Stata Stata Press TexasConstantinides G M (1979) Multiperiod consumption and investment behavior with

convex transactions costs Management Science 25 1127ndash1137Constantinides G M and Richard S F (1978) Existence of optimal simple policies for

discounted-cost inventory and cash management in continuous time Operations Research

26 620ndash636DeAngelo H DeAngelo L and Whited T M (2011) Capital structure dynamics and

transitory debt Journal of Financial Economics 99 235ndash261Dixit A K (1993) The Art of Smooth Pasting Routledge London and New YorkFama E and French K R (2002) Testing trade-off and pecking order predictions about

dividends and debt Review of Financial Studies 15 1ndash33Faulkender M and Petersen M A (2006) Does the source of capital affect capital struc-

ture Review of Financial Studies 19 45ndash79Faulkender M Flannery M J Hankins K W and Smith J M (2012) Cash Flows and

Leverage Adjustments Journal of Financial Economics 103 632ndash646Fischer E Heinkel R and Zechner J (1989) Dynamic capital structure choice theory

and tests Journal of Finance 44 19ndash40Frank M Z and Goyal V K (2008) Trade-off and pecking order theories of

debt in B E Eckbo (ed) Handbook of Corporate Finance Empirical Corporate Finance

Vol 2 Handbook of Finance Series Chapter 12 ElsevierNorth-Holland Amsterdam pp

135ndash202

Frank M Z and Goyal V K (2009) Capital structure decisions which factors are reliably

important Financial Management 38 1ndash37

Garvey G T and Hanka G (1999) Capital structure and corporate control the effect of

antitakeover statutes on firm leverage Journal of Finance 54 519ndash546

Goldstein R Ju N and Leland H (2001) An EBIT-based model of dynamic capital

structure Journal of Business 74 483ndash512

Graham J R and Tucker A L (2006) Tax shelters and corporate debt policy Journal of

Financial Economics 81 563ndash594

Hackbarth D and Mauer D C (2012) Optimal priority structure capital structure and

investment Review of Financial Studies 25 747ndash796

Halling M Yu J and Zechner J (2012) Leverage dynamics over the business cycle

Working paper University of Utah University of New South Wales and Vienna

University of Economics and Business AdministrationHennessy C A and Whited T A (2005) Debt dynamics Journal of Finance 60 1129ndash1165

Huang R and Ritter J R (2009) Testing theories of capital structure and estimating the

speed of adjustment Journal of Financial and Qualitative Analysis 44 237ndash271

1452 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

Hovakimian A Opler T and Titman S (2001) The debt-equity choice Journal of

Financial and Quantitative Analysis 36 1ndash24Jung K Kim Y and Stulz R M (1996) Timing investment opportunities managerial

discretion and the security issue decisions Journal of Financial Economics 42 159ndash185Kane A Marcus A J and McDonald R L (1984) How big is the tax advantage to

debt Journal of Finance 39 841ndash853Kayhan A and Titman S (2007) Firmsrsquo histories and their capital structures Journal of

Financial Economics 83 1ndash32Leary M T (2009) Bank loan supply lender choice and corporate capital structure

Journal of Finance 64 1143ndash1185Leary M T and Roberts M R (2005) Do firms rebalance their capital structures

Journal of Finance 60 2575ndash2619Leland H E (1994) Corporate debt value bond covenants and optimal capital structure

Journal of finance 49 1213ndash1252Lemmon M L Roberts M R and Zender J F (2008) Back to the beginning persist-

ence and the cross-section of corporate capital structure Journal of Finance 63 1575ndash1608Long M S and Malitz I B (1985) The investment-financing nexus some empirical

evidence Midland Corporate Finance Journal 3 53ndash59Mackie-Mason J K (1990) Do taxes affect corporate financing decisions Journal of

Finance 45 1471ndash1493Morellec E Nikolov B and Schurhoff N (2012) Corporate governance and capital

structure dynamics Journal of Finance 67 803ndash848Myers S C (1984) The capital structure puzzle Journal of Finance 39 575ndash592Myers S C (1993) Still searching for optimal capital structure Journal of Applied

Corporate Finance 61 4ndash14Rajan R and Zingales L (1995) What do we know about capital structure some evidence

from international data Journal of Finance 50 1421ndash1460Strebulaev I A (2007) Do tests of capital structure theory mean what they say Journal of

Finance 62 1747ndash1787

Strebulaev I A and Whited T M (2012) Dynamic Models and Structural Estimation in

Corporate Finance Now Publisher Boston and Delft

Titman S and Wessels R (1988) The determinants of capital structure choice Journal of

Finance 43 1ndash21

Welch I (2004) Capital structure and stock returns Journal of Political Economy 112

106ndash131

Welch I (2011) Two common problems in capital structure research the financial-debt-

to-asset ratio and issuing activity versus leverage changes International Review of Finance

11 1ndash17

PROFITSndashLEVERAGE PUZZLEREVISITED 1453

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

values involved In Table V we sort firms according to profits and thentabulate the levels and changes in both debt and equity We do this firstfor all firms and then for small versus large firmsFor the analysis on all firms we observe in Column (1) that debt peaks at

the middle of the distribution This is because firms with medium profitabil-ity are also the largest as seen in Column (8) which reports average assetvalues for different profitability quintiles When we sort first by firm size andthen examine firms with different levels of profitability within size quintileswe find that for small firms debt is roughly independent of profits while forlarge firms less profitable firms have a higher level of debtAs expected Columns (3) and (5) show that more profitable firms have

higher equity values Columns (2) and (7) consider the relationship betweenissuances of debt and equity and firm profitability Debt issuances are sig-nificantly larger for more profitable firms In contrast equity issuances arelarger for low-profit firms Small low-profit firms issue substantially lessequity compared with small high-profit firms Among large firms the low-profit ones tend to issue equity while the high-profit ones tend to repurchaseequity The fact that more profitable firms issue debt and repurchase equitywhile the least profitable firms retire debt and issue equity is consistent withthe predicted relation between profitability and financing decisions under thetrade-off theoryWe also examine two-way sorts by size and profitability and report results

for the smallest and largest quintile of firms For the smallest quintile offirms the change in debt is largely unrelated to profits But for large firmsthere is a positive relation between profits and debt issuances High-profitfirms have a big positive change in debt Low-profit large firms have anegative change in debtColumns (4) and (6) provide an explanation of why the leverage ratio

regression results contradict those from the basic profitability sorts pre-sented here As we can see profitability indirectly affects leverage ratiosby increasing equity values Changes in both the book value of equity andthe market value of equity are positive and large for highly profitable firmsIn contrast these changes are negative for less profitable firms This evidenceillustrates an important issue concerning the use of leverage ratios Suchratios are often interpreted as essentially reflecting the use of debt by thefirm This interpretation while common is empirically misleadingFor the typical firm the change in the value of equity is larger than the

change in debt For example in the third profit quintile for large firmscolumn (7) the mean equity issue is just $9 million but the change in themarket value of equity is $242 million At the same time the mean change indebt is $27 million This suggests that a fair bit of the observed variation in

PROFITSndashLEVERAGE PUZZLEREVISITED 1433

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

the leverage ratios is primarily driven by the changes in the market value ofequity in the denominator rather than by the changes in debt in the numer-ator Since equity issues are often small this implies that the variation in theleverage ratio is primarily driven by internal operations rather than byexternal financing actions This again points to the fact that leverageratios can provide a misleading account of actual patterns in the dataThere is always a potential concern that the averages may be misleading

due to the impact of outliers To address this concern in AppendixTable AIII we present median values of the profitability sorts Theseresults generally reinforce the findings in Table VTable VI again considers the magnitudes of financing activity but this

time the issues are scaled by a measure of firm size That way it is easier to

Table V Magnitude of financing activity

The sample contains nonfinancial firms listed on the annual Compustat files for the period

from 1971 to 2009 The table presents average debt and equity changes in debt and equityand equity issuances (in millions of dollars) for firms sorted on profitability within sizeclasses The table reports information for all firms sorted on profitability and for profit-

ability sorts within the smallest and largest firms The sorts are done annually The tablealso reports results from t-test for the difference in percentage of firms issuing or retiringsecurities in the lowest and the highest profit quintile

D D BVE BVE MVE MVE EqyIss Assets

(1) (2) (3) (4) (5) (6) (7) (8)

Allfirm

s

Low 83 5 78 10 120 5 7 234

2 605 13 509 24 721 8 17 1691

3 1227 9 1123 22 1834 3 9 3551

4 833 23 1166 7 2408 74 9 2994

High 570 53 1171 71 3817 68 36 2581

t Highfrac14Low 31 9 36 9 38 2 16 36

Smallfirm

s Low 25 04 42 01 332 28 33 97

2 29 06 55 02 278 16 21 116

3 31 07 63 01 212 05 12 127

4 30 06 70 05 179 10 08 136

High 24 06 84 14 234 17 07 142

t Highfrac14Low 04 15 153 85 73 11 201 107

Largefirm

s Low 3929 55 2990 152 4671 11 63 10846

2 3826 19 3178 95 5122 66 11 10721

3 3074 27 3498 41 6173 242 9 9697

4 2623 62 3915 45 8473 94 57 9855

High 2180 193 4642 261 14924 202 148 10171

t Highfrac14Low 94 64 108 87 246 12 169 16

1434 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

see whether large firms behave disproportionately when compared withsmall firms Column (1) shows that low-profit firms have more debtas reported in many previous studies In Columns (4) and (6) we seethat low-profit firms have a major decline in both the book value ofequity and the market value of equity The high-profit firms have thereverse experienceMore interesting are the active decisions of the firms In Column (2) we see

that in contrast to the usual impression the low-profit firms are reducingdebt while the high-profit firms are increasing it In Column (7) we see thatthe low-profit firms are also reducing leverage by issuing equity The high-profit firms issue little equity In the lower panel similar results are foundwhen we separate small and large firms Equity issuing is particularly strong

Table VI Magnitude of financing activitymdashscaled levels and changes

The sample contains nonfinancial firms listed on the annual Compustat files for the period

from 1971 to 2009 The table presents average debt and equity changes in debt and equityand equity issuances scaled by assets for firms sorted on profitability within size classes Thetable reports information for all firms sorted on profitability and for profitability sorts

within the smallest and largest firms The sorts are done annually The table also reportsresults from t-test for the difference in percentage of firms issuing or retiring securities inthe lowest and the highest profit quintile

D D BVE BVE MVE MVE EqyIss

(1) (2) (3) (4) (5) (6) (7)

Allfirm

s

Low 386 122 12 871 3427 1868 217

2 285 58 407 119 940 157 28

3 293 12 425 14 742 03 12

4 269 26 461 02 955 27 10

High 192 15 551 47 1675 74 07

t Highfrac14Low 35 49 50 47 127 42 157

Smallfirm

s Low 545 154 841 2145 9688 9128 696

2 684 49 712 1915 5014 2318 288

3 331 253 141 969 3063 954 129

4 267 181 357 175 1673 221 55

High 194 05 543 26 1758 147 41

t Highfrac14Low 49 17 51 36 88 27 99

Largefirm

s Low 350 30 306 58 511 55 09

2 346 14 335 13 523 02 04

3 353 165 332 71 711 29 01

4 280 01 414 04 909 09 05

High 227 15 477 19 1643 27 14

t Highfrac14Low 371 112 453 181 570 57 221

PROFITSndashLEVERAGE PUZZLEREVISITED 1435

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

among the small low-profit firms that are also experiencing sharp reductionsin the value of equity In column (2) we see that both for the large and for thesmall firms net debt increases are primarily found in the highest quintile offirm profitability

6 Debt and Equity Regressions

So far we have demonstrated that in our data the conventional leverageregressions have the usual signs We then demonstrated that in simple sortsof the data strikingly different active financing patterns emerge The nexttask is to reconcile these differencesTable VII presents simple regressions predicting changes in debt book

equity and market equity as well as equity issuances Consistent with thesorts we control for changes in firm size and we also include firm and yeardummies There may be a concern about the speed of a firmrsquos responsesAccordingly we consider the lagged change in profits as the main case butwe also include the contemporaneous change in profits Whether the contem-poraneous change in profits is included or not does not in fact matter Wereport panel-robust standard errors adjusted for clustering at the firm levelThe results in Table VII support the findings in the sorts Firms that have

an increase in profits in 1 year increase debt both in the same year and in thenext year Firms that have an increase in profits experience an increase inboth the book value of equity and the market value of equity Importantlyhowever firms that experience and increase in profits reduce their equityissues All these effects are statistically significant (except for the marketvalue of equity in the column 5 specification) and seem intuitively reason-able In Appendix Table AIV similar regressions to those presented aboveare reported but without firm-fixed effects We find that empirically verylittle changes whether we include firm fixed effects or leave them out Thusthe targeting behavior to the extent that it happens does not account forwhat we are reportingThe next step is to bring the sorts and the simple regressions together into

a common setting We therefore include factors in addition to profitability(i) median industry leverage (ii) the market-to-book assets ratio (iii) tangi-bility of assets and (iv) firm size (measured by the log of assets) Rajan andZingales (1995) show that these factors are related to leverage in G7countries A number of studies have used these factors to estimateleverage targets Frank and Goyal (2009) show that these factors arerobustly related to leverage in the USA In these regressions we use indica-tors for the quintile that the firm is in for each factor

1436 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

Table VIII reports these results In Columns (1) and (2) we examine debtissues and again show that profits positively affect debt issuances The effectsare large and statistically significant at the 1 level Columns (3) and (4)which examine book equity show a strong effect of profits This effect is onlyslightly reduced by the inclusion of other factors Columns (5) and (6) illus-trate that the impact of profits on the change in market equity is also robustto the inclusion of the conventional factors Columns (7) and (8) examineequity issuances Here we again find that profits have a significant negativeimpact on equity issuances As before the results are robust to the inclusionof the conventional factorsIn unreported tables we estimate debt and equity changes using quantile

regressions In terms of the signs and significance of the profitability variable

Table VII Debt and equity changes

The sample contains nonfinancial firms listed on the annual Compustat files for the period

from 1971 to 2009 The table presents fixed effect estimates from regressions of changes indebt (D) changes in the book value of equity (BVE) changes in the market value ofequity (MVE) and net equity issuances (EqyIss) The explanatory variables include

current and lagged change in operating income before depreciation (Profits andProfitst1) and lagged change in the book value of assets (Assetst1) In addition tofixed firm effects the regressions include year indicator variables The reported t-statistics

are corrected for clustering at the firm level

aMean significant at the 1 level

bMean significant at the 5 level

cMean significant at the 10 level

D BVE MVE EqyIss

(1) (2) (3) (4) (5) (6) (7) (8)

Profitst1 022a 025a 015b 035a 004 060b 005b 007a

(33) (31) (25) (35) (03) (25) (25) (53)

Profits 050b 078a 140a 003a

(24) (32) (48) (34)

Assetst1 000 007b 025a 001b

(00) (21) (35) (25)

Constant 167 283a 34 51 1145a 1182a 605a 595a

(14) (30) (04) (05) (32) (29) (69) (74)

Year FE Yes Yes Yes Yes Yes Yes Yes Yes

Firm FE Yes Yes Yes Yes Yes Yes Yes Yes

R2 001 004 001 010 001 005 002 003

N 162056 162056 162130 162130 157550 157550 162154 162154

PROFITSndashLEVERAGE PUZZLEREVISITED 1437

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

the results are qualitatively similar to the OLS results reported in Table VIIIProfitability positively affects debt issuances and negatively affects equityissuances across various quantiles Overall the results in Table VIII showthat there is nothing that is special about using sorts or regressions to explaindebt and equity issues In either case we find that more profitable firms tendto increase their debt experience an increase in the value of equity andrepurchase shares Thus the control factors are not responsible for theusual rejection of the trade-off theory It would appear that what mattersis the size of the issuing activity relative to the organic increases in equityvalue due to profits In other words what matters in the leverage regressionseems to be coming from the presence of E in D

DthornE and not from D

Table VIII Debt and equity issuances profitability and leverage factor quintiles

The sample contains nonfinancial firms listed on the annual Compustat files for the period

from 1971 to 2009 The table presents estimates from regressions of change in debt (D)change in the book value of equity (BVE) change in the market value of equity (MVE)and net equity issuances (EquityIss) The explanatory variables include lagged profitability

quintiles and lagged leverage factor quintiles In addition to fixed firm effects the regres-sions include year indicator variables The reported t-statistics are corrected for clustering atthe firm level

aMean significant at the 1 level

bMean significant at the 5 level

cMean significant at the 10 level

D BVE MVE EqyIss

(1) (2) (3) (4) (5) (6) (7) (8)

ProfitabilityQint 214a 167a 283a 175a 80 429a 53a 55a

(70) (54) (66) (59) (08) (48) (55) (62)

IndLevQint 104a 43 99 20

(29) (15) (10) (10)

Ln(Assets)Qint52 17 1026a 53a

(16) (05) (77) (35)MB

Qint158a 372a 1155a 09

(45) (56) (106) (06)

TangibilityQint 05 177 436c 64a

(01) (15) (18) (31)

Constant 167 96 656b 235 4613a 8610a 113 570a

(08) (04) (26) (04) (61) (58) (11) (31)

Year FE Yes Yes Yes Yes Yes Yes Yes Yes

Firm FE Yes Yes Yes Yes Yes Yes Yes Yes

R2 0002 0002 0003 0004 0013 0014 0013 0014

N 155587 155037 155587 155037 155587 155037 155587 155037

1438 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

7 Are Financial Market Conditions Important

According to the trade-off theory capital structure is determined by con-sidering the impact of costs and benefits of debt The time variation inthe costs and benefits of debt imply that issuance decisions would alsovary over time Furthermore market conditions affect the transactioncosts of adjusting Adverse selection is a greater problem in a cold marketthan in a hot one Accordingly it may be easier to adjust in good marketconditions Related ideas have been developed by Huang and Ritter(2009) Faulkender et al (2012) and Halling Yu and Zechner (2012)Faulkender Flannery Hankins and Smithrsquos paper examines cross-sectionalvariation in adjustment speeds including market valuations andhow they may affect adjustment speeds It is natural to think that ingood market conditions it will be less costly to issue both debt and equityand adjusting leverage to respond to profitability shocks would be mucheasierTo test the importance of market conditions we require a definition of

good times and bad times Our empirical strategy is to estimate good timesversus bad times at the four-digit industry level We define an industry ashaving ldquogood timesrdquo if the median firm in that industry has a market-to-book ratio that is higher than the 67th percentile of the time-series distribu-tion of the industry median market-to-book ratios Conversely an industryis defined as having ldquobad timesrdquo if the median firm in the industry has amarket-to-book ratio that is lower than the 33rd percentile of the time-seriesdistribution of the industry median market-to-book ratiosPanel A of Table IX tabulates issuance activity for profitability sorts in

both good times and bad times As might be expected active debt and equityissuances are larger during good times There is also somewhat more activeswaps between debt and equity during good timesPanel B reports the issuance activity for the smallest and the largest firms

during the good times and bad times Within each size quintile firms aresorted on profitability This is a three-way sort As before we find that thereis much more active use of external markets during good times Small low-profit firms are more likely to issue equity in good times than in bad timesDebt issuances are significantly higher in good times Large high-profitfirms are significantly more likely to issue debt and repurchase equity ingood times than in bad timesPanel C reports the magnitudes of the financial variables rather than the

frequencies During bad times less profitable large firms retire substantialamounts of debt and they show a tendency to issue equity Small firms donot seem to engage in similar debt reduction activities However like the

PROFITSndashLEVERAGE PUZZLEREVISITED 1439

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

Table IX Debt and equity issues in good and bad times

The table reports the frequency and magnitude of financing activity for sorts on profits for subsamples

of firms in good and bad times An industry is defined as having ldquogood timesrdquo if the median firm in

that industry has a market-to-book ratio that is higher than the 67th percentile of the time-series

distribution of industry median market-to-book ratios Conversely an industry is defined as having

bad times if the median firm in that industry has a market-to-book ratio that is lower than the 33rd

percentile of the time-series distribution of industry median market-to-book ratios Panels A and B

report the percentage of firms (i) issuing net debt in excess of 5 of the value of their assets (ii) issuing

net equity in excess of 5 of the value of their assets (iii) issuing debt and repurchasing equity both in

excess of 5 of the value of their assets and (iv) issuing equity and retiring debt both in excess of 5

of the value of their assets Panel C tabulates the mean debt and equity levels and the mean changes in

debt and equity issuances The sample contains nonfinancial firms listed on the annual Compustat files

for the period from 1971 to 2009

Iss Iss Iss D Iss ENet D Net E Rep E Ret D(1) (2) (3) (4)

Panel A sorts on profitability

Badtimes

Low 145 131 01 202 170 49 02 083 173 46 04 074 182 47 03 10

High 177 51 06 10t Highfrac14Low 41 136 33 01

Goodtimes

Low 206 381 01 462 197 148 02 233 212 103 04 184 234 92 10 18

High 216 92 16 14t Highfrac14Low 23 658 138 23

Panel B sorts on firm size and profitability

Badtimes

Smallfirm

s Low 151 318 00 312 182 276 02 253 151 140 02 294 183 73 00 19

High 177 80 00 11t Highfrac14Low 10 90 NA 20

Largefirm

s Low 172 54 01 042 148 39 04 053 150 43 04 044 156 33 03 06

High 147 28 05 03t Highfrac14Low 17 32 20 02

Goodtimes Smallfirm

s Low 213 448 01 542 230 443 01 473 222 325 00 474 203 197 02 38

High 190 153 02 31t Highfrac14Low 19 230 10 39

Largefirm

s Low 203 96 03 122 208 66 09 133 232 54 16 064 215 39 19 05

High 234 36 38 04t Highfrac14Low 29 99 92 38

(continued)

1440 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

larger firms there is a tendency for small firms to issue more equitymdashat leastto the extent possibleDuring good times large profitable firms raise significant amounts of debt

as they experience an increase in the value of their equity Such firms alsoengage in share repurchases Debt issues by smaller firms are much lessaffected by their own profitability during good times and their change indebt is more or less independent of their profits However equity issuanceis primarily found among the low-profit small firms during good timesTable IX shows that empirically profitability has a time-varying impacton leverage choicesIn Table X the basic regressions for changes in debt and for equity

issuance are presented for both good times and bad times Comparinggood times with bad times we see that the effects of profits are muchstronger in good times In bad times the impacts are rather weak statisticallyThis difference is interesting and deserves further study However for ourpurposes recall that the static trade-off theory implies time-varying capital

Table IX Continued

D D BVE BVE MVE MVE EqyIss Assets

(1) (2) (3) (4) (5) (6) (7) (8)

Panel C magnitude of financing activity

Badtimes Smallfirm

s Low 40 03 58 14 305 06 22 140

2 44 05 82 07 240 67 17 180

3 45 04 94 02 176 18 05 187

4 43 09 98 02 137 04 03 188

High 36 10 114 13 209 16 04 199

Largefirm

s Low 3365 73 2523 169 2304 163 94 8972

2 3031 95 2983 35 3284 278 56 9068

3 2930 5 3360 14 3947 308 57 9225

4 2281 6 3817 19 5213 442 7 8902

High 1790 62 4767 13 8799 58 31 9573

Goodtimes

Smallfirm

s Low 24 06 38 05 339 23 39 90

2 26 08 48 01 301 04 25 104

3 30 09 60 04 246 03 17 125

4 28 07 66 08 206 18 12 130

High 23 06 85 16 261 13 08 141

Largefirm

s Low 4532 36 3477 103 6557 260 53 12465

2 4580 58 3527 139 6911 421 23 12522

3 2952 111 3485 2 7586 279 32 9535

4 2752 124 3939 72 10686 33 105 10226

High 2429 228 4819 377 18565 268 222 10830

PROFITSndashLEVERAGE PUZZLEREVISITED 1441

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

structure choices even when the target is time invariant Empirically we doobserve time-varying choices

8 Is Scaling Important

In the preceding section we provided results for unscaled debt and equityissues This seems appropriate to us However it involves making twochanges to the standard regression We would like to determine if both

Table X Debt and equity changes in good and bad timesmdashregressions

The sample contains nonfinancial firms listed on the annual Compustat files for the period

from 1971 to 2009 The table presents estimates from regressions of change in debt (D)and net equity issuances (EqyIss) in good and bad times The explanatory variables includechange in profits lagged change in profits and lagged change in assets The regressions

include both firm and year fixed effects An industry is defined as having ldquogood timesrdquo ifthe median firm in that industry has a market-to-book ratio that is higher than the 67thpercentile of the time-series distribution of industry median market-to-book ratios

Conversely an industry is defined as having bad times if the median firm in thatindustry has a market-to-book ratio that is lower than the 33rd percentile of the time-series distribution of industry median market-to-book ratios The reported t-statistics arecorrected for clustering at the firm level

aMean significant at the 1 level

bMean significant at the 5 level

cMean significant at the 10 level

Good times Bad times

D EqyIss D EqyIss

(1) (2) (3) (4) (5) (6) (7) (8)

Profits 087a 091a 006b 009a 060 065 003 003b

(30) (29) (20) (28) (15) (16) (16) (20)

Profitst1 029c 015a 034c 001

(19) (37) (17) (04)

Assetst1 000 002b 000 001

(02) (22) (00) (07)

Constant 1337 3989 8691a 10001a 1351 1851 4099a 4055a

(03) (09) (55) (37) (09) (13) (80) (78)

Year FE Yes Yes Yes Yes Yes Yes Yes Yes

Firm FE Yes Yes Yes Yes Yes Yes Yes Yes

R2 0071 0080 0018 0045 0178 0212 0034 0041

N 74827 68729 74891 68787 26965 26013 26971 26018

1442 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

changes are crucial Accordingly in this section we take a step back toconsider prior literature in which variables are scaled The questions arehow does scaling affect the results and whyTable XI examines alternative scalings Columns (1) and (2) scale debt and

equity issuances by the total firm net issuance Columns (3) and (4) scaledebt and equity issuances by the sum of debt and book equity Columns (5)and (6) consider the change in debt scaled by the sum of debt and bookequity Columns (7) and (8) consider the change in leverage due to externalfinancing Since we are examining changes we use the changes in thestandard regression factors as controls The primary interest is in the coef-ficient on the change in profitsWe find that a change in profits is associated with an increase in debt

issues But the sign reverses when we scale the debt issues by the sum of debtplus book equity This is telling What is really driving the results is thechange in equity in the denominator not the change in debt in the numer-ator This result is further substantiated in Columns (5) and (6) where thechange in debt to capital ratio is used instead of debt issuanceTo further explore the impact of active issuance relative to passive

experiencing of changes we decompose leverage by separating out theeffect of debt and equity issues from those due to changes in retainedearnings Thus we define leverage change due to external financing as

LevEF frac14Dt1 thorn d

Dt1 thorn Et1 thorn dthorn e

Dt1

Dt1 thorn Et1 eth1THORN

where D is debt E is book value of equity d is debt issues net of retirementsand e is equity issues net of repurchases Garvey and Hanka (1999) use asimilar measure in their study This measure directly captures the netleverage effect of debt and equity issuancesColumns (7) and (8) report results from regressions of the net leverage

effect of debt and equity issues on changes in profitability after controllingfor other leverage factors in differences and year indicator variables Theeffect of profitability on leverage change due to external financing is positiveand statistically significant This comparison suggests that changes in totaldebt and equity are only partly a result of debt and equity issuance decisionsOther balance sheet adjustments complicate the inferences from leverageratio regressionsFor most of our analysis we follow the literature in treating debt as homo-

genous It is of course potentially interesting to ask what happens when wedistinguish between short-term debt and long-term debt Which part of thefirmrsquos debt maturity structure drives the under-adjustment Table XII

PROFITSndashLEVERAGE PUZZLEREVISITED 1443

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

provides results from such decompositions Alternative scalings areconsidered The main results of interest are that the distinction betweenshort- and long-term debt is important Short-term debt issuance is largelyindependent of changes in profits The negative sign on profit change stemsfrom the action of long-term debt This seems intuitively reasonable

Table XI Scaled issuances

The sample contains nonfinancial firms listed on the annual Compustat files for the period

from 1971 to 2009 The table presents estimates from regressions of debt issuance to totalissuance (in columns (1) and (2)) debt issuance to total capitalization (in columns (3) and(4)) change in debt to capitalization ratio (in columns (5) and (6)) and change in leverage

ratio considering only external financing (in columns (7) and (8)) The change in leverageratio considering only external financing is defined as

LevEF frac14Dt1 thorn d

Dt1 thorn Et1 thorn dthorn e

Dt1

Dt1 thorn Et1

where D is debt E is book value of equity d is debt issues net of retirements and e is equity

issues net of repurchases The regressions include lagged leverage factors and year indicatorvariables Some of the specifications include firm fixed effects The reported t-statistics arecorrected for clustering at the firm level

aMean significant at the 1 level

bMean significant at the 5 level

cMean significant at the 10 level

DebtIssTotIss

DebtIssethDthornETHORN D

DthornE LevEF

(1) (2) (3) (4) (5) (6) (7) (8)

Profitt1 004a 000 002a 001c 003a 002 001b 002a

(35) (03) (34) (18) (35) (14) (23) (26)

IndLevt1 004c 001 001 001c 000 000 000 000

(19) (06) (15) (19) (02) (04) (03) (05)

Market=bookt1 001a 001a 000a 000a 000a 000a 000 000c

(47) (85) (68) (60) (55) (36) (05) (18)

Tangt1 004 005 013a 011a 011a 006a 003a 003a

(12) (15) (144) (112) (79) (38) (44) (42)

Assetst1 008a 004a 006a 004a 002a 001a 003a 002a

(97) (45) (283) (161) (53) (35) (147) (101)

Constant 079a 074a 002a 003a 000 000 000b 002a

(497) (442) (91) (111) (00) (05) (22) (93)

Year FE Yes Yes Yes Yes Yes Yes Yes Yes

Firm FE No Yes No Yes No Yes No Yes

R2 0015 0098 0022 0076 0009 0070 0012 0086

N 133905 133905 140552 140552 138229 138229 140572 140572

1444 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

inasmuch as short-term debt is probably more associated with day-to-dayoperations rather than longer term planning considerationsIn Table XIII we examine change in leverage ( D

DthornE) and change inleverage due to external financing (LEVEF) during good and bad timesseparately The traditional negative effect of a change in profits on achange in leverage is found both in good times and in bad timesHowever when we decompose the impact in order to isolate the externalcomponent as before the result disappears In good times the externalfinancing component of a change in profits is weak but still positiveDuring bad times it is not statistically significant This implies that distin-guishing between good and bad times does not fundamentally alter the basic

Table XII Short- and long-term debt issuances

The sample contains nonfinancial firms listed on the annual Compustat files for the period

from 1971 to 2009 The table presents estimates from regressions of short-term debtissuance to total issuance (in columns (1) and (2)) long-term debt issuance to totalissuance (in columns (3) and (4)) short-term debt issuance to total capitalization (in

columns (5) and (6)) and long-term debt to total capitalization (in columns (7) and (8))Some of the specifications include firm fixed effects The reported t-statistics are correctedfor clustering at the firm level

aMean significant at the 1 level

bMean significant at the 5 level

cMean significant at the 10 level

STDebtIssTotIss

LTDebtIssTotIss

STDebtIssethDthornETHORN

LTDebtIssethDthornETHORN

(1) (2) (3) (4) (5) (6) (7) (8)

Profitabilityt1 0002 0003 0042a 0008 0001 0001 0019a 0012a

(02) (04) (30) (05) (14) (10) (45) (28)

IndLevt1 0009 0007 0070b 0033 0001 0002c 0014a 0005

(06) (05) (26) (11) (08) (17) (26) (09)

M=Bt1 0003a 0003a 0003b 0010a 0000 0000 0003a 0002a

(48) (35) (22) (65) (10) (01) (84) (42)

Tangibilityt1 0020 0007 0011 0067 0005a 0006a 0097a 0081a

(12) (03) (03) (16) (34) (33) (108) (89)

Assetst1 0039a 0011b 0034a 0024b 0001a 0001 0046a 0020a

(89) (22) (38) (23) (36) (13) (190) (82)

Constant 0001b 0153a 0798a 0591a 0000c 0006a 0057a 0096a

(25) (144) (470) (288) (17) (124) (292) (234)

Year FE Yes Yes Yes Yes Yes Yes Yes Yes

Firm FE No Yes No Yes No Yes No Yes

R2 0025 0089 0018 0073 0030 0147 0016 0298

N 136146 136146 136140 136140 143031 143031 142847 142847

PROFITSndashLEVERAGE PUZZLEREVISITED 1445

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

Table XIII Externally financed leverage changes in good and bad times

The sample contains nonfinancial firms listed on the annual Compustat files for the period

from 1971 to 2009 The table presents estimates from regressions of change in debt tocapitalization ratio (in columns (1) (2) (5) and (6)) and change in leverage ratio consider-ing only external financing (in columns (3) (4) (7) and (8)) The change in leverage ratio

considering only external financing is defined as

LevEF frac14Dt1 thorn d

Dt1 thorn Et1 thorn dthorn e

Dt1

Dt1 thorn Et1

where D is debt E is book value of equity d is debt issues net of retirements and e is equityissues net of repurchases The regressions include lagged leverage factors and year indicator

variables The reported t-statistics are corrected for clustering at the firm level An industryis defined as having ldquogood timesrdquo if the median firm in that industry has a market-to-bookratio that is higher than the 67th percentile of the time-series distribution of industry

median market-to-book ratios Conversely an industry is defined as having ldquobad timesrdquoif the median firm in that industry has a market-to-book ratio that is lower than the 33rdpercentile of the time-series distribution of industry median market-to-book ratios

aMean significant at the 1 level

bMean significant at the 5 level

cMean significant at the 10 level

Good times Bad times

DDthornE LevEF D

DthornE LevEF

(1) (2) (3) (4) (5) (6) (7) (8)

Profitabiltyt1 0022c 0005 0016c 0025b 0073b 0041 0007 0007

(17) (03) (18) (24) (25) (11) (04) (03)

IndLevt1 0000 0011 0001 0009 0010 0018 0001 0000

(00) (09) (01) (13) (07) (10) (02) (00)

MB

t1

0004a 0003b 0000 0001 0008a 0001 0000 0000

(40) (24) (05) (08) (28) (04) (03) (03)

Tangt1 0097a 0041 0033a 0038b 0127a 0048 0059a 0033

(45) (16) (27) (25) (49) (14) (43) (16)

Assetst1 0014a 0013b 0024a 0019a 0037a 0023c 0036a 0025a

(30) (20) (84) (48) (45) (17) (88) (43)

Constant 0030b 0036 0000 0003 0028c 0014 0021 0001

(20) (16) (00) (07) (17) (06) (11) (01)

Year FE Yes Yes Yes Yes Yes Yes Yes Yes

Firm FE No Yes No Yes No Yes No Yes

R2 0006 0088 0009 0117 0026 0251 0023 0202

N 56820 56820 58200 58200 23462 23462 23511 23511

1446 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

conclusions The traditional leverage result on profits is driven by increasesin equity that are experienced and then partially but not completely offsetby firmsTo the extent that firms are engaging in active financing they are following

the prescriptions of the trade-off theory Thus we have a coherent picture ofhow firms are responding to profits Basically more profitable firms experiencean increase in the value of their equity They respond to this by issuing debtand repurchasing equity But the magnitude of the increase in equity is largerthan the magnitudes of the active steps taken in the debt and equity marketsThis means that the firmsrsquo actual actions are as predicted by the trade-offtheory But it leaves open the question of why they do not act more stronglyThe capital structure literature has long been interested in dynamics and

capital structure adjustment paths In recent years this interest has increas-ingly resulted in explicit dynamic optimizing models such as those ofGoldstein Ju and Leland (2001) Hennessy and Whited (2005) andStrebulaev (2007) These models commonly make specialized assumptionsleading some to question their generalityWe argue that both fixed transactions costs and variable transactions costs

are plausible and realistic in models that pin down firm size The simultaneouspresence of both is important Due to fixed costs firms only rebalance inter-mittently Due to variable adjustment costs the extent of the adjustment isreduced to the point at which an extra unit to adjustment is balanced againstthe extra unit transactions costs This means that both leverage increases andleverage reductions will only go part way toward a static optimumThe key point is that with both fixed and variable transactions costs we

should expect to only see intermittent leverage adjustments When leverage isadjusted it should ldquounder adjustrdquo when compared to a frictionless modelFurther discussion of transactions costs in a capital structure setting isprovided by Leary and Roberts (2005)

9 Conclusion

The connection between corporate profits and capital structure has beenvery influential in the assessment of the trade-off theory The conventionalinterpretation of the evidence has pushed the literature away from the trade-off theory and toward much more complex dynamic models and ideas As aresult it is important to make sure that the evidence has been correctlyinterpreted We show that the problem is with the interpretation of thetest results and a failure to take into account the costs associated with re-sponding to a profit shock

PROFITSndashLEVERAGE PUZZLEREVISITED 1447

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

We document that more profitable firms really do borrow more (not less)They tend to repurchase equity They experience an increase in both thebook value of equity and the market value of equity Less profitable firmsreally do tend to reduce their debt and to issue equityFirm size and market conditions also matter Larger firms tend to be more

active in the debt markets while smaller firms tend to be relatively moreactive in the equity markets During good times there is more use ofexternal financingThe usual profitsndashleverage puzzle result is primarily driven by the increase

in equity that is experienced by the more profitable firms Accordingly thepuzzle should be restated as asking why do firms not take sufficiently largeoffsetting actions to fully undo the change in equity What limits the mag-nitudes of the typical leverage response to profit shocksIn a frictionless model the partial response appears puzzling But there is

good empirical reason to believe that rebalancing entails both fixed andvariable costs and that firm size matters The rebalancing costs can befully avoided by doing nothing Accordingly the firm must decide whetherany given shock is big enough to be worth responding to If it is then thefirm must decide how big a response is called for These technical conditionsare known as ldquovalue matchingrdquo and ldquosmooth pastingrdquo Thus optimizationimplies that some shocks will be ignored Even if the shock is not ignoredthe optimal response will only partially undo the shock The magnitude ofthe leverage response must balance the marginal cost and the marginalbenefit of an extra unit of leverage Since the marginal cost of adjustingleverage is strictly positive the adjustment toward that static leverageoptimum will only go part way This is true both for leverage increasesand for leverage reductionsOne might conjecture that the variable costs would be time varying and

that such costs might have declined over time Since we do not have a fullaccounting of all of the variable costs it is of course impossible to be sureHowever there is some evidence suggesting that at least the directunderwriting fees are remarkably stable over time10

Surely nobody ought to be surprised that more profitable firms are moreinclined to repurchase equity while unprofitable firms tend not to do so Thefact that more profitable firms typically experience an increase in the value of

10 For IPOs Jay Ritter provides important summary evidence as httpbearwarringtonufleduritterIPOs2013Underwritingpdf He finds that for medium-sized deals a remark-

ably large fraction of the deals have a 7 fee year after year Whether this is due tobehavioral factors or some sort of a coordination device is not clear The extent towhich similar intertemporal stability carries over to other forms of external financemerits further study

1448 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

equity is equally natural Were it not for the contrary claims in the literature(ie the profitsndashleverage puzzle) we would argue that nobody ought to findit surprising that more profitable firms are more inclined to issue debt Thefact that leverage adjustments typically only partially offset profit shocksshould also sound quite natural These facts all fit together easily within thetrade-off theory once we take account of both fixed and variable costs ofactively adjusting capital structure

Appendix

Table AI Market leverage regressions

The table presents estimates of the regressions of market leverage on firm characteristics

The sample comes from the annual Compustat files during the period 1971ndash2009 Financialfirms are excluded The leverage is estimated as the ratio of debt over debt plus marketequity The explanatory variables Profitabilityt1

MB

t1

Tangibilityt1 and LnethAssetsTHORNt1are described in Table 1 IndMedianLevt1 is estimated as the median market leverage of allother firms in the same industry The industry is defined at the level of the four-digit SICcode All specifications include the year fixed effects The specifications in column (2)

additionally include the firm fixed effects We report t-statistics where the standard errorsare clustered at the firm level in parentheses

aSignificant at the 1 level

Quantile regressions

(1) (2) (3) (4) (5)

Profitabilityt1 013a 013a 007a 011a 020a

(275) (262) (349) (425) (538)

IndLevt1 046a 020a 032a 058a 063a

(506) (239) (1299) (1630) (1042)MB

t1

003a 001a 001a 002a 004a

(509) (303) (743) (709) (534)

Tangibilityt1 013a 018a 013a 014a 010a

(191) (181) (762) (571) (232)

LnethAssetsTHORNt1 001a 005a 001a 001a lt001a

(112) (258) (682) (342) (29)

Constant 010a 011a 005a 005a 029a

(171) (98) (173) (118) (398)

Year FE Yes Yes Yes Yes Yes

Firm FE No Yes No No No

Regression type OLS FE 25th tile Median 75thtile

Clustered SE Yes Yes No No No

R2 030 067

Pseudo R2 016 023 0201

N 158578 158578 158578 158578 158578

PROFITSndashLEVERAGE PUZZLEREVISITED 1449

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

Table AII Profitability sorts for firms issuing large amounts

Percentage of firms

Iss Ret Iss Iss Rep Iss Iss IssD IssE

D D NetD E E NetE DampE RepE RetD

(1) (2) (3) (4) (5) (6) (7) (8) (9)

Panel A 10 cutoff to define issuing or retiring activity

Allfirm

s

Low 205 150 134 260 05 257 45 00 23

2 207 175 103 77 06 75 11 01 10

3 243 196 101 50 08 47 08 01 09

4 252 186 113 50 13 48 08 02 08

High 227 161 118 60 27 57 07 04 09

t Highfrac14Low 70 38 62 755 226 768 320 90 152

Smallfirm

s Low 230 132 162 380 05 375 82 00 32

2 228 143 165 334 04 333 65 00 26

3 211 146 147 216 03 214 38 00 23

4 190 154 120 126 04 124 15 01 20

High 194 162 114 99 09 97 12 01 15

t Highfrac14Low 42 40 68 342 20 340 169 07 131

Largefirm

s Low 231 200 82 31 06 29 08 01 04

2 231 187 67 17 06 14 03 01 02

3 234 177 77 17 09 14 03 02 03

4 218 158 75 16 17 14 02 03 02

High 218 135 99 19 41 15 03 07 02

t Highfrac14Low 19 106 36 46 135 57 39 58 23

Percentage of firms

Iss Ret Iss Iss Rep Iss Both Iss D Iss E

D D Net D E E Net E DampE Rep E Ret D

(1) (2) (3) (4) (5) (6) (7) (8) (9)

Panel B 15 cutoff to define issuing or retiring activity

Allfirm

s

Low 158 101 100 230 01 228 27 00 15

2 142 115 63 57 01 56 05 00 06

3 165 128 57 32 01 31 03 00 05

4 173 118 65 35 01 34 03 00 05

High 160 102 70 46 04 44 03 00 05

t Highfrac14Low 08 06 141 738 87 746 277 24 131

Smallfirm

s Low 195 95 132 354 01 350 56 00 24

2 185 96 131 303 01 301 44 00 18

3 160 96 104 179 00 176 18 00 15

4 134 99 78 99 01 98 08 00 12

High 133 95 73 82 01 80 06 00 10

t Highfrac14Low 80 00 95 345 12 344 149 12 52

Largefirm

s Low 145 123 42 14 01 12 03 00 01

2 137 114 32 07 01 06 01 00 01

3 142 104 35 06 01 05 00 00 01

4 130 91 36 08 03 07 01 01 01

High 134 81 48 09 06 07 01 01 01

t Highfrac14Low 19 85 17 26 54 33 34 17 11

1450 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

Table AIV Debt and equity changesmdashwithout fixed effects

The sample contains nonfinancial firms listed on the annual Compustat files for the period from 1971

to 2009 The table presents estimates from regressions of change in debt (D) change in the book

value of equity (BVE) change in the market value of equity (MVE) and net equity issuances

(EqyIss) The explanatory variables include current and lagged change in operating income before

depreciation (Profits and Profitst1) and lagged change in book value of assets (Assetst1) The

regressions include year indicator variables The reported t-statistics are corrected for clustering at the

firm level

aMean significant at the 1 level

bMean significant at the 5 level

cMean significant at the 10 level

D BVE MVE EqyIss

(1) (2) (3) (4) (5) (6) (7) (8)

Profitst1 025a 024a 023a 038a 007 066a 007a 009a

(40) (32) (36) (39) (06) (29) (42) (51)Profits 051b 082a 149a 004a

(25) (33) (49) (33)Assetst1 002 005c 024a 001b

(08) (17) (34) (23)Constant 93 176a 45 71 1552a 1704a 130a 140a

(12) (28) (07) (09) (49) (49) (45) (45)Year FE Yes Yes Yes Yes Yes Yes Yes YesR2 001 005 001 011 001 005 002 002N 162056 162056 162130 162130 157550 157550 162154 162154

Table AIII Median profitability sorts

The sample contains nonfinancial firms listed on the annual Compustat files for the period from 1971

to 2009 The table presents median debt and equity median changes in debt and equity and median

equity issuances (EqyIss) in millions of dollars for all firms sorted on profitability and for profitability

sorts within the smallest and largest firms The sorts are done annually

D D BVE BVE MVE MVE EqyIss Assets(1) (2) (3) (4) (5) (6) (7) (8)

Allfirm

s

Low 21 00 75 17 256 20 01 1922 186 02 449 18 598 19 00 11033 816 06 1274 03 1637 05 00 32244 787 02 1541 18 2488 03 00 3556High 273 00 1361 66 3300 14 00 2578

Smallfirm

s Low 05 00 13 07 140 14 06 402 06 00 25 08 114 12 01 563 09 00 32 05 85 07 00 674 12 00 39 01 68 03 00 79High 10 00 51 04 98 01 00 90

Largefirm

s Low 11485 381 10269 371 13227 202 03 333902 13051 393 12786 140 17571 169 05 392463 11554 287 13429 24 21623 305 00 370314 9963 209 15052 194 29424 308 00 36705High 7361 00 16380 613 49142 439 00 34976

PROFITSndashLEVERAGE PUZZLEREVISITED 1451

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

References

Auerbach A J (1985) Real determinants of corporate leverage in B M Friedman (ed)

Corporate Capital Structures in the United States University of Chicago Press Chicago

IllinoisBertrand M and Mullainathan S (2003) Enjoying the quiet life corporate governance and

managerial preferences Journal of Political Economy 111 1043ndash1075Booth L Varouj A Demirguc-Kunt A and Maksimovic V (2001) Capital structures in

developing countries Journal of Finance 56 87ndash130Brennan M J and Schwartz E S (1984) Optimal financial policy and firm valuation

Journal of Finance 39 593ndash607Cameron A C and Trivedi P K (2010) Microeconometrics using Stata Stata Press TexasConstantinides G M (1979) Multiperiod consumption and investment behavior with

convex transactions costs Management Science 25 1127ndash1137Constantinides G M and Richard S F (1978) Existence of optimal simple policies for

discounted-cost inventory and cash management in continuous time Operations Research

26 620ndash636DeAngelo H DeAngelo L and Whited T M (2011) Capital structure dynamics and

transitory debt Journal of Financial Economics 99 235ndash261Dixit A K (1993) The Art of Smooth Pasting Routledge London and New YorkFama E and French K R (2002) Testing trade-off and pecking order predictions about

dividends and debt Review of Financial Studies 15 1ndash33Faulkender M and Petersen M A (2006) Does the source of capital affect capital struc-

ture Review of Financial Studies 19 45ndash79Faulkender M Flannery M J Hankins K W and Smith J M (2012) Cash Flows and

Leverage Adjustments Journal of Financial Economics 103 632ndash646Fischer E Heinkel R and Zechner J (1989) Dynamic capital structure choice theory

and tests Journal of Finance 44 19ndash40Frank M Z and Goyal V K (2008) Trade-off and pecking order theories of

debt in B E Eckbo (ed) Handbook of Corporate Finance Empirical Corporate Finance

Vol 2 Handbook of Finance Series Chapter 12 ElsevierNorth-Holland Amsterdam pp

135ndash202

Frank M Z and Goyal V K (2009) Capital structure decisions which factors are reliably

important Financial Management 38 1ndash37

Garvey G T and Hanka G (1999) Capital structure and corporate control the effect of

antitakeover statutes on firm leverage Journal of Finance 54 519ndash546

Goldstein R Ju N and Leland H (2001) An EBIT-based model of dynamic capital

structure Journal of Business 74 483ndash512

Graham J R and Tucker A L (2006) Tax shelters and corporate debt policy Journal of

Financial Economics 81 563ndash594

Hackbarth D and Mauer D C (2012) Optimal priority structure capital structure and

investment Review of Financial Studies 25 747ndash796

Halling M Yu J and Zechner J (2012) Leverage dynamics over the business cycle

Working paper University of Utah University of New South Wales and Vienna

University of Economics and Business AdministrationHennessy C A and Whited T A (2005) Debt dynamics Journal of Finance 60 1129ndash1165

Huang R and Ritter J R (2009) Testing theories of capital structure and estimating the

speed of adjustment Journal of Financial and Qualitative Analysis 44 237ndash271

1452 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

Hovakimian A Opler T and Titman S (2001) The debt-equity choice Journal of

Financial and Quantitative Analysis 36 1ndash24Jung K Kim Y and Stulz R M (1996) Timing investment opportunities managerial

discretion and the security issue decisions Journal of Financial Economics 42 159ndash185Kane A Marcus A J and McDonald R L (1984) How big is the tax advantage to

debt Journal of Finance 39 841ndash853Kayhan A and Titman S (2007) Firmsrsquo histories and their capital structures Journal of

Financial Economics 83 1ndash32Leary M T (2009) Bank loan supply lender choice and corporate capital structure

Journal of Finance 64 1143ndash1185Leary M T and Roberts M R (2005) Do firms rebalance their capital structures

Journal of Finance 60 2575ndash2619Leland H E (1994) Corporate debt value bond covenants and optimal capital structure

Journal of finance 49 1213ndash1252Lemmon M L Roberts M R and Zender J F (2008) Back to the beginning persist-

ence and the cross-section of corporate capital structure Journal of Finance 63 1575ndash1608Long M S and Malitz I B (1985) The investment-financing nexus some empirical

evidence Midland Corporate Finance Journal 3 53ndash59Mackie-Mason J K (1990) Do taxes affect corporate financing decisions Journal of

Finance 45 1471ndash1493Morellec E Nikolov B and Schurhoff N (2012) Corporate governance and capital

structure dynamics Journal of Finance 67 803ndash848Myers S C (1984) The capital structure puzzle Journal of Finance 39 575ndash592Myers S C (1993) Still searching for optimal capital structure Journal of Applied

Corporate Finance 61 4ndash14Rajan R and Zingales L (1995) What do we know about capital structure some evidence

from international data Journal of Finance 50 1421ndash1460Strebulaev I A (2007) Do tests of capital structure theory mean what they say Journal of

Finance 62 1747ndash1787

Strebulaev I A and Whited T M (2012) Dynamic Models and Structural Estimation in

Corporate Finance Now Publisher Boston and Delft

Titman S and Wessels R (1988) The determinants of capital structure choice Journal of

Finance 43 1ndash21

Welch I (2004) Capital structure and stock returns Journal of Political Economy 112

106ndash131

Welch I (2011) Two common problems in capital structure research the financial-debt-

to-asset ratio and issuing activity versus leverage changes International Review of Finance

11 1ndash17

PROFITSndashLEVERAGE PUZZLEREVISITED 1453

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

the leverage ratios is primarily driven by the changes in the market value ofequity in the denominator rather than by the changes in debt in the numer-ator Since equity issues are often small this implies that the variation in theleverage ratio is primarily driven by internal operations rather than byexternal financing actions This again points to the fact that leverageratios can provide a misleading account of actual patterns in the dataThere is always a potential concern that the averages may be misleading

due to the impact of outliers To address this concern in AppendixTable AIII we present median values of the profitability sorts Theseresults generally reinforce the findings in Table VTable VI again considers the magnitudes of financing activity but this

time the issues are scaled by a measure of firm size That way it is easier to

Table V Magnitude of financing activity

The sample contains nonfinancial firms listed on the annual Compustat files for the period

from 1971 to 2009 The table presents average debt and equity changes in debt and equityand equity issuances (in millions of dollars) for firms sorted on profitability within sizeclasses The table reports information for all firms sorted on profitability and for profit-

ability sorts within the smallest and largest firms The sorts are done annually The tablealso reports results from t-test for the difference in percentage of firms issuing or retiringsecurities in the lowest and the highest profit quintile

D D BVE BVE MVE MVE EqyIss Assets

(1) (2) (3) (4) (5) (6) (7) (8)

Allfirm

s

Low 83 5 78 10 120 5 7 234

2 605 13 509 24 721 8 17 1691

3 1227 9 1123 22 1834 3 9 3551

4 833 23 1166 7 2408 74 9 2994

High 570 53 1171 71 3817 68 36 2581

t Highfrac14Low 31 9 36 9 38 2 16 36

Smallfirm

s Low 25 04 42 01 332 28 33 97

2 29 06 55 02 278 16 21 116

3 31 07 63 01 212 05 12 127

4 30 06 70 05 179 10 08 136

High 24 06 84 14 234 17 07 142

t Highfrac14Low 04 15 153 85 73 11 201 107

Largefirm

s Low 3929 55 2990 152 4671 11 63 10846

2 3826 19 3178 95 5122 66 11 10721

3 3074 27 3498 41 6173 242 9 9697

4 2623 62 3915 45 8473 94 57 9855

High 2180 193 4642 261 14924 202 148 10171

t Highfrac14Low 94 64 108 87 246 12 169 16

1434 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

see whether large firms behave disproportionately when compared withsmall firms Column (1) shows that low-profit firms have more debtas reported in many previous studies In Columns (4) and (6) we seethat low-profit firms have a major decline in both the book value ofequity and the market value of equity The high-profit firms have thereverse experienceMore interesting are the active decisions of the firms In Column (2) we see

that in contrast to the usual impression the low-profit firms are reducingdebt while the high-profit firms are increasing it In Column (7) we see thatthe low-profit firms are also reducing leverage by issuing equity The high-profit firms issue little equity In the lower panel similar results are foundwhen we separate small and large firms Equity issuing is particularly strong

Table VI Magnitude of financing activitymdashscaled levels and changes

The sample contains nonfinancial firms listed on the annual Compustat files for the period

from 1971 to 2009 The table presents average debt and equity changes in debt and equityand equity issuances scaled by assets for firms sorted on profitability within size classes Thetable reports information for all firms sorted on profitability and for profitability sorts

within the smallest and largest firms The sorts are done annually The table also reportsresults from t-test for the difference in percentage of firms issuing or retiring securities inthe lowest and the highest profit quintile

D D BVE BVE MVE MVE EqyIss

(1) (2) (3) (4) (5) (6) (7)

Allfirm

s

Low 386 122 12 871 3427 1868 217

2 285 58 407 119 940 157 28

3 293 12 425 14 742 03 12

4 269 26 461 02 955 27 10

High 192 15 551 47 1675 74 07

t Highfrac14Low 35 49 50 47 127 42 157

Smallfirm

s Low 545 154 841 2145 9688 9128 696

2 684 49 712 1915 5014 2318 288

3 331 253 141 969 3063 954 129

4 267 181 357 175 1673 221 55

High 194 05 543 26 1758 147 41

t Highfrac14Low 49 17 51 36 88 27 99

Largefirm

s Low 350 30 306 58 511 55 09

2 346 14 335 13 523 02 04

3 353 165 332 71 711 29 01

4 280 01 414 04 909 09 05

High 227 15 477 19 1643 27 14

t Highfrac14Low 371 112 453 181 570 57 221

PROFITSndashLEVERAGE PUZZLEREVISITED 1435

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

among the small low-profit firms that are also experiencing sharp reductionsin the value of equity In column (2) we see that both for the large and for thesmall firms net debt increases are primarily found in the highest quintile offirm profitability

6 Debt and Equity Regressions

So far we have demonstrated that in our data the conventional leverageregressions have the usual signs We then demonstrated that in simple sortsof the data strikingly different active financing patterns emerge The nexttask is to reconcile these differencesTable VII presents simple regressions predicting changes in debt book

equity and market equity as well as equity issuances Consistent with thesorts we control for changes in firm size and we also include firm and yeardummies There may be a concern about the speed of a firmrsquos responsesAccordingly we consider the lagged change in profits as the main case butwe also include the contemporaneous change in profits Whether the contem-poraneous change in profits is included or not does not in fact matter Wereport panel-robust standard errors adjusted for clustering at the firm levelThe results in Table VII support the findings in the sorts Firms that have

an increase in profits in 1 year increase debt both in the same year and in thenext year Firms that have an increase in profits experience an increase inboth the book value of equity and the market value of equity Importantlyhowever firms that experience and increase in profits reduce their equityissues All these effects are statistically significant (except for the marketvalue of equity in the column 5 specification) and seem intuitively reason-able In Appendix Table AIV similar regressions to those presented aboveare reported but without firm-fixed effects We find that empirically verylittle changes whether we include firm fixed effects or leave them out Thusthe targeting behavior to the extent that it happens does not account forwhat we are reportingThe next step is to bring the sorts and the simple regressions together into

a common setting We therefore include factors in addition to profitability(i) median industry leverage (ii) the market-to-book assets ratio (iii) tangi-bility of assets and (iv) firm size (measured by the log of assets) Rajan andZingales (1995) show that these factors are related to leverage in G7countries A number of studies have used these factors to estimateleverage targets Frank and Goyal (2009) show that these factors arerobustly related to leverage in the USA In these regressions we use indica-tors for the quintile that the firm is in for each factor

1436 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

Table VIII reports these results In Columns (1) and (2) we examine debtissues and again show that profits positively affect debt issuances The effectsare large and statistically significant at the 1 level Columns (3) and (4)which examine book equity show a strong effect of profits This effect is onlyslightly reduced by the inclusion of other factors Columns (5) and (6) illus-trate that the impact of profits on the change in market equity is also robustto the inclusion of the conventional factors Columns (7) and (8) examineequity issuances Here we again find that profits have a significant negativeimpact on equity issuances As before the results are robust to the inclusionof the conventional factorsIn unreported tables we estimate debt and equity changes using quantile

regressions In terms of the signs and significance of the profitability variable

Table VII Debt and equity changes

The sample contains nonfinancial firms listed on the annual Compustat files for the period

from 1971 to 2009 The table presents fixed effect estimates from regressions of changes indebt (D) changes in the book value of equity (BVE) changes in the market value ofequity (MVE) and net equity issuances (EqyIss) The explanatory variables include

current and lagged change in operating income before depreciation (Profits andProfitst1) and lagged change in the book value of assets (Assetst1) In addition tofixed firm effects the regressions include year indicator variables The reported t-statistics

are corrected for clustering at the firm level

aMean significant at the 1 level

bMean significant at the 5 level

cMean significant at the 10 level

D BVE MVE EqyIss

(1) (2) (3) (4) (5) (6) (7) (8)

Profitst1 022a 025a 015b 035a 004 060b 005b 007a

(33) (31) (25) (35) (03) (25) (25) (53)

Profits 050b 078a 140a 003a

(24) (32) (48) (34)

Assetst1 000 007b 025a 001b

(00) (21) (35) (25)

Constant 167 283a 34 51 1145a 1182a 605a 595a

(14) (30) (04) (05) (32) (29) (69) (74)

Year FE Yes Yes Yes Yes Yes Yes Yes Yes

Firm FE Yes Yes Yes Yes Yes Yes Yes Yes

R2 001 004 001 010 001 005 002 003

N 162056 162056 162130 162130 157550 157550 162154 162154

PROFITSndashLEVERAGE PUZZLEREVISITED 1437

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

the results are qualitatively similar to the OLS results reported in Table VIIIProfitability positively affects debt issuances and negatively affects equityissuances across various quantiles Overall the results in Table VIII showthat there is nothing that is special about using sorts or regressions to explaindebt and equity issues In either case we find that more profitable firms tendto increase their debt experience an increase in the value of equity andrepurchase shares Thus the control factors are not responsible for theusual rejection of the trade-off theory It would appear that what mattersis the size of the issuing activity relative to the organic increases in equityvalue due to profits In other words what matters in the leverage regressionseems to be coming from the presence of E in D

DthornE and not from D

Table VIII Debt and equity issuances profitability and leverage factor quintiles

The sample contains nonfinancial firms listed on the annual Compustat files for the period

from 1971 to 2009 The table presents estimates from regressions of change in debt (D)change in the book value of equity (BVE) change in the market value of equity (MVE)and net equity issuances (EquityIss) The explanatory variables include lagged profitability

quintiles and lagged leverage factor quintiles In addition to fixed firm effects the regres-sions include year indicator variables The reported t-statistics are corrected for clustering atthe firm level

aMean significant at the 1 level

bMean significant at the 5 level

cMean significant at the 10 level

D BVE MVE EqyIss

(1) (2) (3) (4) (5) (6) (7) (8)

ProfitabilityQint 214a 167a 283a 175a 80 429a 53a 55a

(70) (54) (66) (59) (08) (48) (55) (62)

IndLevQint 104a 43 99 20

(29) (15) (10) (10)

Ln(Assets)Qint52 17 1026a 53a

(16) (05) (77) (35)MB

Qint158a 372a 1155a 09

(45) (56) (106) (06)

TangibilityQint 05 177 436c 64a

(01) (15) (18) (31)

Constant 167 96 656b 235 4613a 8610a 113 570a

(08) (04) (26) (04) (61) (58) (11) (31)

Year FE Yes Yes Yes Yes Yes Yes Yes Yes

Firm FE Yes Yes Yes Yes Yes Yes Yes Yes

R2 0002 0002 0003 0004 0013 0014 0013 0014

N 155587 155037 155587 155037 155587 155037 155587 155037

1438 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

7 Are Financial Market Conditions Important

According to the trade-off theory capital structure is determined by con-sidering the impact of costs and benefits of debt The time variation inthe costs and benefits of debt imply that issuance decisions would alsovary over time Furthermore market conditions affect the transactioncosts of adjusting Adverse selection is a greater problem in a cold marketthan in a hot one Accordingly it may be easier to adjust in good marketconditions Related ideas have been developed by Huang and Ritter(2009) Faulkender et al (2012) and Halling Yu and Zechner (2012)Faulkender Flannery Hankins and Smithrsquos paper examines cross-sectionalvariation in adjustment speeds including market valuations andhow they may affect adjustment speeds It is natural to think that ingood market conditions it will be less costly to issue both debt and equityand adjusting leverage to respond to profitability shocks would be mucheasierTo test the importance of market conditions we require a definition of

good times and bad times Our empirical strategy is to estimate good timesversus bad times at the four-digit industry level We define an industry ashaving ldquogood timesrdquo if the median firm in that industry has a market-to-book ratio that is higher than the 67th percentile of the time-series distribu-tion of the industry median market-to-book ratios Conversely an industryis defined as having ldquobad timesrdquo if the median firm in the industry has amarket-to-book ratio that is lower than the 33rd percentile of the time-seriesdistribution of the industry median market-to-book ratiosPanel A of Table IX tabulates issuance activity for profitability sorts in

both good times and bad times As might be expected active debt and equityissuances are larger during good times There is also somewhat more activeswaps between debt and equity during good timesPanel B reports the issuance activity for the smallest and the largest firms

during the good times and bad times Within each size quintile firms aresorted on profitability This is a three-way sort As before we find that thereis much more active use of external markets during good times Small low-profit firms are more likely to issue equity in good times than in bad timesDebt issuances are significantly higher in good times Large high-profitfirms are significantly more likely to issue debt and repurchase equity ingood times than in bad timesPanel C reports the magnitudes of the financial variables rather than the

frequencies During bad times less profitable large firms retire substantialamounts of debt and they show a tendency to issue equity Small firms donot seem to engage in similar debt reduction activities However like the

PROFITSndashLEVERAGE PUZZLEREVISITED 1439

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

Table IX Debt and equity issues in good and bad times

The table reports the frequency and magnitude of financing activity for sorts on profits for subsamples

of firms in good and bad times An industry is defined as having ldquogood timesrdquo if the median firm in

that industry has a market-to-book ratio that is higher than the 67th percentile of the time-series

distribution of industry median market-to-book ratios Conversely an industry is defined as having

bad times if the median firm in that industry has a market-to-book ratio that is lower than the 33rd

percentile of the time-series distribution of industry median market-to-book ratios Panels A and B

report the percentage of firms (i) issuing net debt in excess of 5 of the value of their assets (ii) issuing

net equity in excess of 5 of the value of their assets (iii) issuing debt and repurchasing equity both in

excess of 5 of the value of their assets and (iv) issuing equity and retiring debt both in excess of 5

of the value of their assets Panel C tabulates the mean debt and equity levels and the mean changes in

debt and equity issuances The sample contains nonfinancial firms listed on the annual Compustat files

for the period from 1971 to 2009

Iss Iss Iss D Iss ENet D Net E Rep E Ret D(1) (2) (3) (4)

Panel A sorts on profitability

Badtimes

Low 145 131 01 202 170 49 02 083 173 46 04 074 182 47 03 10

High 177 51 06 10t Highfrac14Low 41 136 33 01

Goodtimes

Low 206 381 01 462 197 148 02 233 212 103 04 184 234 92 10 18

High 216 92 16 14t Highfrac14Low 23 658 138 23

Panel B sorts on firm size and profitability

Badtimes

Smallfirm

s Low 151 318 00 312 182 276 02 253 151 140 02 294 183 73 00 19

High 177 80 00 11t Highfrac14Low 10 90 NA 20

Largefirm

s Low 172 54 01 042 148 39 04 053 150 43 04 044 156 33 03 06

High 147 28 05 03t Highfrac14Low 17 32 20 02

Goodtimes Smallfirm

s Low 213 448 01 542 230 443 01 473 222 325 00 474 203 197 02 38

High 190 153 02 31t Highfrac14Low 19 230 10 39

Largefirm

s Low 203 96 03 122 208 66 09 133 232 54 16 064 215 39 19 05

High 234 36 38 04t Highfrac14Low 29 99 92 38

(continued)

1440 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

larger firms there is a tendency for small firms to issue more equitymdashat leastto the extent possibleDuring good times large profitable firms raise significant amounts of debt

as they experience an increase in the value of their equity Such firms alsoengage in share repurchases Debt issues by smaller firms are much lessaffected by their own profitability during good times and their change indebt is more or less independent of their profits However equity issuanceis primarily found among the low-profit small firms during good timesTable IX shows that empirically profitability has a time-varying impacton leverage choicesIn Table X the basic regressions for changes in debt and for equity

issuance are presented for both good times and bad times Comparinggood times with bad times we see that the effects of profits are muchstronger in good times In bad times the impacts are rather weak statisticallyThis difference is interesting and deserves further study However for ourpurposes recall that the static trade-off theory implies time-varying capital

Table IX Continued

D D BVE BVE MVE MVE EqyIss Assets

(1) (2) (3) (4) (5) (6) (7) (8)

Panel C magnitude of financing activity

Badtimes Smallfirm

s Low 40 03 58 14 305 06 22 140

2 44 05 82 07 240 67 17 180

3 45 04 94 02 176 18 05 187

4 43 09 98 02 137 04 03 188

High 36 10 114 13 209 16 04 199

Largefirm

s Low 3365 73 2523 169 2304 163 94 8972

2 3031 95 2983 35 3284 278 56 9068

3 2930 5 3360 14 3947 308 57 9225

4 2281 6 3817 19 5213 442 7 8902

High 1790 62 4767 13 8799 58 31 9573

Goodtimes

Smallfirm

s Low 24 06 38 05 339 23 39 90

2 26 08 48 01 301 04 25 104

3 30 09 60 04 246 03 17 125

4 28 07 66 08 206 18 12 130

High 23 06 85 16 261 13 08 141

Largefirm

s Low 4532 36 3477 103 6557 260 53 12465

2 4580 58 3527 139 6911 421 23 12522

3 2952 111 3485 2 7586 279 32 9535

4 2752 124 3939 72 10686 33 105 10226

High 2429 228 4819 377 18565 268 222 10830

PROFITSndashLEVERAGE PUZZLEREVISITED 1441

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

structure choices even when the target is time invariant Empirically we doobserve time-varying choices

8 Is Scaling Important

In the preceding section we provided results for unscaled debt and equityissues This seems appropriate to us However it involves making twochanges to the standard regression We would like to determine if both

Table X Debt and equity changes in good and bad timesmdashregressions

The sample contains nonfinancial firms listed on the annual Compustat files for the period

from 1971 to 2009 The table presents estimates from regressions of change in debt (D)and net equity issuances (EqyIss) in good and bad times The explanatory variables includechange in profits lagged change in profits and lagged change in assets The regressions

include both firm and year fixed effects An industry is defined as having ldquogood timesrdquo ifthe median firm in that industry has a market-to-book ratio that is higher than the 67thpercentile of the time-series distribution of industry median market-to-book ratios

Conversely an industry is defined as having bad times if the median firm in thatindustry has a market-to-book ratio that is lower than the 33rd percentile of the time-series distribution of industry median market-to-book ratios The reported t-statistics arecorrected for clustering at the firm level

aMean significant at the 1 level

bMean significant at the 5 level

cMean significant at the 10 level

Good times Bad times

D EqyIss D EqyIss

(1) (2) (3) (4) (5) (6) (7) (8)

Profits 087a 091a 006b 009a 060 065 003 003b

(30) (29) (20) (28) (15) (16) (16) (20)

Profitst1 029c 015a 034c 001

(19) (37) (17) (04)

Assetst1 000 002b 000 001

(02) (22) (00) (07)

Constant 1337 3989 8691a 10001a 1351 1851 4099a 4055a

(03) (09) (55) (37) (09) (13) (80) (78)

Year FE Yes Yes Yes Yes Yes Yes Yes Yes

Firm FE Yes Yes Yes Yes Yes Yes Yes Yes

R2 0071 0080 0018 0045 0178 0212 0034 0041

N 74827 68729 74891 68787 26965 26013 26971 26018

1442 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

changes are crucial Accordingly in this section we take a step back toconsider prior literature in which variables are scaled The questions arehow does scaling affect the results and whyTable XI examines alternative scalings Columns (1) and (2) scale debt and

equity issuances by the total firm net issuance Columns (3) and (4) scaledebt and equity issuances by the sum of debt and book equity Columns (5)and (6) consider the change in debt scaled by the sum of debt and bookequity Columns (7) and (8) consider the change in leverage due to externalfinancing Since we are examining changes we use the changes in thestandard regression factors as controls The primary interest is in the coef-ficient on the change in profitsWe find that a change in profits is associated with an increase in debt

issues But the sign reverses when we scale the debt issues by the sum of debtplus book equity This is telling What is really driving the results is thechange in equity in the denominator not the change in debt in the numer-ator This result is further substantiated in Columns (5) and (6) where thechange in debt to capital ratio is used instead of debt issuanceTo further explore the impact of active issuance relative to passive

experiencing of changes we decompose leverage by separating out theeffect of debt and equity issues from those due to changes in retainedearnings Thus we define leverage change due to external financing as

LevEF frac14Dt1 thorn d

Dt1 thorn Et1 thorn dthorn e

Dt1

Dt1 thorn Et1 eth1THORN

where D is debt E is book value of equity d is debt issues net of retirementsand e is equity issues net of repurchases Garvey and Hanka (1999) use asimilar measure in their study This measure directly captures the netleverage effect of debt and equity issuancesColumns (7) and (8) report results from regressions of the net leverage

effect of debt and equity issues on changes in profitability after controllingfor other leverage factors in differences and year indicator variables Theeffect of profitability on leverage change due to external financing is positiveand statistically significant This comparison suggests that changes in totaldebt and equity are only partly a result of debt and equity issuance decisionsOther balance sheet adjustments complicate the inferences from leverageratio regressionsFor most of our analysis we follow the literature in treating debt as homo-

genous It is of course potentially interesting to ask what happens when wedistinguish between short-term debt and long-term debt Which part of thefirmrsquos debt maturity structure drives the under-adjustment Table XII

PROFITSndashLEVERAGE PUZZLEREVISITED 1443

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

provides results from such decompositions Alternative scalings areconsidered The main results of interest are that the distinction betweenshort- and long-term debt is important Short-term debt issuance is largelyindependent of changes in profits The negative sign on profit change stemsfrom the action of long-term debt This seems intuitively reasonable

Table XI Scaled issuances

The sample contains nonfinancial firms listed on the annual Compustat files for the period

from 1971 to 2009 The table presents estimates from regressions of debt issuance to totalissuance (in columns (1) and (2)) debt issuance to total capitalization (in columns (3) and(4)) change in debt to capitalization ratio (in columns (5) and (6)) and change in leverage

ratio considering only external financing (in columns (7) and (8)) The change in leverageratio considering only external financing is defined as

LevEF frac14Dt1 thorn d

Dt1 thorn Et1 thorn dthorn e

Dt1

Dt1 thorn Et1

where D is debt E is book value of equity d is debt issues net of retirements and e is equity

issues net of repurchases The regressions include lagged leverage factors and year indicatorvariables Some of the specifications include firm fixed effects The reported t-statistics arecorrected for clustering at the firm level

aMean significant at the 1 level

bMean significant at the 5 level

cMean significant at the 10 level

DebtIssTotIss

DebtIssethDthornETHORN D

DthornE LevEF

(1) (2) (3) (4) (5) (6) (7) (8)

Profitt1 004a 000 002a 001c 003a 002 001b 002a

(35) (03) (34) (18) (35) (14) (23) (26)

IndLevt1 004c 001 001 001c 000 000 000 000

(19) (06) (15) (19) (02) (04) (03) (05)

Market=bookt1 001a 001a 000a 000a 000a 000a 000 000c

(47) (85) (68) (60) (55) (36) (05) (18)

Tangt1 004 005 013a 011a 011a 006a 003a 003a

(12) (15) (144) (112) (79) (38) (44) (42)

Assetst1 008a 004a 006a 004a 002a 001a 003a 002a

(97) (45) (283) (161) (53) (35) (147) (101)

Constant 079a 074a 002a 003a 000 000 000b 002a

(497) (442) (91) (111) (00) (05) (22) (93)

Year FE Yes Yes Yes Yes Yes Yes Yes Yes

Firm FE No Yes No Yes No Yes No Yes

R2 0015 0098 0022 0076 0009 0070 0012 0086

N 133905 133905 140552 140552 138229 138229 140572 140572

1444 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

inasmuch as short-term debt is probably more associated with day-to-dayoperations rather than longer term planning considerationsIn Table XIII we examine change in leverage ( D

DthornE) and change inleverage due to external financing (LEVEF) during good and bad timesseparately The traditional negative effect of a change in profits on achange in leverage is found both in good times and in bad timesHowever when we decompose the impact in order to isolate the externalcomponent as before the result disappears In good times the externalfinancing component of a change in profits is weak but still positiveDuring bad times it is not statistically significant This implies that distin-guishing between good and bad times does not fundamentally alter the basic

Table XII Short- and long-term debt issuances

The sample contains nonfinancial firms listed on the annual Compustat files for the period

from 1971 to 2009 The table presents estimates from regressions of short-term debtissuance to total issuance (in columns (1) and (2)) long-term debt issuance to totalissuance (in columns (3) and (4)) short-term debt issuance to total capitalization (in

columns (5) and (6)) and long-term debt to total capitalization (in columns (7) and (8))Some of the specifications include firm fixed effects The reported t-statistics are correctedfor clustering at the firm level

aMean significant at the 1 level

bMean significant at the 5 level

cMean significant at the 10 level

STDebtIssTotIss

LTDebtIssTotIss

STDebtIssethDthornETHORN

LTDebtIssethDthornETHORN

(1) (2) (3) (4) (5) (6) (7) (8)

Profitabilityt1 0002 0003 0042a 0008 0001 0001 0019a 0012a

(02) (04) (30) (05) (14) (10) (45) (28)

IndLevt1 0009 0007 0070b 0033 0001 0002c 0014a 0005

(06) (05) (26) (11) (08) (17) (26) (09)

M=Bt1 0003a 0003a 0003b 0010a 0000 0000 0003a 0002a

(48) (35) (22) (65) (10) (01) (84) (42)

Tangibilityt1 0020 0007 0011 0067 0005a 0006a 0097a 0081a

(12) (03) (03) (16) (34) (33) (108) (89)

Assetst1 0039a 0011b 0034a 0024b 0001a 0001 0046a 0020a

(89) (22) (38) (23) (36) (13) (190) (82)

Constant 0001b 0153a 0798a 0591a 0000c 0006a 0057a 0096a

(25) (144) (470) (288) (17) (124) (292) (234)

Year FE Yes Yes Yes Yes Yes Yes Yes Yes

Firm FE No Yes No Yes No Yes No Yes

R2 0025 0089 0018 0073 0030 0147 0016 0298

N 136146 136146 136140 136140 143031 143031 142847 142847

PROFITSndashLEVERAGE PUZZLEREVISITED 1445

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

Table XIII Externally financed leverage changes in good and bad times

The sample contains nonfinancial firms listed on the annual Compustat files for the period

from 1971 to 2009 The table presents estimates from regressions of change in debt tocapitalization ratio (in columns (1) (2) (5) and (6)) and change in leverage ratio consider-ing only external financing (in columns (3) (4) (7) and (8)) The change in leverage ratio

considering only external financing is defined as

LevEF frac14Dt1 thorn d

Dt1 thorn Et1 thorn dthorn e

Dt1

Dt1 thorn Et1

where D is debt E is book value of equity d is debt issues net of retirements and e is equityissues net of repurchases The regressions include lagged leverage factors and year indicator

variables The reported t-statistics are corrected for clustering at the firm level An industryis defined as having ldquogood timesrdquo if the median firm in that industry has a market-to-bookratio that is higher than the 67th percentile of the time-series distribution of industry

median market-to-book ratios Conversely an industry is defined as having ldquobad timesrdquoif the median firm in that industry has a market-to-book ratio that is lower than the 33rdpercentile of the time-series distribution of industry median market-to-book ratios

aMean significant at the 1 level

bMean significant at the 5 level

cMean significant at the 10 level

Good times Bad times

DDthornE LevEF D

DthornE LevEF

(1) (2) (3) (4) (5) (6) (7) (8)

Profitabiltyt1 0022c 0005 0016c 0025b 0073b 0041 0007 0007

(17) (03) (18) (24) (25) (11) (04) (03)

IndLevt1 0000 0011 0001 0009 0010 0018 0001 0000

(00) (09) (01) (13) (07) (10) (02) (00)

MB

t1

0004a 0003b 0000 0001 0008a 0001 0000 0000

(40) (24) (05) (08) (28) (04) (03) (03)

Tangt1 0097a 0041 0033a 0038b 0127a 0048 0059a 0033

(45) (16) (27) (25) (49) (14) (43) (16)

Assetst1 0014a 0013b 0024a 0019a 0037a 0023c 0036a 0025a

(30) (20) (84) (48) (45) (17) (88) (43)

Constant 0030b 0036 0000 0003 0028c 0014 0021 0001

(20) (16) (00) (07) (17) (06) (11) (01)

Year FE Yes Yes Yes Yes Yes Yes Yes Yes

Firm FE No Yes No Yes No Yes No Yes

R2 0006 0088 0009 0117 0026 0251 0023 0202

N 56820 56820 58200 58200 23462 23462 23511 23511

1446 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

conclusions The traditional leverage result on profits is driven by increasesin equity that are experienced and then partially but not completely offsetby firmsTo the extent that firms are engaging in active financing they are following

the prescriptions of the trade-off theory Thus we have a coherent picture ofhow firms are responding to profits Basically more profitable firms experiencean increase in the value of their equity They respond to this by issuing debtand repurchasing equity But the magnitude of the increase in equity is largerthan the magnitudes of the active steps taken in the debt and equity marketsThis means that the firmsrsquo actual actions are as predicted by the trade-offtheory But it leaves open the question of why they do not act more stronglyThe capital structure literature has long been interested in dynamics and

capital structure adjustment paths In recent years this interest has increas-ingly resulted in explicit dynamic optimizing models such as those ofGoldstein Ju and Leland (2001) Hennessy and Whited (2005) andStrebulaev (2007) These models commonly make specialized assumptionsleading some to question their generalityWe argue that both fixed transactions costs and variable transactions costs

are plausible and realistic in models that pin down firm size The simultaneouspresence of both is important Due to fixed costs firms only rebalance inter-mittently Due to variable adjustment costs the extent of the adjustment isreduced to the point at which an extra unit to adjustment is balanced againstthe extra unit transactions costs This means that both leverage increases andleverage reductions will only go part way toward a static optimumThe key point is that with both fixed and variable transactions costs we

should expect to only see intermittent leverage adjustments When leverage isadjusted it should ldquounder adjustrdquo when compared to a frictionless modelFurther discussion of transactions costs in a capital structure setting isprovided by Leary and Roberts (2005)

9 Conclusion

The connection between corporate profits and capital structure has beenvery influential in the assessment of the trade-off theory The conventionalinterpretation of the evidence has pushed the literature away from the trade-off theory and toward much more complex dynamic models and ideas As aresult it is important to make sure that the evidence has been correctlyinterpreted We show that the problem is with the interpretation of thetest results and a failure to take into account the costs associated with re-sponding to a profit shock

PROFITSndashLEVERAGE PUZZLEREVISITED 1447

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

We document that more profitable firms really do borrow more (not less)They tend to repurchase equity They experience an increase in both thebook value of equity and the market value of equity Less profitable firmsreally do tend to reduce their debt and to issue equityFirm size and market conditions also matter Larger firms tend to be more

active in the debt markets while smaller firms tend to be relatively moreactive in the equity markets During good times there is more use ofexternal financingThe usual profitsndashleverage puzzle result is primarily driven by the increase

in equity that is experienced by the more profitable firms Accordingly thepuzzle should be restated as asking why do firms not take sufficiently largeoffsetting actions to fully undo the change in equity What limits the mag-nitudes of the typical leverage response to profit shocksIn a frictionless model the partial response appears puzzling But there is

good empirical reason to believe that rebalancing entails both fixed andvariable costs and that firm size matters The rebalancing costs can befully avoided by doing nothing Accordingly the firm must decide whetherany given shock is big enough to be worth responding to If it is then thefirm must decide how big a response is called for These technical conditionsare known as ldquovalue matchingrdquo and ldquosmooth pastingrdquo Thus optimizationimplies that some shocks will be ignored Even if the shock is not ignoredthe optimal response will only partially undo the shock The magnitude ofthe leverage response must balance the marginal cost and the marginalbenefit of an extra unit of leverage Since the marginal cost of adjustingleverage is strictly positive the adjustment toward that static leverageoptimum will only go part way This is true both for leverage increasesand for leverage reductionsOne might conjecture that the variable costs would be time varying and

that such costs might have declined over time Since we do not have a fullaccounting of all of the variable costs it is of course impossible to be sureHowever there is some evidence suggesting that at least the directunderwriting fees are remarkably stable over time10

Surely nobody ought to be surprised that more profitable firms are moreinclined to repurchase equity while unprofitable firms tend not to do so Thefact that more profitable firms typically experience an increase in the value of

10 For IPOs Jay Ritter provides important summary evidence as httpbearwarringtonufleduritterIPOs2013Underwritingpdf He finds that for medium-sized deals a remark-

ably large fraction of the deals have a 7 fee year after year Whether this is due tobehavioral factors or some sort of a coordination device is not clear The extent towhich similar intertemporal stability carries over to other forms of external financemerits further study

1448 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

equity is equally natural Were it not for the contrary claims in the literature(ie the profitsndashleverage puzzle) we would argue that nobody ought to findit surprising that more profitable firms are more inclined to issue debt Thefact that leverage adjustments typically only partially offset profit shocksshould also sound quite natural These facts all fit together easily within thetrade-off theory once we take account of both fixed and variable costs ofactively adjusting capital structure

Appendix

Table AI Market leverage regressions

The table presents estimates of the regressions of market leverage on firm characteristics

The sample comes from the annual Compustat files during the period 1971ndash2009 Financialfirms are excluded The leverage is estimated as the ratio of debt over debt plus marketequity The explanatory variables Profitabilityt1

MB

t1

Tangibilityt1 and LnethAssetsTHORNt1are described in Table 1 IndMedianLevt1 is estimated as the median market leverage of allother firms in the same industry The industry is defined at the level of the four-digit SICcode All specifications include the year fixed effects The specifications in column (2)

additionally include the firm fixed effects We report t-statistics where the standard errorsare clustered at the firm level in parentheses

aSignificant at the 1 level

Quantile regressions

(1) (2) (3) (4) (5)

Profitabilityt1 013a 013a 007a 011a 020a

(275) (262) (349) (425) (538)

IndLevt1 046a 020a 032a 058a 063a

(506) (239) (1299) (1630) (1042)MB

t1

003a 001a 001a 002a 004a

(509) (303) (743) (709) (534)

Tangibilityt1 013a 018a 013a 014a 010a

(191) (181) (762) (571) (232)

LnethAssetsTHORNt1 001a 005a 001a 001a lt001a

(112) (258) (682) (342) (29)

Constant 010a 011a 005a 005a 029a

(171) (98) (173) (118) (398)

Year FE Yes Yes Yes Yes Yes

Firm FE No Yes No No No

Regression type OLS FE 25th tile Median 75thtile

Clustered SE Yes Yes No No No

R2 030 067

Pseudo R2 016 023 0201

N 158578 158578 158578 158578 158578

PROFITSndashLEVERAGE PUZZLEREVISITED 1449

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

Table AII Profitability sorts for firms issuing large amounts

Percentage of firms

Iss Ret Iss Iss Rep Iss Iss IssD IssE

D D NetD E E NetE DampE RepE RetD

(1) (2) (3) (4) (5) (6) (7) (8) (9)

Panel A 10 cutoff to define issuing or retiring activity

Allfirm

s

Low 205 150 134 260 05 257 45 00 23

2 207 175 103 77 06 75 11 01 10

3 243 196 101 50 08 47 08 01 09

4 252 186 113 50 13 48 08 02 08

High 227 161 118 60 27 57 07 04 09

t Highfrac14Low 70 38 62 755 226 768 320 90 152

Smallfirm

s Low 230 132 162 380 05 375 82 00 32

2 228 143 165 334 04 333 65 00 26

3 211 146 147 216 03 214 38 00 23

4 190 154 120 126 04 124 15 01 20

High 194 162 114 99 09 97 12 01 15

t Highfrac14Low 42 40 68 342 20 340 169 07 131

Largefirm

s Low 231 200 82 31 06 29 08 01 04

2 231 187 67 17 06 14 03 01 02

3 234 177 77 17 09 14 03 02 03

4 218 158 75 16 17 14 02 03 02

High 218 135 99 19 41 15 03 07 02

t Highfrac14Low 19 106 36 46 135 57 39 58 23

Percentage of firms

Iss Ret Iss Iss Rep Iss Both Iss D Iss E

D D Net D E E Net E DampE Rep E Ret D

(1) (2) (3) (4) (5) (6) (7) (8) (9)

Panel B 15 cutoff to define issuing or retiring activity

Allfirm

s

Low 158 101 100 230 01 228 27 00 15

2 142 115 63 57 01 56 05 00 06

3 165 128 57 32 01 31 03 00 05

4 173 118 65 35 01 34 03 00 05

High 160 102 70 46 04 44 03 00 05

t Highfrac14Low 08 06 141 738 87 746 277 24 131

Smallfirm

s Low 195 95 132 354 01 350 56 00 24

2 185 96 131 303 01 301 44 00 18

3 160 96 104 179 00 176 18 00 15

4 134 99 78 99 01 98 08 00 12

High 133 95 73 82 01 80 06 00 10

t Highfrac14Low 80 00 95 345 12 344 149 12 52

Largefirm

s Low 145 123 42 14 01 12 03 00 01

2 137 114 32 07 01 06 01 00 01

3 142 104 35 06 01 05 00 00 01

4 130 91 36 08 03 07 01 01 01

High 134 81 48 09 06 07 01 01 01

t Highfrac14Low 19 85 17 26 54 33 34 17 11

1450 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

Table AIV Debt and equity changesmdashwithout fixed effects

The sample contains nonfinancial firms listed on the annual Compustat files for the period from 1971

to 2009 The table presents estimates from regressions of change in debt (D) change in the book

value of equity (BVE) change in the market value of equity (MVE) and net equity issuances

(EqyIss) The explanatory variables include current and lagged change in operating income before

depreciation (Profits and Profitst1) and lagged change in book value of assets (Assetst1) The

regressions include year indicator variables The reported t-statistics are corrected for clustering at the

firm level

aMean significant at the 1 level

bMean significant at the 5 level

cMean significant at the 10 level

D BVE MVE EqyIss

(1) (2) (3) (4) (5) (6) (7) (8)

Profitst1 025a 024a 023a 038a 007 066a 007a 009a

(40) (32) (36) (39) (06) (29) (42) (51)Profits 051b 082a 149a 004a

(25) (33) (49) (33)Assetst1 002 005c 024a 001b

(08) (17) (34) (23)Constant 93 176a 45 71 1552a 1704a 130a 140a

(12) (28) (07) (09) (49) (49) (45) (45)Year FE Yes Yes Yes Yes Yes Yes Yes YesR2 001 005 001 011 001 005 002 002N 162056 162056 162130 162130 157550 157550 162154 162154

Table AIII Median profitability sorts

The sample contains nonfinancial firms listed on the annual Compustat files for the period from 1971

to 2009 The table presents median debt and equity median changes in debt and equity and median

equity issuances (EqyIss) in millions of dollars for all firms sorted on profitability and for profitability

sorts within the smallest and largest firms The sorts are done annually

D D BVE BVE MVE MVE EqyIss Assets(1) (2) (3) (4) (5) (6) (7) (8)

Allfirm

s

Low 21 00 75 17 256 20 01 1922 186 02 449 18 598 19 00 11033 816 06 1274 03 1637 05 00 32244 787 02 1541 18 2488 03 00 3556High 273 00 1361 66 3300 14 00 2578

Smallfirm

s Low 05 00 13 07 140 14 06 402 06 00 25 08 114 12 01 563 09 00 32 05 85 07 00 674 12 00 39 01 68 03 00 79High 10 00 51 04 98 01 00 90

Largefirm

s Low 11485 381 10269 371 13227 202 03 333902 13051 393 12786 140 17571 169 05 392463 11554 287 13429 24 21623 305 00 370314 9963 209 15052 194 29424 308 00 36705High 7361 00 16380 613 49142 439 00 34976

PROFITSndashLEVERAGE PUZZLEREVISITED 1451

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

References

Auerbach A J (1985) Real determinants of corporate leverage in B M Friedman (ed)

Corporate Capital Structures in the United States University of Chicago Press Chicago

IllinoisBertrand M and Mullainathan S (2003) Enjoying the quiet life corporate governance and

managerial preferences Journal of Political Economy 111 1043ndash1075Booth L Varouj A Demirguc-Kunt A and Maksimovic V (2001) Capital structures in

developing countries Journal of Finance 56 87ndash130Brennan M J and Schwartz E S (1984) Optimal financial policy and firm valuation

Journal of Finance 39 593ndash607Cameron A C and Trivedi P K (2010) Microeconometrics using Stata Stata Press TexasConstantinides G M (1979) Multiperiod consumption and investment behavior with

convex transactions costs Management Science 25 1127ndash1137Constantinides G M and Richard S F (1978) Existence of optimal simple policies for

discounted-cost inventory and cash management in continuous time Operations Research

26 620ndash636DeAngelo H DeAngelo L and Whited T M (2011) Capital structure dynamics and

transitory debt Journal of Financial Economics 99 235ndash261Dixit A K (1993) The Art of Smooth Pasting Routledge London and New YorkFama E and French K R (2002) Testing trade-off and pecking order predictions about

dividends and debt Review of Financial Studies 15 1ndash33Faulkender M and Petersen M A (2006) Does the source of capital affect capital struc-

ture Review of Financial Studies 19 45ndash79Faulkender M Flannery M J Hankins K W and Smith J M (2012) Cash Flows and

Leverage Adjustments Journal of Financial Economics 103 632ndash646Fischer E Heinkel R and Zechner J (1989) Dynamic capital structure choice theory

and tests Journal of Finance 44 19ndash40Frank M Z and Goyal V K (2008) Trade-off and pecking order theories of

debt in B E Eckbo (ed) Handbook of Corporate Finance Empirical Corporate Finance

Vol 2 Handbook of Finance Series Chapter 12 ElsevierNorth-Holland Amsterdam pp

135ndash202

Frank M Z and Goyal V K (2009) Capital structure decisions which factors are reliably

important Financial Management 38 1ndash37

Garvey G T and Hanka G (1999) Capital structure and corporate control the effect of

antitakeover statutes on firm leverage Journal of Finance 54 519ndash546

Goldstein R Ju N and Leland H (2001) An EBIT-based model of dynamic capital

structure Journal of Business 74 483ndash512

Graham J R and Tucker A L (2006) Tax shelters and corporate debt policy Journal of

Financial Economics 81 563ndash594

Hackbarth D and Mauer D C (2012) Optimal priority structure capital structure and

investment Review of Financial Studies 25 747ndash796

Halling M Yu J and Zechner J (2012) Leverage dynamics over the business cycle

Working paper University of Utah University of New South Wales and Vienna

University of Economics and Business AdministrationHennessy C A and Whited T A (2005) Debt dynamics Journal of Finance 60 1129ndash1165

Huang R and Ritter J R (2009) Testing theories of capital structure and estimating the

speed of adjustment Journal of Financial and Qualitative Analysis 44 237ndash271

1452 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

Hovakimian A Opler T and Titman S (2001) The debt-equity choice Journal of

Financial and Quantitative Analysis 36 1ndash24Jung K Kim Y and Stulz R M (1996) Timing investment opportunities managerial

discretion and the security issue decisions Journal of Financial Economics 42 159ndash185Kane A Marcus A J and McDonald R L (1984) How big is the tax advantage to

debt Journal of Finance 39 841ndash853Kayhan A and Titman S (2007) Firmsrsquo histories and their capital structures Journal of

Financial Economics 83 1ndash32Leary M T (2009) Bank loan supply lender choice and corporate capital structure

Journal of Finance 64 1143ndash1185Leary M T and Roberts M R (2005) Do firms rebalance their capital structures

Journal of Finance 60 2575ndash2619Leland H E (1994) Corporate debt value bond covenants and optimal capital structure

Journal of finance 49 1213ndash1252Lemmon M L Roberts M R and Zender J F (2008) Back to the beginning persist-

ence and the cross-section of corporate capital structure Journal of Finance 63 1575ndash1608Long M S and Malitz I B (1985) The investment-financing nexus some empirical

evidence Midland Corporate Finance Journal 3 53ndash59Mackie-Mason J K (1990) Do taxes affect corporate financing decisions Journal of

Finance 45 1471ndash1493Morellec E Nikolov B and Schurhoff N (2012) Corporate governance and capital

structure dynamics Journal of Finance 67 803ndash848Myers S C (1984) The capital structure puzzle Journal of Finance 39 575ndash592Myers S C (1993) Still searching for optimal capital structure Journal of Applied

Corporate Finance 61 4ndash14Rajan R and Zingales L (1995) What do we know about capital structure some evidence

from international data Journal of Finance 50 1421ndash1460Strebulaev I A (2007) Do tests of capital structure theory mean what they say Journal of

Finance 62 1747ndash1787

Strebulaev I A and Whited T M (2012) Dynamic Models and Structural Estimation in

Corporate Finance Now Publisher Boston and Delft

Titman S and Wessels R (1988) The determinants of capital structure choice Journal of

Finance 43 1ndash21

Welch I (2004) Capital structure and stock returns Journal of Political Economy 112

106ndash131

Welch I (2011) Two common problems in capital structure research the financial-debt-

to-asset ratio and issuing activity versus leverage changes International Review of Finance

11 1ndash17

PROFITSndashLEVERAGE PUZZLEREVISITED 1453

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

see whether large firms behave disproportionately when compared withsmall firms Column (1) shows that low-profit firms have more debtas reported in many previous studies In Columns (4) and (6) we seethat low-profit firms have a major decline in both the book value ofequity and the market value of equity The high-profit firms have thereverse experienceMore interesting are the active decisions of the firms In Column (2) we see

that in contrast to the usual impression the low-profit firms are reducingdebt while the high-profit firms are increasing it In Column (7) we see thatthe low-profit firms are also reducing leverage by issuing equity The high-profit firms issue little equity In the lower panel similar results are foundwhen we separate small and large firms Equity issuing is particularly strong

Table VI Magnitude of financing activitymdashscaled levels and changes

The sample contains nonfinancial firms listed on the annual Compustat files for the period

from 1971 to 2009 The table presents average debt and equity changes in debt and equityand equity issuances scaled by assets for firms sorted on profitability within size classes Thetable reports information for all firms sorted on profitability and for profitability sorts

within the smallest and largest firms The sorts are done annually The table also reportsresults from t-test for the difference in percentage of firms issuing or retiring securities inthe lowest and the highest profit quintile

D D BVE BVE MVE MVE EqyIss

(1) (2) (3) (4) (5) (6) (7)

Allfirm

s

Low 386 122 12 871 3427 1868 217

2 285 58 407 119 940 157 28

3 293 12 425 14 742 03 12

4 269 26 461 02 955 27 10

High 192 15 551 47 1675 74 07

t Highfrac14Low 35 49 50 47 127 42 157

Smallfirm

s Low 545 154 841 2145 9688 9128 696

2 684 49 712 1915 5014 2318 288

3 331 253 141 969 3063 954 129

4 267 181 357 175 1673 221 55

High 194 05 543 26 1758 147 41

t Highfrac14Low 49 17 51 36 88 27 99

Largefirm

s Low 350 30 306 58 511 55 09

2 346 14 335 13 523 02 04

3 353 165 332 71 711 29 01

4 280 01 414 04 909 09 05

High 227 15 477 19 1643 27 14

t Highfrac14Low 371 112 453 181 570 57 221

PROFITSndashLEVERAGE PUZZLEREVISITED 1435

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

among the small low-profit firms that are also experiencing sharp reductionsin the value of equity In column (2) we see that both for the large and for thesmall firms net debt increases are primarily found in the highest quintile offirm profitability

6 Debt and Equity Regressions

So far we have demonstrated that in our data the conventional leverageregressions have the usual signs We then demonstrated that in simple sortsof the data strikingly different active financing patterns emerge The nexttask is to reconcile these differencesTable VII presents simple regressions predicting changes in debt book

equity and market equity as well as equity issuances Consistent with thesorts we control for changes in firm size and we also include firm and yeardummies There may be a concern about the speed of a firmrsquos responsesAccordingly we consider the lagged change in profits as the main case butwe also include the contemporaneous change in profits Whether the contem-poraneous change in profits is included or not does not in fact matter Wereport panel-robust standard errors adjusted for clustering at the firm levelThe results in Table VII support the findings in the sorts Firms that have

an increase in profits in 1 year increase debt both in the same year and in thenext year Firms that have an increase in profits experience an increase inboth the book value of equity and the market value of equity Importantlyhowever firms that experience and increase in profits reduce their equityissues All these effects are statistically significant (except for the marketvalue of equity in the column 5 specification) and seem intuitively reason-able In Appendix Table AIV similar regressions to those presented aboveare reported but without firm-fixed effects We find that empirically verylittle changes whether we include firm fixed effects or leave them out Thusthe targeting behavior to the extent that it happens does not account forwhat we are reportingThe next step is to bring the sorts and the simple regressions together into

a common setting We therefore include factors in addition to profitability(i) median industry leverage (ii) the market-to-book assets ratio (iii) tangi-bility of assets and (iv) firm size (measured by the log of assets) Rajan andZingales (1995) show that these factors are related to leverage in G7countries A number of studies have used these factors to estimateleverage targets Frank and Goyal (2009) show that these factors arerobustly related to leverage in the USA In these regressions we use indica-tors for the quintile that the firm is in for each factor

1436 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

Table VIII reports these results In Columns (1) and (2) we examine debtissues and again show that profits positively affect debt issuances The effectsare large and statistically significant at the 1 level Columns (3) and (4)which examine book equity show a strong effect of profits This effect is onlyslightly reduced by the inclusion of other factors Columns (5) and (6) illus-trate that the impact of profits on the change in market equity is also robustto the inclusion of the conventional factors Columns (7) and (8) examineequity issuances Here we again find that profits have a significant negativeimpact on equity issuances As before the results are robust to the inclusionof the conventional factorsIn unreported tables we estimate debt and equity changes using quantile

regressions In terms of the signs and significance of the profitability variable

Table VII Debt and equity changes

The sample contains nonfinancial firms listed on the annual Compustat files for the period

from 1971 to 2009 The table presents fixed effect estimates from regressions of changes indebt (D) changes in the book value of equity (BVE) changes in the market value ofequity (MVE) and net equity issuances (EqyIss) The explanatory variables include

current and lagged change in operating income before depreciation (Profits andProfitst1) and lagged change in the book value of assets (Assetst1) In addition tofixed firm effects the regressions include year indicator variables The reported t-statistics

are corrected for clustering at the firm level

aMean significant at the 1 level

bMean significant at the 5 level

cMean significant at the 10 level

D BVE MVE EqyIss

(1) (2) (3) (4) (5) (6) (7) (8)

Profitst1 022a 025a 015b 035a 004 060b 005b 007a

(33) (31) (25) (35) (03) (25) (25) (53)

Profits 050b 078a 140a 003a

(24) (32) (48) (34)

Assetst1 000 007b 025a 001b

(00) (21) (35) (25)

Constant 167 283a 34 51 1145a 1182a 605a 595a

(14) (30) (04) (05) (32) (29) (69) (74)

Year FE Yes Yes Yes Yes Yes Yes Yes Yes

Firm FE Yes Yes Yes Yes Yes Yes Yes Yes

R2 001 004 001 010 001 005 002 003

N 162056 162056 162130 162130 157550 157550 162154 162154

PROFITSndashLEVERAGE PUZZLEREVISITED 1437

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

the results are qualitatively similar to the OLS results reported in Table VIIIProfitability positively affects debt issuances and negatively affects equityissuances across various quantiles Overall the results in Table VIII showthat there is nothing that is special about using sorts or regressions to explaindebt and equity issues In either case we find that more profitable firms tendto increase their debt experience an increase in the value of equity andrepurchase shares Thus the control factors are not responsible for theusual rejection of the trade-off theory It would appear that what mattersis the size of the issuing activity relative to the organic increases in equityvalue due to profits In other words what matters in the leverage regressionseems to be coming from the presence of E in D

DthornE and not from D

Table VIII Debt and equity issuances profitability and leverage factor quintiles

The sample contains nonfinancial firms listed on the annual Compustat files for the period

from 1971 to 2009 The table presents estimates from regressions of change in debt (D)change in the book value of equity (BVE) change in the market value of equity (MVE)and net equity issuances (EquityIss) The explanatory variables include lagged profitability

quintiles and lagged leverage factor quintiles In addition to fixed firm effects the regres-sions include year indicator variables The reported t-statistics are corrected for clustering atthe firm level

aMean significant at the 1 level

bMean significant at the 5 level

cMean significant at the 10 level

D BVE MVE EqyIss

(1) (2) (3) (4) (5) (6) (7) (8)

ProfitabilityQint 214a 167a 283a 175a 80 429a 53a 55a

(70) (54) (66) (59) (08) (48) (55) (62)

IndLevQint 104a 43 99 20

(29) (15) (10) (10)

Ln(Assets)Qint52 17 1026a 53a

(16) (05) (77) (35)MB

Qint158a 372a 1155a 09

(45) (56) (106) (06)

TangibilityQint 05 177 436c 64a

(01) (15) (18) (31)

Constant 167 96 656b 235 4613a 8610a 113 570a

(08) (04) (26) (04) (61) (58) (11) (31)

Year FE Yes Yes Yes Yes Yes Yes Yes Yes

Firm FE Yes Yes Yes Yes Yes Yes Yes Yes

R2 0002 0002 0003 0004 0013 0014 0013 0014

N 155587 155037 155587 155037 155587 155037 155587 155037

1438 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

7 Are Financial Market Conditions Important

According to the trade-off theory capital structure is determined by con-sidering the impact of costs and benefits of debt The time variation inthe costs and benefits of debt imply that issuance decisions would alsovary over time Furthermore market conditions affect the transactioncosts of adjusting Adverse selection is a greater problem in a cold marketthan in a hot one Accordingly it may be easier to adjust in good marketconditions Related ideas have been developed by Huang and Ritter(2009) Faulkender et al (2012) and Halling Yu and Zechner (2012)Faulkender Flannery Hankins and Smithrsquos paper examines cross-sectionalvariation in adjustment speeds including market valuations andhow they may affect adjustment speeds It is natural to think that ingood market conditions it will be less costly to issue both debt and equityand adjusting leverage to respond to profitability shocks would be mucheasierTo test the importance of market conditions we require a definition of

good times and bad times Our empirical strategy is to estimate good timesversus bad times at the four-digit industry level We define an industry ashaving ldquogood timesrdquo if the median firm in that industry has a market-to-book ratio that is higher than the 67th percentile of the time-series distribu-tion of the industry median market-to-book ratios Conversely an industryis defined as having ldquobad timesrdquo if the median firm in the industry has amarket-to-book ratio that is lower than the 33rd percentile of the time-seriesdistribution of the industry median market-to-book ratiosPanel A of Table IX tabulates issuance activity for profitability sorts in

both good times and bad times As might be expected active debt and equityissuances are larger during good times There is also somewhat more activeswaps between debt and equity during good timesPanel B reports the issuance activity for the smallest and the largest firms

during the good times and bad times Within each size quintile firms aresorted on profitability This is a three-way sort As before we find that thereis much more active use of external markets during good times Small low-profit firms are more likely to issue equity in good times than in bad timesDebt issuances are significantly higher in good times Large high-profitfirms are significantly more likely to issue debt and repurchase equity ingood times than in bad timesPanel C reports the magnitudes of the financial variables rather than the

frequencies During bad times less profitable large firms retire substantialamounts of debt and they show a tendency to issue equity Small firms donot seem to engage in similar debt reduction activities However like the

PROFITSndashLEVERAGE PUZZLEREVISITED 1439

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

Table IX Debt and equity issues in good and bad times

The table reports the frequency and magnitude of financing activity for sorts on profits for subsamples

of firms in good and bad times An industry is defined as having ldquogood timesrdquo if the median firm in

that industry has a market-to-book ratio that is higher than the 67th percentile of the time-series

distribution of industry median market-to-book ratios Conversely an industry is defined as having

bad times if the median firm in that industry has a market-to-book ratio that is lower than the 33rd

percentile of the time-series distribution of industry median market-to-book ratios Panels A and B

report the percentage of firms (i) issuing net debt in excess of 5 of the value of their assets (ii) issuing

net equity in excess of 5 of the value of their assets (iii) issuing debt and repurchasing equity both in

excess of 5 of the value of their assets and (iv) issuing equity and retiring debt both in excess of 5

of the value of their assets Panel C tabulates the mean debt and equity levels and the mean changes in

debt and equity issuances The sample contains nonfinancial firms listed on the annual Compustat files

for the period from 1971 to 2009

Iss Iss Iss D Iss ENet D Net E Rep E Ret D(1) (2) (3) (4)

Panel A sorts on profitability

Badtimes

Low 145 131 01 202 170 49 02 083 173 46 04 074 182 47 03 10

High 177 51 06 10t Highfrac14Low 41 136 33 01

Goodtimes

Low 206 381 01 462 197 148 02 233 212 103 04 184 234 92 10 18

High 216 92 16 14t Highfrac14Low 23 658 138 23

Panel B sorts on firm size and profitability

Badtimes

Smallfirm

s Low 151 318 00 312 182 276 02 253 151 140 02 294 183 73 00 19

High 177 80 00 11t Highfrac14Low 10 90 NA 20

Largefirm

s Low 172 54 01 042 148 39 04 053 150 43 04 044 156 33 03 06

High 147 28 05 03t Highfrac14Low 17 32 20 02

Goodtimes Smallfirm

s Low 213 448 01 542 230 443 01 473 222 325 00 474 203 197 02 38

High 190 153 02 31t Highfrac14Low 19 230 10 39

Largefirm

s Low 203 96 03 122 208 66 09 133 232 54 16 064 215 39 19 05

High 234 36 38 04t Highfrac14Low 29 99 92 38

(continued)

1440 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

larger firms there is a tendency for small firms to issue more equitymdashat leastto the extent possibleDuring good times large profitable firms raise significant amounts of debt

as they experience an increase in the value of their equity Such firms alsoengage in share repurchases Debt issues by smaller firms are much lessaffected by their own profitability during good times and their change indebt is more or less independent of their profits However equity issuanceis primarily found among the low-profit small firms during good timesTable IX shows that empirically profitability has a time-varying impacton leverage choicesIn Table X the basic regressions for changes in debt and for equity

issuance are presented for both good times and bad times Comparinggood times with bad times we see that the effects of profits are muchstronger in good times In bad times the impacts are rather weak statisticallyThis difference is interesting and deserves further study However for ourpurposes recall that the static trade-off theory implies time-varying capital

Table IX Continued

D D BVE BVE MVE MVE EqyIss Assets

(1) (2) (3) (4) (5) (6) (7) (8)

Panel C magnitude of financing activity

Badtimes Smallfirm

s Low 40 03 58 14 305 06 22 140

2 44 05 82 07 240 67 17 180

3 45 04 94 02 176 18 05 187

4 43 09 98 02 137 04 03 188

High 36 10 114 13 209 16 04 199

Largefirm

s Low 3365 73 2523 169 2304 163 94 8972

2 3031 95 2983 35 3284 278 56 9068

3 2930 5 3360 14 3947 308 57 9225

4 2281 6 3817 19 5213 442 7 8902

High 1790 62 4767 13 8799 58 31 9573

Goodtimes

Smallfirm

s Low 24 06 38 05 339 23 39 90

2 26 08 48 01 301 04 25 104

3 30 09 60 04 246 03 17 125

4 28 07 66 08 206 18 12 130

High 23 06 85 16 261 13 08 141

Largefirm

s Low 4532 36 3477 103 6557 260 53 12465

2 4580 58 3527 139 6911 421 23 12522

3 2952 111 3485 2 7586 279 32 9535

4 2752 124 3939 72 10686 33 105 10226

High 2429 228 4819 377 18565 268 222 10830

PROFITSndashLEVERAGE PUZZLEREVISITED 1441

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

structure choices even when the target is time invariant Empirically we doobserve time-varying choices

8 Is Scaling Important

In the preceding section we provided results for unscaled debt and equityissues This seems appropriate to us However it involves making twochanges to the standard regression We would like to determine if both

Table X Debt and equity changes in good and bad timesmdashregressions

The sample contains nonfinancial firms listed on the annual Compustat files for the period

from 1971 to 2009 The table presents estimates from regressions of change in debt (D)and net equity issuances (EqyIss) in good and bad times The explanatory variables includechange in profits lagged change in profits and lagged change in assets The regressions

include both firm and year fixed effects An industry is defined as having ldquogood timesrdquo ifthe median firm in that industry has a market-to-book ratio that is higher than the 67thpercentile of the time-series distribution of industry median market-to-book ratios

Conversely an industry is defined as having bad times if the median firm in thatindustry has a market-to-book ratio that is lower than the 33rd percentile of the time-series distribution of industry median market-to-book ratios The reported t-statistics arecorrected for clustering at the firm level

aMean significant at the 1 level

bMean significant at the 5 level

cMean significant at the 10 level

Good times Bad times

D EqyIss D EqyIss

(1) (2) (3) (4) (5) (6) (7) (8)

Profits 087a 091a 006b 009a 060 065 003 003b

(30) (29) (20) (28) (15) (16) (16) (20)

Profitst1 029c 015a 034c 001

(19) (37) (17) (04)

Assetst1 000 002b 000 001

(02) (22) (00) (07)

Constant 1337 3989 8691a 10001a 1351 1851 4099a 4055a

(03) (09) (55) (37) (09) (13) (80) (78)

Year FE Yes Yes Yes Yes Yes Yes Yes Yes

Firm FE Yes Yes Yes Yes Yes Yes Yes Yes

R2 0071 0080 0018 0045 0178 0212 0034 0041

N 74827 68729 74891 68787 26965 26013 26971 26018

1442 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

changes are crucial Accordingly in this section we take a step back toconsider prior literature in which variables are scaled The questions arehow does scaling affect the results and whyTable XI examines alternative scalings Columns (1) and (2) scale debt and

equity issuances by the total firm net issuance Columns (3) and (4) scaledebt and equity issuances by the sum of debt and book equity Columns (5)and (6) consider the change in debt scaled by the sum of debt and bookequity Columns (7) and (8) consider the change in leverage due to externalfinancing Since we are examining changes we use the changes in thestandard regression factors as controls The primary interest is in the coef-ficient on the change in profitsWe find that a change in profits is associated with an increase in debt

issues But the sign reverses when we scale the debt issues by the sum of debtplus book equity This is telling What is really driving the results is thechange in equity in the denominator not the change in debt in the numer-ator This result is further substantiated in Columns (5) and (6) where thechange in debt to capital ratio is used instead of debt issuanceTo further explore the impact of active issuance relative to passive

experiencing of changes we decompose leverage by separating out theeffect of debt and equity issues from those due to changes in retainedearnings Thus we define leverage change due to external financing as

LevEF frac14Dt1 thorn d

Dt1 thorn Et1 thorn dthorn e

Dt1

Dt1 thorn Et1 eth1THORN

where D is debt E is book value of equity d is debt issues net of retirementsand e is equity issues net of repurchases Garvey and Hanka (1999) use asimilar measure in their study This measure directly captures the netleverage effect of debt and equity issuancesColumns (7) and (8) report results from regressions of the net leverage

effect of debt and equity issues on changes in profitability after controllingfor other leverage factors in differences and year indicator variables Theeffect of profitability on leverage change due to external financing is positiveand statistically significant This comparison suggests that changes in totaldebt and equity are only partly a result of debt and equity issuance decisionsOther balance sheet adjustments complicate the inferences from leverageratio regressionsFor most of our analysis we follow the literature in treating debt as homo-

genous It is of course potentially interesting to ask what happens when wedistinguish between short-term debt and long-term debt Which part of thefirmrsquos debt maturity structure drives the under-adjustment Table XII

PROFITSndashLEVERAGE PUZZLEREVISITED 1443

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

provides results from such decompositions Alternative scalings areconsidered The main results of interest are that the distinction betweenshort- and long-term debt is important Short-term debt issuance is largelyindependent of changes in profits The negative sign on profit change stemsfrom the action of long-term debt This seems intuitively reasonable

Table XI Scaled issuances

The sample contains nonfinancial firms listed on the annual Compustat files for the period

from 1971 to 2009 The table presents estimates from regressions of debt issuance to totalissuance (in columns (1) and (2)) debt issuance to total capitalization (in columns (3) and(4)) change in debt to capitalization ratio (in columns (5) and (6)) and change in leverage

ratio considering only external financing (in columns (7) and (8)) The change in leverageratio considering only external financing is defined as

LevEF frac14Dt1 thorn d

Dt1 thorn Et1 thorn dthorn e

Dt1

Dt1 thorn Et1

where D is debt E is book value of equity d is debt issues net of retirements and e is equity

issues net of repurchases The regressions include lagged leverage factors and year indicatorvariables Some of the specifications include firm fixed effects The reported t-statistics arecorrected for clustering at the firm level

aMean significant at the 1 level

bMean significant at the 5 level

cMean significant at the 10 level

DebtIssTotIss

DebtIssethDthornETHORN D

DthornE LevEF

(1) (2) (3) (4) (5) (6) (7) (8)

Profitt1 004a 000 002a 001c 003a 002 001b 002a

(35) (03) (34) (18) (35) (14) (23) (26)

IndLevt1 004c 001 001 001c 000 000 000 000

(19) (06) (15) (19) (02) (04) (03) (05)

Market=bookt1 001a 001a 000a 000a 000a 000a 000 000c

(47) (85) (68) (60) (55) (36) (05) (18)

Tangt1 004 005 013a 011a 011a 006a 003a 003a

(12) (15) (144) (112) (79) (38) (44) (42)

Assetst1 008a 004a 006a 004a 002a 001a 003a 002a

(97) (45) (283) (161) (53) (35) (147) (101)

Constant 079a 074a 002a 003a 000 000 000b 002a

(497) (442) (91) (111) (00) (05) (22) (93)

Year FE Yes Yes Yes Yes Yes Yes Yes Yes

Firm FE No Yes No Yes No Yes No Yes

R2 0015 0098 0022 0076 0009 0070 0012 0086

N 133905 133905 140552 140552 138229 138229 140572 140572

1444 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

inasmuch as short-term debt is probably more associated with day-to-dayoperations rather than longer term planning considerationsIn Table XIII we examine change in leverage ( D

DthornE) and change inleverage due to external financing (LEVEF) during good and bad timesseparately The traditional negative effect of a change in profits on achange in leverage is found both in good times and in bad timesHowever when we decompose the impact in order to isolate the externalcomponent as before the result disappears In good times the externalfinancing component of a change in profits is weak but still positiveDuring bad times it is not statistically significant This implies that distin-guishing between good and bad times does not fundamentally alter the basic

Table XII Short- and long-term debt issuances

The sample contains nonfinancial firms listed on the annual Compustat files for the period

from 1971 to 2009 The table presents estimates from regressions of short-term debtissuance to total issuance (in columns (1) and (2)) long-term debt issuance to totalissuance (in columns (3) and (4)) short-term debt issuance to total capitalization (in

columns (5) and (6)) and long-term debt to total capitalization (in columns (7) and (8))Some of the specifications include firm fixed effects The reported t-statistics are correctedfor clustering at the firm level

aMean significant at the 1 level

bMean significant at the 5 level

cMean significant at the 10 level

STDebtIssTotIss

LTDebtIssTotIss

STDebtIssethDthornETHORN

LTDebtIssethDthornETHORN

(1) (2) (3) (4) (5) (6) (7) (8)

Profitabilityt1 0002 0003 0042a 0008 0001 0001 0019a 0012a

(02) (04) (30) (05) (14) (10) (45) (28)

IndLevt1 0009 0007 0070b 0033 0001 0002c 0014a 0005

(06) (05) (26) (11) (08) (17) (26) (09)

M=Bt1 0003a 0003a 0003b 0010a 0000 0000 0003a 0002a

(48) (35) (22) (65) (10) (01) (84) (42)

Tangibilityt1 0020 0007 0011 0067 0005a 0006a 0097a 0081a

(12) (03) (03) (16) (34) (33) (108) (89)

Assetst1 0039a 0011b 0034a 0024b 0001a 0001 0046a 0020a

(89) (22) (38) (23) (36) (13) (190) (82)

Constant 0001b 0153a 0798a 0591a 0000c 0006a 0057a 0096a

(25) (144) (470) (288) (17) (124) (292) (234)

Year FE Yes Yes Yes Yes Yes Yes Yes Yes

Firm FE No Yes No Yes No Yes No Yes

R2 0025 0089 0018 0073 0030 0147 0016 0298

N 136146 136146 136140 136140 143031 143031 142847 142847

PROFITSndashLEVERAGE PUZZLEREVISITED 1445

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

Table XIII Externally financed leverage changes in good and bad times

The sample contains nonfinancial firms listed on the annual Compustat files for the period

from 1971 to 2009 The table presents estimates from regressions of change in debt tocapitalization ratio (in columns (1) (2) (5) and (6)) and change in leverage ratio consider-ing only external financing (in columns (3) (4) (7) and (8)) The change in leverage ratio

considering only external financing is defined as

LevEF frac14Dt1 thorn d

Dt1 thorn Et1 thorn dthorn e

Dt1

Dt1 thorn Et1

where D is debt E is book value of equity d is debt issues net of retirements and e is equityissues net of repurchases The regressions include lagged leverage factors and year indicator

variables The reported t-statistics are corrected for clustering at the firm level An industryis defined as having ldquogood timesrdquo if the median firm in that industry has a market-to-bookratio that is higher than the 67th percentile of the time-series distribution of industry

median market-to-book ratios Conversely an industry is defined as having ldquobad timesrdquoif the median firm in that industry has a market-to-book ratio that is lower than the 33rdpercentile of the time-series distribution of industry median market-to-book ratios

aMean significant at the 1 level

bMean significant at the 5 level

cMean significant at the 10 level

Good times Bad times

DDthornE LevEF D

DthornE LevEF

(1) (2) (3) (4) (5) (6) (7) (8)

Profitabiltyt1 0022c 0005 0016c 0025b 0073b 0041 0007 0007

(17) (03) (18) (24) (25) (11) (04) (03)

IndLevt1 0000 0011 0001 0009 0010 0018 0001 0000

(00) (09) (01) (13) (07) (10) (02) (00)

MB

t1

0004a 0003b 0000 0001 0008a 0001 0000 0000

(40) (24) (05) (08) (28) (04) (03) (03)

Tangt1 0097a 0041 0033a 0038b 0127a 0048 0059a 0033

(45) (16) (27) (25) (49) (14) (43) (16)

Assetst1 0014a 0013b 0024a 0019a 0037a 0023c 0036a 0025a

(30) (20) (84) (48) (45) (17) (88) (43)

Constant 0030b 0036 0000 0003 0028c 0014 0021 0001

(20) (16) (00) (07) (17) (06) (11) (01)

Year FE Yes Yes Yes Yes Yes Yes Yes Yes

Firm FE No Yes No Yes No Yes No Yes

R2 0006 0088 0009 0117 0026 0251 0023 0202

N 56820 56820 58200 58200 23462 23462 23511 23511

1446 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

conclusions The traditional leverage result on profits is driven by increasesin equity that are experienced and then partially but not completely offsetby firmsTo the extent that firms are engaging in active financing they are following

the prescriptions of the trade-off theory Thus we have a coherent picture ofhow firms are responding to profits Basically more profitable firms experiencean increase in the value of their equity They respond to this by issuing debtand repurchasing equity But the magnitude of the increase in equity is largerthan the magnitudes of the active steps taken in the debt and equity marketsThis means that the firmsrsquo actual actions are as predicted by the trade-offtheory But it leaves open the question of why they do not act more stronglyThe capital structure literature has long been interested in dynamics and

capital structure adjustment paths In recent years this interest has increas-ingly resulted in explicit dynamic optimizing models such as those ofGoldstein Ju and Leland (2001) Hennessy and Whited (2005) andStrebulaev (2007) These models commonly make specialized assumptionsleading some to question their generalityWe argue that both fixed transactions costs and variable transactions costs

are plausible and realistic in models that pin down firm size The simultaneouspresence of both is important Due to fixed costs firms only rebalance inter-mittently Due to variable adjustment costs the extent of the adjustment isreduced to the point at which an extra unit to adjustment is balanced againstthe extra unit transactions costs This means that both leverage increases andleverage reductions will only go part way toward a static optimumThe key point is that with both fixed and variable transactions costs we

should expect to only see intermittent leverage adjustments When leverage isadjusted it should ldquounder adjustrdquo when compared to a frictionless modelFurther discussion of transactions costs in a capital structure setting isprovided by Leary and Roberts (2005)

9 Conclusion

The connection between corporate profits and capital structure has beenvery influential in the assessment of the trade-off theory The conventionalinterpretation of the evidence has pushed the literature away from the trade-off theory and toward much more complex dynamic models and ideas As aresult it is important to make sure that the evidence has been correctlyinterpreted We show that the problem is with the interpretation of thetest results and a failure to take into account the costs associated with re-sponding to a profit shock

PROFITSndashLEVERAGE PUZZLEREVISITED 1447

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

We document that more profitable firms really do borrow more (not less)They tend to repurchase equity They experience an increase in both thebook value of equity and the market value of equity Less profitable firmsreally do tend to reduce their debt and to issue equityFirm size and market conditions also matter Larger firms tend to be more

active in the debt markets while smaller firms tend to be relatively moreactive in the equity markets During good times there is more use ofexternal financingThe usual profitsndashleverage puzzle result is primarily driven by the increase

in equity that is experienced by the more profitable firms Accordingly thepuzzle should be restated as asking why do firms not take sufficiently largeoffsetting actions to fully undo the change in equity What limits the mag-nitudes of the typical leverage response to profit shocksIn a frictionless model the partial response appears puzzling But there is

good empirical reason to believe that rebalancing entails both fixed andvariable costs and that firm size matters The rebalancing costs can befully avoided by doing nothing Accordingly the firm must decide whetherany given shock is big enough to be worth responding to If it is then thefirm must decide how big a response is called for These technical conditionsare known as ldquovalue matchingrdquo and ldquosmooth pastingrdquo Thus optimizationimplies that some shocks will be ignored Even if the shock is not ignoredthe optimal response will only partially undo the shock The magnitude ofthe leverage response must balance the marginal cost and the marginalbenefit of an extra unit of leverage Since the marginal cost of adjustingleverage is strictly positive the adjustment toward that static leverageoptimum will only go part way This is true both for leverage increasesand for leverage reductionsOne might conjecture that the variable costs would be time varying and

that such costs might have declined over time Since we do not have a fullaccounting of all of the variable costs it is of course impossible to be sureHowever there is some evidence suggesting that at least the directunderwriting fees are remarkably stable over time10

Surely nobody ought to be surprised that more profitable firms are moreinclined to repurchase equity while unprofitable firms tend not to do so Thefact that more profitable firms typically experience an increase in the value of

10 For IPOs Jay Ritter provides important summary evidence as httpbearwarringtonufleduritterIPOs2013Underwritingpdf He finds that for medium-sized deals a remark-

ably large fraction of the deals have a 7 fee year after year Whether this is due tobehavioral factors or some sort of a coordination device is not clear The extent towhich similar intertemporal stability carries over to other forms of external financemerits further study

1448 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

equity is equally natural Were it not for the contrary claims in the literature(ie the profitsndashleverage puzzle) we would argue that nobody ought to findit surprising that more profitable firms are more inclined to issue debt Thefact that leverage adjustments typically only partially offset profit shocksshould also sound quite natural These facts all fit together easily within thetrade-off theory once we take account of both fixed and variable costs ofactively adjusting capital structure

Appendix

Table AI Market leverage regressions

The table presents estimates of the regressions of market leverage on firm characteristics

The sample comes from the annual Compustat files during the period 1971ndash2009 Financialfirms are excluded The leverage is estimated as the ratio of debt over debt plus marketequity The explanatory variables Profitabilityt1

MB

t1

Tangibilityt1 and LnethAssetsTHORNt1are described in Table 1 IndMedianLevt1 is estimated as the median market leverage of allother firms in the same industry The industry is defined at the level of the four-digit SICcode All specifications include the year fixed effects The specifications in column (2)

additionally include the firm fixed effects We report t-statistics where the standard errorsare clustered at the firm level in parentheses

aSignificant at the 1 level

Quantile regressions

(1) (2) (3) (4) (5)

Profitabilityt1 013a 013a 007a 011a 020a

(275) (262) (349) (425) (538)

IndLevt1 046a 020a 032a 058a 063a

(506) (239) (1299) (1630) (1042)MB

t1

003a 001a 001a 002a 004a

(509) (303) (743) (709) (534)

Tangibilityt1 013a 018a 013a 014a 010a

(191) (181) (762) (571) (232)

LnethAssetsTHORNt1 001a 005a 001a 001a lt001a

(112) (258) (682) (342) (29)

Constant 010a 011a 005a 005a 029a

(171) (98) (173) (118) (398)

Year FE Yes Yes Yes Yes Yes

Firm FE No Yes No No No

Regression type OLS FE 25th tile Median 75thtile

Clustered SE Yes Yes No No No

R2 030 067

Pseudo R2 016 023 0201

N 158578 158578 158578 158578 158578

PROFITSndashLEVERAGE PUZZLEREVISITED 1449

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

Table AII Profitability sorts for firms issuing large amounts

Percentage of firms

Iss Ret Iss Iss Rep Iss Iss IssD IssE

D D NetD E E NetE DampE RepE RetD

(1) (2) (3) (4) (5) (6) (7) (8) (9)

Panel A 10 cutoff to define issuing or retiring activity

Allfirm

s

Low 205 150 134 260 05 257 45 00 23

2 207 175 103 77 06 75 11 01 10

3 243 196 101 50 08 47 08 01 09

4 252 186 113 50 13 48 08 02 08

High 227 161 118 60 27 57 07 04 09

t Highfrac14Low 70 38 62 755 226 768 320 90 152

Smallfirm

s Low 230 132 162 380 05 375 82 00 32

2 228 143 165 334 04 333 65 00 26

3 211 146 147 216 03 214 38 00 23

4 190 154 120 126 04 124 15 01 20

High 194 162 114 99 09 97 12 01 15

t Highfrac14Low 42 40 68 342 20 340 169 07 131

Largefirm

s Low 231 200 82 31 06 29 08 01 04

2 231 187 67 17 06 14 03 01 02

3 234 177 77 17 09 14 03 02 03

4 218 158 75 16 17 14 02 03 02

High 218 135 99 19 41 15 03 07 02

t Highfrac14Low 19 106 36 46 135 57 39 58 23

Percentage of firms

Iss Ret Iss Iss Rep Iss Both Iss D Iss E

D D Net D E E Net E DampE Rep E Ret D

(1) (2) (3) (4) (5) (6) (7) (8) (9)

Panel B 15 cutoff to define issuing or retiring activity

Allfirm

s

Low 158 101 100 230 01 228 27 00 15

2 142 115 63 57 01 56 05 00 06

3 165 128 57 32 01 31 03 00 05

4 173 118 65 35 01 34 03 00 05

High 160 102 70 46 04 44 03 00 05

t Highfrac14Low 08 06 141 738 87 746 277 24 131

Smallfirm

s Low 195 95 132 354 01 350 56 00 24

2 185 96 131 303 01 301 44 00 18

3 160 96 104 179 00 176 18 00 15

4 134 99 78 99 01 98 08 00 12

High 133 95 73 82 01 80 06 00 10

t Highfrac14Low 80 00 95 345 12 344 149 12 52

Largefirm

s Low 145 123 42 14 01 12 03 00 01

2 137 114 32 07 01 06 01 00 01

3 142 104 35 06 01 05 00 00 01

4 130 91 36 08 03 07 01 01 01

High 134 81 48 09 06 07 01 01 01

t Highfrac14Low 19 85 17 26 54 33 34 17 11

1450 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

Table AIV Debt and equity changesmdashwithout fixed effects

The sample contains nonfinancial firms listed on the annual Compustat files for the period from 1971

to 2009 The table presents estimates from regressions of change in debt (D) change in the book

value of equity (BVE) change in the market value of equity (MVE) and net equity issuances

(EqyIss) The explanatory variables include current and lagged change in operating income before

depreciation (Profits and Profitst1) and lagged change in book value of assets (Assetst1) The

regressions include year indicator variables The reported t-statistics are corrected for clustering at the

firm level

aMean significant at the 1 level

bMean significant at the 5 level

cMean significant at the 10 level

D BVE MVE EqyIss

(1) (2) (3) (4) (5) (6) (7) (8)

Profitst1 025a 024a 023a 038a 007 066a 007a 009a

(40) (32) (36) (39) (06) (29) (42) (51)Profits 051b 082a 149a 004a

(25) (33) (49) (33)Assetst1 002 005c 024a 001b

(08) (17) (34) (23)Constant 93 176a 45 71 1552a 1704a 130a 140a

(12) (28) (07) (09) (49) (49) (45) (45)Year FE Yes Yes Yes Yes Yes Yes Yes YesR2 001 005 001 011 001 005 002 002N 162056 162056 162130 162130 157550 157550 162154 162154

Table AIII Median profitability sorts

The sample contains nonfinancial firms listed on the annual Compustat files for the period from 1971

to 2009 The table presents median debt and equity median changes in debt and equity and median

equity issuances (EqyIss) in millions of dollars for all firms sorted on profitability and for profitability

sorts within the smallest and largest firms The sorts are done annually

D D BVE BVE MVE MVE EqyIss Assets(1) (2) (3) (4) (5) (6) (7) (8)

Allfirm

s

Low 21 00 75 17 256 20 01 1922 186 02 449 18 598 19 00 11033 816 06 1274 03 1637 05 00 32244 787 02 1541 18 2488 03 00 3556High 273 00 1361 66 3300 14 00 2578

Smallfirm

s Low 05 00 13 07 140 14 06 402 06 00 25 08 114 12 01 563 09 00 32 05 85 07 00 674 12 00 39 01 68 03 00 79High 10 00 51 04 98 01 00 90

Largefirm

s Low 11485 381 10269 371 13227 202 03 333902 13051 393 12786 140 17571 169 05 392463 11554 287 13429 24 21623 305 00 370314 9963 209 15052 194 29424 308 00 36705High 7361 00 16380 613 49142 439 00 34976

PROFITSndashLEVERAGE PUZZLEREVISITED 1451

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

References

Auerbach A J (1985) Real determinants of corporate leverage in B M Friedman (ed)

Corporate Capital Structures in the United States University of Chicago Press Chicago

IllinoisBertrand M and Mullainathan S (2003) Enjoying the quiet life corporate governance and

managerial preferences Journal of Political Economy 111 1043ndash1075Booth L Varouj A Demirguc-Kunt A and Maksimovic V (2001) Capital structures in

developing countries Journal of Finance 56 87ndash130Brennan M J and Schwartz E S (1984) Optimal financial policy and firm valuation

Journal of Finance 39 593ndash607Cameron A C and Trivedi P K (2010) Microeconometrics using Stata Stata Press TexasConstantinides G M (1979) Multiperiod consumption and investment behavior with

convex transactions costs Management Science 25 1127ndash1137Constantinides G M and Richard S F (1978) Existence of optimal simple policies for

discounted-cost inventory and cash management in continuous time Operations Research

26 620ndash636DeAngelo H DeAngelo L and Whited T M (2011) Capital structure dynamics and

transitory debt Journal of Financial Economics 99 235ndash261Dixit A K (1993) The Art of Smooth Pasting Routledge London and New YorkFama E and French K R (2002) Testing trade-off and pecking order predictions about

dividends and debt Review of Financial Studies 15 1ndash33Faulkender M and Petersen M A (2006) Does the source of capital affect capital struc-

ture Review of Financial Studies 19 45ndash79Faulkender M Flannery M J Hankins K W and Smith J M (2012) Cash Flows and

Leverage Adjustments Journal of Financial Economics 103 632ndash646Fischer E Heinkel R and Zechner J (1989) Dynamic capital structure choice theory

and tests Journal of Finance 44 19ndash40Frank M Z and Goyal V K (2008) Trade-off and pecking order theories of

debt in B E Eckbo (ed) Handbook of Corporate Finance Empirical Corporate Finance

Vol 2 Handbook of Finance Series Chapter 12 ElsevierNorth-Holland Amsterdam pp

135ndash202

Frank M Z and Goyal V K (2009) Capital structure decisions which factors are reliably

important Financial Management 38 1ndash37

Garvey G T and Hanka G (1999) Capital structure and corporate control the effect of

antitakeover statutes on firm leverage Journal of Finance 54 519ndash546

Goldstein R Ju N and Leland H (2001) An EBIT-based model of dynamic capital

structure Journal of Business 74 483ndash512

Graham J R and Tucker A L (2006) Tax shelters and corporate debt policy Journal of

Financial Economics 81 563ndash594

Hackbarth D and Mauer D C (2012) Optimal priority structure capital structure and

investment Review of Financial Studies 25 747ndash796

Halling M Yu J and Zechner J (2012) Leverage dynamics over the business cycle

Working paper University of Utah University of New South Wales and Vienna

University of Economics and Business AdministrationHennessy C A and Whited T A (2005) Debt dynamics Journal of Finance 60 1129ndash1165

Huang R and Ritter J R (2009) Testing theories of capital structure and estimating the

speed of adjustment Journal of Financial and Qualitative Analysis 44 237ndash271

1452 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

Hovakimian A Opler T and Titman S (2001) The debt-equity choice Journal of

Financial and Quantitative Analysis 36 1ndash24Jung K Kim Y and Stulz R M (1996) Timing investment opportunities managerial

discretion and the security issue decisions Journal of Financial Economics 42 159ndash185Kane A Marcus A J and McDonald R L (1984) How big is the tax advantage to

debt Journal of Finance 39 841ndash853Kayhan A and Titman S (2007) Firmsrsquo histories and their capital structures Journal of

Financial Economics 83 1ndash32Leary M T (2009) Bank loan supply lender choice and corporate capital structure

Journal of Finance 64 1143ndash1185Leary M T and Roberts M R (2005) Do firms rebalance their capital structures

Journal of Finance 60 2575ndash2619Leland H E (1994) Corporate debt value bond covenants and optimal capital structure

Journal of finance 49 1213ndash1252Lemmon M L Roberts M R and Zender J F (2008) Back to the beginning persist-

ence and the cross-section of corporate capital structure Journal of Finance 63 1575ndash1608Long M S and Malitz I B (1985) The investment-financing nexus some empirical

evidence Midland Corporate Finance Journal 3 53ndash59Mackie-Mason J K (1990) Do taxes affect corporate financing decisions Journal of

Finance 45 1471ndash1493Morellec E Nikolov B and Schurhoff N (2012) Corporate governance and capital

structure dynamics Journal of Finance 67 803ndash848Myers S C (1984) The capital structure puzzle Journal of Finance 39 575ndash592Myers S C (1993) Still searching for optimal capital structure Journal of Applied

Corporate Finance 61 4ndash14Rajan R and Zingales L (1995) What do we know about capital structure some evidence

from international data Journal of Finance 50 1421ndash1460Strebulaev I A (2007) Do tests of capital structure theory mean what they say Journal of

Finance 62 1747ndash1787

Strebulaev I A and Whited T M (2012) Dynamic Models and Structural Estimation in

Corporate Finance Now Publisher Boston and Delft

Titman S and Wessels R (1988) The determinants of capital structure choice Journal of

Finance 43 1ndash21

Welch I (2004) Capital structure and stock returns Journal of Political Economy 112

106ndash131

Welch I (2011) Two common problems in capital structure research the financial-debt-

to-asset ratio and issuing activity versus leverage changes International Review of Finance

11 1ndash17

PROFITSndashLEVERAGE PUZZLEREVISITED 1453

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

among the small low-profit firms that are also experiencing sharp reductionsin the value of equity In column (2) we see that both for the large and for thesmall firms net debt increases are primarily found in the highest quintile offirm profitability

6 Debt and Equity Regressions

So far we have demonstrated that in our data the conventional leverageregressions have the usual signs We then demonstrated that in simple sortsof the data strikingly different active financing patterns emerge The nexttask is to reconcile these differencesTable VII presents simple regressions predicting changes in debt book

equity and market equity as well as equity issuances Consistent with thesorts we control for changes in firm size and we also include firm and yeardummies There may be a concern about the speed of a firmrsquos responsesAccordingly we consider the lagged change in profits as the main case butwe also include the contemporaneous change in profits Whether the contem-poraneous change in profits is included or not does not in fact matter Wereport panel-robust standard errors adjusted for clustering at the firm levelThe results in Table VII support the findings in the sorts Firms that have

an increase in profits in 1 year increase debt both in the same year and in thenext year Firms that have an increase in profits experience an increase inboth the book value of equity and the market value of equity Importantlyhowever firms that experience and increase in profits reduce their equityissues All these effects are statistically significant (except for the marketvalue of equity in the column 5 specification) and seem intuitively reason-able In Appendix Table AIV similar regressions to those presented aboveare reported but without firm-fixed effects We find that empirically verylittle changes whether we include firm fixed effects or leave them out Thusthe targeting behavior to the extent that it happens does not account forwhat we are reportingThe next step is to bring the sorts and the simple regressions together into

a common setting We therefore include factors in addition to profitability(i) median industry leverage (ii) the market-to-book assets ratio (iii) tangi-bility of assets and (iv) firm size (measured by the log of assets) Rajan andZingales (1995) show that these factors are related to leverage in G7countries A number of studies have used these factors to estimateleverage targets Frank and Goyal (2009) show that these factors arerobustly related to leverage in the USA In these regressions we use indica-tors for the quintile that the firm is in for each factor

1436 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

Table VIII reports these results In Columns (1) and (2) we examine debtissues and again show that profits positively affect debt issuances The effectsare large and statistically significant at the 1 level Columns (3) and (4)which examine book equity show a strong effect of profits This effect is onlyslightly reduced by the inclusion of other factors Columns (5) and (6) illus-trate that the impact of profits on the change in market equity is also robustto the inclusion of the conventional factors Columns (7) and (8) examineequity issuances Here we again find that profits have a significant negativeimpact on equity issuances As before the results are robust to the inclusionof the conventional factorsIn unreported tables we estimate debt and equity changes using quantile

regressions In terms of the signs and significance of the profitability variable

Table VII Debt and equity changes

The sample contains nonfinancial firms listed on the annual Compustat files for the period

from 1971 to 2009 The table presents fixed effect estimates from regressions of changes indebt (D) changes in the book value of equity (BVE) changes in the market value ofequity (MVE) and net equity issuances (EqyIss) The explanatory variables include

current and lagged change in operating income before depreciation (Profits andProfitst1) and lagged change in the book value of assets (Assetst1) In addition tofixed firm effects the regressions include year indicator variables The reported t-statistics

are corrected for clustering at the firm level

aMean significant at the 1 level

bMean significant at the 5 level

cMean significant at the 10 level

D BVE MVE EqyIss

(1) (2) (3) (4) (5) (6) (7) (8)

Profitst1 022a 025a 015b 035a 004 060b 005b 007a

(33) (31) (25) (35) (03) (25) (25) (53)

Profits 050b 078a 140a 003a

(24) (32) (48) (34)

Assetst1 000 007b 025a 001b

(00) (21) (35) (25)

Constant 167 283a 34 51 1145a 1182a 605a 595a

(14) (30) (04) (05) (32) (29) (69) (74)

Year FE Yes Yes Yes Yes Yes Yes Yes Yes

Firm FE Yes Yes Yes Yes Yes Yes Yes Yes

R2 001 004 001 010 001 005 002 003

N 162056 162056 162130 162130 157550 157550 162154 162154

PROFITSndashLEVERAGE PUZZLEREVISITED 1437

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

the results are qualitatively similar to the OLS results reported in Table VIIIProfitability positively affects debt issuances and negatively affects equityissuances across various quantiles Overall the results in Table VIII showthat there is nothing that is special about using sorts or regressions to explaindebt and equity issues In either case we find that more profitable firms tendto increase their debt experience an increase in the value of equity andrepurchase shares Thus the control factors are not responsible for theusual rejection of the trade-off theory It would appear that what mattersis the size of the issuing activity relative to the organic increases in equityvalue due to profits In other words what matters in the leverage regressionseems to be coming from the presence of E in D

DthornE and not from D

Table VIII Debt and equity issuances profitability and leverage factor quintiles

The sample contains nonfinancial firms listed on the annual Compustat files for the period

from 1971 to 2009 The table presents estimates from regressions of change in debt (D)change in the book value of equity (BVE) change in the market value of equity (MVE)and net equity issuances (EquityIss) The explanatory variables include lagged profitability

quintiles and lagged leverage factor quintiles In addition to fixed firm effects the regres-sions include year indicator variables The reported t-statistics are corrected for clustering atthe firm level

aMean significant at the 1 level

bMean significant at the 5 level

cMean significant at the 10 level

D BVE MVE EqyIss

(1) (2) (3) (4) (5) (6) (7) (8)

ProfitabilityQint 214a 167a 283a 175a 80 429a 53a 55a

(70) (54) (66) (59) (08) (48) (55) (62)

IndLevQint 104a 43 99 20

(29) (15) (10) (10)

Ln(Assets)Qint52 17 1026a 53a

(16) (05) (77) (35)MB

Qint158a 372a 1155a 09

(45) (56) (106) (06)

TangibilityQint 05 177 436c 64a

(01) (15) (18) (31)

Constant 167 96 656b 235 4613a 8610a 113 570a

(08) (04) (26) (04) (61) (58) (11) (31)

Year FE Yes Yes Yes Yes Yes Yes Yes Yes

Firm FE Yes Yes Yes Yes Yes Yes Yes Yes

R2 0002 0002 0003 0004 0013 0014 0013 0014

N 155587 155037 155587 155037 155587 155037 155587 155037

1438 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

7 Are Financial Market Conditions Important

According to the trade-off theory capital structure is determined by con-sidering the impact of costs and benefits of debt The time variation inthe costs and benefits of debt imply that issuance decisions would alsovary over time Furthermore market conditions affect the transactioncosts of adjusting Adverse selection is a greater problem in a cold marketthan in a hot one Accordingly it may be easier to adjust in good marketconditions Related ideas have been developed by Huang and Ritter(2009) Faulkender et al (2012) and Halling Yu and Zechner (2012)Faulkender Flannery Hankins and Smithrsquos paper examines cross-sectionalvariation in adjustment speeds including market valuations andhow they may affect adjustment speeds It is natural to think that ingood market conditions it will be less costly to issue both debt and equityand adjusting leverage to respond to profitability shocks would be mucheasierTo test the importance of market conditions we require a definition of

good times and bad times Our empirical strategy is to estimate good timesversus bad times at the four-digit industry level We define an industry ashaving ldquogood timesrdquo if the median firm in that industry has a market-to-book ratio that is higher than the 67th percentile of the time-series distribu-tion of the industry median market-to-book ratios Conversely an industryis defined as having ldquobad timesrdquo if the median firm in the industry has amarket-to-book ratio that is lower than the 33rd percentile of the time-seriesdistribution of the industry median market-to-book ratiosPanel A of Table IX tabulates issuance activity for profitability sorts in

both good times and bad times As might be expected active debt and equityissuances are larger during good times There is also somewhat more activeswaps between debt and equity during good timesPanel B reports the issuance activity for the smallest and the largest firms

during the good times and bad times Within each size quintile firms aresorted on profitability This is a three-way sort As before we find that thereis much more active use of external markets during good times Small low-profit firms are more likely to issue equity in good times than in bad timesDebt issuances are significantly higher in good times Large high-profitfirms are significantly more likely to issue debt and repurchase equity ingood times than in bad timesPanel C reports the magnitudes of the financial variables rather than the

frequencies During bad times less profitable large firms retire substantialamounts of debt and they show a tendency to issue equity Small firms donot seem to engage in similar debt reduction activities However like the

PROFITSndashLEVERAGE PUZZLEREVISITED 1439

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

Table IX Debt and equity issues in good and bad times

The table reports the frequency and magnitude of financing activity for sorts on profits for subsamples

of firms in good and bad times An industry is defined as having ldquogood timesrdquo if the median firm in

that industry has a market-to-book ratio that is higher than the 67th percentile of the time-series

distribution of industry median market-to-book ratios Conversely an industry is defined as having

bad times if the median firm in that industry has a market-to-book ratio that is lower than the 33rd

percentile of the time-series distribution of industry median market-to-book ratios Panels A and B

report the percentage of firms (i) issuing net debt in excess of 5 of the value of their assets (ii) issuing

net equity in excess of 5 of the value of their assets (iii) issuing debt and repurchasing equity both in

excess of 5 of the value of their assets and (iv) issuing equity and retiring debt both in excess of 5

of the value of their assets Panel C tabulates the mean debt and equity levels and the mean changes in

debt and equity issuances The sample contains nonfinancial firms listed on the annual Compustat files

for the period from 1971 to 2009

Iss Iss Iss D Iss ENet D Net E Rep E Ret D(1) (2) (3) (4)

Panel A sorts on profitability

Badtimes

Low 145 131 01 202 170 49 02 083 173 46 04 074 182 47 03 10

High 177 51 06 10t Highfrac14Low 41 136 33 01

Goodtimes

Low 206 381 01 462 197 148 02 233 212 103 04 184 234 92 10 18

High 216 92 16 14t Highfrac14Low 23 658 138 23

Panel B sorts on firm size and profitability

Badtimes

Smallfirm

s Low 151 318 00 312 182 276 02 253 151 140 02 294 183 73 00 19

High 177 80 00 11t Highfrac14Low 10 90 NA 20

Largefirm

s Low 172 54 01 042 148 39 04 053 150 43 04 044 156 33 03 06

High 147 28 05 03t Highfrac14Low 17 32 20 02

Goodtimes Smallfirm

s Low 213 448 01 542 230 443 01 473 222 325 00 474 203 197 02 38

High 190 153 02 31t Highfrac14Low 19 230 10 39

Largefirm

s Low 203 96 03 122 208 66 09 133 232 54 16 064 215 39 19 05

High 234 36 38 04t Highfrac14Low 29 99 92 38

(continued)

1440 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

larger firms there is a tendency for small firms to issue more equitymdashat leastto the extent possibleDuring good times large profitable firms raise significant amounts of debt

as they experience an increase in the value of their equity Such firms alsoengage in share repurchases Debt issues by smaller firms are much lessaffected by their own profitability during good times and their change indebt is more or less independent of their profits However equity issuanceis primarily found among the low-profit small firms during good timesTable IX shows that empirically profitability has a time-varying impacton leverage choicesIn Table X the basic regressions for changes in debt and for equity

issuance are presented for both good times and bad times Comparinggood times with bad times we see that the effects of profits are muchstronger in good times In bad times the impacts are rather weak statisticallyThis difference is interesting and deserves further study However for ourpurposes recall that the static trade-off theory implies time-varying capital

Table IX Continued

D D BVE BVE MVE MVE EqyIss Assets

(1) (2) (3) (4) (5) (6) (7) (8)

Panel C magnitude of financing activity

Badtimes Smallfirm

s Low 40 03 58 14 305 06 22 140

2 44 05 82 07 240 67 17 180

3 45 04 94 02 176 18 05 187

4 43 09 98 02 137 04 03 188

High 36 10 114 13 209 16 04 199

Largefirm

s Low 3365 73 2523 169 2304 163 94 8972

2 3031 95 2983 35 3284 278 56 9068

3 2930 5 3360 14 3947 308 57 9225

4 2281 6 3817 19 5213 442 7 8902

High 1790 62 4767 13 8799 58 31 9573

Goodtimes

Smallfirm

s Low 24 06 38 05 339 23 39 90

2 26 08 48 01 301 04 25 104

3 30 09 60 04 246 03 17 125

4 28 07 66 08 206 18 12 130

High 23 06 85 16 261 13 08 141

Largefirm

s Low 4532 36 3477 103 6557 260 53 12465

2 4580 58 3527 139 6911 421 23 12522

3 2952 111 3485 2 7586 279 32 9535

4 2752 124 3939 72 10686 33 105 10226

High 2429 228 4819 377 18565 268 222 10830

PROFITSndashLEVERAGE PUZZLEREVISITED 1441

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

structure choices even when the target is time invariant Empirically we doobserve time-varying choices

8 Is Scaling Important

In the preceding section we provided results for unscaled debt and equityissues This seems appropriate to us However it involves making twochanges to the standard regression We would like to determine if both

Table X Debt and equity changes in good and bad timesmdashregressions

The sample contains nonfinancial firms listed on the annual Compustat files for the period

from 1971 to 2009 The table presents estimates from regressions of change in debt (D)and net equity issuances (EqyIss) in good and bad times The explanatory variables includechange in profits lagged change in profits and lagged change in assets The regressions

include both firm and year fixed effects An industry is defined as having ldquogood timesrdquo ifthe median firm in that industry has a market-to-book ratio that is higher than the 67thpercentile of the time-series distribution of industry median market-to-book ratios

Conversely an industry is defined as having bad times if the median firm in thatindustry has a market-to-book ratio that is lower than the 33rd percentile of the time-series distribution of industry median market-to-book ratios The reported t-statistics arecorrected for clustering at the firm level

aMean significant at the 1 level

bMean significant at the 5 level

cMean significant at the 10 level

Good times Bad times

D EqyIss D EqyIss

(1) (2) (3) (4) (5) (6) (7) (8)

Profits 087a 091a 006b 009a 060 065 003 003b

(30) (29) (20) (28) (15) (16) (16) (20)

Profitst1 029c 015a 034c 001

(19) (37) (17) (04)

Assetst1 000 002b 000 001

(02) (22) (00) (07)

Constant 1337 3989 8691a 10001a 1351 1851 4099a 4055a

(03) (09) (55) (37) (09) (13) (80) (78)

Year FE Yes Yes Yes Yes Yes Yes Yes Yes

Firm FE Yes Yes Yes Yes Yes Yes Yes Yes

R2 0071 0080 0018 0045 0178 0212 0034 0041

N 74827 68729 74891 68787 26965 26013 26971 26018

1442 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

changes are crucial Accordingly in this section we take a step back toconsider prior literature in which variables are scaled The questions arehow does scaling affect the results and whyTable XI examines alternative scalings Columns (1) and (2) scale debt and

equity issuances by the total firm net issuance Columns (3) and (4) scaledebt and equity issuances by the sum of debt and book equity Columns (5)and (6) consider the change in debt scaled by the sum of debt and bookequity Columns (7) and (8) consider the change in leverage due to externalfinancing Since we are examining changes we use the changes in thestandard regression factors as controls The primary interest is in the coef-ficient on the change in profitsWe find that a change in profits is associated with an increase in debt

issues But the sign reverses when we scale the debt issues by the sum of debtplus book equity This is telling What is really driving the results is thechange in equity in the denominator not the change in debt in the numer-ator This result is further substantiated in Columns (5) and (6) where thechange in debt to capital ratio is used instead of debt issuanceTo further explore the impact of active issuance relative to passive

experiencing of changes we decompose leverage by separating out theeffect of debt and equity issues from those due to changes in retainedearnings Thus we define leverage change due to external financing as

LevEF frac14Dt1 thorn d

Dt1 thorn Et1 thorn dthorn e

Dt1

Dt1 thorn Et1 eth1THORN

where D is debt E is book value of equity d is debt issues net of retirementsand e is equity issues net of repurchases Garvey and Hanka (1999) use asimilar measure in their study This measure directly captures the netleverage effect of debt and equity issuancesColumns (7) and (8) report results from regressions of the net leverage

effect of debt and equity issues on changes in profitability after controllingfor other leverage factors in differences and year indicator variables Theeffect of profitability on leverage change due to external financing is positiveand statistically significant This comparison suggests that changes in totaldebt and equity are only partly a result of debt and equity issuance decisionsOther balance sheet adjustments complicate the inferences from leverageratio regressionsFor most of our analysis we follow the literature in treating debt as homo-

genous It is of course potentially interesting to ask what happens when wedistinguish between short-term debt and long-term debt Which part of thefirmrsquos debt maturity structure drives the under-adjustment Table XII

PROFITSndashLEVERAGE PUZZLEREVISITED 1443

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

provides results from such decompositions Alternative scalings areconsidered The main results of interest are that the distinction betweenshort- and long-term debt is important Short-term debt issuance is largelyindependent of changes in profits The negative sign on profit change stemsfrom the action of long-term debt This seems intuitively reasonable

Table XI Scaled issuances

The sample contains nonfinancial firms listed on the annual Compustat files for the period

from 1971 to 2009 The table presents estimates from regressions of debt issuance to totalissuance (in columns (1) and (2)) debt issuance to total capitalization (in columns (3) and(4)) change in debt to capitalization ratio (in columns (5) and (6)) and change in leverage

ratio considering only external financing (in columns (7) and (8)) The change in leverageratio considering only external financing is defined as

LevEF frac14Dt1 thorn d

Dt1 thorn Et1 thorn dthorn e

Dt1

Dt1 thorn Et1

where D is debt E is book value of equity d is debt issues net of retirements and e is equity

issues net of repurchases The regressions include lagged leverage factors and year indicatorvariables Some of the specifications include firm fixed effects The reported t-statistics arecorrected for clustering at the firm level

aMean significant at the 1 level

bMean significant at the 5 level

cMean significant at the 10 level

DebtIssTotIss

DebtIssethDthornETHORN D

DthornE LevEF

(1) (2) (3) (4) (5) (6) (7) (8)

Profitt1 004a 000 002a 001c 003a 002 001b 002a

(35) (03) (34) (18) (35) (14) (23) (26)

IndLevt1 004c 001 001 001c 000 000 000 000

(19) (06) (15) (19) (02) (04) (03) (05)

Market=bookt1 001a 001a 000a 000a 000a 000a 000 000c

(47) (85) (68) (60) (55) (36) (05) (18)

Tangt1 004 005 013a 011a 011a 006a 003a 003a

(12) (15) (144) (112) (79) (38) (44) (42)

Assetst1 008a 004a 006a 004a 002a 001a 003a 002a

(97) (45) (283) (161) (53) (35) (147) (101)

Constant 079a 074a 002a 003a 000 000 000b 002a

(497) (442) (91) (111) (00) (05) (22) (93)

Year FE Yes Yes Yes Yes Yes Yes Yes Yes

Firm FE No Yes No Yes No Yes No Yes

R2 0015 0098 0022 0076 0009 0070 0012 0086

N 133905 133905 140552 140552 138229 138229 140572 140572

1444 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

inasmuch as short-term debt is probably more associated with day-to-dayoperations rather than longer term planning considerationsIn Table XIII we examine change in leverage ( D

DthornE) and change inleverage due to external financing (LEVEF) during good and bad timesseparately The traditional negative effect of a change in profits on achange in leverage is found both in good times and in bad timesHowever when we decompose the impact in order to isolate the externalcomponent as before the result disappears In good times the externalfinancing component of a change in profits is weak but still positiveDuring bad times it is not statistically significant This implies that distin-guishing between good and bad times does not fundamentally alter the basic

Table XII Short- and long-term debt issuances

The sample contains nonfinancial firms listed on the annual Compustat files for the period

from 1971 to 2009 The table presents estimates from regressions of short-term debtissuance to total issuance (in columns (1) and (2)) long-term debt issuance to totalissuance (in columns (3) and (4)) short-term debt issuance to total capitalization (in

columns (5) and (6)) and long-term debt to total capitalization (in columns (7) and (8))Some of the specifications include firm fixed effects The reported t-statistics are correctedfor clustering at the firm level

aMean significant at the 1 level

bMean significant at the 5 level

cMean significant at the 10 level

STDebtIssTotIss

LTDebtIssTotIss

STDebtIssethDthornETHORN

LTDebtIssethDthornETHORN

(1) (2) (3) (4) (5) (6) (7) (8)

Profitabilityt1 0002 0003 0042a 0008 0001 0001 0019a 0012a

(02) (04) (30) (05) (14) (10) (45) (28)

IndLevt1 0009 0007 0070b 0033 0001 0002c 0014a 0005

(06) (05) (26) (11) (08) (17) (26) (09)

M=Bt1 0003a 0003a 0003b 0010a 0000 0000 0003a 0002a

(48) (35) (22) (65) (10) (01) (84) (42)

Tangibilityt1 0020 0007 0011 0067 0005a 0006a 0097a 0081a

(12) (03) (03) (16) (34) (33) (108) (89)

Assetst1 0039a 0011b 0034a 0024b 0001a 0001 0046a 0020a

(89) (22) (38) (23) (36) (13) (190) (82)

Constant 0001b 0153a 0798a 0591a 0000c 0006a 0057a 0096a

(25) (144) (470) (288) (17) (124) (292) (234)

Year FE Yes Yes Yes Yes Yes Yes Yes Yes

Firm FE No Yes No Yes No Yes No Yes

R2 0025 0089 0018 0073 0030 0147 0016 0298

N 136146 136146 136140 136140 143031 143031 142847 142847

PROFITSndashLEVERAGE PUZZLEREVISITED 1445

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

Table XIII Externally financed leverage changes in good and bad times

The sample contains nonfinancial firms listed on the annual Compustat files for the period

from 1971 to 2009 The table presents estimates from regressions of change in debt tocapitalization ratio (in columns (1) (2) (5) and (6)) and change in leverage ratio consider-ing only external financing (in columns (3) (4) (7) and (8)) The change in leverage ratio

considering only external financing is defined as

LevEF frac14Dt1 thorn d

Dt1 thorn Et1 thorn dthorn e

Dt1

Dt1 thorn Et1

where D is debt E is book value of equity d is debt issues net of retirements and e is equityissues net of repurchases The regressions include lagged leverage factors and year indicator

variables The reported t-statistics are corrected for clustering at the firm level An industryis defined as having ldquogood timesrdquo if the median firm in that industry has a market-to-bookratio that is higher than the 67th percentile of the time-series distribution of industry

median market-to-book ratios Conversely an industry is defined as having ldquobad timesrdquoif the median firm in that industry has a market-to-book ratio that is lower than the 33rdpercentile of the time-series distribution of industry median market-to-book ratios

aMean significant at the 1 level

bMean significant at the 5 level

cMean significant at the 10 level

Good times Bad times

DDthornE LevEF D

DthornE LevEF

(1) (2) (3) (4) (5) (6) (7) (8)

Profitabiltyt1 0022c 0005 0016c 0025b 0073b 0041 0007 0007

(17) (03) (18) (24) (25) (11) (04) (03)

IndLevt1 0000 0011 0001 0009 0010 0018 0001 0000

(00) (09) (01) (13) (07) (10) (02) (00)

MB

t1

0004a 0003b 0000 0001 0008a 0001 0000 0000

(40) (24) (05) (08) (28) (04) (03) (03)

Tangt1 0097a 0041 0033a 0038b 0127a 0048 0059a 0033

(45) (16) (27) (25) (49) (14) (43) (16)

Assetst1 0014a 0013b 0024a 0019a 0037a 0023c 0036a 0025a

(30) (20) (84) (48) (45) (17) (88) (43)

Constant 0030b 0036 0000 0003 0028c 0014 0021 0001

(20) (16) (00) (07) (17) (06) (11) (01)

Year FE Yes Yes Yes Yes Yes Yes Yes Yes

Firm FE No Yes No Yes No Yes No Yes

R2 0006 0088 0009 0117 0026 0251 0023 0202

N 56820 56820 58200 58200 23462 23462 23511 23511

1446 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

conclusions The traditional leverage result on profits is driven by increasesin equity that are experienced and then partially but not completely offsetby firmsTo the extent that firms are engaging in active financing they are following

the prescriptions of the trade-off theory Thus we have a coherent picture ofhow firms are responding to profits Basically more profitable firms experiencean increase in the value of their equity They respond to this by issuing debtand repurchasing equity But the magnitude of the increase in equity is largerthan the magnitudes of the active steps taken in the debt and equity marketsThis means that the firmsrsquo actual actions are as predicted by the trade-offtheory But it leaves open the question of why they do not act more stronglyThe capital structure literature has long been interested in dynamics and

capital structure adjustment paths In recent years this interest has increas-ingly resulted in explicit dynamic optimizing models such as those ofGoldstein Ju and Leland (2001) Hennessy and Whited (2005) andStrebulaev (2007) These models commonly make specialized assumptionsleading some to question their generalityWe argue that both fixed transactions costs and variable transactions costs

are plausible and realistic in models that pin down firm size The simultaneouspresence of both is important Due to fixed costs firms only rebalance inter-mittently Due to variable adjustment costs the extent of the adjustment isreduced to the point at which an extra unit to adjustment is balanced againstthe extra unit transactions costs This means that both leverage increases andleverage reductions will only go part way toward a static optimumThe key point is that with both fixed and variable transactions costs we

should expect to only see intermittent leverage adjustments When leverage isadjusted it should ldquounder adjustrdquo when compared to a frictionless modelFurther discussion of transactions costs in a capital structure setting isprovided by Leary and Roberts (2005)

9 Conclusion

The connection between corporate profits and capital structure has beenvery influential in the assessment of the trade-off theory The conventionalinterpretation of the evidence has pushed the literature away from the trade-off theory and toward much more complex dynamic models and ideas As aresult it is important to make sure that the evidence has been correctlyinterpreted We show that the problem is with the interpretation of thetest results and a failure to take into account the costs associated with re-sponding to a profit shock

PROFITSndashLEVERAGE PUZZLEREVISITED 1447

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

We document that more profitable firms really do borrow more (not less)They tend to repurchase equity They experience an increase in both thebook value of equity and the market value of equity Less profitable firmsreally do tend to reduce their debt and to issue equityFirm size and market conditions also matter Larger firms tend to be more

active in the debt markets while smaller firms tend to be relatively moreactive in the equity markets During good times there is more use ofexternal financingThe usual profitsndashleverage puzzle result is primarily driven by the increase

in equity that is experienced by the more profitable firms Accordingly thepuzzle should be restated as asking why do firms not take sufficiently largeoffsetting actions to fully undo the change in equity What limits the mag-nitudes of the typical leverage response to profit shocksIn a frictionless model the partial response appears puzzling But there is

good empirical reason to believe that rebalancing entails both fixed andvariable costs and that firm size matters The rebalancing costs can befully avoided by doing nothing Accordingly the firm must decide whetherany given shock is big enough to be worth responding to If it is then thefirm must decide how big a response is called for These technical conditionsare known as ldquovalue matchingrdquo and ldquosmooth pastingrdquo Thus optimizationimplies that some shocks will be ignored Even if the shock is not ignoredthe optimal response will only partially undo the shock The magnitude ofthe leverage response must balance the marginal cost and the marginalbenefit of an extra unit of leverage Since the marginal cost of adjustingleverage is strictly positive the adjustment toward that static leverageoptimum will only go part way This is true both for leverage increasesand for leverage reductionsOne might conjecture that the variable costs would be time varying and

that such costs might have declined over time Since we do not have a fullaccounting of all of the variable costs it is of course impossible to be sureHowever there is some evidence suggesting that at least the directunderwriting fees are remarkably stable over time10

Surely nobody ought to be surprised that more profitable firms are moreinclined to repurchase equity while unprofitable firms tend not to do so Thefact that more profitable firms typically experience an increase in the value of

10 For IPOs Jay Ritter provides important summary evidence as httpbearwarringtonufleduritterIPOs2013Underwritingpdf He finds that for medium-sized deals a remark-

ably large fraction of the deals have a 7 fee year after year Whether this is due tobehavioral factors or some sort of a coordination device is not clear The extent towhich similar intertemporal stability carries over to other forms of external financemerits further study

1448 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

equity is equally natural Were it not for the contrary claims in the literature(ie the profitsndashleverage puzzle) we would argue that nobody ought to findit surprising that more profitable firms are more inclined to issue debt Thefact that leverage adjustments typically only partially offset profit shocksshould also sound quite natural These facts all fit together easily within thetrade-off theory once we take account of both fixed and variable costs ofactively adjusting capital structure

Appendix

Table AI Market leverage regressions

The table presents estimates of the regressions of market leverage on firm characteristics

The sample comes from the annual Compustat files during the period 1971ndash2009 Financialfirms are excluded The leverage is estimated as the ratio of debt over debt plus marketequity The explanatory variables Profitabilityt1

MB

t1

Tangibilityt1 and LnethAssetsTHORNt1are described in Table 1 IndMedianLevt1 is estimated as the median market leverage of allother firms in the same industry The industry is defined at the level of the four-digit SICcode All specifications include the year fixed effects The specifications in column (2)

additionally include the firm fixed effects We report t-statistics where the standard errorsare clustered at the firm level in parentheses

aSignificant at the 1 level

Quantile regressions

(1) (2) (3) (4) (5)

Profitabilityt1 013a 013a 007a 011a 020a

(275) (262) (349) (425) (538)

IndLevt1 046a 020a 032a 058a 063a

(506) (239) (1299) (1630) (1042)MB

t1

003a 001a 001a 002a 004a

(509) (303) (743) (709) (534)

Tangibilityt1 013a 018a 013a 014a 010a

(191) (181) (762) (571) (232)

LnethAssetsTHORNt1 001a 005a 001a 001a lt001a

(112) (258) (682) (342) (29)

Constant 010a 011a 005a 005a 029a

(171) (98) (173) (118) (398)

Year FE Yes Yes Yes Yes Yes

Firm FE No Yes No No No

Regression type OLS FE 25th tile Median 75thtile

Clustered SE Yes Yes No No No

R2 030 067

Pseudo R2 016 023 0201

N 158578 158578 158578 158578 158578

PROFITSndashLEVERAGE PUZZLEREVISITED 1449

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

Table AII Profitability sorts for firms issuing large amounts

Percentage of firms

Iss Ret Iss Iss Rep Iss Iss IssD IssE

D D NetD E E NetE DampE RepE RetD

(1) (2) (3) (4) (5) (6) (7) (8) (9)

Panel A 10 cutoff to define issuing or retiring activity

Allfirm

s

Low 205 150 134 260 05 257 45 00 23

2 207 175 103 77 06 75 11 01 10

3 243 196 101 50 08 47 08 01 09

4 252 186 113 50 13 48 08 02 08

High 227 161 118 60 27 57 07 04 09

t Highfrac14Low 70 38 62 755 226 768 320 90 152

Smallfirm

s Low 230 132 162 380 05 375 82 00 32

2 228 143 165 334 04 333 65 00 26

3 211 146 147 216 03 214 38 00 23

4 190 154 120 126 04 124 15 01 20

High 194 162 114 99 09 97 12 01 15

t Highfrac14Low 42 40 68 342 20 340 169 07 131

Largefirm

s Low 231 200 82 31 06 29 08 01 04

2 231 187 67 17 06 14 03 01 02

3 234 177 77 17 09 14 03 02 03

4 218 158 75 16 17 14 02 03 02

High 218 135 99 19 41 15 03 07 02

t Highfrac14Low 19 106 36 46 135 57 39 58 23

Percentage of firms

Iss Ret Iss Iss Rep Iss Both Iss D Iss E

D D Net D E E Net E DampE Rep E Ret D

(1) (2) (3) (4) (5) (6) (7) (8) (9)

Panel B 15 cutoff to define issuing or retiring activity

Allfirm

s

Low 158 101 100 230 01 228 27 00 15

2 142 115 63 57 01 56 05 00 06

3 165 128 57 32 01 31 03 00 05

4 173 118 65 35 01 34 03 00 05

High 160 102 70 46 04 44 03 00 05

t Highfrac14Low 08 06 141 738 87 746 277 24 131

Smallfirm

s Low 195 95 132 354 01 350 56 00 24

2 185 96 131 303 01 301 44 00 18

3 160 96 104 179 00 176 18 00 15

4 134 99 78 99 01 98 08 00 12

High 133 95 73 82 01 80 06 00 10

t Highfrac14Low 80 00 95 345 12 344 149 12 52

Largefirm

s Low 145 123 42 14 01 12 03 00 01

2 137 114 32 07 01 06 01 00 01

3 142 104 35 06 01 05 00 00 01

4 130 91 36 08 03 07 01 01 01

High 134 81 48 09 06 07 01 01 01

t Highfrac14Low 19 85 17 26 54 33 34 17 11

1450 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

Table AIV Debt and equity changesmdashwithout fixed effects

The sample contains nonfinancial firms listed on the annual Compustat files for the period from 1971

to 2009 The table presents estimates from regressions of change in debt (D) change in the book

value of equity (BVE) change in the market value of equity (MVE) and net equity issuances

(EqyIss) The explanatory variables include current and lagged change in operating income before

depreciation (Profits and Profitst1) and lagged change in book value of assets (Assetst1) The

regressions include year indicator variables The reported t-statistics are corrected for clustering at the

firm level

aMean significant at the 1 level

bMean significant at the 5 level

cMean significant at the 10 level

D BVE MVE EqyIss

(1) (2) (3) (4) (5) (6) (7) (8)

Profitst1 025a 024a 023a 038a 007 066a 007a 009a

(40) (32) (36) (39) (06) (29) (42) (51)Profits 051b 082a 149a 004a

(25) (33) (49) (33)Assetst1 002 005c 024a 001b

(08) (17) (34) (23)Constant 93 176a 45 71 1552a 1704a 130a 140a

(12) (28) (07) (09) (49) (49) (45) (45)Year FE Yes Yes Yes Yes Yes Yes Yes YesR2 001 005 001 011 001 005 002 002N 162056 162056 162130 162130 157550 157550 162154 162154

Table AIII Median profitability sorts

The sample contains nonfinancial firms listed on the annual Compustat files for the period from 1971

to 2009 The table presents median debt and equity median changes in debt and equity and median

equity issuances (EqyIss) in millions of dollars for all firms sorted on profitability and for profitability

sorts within the smallest and largest firms The sorts are done annually

D D BVE BVE MVE MVE EqyIss Assets(1) (2) (3) (4) (5) (6) (7) (8)

Allfirm

s

Low 21 00 75 17 256 20 01 1922 186 02 449 18 598 19 00 11033 816 06 1274 03 1637 05 00 32244 787 02 1541 18 2488 03 00 3556High 273 00 1361 66 3300 14 00 2578

Smallfirm

s Low 05 00 13 07 140 14 06 402 06 00 25 08 114 12 01 563 09 00 32 05 85 07 00 674 12 00 39 01 68 03 00 79High 10 00 51 04 98 01 00 90

Largefirm

s Low 11485 381 10269 371 13227 202 03 333902 13051 393 12786 140 17571 169 05 392463 11554 287 13429 24 21623 305 00 370314 9963 209 15052 194 29424 308 00 36705High 7361 00 16380 613 49142 439 00 34976

PROFITSndashLEVERAGE PUZZLEREVISITED 1451

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

References

Auerbach A J (1985) Real determinants of corporate leverage in B M Friedman (ed)

Corporate Capital Structures in the United States University of Chicago Press Chicago

IllinoisBertrand M and Mullainathan S (2003) Enjoying the quiet life corporate governance and

managerial preferences Journal of Political Economy 111 1043ndash1075Booth L Varouj A Demirguc-Kunt A and Maksimovic V (2001) Capital structures in

developing countries Journal of Finance 56 87ndash130Brennan M J and Schwartz E S (1984) Optimal financial policy and firm valuation

Journal of Finance 39 593ndash607Cameron A C and Trivedi P K (2010) Microeconometrics using Stata Stata Press TexasConstantinides G M (1979) Multiperiod consumption and investment behavior with

convex transactions costs Management Science 25 1127ndash1137Constantinides G M and Richard S F (1978) Existence of optimal simple policies for

discounted-cost inventory and cash management in continuous time Operations Research

26 620ndash636DeAngelo H DeAngelo L and Whited T M (2011) Capital structure dynamics and

transitory debt Journal of Financial Economics 99 235ndash261Dixit A K (1993) The Art of Smooth Pasting Routledge London and New YorkFama E and French K R (2002) Testing trade-off and pecking order predictions about

dividends and debt Review of Financial Studies 15 1ndash33Faulkender M and Petersen M A (2006) Does the source of capital affect capital struc-

ture Review of Financial Studies 19 45ndash79Faulkender M Flannery M J Hankins K W and Smith J M (2012) Cash Flows and

Leverage Adjustments Journal of Financial Economics 103 632ndash646Fischer E Heinkel R and Zechner J (1989) Dynamic capital structure choice theory

and tests Journal of Finance 44 19ndash40Frank M Z and Goyal V K (2008) Trade-off and pecking order theories of

debt in B E Eckbo (ed) Handbook of Corporate Finance Empirical Corporate Finance

Vol 2 Handbook of Finance Series Chapter 12 ElsevierNorth-Holland Amsterdam pp

135ndash202

Frank M Z and Goyal V K (2009) Capital structure decisions which factors are reliably

important Financial Management 38 1ndash37

Garvey G T and Hanka G (1999) Capital structure and corporate control the effect of

antitakeover statutes on firm leverage Journal of Finance 54 519ndash546

Goldstein R Ju N and Leland H (2001) An EBIT-based model of dynamic capital

structure Journal of Business 74 483ndash512

Graham J R and Tucker A L (2006) Tax shelters and corporate debt policy Journal of

Financial Economics 81 563ndash594

Hackbarth D and Mauer D C (2012) Optimal priority structure capital structure and

investment Review of Financial Studies 25 747ndash796

Halling M Yu J and Zechner J (2012) Leverage dynamics over the business cycle

Working paper University of Utah University of New South Wales and Vienna

University of Economics and Business AdministrationHennessy C A and Whited T A (2005) Debt dynamics Journal of Finance 60 1129ndash1165

Huang R and Ritter J R (2009) Testing theories of capital structure and estimating the

speed of adjustment Journal of Financial and Qualitative Analysis 44 237ndash271

1452 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

Hovakimian A Opler T and Titman S (2001) The debt-equity choice Journal of

Financial and Quantitative Analysis 36 1ndash24Jung K Kim Y and Stulz R M (1996) Timing investment opportunities managerial

discretion and the security issue decisions Journal of Financial Economics 42 159ndash185Kane A Marcus A J and McDonald R L (1984) How big is the tax advantage to

debt Journal of Finance 39 841ndash853Kayhan A and Titman S (2007) Firmsrsquo histories and their capital structures Journal of

Financial Economics 83 1ndash32Leary M T (2009) Bank loan supply lender choice and corporate capital structure

Journal of Finance 64 1143ndash1185Leary M T and Roberts M R (2005) Do firms rebalance their capital structures

Journal of Finance 60 2575ndash2619Leland H E (1994) Corporate debt value bond covenants and optimal capital structure

Journal of finance 49 1213ndash1252Lemmon M L Roberts M R and Zender J F (2008) Back to the beginning persist-

ence and the cross-section of corporate capital structure Journal of Finance 63 1575ndash1608Long M S and Malitz I B (1985) The investment-financing nexus some empirical

evidence Midland Corporate Finance Journal 3 53ndash59Mackie-Mason J K (1990) Do taxes affect corporate financing decisions Journal of

Finance 45 1471ndash1493Morellec E Nikolov B and Schurhoff N (2012) Corporate governance and capital

structure dynamics Journal of Finance 67 803ndash848Myers S C (1984) The capital structure puzzle Journal of Finance 39 575ndash592Myers S C (1993) Still searching for optimal capital structure Journal of Applied

Corporate Finance 61 4ndash14Rajan R and Zingales L (1995) What do we know about capital structure some evidence

from international data Journal of Finance 50 1421ndash1460Strebulaev I A (2007) Do tests of capital structure theory mean what they say Journal of

Finance 62 1747ndash1787

Strebulaev I A and Whited T M (2012) Dynamic Models and Structural Estimation in

Corporate Finance Now Publisher Boston and Delft

Titman S and Wessels R (1988) The determinants of capital structure choice Journal of

Finance 43 1ndash21

Welch I (2004) Capital structure and stock returns Journal of Political Economy 112

106ndash131

Welch I (2011) Two common problems in capital structure research the financial-debt-

to-asset ratio and issuing activity versus leverage changes International Review of Finance

11 1ndash17

PROFITSndashLEVERAGE PUZZLEREVISITED 1453

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

Table VIII reports these results In Columns (1) and (2) we examine debtissues and again show that profits positively affect debt issuances The effectsare large and statistically significant at the 1 level Columns (3) and (4)which examine book equity show a strong effect of profits This effect is onlyslightly reduced by the inclusion of other factors Columns (5) and (6) illus-trate that the impact of profits on the change in market equity is also robustto the inclusion of the conventional factors Columns (7) and (8) examineequity issuances Here we again find that profits have a significant negativeimpact on equity issuances As before the results are robust to the inclusionof the conventional factorsIn unreported tables we estimate debt and equity changes using quantile

regressions In terms of the signs and significance of the profitability variable

Table VII Debt and equity changes

The sample contains nonfinancial firms listed on the annual Compustat files for the period

from 1971 to 2009 The table presents fixed effect estimates from regressions of changes indebt (D) changes in the book value of equity (BVE) changes in the market value ofequity (MVE) and net equity issuances (EqyIss) The explanatory variables include

current and lagged change in operating income before depreciation (Profits andProfitst1) and lagged change in the book value of assets (Assetst1) In addition tofixed firm effects the regressions include year indicator variables The reported t-statistics

are corrected for clustering at the firm level

aMean significant at the 1 level

bMean significant at the 5 level

cMean significant at the 10 level

D BVE MVE EqyIss

(1) (2) (3) (4) (5) (6) (7) (8)

Profitst1 022a 025a 015b 035a 004 060b 005b 007a

(33) (31) (25) (35) (03) (25) (25) (53)

Profits 050b 078a 140a 003a

(24) (32) (48) (34)

Assetst1 000 007b 025a 001b

(00) (21) (35) (25)

Constant 167 283a 34 51 1145a 1182a 605a 595a

(14) (30) (04) (05) (32) (29) (69) (74)

Year FE Yes Yes Yes Yes Yes Yes Yes Yes

Firm FE Yes Yes Yes Yes Yes Yes Yes Yes

R2 001 004 001 010 001 005 002 003

N 162056 162056 162130 162130 157550 157550 162154 162154

PROFITSndashLEVERAGE PUZZLEREVISITED 1437

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

the results are qualitatively similar to the OLS results reported in Table VIIIProfitability positively affects debt issuances and negatively affects equityissuances across various quantiles Overall the results in Table VIII showthat there is nothing that is special about using sorts or regressions to explaindebt and equity issues In either case we find that more profitable firms tendto increase their debt experience an increase in the value of equity andrepurchase shares Thus the control factors are not responsible for theusual rejection of the trade-off theory It would appear that what mattersis the size of the issuing activity relative to the organic increases in equityvalue due to profits In other words what matters in the leverage regressionseems to be coming from the presence of E in D

DthornE and not from D

Table VIII Debt and equity issuances profitability and leverage factor quintiles

The sample contains nonfinancial firms listed on the annual Compustat files for the period

from 1971 to 2009 The table presents estimates from regressions of change in debt (D)change in the book value of equity (BVE) change in the market value of equity (MVE)and net equity issuances (EquityIss) The explanatory variables include lagged profitability

quintiles and lagged leverage factor quintiles In addition to fixed firm effects the regres-sions include year indicator variables The reported t-statistics are corrected for clustering atthe firm level

aMean significant at the 1 level

bMean significant at the 5 level

cMean significant at the 10 level

D BVE MVE EqyIss

(1) (2) (3) (4) (5) (6) (7) (8)

ProfitabilityQint 214a 167a 283a 175a 80 429a 53a 55a

(70) (54) (66) (59) (08) (48) (55) (62)

IndLevQint 104a 43 99 20

(29) (15) (10) (10)

Ln(Assets)Qint52 17 1026a 53a

(16) (05) (77) (35)MB

Qint158a 372a 1155a 09

(45) (56) (106) (06)

TangibilityQint 05 177 436c 64a

(01) (15) (18) (31)

Constant 167 96 656b 235 4613a 8610a 113 570a

(08) (04) (26) (04) (61) (58) (11) (31)

Year FE Yes Yes Yes Yes Yes Yes Yes Yes

Firm FE Yes Yes Yes Yes Yes Yes Yes Yes

R2 0002 0002 0003 0004 0013 0014 0013 0014

N 155587 155037 155587 155037 155587 155037 155587 155037

1438 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

7 Are Financial Market Conditions Important

According to the trade-off theory capital structure is determined by con-sidering the impact of costs and benefits of debt The time variation inthe costs and benefits of debt imply that issuance decisions would alsovary over time Furthermore market conditions affect the transactioncosts of adjusting Adverse selection is a greater problem in a cold marketthan in a hot one Accordingly it may be easier to adjust in good marketconditions Related ideas have been developed by Huang and Ritter(2009) Faulkender et al (2012) and Halling Yu and Zechner (2012)Faulkender Flannery Hankins and Smithrsquos paper examines cross-sectionalvariation in adjustment speeds including market valuations andhow they may affect adjustment speeds It is natural to think that ingood market conditions it will be less costly to issue both debt and equityand adjusting leverage to respond to profitability shocks would be mucheasierTo test the importance of market conditions we require a definition of

good times and bad times Our empirical strategy is to estimate good timesversus bad times at the four-digit industry level We define an industry ashaving ldquogood timesrdquo if the median firm in that industry has a market-to-book ratio that is higher than the 67th percentile of the time-series distribu-tion of the industry median market-to-book ratios Conversely an industryis defined as having ldquobad timesrdquo if the median firm in the industry has amarket-to-book ratio that is lower than the 33rd percentile of the time-seriesdistribution of the industry median market-to-book ratiosPanel A of Table IX tabulates issuance activity for profitability sorts in

both good times and bad times As might be expected active debt and equityissuances are larger during good times There is also somewhat more activeswaps between debt and equity during good timesPanel B reports the issuance activity for the smallest and the largest firms

during the good times and bad times Within each size quintile firms aresorted on profitability This is a three-way sort As before we find that thereis much more active use of external markets during good times Small low-profit firms are more likely to issue equity in good times than in bad timesDebt issuances are significantly higher in good times Large high-profitfirms are significantly more likely to issue debt and repurchase equity ingood times than in bad timesPanel C reports the magnitudes of the financial variables rather than the

frequencies During bad times less profitable large firms retire substantialamounts of debt and they show a tendency to issue equity Small firms donot seem to engage in similar debt reduction activities However like the

PROFITSndashLEVERAGE PUZZLEREVISITED 1439

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

Table IX Debt and equity issues in good and bad times

The table reports the frequency and magnitude of financing activity for sorts on profits for subsamples

of firms in good and bad times An industry is defined as having ldquogood timesrdquo if the median firm in

that industry has a market-to-book ratio that is higher than the 67th percentile of the time-series

distribution of industry median market-to-book ratios Conversely an industry is defined as having

bad times if the median firm in that industry has a market-to-book ratio that is lower than the 33rd

percentile of the time-series distribution of industry median market-to-book ratios Panels A and B

report the percentage of firms (i) issuing net debt in excess of 5 of the value of their assets (ii) issuing

net equity in excess of 5 of the value of their assets (iii) issuing debt and repurchasing equity both in

excess of 5 of the value of their assets and (iv) issuing equity and retiring debt both in excess of 5

of the value of their assets Panel C tabulates the mean debt and equity levels and the mean changes in

debt and equity issuances The sample contains nonfinancial firms listed on the annual Compustat files

for the period from 1971 to 2009

Iss Iss Iss D Iss ENet D Net E Rep E Ret D(1) (2) (3) (4)

Panel A sorts on profitability

Badtimes

Low 145 131 01 202 170 49 02 083 173 46 04 074 182 47 03 10

High 177 51 06 10t Highfrac14Low 41 136 33 01

Goodtimes

Low 206 381 01 462 197 148 02 233 212 103 04 184 234 92 10 18

High 216 92 16 14t Highfrac14Low 23 658 138 23

Panel B sorts on firm size and profitability

Badtimes

Smallfirm

s Low 151 318 00 312 182 276 02 253 151 140 02 294 183 73 00 19

High 177 80 00 11t Highfrac14Low 10 90 NA 20

Largefirm

s Low 172 54 01 042 148 39 04 053 150 43 04 044 156 33 03 06

High 147 28 05 03t Highfrac14Low 17 32 20 02

Goodtimes Smallfirm

s Low 213 448 01 542 230 443 01 473 222 325 00 474 203 197 02 38

High 190 153 02 31t Highfrac14Low 19 230 10 39

Largefirm

s Low 203 96 03 122 208 66 09 133 232 54 16 064 215 39 19 05

High 234 36 38 04t Highfrac14Low 29 99 92 38

(continued)

1440 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

larger firms there is a tendency for small firms to issue more equitymdashat leastto the extent possibleDuring good times large profitable firms raise significant amounts of debt

as they experience an increase in the value of their equity Such firms alsoengage in share repurchases Debt issues by smaller firms are much lessaffected by their own profitability during good times and their change indebt is more or less independent of their profits However equity issuanceis primarily found among the low-profit small firms during good timesTable IX shows that empirically profitability has a time-varying impacton leverage choicesIn Table X the basic regressions for changes in debt and for equity

issuance are presented for both good times and bad times Comparinggood times with bad times we see that the effects of profits are muchstronger in good times In bad times the impacts are rather weak statisticallyThis difference is interesting and deserves further study However for ourpurposes recall that the static trade-off theory implies time-varying capital

Table IX Continued

D D BVE BVE MVE MVE EqyIss Assets

(1) (2) (3) (4) (5) (6) (7) (8)

Panel C magnitude of financing activity

Badtimes Smallfirm

s Low 40 03 58 14 305 06 22 140

2 44 05 82 07 240 67 17 180

3 45 04 94 02 176 18 05 187

4 43 09 98 02 137 04 03 188

High 36 10 114 13 209 16 04 199

Largefirm

s Low 3365 73 2523 169 2304 163 94 8972

2 3031 95 2983 35 3284 278 56 9068

3 2930 5 3360 14 3947 308 57 9225

4 2281 6 3817 19 5213 442 7 8902

High 1790 62 4767 13 8799 58 31 9573

Goodtimes

Smallfirm

s Low 24 06 38 05 339 23 39 90

2 26 08 48 01 301 04 25 104

3 30 09 60 04 246 03 17 125

4 28 07 66 08 206 18 12 130

High 23 06 85 16 261 13 08 141

Largefirm

s Low 4532 36 3477 103 6557 260 53 12465

2 4580 58 3527 139 6911 421 23 12522

3 2952 111 3485 2 7586 279 32 9535

4 2752 124 3939 72 10686 33 105 10226

High 2429 228 4819 377 18565 268 222 10830

PROFITSndashLEVERAGE PUZZLEREVISITED 1441

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

structure choices even when the target is time invariant Empirically we doobserve time-varying choices

8 Is Scaling Important

In the preceding section we provided results for unscaled debt and equityissues This seems appropriate to us However it involves making twochanges to the standard regression We would like to determine if both

Table X Debt and equity changes in good and bad timesmdashregressions

The sample contains nonfinancial firms listed on the annual Compustat files for the period

from 1971 to 2009 The table presents estimates from regressions of change in debt (D)and net equity issuances (EqyIss) in good and bad times The explanatory variables includechange in profits lagged change in profits and lagged change in assets The regressions

include both firm and year fixed effects An industry is defined as having ldquogood timesrdquo ifthe median firm in that industry has a market-to-book ratio that is higher than the 67thpercentile of the time-series distribution of industry median market-to-book ratios

Conversely an industry is defined as having bad times if the median firm in thatindustry has a market-to-book ratio that is lower than the 33rd percentile of the time-series distribution of industry median market-to-book ratios The reported t-statistics arecorrected for clustering at the firm level

aMean significant at the 1 level

bMean significant at the 5 level

cMean significant at the 10 level

Good times Bad times

D EqyIss D EqyIss

(1) (2) (3) (4) (5) (6) (7) (8)

Profits 087a 091a 006b 009a 060 065 003 003b

(30) (29) (20) (28) (15) (16) (16) (20)

Profitst1 029c 015a 034c 001

(19) (37) (17) (04)

Assetst1 000 002b 000 001

(02) (22) (00) (07)

Constant 1337 3989 8691a 10001a 1351 1851 4099a 4055a

(03) (09) (55) (37) (09) (13) (80) (78)

Year FE Yes Yes Yes Yes Yes Yes Yes Yes

Firm FE Yes Yes Yes Yes Yes Yes Yes Yes

R2 0071 0080 0018 0045 0178 0212 0034 0041

N 74827 68729 74891 68787 26965 26013 26971 26018

1442 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

changes are crucial Accordingly in this section we take a step back toconsider prior literature in which variables are scaled The questions arehow does scaling affect the results and whyTable XI examines alternative scalings Columns (1) and (2) scale debt and

equity issuances by the total firm net issuance Columns (3) and (4) scaledebt and equity issuances by the sum of debt and book equity Columns (5)and (6) consider the change in debt scaled by the sum of debt and bookequity Columns (7) and (8) consider the change in leverage due to externalfinancing Since we are examining changes we use the changes in thestandard regression factors as controls The primary interest is in the coef-ficient on the change in profitsWe find that a change in profits is associated with an increase in debt

issues But the sign reverses when we scale the debt issues by the sum of debtplus book equity This is telling What is really driving the results is thechange in equity in the denominator not the change in debt in the numer-ator This result is further substantiated in Columns (5) and (6) where thechange in debt to capital ratio is used instead of debt issuanceTo further explore the impact of active issuance relative to passive

experiencing of changes we decompose leverage by separating out theeffect of debt and equity issues from those due to changes in retainedearnings Thus we define leverage change due to external financing as

LevEF frac14Dt1 thorn d

Dt1 thorn Et1 thorn dthorn e

Dt1

Dt1 thorn Et1 eth1THORN

where D is debt E is book value of equity d is debt issues net of retirementsand e is equity issues net of repurchases Garvey and Hanka (1999) use asimilar measure in their study This measure directly captures the netleverage effect of debt and equity issuancesColumns (7) and (8) report results from regressions of the net leverage

effect of debt and equity issues on changes in profitability after controllingfor other leverage factors in differences and year indicator variables Theeffect of profitability on leverage change due to external financing is positiveand statistically significant This comparison suggests that changes in totaldebt and equity are only partly a result of debt and equity issuance decisionsOther balance sheet adjustments complicate the inferences from leverageratio regressionsFor most of our analysis we follow the literature in treating debt as homo-

genous It is of course potentially interesting to ask what happens when wedistinguish between short-term debt and long-term debt Which part of thefirmrsquos debt maturity structure drives the under-adjustment Table XII

PROFITSndashLEVERAGE PUZZLEREVISITED 1443

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

provides results from such decompositions Alternative scalings areconsidered The main results of interest are that the distinction betweenshort- and long-term debt is important Short-term debt issuance is largelyindependent of changes in profits The negative sign on profit change stemsfrom the action of long-term debt This seems intuitively reasonable

Table XI Scaled issuances

The sample contains nonfinancial firms listed on the annual Compustat files for the period

from 1971 to 2009 The table presents estimates from regressions of debt issuance to totalissuance (in columns (1) and (2)) debt issuance to total capitalization (in columns (3) and(4)) change in debt to capitalization ratio (in columns (5) and (6)) and change in leverage

ratio considering only external financing (in columns (7) and (8)) The change in leverageratio considering only external financing is defined as

LevEF frac14Dt1 thorn d

Dt1 thorn Et1 thorn dthorn e

Dt1

Dt1 thorn Et1

where D is debt E is book value of equity d is debt issues net of retirements and e is equity

issues net of repurchases The regressions include lagged leverage factors and year indicatorvariables Some of the specifications include firm fixed effects The reported t-statistics arecorrected for clustering at the firm level

aMean significant at the 1 level

bMean significant at the 5 level

cMean significant at the 10 level

DebtIssTotIss

DebtIssethDthornETHORN D

DthornE LevEF

(1) (2) (3) (4) (5) (6) (7) (8)

Profitt1 004a 000 002a 001c 003a 002 001b 002a

(35) (03) (34) (18) (35) (14) (23) (26)

IndLevt1 004c 001 001 001c 000 000 000 000

(19) (06) (15) (19) (02) (04) (03) (05)

Market=bookt1 001a 001a 000a 000a 000a 000a 000 000c

(47) (85) (68) (60) (55) (36) (05) (18)

Tangt1 004 005 013a 011a 011a 006a 003a 003a

(12) (15) (144) (112) (79) (38) (44) (42)

Assetst1 008a 004a 006a 004a 002a 001a 003a 002a

(97) (45) (283) (161) (53) (35) (147) (101)

Constant 079a 074a 002a 003a 000 000 000b 002a

(497) (442) (91) (111) (00) (05) (22) (93)

Year FE Yes Yes Yes Yes Yes Yes Yes Yes

Firm FE No Yes No Yes No Yes No Yes

R2 0015 0098 0022 0076 0009 0070 0012 0086

N 133905 133905 140552 140552 138229 138229 140572 140572

1444 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

inasmuch as short-term debt is probably more associated with day-to-dayoperations rather than longer term planning considerationsIn Table XIII we examine change in leverage ( D

DthornE) and change inleverage due to external financing (LEVEF) during good and bad timesseparately The traditional negative effect of a change in profits on achange in leverage is found both in good times and in bad timesHowever when we decompose the impact in order to isolate the externalcomponent as before the result disappears In good times the externalfinancing component of a change in profits is weak but still positiveDuring bad times it is not statistically significant This implies that distin-guishing between good and bad times does not fundamentally alter the basic

Table XII Short- and long-term debt issuances

The sample contains nonfinancial firms listed on the annual Compustat files for the period

from 1971 to 2009 The table presents estimates from regressions of short-term debtissuance to total issuance (in columns (1) and (2)) long-term debt issuance to totalissuance (in columns (3) and (4)) short-term debt issuance to total capitalization (in

columns (5) and (6)) and long-term debt to total capitalization (in columns (7) and (8))Some of the specifications include firm fixed effects The reported t-statistics are correctedfor clustering at the firm level

aMean significant at the 1 level

bMean significant at the 5 level

cMean significant at the 10 level

STDebtIssTotIss

LTDebtIssTotIss

STDebtIssethDthornETHORN

LTDebtIssethDthornETHORN

(1) (2) (3) (4) (5) (6) (7) (8)

Profitabilityt1 0002 0003 0042a 0008 0001 0001 0019a 0012a

(02) (04) (30) (05) (14) (10) (45) (28)

IndLevt1 0009 0007 0070b 0033 0001 0002c 0014a 0005

(06) (05) (26) (11) (08) (17) (26) (09)

M=Bt1 0003a 0003a 0003b 0010a 0000 0000 0003a 0002a

(48) (35) (22) (65) (10) (01) (84) (42)

Tangibilityt1 0020 0007 0011 0067 0005a 0006a 0097a 0081a

(12) (03) (03) (16) (34) (33) (108) (89)

Assetst1 0039a 0011b 0034a 0024b 0001a 0001 0046a 0020a

(89) (22) (38) (23) (36) (13) (190) (82)

Constant 0001b 0153a 0798a 0591a 0000c 0006a 0057a 0096a

(25) (144) (470) (288) (17) (124) (292) (234)

Year FE Yes Yes Yes Yes Yes Yes Yes Yes

Firm FE No Yes No Yes No Yes No Yes

R2 0025 0089 0018 0073 0030 0147 0016 0298

N 136146 136146 136140 136140 143031 143031 142847 142847

PROFITSndashLEVERAGE PUZZLEREVISITED 1445

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

Table XIII Externally financed leverage changes in good and bad times

The sample contains nonfinancial firms listed on the annual Compustat files for the period

from 1971 to 2009 The table presents estimates from regressions of change in debt tocapitalization ratio (in columns (1) (2) (5) and (6)) and change in leverage ratio consider-ing only external financing (in columns (3) (4) (7) and (8)) The change in leverage ratio

considering only external financing is defined as

LevEF frac14Dt1 thorn d

Dt1 thorn Et1 thorn dthorn e

Dt1

Dt1 thorn Et1

where D is debt E is book value of equity d is debt issues net of retirements and e is equityissues net of repurchases The regressions include lagged leverage factors and year indicator

variables The reported t-statistics are corrected for clustering at the firm level An industryis defined as having ldquogood timesrdquo if the median firm in that industry has a market-to-bookratio that is higher than the 67th percentile of the time-series distribution of industry

median market-to-book ratios Conversely an industry is defined as having ldquobad timesrdquoif the median firm in that industry has a market-to-book ratio that is lower than the 33rdpercentile of the time-series distribution of industry median market-to-book ratios

aMean significant at the 1 level

bMean significant at the 5 level

cMean significant at the 10 level

Good times Bad times

DDthornE LevEF D

DthornE LevEF

(1) (2) (3) (4) (5) (6) (7) (8)

Profitabiltyt1 0022c 0005 0016c 0025b 0073b 0041 0007 0007

(17) (03) (18) (24) (25) (11) (04) (03)

IndLevt1 0000 0011 0001 0009 0010 0018 0001 0000

(00) (09) (01) (13) (07) (10) (02) (00)

MB

t1

0004a 0003b 0000 0001 0008a 0001 0000 0000

(40) (24) (05) (08) (28) (04) (03) (03)

Tangt1 0097a 0041 0033a 0038b 0127a 0048 0059a 0033

(45) (16) (27) (25) (49) (14) (43) (16)

Assetst1 0014a 0013b 0024a 0019a 0037a 0023c 0036a 0025a

(30) (20) (84) (48) (45) (17) (88) (43)

Constant 0030b 0036 0000 0003 0028c 0014 0021 0001

(20) (16) (00) (07) (17) (06) (11) (01)

Year FE Yes Yes Yes Yes Yes Yes Yes Yes

Firm FE No Yes No Yes No Yes No Yes

R2 0006 0088 0009 0117 0026 0251 0023 0202

N 56820 56820 58200 58200 23462 23462 23511 23511

1446 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

conclusions The traditional leverage result on profits is driven by increasesin equity that are experienced and then partially but not completely offsetby firmsTo the extent that firms are engaging in active financing they are following

the prescriptions of the trade-off theory Thus we have a coherent picture ofhow firms are responding to profits Basically more profitable firms experiencean increase in the value of their equity They respond to this by issuing debtand repurchasing equity But the magnitude of the increase in equity is largerthan the magnitudes of the active steps taken in the debt and equity marketsThis means that the firmsrsquo actual actions are as predicted by the trade-offtheory But it leaves open the question of why they do not act more stronglyThe capital structure literature has long been interested in dynamics and

capital structure adjustment paths In recent years this interest has increas-ingly resulted in explicit dynamic optimizing models such as those ofGoldstein Ju and Leland (2001) Hennessy and Whited (2005) andStrebulaev (2007) These models commonly make specialized assumptionsleading some to question their generalityWe argue that both fixed transactions costs and variable transactions costs

are plausible and realistic in models that pin down firm size The simultaneouspresence of both is important Due to fixed costs firms only rebalance inter-mittently Due to variable adjustment costs the extent of the adjustment isreduced to the point at which an extra unit to adjustment is balanced againstthe extra unit transactions costs This means that both leverage increases andleverage reductions will only go part way toward a static optimumThe key point is that with both fixed and variable transactions costs we

should expect to only see intermittent leverage adjustments When leverage isadjusted it should ldquounder adjustrdquo when compared to a frictionless modelFurther discussion of transactions costs in a capital structure setting isprovided by Leary and Roberts (2005)

9 Conclusion

The connection between corporate profits and capital structure has beenvery influential in the assessment of the trade-off theory The conventionalinterpretation of the evidence has pushed the literature away from the trade-off theory and toward much more complex dynamic models and ideas As aresult it is important to make sure that the evidence has been correctlyinterpreted We show that the problem is with the interpretation of thetest results and a failure to take into account the costs associated with re-sponding to a profit shock

PROFITSndashLEVERAGE PUZZLEREVISITED 1447

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

We document that more profitable firms really do borrow more (not less)They tend to repurchase equity They experience an increase in both thebook value of equity and the market value of equity Less profitable firmsreally do tend to reduce their debt and to issue equityFirm size and market conditions also matter Larger firms tend to be more

active in the debt markets while smaller firms tend to be relatively moreactive in the equity markets During good times there is more use ofexternal financingThe usual profitsndashleverage puzzle result is primarily driven by the increase

in equity that is experienced by the more profitable firms Accordingly thepuzzle should be restated as asking why do firms not take sufficiently largeoffsetting actions to fully undo the change in equity What limits the mag-nitudes of the typical leverage response to profit shocksIn a frictionless model the partial response appears puzzling But there is

good empirical reason to believe that rebalancing entails both fixed andvariable costs and that firm size matters The rebalancing costs can befully avoided by doing nothing Accordingly the firm must decide whetherany given shock is big enough to be worth responding to If it is then thefirm must decide how big a response is called for These technical conditionsare known as ldquovalue matchingrdquo and ldquosmooth pastingrdquo Thus optimizationimplies that some shocks will be ignored Even if the shock is not ignoredthe optimal response will only partially undo the shock The magnitude ofthe leverage response must balance the marginal cost and the marginalbenefit of an extra unit of leverage Since the marginal cost of adjustingleverage is strictly positive the adjustment toward that static leverageoptimum will only go part way This is true both for leverage increasesand for leverage reductionsOne might conjecture that the variable costs would be time varying and

that such costs might have declined over time Since we do not have a fullaccounting of all of the variable costs it is of course impossible to be sureHowever there is some evidence suggesting that at least the directunderwriting fees are remarkably stable over time10

Surely nobody ought to be surprised that more profitable firms are moreinclined to repurchase equity while unprofitable firms tend not to do so Thefact that more profitable firms typically experience an increase in the value of

10 For IPOs Jay Ritter provides important summary evidence as httpbearwarringtonufleduritterIPOs2013Underwritingpdf He finds that for medium-sized deals a remark-

ably large fraction of the deals have a 7 fee year after year Whether this is due tobehavioral factors or some sort of a coordination device is not clear The extent towhich similar intertemporal stability carries over to other forms of external financemerits further study

1448 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

equity is equally natural Were it not for the contrary claims in the literature(ie the profitsndashleverage puzzle) we would argue that nobody ought to findit surprising that more profitable firms are more inclined to issue debt Thefact that leverage adjustments typically only partially offset profit shocksshould also sound quite natural These facts all fit together easily within thetrade-off theory once we take account of both fixed and variable costs ofactively adjusting capital structure

Appendix

Table AI Market leverage regressions

The table presents estimates of the regressions of market leverage on firm characteristics

The sample comes from the annual Compustat files during the period 1971ndash2009 Financialfirms are excluded The leverage is estimated as the ratio of debt over debt plus marketequity The explanatory variables Profitabilityt1

MB

t1

Tangibilityt1 and LnethAssetsTHORNt1are described in Table 1 IndMedianLevt1 is estimated as the median market leverage of allother firms in the same industry The industry is defined at the level of the four-digit SICcode All specifications include the year fixed effects The specifications in column (2)

additionally include the firm fixed effects We report t-statistics where the standard errorsare clustered at the firm level in parentheses

aSignificant at the 1 level

Quantile regressions

(1) (2) (3) (4) (5)

Profitabilityt1 013a 013a 007a 011a 020a

(275) (262) (349) (425) (538)

IndLevt1 046a 020a 032a 058a 063a

(506) (239) (1299) (1630) (1042)MB

t1

003a 001a 001a 002a 004a

(509) (303) (743) (709) (534)

Tangibilityt1 013a 018a 013a 014a 010a

(191) (181) (762) (571) (232)

LnethAssetsTHORNt1 001a 005a 001a 001a lt001a

(112) (258) (682) (342) (29)

Constant 010a 011a 005a 005a 029a

(171) (98) (173) (118) (398)

Year FE Yes Yes Yes Yes Yes

Firm FE No Yes No No No

Regression type OLS FE 25th tile Median 75thtile

Clustered SE Yes Yes No No No

R2 030 067

Pseudo R2 016 023 0201

N 158578 158578 158578 158578 158578

PROFITSndashLEVERAGE PUZZLEREVISITED 1449

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

Table AII Profitability sorts for firms issuing large amounts

Percentage of firms

Iss Ret Iss Iss Rep Iss Iss IssD IssE

D D NetD E E NetE DampE RepE RetD

(1) (2) (3) (4) (5) (6) (7) (8) (9)

Panel A 10 cutoff to define issuing or retiring activity

Allfirm

s

Low 205 150 134 260 05 257 45 00 23

2 207 175 103 77 06 75 11 01 10

3 243 196 101 50 08 47 08 01 09

4 252 186 113 50 13 48 08 02 08

High 227 161 118 60 27 57 07 04 09

t Highfrac14Low 70 38 62 755 226 768 320 90 152

Smallfirm

s Low 230 132 162 380 05 375 82 00 32

2 228 143 165 334 04 333 65 00 26

3 211 146 147 216 03 214 38 00 23

4 190 154 120 126 04 124 15 01 20

High 194 162 114 99 09 97 12 01 15

t Highfrac14Low 42 40 68 342 20 340 169 07 131

Largefirm

s Low 231 200 82 31 06 29 08 01 04

2 231 187 67 17 06 14 03 01 02

3 234 177 77 17 09 14 03 02 03

4 218 158 75 16 17 14 02 03 02

High 218 135 99 19 41 15 03 07 02

t Highfrac14Low 19 106 36 46 135 57 39 58 23

Percentage of firms

Iss Ret Iss Iss Rep Iss Both Iss D Iss E

D D Net D E E Net E DampE Rep E Ret D

(1) (2) (3) (4) (5) (6) (7) (8) (9)

Panel B 15 cutoff to define issuing or retiring activity

Allfirm

s

Low 158 101 100 230 01 228 27 00 15

2 142 115 63 57 01 56 05 00 06

3 165 128 57 32 01 31 03 00 05

4 173 118 65 35 01 34 03 00 05

High 160 102 70 46 04 44 03 00 05

t Highfrac14Low 08 06 141 738 87 746 277 24 131

Smallfirm

s Low 195 95 132 354 01 350 56 00 24

2 185 96 131 303 01 301 44 00 18

3 160 96 104 179 00 176 18 00 15

4 134 99 78 99 01 98 08 00 12

High 133 95 73 82 01 80 06 00 10

t Highfrac14Low 80 00 95 345 12 344 149 12 52

Largefirm

s Low 145 123 42 14 01 12 03 00 01

2 137 114 32 07 01 06 01 00 01

3 142 104 35 06 01 05 00 00 01

4 130 91 36 08 03 07 01 01 01

High 134 81 48 09 06 07 01 01 01

t Highfrac14Low 19 85 17 26 54 33 34 17 11

1450 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

Table AIV Debt and equity changesmdashwithout fixed effects

The sample contains nonfinancial firms listed on the annual Compustat files for the period from 1971

to 2009 The table presents estimates from regressions of change in debt (D) change in the book

value of equity (BVE) change in the market value of equity (MVE) and net equity issuances

(EqyIss) The explanatory variables include current and lagged change in operating income before

depreciation (Profits and Profitst1) and lagged change in book value of assets (Assetst1) The

regressions include year indicator variables The reported t-statistics are corrected for clustering at the

firm level

aMean significant at the 1 level

bMean significant at the 5 level

cMean significant at the 10 level

D BVE MVE EqyIss

(1) (2) (3) (4) (5) (6) (7) (8)

Profitst1 025a 024a 023a 038a 007 066a 007a 009a

(40) (32) (36) (39) (06) (29) (42) (51)Profits 051b 082a 149a 004a

(25) (33) (49) (33)Assetst1 002 005c 024a 001b

(08) (17) (34) (23)Constant 93 176a 45 71 1552a 1704a 130a 140a

(12) (28) (07) (09) (49) (49) (45) (45)Year FE Yes Yes Yes Yes Yes Yes Yes YesR2 001 005 001 011 001 005 002 002N 162056 162056 162130 162130 157550 157550 162154 162154

Table AIII Median profitability sorts

The sample contains nonfinancial firms listed on the annual Compustat files for the period from 1971

to 2009 The table presents median debt and equity median changes in debt and equity and median

equity issuances (EqyIss) in millions of dollars for all firms sorted on profitability and for profitability

sorts within the smallest and largest firms The sorts are done annually

D D BVE BVE MVE MVE EqyIss Assets(1) (2) (3) (4) (5) (6) (7) (8)

Allfirm

s

Low 21 00 75 17 256 20 01 1922 186 02 449 18 598 19 00 11033 816 06 1274 03 1637 05 00 32244 787 02 1541 18 2488 03 00 3556High 273 00 1361 66 3300 14 00 2578

Smallfirm

s Low 05 00 13 07 140 14 06 402 06 00 25 08 114 12 01 563 09 00 32 05 85 07 00 674 12 00 39 01 68 03 00 79High 10 00 51 04 98 01 00 90

Largefirm

s Low 11485 381 10269 371 13227 202 03 333902 13051 393 12786 140 17571 169 05 392463 11554 287 13429 24 21623 305 00 370314 9963 209 15052 194 29424 308 00 36705High 7361 00 16380 613 49142 439 00 34976

PROFITSndashLEVERAGE PUZZLEREVISITED 1451

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

References

Auerbach A J (1985) Real determinants of corporate leverage in B M Friedman (ed)

Corporate Capital Structures in the United States University of Chicago Press Chicago

IllinoisBertrand M and Mullainathan S (2003) Enjoying the quiet life corporate governance and

managerial preferences Journal of Political Economy 111 1043ndash1075Booth L Varouj A Demirguc-Kunt A and Maksimovic V (2001) Capital structures in

developing countries Journal of Finance 56 87ndash130Brennan M J and Schwartz E S (1984) Optimal financial policy and firm valuation

Journal of Finance 39 593ndash607Cameron A C and Trivedi P K (2010) Microeconometrics using Stata Stata Press TexasConstantinides G M (1979) Multiperiod consumption and investment behavior with

convex transactions costs Management Science 25 1127ndash1137Constantinides G M and Richard S F (1978) Existence of optimal simple policies for

discounted-cost inventory and cash management in continuous time Operations Research

26 620ndash636DeAngelo H DeAngelo L and Whited T M (2011) Capital structure dynamics and

transitory debt Journal of Financial Economics 99 235ndash261Dixit A K (1993) The Art of Smooth Pasting Routledge London and New YorkFama E and French K R (2002) Testing trade-off and pecking order predictions about

dividends and debt Review of Financial Studies 15 1ndash33Faulkender M and Petersen M A (2006) Does the source of capital affect capital struc-

ture Review of Financial Studies 19 45ndash79Faulkender M Flannery M J Hankins K W and Smith J M (2012) Cash Flows and

Leverage Adjustments Journal of Financial Economics 103 632ndash646Fischer E Heinkel R and Zechner J (1989) Dynamic capital structure choice theory

and tests Journal of Finance 44 19ndash40Frank M Z and Goyal V K (2008) Trade-off and pecking order theories of

debt in B E Eckbo (ed) Handbook of Corporate Finance Empirical Corporate Finance

Vol 2 Handbook of Finance Series Chapter 12 ElsevierNorth-Holland Amsterdam pp

135ndash202

Frank M Z and Goyal V K (2009) Capital structure decisions which factors are reliably

important Financial Management 38 1ndash37

Garvey G T and Hanka G (1999) Capital structure and corporate control the effect of

antitakeover statutes on firm leverage Journal of Finance 54 519ndash546

Goldstein R Ju N and Leland H (2001) An EBIT-based model of dynamic capital

structure Journal of Business 74 483ndash512

Graham J R and Tucker A L (2006) Tax shelters and corporate debt policy Journal of

Financial Economics 81 563ndash594

Hackbarth D and Mauer D C (2012) Optimal priority structure capital structure and

investment Review of Financial Studies 25 747ndash796

Halling M Yu J and Zechner J (2012) Leverage dynamics over the business cycle

Working paper University of Utah University of New South Wales and Vienna

University of Economics and Business AdministrationHennessy C A and Whited T A (2005) Debt dynamics Journal of Finance 60 1129ndash1165

Huang R and Ritter J R (2009) Testing theories of capital structure and estimating the

speed of adjustment Journal of Financial and Qualitative Analysis 44 237ndash271

1452 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

Hovakimian A Opler T and Titman S (2001) The debt-equity choice Journal of

Financial and Quantitative Analysis 36 1ndash24Jung K Kim Y and Stulz R M (1996) Timing investment opportunities managerial

discretion and the security issue decisions Journal of Financial Economics 42 159ndash185Kane A Marcus A J and McDonald R L (1984) How big is the tax advantage to

debt Journal of Finance 39 841ndash853Kayhan A and Titman S (2007) Firmsrsquo histories and their capital structures Journal of

Financial Economics 83 1ndash32Leary M T (2009) Bank loan supply lender choice and corporate capital structure

Journal of Finance 64 1143ndash1185Leary M T and Roberts M R (2005) Do firms rebalance their capital structures

Journal of Finance 60 2575ndash2619Leland H E (1994) Corporate debt value bond covenants and optimal capital structure

Journal of finance 49 1213ndash1252Lemmon M L Roberts M R and Zender J F (2008) Back to the beginning persist-

ence and the cross-section of corporate capital structure Journal of Finance 63 1575ndash1608Long M S and Malitz I B (1985) The investment-financing nexus some empirical

evidence Midland Corporate Finance Journal 3 53ndash59Mackie-Mason J K (1990) Do taxes affect corporate financing decisions Journal of

Finance 45 1471ndash1493Morellec E Nikolov B and Schurhoff N (2012) Corporate governance and capital

structure dynamics Journal of Finance 67 803ndash848Myers S C (1984) The capital structure puzzle Journal of Finance 39 575ndash592Myers S C (1993) Still searching for optimal capital structure Journal of Applied

Corporate Finance 61 4ndash14Rajan R and Zingales L (1995) What do we know about capital structure some evidence

from international data Journal of Finance 50 1421ndash1460Strebulaev I A (2007) Do tests of capital structure theory mean what they say Journal of

Finance 62 1747ndash1787

Strebulaev I A and Whited T M (2012) Dynamic Models and Structural Estimation in

Corporate Finance Now Publisher Boston and Delft

Titman S and Wessels R (1988) The determinants of capital structure choice Journal of

Finance 43 1ndash21

Welch I (2004) Capital structure and stock returns Journal of Political Economy 112

106ndash131

Welch I (2011) Two common problems in capital structure research the financial-debt-

to-asset ratio and issuing activity versus leverage changes International Review of Finance

11 1ndash17

PROFITSndashLEVERAGE PUZZLEREVISITED 1453

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

the results are qualitatively similar to the OLS results reported in Table VIIIProfitability positively affects debt issuances and negatively affects equityissuances across various quantiles Overall the results in Table VIII showthat there is nothing that is special about using sorts or regressions to explaindebt and equity issues In either case we find that more profitable firms tendto increase their debt experience an increase in the value of equity andrepurchase shares Thus the control factors are not responsible for theusual rejection of the trade-off theory It would appear that what mattersis the size of the issuing activity relative to the organic increases in equityvalue due to profits In other words what matters in the leverage regressionseems to be coming from the presence of E in D

DthornE and not from D

Table VIII Debt and equity issuances profitability and leverage factor quintiles

The sample contains nonfinancial firms listed on the annual Compustat files for the period

from 1971 to 2009 The table presents estimates from regressions of change in debt (D)change in the book value of equity (BVE) change in the market value of equity (MVE)and net equity issuances (EquityIss) The explanatory variables include lagged profitability

quintiles and lagged leverage factor quintiles In addition to fixed firm effects the regres-sions include year indicator variables The reported t-statistics are corrected for clustering atthe firm level

aMean significant at the 1 level

bMean significant at the 5 level

cMean significant at the 10 level

D BVE MVE EqyIss

(1) (2) (3) (4) (5) (6) (7) (8)

ProfitabilityQint 214a 167a 283a 175a 80 429a 53a 55a

(70) (54) (66) (59) (08) (48) (55) (62)

IndLevQint 104a 43 99 20

(29) (15) (10) (10)

Ln(Assets)Qint52 17 1026a 53a

(16) (05) (77) (35)MB

Qint158a 372a 1155a 09

(45) (56) (106) (06)

TangibilityQint 05 177 436c 64a

(01) (15) (18) (31)

Constant 167 96 656b 235 4613a 8610a 113 570a

(08) (04) (26) (04) (61) (58) (11) (31)

Year FE Yes Yes Yes Yes Yes Yes Yes Yes

Firm FE Yes Yes Yes Yes Yes Yes Yes Yes

R2 0002 0002 0003 0004 0013 0014 0013 0014

N 155587 155037 155587 155037 155587 155037 155587 155037

1438 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

7 Are Financial Market Conditions Important

According to the trade-off theory capital structure is determined by con-sidering the impact of costs and benefits of debt The time variation inthe costs and benefits of debt imply that issuance decisions would alsovary over time Furthermore market conditions affect the transactioncosts of adjusting Adverse selection is a greater problem in a cold marketthan in a hot one Accordingly it may be easier to adjust in good marketconditions Related ideas have been developed by Huang and Ritter(2009) Faulkender et al (2012) and Halling Yu and Zechner (2012)Faulkender Flannery Hankins and Smithrsquos paper examines cross-sectionalvariation in adjustment speeds including market valuations andhow they may affect adjustment speeds It is natural to think that ingood market conditions it will be less costly to issue both debt and equityand adjusting leverage to respond to profitability shocks would be mucheasierTo test the importance of market conditions we require a definition of

good times and bad times Our empirical strategy is to estimate good timesversus bad times at the four-digit industry level We define an industry ashaving ldquogood timesrdquo if the median firm in that industry has a market-to-book ratio that is higher than the 67th percentile of the time-series distribu-tion of the industry median market-to-book ratios Conversely an industryis defined as having ldquobad timesrdquo if the median firm in the industry has amarket-to-book ratio that is lower than the 33rd percentile of the time-seriesdistribution of the industry median market-to-book ratiosPanel A of Table IX tabulates issuance activity for profitability sorts in

both good times and bad times As might be expected active debt and equityissuances are larger during good times There is also somewhat more activeswaps between debt and equity during good timesPanel B reports the issuance activity for the smallest and the largest firms

during the good times and bad times Within each size quintile firms aresorted on profitability This is a three-way sort As before we find that thereis much more active use of external markets during good times Small low-profit firms are more likely to issue equity in good times than in bad timesDebt issuances are significantly higher in good times Large high-profitfirms are significantly more likely to issue debt and repurchase equity ingood times than in bad timesPanel C reports the magnitudes of the financial variables rather than the

frequencies During bad times less profitable large firms retire substantialamounts of debt and they show a tendency to issue equity Small firms donot seem to engage in similar debt reduction activities However like the

PROFITSndashLEVERAGE PUZZLEREVISITED 1439

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

Table IX Debt and equity issues in good and bad times

The table reports the frequency and magnitude of financing activity for sorts on profits for subsamples

of firms in good and bad times An industry is defined as having ldquogood timesrdquo if the median firm in

that industry has a market-to-book ratio that is higher than the 67th percentile of the time-series

distribution of industry median market-to-book ratios Conversely an industry is defined as having

bad times if the median firm in that industry has a market-to-book ratio that is lower than the 33rd

percentile of the time-series distribution of industry median market-to-book ratios Panels A and B

report the percentage of firms (i) issuing net debt in excess of 5 of the value of their assets (ii) issuing

net equity in excess of 5 of the value of their assets (iii) issuing debt and repurchasing equity both in

excess of 5 of the value of their assets and (iv) issuing equity and retiring debt both in excess of 5

of the value of their assets Panel C tabulates the mean debt and equity levels and the mean changes in

debt and equity issuances The sample contains nonfinancial firms listed on the annual Compustat files

for the period from 1971 to 2009

Iss Iss Iss D Iss ENet D Net E Rep E Ret D(1) (2) (3) (4)

Panel A sorts on profitability

Badtimes

Low 145 131 01 202 170 49 02 083 173 46 04 074 182 47 03 10

High 177 51 06 10t Highfrac14Low 41 136 33 01

Goodtimes

Low 206 381 01 462 197 148 02 233 212 103 04 184 234 92 10 18

High 216 92 16 14t Highfrac14Low 23 658 138 23

Panel B sorts on firm size and profitability

Badtimes

Smallfirm

s Low 151 318 00 312 182 276 02 253 151 140 02 294 183 73 00 19

High 177 80 00 11t Highfrac14Low 10 90 NA 20

Largefirm

s Low 172 54 01 042 148 39 04 053 150 43 04 044 156 33 03 06

High 147 28 05 03t Highfrac14Low 17 32 20 02

Goodtimes Smallfirm

s Low 213 448 01 542 230 443 01 473 222 325 00 474 203 197 02 38

High 190 153 02 31t Highfrac14Low 19 230 10 39

Largefirm

s Low 203 96 03 122 208 66 09 133 232 54 16 064 215 39 19 05

High 234 36 38 04t Highfrac14Low 29 99 92 38

(continued)

1440 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

larger firms there is a tendency for small firms to issue more equitymdashat leastto the extent possibleDuring good times large profitable firms raise significant amounts of debt

as they experience an increase in the value of their equity Such firms alsoengage in share repurchases Debt issues by smaller firms are much lessaffected by their own profitability during good times and their change indebt is more or less independent of their profits However equity issuanceis primarily found among the low-profit small firms during good timesTable IX shows that empirically profitability has a time-varying impacton leverage choicesIn Table X the basic regressions for changes in debt and for equity

issuance are presented for both good times and bad times Comparinggood times with bad times we see that the effects of profits are muchstronger in good times In bad times the impacts are rather weak statisticallyThis difference is interesting and deserves further study However for ourpurposes recall that the static trade-off theory implies time-varying capital

Table IX Continued

D D BVE BVE MVE MVE EqyIss Assets

(1) (2) (3) (4) (5) (6) (7) (8)

Panel C magnitude of financing activity

Badtimes Smallfirm

s Low 40 03 58 14 305 06 22 140

2 44 05 82 07 240 67 17 180

3 45 04 94 02 176 18 05 187

4 43 09 98 02 137 04 03 188

High 36 10 114 13 209 16 04 199

Largefirm

s Low 3365 73 2523 169 2304 163 94 8972

2 3031 95 2983 35 3284 278 56 9068

3 2930 5 3360 14 3947 308 57 9225

4 2281 6 3817 19 5213 442 7 8902

High 1790 62 4767 13 8799 58 31 9573

Goodtimes

Smallfirm

s Low 24 06 38 05 339 23 39 90

2 26 08 48 01 301 04 25 104

3 30 09 60 04 246 03 17 125

4 28 07 66 08 206 18 12 130

High 23 06 85 16 261 13 08 141

Largefirm

s Low 4532 36 3477 103 6557 260 53 12465

2 4580 58 3527 139 6911 421 23 12522

3 2952 111 3485 2 7586 279 32 9535

4 2752 124 3939 72 10686 33 105 10226

High 2429 228 4819 377 18565 268 222 10830

PROFITSndashLEVERAGE PUZZLEREVISITED 1441

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

structure choices even when the target is time invariant Empirically we doobserve time-varying choices

8 Is Scaling Important

In the preceding section we provided results for unscaled debt and equityissues This seems appropriate to us However it involves making twochanges to the standard regression We would like to determine if both

Table X Debt and equity changes in good and bad timesmdashregressions

The sample contains nonfinancial firms listed on the annual Compustat files for the period

from 1971 to 2009 The table presents estimates from regressions of change in debt (D)and net equity issuances (EqyIss) in good and bad times The explanatory variables includechange in profits lagged change in profits and lagged change in assets The regressions

include both firm and year fixed effects An industry is defined as having ldquogood timesrdquo ifthe median firm in that industry has a market-to-book ratio that is higher than the 67thpercentile of the time-series distribution of industry median market-to-book ratios

Conversely an industry is defined as having bad times if the median firm in thatindustry has a market-to-book ratio that is lower than the 33rd percentile of the time-series distribution of industry median market-to-book ratios The reported t-statistics arecorrected for clustering at the firm level

aMean significant at the 1 level

bMean significant at the 5 level

cMean significant at the 10 level

Good times Bad times

D EqyIss D EqyIss

(1) (2) (3) (4) (5) (6) (7) (8)

Profits 087a 091a 006b 009a 060 065 003 003b

(30) (29) (20) (28) (15) (16) (16) (20)

Profitst1 029c 015a 034c 001

(19) (37) (17) (04)

Assetst1 000 002b 000 001

(02) (22) (00) (07)

Constant 1337 3989 8691a 10001a 1351 1851 4099a 4055a

(03) (09) (55) (37) (09) (13) (80) (78)

Year FE Yes Yes Yes Yes Yes Yes Yes Yes

Firm FE Yes Yes Yes Yes Yes Yes Yes Yes

R2 0071 0080 0018 0045 0178 0212 0034 0041

N 74827 68729 74891 68787 26965 26013 26971 26018

1442 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

changes are crucial Accordingly in this section we take a step back toconsider prior literature in which variables are scaled The questions arehow does scaling affect the results and whyTable XI examines alternative scalings Columns (1) and (2) scale debt and

equity issuances by the total firm net issuance Columns (3) and (4) scaledebt and equity issuances by the sum of debt and book equity Columns (5)and (6) consider the change in debt scaled by the sum of debt and bookequity Columns (7) and (8) consider the change in leverage due to externalfinancing Since we are examining changes we use the changes in thestandard regression factors as controls The primary interest is in the coef-ficient on the change in profitsWe find that a change in profits is associated with an increase in debt

issues But the sign reverses when we scale the debt issues by the sum of debtplus book equity This is telling What is really driving the results is thechange in equity in the denominator not the change in debt in the numer-ator This result is further substantiated in Columns (5) and (6) where thechange in debt to capital ratio is used instead of debt issuanceTo further explore the impact of active issuance relative to passive

experiencing of changes we decompose leverage by separating out theeffect of debt and equity issues from those due to changes in retainedearnings Thus we define leverage change due to external financing as

LevEF frac14Dt1 thorn d

Dt1 thorn Et1 thorn dthorn e

Dt1

Dt1 thorn Et1 eth1THORN

where D is debt E is book value of equity d is debt issues net of retirementsand e is equity issues net of repurchases Garvey and Hanka (1999) use asimilar measure in their study This measure directly captures the netleverage effect of debt and equity issuancesColumns (7) and (8) report results from regressions of the net leverage

effect of debt and equity issues on changes in profitability after controllingfor other leverage factors in differences and year indicator variables Theeffect of profitability on leverage change due to external financing is positiveand statistically significant This comparison suggests that changes in totaldebt and equity are only partly a result of debt and equity issuance decisionsOther balance sheet adjustments complicate the inferences from leverageratio regressionsFor most of our analysis we follow the literature in treating debt as homo-

genous It is of course potentially interesting to ask what happens when wedistinguish between short-term debt and long-term debt Which part of thefirmrsquos debt maturity structure drives the under-adjustment Table XII

PROFITSndashLEVERAGE PUZZLEREVISITED 1443

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

provides results from such decompositions Alternative scalings areconsidered The main results of interest are that the distinction betweenshort- and long-term debt is important Short-term debt issuance is largelyindependent of changes in profits The negative sign on profit change stemsfrom the action of long-term debt This seems intuitively reasonable

Table XI Scaled issuances

The sample contains nonfinancial firms listed on the annual Compustat files for the period

from 1971 to 2009 The table presents estimates from regressions of debt issuance to totalissuance (in columns (1) and (2)) debt issuance to total capitalization (in columns (3) and(4)) change in debt to capitalization ratio (in columns (5) and (6)) and change in leverage

ratio considering only external financing (in columns (7) and (8)) The change in leverageratio considering only external financing is defined as

LevEF frac14Dt1 thorn d

Dt1 thorn Et1 thorn dthorn e

Dt1

Dt1 thorn Et1

where D is debt E is book value of equity d is debt issues net of retirements and e is equity

issues net of repurchases The regressions include lagged leverage factors and year indicatorvariables Some of the specifications include firm fixed effects The reported t-statistics arecorrected for clustering at the firm level

aMean significant at the 1 level

bMean significant at the 5 level

cMean significant at the 10 level

DebtIssTotIss

DebtIssethDthornETHORN D

DthornE LevEF

(1) (2) (3) (4) (5) (6) (7) (8)

Profitt1 004a 000 002a 001c 003a 002 001b 002a

(35) (03) (34) (18) (35) (14) (23) (26)

IndLevt1 004c 001 001 001c 000 000 000 000

(19) (06) (15) (19) (02) (04) (03) (05)

Market=bookt1 001a 001a 000a 000a 000a 000a 000 000c

(47) (85) (68) (60) (55) (36) (05) (18)

Tangt1 004 005 013a 011a 011a 006a 003a 003a

(12) (15) (144) (112) (79) (38) (44) (42)

Assetst1 008a 004a 006a 004a 002a 001a 003a 002a

(97) (45) (283) (161) (53) (35) (147) (101)

Constant 079a 074a 002a 003a 000 000 000b 002a

(497) (442) (91) (111) (00) (05) (22) (93)

Year FE Yes Yes Yes Yes Yes Yes Yes Yes

Firm FE No Yes No Yes No Yes No Yes

R2 0015 0098 0022 0076 0009 0070 0012 0086

N 133905 133905 140552 140552 138229 138229 140572 140572

1444 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

inasmuch as short-term debt is probably more associated with day-to-dayoperations rather than longer term planning considerationsIn Table XIII we examine change in leverage ( D

DthornE) and change inleverage due to external financing (LEVEF) during good and bad timesseparately The traditional negative effect of a change in profits on achange in leverage is found both in good times and in bad timesHowever when we decompose the impact in order to isolate the externalcomponent as before the result disappears In good times the externalfinancing component of a change in profits is weak but still positiveDuring bad times it is not statistically significant This implies that distin-guishing between good and bad times does not fundamentally alter the basic

Table XII Short- and long-term debt issuances

The sample contains nonfinancial firms listed on the annual Compustat files for the period

from 1971 to 2009 The table presents estimates from regressions of short-term debtissuance to total issuance (in columns (1) and (2)) long-term debt issuance to totalissuance (in columns (3) and (4)) short-term debt issuance to total capitalization (in

columns (5) and (6)) and long-term debt to total capitalization (in columns (7) and (8))Some of the specifications include firm fixed effects The reported t-statistics are correctedfor clustering at the firm level

aMean significant at the 1 level

bMean significant at the 5 level

cMean significant at the 10 level

STDebtIssTotIss

LTDebtIssTotIss

STDebtIssethDthornETHORN

LTDebtIssethDthornETHORN

(1) (2) (3) (4) (5) (6) (7) (8)

Profitabilityt1 0002 0003 0042a 0008 0001 0001 0019a 0012a

(02) (04) (30) (05) (14) (10) (45) (28)

IndLevt1 0009 0007 0070b 0033 0001 0002c 0014a 0005

(06) (05) (26) (11) (08) (17) (26) (09)

M=Bt1 0003a 0003a 0003b 0010a 0000 0000 0003a 0002a

(48) (35) (22) (65) (10) (01) (84) (42)

Tangibilityt1 0020 0007 0011 0067 0005a 0006a 0097a 0081a

(12) (03) (03) (16) (34) (33) (108) (89)

Assetst1 0039a 0011b 0034a 0024b 0001a 0001 0046a 0020a

(89) (22) (38) (23) (36) (13) (190) (82)

Constant 0001b 0153a 0798a 0591a 0000c 0006a 0057a 0096a

(25) (144) (470) (288) (17) (124) (292) (234)

Year FE Yes Yes Yes Yes Yes Yes Yes Yes

Firm FE No Yes No Yes No Yes No Yes

R2 0025 0089 0018 0073 0030 0147 0016 0298

N 136146 136146 136140 136140 143031 143031 142847 142847

PROFITSndashLEVERAGE PUZZLEREVISITED 1445

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

Table XIII Externally financed leverage changes in good and bad times

The sample contains nonfinancial firms listed on the annual Compustat files for the period

from 1971 to 2009 The table presents estimates from regressions of change in debt tocapitalization ratio (in columns (1) (2) (5) and (6)) and change in leverage ratio consider-ing only external financing (in columns (3) (4) (7) and (8)) The change in leverage ratio

considering only external financing is defined as

LevEF frac14Dt1 thorn d

Dt1 thorn Et1 thorn dthorn e

Dt1

Dt1 thorn Et1

where D is debt E is book value of equity d is debt issues net of retirements and e is equityissues net of repurchases The regressions include lagged leverage factors and year indicator

variables The reported t-statistics are corrected for clustering at the firm level An industryis defined as having ldquogood timesrdquo if the median firm in that industry has a market-to-bookratio that is higher than the 67th percentile of the time-series distribution of industry

median market-to-book ratios Conversely an industry is defined as having ldquobad timesrdquoif the median firm in that industry has a market-to-book ratio that is lower than the 33rdpercentile of the time-series distribution of industry median market-to-book ratios

aMean significant at the 1 level

bMean significant at the 5 level

cMean significant at the 10 level

Good times Bad times

DDthornE LevEF D

DthornE LevEF

(1) (2) (3) (4) (5) (6) (7) (8)

Profitabiltyt1 0022c 0005 0016c 0025b 0073b 0041 0007 0007

(17) (03) (18) (24) (25) (11) (04) (03)

IndLevt1 0000 0011 0001 0009 0010 0018 0001 0000

(00) (09) (01) (13) (07) (10) (02) (00)

MB

t1

0004a 0003b 0000 0001 0008a 0001 0000 0000

(40) (24) (05) (08) (28) (04) (03) (03)

Tangt1 0097a 0041 0033a 0038b 0127a 0048 0059a 0033

(45) (16) (27) (25) (49) (14) (43) (16)

Assetst1 0014a 0013b 0024a 0019a 0037a 0023c 0036a 0025a

(30) (20) (84) (48) (45) (17) (88) (43)

Constant 0030b 0036 0000 0003 0028c 0014 0021 0001

(20) (16) (00) (07) (17) (06) (11) (01)

Year FE Yes Yes Yes Yes Yes Yes Yes Yes

Firm FE No Yes No Yes No Yes No Yes

R2 0006 0088 0009 0117 0026 0251 0023 0202

N 56820 56820 58200 58200 23462 23462 23511 23511

1446 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

conclusions The traditional leverage result on profits is driven by increasesin equity that are experienced and then partially but not completely offsetby firmsTo the extent that firms are engaging in active financing they are following

the prescriptions of the trade-off theory Thus we have a coherent picture ofhow firms are responding to profits Basically more profitable firms experiencean increase in the value of their equity They respond to this by issuing debtand repurchasing equity But the magnitude of the increase in equity is largerthan the magnitudes of the active steps taken in the debt and equity marketsThis means that the firmsrsquo actual actions are as predicted by the trade-offtheory But it leaves open the question of why they do not act more stronglyThe capital structure literature has long been interested in dynamics and

capital structure adjustment paths In recent years this interest has increas-ingly resulted in explicit dynamic optimizing models such as those ofGoldstein Ju and Leland (2001) Hennessy and Whited (2005) andStrebulaev (2007) These models commonly make specialized assumptionsleading some to question their generalityWe argue that both fixed transactions costs and variable transactions costs

are plausible and realistic in models that pin down firm size The simultaneouspresence of both is important Due to fixed costs firms only rebalance inter-mittently Due to variable adjustment costs the extent of the adjustment isreduced to the point at which an extra unit to adjustment is balanced againstthe extra unit transactions costs This means that both leverage increases andleverage reductions will only go part way toward a static optimumThe key point is that with both fixed and variable transactions costs we

should expect to only see intermittent leverage adjustments When leverage isadjusted it should ldquounder adjustrdquo when compared to a frictionless modelFurther discussion of transactions costs in a capital structure setting isprovided by Leary and Roberts (2005)

9 Conclusion

The connection between corporate profits and capital structure has beenvery influential in the assessment of the trade-off theory The conventionalinterpretation of the evidence has pushed the literature away from the trade-off theory and toward much more complex dynamic models and ideas As aresult it is important to make sure that the evidence has been correctlyinterpreted We show that the problem is with the interpretation of thetest results and a failure to take into account the costs associated with re-sponding to a profit shock

PROFITSndashLEVERAGE PUZZLEREVISITED 1447

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

We document that more profitable firms really do borrow more (not less)They tend to repurchase equity They experience an increase in both thebook value of equity and the market value of equity Less profitable firmsreally do tend to reduce their debt and to issue equityFirm size and market conditions also matter Larger firms tend to be more

active in the debt markets while smaller firms tend to be relatively moreactive in the equity markets During good times there is more use ofexternal financingThe usual profitsndashleverage puzzle result is primarily driven by the increase

in equity that is experienced by the more profitable firms Accordingly thepuzzle should be restated as asking why do firms not take sufficiently largeoffsetting actions to fully undo the change in equity What limits the mag-nitudes of the typical leverage response to profit shocksIn a frictionless model the partial response appears puzzling But there is

good empirical reason to believe that rebalancing entails both fixed andvariable costs and that firm size matters The rebalancing costs can befully avoided by doing nothing Accordingly the firm must decide whetherany given shock is big enough to be worth responding to If it is then thefirm must decide how big a response is called for These technical conditionsare known as ldquovalue matchingrdquo and ldquosmooth pastingrdquo Thus optimizationimplies that some shocks will be ignored Even if the shock is not ignoredthe optimal response will only partially undo the shock The magnitude ofthe leverage response must balance the marginal cost and the marginalbenefit of an extra unit of leverage Since the marginal cost of adjustingleverage is strictly positive the adjustment toward that static leverageoptimum will only go part way This is true both for leverage increasesand for leverage reductionsOne might conjecture that the variable costs would be time varying and

that such costs might have declined over time Since we do not have a fullaccounting of all of the variable costs it is of course impossible to be sureHowever there is some evidence suggesting that at least the directunderwriting fees are remarkably stable over time10

Surely nobody ought to be surprised that more profitable firms are moreinclined to repurchase equity while unprofitable firms tend not to do so Thefact that more profitable firms typically experience an increase in the value of

10 For IPOs Jay Ritter provides important summary evidence as httpbearwarringtonufleduritterIPOs2013Underwritingpdf He finds that for medium-sized deals a remark-

ably large fraction of the deals have a 7 fee year after year Whether this is due tobehavioral factors or some sort of a coordination device is not clear The extent towhich similar intertemporal stability carries over to other forms of external financemerits further study

1448 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

equity is equally natural Were it not for the contrary claims in the literature(ie the profitsndashleverage puzzle) we would argue that nobody ought to findit surprising that more profitable firms are more inclined to issue debt Thefact that leverage adjustments typically only partially offset profit shocksshould also sound quite natural These facts all fit together easily within thetrade-off theory once we take account of both fixed and variable costs ofactively adjusting capital structure

Appendix

Table AI Market leverage regressions

The table presents estimates of the regressions of market leverage on firm characteristics

The sample comes from the annual Compustat files during the period 1971ndash2009 Financialfirms are excluded The leverage is estimated as the ratio of debt over debt plus marketequity The explanatory variables Profitabilityt1

MB

t1

Tangibilityt1 and LnethAssetsTHORNt1are described in Table 1 IndMedianLevt1 is estimated as the median market leverage of allother firms in the same industry The industry is defined at the level of the four-digit SICcode All specifications include the year fixed effects The specifications in column (2)

additionally include the firm fixed effects We report t-statistics where the standard errorsare clustered at the firm level in parentheses

aSignificant at the 1 level

Quantile regressions

(1) (2) (3) (4) (5)

Profitabilityt1 013a 013a 007a 011a 020a

(275) (262) (349) (425) (538)

IndLevt1 046a 020a 032a 058a 063a

(506) (239) (1299) (1630) (1042)MB

t1

003a 001a 001a 002a 004a

(509) (303) (743) (709) (534)

Tangibilityt1 013a 018a 013a 014a 010a

(191) (181) (762) (571) (232)

LnethAssetsTHORNt1 001a 005a 001a 001a lt001a

(112) (258) (682) (342) (29)

Constant 010a 011a 005a 005a 029a

(171) (98) (173) (118) (398)

Year FE Yes Yes Yes Yes Yes

Firm FE No Yes No No No

Regression type OLS FE 25th tile Median 75thtile

Clustered SE Yes Yes No No No

R2 030 067

Pseudo R2 016 023 0201

N 158578 158578 158578 158578 158578

PROFITSndashLEVERAGE PUZZLEREVISITED 1449

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

Table AII Profitability sorts for firms issuing large amounts

Percentage of firms

Iss Ret Iss Iss Rep Iss Iss IssD IssE

D D NetD E E NetE DampE RepE RetD

(1) (2) (3) (4) (5) (6) (7) (8) (9)

Panel A 10 cutoff to define issuing or retiring activity

Allfirm

s

Low 205 150 134 260 05 257 45 00 23

2 207 175 103 77 06 75 11 01 10

3 243 196 101 50 08 47 08 01 09

4 252 186 113 50 13 48 08 02 08

High 227 161 118 60 27 57 07 04 09

t Highfrac14Low 70 38 62 755 226 768 320 90 152

Smallfirm

s Low 230 132 162 380 05 375 82 00 32

2 228 143 165 334 04 333 65 00 26

3 211 146 147 216 03 214 38 00 23

4 190 154 120 126 04 124 15 01 20

High 194 162 114 99 09 97 12 01 15

t Highfrac14Low 42 40 68 342 20 340 169 07 131

Largefirm

s Low 231 200 82 31 06 29 08 01 04

2 231 187 67 17 06 14 03 01 02

3 234 177 77 17 09 14 03 02 03

4 218 158 75 16 17 14 02 03 02

High 218 135 99 19 41 15 03 07 02

t Highfrac14Low 19 106 36 46 135 57 39 58 23

Percentage of firms

Iss Ret Iss Iss Rep Iss Both Iss D Iss E

D D Net D E E Net E DampE Rep E Ret D

(1) (2) (3) (4) (5) (6) (7) (8) (9)

Panel B 15 cutoff to define issuing or retiring activity

Allfirm

s

Low 158 101 100 230 01 228 27 00 15

2 142 115 63 57 01 56 05 00 06

3 165 128 57 32 01 31 03 00 05

4 173 118 65 35 01 34 03 00 05

High 160 102 70 46 04 44 03 00 05

t Highfrac14Low 08 06 141 738 87 746 277 24 131

Smallfirm

s Low 195 95 132 354 01 350 56 00 24

2 185 96 131 303 01 301 44 00 18

3 160 96 104 179 00 176 18 00 15

4 134 99 78 99 01 98 08 00 12

High 133 95 73 82 01 80 06 00 10

t Highfrac14Low 80 00 95 345 12 344 149 12 52

Largefirm

s Low 145 123 42 14 01 12 03 00 01

2 137 114 32 07 01 06 01 00 01

3 142 104 35 06 01 05 00 00 01

4 130 91 36 08 03 07 01 01 01

High 134 81 48 09 06 07 01 01 01

t Highfrac14Low 19 85 17 26 54 33 34 17 11

1450 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

Table AIV Debt and equity changesmdashwithout fixed effects

The sample contains nonfinancial firms listed on the annual Compustat files for the period from 1971

to 2009 The table presents estimates from regressions of change in debt (D) change in the book

value of equity (BVE) change in the market value of equity (MVE) and net equity issuances

(EqyIss) The explanatory variables include current and lagged change in operating income before

depreciation (Profits and Profitst1) and lagged change in book value of assets (Assetst1) The

regressions include year indicator variables The reported t-statistics are corrected for clustering at the

firm level

aMean significant at the 1 level

bMean significant at the 5 level

cMean significant at the 10 level

D BVE MVE EqyIss

(1) (2) (3) (4) (5) (6) (7) (8)

Profitst1 025a 024a 023a 038a 007 066a 007a 009a

(40) (32) (36) (39) (06) (29) (42) (51)Profits 051b 082a 149a 004a

(25) (33) (49) (33)Assetst1 002 005c 024a 001b

(08) (17) (34) (23)Constant 93 176a 45 71 1552a 1704a 130a 140a

(12) (28) (07) (09) (49) (49) (45) (45)Year FE Yes Yes Yes Yes Yes Yes Yes YesR2 001 005 001 011 001 005 002 002N 162056 162056 162130 162130 157550 157550 162154 162154

Table AIII Median profitability sorts

The sample contains nonfinancial firms listed on the annual Compustat files for the period from 1971

to 2009 The table presents median debt and equity median changes in debt and equity and median

equity issuances (EqyIss) in millions of dollars for all firms sorted on profitability and for profitability

sorts within the smallest and largest firms The sorts are done annually

D D BVE BVE MVE MVE EqyIss Assets(1) (2) (3) (4) (5) (6) (7) (8)

Allfirm

s

Low 21 00 75 17 256 20 01 1922 186 02 449 18 598 19 00 11033 816 06 1274 03 1637 05 00 32244 787 02 1541 18 2488 03 00 3556High 273 00 1361 66 3300 14 00 2578

Smallfirm

s Low 05 00 13 07 140 14 06 402 06 00 25 08 114 12 01 563 09 00 32 05 85 07 00 674 12 00 39 01 68 03 00 79High 10 00 51 04 98 01 00 90

Largefirm

s Low 11485 381 10269 371 13227 202 03 333902 13051 393 12786 140 17571 169 05 392463 11554 287 13429 24 21623 305 00 370314 9963 209 15052 194 29424 308 00 36705High 7361 00 16380 613 49142 439 00 34976

PROFITSndashLEVERAGE PUZZLEREVISITED 1451

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

References

Auerbach A J (1985) Real determinants of corporate leverage in B M Friedman (ed)

Corporate Capital Structures in the United States University of Chicago Press Chicago

IllinoisBertrand M and Mullainathan S (2003) Enjoying the quiet life corporate governance and

managerial preferences Journal of Political Economy 111 1043ndash1075Booth L Varouj A Demirguc-Kunt A and Maksimovic V (2001) Capital structures in

developing countries Journal of Finance 56 87ndash130Brennan M J and Schwartz E S (1984) Optimal financial policy and firm valuation

Journal of Finance 39 593ndash607Cameron A C and Trivedi P K (2010) Microeconometrics using Stata Stata Press TexasConstantinides G M (1979) Multiperiod consumption and investment behavior with

convex transactions costs Management Science 25 1127ndash1137Constantinides G M and Richard S F (1978) Existence of optimal simple policies for

discounted-cost inventory and cash management in continuous time Operations Research

26 620ndash636DeAngelo H DeAngelo L and Whited T M (2011) Capital structure dynamics and

transitory debt Journal of Financial Economics 99 235ndash261Dixit A K (1993) The Art of Smooth Pasting Routledge London and New YorkFama E and French K R (2002) Testing trade-off and pecking order predictions about

dividends and debt Review of Financial Studies 15 1ndash33Faulkender M and Petersen M A (2006) Does the source of capital affect capital struc-

ture Review of Financial Studies 19 45ndash79Faulkender M Flannery M J Hankins K W and Smith J M (2012) Cash Flows and

Leverage Adjustments Journal of Financial Economics 103 632ndash646Fischer E Heinkel R and Zechner J (1989) Dynamic capital structure choice theory

and tests Journal of Finance 44 19ndash40Frank M Z and Goyal V K (2008) Trade-off and pecking order theories of

debt in B E Eckbo (ed) Handbook of Corporate Finance Empirical Corporate Finance

Vol 2 Handbook of Finance Series Chapter 12 ElsevierNorth-Holland Amsterdam pp

135ndash202

Frank M Z and Goyal V K (2009) Capital structure decisions which factors are reliably

important Financial Management 38 1ndash37

Garvey G T and Hanka G (1999) Capital structure and corporate control the effect of

antitakeover statutes on firm leverage Journal of Finance 54 519ndash546

Goldstein R Ju N and Leland H (2001) An EBIT-based model of dynamic capital

structure Journal of Business 74 483ndash512

Graham J R and Tucker A L (2006) Tax shelters and corporate debt policy Journal of

Financial Economics 81 563ndash594

Hackbarth D and Mauer D C (2012) Optimal priority structure capital structure and

investment Review of Financial Studies 25 747ndash796

Halling M Yu J and Zechner J (2012) Leverage dynamics over the business cycle

Working paper University of Utah University of New South Wales and Vienna

University of Economics and Business AdministrationHennessy C A and Whited T A (2005) Debt dynamics Journal of Finance 60 1129ndash1165

Huang R and Ritter J R (2009) Testing theories of capital structure and estimating the

speed of adjustment Journal of Financial and Qualitative Analysis 44 237ndash271

1452 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

Hovakimian A Opler T and Titman S (2001) The debt-equity choice Journal of

Financial and Quantitative Analysis 36 1ndash24Jung K Kim Y and Stulz R M (1996) Timing investment opportunities managerial

discretion and the security issue decisions Journal of Financial Economics 42 159ndash185Kane A Marcus A J and McDonald R L (1984) How big is the tax advantage to

debt Journal of Finance 39 841ndash853Kayhan A and Titman S (2007) Firmsrsquo histories and their capital structures Journal of

Financial Economics 83 1ndash32Leary M T (2009) Bank loan supply lender choice and corporate capital structure

Journal of Finance 64 1143ndash1185Leary M T and Roberts M R (2005) Do firms rebalance their capital structures

Journal of Finance 60 2575ndash2619Leland H E (1994) Corporate debt value bond covenants and optimal capital structure

Journal of finance 49 1213ndash1252Lemmon M L Roberts M R and Zender J F (2008) Back to the beginning persist-

ence and the cross-section of corporate capital structure Journal of Finance 63 1575ndash1608Long M S and Malitz I B (1985) The investment-financing nexus some empirical

evidence Midland Corporate Finance Journal 3 53ndash59Mackie-Mason J K (1990) Do taxes affect corporate financing decisions Journal of

Finance 45 1471ndash1493Morellec E Nikolov B and Schurhoff N (2012) Corporate governance and capital

structure dynamics Journal of Finance 67 803ndash848Myers S C (1984) The capital structure puzzle Journal of Finance 39 575ndash592Myers S C (1993) Still searching for optimal capital structure Journal of Applied

Corporate Finance 61 4ndash14Rajan R and Zingales L (1995) What do we know about capital structure some evidence

from international data Journal of Finance 50 1421ndash1460Strebulaev I A (2007) Do tests of capital structure theory mean what they say Journal of

Finance 62 1747ndash1787

Strebulaev I A and Whited T M (2012) Dynamic Models and Structural Estimation in

Corporate Finance Now Publisher Boston and Delft

Titman S and Wessels R (1988) The determinants of capital structure choice Journal of

Finance 43 1ndash21

Welch I (2004) Capital structure and stock returns Journal of Political Economy 112

106ndash131

Welch I (2011) Two common problems in capital structure research the financial-debt-

to-asset ratio and issuing activity versus leverage changes International Review of Finance

11 1ndash17

PROFITSndashLEVERAGE PUZZLEREVISITED 1453

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

7 Are Financial Market Conditions Important

According to the trade-off theory capital structure is determined by con-sidering the impact of costs and benefits of debt The time variation inthe costs and benefits of debt imply that issuance decisions would alsovary over time Furthermore market conditions affect the transactioncosts of adjusting Adverse selection is a greater problem in a cold marketthan in a hot one Accordingly it may be easier to adjust in good marketconditions Related ideas have been developed by Huang and Ritter(2009) Faulkender et al (2012) and Halling Yu and Zechner (2012)Faulkender Flannery Hankins and Smithrsquos paper examines cross-sectionalvariation in adjustment speeds including market valuations andhow they may affect adjustment speeds It is natural to think that ingood market conditions it will be less costly to issue both debt and equityand adjusting leverage to respond to profitability shocks would be mucheasierTo test the importance of market conditions we require a definition of

good times and bad times Our empirical strategy is to estimate good timesversus bad times at the four-digit industry level We define an industry ashaving ldquogood timesrdquo if the median firm in that industry has a market-to-book ratio that is higher than the 67th percentile of the time-series distribu-tion of the industry median market-to-book ratios Conversely an industryis defined as having ldquobad timesrdquo if the median firm in the industry has amarket-to-book ratio that is lower than the 33rd percentile of the time-seriesdistribution of the industry median market-to-book ratiosPanel A of Table IX tabulates issuance activity for profitability sorts in

both good times and bad times As might be expected active debt and equityissuances are larger during good times There is also somewhat more activeswaps between debt and equity during good timesPanel B reports the issuance activity for the smallest and the largest firms

during the good times and bad times Within each size quintile firms aresorted on profitability This is a three-way sort As before we find that thereis much more active use of external markets during good times Small low-profit firms are more likely to issue equity in good times than in bad timesDebt issuances are significantly higher in good times Large high-profitfirms are significantly more likely to issue debt and repurchase equity ingood times than in bad timesPanel C reports the magnitudes of the financial variables rather than the

frequencies During bad times less profitable large firms retire substantialamounts of debt and they show a tendency to issue equity Small firms donot seem to engage in similar debt reduction activities However like the

PROFITSndashLEVERAGE PUZZLEREVISITED 1439

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

Table IX Debt and equity issues in good and bad times

The table reports the frequency and magnitude of financing activity for sorts on profits for subsamples

of firms in good and bad times An industry is defined as having ldquogood timesrdquo if the median firm in

that industry has a market-to-book ratio that is higher than the 67th percentile of the time-series

distribution of industry median market-to-book ratios Conversely an industry is defined as having

bad times if the median firm in that industry has a market-to-book ratio that is lower than the 33rd

percentile of the time-series distribution of industry median market-to-book ratios Panels A and B

report the percentage of firms (i) issuing net debt in excess of 5 of the value of their assets (ii) issuing

net equity in excess of 5 of the value of their assets (iii) issuing debt and repurchasing equity both in

excess of 5 of the value of their assets and (iv) issuing equity and retiring debt both in excess of 5

of the value of their assets Panel C tabulates the mean debt and equity levels and the mean changes in

debt and equity issuances The sample contains nonfinancial firms listed on the annual Compustat files

for the period from 1971 to 2009

Iss Iss Iss D Iss ENet D Net E Rep E Ret D(1) (2) (3) (4)

Panel A sorts on profitability

Badtimes

Low 145 131 01 202 170 49 02 083 173 46 04 074 182 47 03 10

High 177 51 06 10t Highfrac14Low 41 136 33 01

Goodtimes

Low 206 381 01 462 197 148 02 233 212 103 04 184 234 92 10 18

High 216 92 16 14t Highfrac14Low 23 658 138 23

Panel B sorts on firm size and profitability

Badtimes

Smallfirm

s Low 151 318 00 312 182 276 02 253 151 140 02 294 183 73 00 19

High 177 80 00 11t Highfrac14Low 10 90 NA 20

Largefirm

s Low 172 54 01 042 148 39 04 053 150 43 04 044 156 33 03 06

High 147 28 05 03t Highfrac14Low 17 32 20 02

Goodtimes Smallfirm

s Low 213 448 01 542 230 443 01 473 222 325 00 474 203 197 02 38

High 190 153 02 31t Highfrac14Low 19 230 10 39

Largefirm

s Low 203 96 03 122 208 66 09 133 232 54 16 064 215 39 19 05

High 234 36 38 04t Highfrac14Low 29 99 92 38

(continued)

1440 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

larger firms there is a tendency for small firms to issue more equitymdashat leastto the extent possibleDuring good times large profitable firms raise significant amounts of debt

as they experience an increase in the value of their equity Such firms alsoengage in share repurchases Debt issues by smaller firms are much lessaffected by their own profitability during good times and their change indebt is more or less independent of their profits However equity issuanceis primarily found among the low-profit small firms during good timesTable IX shows that empirically profitability has a time-varying impacton leverage choicesIn Table X the basic regressions for changes in debt and for equity

issuance are presented for both good times and bad times Comparinggood times with bad times we see that the effects of profits are muchstronger in good times In bad times the impacts are rather weak statisticallyThis difference is interesting and deserves further study However for ourpurposes recall that the static trade-off theory implies time-varying capital

Table IX Continued

D D BVE BVE MVE MVE EqyIss Assets

(1) (2) (3) (4) (5) (6) (7) (8)

Panel C magnitude of financing activity

Badtimes Smallfirm

s Low 40 03 58 14 305 06 22 140

2 44 05 82 07 240 67 17 180

3 45 04 94 02 176 18 05 187

4 43 09 98 02 137 04 03 188

High 36 10 114 13 209 16 04 199

Largefirm

s Low 3365 73 2523 169 2304 163 94 8972

2 3031 95 2983 35 3284 278 56 9068

3 2930 5 3360 14 3947 308 57 9225

4 2281 6 3817 19 5213 442 7 8902

High 1790 62 4767 13 8799 58 31 9573

Goodtimes

Smallfirm

s Low 24 06 38 05 339 23 39 90

2 26 08 48 01 301 04 25 104

3 30 09 60 04 246 03 17 125

4 28 07 66 08 206 18 12 130

High 23 06 85 16 261 13 08 141

Largefirm

s Low 4532 36 3477 103 6557 260 53 12465

2 4580 58 3527 139 6911 421 23 12522

3 2952 111 3485 2 7586 279 32 9535

4 2752 124 3939 72 10686 33 105 10226

High 2429 228 4819 377 18565 268 222 10830

PROFITSndashLEVERAGE PUZZLEREVISITED 1441

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

structure choices even when the target is time invariant Empirically we doobserve time-varying choices

8 Is Scaling Important

In the preceding section we provided results for unscaled debt and equityissues This seems appropriate to us However it involves making twochanges to the standard regression We would like to determine if both

Table X Debt and equity changes in good and bad timesmdashregressions

The sample contains nonfinancial firms listed on the annual Compustat files for the period

from 1971 to 2009 The table presents estimates from regressions of change in debt (D)and net equity issuances (EqyIss) in good and bad times The explanatory variables includechange in profits lagged change in profits and lagged change in assets The regressions

include both firm and year fixed effects An industry is defined as having ldquogood timesrdquo ifthe median firm in that industry has a market-to-book ratio that is higher than the 67thpercentile of the time-series distribution of industry median market-to-book ratios

Conversely an industry is defined as having bad times if the median firm in thatindustry has a market-to-book ratio that is lower than the 33rd percentile of the time-series distribution of industry median market-to-book ratios The reported t-statistics arecorrected for clustering at the firm level

aMean significant at the 1 level

bMean significant at the 5 level

cMean significant at the 10 level

Good times Bad times

D EqyIss D EqyIss

(1) (2) (3) (4) (5) (6) (7) (8)

Profits 087a 091a 006b 009a 060 065 003 003b

(30) (29) (20) (28) (15) (16) (16) (20)

Profitst1 029c 015a 034c 001

(19) (37) (17) (04)

Assetst1 000 002b 000 001

(02) (22) (00) (07)

Constant 1337 3989 8691a 10001a 1351 1851 4099a 4055a

(03) (09) (55) (37) (09) (13) (80) (78)

Year FE Yes Yes Yes Yes Yes Yes Yes Yes

Firm FE Yes Yes Yes Yes Yes Yes Yes Yes

R2 0071 0080 0018 0045 0178 0212 0034 0041

N 74827 68729 74891 68787 26965 26013 26971 26018

1442 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

changes are crucial Accordingly in this section we take a step back toconsider prior literature in which variables are scaled The questions arehow does scaling affect the results and whyTable XI examines alternative scalings Columns (1) and (2) scale debt and

equity issuances by the total firm net issuance Columns (3) and (4) scaledebt and equity issuances by the sum of debt and book equity Columns (5)and (6) consider the change in debt scaled by the sum of debt and bookequity Columns (7) and (8) consider the change in leverage due to externalfinancing Since we are examining changes we use the changes in thestandard regression factors as controls The primary interest is in the coef-ficient on the change in profitsWe find that a change in profits is associated with an increase in debt

issues But the sign reverses when we scale the debt issues by the sum of debtplus book equity This is telling What is really driving the results is thechange in equity in the denominator not the change in debt in the numer-ator This result is further substantiated in Columns (5) and (6) where thechange in debt to capital ratio is used instead of debt issuanceTo further explore the impact of active issuance relative to passive

experiencing of changes we decompose leverage by separating out theeffect of debt and equity issues from those due to changes in retainedearnings Thus we define leverage change due to external financing as

LevEF frac14Dt1 thorn d

Dt1 thorn Et1 thorn dthorn e

Dt1

Dt1 thorn Et1 eth1THORN

where D is debt E is book value of equity d is debt issues net of retirementsand e is equity issues net of repurchases Garvey and Hanka (1999) use asimilar measure in their study This measure directly captures the netleverage effect of debt and equity issuancesColumns (7) and (8) report results from regressions of the net leverage

effect of debt and equity issues on changes in profitability after controllingfor other leverage factors in differences and year indicator variables Theeffect of profitability on leverage change due to external financing is positiveand statistically significant This comparison suggests that changes in totaldebt and equity are only partly a result of debt and equity issuance decisionsOther balance sheet adjustments complicate the inferences from leverageratio regressionsFor most of our analysis we follow the literature in treating debt as homo-

genous It is of course potentially interesting to ask what happens when wedistinguish between short-term debt and long-term debt Which part of thefirmrsquos debt maturity structure drives the under-adjustment Table XII

PROFITSndashLEVERAGE PUZZLEREVISITED 1443

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

provides results from such decompositions Alternative scalings areconsidered The main results of interest are that the distinction betweenshort- and long-term debt is important Short-term debt issuance is largelyindependent of changes in profits The negative sign on profit change stemsfrom the action of long-term debt This seems intuitively reasonable

Table XI Scaled issuances

The sample contains nonfinancial firms listed on the annual Compustat files for the period

from 1971 to 2009 The table presents estimates from regressions of debt issuance to totalissuance (in columns (1) and (2)) debt issuance to total capitalization (in columns (3) and(4)) change in debt to capitalization ratio (in columns (5) and (6)) and change in leverage

ratio considering only external financing (in columns (7) and (8)) The change in leverageratio considering only external financing is defined as

LevEF frac14Dt1 thorn d

Dt1 thorn Et1 thorn dthorn e

Dt1

Dt1 thorn Et1

where D is debt E is book value of equity d is debt issues net of retirements and e is equity

issues net of repurchases The regressions include lagged leverage factors and year indicatorvariables Some of the specifications include firm fixed effects The reported t-statistics arecorrected for clustering at the firm level

aMean significant at the 1 level

bMean significant at the 5 level

cMean significant at the 10 level

DebtIssTotIss

DebtIssethDthornETHORN D

DthornE LevEF

(1) (2) (3) (4) (5) (6) (7) (8)

Profitt1 004a 000 002a 001c 003a 002 001b 002a

(35) (03) (34) (18) (35) (14) (23) (26)

IndLevt1 004c 001 001 001c 000 000 000 000

(19) (06) (15) (19) (02) (04) (03) (05)

Market=bookt1 001a 001a 000a 000a 000a 000a 000 000c

(47) (85) (68) (60) (55) (36) (05) (18)

Tangt1 004 005 013a 011a 011a 006a 003a 003a

(12) (15) (144) (112) (79) (38) (44) (42)

Assetst1 008a 004a 006a 004a 002a 001a 003a 002a

(97) (45) (283) (161) (53) (35) (147) (101)

Constant 079a 074a 002a 003a 000 000 000b 002a

(497) (442) (91) (111) (00) (05) (22) (93)

Year FE Yes Yes Yes Yes Yes Yes Yes Yes

Firm FE No Yes No Yes No Yes No Yes

R2 0015 0098 0022 0076 0009 0070 0012 0086

N 133905 133905 140552 140552 138229 138229 140572 140572

1444 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

inasmuch as short-term debt is probably more associated with day-to-dayoperations rather than longer term planning considerationsIn Table XIII we examine change in leverage ( D

DthornE) and change inleverage due to external financing (LEVEF) during good and bad timesseparately The traditional negative effect of a change in profits on achange in leverage is found both in good times and in bad timesHowever when we decompose the impact in order to isolate the externalcomponent as before the result disappears In good times the externalfinancing component of a change in profits is weak but still positiveDuring bad times it is not statistically significant This implies that distin-guishing between good and bad times does not fundamentally alter the basic

Table XII Short- and long-term debt issuances

The sample contains nonfinancial firms listed on the annual Compustat files for the period

from 1971 to 2009 The table presents estimates from regressions of short-term debtissuance to total issuance (in columns (1) and (2)) long-term debt issuance to totalissuance (in columns (3) and (4)) short-term debt issuance to total capitalization (in

columns (5) and (6)) and long-term debt to total capitalization (in columns (7) and (8))Some of the specifications include firm fixed effects The reported t-statistics are correctedfor clustering at the firm level

aMean significant at the 1 level

bMean significant at the 5 level

cMean significant at the 10 level

STDebtIssTotIss

LTDebtIssTotIss

STDebtIssethDthornETHORN

LTDebtIssethDthornETHORN

(1) (2) (3) (4) (5) (6) (7) (8)

Profitabilityt1 0002 0003 0042a 0008 0001 0001 0019a 0012a

(02) (04) (30) (05) (14) (10) (45) (28)

IndLevt1 0009 0007 0070b 0033 0001 0002c 0014a 0005

(06) (05) (26) (11) (08) (17) (26) (09)

M=Bt1 0003a 0003a 0003b 0010a 0000 0000 0003a 0002a

(48) (35) (22) (65) (10) (01) (84) (42)

Tangibilityt1 0020 0007 0011 0067 0005a 0006a 0097a 0081a

(12) (03) (03) (16) (34) (33) (108) (89)

Assetst1 0039a 0011b 0034a 0024b 0001a 0001 0046a 0020a

(89) (22) (38) (23) (36) (13) (190) (82)

Constant 0001b 0153a 0798a 0591a 0000c 0006a 0057a 0096a

(25) (144) (470) (288) (17) (124) (292) (234)

Year FE Yes Yes Yes Yes Yes Yes Yes Yes

Firm FE No Yes No Yes No Yes No Yes

R2 0025 0089 0018 0073 0030 0147 0016 0298

N 136146 136146 136140 136140 143031 143031 142847 142847

PROFITSndashLEVERAGE PUZZLEREVISITED 1445

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

Table XIII Externally financed leverage changes in good and bad times

The sample contains nonfinancial firms listed on the annual Compustat files for the period

from 1971 to 2009 The table presents estimates from regressions of change in debt tocapitalization ratio (in columns (1) (2) (5) and (6)) and change in leverage ratio consider-ing only external financing (in columns (3) (4) (7) and (8)) The change in leverage ratio

considering only external financing is defined as

LevEF frac14Dt1 thorn d

Dt1 thorn Et1 thorn dthorn e

Dt1

Dt1 thorn Et1

where D is debt E is book value of equity d is debt issues net of retirements and e is equityissues net of repurchases The regressions include lagged leverage factors and year indicator

variables The reported t-statistics are corrected for clustering at the firm level An industryis defined as having ldquogood timesrdquo if the median firm in that industry has a market-to-bookratio that is higher than the 67th percentile of the time-series distribution of industry

median market-to-book ratios Conversely an industry is defined as having ldquobad timesrdquoif the median firm in that industry has a market-to-book ratio that is lower than the 33rdpercentile of the time-series distribution of industry median market-to-book ratios

aMean significant at the 1 level

bMean significant at the 5 level

cMean significant at the 10 level

Good times Bad times

DDthornE LevEF D

DthornE LevEF

(1) (2) (3) (4) (5) (6) (7) (8)

Profitabiltyt1 0022c 0005 0016c 0025b 0073b 0041 0007 0007

(17) (03) (18) (24) (25) (11) (04) (03)

IndLevt1 0000 0011 0001 0009 0010 0018 0001 0000

(00) (09) (01) (13) (07) (10) (02) (00)

MB

t1

0004a 0003b 0000 0001 0008a 0001 0000 0000

(40) (24) (05) (08) (28) (04) (03) (03)

Tangt1 0097a 0041 0033a 0038b 0127a 0048 0059a 0033

(45) (16) (27) (25) (49) (14) (43) (16)

Assetst1 0014a 0013b 0024a 0019a 0037a 0023c 0036a 0025a

(30) (20) (84) (48) (45) (17) (88) (43)

Constant 0030b 0036 0000 0003 0028c 0014 0021 0001

(20) (16) (00) (07) (17) (06) (11) (01)

Year FE Yes Yes Yes Yes Yes Yes Yes Yes

Firm FE No Yes No Yes No Yes No Yes

R2 0006 0088 0009 0117 0026 0251 0023 0202

N 56820 56820 58200 58200 23462 23462 23511 23511

1446 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

conclusions The traditional leverage result on profits is driven by increasesin equity that are experienced and then partially but not completely offsetby firmsTo the extent that firms are engaging in active financing they are following

the prescriptions of the trade-off theory Thus we have a coherent picture ofhow firms are responding to profits Basically more profitable firms experiencean increase in the value of their equity They respond to this by issuing debtand repurchasing equity But the magnitude of the increase in equity is largerthan the magnitudes of the active steps taken in the debt and equity marketsThis means that the firmsrsquo actual actions are as predicted by the trade-offtheory But it leaves open the question of why they do not act more stronglyThe capital structure literature has long been interested in dynamics and

capital structure adjustment paths In recent years this interest has increas-ingly resulted in explicit dynamic optimizing models such as those ofGoldstein Ju and Leland (2001) Hennessy and Whited (2005) andStrebulaev (2007) These models commonly make specialized assumptionsleading some to question their generalityWe argue that both fixed transactions costs and variable transactions costs

are plausible and realistic in models that pin down firm size The simultaneouspresence of both is important Due to fixed costs firms only rebalance inter-mittently Due to variable adjustment costs the extent of the adjustment isreduced to the point at which an extra unit to adjustment is balanced againstthe extra unit transactions costs This means that both leverage increases andleverage reductions will only go part way toward a static optimumThe key point is that with both fixed and variable transactions costs we

should expect to only see intermittent leverage adjustments When leverage isadjusted it should ldquounder adjustrdquo when compared to a frictionless modelFurther discussion of transactions costs in a capital structure setting isprovided by Leary and Roberts (2005)

9 Conclusion

The connection between corporate profits and capital structure has beenvery influential in the assessment of the trade-off theory The conventionalinterpretation of the evidence has pushed the literature away from the trade-off theory and toward much more complex dynamic models and ideas As aresult it is important to make sure that the evidence has been correctlyinterpreted We show that the problem is with the interpretation of thetest results and a failure to take into account the costs associated with re-sponding to a profit shock

PROFITSndashLEVERAGE PUZZLEREVISITED 1447

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

We document that more profitable firms really do borrow more (not less)They tend to repurchase equity They experience an increase in both thebook value of equity and the market value of equity Less profitable firmsreally do tend to reduce their debt and to issue equityFirm size and market conditions also matter Larger firms tend to be more

active in the debt markets while smaller firms tend to be relatively moreactive in the equity markets During good times there is more use ofexternal financingThe usual profitsndashleverage puzzle result is primarily driven by the increase

in equity that is experienced by the more profitable firms Accordingly thepuzzle should be restated as asking why do firms not take sufficiently largeoffsetting actions to fully undo the change in equity What limits the mag-nitudes of the typical leverage response to profit shocksIn a frictionless model the partial response appears puzzling But there is

good empirical reason to believe that rebalancing entails both fixed andvariable costs and that firm size matters The rebalancing costs can befully avoided by doing nothing Accordingly the firm must decide whetherany given shock is big enough to be worth responding to If it is then thefirm must decide how big a response is called for These technical conditionsare known as ldquovalue matchingrdquo and ldquosmooth pastingrdquo Thus optimizationimplies that some shocks will be ignored Even if the shock is not ignoredthe optimal response will only partially undo the shock The magnitude ofthe leverage response must balance the marginal cost and the marginalbenefit of an extra unit of leverage Since the marginal cost of adjustingleverage is strictly positive the adjustment toward that static leverageoptimum will only go part way This is true both for leverage increasesand for leverage reductionsOne might conjecture that the variable costs would be time varying and

that such costs might have declined over time Since we do not have a fullaccounting of all of the variable costs it is of course impossible to be sureHowever there is some evidence suggesting that at least the directunderwriting fees are remarkably stable over time10

Surely nobody ought to be surprised that more profitable firms are moreinclined to repurchase equity while unprofitable firms tend not to do so Thefact that more profitable firms typically experience an increase in the value of

10 For IPOs Jay Ritter provides important summary evidence as httpbearwarringtonufleduritterIPOs2013Underwritingpdf He finds that for medium-sized deals a remark-

ably large fraction of the deals have a 7 fee year after year Whether this is due tobehavioral factors or some sort of a coordination device is not clear The extent towhich similar intertemporal stability carries over to other forms of external financemerits further study

1448 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

equity is equally natural Were it not for the contrary claims in the literature(ie the profitsndashleverage puzzle) we would argue that nobody ought to findit surprising that more profitable firms are more inclined to issue debt Thefact that leverage adjustments typically only partially offset profit shocksshould also sound quite natural These facts all fit together easily within thetrade-off theory once we take account of both fixed and variable costs ofactively adjusting capital structure

Appendix

Table AI Market leverage regressions

The table presents estimates of the regressions of market leverage on firm characteristics

The sample comes from the annual Compustat files during the period 1971ndash2009 Financialfirms are excluded The leverage is estimated as the ratio of debt over debt plus marketequity The explanatory variables Profitabilityt1

MB

t1

Tangibilityt1 and LnethAssetsTHORNt1are described in Table 1 IndMedianLevt1 is estimated as the median market leverage of allother firms in the same industry The industry is defined at the level of the four-digit SICcode All specifications include the year fixed effects The specifications in column (2)

additionally include the firm fixed effects We report t-statistics where the standard errorsare clustered at the firm level in parentheses

aSignificant at the 1 level

Quantile regressions

(1) (2) (3) (4) (5)

Profitabilityt1 013a 013a 007a 011a 020a

(275) (262) (349) (425) (538)

IndLevt1 046a 020a 032a 058a 063a

(506) (239) (1299) (1630) (1042)MB

t1

003a 001a 001a 002a 004a

(509) (303) (743) (709) (534)

Tangibilityt1 013a 018a 013a 014a 010a

(191) (181) (762) (571) (232)

LnethAssetsTHORNt1 001a 005a 001a 001a lt001a

(112) (258) (682) (342) (29)

Constant 010a 011a 005a 005a 029a

(171) (98) (173) (118) (398)

Year FE Yes Yes Yes Yes Yes

Firm FE No Yes No No No

Regression type OLS FE 25th tile Median 75thtile

Clustered SE Yes Yes No No No

R2 030 067

Pseudo R2 016 023 0201

N 158578 158578 158578 158578 158578

PROFITSndashLEVERAGE PUZZLEREVISITED 1449

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

Table AII Profitability sorts for firms issuing large amounts

Percentage of firms

Iss Ret Iss Iss Rep Iss Iss IssD IssE

D D NetD E E NetE DampE RepE RetD

(1) (2) (3) (4) (5) (6) (7) (8) (9)

Panel A 10 cutoff to define issuing or retiring activity

Allfirm

s

Low 205 150 134 260 05 257 45 00 23

2 207 175 103 77 06 75 11 01 10

3 243 196 101 50 08 47 08 01 09

4 252 186 113 50 13 48 08 02 08

High 227 161 118 60 27 57 07 04 09

t Highfrac14Low 70 38 62 755 226 768 320 90 152

Smallfirm

s Low 230 132 162 380 05 375 82 00 32

2 228 143 165 334 04 333 65 00 26

3 211 146 147 216 03 214 38 00 23

4 190 154 120 126 04 124 15 01 20

High 194 162 114 99 09 97 12 01 15

t Highfrac14Low 42 40 68 342 20 340 169 07 131

Largefirm

s Low 231 200 82 31 06 29 08 01 04

2 231 187 67 17 06 14 03 01 02

3 234 177 77 17 09 14 03 02 03

4 218 158 75 16 17 14 02 03 02

High 218 135 99 19 41 15 03 07 02

t Highfrac14Low 19 106 36 46 135 57 39 58 23

Percentage of firms

Iss Ret Iss Iss Rep Iss Both Iss D Iss E

D D Net D E E Net E DampE Rep E Ret D

(1) (2) (3) (4) (5) (6) (7) (8) (9)

Panel B 15 cutoff to define issuing or retiring activity

Allfirm

s

Low 158 101 100 230 01 228 27 00 15

2 142 115 63 57 01 56 05 00 06

3 165 128 57 32 01 31 03 00 05

4 173 118 65 35 01 34 03 00 05

High 160 102 70 46 04 44 03 00 05

t Highfrac14Low 08 06 141 738 87 746 277 24 131

Smallfirm

s Low 195 95 132 354 01 350 56 00 24

2 185 96 131 303 01 301 44 00 18

3 160 96 104 179 00 176 18 00 15

4 134 99 78 99 01 98 08 00 12

High 133 95 73 82 01 80 06 00 10

t Highfrac14Low 80 00 95 345 12 344 149 12 52

Largefirm

s Low 145 123 42 14 01 12 03 00 01

2 137 114 32 07 01 06 01 00 01

3 142 104 35 06 01 05 00 00 01

4 130 91 36 08 03 07 01 01 01

High 134 81 48 09 06 07 01 01 01

t Highfrac14Low 19 85 17 26 54 33 34 17 11

1450 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

Table AIV Debt and equity changesmdashwithout fixed effects

The sample contains nonfinancial firms listed on the annual Compustat files for the period from 1971

to 2009 The table presents estimates from regressions of change in debt (D) change in the book

value of equity (BVE) change in the market value of equity (MVE) and net equity issuances

(EqyIss) The explanatory variables include current and lagged change in operating income before

depreciation (Profits and Profitst1) and lagged change in book value of assets (Assetst1) The

regressions include year indicator variables The reported t-statistics are corrected for clustering at the

firm level

aMean significant at the 1 level

bMean significant at the 5 level

cMean significant at the 10 level

D BVE MVE EqyIss

(1) (2) (3) (4) (5) (6) (7) (8)

Profitst1 025a 024a 023a 038a 007 066a 007a 009a

(40) (32) (36) (39) (06) (29) (42) (51)Profits 051b 082a 149a 004a

(25) (33) (49) (33)Assetst1 002 005c 024a 001b

(08) (17) (34) (23)Constant 93 176a 45 71 1552a 1704a 130a 140a

(12) (28) (07) (09) (49) (49) (45) (45)Year FE Yes Yes Yes Yes Yes Yes Yes YesR2 001 005 001 011 001 005 002 002N 162056 162056 162130 162130 157550 157550 162154 162154

Table AIII Median profitability sorts

The sample contains nonfinancial firms listed on the annual Compustat files for the period from 1971

to 2009 The table presents median debt and equity median changes in debt and equity and median

equity issuances (EqyIss) in millions of dollars for all firms sorted on profitability and for profitability

sorts within the smallest and largest firms The sorts are done annually

D D BVE BVE MVE MVE EqyIss Assets(1) (2) (3) (4) (5) (6) (7) (8)

Allfirm

s

Low 21 00 75 17 256 20 01 1922 186 02 449 18 598 19 00 11033 816 06 1274 03 1637 05 00 32244 787 02 1541 18 2488 03 00 3556High 273 00 1361 66 3300 14 00 2578

Smallfirm

s Low 05 00 13 07 140 14 06 402 06 00 25 08 114 12 01 563 09 00 32 05 85 07 00 674 12 00 39 01 68 03 00 79High 10 00 51 04 98 01 00 90

Largefirm

s Low 11485 381 10269 371 13227 202 03 333902 13051 393 12786 140 17571 169 05 392463 11554 287 13429 24 21623 305 00 370314 9963 209 15052 194 29424 308 00 36705High 7361 00 16380 613 49142 439 00 34976

PROFITSndashLEVERAGE PUZZLEREVISITED 1451

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

References

Auerbach A J (1985) Real determinants of corporate leverage in B M Friedman (ed)

Corporate Capital Structures in the United States University of Chicago Press Chicago

IllinoisBertrand M and Mullainathan S (2003) Enjoying the quiet life corporate governance and

managerial preferences Journal of Political Economy 111 1043ndash1075Booth L Varouj A Demirguc-Kunt A and Maksimovic V (2001) Capital structures in

developing countries Journal of Finance 56 87ndash130Brennan M J and Schwartz E S (1984) Optimal financial policy and firm valuation

Journal of Finance 39 593ndash607Cameron A C and Trivedi P K (2010) Microeconometrics using Stata Stata Press TexasConstantinides G M (1979) Multiperiod consumption and investment behavior with

convex transactions costs Management Science 25 1127ndash1137Constantinides G M and Richard S F (1978) Existence of optimal simple policies for

discounted-cost inventory and cash management in continuous time Operations Research

26 620ndash636DeAngelo H DeAngelo L and Whited T M (2011) Capital structure dynamics and

transitory debt Journal of Financial Economics 99 235ndash261Dixit A K (1993) The Art of Smooth Pasting Routledge London and New YorkFama E and French K R (2002) Testing trade-off and pecking order predictions about

dividends and debt Review of Financial Studies 15 1ndash33Faulkender M and Petersen M A (2006) Does the source of capital affect capital struc-

ture Review of Financial Studies 19 45ndash79Faulkender M Flannery M J Hankins K W and Smith J M (2012) Cash Flows and

Leverage Adjustments Journal of Financial Economics 103 632ndash646Fischer E Heinkel R and Zechner J (1989) Dynamic capital structure choice theory

and tests Journal of Finance 44 19ndash40Frank M Z and Goyal V K (2008) Trade-off and pecking order theories of

debt in B E Eckbo (ed) Handbook of Corporate Finance Empirical Corporate Finance

Vol 2 Handbook of Finance Series Chapter 12 ElsevierNorth-Holland Amsterdam pp

135ndash202

Frank M Z and Goyal V K (2009) Capital structure decisions which factors are reliably

important Financial Management 38 1ndash37

Garvey G T and Hanka G (1999) Capital structure and corporate control the effect of

antitakeover statutes on firm leverage Journal of Finance 54 519ndash546

Goldstein R Ju N and Leland H (2001) An EBIT-based model of dynamic capital

structure Journal of Business 74 483ndash512

Graham J R and Tucker A L (2006) Tax shelters and corporate debt policy Journal of

Financial Economics 81 563ndash594

Hackbarth D and Mauer D C (2012) Optimal priority structure capital structure and

investment Review of Financial Studies 25 747ndash796

Halling M Yu J and Zechner J (2012) Leverage dynamics over the business cycle

Working paper University of Utah University of New South Wales and Vienna

University of Economics and Business AdministrationHennessy C A and Whited T A (2005) Debt dynamics Journal of Finance 60 1129ndash1165

Huang R and Ritter J R (2009) Testing theories of capital structure and estimating the

speed of adjustment Journal of Financial and Qualitative Analysis 44 237ndash271

1452 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

Hovakimian A Opler T and Titman S (2001) The debt-equity choice Journal of

Financial and Quantitative Analysis 36 1ndash24Jung K Kim Y and Stulz R M (1996) Timing investment opportunities managerial

discretion and the security issue decisions Journal of Financial Economics 42 159ndash185Kane A Marcus A J and McDonald R L (1984) How big is the tax advantage to

debt Journal of Finance 39 841ndash853Kayhan A and Titman S (2007) Firmsrsquo histories and their capital structures Journal of

Financial Economics 83 1ndash32Leary M T (2009) Bank loan supply lender choice and corporate capital structure

Journal of Finance 64 1143ndash1185Leary M T and Roberts M R (2005) Do firms rebalance their capital structures

Journal of Finance 60 2575ndash2619Leland H E (1994) Corporate debt value bond covenants and optimal capital structure

Journal of finance 49 1213ndash1252Lemmon M L Roberts M R and Zender J F (2008) Back to the beginning persist-

ence and the cross-section of corporate capital structure Journal of Finance 63 1575ndash1608Long M S and Malitz I B (1985) The investment-financing nexus some empirical

evidence Midland Corporate Finance Journal 3 53ndash59Mackie-Mason J K (1990) Do taxes affect corporate financing decisions Journal of

Finance 45 1471ndash1493Morellec E Nikolov B and Schurhoff N (2012) Corporate governance and capital

structure dynamics Journal of Finance 67 803ndash848Myers S C (1984) The capital structure puzzle Journal of Finance 39 575ndash592Myers S C (1993) Still searching for optimal capital structure Journal of Applied

Corporate Finance 61 4ndash14Rajan R and Zingales L (1995) What do we know about capital structure some evidence

from international data Journal of Finance 50 1421ndash1460Strebulaev I A (2007) Do tests of capital structure theory mean what they say Journal of

Finance 62 1747ndash1787

Strebulaev I A and Whited T M (2012) Dynamic Models and Structural Estimation in

Corporate Finance Now Publisher Boston and Delft

Titman S and Wessels R (1988) The determinants of capital structure choice Journal of

Finance 43 1ndash21

Welch I (2004) Capital structure and stock returns Journal of Political Economy 112

106ndash131

Welch I (2011) Two common problems in capital structure research the financial-debt-

to-asset ratio and issuing activity versus leverage changes International Review of Finance

11 1ndash17

PROFITSndashLEVERAGE PUZZLEREVISITED 1453

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

Table IX Debt and equity issues in good and bad times

The table reports the frequency and magnitude of financing activity for sorts on profits for subsamples

of firms in good and bad times An industry is defined as having ldquogood timesrdquo if the median firm in

that industry has a market-to-book ratio that is higher than the 67th percentile of the time-series

distribution of industry median market-to-book ratios Conversely an industry is defined as having

bad times if the median firm in that industry has a market-to-book ratio that is lower than the 33rd

percentile of the time-series distribution of industry median market-to-book ratios Panels A and B

report the percentage of firms (i) issuing net debt in excess of 5 of the value of their assets (ii) issuing

net equity in excess of 5 of the value of their assets (iii) issuing debt and repurchasing equity both in

excess of 5 of the value of their assets and (iv) issuing equity and retiring debt both in excess of 5

of the value of their assets Panel C tabulates the mean debt and equity levels and the mean changes in

debt and equity issuances The sample contains nonfinancial firms listed on the annual Compustat files

for the period from 1971 to 2009

Iss Iss Iss D Iss ENet D Net E Rep E Ret D(1) (2) (3) (4)

Panel A sorts on profitability

Badtimes

Low 145 131 01 202 170 49 02 083 173 46 04 074 182 47 03 10

High 177 51 06 10t Highfrac14Low 41 136 33 01

Goodtimes

Low 206 381 01 462 197 148 02 233 212 103 04 184 234 92 10 18

High 216 92 16 14t Highfrac14Low 23 658 138 23

Panel B sorts on firm size and profitability

Badtimes

Smallfirm

s Low 151 318 00 312 182 276 02 253 151 140 02 294 183 73 00 19

High 177 80 00 11t Highfrac14Low 10 90 NA 20

Largefirm

s Low 172 54 01 042 148 39 04 053 150 43 04 044 156 33 03 06

High 147 28 05 03t Highfrac14Low 17 32 20 02

Goodtimes Smallfirm

s Low 213 448 01 542 230 443 01 473 222 325 00 474 203 197 02 38

High 190 153 02 31t Highfrac14Low 19 230 10 39

Largefirm

s Low 203 96 03 122 208 66 09 133 232 54 16 064 215 39 19 05

High 234 36 38 04t Highfrac14Low 29 99 92 38

(continued)

1440 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

larger firms there is a tendency for small firms to issue more equitymdashat leastto the extent possibleDuring good times large profitable firms raise significant amounts of debt

as they experience an increase in the value of their equity Such firms alsoengage in share repurchases Debt issues by smaller firms are much lessaffected by their own profitability during good times and their change indebt is more or less independent of their profits However equity issuanceis primarily found among the low-profit small firms during good timesTable IX shows that empirically profitability has a time-varying impacton leverage choicesIn Table X the basic regressions for changes in debt and for equity

issuance are presented for both good times and bad times Comparinggood times with bad times we see that the effects of profits are muchstronger in good times In bad times the impacts are rather weak statisticallyThis difference is interesting and deserves further study However for ourpurposes recall that the static trade-off theory implies time-varying capital

Table IX Continued

D D BVE BVE MVE MVE EqyIss Assets

(1) (2) (3) (4) (5) (6) (7) (8)

Panel C magnitude of financing activity

Badtimes Smallfirm

s Low 40 03 58 14 305 06 22 140

2 44 05 82 07 240 67 17 180

3 45 04 94 02 176 18 05 187

4 43 09 98 02 137 04 03 188

High 36 10 114 13 209 16 04 199

Largefirm

s Low 3365 73 2523 169 2304 163 94 8972

2 3031 95 2983 35 3284 278 56 9068

3 2930 5 3360 14 3947 308 57 9225

4 2281 6 3817 19 5213 442 7 8902

High 1790 62 4767 13 8799 58 31 9573

Goodtimes

Smallfirm

s Low 24 06 38 05 339 23 39 90

2 26 08 48 01 301 04 25 104

3 30 09 60 04 246 03 17 125

4 28 07 66 08 206 18 12 130

High 23 06 85 16 261 13 08 141

Largefirm

s Low 4532 36 3477 103 6557 260 53 12465

2 4580 58 3527 139 6911 421 23 12522

3 2952 111 3485 2 7586 279 32 9535

4 2752 124 3939 72 10686 33 105 10226

High 2429 228 4819 377 18565 268 222 10830

PROFITSndashLEVERAGE PUZZLEREVISITED 1441

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

structure choices even when the target is time invariant Empirically we doobserve time-varying choices

8 Is Scaling Important

In the preceding section we provided results for unscaled debt and equityissues This seems appropriate to us However it involves making twochanges to the standard regression We would like to determine if both

Table X Debt and equity changes in good and bad timesmdashregressions

The sample contains nonfinancial firms listed on the annual Compustat files for the period

from 1971 to 2009 The table presents estimates from regressions of change in debt (D)and net equity issuances (EqyIss) in good and bad times The explanatory variables includechange in profits lagged change in profits and lagged change in assets The regressions

include both firm and year fixed effects An industry is defined as having ldquogood timesrdquo ifthe median firm in that industry has a market-to-book ratio that is higher than the 67thpercentile of the time-series distribution of industry median market-to-book ratios

Conversely an industry is defined as having bad times if the median firm in thatindustry has a market-to-book ratio that is lower than the 33rd percentile of the time-series distribution of industry median market-to-book ratios The reported t-statistics arecorrected for clustering at the firm level

aMean significant at the 1 level

bMean significant at the 5 level

cMean significant at the 10 level

Good times Bad times

D EqyIss D EqyIss

(1) (2) (3) (4) (5) (6) (7) (8)

Profits 087a 091a 006b 009a 060 065 003 003b

(30) (29) (20) (28) (15) (16) (16) (20)

Profitst1 029c 015a 034c 001

(19) (37) (17) (04)

Assetst1 000 002b 000 001

(02) (22) (00) (07)

Constant 1337 3989 8691a 10001a 1351 1851 4099a 4055a

(03) (09) (55) (37) (09) (13) (80) (78)

Year FE Yes Yes Yes Yes Yes Yes Yes Yes

Firm FE Yes Yes Yes Yes Yes Yes Yes Yes

R2 0071 0080 0018 0045 0178 0212 0034 0041

N 74827 68729 74891 68787 26965 26013 26971 26018

1442 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

changes are crucial Accordingly in this section we take a step back toconsider prior literature in which variables are scaled The questions arehow does scaling affect the results and whyTable XI examines alternative scalings Columns (1) and (2) scale debt and

equity issuances by the total firm net issuance Columns (3) and (4) scaledebt and equity issuances by the sum of debt and book equity Columns (5)and (6) consider the change in debt scaled by the sum of debt and bookequity Columns (7) and (8) consider the change in leverage due to externalfinancing Since we are examining changes we use the changes in thestandard regression factors as controls The primary interest is in the coef-ficient on the change in profitsWe find that a change in profits is associated with an increase in debt

issues But the sign reverses when we scale the debt issues by the sum of debtplus book equity This is telling What is really driving the results is thechange in equity in the denominator not the change in debt in the numer-ator This result is further substantiated in Columns (5) and (6) where thechange in debt to capital ratio is used instead of debt issuanceTo further explore the impact of active issuance relative to passive

experiencing of changes we decompose leverage by separating out theeffect of debt and equity issues from those due to changes in retainedearnings Thus we define leverage change due to external financing as

LevEF frac14Dt1 thorn d

Dt1 thorn Et1 thorn dthorn e

Dt1

Dt1 thorn Et1 eth1THORN

where D is debt E is book value of equity d is debt issues net of retirementsand e is equity issues net of repurchases Garvey and Hanka (1999) use asimilar measure in their study This measure directly captures the netleverage effect of debt and equity issuancesColumns (7) and (8) report results from regressions of the net leverage

effect of debt and equity issues on changes in profitability after controllingfor other leverage factors in differences and year indicator variables Theeffect of profitability on leverage change due to external financing is positiveand statistically significant This comparison suggests that changes in totaldebt and equity are only partly a result of debt and equity issuance decisionsOther balance sheet adjustments complicate the inferences from leverageratio regressionsFor most of our analysis we follow the literature in treating debt as homo-

genous It is of course potentially interesting to ask what happens when wedistinguish between short-term debt and long-term debt Which part of thefirmrsquos debt maturity structure drives the under-adjustment Table XII

PROFITSndashLEVERAGE PUZZLEREVISITED 1443

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

provides results from such decompositions Alternative scalings areconsidered The main results of interest are that the distinction betweenshort- and long-term debt is important Short-term debt issuance is largelyindependent of changes in profits The negative sign on profit change stemsfrom the action of long-term debt This seems intuitively reasonable

Table XI Scaled issuances

The sample contains nonfinancial firms listed on the annual Compustat files for the period

from 1971 to 2009 The table presents estimates from regressions of debt issuance to totalissuance (in columns (1) and (2)) debt issuance to total capitalization (in columns (3) and(4)) change in debt to capitalization ratio (in columns (5) and (6)) and change in leverage

ratio considering only external financing (in columns (7) and (8)) The change in leverageratio considering only external financing is defined as

LevEF frac14Dt1 thorn d

Dt1 thorn Et1 thorn dthorn e

Dt1

Dt1 thorn Et1

where D is debt E is book value of equity d is debt issues net of retirements and e is equity

issues net of repurchases The regressions include lagged leverage factors and year indicatorvariables Some of the specifications include firm fixed effects The reported t-statistics arecorrected for clustering at the firm level

aMean significant at the 1 level

bMean significant at the 5 level

cMean significant at the 10 level

DebtIssTotIss

DebtIssethDthornETHORN D

DthornE LevEF

(1) (2) (3) (4) (5) (6) (7) (8)

Profitt1 004a 000 002a 001c 003a 002 001b 002a

(35) (03) (34) (18) (35) (14) (23) (26)

IndLevt1 004c 001 001 001c 000 000 000 000

(19) (06) (15) (19) (02) (04) (03) (05)

Market=bookt1 001a 001a 000a 000a 000a 000a 000 000c

(47) (85) (68) (60) (55) (36) (05) (18)

Tangt1 004 005 013a 011a 011a 006a 003a 003a

(12) (15) (144) (112) (79) (38) (44) (42)

Assetst1 008a 004a 006a 004a 002a 001a 003a 002a

(97) (45) (283) (161) (53) (35) (147) (101)

Constant 079a 074a 002a 003a 000 000 000b 002a

(497) (442) (91) (111) (00) (05) (22) (93)

Year FE Yes Yes Yes Yes Yes Yes Yes Yes

Firm FE No Yes No Yes No Yes No Yes

R2 0015 0098 0022 0076 0009 0070 0012 0086

N 133905 133905 140552 140552 138229 138229 140572 140572

1444 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

inasmuch as short-term debt is probably more associated with day-to-dayoperations rather than longer term planning considerationsIn Table XIII we examine change in leverage ( D

DthornE) and change inleverage due to external financing (LEVEF) during good and bad timesseparately The traditional negative effect of a change in profits on achange in leverage is found both in good times and in bad timesHowever when we decompose the impact in order to isolate the externalcomponent as before the result disappears In good times the externalfinancing component of a change in profits is weak but still positiveDuring bad times it is not statistically significant This implies that distin-guishing between good and bad times does not fundamentally alter the basic

Table XII Short- and long-term debt issuances

The sample contains nonfinancial firms listed on the annual Compustat files for the period

from 1971 to 2009 The table presents estimates from regressions of short-term debtissuance to total issuance (in columns (1) and (2)) long-term debt issuance to totalissuance (in columns (3) and (4)) short-term debt issuance to total capitalization (in

columns (5) and (6)) and long-term debt to total capitalization (in columns (7) and (8))Some of the specifications include firm fixed effects The reported t-statistics are correctedfor clustering at the firm level

aMean significant at the 1 level

bMean significant at the 5 level

cMean significant at the 10 level

STDebtIssTotIss

LTDebtIssTotIss

STDebtIssethDthornETHORN

LTDebtIssethDthornETHORN

(1) (2) (3) (4) (5) (6) (7) (8)

Profitabilityt1 0002 0003 0042a 0008 0001 0001 0019a 0012a

(02) (04) (30) (05) (14) (10) (45) (28)

IndLevt1 0009 0007 0070b 0033 0001 0002c 0014a 0005

(06) (05) (26) (11) (08) (17) (26) (09)

M=Bt1 0003a 0003a 0003b 0010a 0000 0000 0003a 0002a

(48) (35) (22) (65) (10) (01) (84) (42)

Tangibilityt1 0020 0007 0011 0067 0005a 0006a 0097a 0081a

(12) (03) (03) (16) (34) (33) (108) (89)

Assetst1 0039a 0011b 0034a 0024b 0001a 0001 0046a 0020a

(89) (22) (38) (23) (36) (13) (190) (82)

Constant 0001b 0153a 0798a 0591a 0000c 0006a 0057a 0096a

(25) (144) (470) (288) (17) (124) (292) (234)

Year FE Yes Yes Yes Yes Yes Yes Yes Yes

Firm FE No Yes No Yes No Yes No Yes

R2 0025 0089 0018 0073 0030 0147 0016 0298

N 136146 136146 136140 136140 143031 143031 142847 142847

PROFITSndashLEVERAGE PUZZLEREVISITED 1445

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

Table XIII Externally financed leverage changes in good and bad times

The sample contains nonfinancial firms listed on the annual Compustat files for the period

from 1971 to 2009 The table presents estimates from regressions of change in debt tocapitalization ratio (in columns (1) (2) (5) and (6)) and change in leverage ratio consider-ing only external financing (in columns (3) (4) (7) and (8)) The change in leverage ratio

considering only external financing is defined as

LevEF frac14Dt1 thorn d

Dt1 thorn Et1 thorn dthorn e

Dt1

Dt1 thorn Et1

where D is debt E is book value of equity d is debt issues net of retirements and e is equityissues net of repurchases The regressions include lagged leverage factors and year indicator

variables The reported t-statistics are corrected for clustering at the firm level An industryis defined as having ldquogood timesrdquo if the median firm in that industry has a market-to-bookratio that is higher than the 67th percentile of the time-series distribution of industry

median market-to-book ratios Conversely an industry is defined as having ldquobad timesrdquoif the median firm in that industry has a market-to-book ratio that is lower than the 33rdpercentile of the time-series distribution of industry median market-to-book ratios

aMean significant at the 1 level

bMean significant at the 5 level

cMean significant at the 10 level

Good times Bad times

DDthornE LevEF D

DthornE LevEF

(1) (2) (3) (4) (5) (6) (7) (8)

Profitabiltyt1 0022c 0005 0016c 0025b 0073b 0041 0007 0007

(17) (03) (18) (24) (25) (11) (04) (03)

IndLevt1 0000 0011 0001 0009 0010 0018 0001 0000

(00) (09) (01) (13) (07) (10) (02) (00)

MB

t1

0004a 0003b 0000 0001 0008a 0001 0000 0000

(40) (24) (05) (08) (28) (04) (03) (03)

Tangt1 0097a 0041 0033a 0038b 0127a 0048 0059a 0033

(45) (16) (27) (25) (49) (14) (43) (16)

Assetst1 0014a 0013b 0024a 0019a 0037a 0023c 0036a 0025a

(30) (20) (84) (48) (45) (17) (88) (43)

Constant 0030b 0036 0000 0003 0028c 0014 0021 0001

(20) (16) (00) (07) (17) (06) (11) (01)

Year FE Yes Yes Yes Yes Yes Yes Yes Yes

Firm FE No Yes No Yes No Yes No Yes

R2 0006 0088 0009 0117 0026 0251 0023 0202

N 56820 56820 58200 58200 23462 23462 23511 23511

1446 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

conclusions The traditional leverage result on profits is driven by increasesin equity that are experienced and then partially but not completely offsetby firmsTo the extent that firms are engaging in active financing they are following

the prescriptions of the trade-off theory Thus we have a coherent picture ofhow firms are responding to profits Basically more profitable firms experiencean increase in the value of their equity They respond to this by issuing debtand repurchasing equity But the magnitude of the increase in equity is largerthan the magnitudes of the active steps taken in the debt and equity marketsThis means that the firmsrsquo actual actions are as predicted by the trade-offtheory But it leaves open the question of why they do not act more stronglyThe capital structure literature has long been interested in dynamics and

capital structure adjustment paths In recent years this interest has increas-ingly resulted in explicit dynamic optimizing models such as those ofGoldstein Ju and Leland (2001) Hennessy and Whited (2005) andStrebulaev (2007) These models commonly make specialized assumptionsleading some to question their generalityWe argue that both fixed transactions costs and variable transactions costs

are plausible and realistic in models that pin down firm size The simultaneouspresence of both is important Due to fixed costs firms only rebalance inter-mittently Due to variable adjustment costs the extent of the adjustment isreduced to the point at which an extra unit to adjustment is balanced againstthe extra unit transactions costs This means that both leverage increases andleverage reductions will only go part way toward a static optimumThe key point is that with both fixed and variable transactions costs we

should expect to only see intermittent leverage adjustments When leverage isadjusted it should ldquounder adjustrdquo when compared to a frictionless modelFurther discussion of transactions costs in a capital structure setting isprovided by Leary and Roberts (2005)

9 Conclusion

The connection between corporate profits and capital structure has beenvery influential in the assessment of the trade-off theory The conventionalinterpretation of the evidence has pushed the literature away from the trade-off theory and toward much more complex dynamic models and ideas As aresult it is important to make sure that the evidence has been correctlyinterpreted We show that the problem is with the interpretation of thetest results and a failure to take into account the costs associated with re-sponding to a profit shock

PROFITSndashLEVERAGE PUZZLEREVISITED 1447

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

We document that more profitable firms really do borrow more (not less)They tend to repurchase equity They experience an increase in both thebook value of equity and the market value of equity Less profitable firmsreally do tend to reduce their debt and to issue equityFirm size and market conditions also matter Larger firms tend to be more

active in the debt markets while smaller firms tend to be relatively moreactive in the equity markets During good times there is more use ofexternal financingThe usual profitsndashleverage puzzle result is primarily driven by the increase

in equity that is experienced by the more profitable firms Accordingly thepuzzle should be restated as asking why do firms not take sufficiently largeoffsetting actions to fully undo the change in equity What limits the mag-nitudes of the typical leverage response to profit shocksIn a frictionless model the partial response appears puzzling But there is

good empirical reason to believe that rebalancing entails both fixed andvariable costs and that firm size matters The rebalancing costs can befully avoided by doing nothing Accordingly the firm must decide whetherany given shock is big enough to be worth responding to If it is then thefirm must decide how big a response is called for These technical conditionsare known as ldquovalue matchingrdquo and ldquosmooth pastingrdquo Thus optimizationimplies that some shocks will be ignored Even if the shock is not ignoredthe optimal response will only partially undo the shock The magnitude ofthe leverage response must balance the marginal cost and the marginalbenefit of an extra unit of leverage Since the marginal cost of adjustingleverage is strictly positive the adjustment toward that static leverageoptimum will only go part way This is true both for leverage increasesand for leverage reductionsOne might conjecture that the variable costs would be time varying and

that such costs might have declined over time Since we do not have a fullaccounting of all of the variable costs it is of course impossible to be sureHowever there is some evidence suggesting that at least the directunderwriting fees are remarkably stable over time10

Surely nobody ought to be surprised that more profitable firms are moreinclined to repurchase equity while unprofitable firms tend not to do so Thefact that more profitable firms typically experience an increase in the value of

10 For IPOs Jay Ritter provides important summary evidence as httpbearwarringtonufleduritterIPOs2013Underwritingpdf He finds that for medium-sized deals a remark-

ably large fraction of the deals have a 7 fee year after year Whether this is due tobehavioral factors or some sort of a coordination device is not clear The extent towhich similar intertemporal stability carries over to other forms of external financemerits further study

1448 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

equity is equally natural Were it not for the contrary claims in the literature(ie the profitsndashleverage puzzle) we would argue that nobody ought to findit surprising that more profitable firms are more inclined to issue debt Thefact that leverage adjustments typically only partially offset profit shocksshould also sound quite natural These facts all fit together easily within thetrade-off theory once we take account of both fixed and variable costs ofactively adjusting capital structure

Appendix

Table AI Market leverage regressions

The table presents estimates of the regressions of market leverage on firm characteristics

The sample comes from the annual Compustat files during the period 1971ndash2009 Financialfirms are excluded The leverage is estimated as the ratio of debt over debt plus marketequity The explanatory variables Profitabilityt1

MB

t1

Tangibilityt1 and LnethAssetsTHORNt1are described in Table 1 IndMedianLevt1 is estimated as the median market leverage of allother firms in the same industry The industry is defined at the level of the four-digit SICcode All specifications include the year fixed effects The specifications in column (2)

additionally include the firm fixed effects We report t-statistics where the standard errorsare clustered at the firm level in parentheses

aSignificant at the 1 level

Quantile regressions

(1) (2) (3) (4) (5)

Profitabilityt1 013a 013a 007a 011a 020a

(275) (262) (349) (425) (538)

IndLevt1 046a 020a 032a 058a 063a

(506) (239) (1299) (1630) (1042)MB

t1

003a 001a 001a 002a 004a

(509) (303) (743) (709) (534)

Tangibilityt1 013a 018a 013a 014a 010a

(191) (181) (762) (571) (232)

LnethAssetsTHORNt1 001a 005a 001a 001a lt001a

(112) (258) (682) (342) (29)

Constant 010a 011a 005a 005a 029a

(171) (98) (173) (118) (398)

Year FE Yes Yes Yes Yes Yes

Firm FE No Yes No No No

Regression type OLS FE 25th tile Median 75thtile

Clustered SE Yes Yes No No No

R2 030 067

Pseudo R2 016 023 0201

N 158578 158578 158578 158578 158578

PROFITSndashLEVERAGE PUZZLEREVISITED 1449

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

Table AII Profitability sorts for firms issuing large amounts

Percentage of firms

Iss Ret Iss Iss Rep Iss Iss IssD IssE

D D NetD E E NetE DampE RepE RetD

(1) (2) (3) (4) (5) (6) (7) (8) (9)

Panel A 10 cutoff to define issuing or retiring activity

Allfirm

s

Low 205 150 134 260 05 257 45 00 23

2 207 175 103 77 06 75 11 01 10

3 243 196 101 50 08 47 08 01 09

4 252 186 113 50 13 48 08 02 08

High 227 161 118 60 27 57 07 04 09

t Highfrac14Low 70 38 62 755 226 768 320 90 152

Smallfirm

s Low 230 132 162 380 05 375 82 00 32

2 228 143 165 334 04 333 65 00 26

3 211 146 147 216 03 214 38 00 23

4 190 154 120 126 04 124 15 01 20

High 194 162 114 99 09 97 12 01 15

t Highfrac14Low 42 40 68 342 20 340 169 07 131

Largefirm

s Low 231 200 82 31 06 29 08 01 04

2 231 187 67 17 06 14 03 01 02

3 234 177 77 17 09 14 03 02 03

4 218 158 75 16 17 14 02 03 02

High 218 135 99 19 41 15 03 07 02

t Highfrac14Low 19 106 36 46 135 57 39 58 23

Percentage of firms

Iss Ret Iss Iss Rep Iss Both Iss D Iss E

D D Net D E E Net E DampE Rep E Ret D

(1) (2) (3) (4) (5) (6) (7) (8) (9)

Panel B 15 cutoff to define issuing or retiring activity

Allfirm

s

Low 158 101 100 230 01 228 27 00 15

2 142 115 63 57 01 56 05 00 06

3 165 128 57 32 01 31 03 00 05

4 173 118 65 35 01 34 03 00 05

High 160 102 70 46 04 44 03 00 05

t Highfrac14Low 08 06 141 738 87 746 277 24 131

Smallfirm

s Low 195 95 132 354 01 350 56 00 24

2 185 96 131 303 01 301 44 00 18

3 160 96 104 179 00 176 18 00 15

4 134 99 78 99 01 98 08 00 12

High 133 95 73 82 01 80 06 00 10

t Highfrac14Low 80 00 95 345 12 344 149 12 52

Largefirm

s Low 145 123 42 14 01 12 03 00 01

2 137 114 32 07 01 06 01 00 01

3 142 104 35 06 01 05 00 00 01

4 130 91 36 08 03 07 01 01 01

High 134 81 48 09 06 07 01 01 01

t Highfrac14Low 19 85 17 26 54 33 34 17 11

1450 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

Table AIV Debt and equity changesmdashwithout fixed effects

The sample contains nonfinancial firms listed on the annual Compustat files for the period from 1971

to 2009 The table presents estimates from regressions of change in debt (D) change in the book

value of equity (BVE) change in the market value of equity (MVE) and net equity issuances

(EqyIss) The explanatory variables include current and lagged change in operating income before

depreciation (Profits and Profitst1) and lagged change in book value of assets (Assetst1) The

regressions include year indicator variables The reported t-statistics are corrected for clustering at the

firm level

aMean significant at the 1 level

bMean significant at the 5 level

cMean significant at the 10 level

D BVE MVE EqyIss

(1) (2) (3) (4) (5) (6) (7) (8)

Profitst1 025a 024a 023a 038a 007 066a 007a 009a

(40) (32) (36) (39) (06) (29) (42) (51)Profits 051b 082a 149a 004a

(25) (33) (49) (33)Assetst1 002 005c 024a 001b

(08) (17) (34) (23)Constant 93 176a 45 71 1552a 1704a 130a 140a

(12) (28) (07) (09) (49) (49) (45) (45)Year FE Yes Yes Yes Yes Yes Yes Yes YesR2 001 005 001 011 001 005 002 002N 162056 162056 162130 162130 157550 157550 162154 162154

Table AIII Median profitability sorts

The sample contains nonfinancial firms listed on the annual Compustat files for the period from 1971

to 2009 The table presents median debt and equity median changes in debt and equity and median

equity issuances (EqyIss) in millions of dollars for all firms sorted on profitability and for profitability

sorts within the smallest and largest firms The sorts are done annually

D D BVE BVE MVE MVE EqyIss Assets(1) (2) (3) (4) (5) (6) (7) (8)

Allfirm

s

Low 21 00 75 17 256 20 01 1922 186 02 449 18 598 19 00 11033 816 06 1274 03 1637 05 00 32244 787 02 1541 18 2488 03 00 3556High 273 00 1361 66 3300 14 00 2578

Smallfirm

s Low 05 00 13 07 140 14 06 402 06 00 25 08 114 12 01 563 09 00 32 05 85 07 00 674 12 00 39 01 68 03 00 79High 10 00 51 04 98 01 00 90

Largefirm

s Low 11485 381 10269 371 13227 202 03 333902 13051 393 12786 140 17571 169 05 392463 11554 287 13429 24 21623 305 00 370314 9963 209 15052 194 29424 308 00 36705High 7361 00 16380 613 49142 439 00 34976

PROFITSndashLEVERAGE PUZZLEREVISITED 1451

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

References

Auerbach A J (1985) Real determinants of corporate leverage in B M Friedman (ed)

Corporate Capital Structures in the United States University of Chicago Press Chicago

IllinoisBertrand M and Mullainathan S (2003) Enjoying the quiet life corporate governance and

managerial preferences Journal of Political Economy 111 1043ndash1075Booth L Varouj A Demirguc-Kunt A and Maksimovic V (2001) Capital structures in

developing countries Journal of Finance 56 87ndash130Brennan M J and Schwartz E S (1984) Optimal financial policy and firm valuation

Journal of Finance 39 593ndash607Cameron A C and Trivedi P K (2010) Microeconometrics using Stata Stata Press TexasConstantinides G M (1979) Multiperiod consumption and investment behavior with

convex transactions costs Management Science 25 1127ndash1137Constantinides G M and Richard S F (1978) Existence of optimal simple policies for

discounted-cost inventory and cash management in continuous time Operations Research

26 620ndash636DeAngelo H DeAngelo L and Whited T M (2011) Capital structure dynamics and

transitory debt Journal of Financial Economics 99 235ndash261Dixit A K (1993) The Art of Smooth Pasting Routledge London and New YorkFama E and French K R (2002) Testing trade-off and pecking order predictions about

dividends and debt Review of Financial Studies 15 1ndash33Faulkender M and Petersen M A (2006) Does the source of capital affect capital struc-

ture Review of Financial Studies 19 45ndash79Faulkender M Flannery M J Hankins K W and Smith J M (2012) Cash Flows and

Leverage Adjustments Journal of Financial Economics 103 632ndash646Fischer E Heinkel R and Zechner J (1989) Dynamic capital structure choice theory

and tests Journal of Finance 44 19ndash40Frank M Z and Goyal V K (2008) Trade-off and pecking order theories of

debt in B E Eckbo (ed) Handbook of Corporate Finance Empirical Corporate Finance

Vol 2 Handbook of Finance Series Chapter 12 ElsevierNorth-Holland Amsterdam pp

135ndash202

Frank M Z and Goyal V K (2009) Capital structure decisions which factors are reliably

important Financial Management 38 1ndash37

Garvey G T and Hanka G (1999) Capital structure and corporate control the effect of

antitakeover statutes on firm leverage Journal of Finance 54 519ndash546

Goldstein R Ju N and Leland H (2001) An EBIT-based model of dynamic capital

structure Journal of Business 74 483ndash512

Graham J R and Tucker A L (2006) Tax shelters and corporate debt policy Journal of

Financial Economics 81 563ndash594

Hackbarth D and Mauer D C (2012) Optimal priority structure capital structure and

investment Review of Financial Studies 25 747ndash796

Halling M Yu J and Zechner J (2012) Leverage dynamics over the business cycle

Working paper University of Utah University of New South Wales and Vienna

University of Economics and Business AdministrationHennessy C A and Whited T A (2005) Debt dynamics Journal of Finance 60 1129ndash1165

Huang R and Ritter J R (2009) Testing theories of capital structure and estimating the

speed of adjustment Journal of Financial and Qualitative Analysis 44 237ndash271

1452 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

Hovakimian A Opler T and Titman S (2001) The debt-equity choice Journal of

Financial and Quantitative Analysis 36 1ndash24Jung K Kim Y and Stulz R M (1996) Timing investment opportunities managerial

discretion and the security issue decisions Journal of Financial Economics 42 159ndash185Kane A Marcus A J and McDonald R L (1984) How big is the tax advantage to

debt Journal of Finance 39 841ndash853Kayhan A and Titman S (2007) Firmsrsquo histories and their capital structures Journal of

Financial Economics 83 1ndash32Leary M T (2009) Bank loan supply lender choice and corporate capital structure

Journal of Finance 64 1143ndash1185Leary M T and Roberts M R (2005) Do firms rebalance their capital structures

Journal of Finance 60 2575ndash2619Leland H E (1994) Corporate debt value bond covenants and optimal capital structure

Journal of finance 49 1213ndash1252Lemmon M L Roberts M R and Zender J F (2008) Back to the beginning persist-

ence and the cross-section of corporate capital structure Journal of Finance 63 1575ndash1608Long M S and Malitz I B (1985) The investment-financing nexus some empirical

evidence Midland Corporate Finance Journal 3 53ndash59Mackie-Mason J K (1990) Do taxes affect corporate financing decisions Journal of

Finance 45 1471ndash1493Morellec E Nikolov B and Schurhoff N (2012) Corporate governance and capital

structure dynamics Journal of Finance 67 803ndash848Myers S C (1984) The capital structure puzzle Journal of Finance 39 575ndash592Myers S C (1993) Still searching for optimal capital structure Journal of Applied

Corporate Finance 61 4ndash14Rajan R and Zingales L (1995) What do we know about capital structure some evidence

from international data Journal of Finance 50 1421ndash1460Strebulaev I A (2007) Do tests of capital structure theory mean what they say Journal of

Finance 62 1747ndash1787

Strebulaev I A and Whited T M (2012) Dynamic Models and Structural Estimation in

Corporate Finance Now Publisher Boston and Delft

Titman S and Wessels R (1988) The determinants of capital structure choice Journal of

Finance 43 1ndash21

Welch I (2004) Capital structure and stock returns Journal of Political Economy 112

106ndash131

Welch I (2011) Two common problems in capital structure research the financial-debt-

to-asset ratio and issuing activity versus leverage changes International Review of Finance

11 1ndash17

PROFITSndashLEVERAGE PUZZLEREVISITED 1453

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

larger firms there is a tendency for small firms to issue more equitymdashat leastto the extent possibleDuring good times large profitable firms raise significant amounts of debt

as they experience an increase in the value of their equity Such firms alsoengage in share repurchases Debt issues by smaller firms are much lessaffected by their own profitability during good times and their change indebt is more or less independent of their profits However equity issuanceis primarily found among the low-profit small firms during good timesTable IX shows that empirically profitability has a time-varying impacton leverage choicesIn Table X the basic regressions for changes in debt and for equity

issuance are presented for both good times and bad times Comparinggood times with bad times we see that the effects of profits are muchstronger in good times In bad times the impacts are rather weak statisticallyThis difference is interesting and deserves further study However for ourpurposes recall that the static trade-off theory implies time-varying capital

Table IX Continued

D D BVE BVE MVE MVE EqyIss Assets

(1) (2) (3) (4) (5) (6) (7) (8)

Panel C magnitude of financing activity

Badtimes Smallfirm

s Low 40 03 58 14 305 06 22 140

2 44 05 82 07 240 67 17 180

3 45 04 94 02 176 18 05 187

4 43 09 98 02 137 04 03 188

High 36 10 114 13 209 16 04 199

Largefirm

s Low 3365 73 2523 169 2304 163 94 8972

2 3031 95 2983 35 3284 278 56 9068

3 2930 5 3360 14 3947 308 57 9225

4 2281 6 3817 19 5213 442 7 8902

High 1790 62 4767 13 8799 58 31 9573

Goodtimes

Smallfirm

s Low 24 06 38 05 339 23 39 90

2 26 08 48 01 301 04 25 104

3 30 09 60 04 246 03 17 125

4 28 07 66 08 206 18 12 130

High 23 06 85 16 261 13 08 141

Largefirm

s Low 4532 36 3477 103 6557 260 53 12465

2 4580 58 3527 139 6911 421 23 12522

3 2952 111 3485 2 7586 279 32 9535

4 2752 124 3939 72 10686 33 105 10226

High 2429 228 4819 377 18565 268 222 10830

PROFITSndashLEVERAGE PUZZLEREVISITED 1441

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

structure choices even when the target is time invariant Empirically we doobserve time-varying choices

8 Is Scaling Important

In the preceding section we provided results for unscaled debt and equityissues This seems appropriate to us However it involves making twochanges to the standard regression We would like to determine if both

Table X Debt and equity changes in good and bad timesmdashregressions

The sample contains nonfinancial firms listed on the annual Compustat files for the period

from 1971 to 2009 The table presents estimates from regressions of change in debt (D)and net equity issuances (EqyIss) in good and bad times The explanatory variables includechange in profits lagged change in profits and lagged change in assets The regressions

include both firm and year fixed effects An industry is defined as having ldquogood timesrdquo ifthe median firm in that industry has a market-to-book ratio that is higher than the 67thpercentile of the time-series distribution of industry median market-to-book ratios

Conversely an industry is defined as having bad times if the median firm in thatindustry has a market-to-book ratio that is lower than the 33rd percentile of the time-series distribution of industry median market-to-book ratios The reported t-statistics arecorrected for clustering at the firm level

aMean significant at the 1 level

bMean significant at the 5 level

cMean significant at the 10 level

Good times Bad times

D EqyIss D EqyIss

(1) (2) (3) (4) (5) (6) (7) (8)

Profits 087a 091a 006b 009a 060 065 003 003b

(30) (29) (20) (28) (15) (16) (16) (20)

Profitst1 029c 015a 034c 001

(19) (37) (17) (04)

Assetst1 000 002b 000 001

(02) (22) (00) (07)

Constant 1337 3989 8691a 10001a 1351 1851 4099a 4055a

(03) (09) (55) (37) (09) (13) (80) (78)

Year FE Yes Yes Yes Yes Yes Yes Yes Yes

Firm FE Yes Yes Yes Yes Yes Yes Yes Yes

R2 0071 0080 0018 0045 0178 0212 0034 0041

N 74827 68729 74891 68787 26965 26013 26971 26018

1442 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

changes are crucial Accordingly in this section we take a step back toconsider prior literature in which variables are scaled The questions arehow does scaling affect the results and whyTable XI examines alternative scalings Columns (1) and (2) scale debt and

equity issuances by the total firm net issuance Columns (3) and (4) scaledebt and equity issuances by the sum of debt and book equity Columns (5)and (6) consider the change in debt scaled by the sum of debt and bookequity Columns (7) and (8) consider the change in leverage due to externalfinancing Since we are examining changes we use the changes in thestandard regression factors as controls The primary interest is in the coef-ficient on the change in profitsWe find that a change in profits is associated with an increase in debt

issues But the sign reverses when we scale the debt issues by the sum of debtplus book equity This is telling What is really driving the results is thechange in equity in the denominator not the change in debt in the numer-ator This result is further substantiated in Columns (5) and (6) where thechange in debt to capital ratio is used instead of debt issuanceTo further explore the impact of active issuance relative to passive

experiencing of changes we decompose leverage by separating out theeffect of debt and equity issues from those due to changes in retainedearnings Thus we define leverage change due to external financing as

LevEF frac14Dt1 thorn d

Dt1 thorn Et1 thorn dthorn e

Dt1

Dt1 thorn Et1 eth1THORN

where D is debt E is book value of equity d is debt issues net of retirementsand e is equity issues net of repurchases Garvey and Hanka (1999) use asimilar measure in their study This measure directly captures the netleverage effect of debt and equity issuancesColumns (7) and (8) report results from regressions of the net leverage

effect of debt and equity issues on changes in profitability after controllingfor other leverage factors in differences and year indicator variables Theeffect of profitability on leverage change due to external financing is positiveand statistically significant This comparison suggests that changes in totaldebt and equity are only partly a result of debt and equity issuance decisionsOther balance sheet adjustments complicate the inferences from leverageratio regressionsFor most of our analysis we follow the literature in treating debt as homo-

genous It is of course potentially interesting to ask what happens when wedistinguish between short-term debt and long-term debt Which part of thefirmrsquos debt maturity structure drives the under-adjustment Table XII

PROFITSndashLEVERAGE PUZZLEREVISITED 1443

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

provides results from such decompositions Alternative scalings areconsidered The main results of interest are that the distinction betweenshort- and long-term debt is important Short-term debt issuance is largelyindependent of changes in profits The negative sign on profit change stemsfrom the action of long-term debt This seems intuitively reasonable

Table XI Scaled issuances

The sample contains nonfinancial firms listed on the annual Compustat files for the period

from 1971 to 2009 The table presents estimates from regressions of debt issuance to totalissuance (in columns (1) and (2)) debt issuance to total capitalization (in columns (3) and(4)) change in debt to capitalization ratio (in columns (5) and (6)) and change in leverage

ratio considering only external financing (in columns (7) and (8)) The change in leverageratio considering only external financing is defined as

LevEF frac14Dt1 thorn d

Dt1 thorn Et1 thorn dthorn e

Dt1

Dt1 thorn Et1

where D is debt E is book value of equity d is debt issues net of retirements and e is equity

issues net of repurchases The regressions include lagged leverage factors and year indicatorvariables Some of the specifications include firm fixed effects The reported t-statistics arecorrected for clustering at the firm level

aMean significant at the 1 level

bMean significant at the 5 level

cMean significant at the 10 level

DebtIssTotIss

DebtIssethDthornETHORN D

DthornE LevEF

(1) (2) (3) (4) (5) (6) (7) (8)

Profitt1 004a 000 002a 001c 003a 002 001b 002a

(35) (03) (34) (18) (35) (14) (23) (26)

IndLevt1 004c 001 001 001c 000 000 000 000

(19) (06) (15) (19) (02) (04) (03) (05)

Market=bookt1 001a 001a 000a 000a 000a 000a 000 000c

(47) (85) (68) (60) (55) (36) (05) (18)

Tangt1 004 005 013a 011a 011a 006a 003a 003a

(12) (15) (144) (112) (79) (38) (44) (42)

Assetst1 008a 004a 006a 004a 002a 001a 003a 002a

(97) (45) (283) (161) (53) (35) (147) (101)

Constant 079a 074a 002a 003a 000 000 000b 002a

(497) (442) (91) (111) (00) (05) (22) (93)

Year FE Yes Yes Yes Yes Yes Yes Yes Yes

Firm FE No Yes No Yes No Yes No Yes

R2 0015 0098 0022 0076 0009 0070 0012 0086

N 133905 133905 140552 140552 138229 138229 140572 140572

1444 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

inasmuch as short-term debt is probably more associated with day-to-dayoperations rather than longer term planning considerationsIn Table XIII we examine change in leverage ( D

DthornE) and change inleverage due to external financing (LEVEF) during good and bad timesseparately The traditional negative effect of a change in profits on achange in leverage is found both in good times and in bad timesHowever when we decompose the impact in order to isolate the externalcomponent as before the result disappears In good times the externalfinancing component of a change in profits is weak but still positiveDuring bad times it is not statistically significant This implies that distin-guishing between good and bad times does not fundamentally alter the basic

Table XII Short- and long-term debt issuances

The sample contains nonfinancial firms listed on the annual Compustat files for the period

from 1971 to 2009 The table presents estimates from regressions of short-term debtissuance to total issuance (in columns (1) and (2)) long-term debt issuance to totalissuance (in columns (3) and (4)) short-term debt issuance to total capitalization (in

columns (5) and (6)) and long-term debt to total capitalization (in columns (7) and (8))Some of the specifications include firm fixed effects The reported t-statistics are correctedfor clustering at the firm level

aMean significant at the 1 level

bMean significant at the 5 level

cMean significant at the 10 level

STDebtIssTotIss

LTDebtIssTotIss

STDebtIssethDthornETHORN

LTDebtIssethDthornETHORN

(1) (2) (3) (4) (5) (6) (7) (8)

Profitabilityt1 0002 0003 0042a 0008 0001 0001 0019a 0012a

(02) (04) (30) (05) (14) (10) (45) (28)

IndLevt1 0009 0007 0070b 0033 0001 0002c 0014a 0005

(06) (05) (26) (11) (08) (17) (26) (09)

M=Bt1 0003a 0003a 0003b 0010a 0000 0000 0003a 0002a

(48) (35) (22) (65) (10) (01) (84) (42)

Tangibilityt1 0020 0007 0011 0067 0005a 0006a 0097a 0081a

(12) (03) (03) (16) (34) (33) (108) (89)

Assetst1 0039a 0011b 0034a 0024b 0001a 0001 0046a 0020a

(89) (22) (38) (23) (36) (13) (190) (82)

Constant 0001b 0153a 0798a 0591a 0000c 0006a 0057a 0096a

(25) (144) (470) (288) (17) (124) (292) (234)

Year FE Yes Yes Yes Yes Yes Yes Yes Yes

Firm FE No Yes No Yes No Yes No Yes

R2 0025 0089 0018 0073 0030 0147 0016 0298

N 136146 136146 136140 136140 143031 143031 142847 142847

PROFITSndashLEVERAGE PUZZLEREVISITED 1445

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

Table XIII Externally financed leverage changes in good and bad times

The sample contains nonfinancial firms listed on the annual Compustat files for the period

from 1971 to 2009 The table presents estimates from regressions of change in debt tocapitalization ratio (in columns (1) (2) (5) and (6)) and change in leverage ratio consider-ing only external financing (in columns (3) (4) (7) and (8)) The change in leverage ratio

considering only external financing is defined as

LevEF frac14Dt1 thorn d

Dt1 thorn Et1 thorn dthorn e

Dt1

Dt1 thorn Et1

where D is debt E is book value of equity d is debt issues net of retirements and e is equityissues net of repurchases The regressions include lagged leverage factors and year indicator

variables The reported t-statistics are corrected for clustering at the firm level An industryis defined as having ldquogood timesrdquo if the median firm in that industry has a market-to-bookratio that is higher than the 67th percentile of the time-series distribution of industry

median market-to-book ratios Conversely an industry is defined as having ldquobad timesrdquoif the median firm in that industry has a market-to-book ratio that is lower than the 33rdpercentile of the time-series distribution of industry median market-to-book ratios

aMean significant at the 1 level

bMean significant at the 5 level

cMean significant at the 10 level

Good times Bad times

DDthornE LevEF D

DthornE LevEF

(1) (2) (3) (4) (5) (6) (7) (8)

Profitabiltyt1 0022c 0005 0016c 0025b 0073b 0041 0007 0007

(17) (03) (18) (24) (25) (11) (04) (03)

IndLevt1 0000 0011 0001 0009 0010 0018 0001 0000

(00) (09) (01) (13) (07) (10) (02) (00)

MB

t1

0004a 0003b 0000 0001 0008a 0001 0000 0000

(40) (24) (05) (08) (28) (04) (03) (03)

Tangt1 0097a 0041 0033a 0038b 0127a 0048 0059a 0033

(45) (16) (27) (25) (49) (14) (43) (16)

Assetst1 0014a 0013b 0024a 0019a 0037a 0023c 0036a 0025a

(30) (20) (84) (48) (45) (17) (88) (43)

Constant 0030b 0036 0000 0003 0028c 0014 0021 0001

(20) (16) (00) (07) (17) (06) (11) (01)

Year FE Yes Yes Yes Yes Yes Yes Yes Yes

Firm FE No Yes No Yes No Yes No Yes

R2 0006 0088 0009 0117 0026 0251 0023 0202

N 56820 56820 58200 58200 23462 23462 23511 23511

1446 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

conclusions The traditional leverage result on profits is driven by increasesin equity that are experienced and then partially but not completely offsetby firmsTo the extent that firms are engaging in active financing they are following

the prescriptions of the trade-off theory Thus we have a coherent picture ofhow firms are responding to profits Basically more profitable firms experiencean increase in the value of their equity They respond to this by issuing debtand repurchasing equity But the magnitude of the increase in equity is largerthan the magnitudes of the active steps taken in the debt and equity marketsThis means that the firmsrsquo actual actions are as predicted by the trade-offtheory But it leaves open the question of why they do not act more stronglyThe capital structure literature has long been interested in dynamics and

capital structure adjustment paths In recent years this interest has increas-ingly resulted in explicit dynamic optimizing models such as those ofGoldstein Ju and Leland (2001) Hennessy and Whited (2005) andStrebulaev (2007) These models commonly make specialized assumptionsleading some to question their generalityWe argue that both fixed transactions costs and variable transactions costs

are plausible and realistic in models that pin down firm size The simultaneouspresence of both is important Due to fixed costs firms only rebalance inter-mittently Due to variable adjustment costs the extent of the adjustment isreduced to the point at which an extra unit to adjustment is balanced againstthe extra unit transactions costs This means that both leverage increases andleverage reductions will only go part way toward a static optimumThe key point is that with both fixed and variable transactions costs we

should expect to only see intermittent leverage adjustments When leverage isadjusted it should ldquounder adjustrdquo when compared to a frictionless modelFurther discussion of transactions costs in a capital structure setting isprovided by Leary and Roberts (2005)

9 Conclusion

The connection between corporate profits and capital structure has beenvery influential in the assessment of the trade-off theory The conventionalinterpretation of the evidence has pushed the literature away from the trade-off theory and toward much more complex dynamic models and ideas As aresult it is important to make sure that the evidence has been correctlyinterpreted We show that the problem is with the interpretation of thetest results and a failure to take into account the costs associated with re-sponding to a profit shock

PROFITSndashLEVERAGE PUZZLEREVISITED 1447

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

We document that more profitable firms really do borrow more (not less)They tend to repurchase equity They experience an increase in both thebook value of equity and the market value of equity Less profitable firmsreally do tend to reduce their debt and to issue equityFirm size and market conditions also matter Larger firms tend to be more

active in the debt markets while smaller firms tend to be relatively moreactive in the equity markets During good times there is more use ofexternal financingThe usual profitsndashleverage puzzle result is primarily driven by the increase

in equity that is experienced by the more profitable firms Accordingly thepuzzle should be restated as asking why do firms not take sufficiently largeoffsetting actions to fully undo the change in equity What limits the mag-nitudes of the typical leverage response to profit shocksIn a frictionless model the partial response appears puzzling But there is

good empirical reason to believe that rebalancing entails both fixed andvariable costs and that firm size matters The rebalancing costs can befully avoided by doing nothing Accordingly the firm must decide whetherany given shock is big enough to be worth responding to If it is then thefirm must decide how big a response is called for These technical conditionsare known as ldquovalue matchingrdquo and ldquosmooth pastingrdquo Thus optimizationimplies that some shocks will be ignored Even if the shock is not ignoredthe optimal response will only partially undo the shock The magnitude ofthe leverage response must balance the marginal cost and the marginalbenefit of an extra unit of leverage Since the marginal cost of adjustingleverage is strictly positive the adjustment toward that static leverageoptimum will only go part way This is true both for leverage increasesand for leverage reductionsOne might conjecture that the variable costs would be time varying and

that such costs might have declined over time Since we do not have a fullaccounting of all of the variable costs it is of course impossible to be sureHowever there is some evidence suggesting that at least the directunderwriting fees are remarkably stable over time10

Surely nobody ought to be surprised that more profitable firms are moreinclined to repurchase equity while unprofitable firms tend not to do so Thefact that more profitable firms typically experience an increase in the value of

10 For IPOs Jay Ritter provides important summary evidence as httpbearwarringtonufleduritterIPOs2013Underwritingpdf He finds that for medium-sized deals a remark-

ably large fraction of the deals have a 7 fee year after year Whether this is due tobehavioral factors or some sort of a coordination device is not clear The extent towhich similar intertemporal stability carries over to other forms of external financemerits further study

1448 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

equity is equally natural Were it not for the contrary claims in the literature(ie the profitsndashleverage puzzle) we would argue that nobody ought to findit surprising that more profitable firms are more inclined to issue debt Thefact that leverage adjustments typically only partially offset profit shocksshould also sound quite natural These facts all fit together easily within thetrade-off theory once we take account of both fixed and variable costs ofactively adjusting capital structure

Appendix

Table AI Market leverage regressions

The table presents estimates of the regressions of market leverage on firm characteristics

The sample comes from the annual Compustat files during the period 1971ndash2009 Financialfirms are excluded The leverage is estimated as the ratio of debt over debt plus marketequity The explanatory variables Profitabilityt1

MB

t1

Tangibilityt1 and LnethAssetsTHORNt1are described in Table 1 IndMedianLevt1 is estimated as the median market leverage of allother firms in the same industry The industry is defined at the level of the four-digit SICcode All specifications include the year fixed effects The specifications in column (2)

additionally include the firm fixed effects We report t-statistics where the standard errorsare clustered at the firm level in parentheses

aSignificant at the 1 level

Quantile regressions

(1) (2) (3) (4) (5)

Profitabilityt1 013a 013a 007a 011a 020a

(275) (262) (349) (425) (538)

IndLevt1 046a 020a 032a 058a 063a

(506) (239) (1299) (1630) (1042)MB

t1

003a 001a 001a 002a 004a

(509) (303) (743) (709) (534)

Tangibilityt1 013a 018a 013a 014a 010a

(191) (181) (762) (571) (232)

LnethAssetsTHORNt1 001a 005a 001a 001a lt001a

(112) (258) (682) (342) (29)

Constant 010a 011a 005a 005a 029a

(171) (98) (173) (118) (398)

Year FE Yes Yes Yes Yes Yes

Firm FE No Yes No No No

Regression type OLS FE 25th tile Median 75thtile

Clustered SE Yes Yes No No No

R2 030 067

Pseudo R2 016 023 0201

N 158578 158578 158578 158578 158578

PROFITSndashLEVERAGE PUZZLEREVISITED 1449

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

Table AII Profitability sorts for firms issuing large amounts

Percentage of firms

Iss Ret Iss Iss Rep Iss Iss IssD IssE

D D NetD E E NetE DampE RepE RetD

(1) (2) (3) (4) (5) (6) (7) (8) (9)

Panel A 10 cutoff to define issuing or retiring activity

Allfirm

s

Low 205 150 134 260 05 257 45 00 23

2 207 175 103 77 06 75 11 01 10

3 243 196 101 50 08 47 08 01 09

4 252 186 113 50 13 48 08 02 08

High 227 161 118 60 27 57 07 04 09

t Highfrac14Low 70 38 62 755 226 768 320 90 152

Smallfirm

s Low 230 132 162 380 05 375 82 00 32

2 228 143 165 334 04 333 65 00 26

3 211 146 147 216 03 214 38 00 23

4 190 154 120 126 04 124 15 01 20

High 194 162 114 99 09 97 12 01 15

t Highfrac14Low 42 40 68 342 20 340 169 07 131

Largefirm

s Low 231 200 82 31 06 29 08 01 04

2 231 187 67 17 06 14 03 01 02

3 234 177 77 17 09 14 03 02 03

4 218 158 75 16 17 14 02 03 02

High 218 135 99 19 41 15 03 07 02

t Highfrac14Low 19 106 36 46 135 57 39 58 23

Percentage of firms

Iss Ret Iss Iss Rep Iss Both Iss D Iss E

D D Net D E E Net E DampE Rep E Ret D

(1) (2) (3) (4) (5) (6) (7) (8) (9)

Panel B 15 cutoff to define issuing or retiring activity

Allfirm

s

Low 158 101 100 230 01 228 27 00 15

2 142 115 63 57 01 56 05 00 06

3 165 128 57 32 01 31 03 00 05

4 173 118 65 35 01 34 03 00 05

High 160 102 70 46 04 44 03 00 05

t Highfrac14Low 08 06 141 738 87 746 277 24 131

Smallfirm

s Low 195 95 132 354 01 350 56 00 24

2 185 96 131 303 01 301 44 00 18

3 160 96 104 179 00 176 18 00 15

4 134 99 78 99 01 98 08 00 12

High 133 95 73 82 01 80 06 00 10

t Highfrac14Low 80 00 95 345 12 344 149 12 52

Largefirm

s Low 145 123 42 14 01 12 03 00 01

2 137 114 32 07 01 06 01 00 01

3 142 104 35 06 01 05 00 00 01

4 130 91 36 08 03 07 01 01 01

High 134 81 48 09 06 07 01 01 01

t Highfrac14Low 19 85 17 26 54 33 34 17 11

1450 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

Table AIV Debt and equity changesmdashwithout fixed effects

The sample contains nonfinancial firms listed on the annual Compustat files for the period from 1971

to 2009 The table presents estimates from regressions of change in debt (D) change in the book

value of equity (BVE) change in the market value of equity (MVE) and net equity issuances

(EqyIss) The explanatory variables include current and lagged change in operating income before

depreciation (Profits and Profitst1) and lagged change in book value of assets (Assetst1) The

regressions include year indicator variables The reported t-statistics are corrected for clustering at the

firm level

aMean significant at the 1 level

bMean significant at the 5 level

cMean significant at the 10 level

D BVE MVE EqyIss

(1) (2) (3) (4) (5) (6) (7) (8)

Profitst1 025a 024a 023a 038a 007 066a 007a 009a

(40) (32) (36) (39) (06) (29) (42) (51)Profits 051b 082a 149a 004a

(25) (33) (49) (33)Assetst1 002 005c 024a 001b

(08) (17) (34) (23)Constant 93 176a 45 71 1552a 1704a 130a 140a

(12) (28) (07) (09) (49) (49) (45) (45)Year FE Yes Yes Yes Yes Yes Yes Yes YesR2 001 005 001 011 001 005 002 002N 162056 162056 162130 162130 157550 157550 162154 162154

Table AIII Median profitability sorts

The sample contains nonfinancial firms listed on the annual Compustat files for the period from 1971

to 2009 The table presents median debt and equity median changes in debt and equity and median

equity issuances (EqyIss) in millions of dollars for all firms sorted on profitability and for profitability

sorts within the smallest and largest firms The sorts are done annually

D D BVE BVE MVE MVE EqyIss Assets(1) (2) (3) (4) (5) (6) (7) (8)

Allfirm

s

Low 21 00 75 17 256 20 01 1922 186 02 449 18 598 19 00 11033 816 06 1274 03 1637 05 00 32244 787 02 1541 18 2488 03 00 3556High 273 00 1361 66 3300 14 00 2578

Smallfirm

s Low 05 00 13 07 140 14 06 402 06 00 25 08 114 12 01 563 09 00 32 05 85 07 00 674 12 00 39 01 68 03 00 79High 10 00 51 04 98 01 00 90

Largefirm

s Low 11485 381 10269 371 13227 202 03 333902 13051 393 12786 140 17571 169 05 392463 11554 287 13429 24 21623 305 00 370314 9963 209 15052 194 29424 308 00 36705High 7361 00 16380 613 49142 439 00 34976

PROFITSndashLEVERAGE PUZZLEREVISITED 1451

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

References

Auerbach A J (1985) Real determinants of corporate leverage in B M Friedman (ed)

Corporate Capital Structures in the United States University of Chicago Press Chicago

IllinoisBertrand M and Mullainathan S (2003) Enjoying the quiet life corporate governance and

managerial preferences Journal of Political Economy 111 1043ndash1075Booth L Varouj A Demirguc-Kunt A and Maksimovic V (2001) Capital structures in

developing countries Journal of Finance 56 87ndash130Brennan M J and Schwartz E S (1984) Optimal financial policy and firm valuation

Journal of Finance 39 593ndash607Cameron A C and Trivedi P K (2010) Microeconometrics using Stata Stata Press TexasConstantinides G M (1979) Multiperiod consumption and investment behavior with

convex transactions costs Management Science 25 1127ndash1137Constantinides G M and Richard S F (1978) Existence of optimal simple policies for

discounted-cost inventory and cash management in continuous time Operations Research

26 620ndash636DeAngelo H DeAngelo L and Whited T M (2011) Capital structure dynamics and

transitory debt Journal of Financial Economics 99 235ndash261Dixit A K (1993) The Art of Smooth Pasting Routledge London and New YorkFama E and French K R (2002) Testing trade-off and pecking order predictions about

dividends and debt Review of Financial Studies 15 1ndash33Faulkender M and Petersen M A (2006) Does the source of capital affect capital struc-

ture Review of Financial Studies 19 45ndash79Faulkender M Flannery M J Hankins K W and Smith J M (2012) Cash Flows and

Leverage Adjustments Journal of Financial Economics 103 632ndash646Fischer E Heinkel R and Zechner J (1989) Dynamic capital structure choice theory

and tests Journal of Finance 44 19ndash40Frank M Z and Goyal V K (2008) Trade-off and pecking order theories of

debt in B E Eckbo (ed) Handbook of Corporate Finance Empirical Corporate Finance

Vol 2 Handbook of Finance Series Chapter 12 ElsevierNorth-Holland Amsterdam pp

135ndash202

Frank M Z and Goyal V K (2009) Capital structure decisions which factors are reliably

important Financial Management 38 1ndash37

Garvey G T and Hanka G (1999) Capital structure and corporate control the effect of

antitakeover statutes on firm leverage Journal of Finance 54 519ndash546

Goldstein R Ju N and Leland H (2001) An EBIT-based model of dynamic capital

structure Journal of Business 74 483ndash512

Graham J R and Tucker A L (2006) Tax shelters and corporate debt policy Journal of

Financial Economics 81 563ndash594

Hackbarth D and Mauer D C (2012) Optimal priority structure capital structure and

investment Review of Financial Studies 25 747ndash796

Halling M Yu J and Zechner J (2012) Leverage dynamics over the business cycle

Working paper University of Utah University of New South Wales and Vienna

University of Economics and Business AdministrationHennessy C A and Whited T A (2005) Debt dynamics Journal of Finance 60 1129ndash1165

Huang R and Ritter J R (2009) Testing theories of capital structure and estimating the

speed of adjustment Journal of Financial and Qualitative Analysis 44 237ndash271

1452 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

Hovakimian A Opler T and Titman S (2001) The debt-equity choice Journal of

Financial and Quantitative Analysis 36 1ndash24Jung K Kim Y and Stulz R M (1996) Timing investment opportunities managerial

discretion and the security issue decisions Journal of Financial Economics 42 159ndash185Kane A Marcus A J and McDonald R L (1984) How big is the tax advantage to

debt Journal of Finance 39 841ndash853Kayhan A and Titman S (2007) Firmsrsquo histories and their capital structures Journal of

Financial Economics 83 1ndash32Leary M T (2009) Bank loan supply lender choice and corporate capital structure

Journal of Finance 64 1143ndash1185Leary M T and Roberts M R (2005) Do firms rebalance their capital structures

Journal of Finance 60 2575ndash2619Leland H E (1994) Corporate debt value bond covenants and optimal capital structure

Journal of finance 49 1213ndash1252Lemmon M L Roberts M R and Zender J F (2008) Back to the beginning persist-

ence and the cross-section of corporate capital structure Journal of Finance 63 1575ndash1608Long M S and Malitz I B (1985) The investment-financing nexus some empirical

evidence Midland Corporate Finance Journal 3 53ndash59Mackie-Mason J K (1990) Do taxes affect corporate financing decisions Journal of

Finance 45 1471ndash1493Morellec E Nikolov B and Schurhoff N (2012) Corporate governance and capital

structure dynamics Journal of Finance 67 803ndash848Myers S C (1984) The capital structure puzzle Journal of Finance 39 575ndash592Myers S C (1993) Still searching for optimal capital structure Journal of Applied

Corporate Finance 61 4ndash14Rajan R and Zingales L (1995) What do we know about capital structure some evidence

from international data Journal of Finance 50 1421ndash1460Strebulaev I A (2007) Do tests of capital structure theory mean what they say Journal of

Finance 62 1747ndash1787

Strebulaev I A and Whited T M (2012) Dynamic Models and Structural Estimation in

Corporate Finance Now Publisher Boston and Delft

Titman S and Wessels R (1988) The determinants of capital structure choice Journal of

Finance 43 1ndash21

Welch I (2004) Capital structure and stock returns Journal of Political Economy 112

106ndash131

Welch I (2011) Two common problems in capital structure research the financial-debt-

to-asset ratio and issuing activity versus leverage changes International Review of Finance

11 1ndash17

PROFITSndashLEVERAGE PUZZLEREVISITED 1453

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

structure choices even when the target is time invariant Empirically we doobserve time-varying choices

8 Is Scaling Important

In the preceding section we provided results for unscaled debt and equityissues This seems appropriate to us However it involves making twochanges to the standard regression We would like to determine if both

Table X Debt and equity changes in good and bad timesmdashregressions

The sample contains nonfinancial firms listed on the annual Compustat files for the period

from 1971 to 2009 The table presents estimates from regressions of change in debt (D)and net equity issuances (EqyIss) in good and bad times The explanatory variables includechange in profits lagged change in profits and lagged change in assets The regressions

include both firm and year fixed effects An industry is defined as having ldquogood timesrdquo ifthe median firm in that industry has a market-to-book ratio that is higher than the 67thpercentile of the time-series distribution of industry median market-to-book ratios

Conversely an industry is defined as having bad times if the median firm in thatindustry has a market-to-book ratio that is lower than the 33rd percentile of the time-series distribution of industry median market-to-book ratios The reported t-statistics arecorrected for clustering at the firm level

aMean significant at the 1 level

bMean significant at the 5 level

cMean significant at the 10 level

Good times Bad times

D EqyIss D EqyIss

(1) (2) (3) (4) (5) (6) (7) (8)

Profits 087a 091a 006b 009a 060 065 003 003b

(30) (29) (20) (28) (15) (16) (16) (20)

Profitst1 029c 015a 034c 001

(19) (37) (17) (04)

Assetst1 000 002b 000 001

(02) (22) (00) (07)

Constant 1337 3989 8691a 10001a 1351 1851 4099a 4055a

(03) (09) (55) (37) (09) (13) (80) (78)

Year FE Yes Yes Yes Yes Yes Yes Yes Yes

Firm FE Yes Yes Yes Yes Yes Yes Yes Yes

R2 0071 0080 0018 0045 0178 0212 0034 0041

N 74827 68729 74891 68787 26965 26013 26971 26018

1442 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

changes are crucial Accordingly in this section we take a step back toconsider prior literature in which variables are scaled The questions arehow does scaling affect the results and whyTable XI examines alternative scalings Columns (1) and (2) scale debt and

equity issuances by the total firm net issuance Columns (3) and (4) scaledebt and equity issuances by the sum of debt and book equity Columns (5)and (6) consider the change in debt scaled by the sum of debt and bookequity Columns (7) and (8) consider the change in leverage due to externalfinancing Since we are examining changes we use the changes in thestandard regression factors as controls The primary interest is in the coef-ficient on the change in profitsWe find that a change in profits is associated with an increase in debt

issues But the sign reverses when we scale the debt issues by the sum of debtplus book equity This is telling What is really driving the results is thechange in equity in the denominator not the change in debt in the numer-ator This result is further substantiated in Columns (5) and (6) where thechange in debt to capital ratio is used instead of debt issuanceTo further explore the impact of active issuance relative to passive

experiencing of changes we decompose leverage by separating out theeffect of debt and equity issues from those due to changes in retainedearnings Thus we define leverage change due to external financing as

LevEF frac14Dt1 thorn d

Dt1 thorn Et1 thorn dthorn e

Dt1

Dt1 thorn Et1 eth1THORN

where D is debt E is book value of equity d is debt issues net of retirementsand e is equity issues net of repurchases Garvey and Hanka (1999) use asimilar measure in their study This measure directly captures the netleverage effect of debt and equity issuancesColumns (7) and (8) report results from regressions of the net leverage

effect of debt and equity issues on changes in profitability after controllingfor other leverage factors in differences and year indicator variables Theeffect of profitability on leverage change due to external financing is positiveand statistically significant This comparison suggests that changes in totaldebt and equity are only partly a result of debt and equity issuance decisionsOther balance sheet adjustments complicate the inferences from leverageratio regressionsFor most of our analysis we follow the literature in treating debt as homo-

genous It is of course potentially interesting to ask what happens when wedistinguish between short-term debt and long-term debt Which part of thefirmrsquos debt maturity structure drives the under-adjustment Table XII

PROFITSndashLEVERAGE PUZZLEREVISITED 1443

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

provides results from such decompositions Alternative scalings areconsidered The main results of interest are that the distinction betweenshort- and long-term debt is important Short-term debt issuance is largelyindependent of changes in profits The negative sign on profit change stemsfrom the action of long-term debt This seems intuitively reasonable

Table XI Scaled issuances

The sample contains nonfinancial firms listed on the annual Compustat files for the period

from 1971 to 2009 The table presents estimates from regressions of debt issuance to totalissuance (in columns (1) and (2)) debt issuance to total capitalization (in columns (3) and(4)) change in debt to capitalization ratio (in columns (5) and (6)) and change in leverage

ratio considering only external financing (in columns (7) and (8)) The change in leverageratio considering only external financing is defined as

LevEF frac14Dt1 thorn d

Dt1 thorn Et1 thorn dthorn e

Dt1

Dt1 thorn Et1

where D is debt E is book value of equity d is debt issues net of retirements and e is equity

issues net of repurchases The regressions include lagged leverage factors and year indicatorvariables Some of the specifications include firm fixed effects The reported t-statistics arecorrected for clustering at the firm level

aMean significant at the 1 level

bMean significant at the 5 level

cMean significant at the 10 level

DebtIssTotIss

DebtIssethDthornETHORN D

DthornE LevEF

(1) (2) (3) (4) (5) (6) (7) (8)

Profitt1 004a 000 002a 001c 003a 002 001b 002a

(35) (03) (34) (18) (35) (14) (23) (26)

IndLevt1 004c 001 001 001c 000 000 000 000

(19) (06) (15) (19) (02) (04) (03) (05)

Market=bookt1 001a 001a 000a 000a 000a 000a 000 000c

(47) (85) (68) (60) (55) (36) (05) (18)

Tangt1 004 005 013a 011a 011a 006a 003a 003a

(12) (15) (144) (112) (79) (38) (44) (42)

Assetst1 008a 004a 006a 004a 002a 001a 003a 002a

(97) (45) (283) (161) (53) (35) (147) (101)

Constant 079a 074a 002a 003a 000 000 000b 002a

(497) (442) (91) (111) (00) (05) (22) (93)

Year FE Yes Yes Yes Yes Yes Yes Yes Yes

Firm FE No Yes No Yes No Yes No Yes

R2 0015 0098 0022 0076 0009 0070 0012 0086

N 133905 133905 140552 140552 138229 138229 140572 140572

1444 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

inasmuch as short-term debt is probably more associated with day-to-dayoperations rather than longer term planning considerationsIn Table XIII we examine change in leverage ( D

DthornE) and change inleverage due to external financing (LEVEF) during good and bad timesseparately The traditional negative effect of a change in profits on achange in leverage is found both in good times and in bad timesHowever when we decompose the impact in order to isolate the externalcomponent as before the result disappears In good times the externalfinancing component of a change in profits is weak but still positiveDuring bad times it is not statistically significant This implies that distin-guishing between good and bad times does not fundamentally alter the basic

Table XII Short- and long-term debt issuances

The sample contains nonfinancial firms listed on the annual Compustat files for the period

from 1971 to 2009 The table presents estimates from regressions of short-term debtissuance to total issuance (in columns (1) and (2)) long-term debt issuance to totalissuance (in columns (3) and (4)) short-term debt issuance to total capitalization (in

columns (5) and (6)) and long-term debt to total capitalization (in columns (7) and (8))Some of the specifications include firm fixed effects The reported t-statistics are correctedfor clustering at the firm level

aMean significant at the 1 level

bMean significant at the 5 level

cMean significant at the 10 level

STDebtIssTotIss

LTDebtIssTotIss

STDebtIssethDthornETHORN

LTDebtIssethDthornETHORN

(1) (2) (3) (4) (5) (6) (7) (8)

Profitabilityt1 0002 0003 0042a 0008 0001 0001 0019a 0012a

(02) (04) (30) (05) (14) (10) (45) (28)

IndLevt1 0009 0007 0070b 0033 0001 0002c 0014a 0005

(06) (05) (26) (11) (08) (17) (26) (09)

M=Bt1 0003a 0003a 0003b 0010a 0000 0000 0003a 0002a

(48) (35) (22) (65) (10) (01) (84) (42)

Tangibilityt1 0020 0007 0011 0067 0005a 0006a 0097a 0081a

(12) (03) (03) (16) (34) (33) (108) (89)

Assetst1 0039a 0011b 0034a 0024b 0001a 0001 0046a 0020a

(89) (22) (38) (23) (36) (13) (190) (82)

Constant 0001b 0153a 0798a 0591a 0000c 0006a 0057a 0096a

(25) (144) (470) (288) (17) (124) (292) (234)

Year FE Yes Yes Yes Yes Yes Yes Yes Yes

Firm FE No Yes No Yes No Yes No Yes

R2 0025 0089 0018 0073 0030 0147 0016 0298

N 136146 136146 136140 136140 143031 143031 142847 142847

PROFITSndashLEVERAGE PUZZLEREVISITED 1445

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

Table XIII Externally financed leverage changes in good and bad times

The sample contains nonfinancial firms listed on the annual Compustat files for the period

from 1971 to 2009 The table presents estimates from regressions of change in debt tocapitalization ratio (in columns (1) (2) (5) and (6)) and change in leverage ratio consider-ing only external financing (in columns (3) (4) (7) and (8)) The change in leverage ratio

considering only external financing is defined as

LevEF frac14Dt1 thorn d

Dt1 thorn Et1 thorn dthorn e

Dt1

Dt1 thorn Et1

where D is debt E is book value of equity d is debt issues net of retirements and e is equityissues net of repurchases The regressions include lagged leverage factors and year indicator

variables The reported t-statistics are corrected for clustering at the firm level An industryis defined as having ldquogood timesrdquo if the median firm in that industry has a market-to-bookratio that is higher than the 67th percentile of the time-series distribution of industry

median market-to-book ratios Conversely an industry is defined as having ldquobad timesrdquoif the median firm in that industry has a market-to-book ratio that is lower than the 33rdpercentile of the time-series distribution of industry median market-to-book ratios

aMean significant at the 1 level

bMean significant at the 5 level

cMean significant at the 10 level

Good times Bad times

DDthornE LevEF D

DthornE LevEF

(1) (2) (3) (4) (5) (6) (7) (8)

Profitabiltyt1 0022c 0005 0016c 0025b 0073b 0041 0007 0007

(17) (03) (18) (24) (25) (11) (04) (03)

IndLevt1 0000 0011 0001 0009 0010 0018 0001 0000

(00) (09) (01) (13) (07) (10) (02) (00)

MB

t1

0004a 0003b 0000 0001 0008a 0001 0000 0000

(40) (24) (05) (08) (28) (04) (03) (03)

Tangt1 0097a 0041 0033a 0038b 0127a 0048 0059a 0033

(45) (16) (27) (25) (49) (14) (43) (16)

Assetst1 0014a 0013b 0024a 0019a 0037a 0023c 0036a 0025a

(30) (20) (84) (48) (45) (17) (88) (43)

Constant 0030b 0036 0000 0003 0028c 0014 0021 0001

(20) (16) (00) (07) (17) (06) (11) (01)

Year FE Yes Yes Yes Yes Yes Yes Yes Yes

Firm FE No Yes No Yes No Yes No Yes

R2 0006 0088 0009 0117 0026 0251 0023 0202

N 56820 56820 58200 58200 23462 23462 23511 23511

1446 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

conclusions The traditional leverage result on profits is driven by increasesin equity that are experienced and then partially but not completely offsetby firmsTo the extent that firms are engaging in active financing they are following

the prescriptions of the trade-off theory Thus we have a coherent picture ofhow firms are responding to profits Basically more profitable firms experiencean increase in the value of their equity They respond to this by issuing debtand repurchasing equity But the magnitude of the increase in equity is largerthan the magnitudes of the active steps taken in the debt and equity marketsThis means that the firmsrsquo actual actions are as predicted by the trade-offtheory But it leaves open the question of why they do not act more stronglyThe capital structure literature has long been interested in dynamics and

capital structure adjustment paths In recent years this interest has increas-ingly resulted in explicit dynamic optimizing models such as those ofGoldstein Ju and Leland (2001) Hennessy and Whited (2005) andStrebulaev (2007) These models commonly make specialized assumptionsleading some to question their generalityWe argue that both fixed transactions costs and variable transactions costs

are plausible and realistic in models that pin down firm size The simultaneouspresence of both is important Due to fixed costs firms only rebalance inter-mittently Due to variable adjustment costs the extent of the adjustment isreduced to the point at which an extra unit to adjustment is balanced againstthe extra unit transactions costs This means that both leverage increases andleverage reductions will only go part way toward a static optimumThe key point is that with both fixed and variable transactions costs we

should expect to only see intermittent leverage adjustments When leverage isadjusted it should ldquounder adjustrdquo when compared to a frictionless modelFurther discussion of transactions costs in a capital structure setting isprovided by Leary and Roberts (2005)

9 Conclusion

The connection between corporate profits and capital structure has beenvery influential in the assessment of the trade-off theory The conventionalinterpretation of the evidence has pushed the literature away from the trade-off theory and toward much more complex dynamic models and ideas As aresult it is important to make sure that the evidence has been correctlyinterpreted We show that the problem is with the interpretation of thetest results and a failure to take into account the costs associated with re-sponding to a profit shock

PROFITSndashLEVERAGE PUZZLEREVISITED 1447

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

We document that more profitable firms really do borrow more (not less)They tend to repurchase equity They experience an increase in both thebook value of equity and the market value of equity Less profitable firmsreally do tend to reduce their debt and to issue equityFirm size and market conditions also matter Larger firms tend to be more

active in the debt markets while smaller firms tend to be relatively moreactive in the equity markets During good times there is more use ofexternal financingThe usual profitsndashleverage puzzle result is primarily driven by the increase

in equity that is experienced by the more profitable firms Accordingly thepuzzle should be restated as asking why do firms not take sufficiently largeoffsetting actions to fully undo the change in equity What limits the mag-nitudes of the typical leverage response to profit shocksIn a frictionless model the partial response appears puzzling But there is

good empirical reason to believe that rebalancing entails both fixed andvariable costs and that firm size matters The rebalancing costs can befully avoided by doing nothing Accordingly the firm must decide whetherany given shock is big enough to be worth responding to If it is then thefirm must decide how big a response is called for These technical conditionsare known as ldquovalue matchingrdquo and ldquosmooth pastingrdquo Thus optimizationimplies that some shocks will be ignored Even if the shock is not ignoredthe optimal response will only partially undo the shock The magnitude ofthe leverage response must balance the marginal cost and the marginalbenefit of an extra unit of leverage Since the marginal cost of adjustingleverage is strictly positive the adjustment toward that static leverageoptimum will only go part way This is true both for leverage increasesand for leverage reductionsOne might conjecture that the variable costs would be time varying and

that such costs might have declined over time Since we do not have a fullaccounting of all of the variable costs it is of course impossible to be sureHowever there is some evidence suggesting that at least the directunderwriting fees are remarkably stable over time10

Surely nobody ought to be surprised that more profitable firms are moreinclined to repurchase equity while unprofitable firms tend not to do so Thefact that more profitable firms typically experience an increase in the value of

10 For IPOs Jay Ritter provides important summary evidence as httpbearwarringtonufleduritterIPOs2013Underwritingpdf He finds that for medium-sized deals a remark-

ably large fraction of the deals have a 7 fee year after year Whether this is due tobehavioral factors or some sort of a coordination device is not clear The extent towhich similar intertemporal stability carries over to other forms of external financemerits further study

1448 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

equity is equally natural Were it not for the contrary claims in the literature(ie the profitsndashleverage puzzle) we would argue that nobody ought to findit surprising that more profitable firms are more inclined to issue debt Thefact that leverage adjustments typically only partially offset profit shocksshould also sound quite natural These facts all fit together easily within thetrade-off theory once we take account of both fixed and variable costs ofactively adjusting capital structure

Appendix

Table AI Market leverage regressions

The table presents estimates of the regressions of market leverage on firm characteristics

The sample comes from the annual Compustat files during the period 1971ndash2009 Financialfirms are excluded The leverage is estimated as the ratio of debt over debt plus marketequity The explanatory variables Profitabilityt1

MB

t1

Tangibilityt1 and LnethAssetsTHORNt1are described in Table 1 IndMedianLevt1 is estimated as the median market leverage of allother firms in the same industry The industry is defined at the level of the four-digit SICcode All specifications include the year fixed effects The specifications in column (2)

additionally include the firm fixed effects We report t-statistics where the standard errorsare clustered at the firm level in parentheses

aSignificant at the 1 level

Quantile regressions

(1) (2) (3) (4) (5)

Profitabilityt1 013a 013a 007a 011a 020a

(275) (262) (349) (425) (538)

IndLevt1 046a 020a 032a 058a 063a

(506) (239) (1299) (1630) (1042)MB

t1

003a 001a 001a 002a 004a

(509) (303) (743) (709) (534)

Tangibilityt1 013a 018a 013a 014a 010a

(191) (181) (762) (571) (232)

LnethAssetsTHORNt1 001a 005a 001a 001a lt001a

(112) (258) (682) (342) (29)

Constant 010a 011a 005a 005a 029a

(171) (98) (173) (118) (398)

Year FE Yes Yes Yes Yes Yes

Firm FE No Yes No No No

Regression type OLS FE 25th tile Median 75thtile

Clustered SE Yes Yes No No No

R2 030 067

Pseudo R2 016 023 0201

N 158578 158578 158578 158578 158578

PROFITSndashLEVERAGE PUZZLEREVISITED 1449

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

Table AII Profitability sorts for firms issuing large amounts

Percentage of firms

Iss Ret Iss Iss Rep Iss Iss IssD IssE

D D NetD E E NetE DampE RepE RetD

(1) (2) (3) (4) (5) (6) (7) (8) (9)

Panel A 10 cutoff to define issuing or retiring activity

Allfirm

s

Low 205 150 134 260 05 257 45 00 23

2 207 175 103 77 06 75 11 01 10

3 243 196 101 50 08 47 08 01 09

4 252 186 113 50 13 48 08 02 08

High 227 161 118 60 27 57 07 04 09

t Highfrac14Low 70 38 62 755 226 768 320 90 152

Smallfirm

s Low 230 132 162 380 05 375 82 00 32

2 228 143 165 334 04 333 65 00 26

3 211 146 147 216 03 214 38 00 23

4 190 154 120 126 04 124 15 01 20

High 194 162 114 99 09 97 12 01 15

t Highfrac14Low 42 40 68 342 20 340 169 07 131

Largefirm

s Low 231 200 82 31 06 29 08 01 04

2 231 187 67 17 06 14 03 01 02

3 234 177 77 17 09 14 03 02 03

4 218 158 75 16 17 14 02 03 02

High 218 135 99 19 41 15 03 07 02

t Highfrac14Low 19 106 36 46 135 57 39 58 23

Percentage of firms

Iss Ret Iss Iss Rep Iss Both Iss D Iss E

D D Net D E E Net E DampE Rep E Ret D

(1) (2) (3) (4) (5) (6) (7) (8) (9)

Panel B 15 cutoff to define issuing or retiring activity

Allfirm

s

Low 158 101 100 230 01 228 27 00 15

2 142 115 63 57 01 56 05 00 06

3 165 128 57 32 01 31 03 00 05

4 173 118 65 35 01 34 03 00 05

High 160 102 70 46 04 44 03 00 05

t Highfrac14Low 08 06 141 738 87 746 277 24 131

Smallfirm

s Low 195 95 132 354 01 350 56 00 24

2 185 96 131 303 01 301 44 00 18

3 160 96 104 179 00 176 18 00 15

4 134 99 78 99 01 98 08 00 12

High 133 95 73 82 01 80 06 00 10

t Highfrac14Low 80 00 95 345 12 344 149 12 52

Largefirm

s Low 145 123 42 14 01 12 03 00 01

2 137 114 32 07 01 06 01 00 01

3 142 104 35 06 01 05 00 00 01

4 130 91 36 08 03 07 01 01 01

High 134 81 48 09 06 07 01 01 01

t Highfrac14Low 19 85 17 26 54 33 34 17 11

1450 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

Table AIV Debt and equity changesmdashwithout fixed effects

The sample contains nonfinancial firms listed on the annual Compustat files for the period from 1971

to 2009 The table presents estimates from regressions of change in debt (D) change in the book

value of equity (BVE) change in the market value of equity (MVE) and net equity issuances

(EqyIss) The explanatory variables include current and lagged change in operating income before

depreciation (Profits and Profitst1) and lagged change in book value of assets (Assetst1) The

regressions include year indicator variables The reported t-statistics are corrected for clustering at the

firm level

aMean significant at the 1 level

bMean significant at the 5 level

cMean significant at the 10 level

D BVE MVE EqyIss

(1) (2) (3) (4) (5) (6) (7) (8)

Profitst1 025a 024a 023a 038a 007 066a 007a 009a

(40) (32) (36) (39) (06) (29) (42) (51)Profits 051b 082a 149a 004a

(25) (33) (49) (33)Assetst1 002 005c 024a 001b

(08) (17) (34) (23)Constant 93 176a 45 71 1552a 1704a 130a 140a

(12) (28) (07) (09) (49) (49) (45) (45)Year FE Yes Yes Yes Yes Yes Yes Yes YesR2 001 005 001 011 001 005 002 002N 162056 162056 162130 162130 157550 157550 162154 162154

Table AIII Median profitability sorts

The sample contains nonfinancial firms listed on the annual Compustat files for the period from 1971

to 2009 The table presents median debt and equity median changes in debt and equity and median

equity issuances (EqyIss) in millions of dollars for all firms sorted on profitability and for profitability

sorts within the smallest and largest firms The sorts are done annually

D D BVE BVE MVE MVE EqyIss Assets(1) (2) (3) (4) (5) (6) (7) (8)

Allfirm

s

Low 21 00 75 17 256 20 01 1922 186 02 449 18 598 19 00 11033 816 06 1274 03 1637 05 00 32244 787 02 1541 18 2488 03 00 3556High 273 00 1361 66 3300 14 00 2578

Smallfirm

s Low 05 00 13 07 140 14 06 402 06 00 25 08 114 12 01 563 09 00 32 05 85 07 00 674 12 00 39 01 68 03 00 79High 10 00 51 04 98 01 00 90

Largefirm

s Low 11485 381 10269 371 13227 202 03 333902 13051 393 12786 140 17571 169 05 392463 11554 287 13429 24 21623 305 00 370314 9963 209 15052 194 29424 308 00 36705High 7361 00 16380 613 49142 439 00 34976

PROFITSndashLEVERAGE PUZZLEREVISITED 1451

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

References

Auerbach A J (1985) Real determinants of corporate leverage in B M Friedman (ed)

Corporate Capital Structures in the United States University of Chicago Press Chicago

IllinoisBertrand M and Mullainathan S (2003) Enjoying the quiet life corporate governance and

managerial preferences Journal of Political Economy 111 1043ndash1075Booth L Varouj A Demirguc-Kunt A and Maksimovic V (2001) Capital structures in

developing countries Journal of Finance 56 87ndash130Brennan M J and Schwartz E S (1984) Optimal financial policy and firm valuation

Journal of Finance 39 593ndash607Cameron A C and Trivedi P K (2010) Microeconometrics using Stata Stata Press TexasConstantinides G M (1979) Multiperiod consumption and investment behavior with

convex transactions costs Management Science 25 1127ndash1137Constantinides G M and Richard S F (1978) Existence of optimal simple policies for

discounted-cost inventory and cash management in continuous time Operations Research

26 620ndash636DeAngelo H DeAngelo L and Whited T M (2011) Capital structure dynamics and

transitory debt Journal of Financial Economics 99 235ndash261Dixit A K (1993) The Art of Smooth Pasting Routledge London and New YorkFama E and French K R (2002) Testing trade-off and pecking order predictions about

dividends and debt Review of Financial Studies 15 1ndash33Faulkender M and Petersen M A (2006) Does the source of capital affect capital struc-

ture Review of Financial Studies 19 45ndash79Faulkender M Flannery M J Hankins K W and Smith J M (2012) Cash Flows and

Leverage Adjustments Journal of Financial Economics 103 632ndash646Fischer E Heinkel R and Zechner J (1989) Dynamic capital structure choice theory

and tests Journal of Finance 44 19ndash40Frank M Z and Goyal V K (2008) Trade-off and pecking order theories of

debt in B E Eckbo (ed) Handbook of Corporate Finance Empirical Corporate Finance

Vol 2 Handbook of Finance Series Chapter 12 ElsevierNorth-Holland Amsterdam pp

135ndash202

Frank M Z and Goyal V K (2009) Capital structure decisions which factors are reliably

important Financial Management 38 1ndash37

Garvey G T and Hanka G (1999) Capital structure and corporate control the effect of

antitakeover statutes on firm leverage Journal of Finance 54 519ndash546

Goldstein R Ju N and Leland H (2001) An EBIT-based model of dynamic capital

structure Journal of Business 74 483ndash512

Graham J R and Tucker A L (2006) Tax shelters and corporate debt policy Journal of

Financial Economics 81 563ndash594

Hackbarth D and Mauer D C (2012) Optimal priority structure capital structure and

investment Review of Financial Studies 25 747ndash796

Halling M Yu J and Zechner J (2012) Leverage dynamics over the business cycle

Working paper University of Utah University of New South Wales and Vienna

University of Economics and Business AdministrationHennessy C A and Whited T A (2005) Debt dynamics Journal of Finance 60 1129ndash1165

Huang R and Ritter J R (2009) Testing theories of capital structure and estimating the

speed of adjustment Journal of Financial and Qualitative Analysis 44 237ndash271

1452 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

Hovakimian A Opler T and Titman S (2001) The debt-equity choice Journal of

Financial and Quantitative Analysis 36 1ndash24Jung K Kim Y and Stulz R M (1996) Timing investment opportunities managerial

discretion and the security issue decisions Journal of Financial Economics 42 159ndash185Kane A Marcus A J and McDonald R L (1984) How big is the tax advantage to

debt Journal of Finance 39 841ndash853Kayhan A and Titman S (2007) Firmsrsquo histories and their capital structures Journal of

Financial Economics 83 1ndash32Leary M T (2009) Bank loan supply lender choice and corporate capital structure

Journal of Finance 64 1143ndash1185Leary M T and Roberts M R (2005) Do firms rebalance their capital structures

Journal of Finance 60 2575ndash2619Leland H E (1994) Corporate debt value bond covenants and optimal capital structure

Journal of finance 49 1213ndash1252Lemmon M L Roberts M R and Zender J F (2008) Back to the beginning persist-

ence and the cross-section of corporate capital structure Journal of Finance 63 1575ndash1608Long M S and Malitz I B (1985) The investment-financing nexus some empirical

evidence Midland Corporate Finance Journal 3 53ndash59Mackie-Mason J K (1990) Do taxes affect corporate financing decisions Journal of

Finance 45 1471ndash1493Morellec E Nikolov B and Schurhoff N (2012) Corporate governance and capital

structure dynamics Journal of Finance 67 803ndash848Myers S C (1984) The capital structure puzzle Journal of Finance 39 575ndash592Myers S C (1993) Still searching for optimal capital structure Journal of Applied

Corporate Finance 61 4ndash14Rajan R and Zingales L (1995) What do we know about capital structure some evidence

from international data Journal of Finance 50 1421ndash1460Strebulaev I A (2007) Do tests of capital structure theory mean what they say Journal of

Finance 62 1747ndash1787

Strebulaev I A and Whited T M (2012) Dynamic Models and Structural Estimation in

Corporate Finance Now Publisher Boston and Delft

Titman S and Wessels R (1988) The determinants of capital structure choice Journal of

Finance 43 1ndash21

Welch I (2004) Capital structure and stock returns Journal of Political Economy 112

106ndash131

Welch I (2011) Two common problems in capital structure research the financial-debt-

to-asset ratio and issuing activity versus leverage changes International Review of Finance

11 1ndash17

PROFITSndashLEVERAGE PUZZLEREVISITED 1453

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

changes are crucial Accordingly in this section we take a step back toconsider prior literature in which variables are scaled The questions arehow does scaling affect the results and whyTable XI examines alternative scalings Columns (1) and (2) scale debt and

equity issuances by the total firm net issuance Columns (3) and (4) scaledebt and equity issuances by the sum of debt and book equity Columns (5)and (6) consider the change in debt scaled by the sum of debt and bookequity Columns (7) and (8) consider the change in leverage due to externalfinancing Since we are examining changes we use the changes in thestandard regression factors as controls The primary interest is in the coef-ficient on the change in profitsWe find that a change in profits is associated with an increase in debt

issues But the sign reverses when we scale the debt issues by the sum of debtplus book equity This is telling What is really driving the results is thechange in equity in the denominator not the change in debt in the numer-ator This result is further substantiated in Columns (5) and (6) where thechange in debt to capital ratio is used instead of debt issuanceTo further explore the impact of active issuance relative to passive

experiencing of changes we decompose leverage by separating out theeffect of debt and equity issues from those due to changes in retainedearnings Thus we define leverage change due to external financing as

LevEF frac14Dt1 thorn d

Dt1 thorn Et1 thorn dthorn e

Dt1

Dt1 thorn Et1 eth1THORN

where D is debt E is book value of equity d is debt issues net of retirementsand e is equity issues net of repurchases Garvey and Hanka (1999) use asimilar measure in their study This measure directly captures the netleverage effect of debt and equity issuancesColumns (7) and (8) report results from regressions of the net leverage

effect of debt and equity issues on changes in profitability after controllingfor other leverage factors in differences and year indicator variables Theeffect of profitability on leverage change due to external financing is positiveand statistically significant This comparison suggests that changes in totaldebt and equity are only partly a result of debt and equity issuance decisionsOther balance sheet adjustments complicate the inferences from leverageratio regressionsFor most of our analysis we follow the literature in treating debt as homo-

genous It is of course potentially interesting to ask what happens when wedistinguish between short-term debt and long-term debt Which part of thefirmrsquos debt maturity structure drives the under-adjustment Table XII

PROFITSndashLEVERAGE PUZZLEREVISITED 1443

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

provides results from such decompositions Alternative scalings areconsidered The main results of interest are that the distinction betweenshort- and long-term debt is important Short-term debt issuance is largelyindependent of changes in profits The negative sign on profit change stemsfrom the action of long-term debt This seems intuitively reasonable

Table XI Scaled issuances

The sample contains nonfinancial firms listed on the annual Compustat files for the period

from 1971 to 2009 The table presents estimates from regressions of debt issuance to totalissuance (in columns (1) and (2)) debt issuance to total capitalization (in columns (3) and(4)) change in debt to capitalization ratio (in columns (5) and (6)) and change in leverage

ratio considering only external financing (in columns (7) and (8)) The change in leverageratio considering only external financing is defined as

LevEF frac14Dt1 thorn d

Dt1 thorn Et1 thorn dthorn e

Dt1

Dt1 thorn Et1

where D is debt E is book value of equity d is debt issues net of retirements and e is equity

issues net of repurchases The regressions include lagged leverage factors and year indicatorvariables Some of the specifications include firm fixed effects The reported t-statistics arecorrected for clustering at the firm level

aMean significant at the 1 level

bMean significant at the 5 level

cMean significant at the 10 level

DebtIssTotIss

DebtIssethDthornETHORN D

DthornE LevEF

(1) (2) (3) (4) (5) (6) (7) (8)

Profitt1 004a 000 002a 001c 003a 002 001b 002a

(35) (03) (34) (18) (35) (14) (23) (26)

IndLevt1 004c 001 001 001c 000 000 000 000

(19) (06) (15) (19) (02) (04) (03) (05)

Market=bookt1 001a 001a 000a 000a 000a 000a 000 000c

(47) (85) (68) (60) (55) (36) (05) (18)

Tangt1 004 005 013a 011a 011a 006a 003a 003a

(12) (15) (144) (112) (79) (38) (44) (42)

Assetst1 008a 004a 006a 004a 002a 001a 003a 002a

(97) (45) (283) (161) (53) (35) (147) (101)

Constant 079a 074a 002a 003a 000 000 000b 002a

(497) (442) (91) (111) (00) (05) (22) (93)

Year FE Yes Yes Yes Yes Yes Yes Yes Yes

Firm FE No Yes No Yes No Yes No Yes

R2 0015 0098 0022 0076 0009 0070 0012 0086

N 133905 133905 140552 140552 138229 138229 140572 140572

1444 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

inasmuch as short-term debt is probably more associated with day-to-dayoperations rather than longer term planning considerationsIn Table XIII we examine change in leverage ( D

DthornE) and change inleverage due to external financing (LEVEF) during good and bad timesseparately The traditional negative effect of a change in profits on achange in leverage is found both in good times and in bad timesHowever when we decompose the impact in order to isolate the externalcomponent as before the result disappears In good times the externalfinancing component of a change in profits is weak but still positiveDuring bad times it is not statistically significant This implies that distin-guishing between good and bad times does not fundamentally alter the basic

Table XII Short- and long-term debt issuances

The sample contains nonfinancial firms listed on the annual Compustat files for the period

from 1971 to 2009 The table presents estimates from regressions of short-term debtissuance to total issuance (in columns (1) and (2)) long-term debt issuance to totalissuance (in columns (3) and (4)) short-term debt issuance to total capitalization (in

columns (5) and (6)) and long-term debt to total capitalization (in columns (7) and (8))Some of the specifications include firm fixed effects The reported t-statistics are correctedfor clustering at the firm level

aMean significant at the 1 level

bMean significant at the 5 level

cMean significant at the 10 level

STDebtIssTotIss

LTDebtIssTotIss

STDebtIssethDthornETHORN

LTDebtIssethDthornETHORN

(1) (2) (3) (4) (5) (6) (7) (8)

Profitabilityt1 0002 0003 0042a 0008 0001 0001 0019a 0012a

(02) (04) (30) (05) (14) (10) (45) (28)

IndLevt1 0009 0007 0070b 0033 0001 0002c 0014a 0005

(06) (05) (26) (11) (08) (17) (26) (09)

M=Bt1 0003a 0003a 0003b 0010a 0000 0000 0003a 0002a

(48) (35) (22) (65) (10) (01) (84) (42)

Tangibilityt1 0020 0007 0011 0067 0005a 0006a 0097a 0081a

(12) (03) (03) (16) (34) (33) (108) (89)

Assetst1 0039a 0011b 0034a 0024b 0001a 0001 0046a 0020a

(89) (22) (38) (23) (36) (13) (190) (82)

Constant 0001b 0153a 0798a 0591a 0000c 0006a 0057a 0096a

(25) (144) (470) (288) (17) (124) (292) (234)

Year FE Yes Yes Yes Yes Yes Yes Yes Yes

Firm FE No Yes No Yes No Yes No Yes

R2 0025 0089 0018 0073 0030 0147 0016 0298

N 136146 136146 136140 136140 143031 143031 142847 142847

PROFITSndashLEVERAGE PUZZLEREVISITED 1445

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

Table XIII Externally financed leverage changes in good and bad times

The sample contains nonfinancial firms listed on the annual Compustat files for the period

from 1971 to 2009 The table presents estimates from regressions of change in debt tocapitalization ratio (in columns (1) (2) (5) and (6)) and change in leverage ratio consider-ing only external financing (in columns (3) (4) (7) and (8)) The change in leverage ratio

considering only external financing is defined as

LevEF frac14Dt1 thorn d

Dt1 thorn Et1 thorn dthorn e

Dt1

Dt1 thorn Et1

where D is debt E is book value of equity d is debt issues net of retirements and e is equityissues net of repurchases The regressions include lagged leverage factors and year indicator

variables The reported t-statistics are corrected for clustering at the firm level An industryis defined as having ldquogood timesrdquo if the median firm in that industry has a market-to-bookratio that is higher than the 67th percentile of the time-series distribution of industry

median market-to-book ratios Conversely an industry is defined as having ldquobad timesrdquoif the median firm in that industry has a market-to-book ratio that is lower than the 33rdpercentile of the time-series distribution of industry median market-to-book ratios

aMean significant at the 1 level

bMean significant at the 5 level

cMean significant at the 10 level

Good times Bad times

DDthornE LevEF D

DthornE LevEF

(1) (2) (3) (4) (5) (6) (7) (8)

Profitabiltyt1 0022c 0005 0016c 0025b 0073b 0041 0007 0007

(17) (03) (18) (24) (25) (11) (04) (03)

IndLevt1 0000 0011 0001 0009 0010 0018 0001 0000

(00) (09) (01) (13) (07) (10) (02) (00)

MB

t1

0004a 0003b 0000 0001 0008a 0001 0000 0000

(40) (24) (05) (08) (28) (04) (03) (03)

Tangt1 0097a 0041 0033a 0038b 0127a 0048 0059a 0033

(45) (16) (27) (25) (49) (14) (43) (16)

Assetst1 0014a 0013b 0024a 0019a 0037a 0023c 0036a 0025a

(30) (20) (84) (48) (45) (17) (88) (43)

Constant 0030b 0036 0000 0003 0028c 0014 0021 0001

(20) (16) (00) (07) (17) (06) (11) (01)

Year FE Yes Yes Yes Yes Yes Yes Yes Yes

Firm FE No Yes No Yes No Yes No Yes

R2 0006 0088 0009 0117 0026 0251 0023 0202

N 56820 56820 58200 58200 23462 23462 23511 23511

1446 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

conclusions The traditional leverage result on profits is driven by increasesin equity that are experienced and then partially but not completely offsetby firmsTo the extent that firms are engaging in active financing they are following

the prescriptions of the trade-off theory Thus we have a coherent picture ofhow firms are responding to profits Basically more profitable firms experiencean increase in the value of their equity They respond to this by issuing debtand repurchasing equity But the magnitude of the increase in equity is largerthan the magnitudes of the active steps taken in the debt and equity marketsThis means that the firmsrsquo actual actions are as predicted by the trade-offtheory But it leaves open the question of why they do not act more stronglyThe capital structure literature has long been interested in dynamics and

capital structure adjustment paths In recent years this interest has increas-ingly resulted in explicit dynamic optimizing models such as those ofGoldstein Ju and Leland (2001) Hennessy and Whited (2005) andStrebulaev (2007) These models commonly make specialized assumptionsleading some to question their generalityWe argue that both fixed transactions costs and variable transactions costs

are plausible and realistic in models that pin down firm size The simultaneouspresence of both is important Due to fixed costs firms only rebalance inter-mittently Due to variable adjustment costs the extent of the adjustment isreduced to the point at which an extra unit to adjustment is balanced againstthe extra unit transactions costs This means that both leverage increases andleverage reductions will only go part way toward a static optimumThe key point is that with both fixed and variable transactions costs we

should expect to only see intermittent leverage adjustments When leverage isadjusted it should ldquounder adjustrdquo when compared to a frictionless modelFurther discussion of transactions costs in a capital structure setting isprovided by Leary and Roberts (2005)

9 Conclusion

The connection between corporate profits and capital structure has beenvery influential in the assessment of the trade-off theory The conventionalinterpretation of the evidence has pushed the literature away from the trade-off theory and toward much more complex dynamic models and ideas As aresult it is important to make sure that the evidence has been correctlyinterpreted We show that the problem is with the interpretation of thetest results and a failure to take into account the costs associated with re-sponding to a profit shock

PROFITSndashLEVERAGE PUZZLEREVISITED 1447

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

We document that more profitable firms really do borrow more (not less)They tend to repurchase equity They experience an increase in both thebook value of equity and the market value of equity Less profitable firmsreally do tend to reduce their debt and to issue equityFirm size and market conditions also matter Larger firms tend to be more

active in the debt markets while smaller firms tend to be relatively moreactive in the equity markets During good times there is more use ofexternal financingThe usual profitsndashleverage puzzle result is primarily driven by the increase

in equity that is experienced by the more profitable firms Accordingly thepuzzle should be restated as asking why do firms not take sufficiently largeoffsetting actions to fully undo the change in equity What limits the mag-nitudes of the typical leverage response to profit shocksIn a frictionless model the partial response appears puzzling But there is

good empirical reason to believe that rebalancing entails both fixed andvariable costs and that firm size matters The rebalancing costs can befully avoided by doing nothing Accordingly the firm must decide whetherany given shock is big enough to be worth responding to If it is then thefirm must decide how big a response is called for These technical conditionsare known as ldquovalue matchingrdquo and ldquosmooth pastingrdquo Thus optimizationimplies that some shocks will be ignored Even if the shock is not ignoredthe optimal response will only partially undo the shock The magnitude ofthe leverage response must balance the marginal cost and the marginalbenefit of an extra unit of leverage Since the marginal cost of adjustingleverage is strictly positive the adjustment toward that static leverageoptimum will only go part way This is true both for leverage increasesand for leverage reductionsOne might conjecture that the variable costs would be time varying and

that such costs might have declined over time Since we do not have a fullaccounting of all of the variable costs it is of course impossible to be sureHowever there is some evidence suggesting that at least the directunderwriting fees are remarkably stable over time10

Surely nobody ought to be surprised that more profitable firms are moreinclined to repurchase equity while unprofitable firms tend not to do so Thefact that more profitable firms typically experience an increase in the value of

10 For IPOs Jay Ritter provides important summary evidence as httpbearwarringtonufleduritterIPOs2013Underwritingpdf He finds that for medium-sized deals a remark-

ably large fraction of the deals have a 7 fee year after year Whether this is due tobehavioral factors or some sort of a coordination device is not clear The extent towhich similar intertemporal stability carries over to other forms of external financemerits further study

1448 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

equity is equally natural Were it not for the contrary claims in the literature(ie the profitsndashleverage puzzle) we would argue that nobody ought to findit surprising that more profitable firms are more inclined to issue debt Thefact that leverage adjustments typically only partially offset profit shocksshould also sound quite natural These facts all fit together easily within thetrade-off theory once we take account of both fixed and variable costs ofactively adjusting capital structure

Appendix

Table AI Market leverage regressions

The table presents estimates of the regressions of market leverage on firm characteristics

The sample comes from the annual Compustat files during the period 1971ndash2009 Financialfirms are excluded The leverage is estimated as the ratio of debt over debt plus marketequity The explanatory variables Profitabilityt1

MB

t1

Tangibilityt1 and LnethAssetsTHORNt1are described in Table 1 IndMedianLevt1 is estimated as the median market leverage of allother firms in the same industry The industry is defined at the level of the four-digit SICcode All specifications include the year fixed effects The specifications in column (2)

additionally include the firm fixed effects We report t-statistics where the standard errorsare clustered at the firm level in parentheses

aSignificant at the 1 level

Quantile regressions

(1) (2) (3) (4) (5)

Profitabilityt1 013a 013a 007a 011a 020a

(275) (262) (349) (425) (538)

IndLevt1 046a 020a 032a 058a 063a

(506) (239) (1299) (1630) (1042)MB

t1

003a 001a 001a 002a 004a

(509) (303) (743) (709) (534)

Tangibilityt1 013a 018a 013a 014a 010a

(191) (181) (762) (571) (232)

LnethAssetsTHORNt1 001a 005a 001a 001a lt001a

(112) (258) (682) (342) (29)

Constant 010a 011a 005a 005a 029a

(171) (98) (173) (118) (398)

Year FE Yes Yes Yes Yes Yes

Firm FE No Yes No No No

Regression type OLS FE 25th tile Median 75thtile

Clustered SE Yes Yes No No No

R2 030 067

Pseudo R2 016 023 0201

N 158578 158578 158578 158578 158578

PROFITSndashLEVERAGE PUZZLEREVISITED 1449

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

Table AII Profitability sorts for firms issuing large amounts

Percentage of firms

Iss Ret Iss Iss Rep Iss Iss IssD IssE

D D NetD E E NetE DampE RepE RetD

(1) (2) (3) (4) (5) (6) (7) (8) (9)

Panel A 10 cutoff to define issuing or retiring activity

Allfirm

s

Low 205 150 134 260 05 257 45 00 23

2 207 175 103 77 06 75 11 01 10

3 243 196 101 50 08 47 08 01 09

4 252 186 113 50 13 48 08 02 08

High 227 161 118 60 27 57 07 04 09

t Highfrac14Low 70 38 62 755 226 768 320 90 152

Smallfirm

s Low 230 132 162 380 05 375 82 00 32

2 228 143 165 334 04 333 65 00 26

3 211 146 147 216 03 214 38 00 23

4 190 154 120 126 04 124 15 01 20

High 194 162 114 99 09 97 12 01 15

t Highfrac14Low 42 40 68 342 20 340 169 07 131

Largefirm

s Low 231 200 82 31 06 29 08 01 04

2 231 187 67 17 06 14 03 01 02

3 234 177 77 17 09 14 03 02 03

4 218 158 75 16 17 14 02 03 02

High 218 135 99 19 41 15 03 07 02

t Highfrac14Low 19 106 36 46 135 57 39 58 23

Percentage of firms

Iss Ret Iss Iss Rep Iss Both Iss D Iss E

D D Net D E E Net E DampE Rep E Ret D

(1) (2) (3) (4) (5) (6) (7) (8) (9)

Panel B 15 cutoff to define issuing or retiring activity

Allfirm

s

Low 158 101 100 230 01 228 27 00 15

2 142 115 63 57 01 56 05 00 06

3 165 128 57 32 01 31 03 00 05

4 173 118 65 35 01 34 03 00 05

High 160 102 70 46 04 44 03 00 05

t Highfrac14Low 08 06 141 738 87 746 277 24 131

Smallfirm

s Low 195 95 132 354 01 350 56 00 24

2 185 96 131 303 01 301 44 00 18

3 160 96 104 179 00 176 18 00 15

4 134 99 78 99 01 98 08 00 12

High 133 95 73 82 01 80 06 00 10

t Highfrac14Low 80 00 95 345 12 344 149 12 52

Largefirm

s Low 145 123 42 14 01 12 03 00 01

2 137 114 32 07 01 06 01 00 01

3 142 104 35 06 01 05 00 00 01

4 130 91 36 08 03 07 01 01 01

High 134 81 48 09 06 07 01 01 01

t Highfrac14Low 19 85 17 26 54 33 34 17 11

1450 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

Table AIV Debt and equity changesmdashwithout fixed effects

The sample contains nonfinancial firms listed on the annual Compustat files for the period from 1971

to 2009 The table presents estimates from regressions of change in debt (D) change in the book

value of equity (BVE) change in the market value of equity (MVE) and net equity issuances

(EqyIss) The explanatory variables include current and lagged change in operating income before

depreciation (Profits and Profitst1) and lagged change in book value of assets (Assetst1) The

regressions include year indicator variables The reported t-statistics are corrected for clustering at the

firm level

aMean significant at the 1 level

bMean significant at the 5 level

cMean significant at the 10 level

D BVE MVE EqyIss

(1) (2) (3) (4) (5) (6) (7) (8)

Profitst1 025a 024a 023a 038a 007 066a 007a 009a

(40) (32) (36) (39) (06) (29) (42) (51)Profits 051b 082a 149a 004a

(25) (33) (49) (33)Assetst1 002 005c 024a 001b

(08) (17) (34) (23)Constant 93 176a 45 71 1552a 1704a 130a 140a

(12) (28) (07) (09) (49) (49) (45) (45)Year FE Yes Yes Yes Yes Yes Yes Yes YesR2 001 005 001 011 001 005 002 002N 162056 162056 162130 162130 157550 157550 162154 162154

Table AIII Median profitability sorts

The sample contains nonfinancial firms listed on the annual Compustat files for the period from 1971

to 2009 The table presents median debt and equity median changes in debt and equity and median

equity issuances (EqyIss) in millions of dollars for all firms sorted on profitability and for profitability

sorts within the smallest and largest firms The sorts are done annually

D D BVE BVE MVE MVE EqyIss Assets(1) (2) (3) (4) (5) (6) (7) (8)

Allfirm

s

Low 21 00 75 17 256 20 01 1922 186 02 449 18 598 19 00 11033 816 06 1274 03 1637 05 00 32244 787 02 1541 18 2488 03 00 3556High 273 00 1361 66 3300 14 00 2578

Smallfirm

s Low 05 00 13 07 140 14 06 402 06 00 25 08 114 12 01 563 09 00 32 05 85 07 00 674 12 00 39 01 68 03 00 79High 10 00 51 04 98 01 00 90

Largefirm

s Low 11485 381 10269 371 13227 202 03 333902 13051 393 12786 140 17571 169 05 392463 11554 287 13429 24 21623 305 00 370314 9963 209 15052 194 29424 308 00 36705High 7361 00 16380 613 49142 439 00 34976

PROFITSndashLEVERAGE PUZZLEREVISITED 1451

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

References

Auerbach A J (1985) Real determinants of corporate leverage in B M Friedman (ed)

Corporate Capital Structures in the United States University of Chicago Press Chicago

IllinoisBertrand M and Mullainathan S (2003) Enjoying the quiet life corporate governance and

managerial preferences Journal of Political Economy 111 1043ndash1075Booth L Varouj A Demirguc-Kunt A and Maksimovic V (2001) Capital structures in

developing countries Journal of Finance 56 87ndash130Brennan M J and Schwartz E S (1984) Optimal financial policy and firm valuation

Journal of Finance 39 593ndash607Cameron A C and Trivedi P K (2010) Microeconometrics using Stata Stata Press TexasConstantinides G M (1979) Multiperiod consumption and investment behavior with

convex transactions costs Management Science 25 1127ndash1137Constantinides G M and Richard S F (1978) Existence of optimal simple policies for

discounted-cost inventory and cash management in continuous time Operations Research

26 620ndash636DeAngelo H DeAngelo L and Whited T M (2011) Capital structure dynamics and

transitory debt Journal of Financial Economics 99 235ndash261Dixit A K (1993) The Art of Smooth Pasting Routledge London and New YorkFama E and French K R (2002) Testing trade-off and pecking order predictions about

dividends and debt Review of Financial Studies 15 1ndash33Faulkender M and Petersen M A (2006) Does the source of capital affect capital struc-

ture Review of Financial Studies 19 45ndash79Faulkender M Flannery M J Hankins K W and Smith J M (2012) Cash Flows and

Leverage Adjustments Journal of Financial Economics 103 632ndash646Fischer E Heinkel R and Zechner J (1989) Dynamic capital structure choice theory

and tests Journal of Finance 44 19ndash40Frank M Z and Goyal V K (2008) Trade-off and pecking order theories of

debt in B E Eckbo (ed) Handbook of Corporate Finance Empirical Corporate Finance

Vol 2 Handbook of Finance Series Chapter 12 ElsevierNorth-Holland Amsterdam pp

135ndash202

Frank M Z and Goyal V K (2009) Capital structure decisions which factors are reliably

important Financial Management 38 1ndash37

Garvey G T and Hanka G (1999) Capital structure and corporate control the effect of

antitakeover statutes on firm leverage Journal of Finance 54 519ndash546

Goldstein R Ju N and Leland H (2001) An EBIT-based model of dynamic capital

structure Journal of Business 74 483ndash512

Graham J R and Tucker A L (2006) Tax shelters and corporate debt policy Journal of

Financial Economics 81 563ndash594

Hackbarth D and Mauer D C (2012) Optimal priority structure capital structure and

investment Review of Financial Studies 25 747ndash796

Halling M Yu J and Zechner J (2012) Leverage dynamics over the business cycle

Working paper University of Utah University of New South Wales and Vienna

University of Economics and Business AdministrationHennessy C A and Whited T A (2005) Debt dynamics Journal of Finance 60 1129ndash1165

Huang R and Ritter J R (2009) Testing theories of capital structure and estimating the

speed of adjustment Journal of Financial and Qualitative Analysis 44 237ndash271

1452 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

Hovakimian A Opler T and Titman S (2001) The debt-equity choice Journal of

Financial and Quantitative Analysis 36 1ndash24Jung K Kim Y and Stulz R M (1996) Timing investment opportunities managerial

discretion and the security issue decisions Journal of Financial Economics 42 159ndash185Kane A Marcus A J and McDonald R L (1984) How big is the tax advantage to

debt Journal of Finance 39 841ndash853Kayhan A and Titman S (2007) Firmsrsquo histories and their capital structures Journal of

Financial Economics 83 1ndash32Leary M T (2009) Bank loan supply lender choice and corporate capital structure

Journal of Finance 64 1143ndash1185Leary M T and Roberts M R (2005) Do firms rebalance their capital structures

Journal of Finance 60 2575ndash2619Leland H E (1994) Corporate debt value bond covenants and optimal capital structure

Journal of finance 49 1213ndash1252Lemmon M L Roberts M R and Zender J F (2008) Back to the beginning persist-

ence and the cross-section of corporate capital structure Journal of Finance 63 1575ndash1608Long M S and Malitz I B (1985) The investment-financing nexus some empirical

evidence Midland Corporate Finance Journal 3 53ndash59Mackie-Mason J K (1990) Do taxes affect corporate financing decisions Journal of

Finance 45 1471ndash1493Morellec E Nikolov B and Schurhoff N (2012) Corporate governance and capital

structure dynamics Journal of Finance 67 803ndash848Myers S C (1984) The capital structure puzzle Journal of Finance 39 575ndash592Myers S C (1993) Still searching for optimal capital structure Journal of Applied

Corporate Finance 61 4ndash14Rajan R and Zingales L (1995) What do we know about capital structure some evidence

from international data Journal of Finance 50 1421ndash1460Strebulaev I A (2007) Do tests of capital structure theory mean what they say Journal of

Finance 62 1747ndash1787

Strebulaev I A and Whited T M (2012) Dynamic Models and Structural Estimation in

Corporate Finance Now Publisher Boston and Delft

Titman S and Wessels R (1988) The determinants of capital structure choice Journal of

Finance 43 1ndash21

Welch I (2004) Capital structure and stock returns Journal of Political Economy 112

106ndash131

Welch I (2011) Two common problems in capital structure research the financial-debt-

to-asset ratio and issuing activity versus leverage changes International Review of Finance

11 1ndash17

PROFITSndashLEVERAGE PUZZLEREVISITED 1453

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

provides results from such decompositions Alternative scalings areconsidered The main results of interest are that the distinction betweenshort- and long-term debt is important Short-term debt issuance is largelyindependent of changes in profits The negative sign on profit change stemsfrom the action of long-term debt This seems intuitively reasonable

Table XI Scaled issuances

The sample contains nonfinancial firms listed on the annual Compustat files for the period

from 1971 to 2009 The table presents estimates from regressions of debt issuance to totalissuance (in columns (1) and (2)) debt issuance to total capitalization (in columns (3) and(4)) change in debt to capitalization ratio (in columns (5) and (6)) and change in leverage

ratio considering only external financing (in columns (7) and (8)) The change in leverageratio considering only external financing is defined as

LevEF frac14Dt1 thorn d

Dt1 thorn Et1 thorn dthorn e

Dt1

Dt1 thorn Et1

where D is debt E is book value of equity d is debt issues net of retirements and e is equity

issues net of repurchases The regressions include lagged leverage factors and year indicatorvariables Some of the specifications include firm fixed effects The reported t-statistics arecorrected for clustering at the firm level

aMean significant at the 1 level

bMean significant at the 5 level

cMean significant at the 10 level

DebtIssTotIss

DebtIssethDthornETHORN D

DthornE LevEF

(1) (2) (3) (4) (5) (6) (7) (8)

Profitt1 004a 000 002a 001c 003a 002 001b 002a

(35) (03) (34) (18) (35) (14) (23) (26)

IndLevt1 004c 001 001 001c 000 000 000 000

(19) (06) (15) (19) (02) (04) (03) (05)

Market=bookt1 001a 001a 000a 000a 000a 000a 000 000c

(47) (85) (68) (60) (55) (36) (05) (18)

Tangt1 004 005 013a 011a 011a 006a 003a 003a

(12) (15) (144) (112) (79) (38) (44) (42)

Assetst1 008a 004a 006a 004a 002a 001a 003a 002a

(97) (45) (283) (161) (53) (35) (147) (101)

Constant 079a 074a 002a 003a 000 000 000b 002a

(497) (442) (91) (111) (00) (05) (22) (93)

Year FE Yes Yes Yes Yes Yes Yes Yes Yes

Firm FE No Yes No Yes No Yes No Yes

R2 0015 0098 0022 0076 0009 0070 0012 0086

N 133905 133905 140552 140552 138229 138229 140572 140572

1444 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

inasmuch as short-term debt is probably more associated with day-to-dayoperations rather than longer term planning considerationsIn Table XIII we examine change in leverage ( D

DthornE) and change inleverage due to external financing (LEVEF) during good and bad timesseparately The traditional negative effect of a change in profits on achange in leverage is found both in good times and in bad timesHowever when we decompose the impact in order to isolate the externalcomponent as before the result disappears In good times the externalfinancing component of a change in profits is weak but still positiveDuring bad times it is not statistically significant This implies that distin-guishing between good and bad times does not fundamentally alter the basic

Table XII Short- and long-term debt issuances

The sample contains nonfinancial firms listed on the annual Compustat files for the period

from 1971 to 2009 The table presents estimates from regressions of short-term debtissuance to total issuance (in columns (1) and (2)) long-term debt issuance to totalissuance (in columns (3) and (4)) short-term debt issuance to total capitalization (in

columns (5) and (6)) and long-term debt to total capitalization (in columns (7) and (8))Some of the specifications include firm fixed effects The reported t-statistics are correctedfor clustering at the firm level

aMean significant at the 1 level

bMean significant at the 5 level

cMean significant at the 10 level

STDebtIssTotIss

LTDebtIssTotIss

STDebtIssethDthornETHORN

LTDebtIssethDthornETHORN

(1) (2) (3) (4) (5) (6) (7) (8)

Profitabilityt1 0002 0003 0042a 0008 0001 0001 0019a 0012a

(02) (04) (30) (05) (14) (10) (45) (28)

IndLevt1 0009 0007 0070b 0033 0001 0002c 0014a 0005

(06) (05) (26) (11) (08) (17) (26) (09)

M=Bt1 0003a 0003a 0003b 0010a 0000 0000 0003a 0002a

(48) (35) (22) (65) (10) (01) (84) (42)

Tangibilityt1 0020 0007 0011 0067 0005a 0006a 0097a 0081a

(12) (03) (03) (16) (34) (33) (108) (89)

Assetst1 0039a 0011b 0034a 0024b 0001a 0001 0046a 0020a

(89) (22) (38) (23) (36) (13) (190) (82)

Constant 0001b 0153a 0798a 0591a 0000c 0006a 0057a 0096a

(25) (144) (470) (288) (17) (124) (292) (234)

Year FE Yes Yes Yes Yes Yes Yes Yes Yes

Firm FE No Yes No Yes No Yes No Yes

R2 0025 0089 0018 0073 0030 0147 0016 0298

N 136146 136146 136140 136140 143031 143031 142847 142847

PROFITSndashLEVERAGE PUZZLEREVISITED 1445

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

Table XIII Externally financed leverage changes in good and bad times

The sample contains nonfinancial firms listed on the annual Compustat files for the period

from 1971 to 2009 The table presents estimates from regressions of change in debt tocapitalization ratio (in columns (1) (2) (5) and (6)) and change in leverage ratio consider-ing only external financing (in columns (3) (4) (7) and (8)) The change in leverage ratio

considering only external financing is defined as

LevEF frac14Dt1 thorn d

Dt1 thorn Et1 thorn dthorn e

Dt1

Dt1 thorn Et1

where D is debt E is book value of equity d is debt issues net of retirements and e is equityissues net of repurchases The regressions include lagged leverage factors and year indicator

variables The reported t-statistics are corrected for clustering at the firm level An industryis defined as having ldquogood timesrdquo if the median firm in that industry has a market-to-bookratio that is higher than the 67th percentile of the time-series distribution of industry

median market-to-book ratios Conversely an industry is defined as having ldquobad timesrdquoif the median firm in that industry has a market-to-book ratio that is lower than the 33rdpercentile of the time-series distribution of industry median market-to-book ratios

aMean significant at the 1 level

bMean significant at the 5 level

cMean significant at the 10 level

Good times Bad times

DDthornE LevEF D

DthornE LevEF

(1) (2) (3) (4) (5) (6) (7) (8)

Profitabiltyt1 0022c 0005 0016c 0025b 0073b 0041 0007 0007

(17) (03) (18) (24) (25) (11) (04) (03)

IndLevt1 0000 0011 0001 0009 0010 0018 0001 0000

(00) (09) (01) (13) (07) (10) (02) (00)

MB

t1

0004a 0003b 0000 0001 0008a 0001 0000 0000

(40) (24) (05) (08) (28) (04) (03) (03)

Tangt1 0097a 0041 0033a 0038b 0127a 0048 0059a 0033

(45) (16) (27) (25) (49) (14) (43) (16)

Assetst1 0014a 0013b 0024a 0019a 0037a 0023c 0036a 0025a

(30) (20) (84) (48) (45) (17) (88) (43)

Constant 0030b 0036 0000 0003 0028c 0014 0021 0001

(20) (16) (00) (07) (17) (06) (11) (01)

Year FE Yes Yes Yes Yes Yes Yes Yes Yes

Firm FE No Yes No Yes No Yes No Yes

R2 0006 0088 0009 0117 0026 0251 0023 0202

N 56820 56820 58200 58200 23462 23462 23511 23511

1446 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

conclusions The traditional leverage result on profits is driven by increasesin equity that are experienced and then partially but not completely offsetby firmsTo the extent that firms are engaging in active financing they are following

the prescriptions of the trade-off theory Thus we have a coherent picture ofhow firms are responding to profits Basically more profitable firms experiencean increase in the value of their equity They respond to this by issuing debtand repurchasing equity But the magnitude of the increase in equity is largerthan the magnitudes of the active steps taken in the debt and equity marketsThis means that the firmsrsquo actual actions are as predicted by the trade-offtheory But it leaves open the question of why they do not act more stronglyThe capital structure literature has long been interested in dynamics and

capital structure adjustment paths In recent years this interest has increas-ingly resulted in explicit dynamic optimizing models such as those ofGoldstein Ju and Leland (2001) Hennessy and Whited (2005) andStrebulaev (2007) These models commonly make specialized assumptionsleading some to question their generalityWe argue that both fixed transactions costs and variable transactions costs

are plausible and realistic in models that pin down firm size The simultaneouspresence of both is important Due to fixed costs firms only rebalance inter-mittently Due to variable adjustment costs the extent of the adjustment isreduced to the point at which an extra unit to adjustment is balanced againstthe extra unit transactions costs This means that both leverage increases andleverage reductions will only go part way toward a static optimumThe key point is that with both fixed and variable transactions costs we

should expect to only see intermittent leverage adjustments When leverage isadjusted it should ldquounder adjustrdquo when compared to a frictionless modelFurther discussion of transactions costs in a capital structure setting isprovided by Leary and Roberts (2005)

9 Conclusion

The connection between corporate profits and capital structure has beenvery influential in the assessment of the trade-off theory The conventionalinterpretation of the evidence has pushed the literature away from the trade-off theory and toward much more complex dynamic models and ideas As aresult it is important to make sure that the evidence has been correctlyinterpreted We show that the problem is with the interpretation of thetest results and a failure to take into account the costs associated with re-sponding to a profit shock

PROFITSndashLEVERAGE PUZZLEREVISITED 1447

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

We document that more profitable firms really do borrow more (not less)They tend to repurchase equity They experience an increase in both thebook value of equity and the market value of equity Less profitable firmsreally do tend to reduce their debt and to issue equityFirm size and market conditions also matter Larger firms tend to be more

active in the debt markets while smaller firms tend to be relatively moreactive in the equity markets During good times there is more use ofexternal financingThe usual profitsndashleverage puzzle result is primarily driven by the increase

in equity that is experienced by the more profitable firms Accordingly thepuzzle should be restated as asking why do firms not take sufficiently largeoffsetting actions to fully undo the change in equity What limits the mag-nitudes of the typical leverage response to profit shocksIn a frictionless model the partial response appears puzzling But there is

good empirical reason to believe that rebalancing entails both fixed andvariable costs and that firm size matters The rebalancing costs can befully avoided by doing nothing Accordingly the firm must decide whetherany given shock is big enough to be worth responding to If it is then thefirm must decide how big a response is called for These technical conditionsare known as ldquovalue matchingrdquo and ldquosmooth pastingrdquo Thus optimizationimplies that some shocks will be ignored Even if the shock is not ignoredthe optimal response will only partially undo the shock The magnitude ofthe leverage response must balance the marginal cost and the marginalbenefit of an extra unit of leverage Since the marginal cost of adjustingleverage is strictly positive the adjustment toward that static leverageoptimum will only go part way This is true both for leverage increasesand for leverage reductionsOne might conjecture that the variable costs would be time varying and

that such costs might have declined over time Since we do not have a fullaccounting of all of the variable costs it is of course impossible to be sureHowever there is some evidence suggesting that at least the directunderwriting fees are remarkably stable over time10

Surely nobody ought to be surprised that more profitable firms are moreinclined to repurchase equity while unprofitable firms tend not to do so Thefact that more profitable firms typically experience an increase in the value of

10 For IPOs Jay Ritter provides important summary evidence as httpbearwarringtonufleduritterIPOs2013Underwritingpdf He finds that for medium-sized deals a remark-

ably large fraction of the deals have a 7 fee year after year Whether this is due tobehavioral factors or some sort of a coordination device is not clear The extent towhich similar intertemporal stability carries over to other forms of external financemerits further study

1448 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

equity is equally natural Were it not for the contrary claims in the literature(ie the profitsndashleverage puzzle) we would argue that nobody ought to findit surprising that more profitable firms are more inclined to issue debt Thefact that leverage adjustments typically only partially offset profit shocksshould also sound quite natural These facts all fit together easily within thetrade-off theory once we take account of both fixed and variable costs ofactively adjusting capital structure

Appendix

Table AI Market leverage regressions

The table presents estimates of the regressions of market leverage on firm characteristics

The sample comes from the annual Compustat files during the period 1971ndash2009 Financialfirms are excluded The leverage is estimated as the ratio of debt over debt plus marketequity The explanatory variables Profitabilityt1

MB

t1

Tangibilityt1 and LnethAssetsTHORNt1are described in Table 1 IndMedianLevt1 is estimated as the median market leverage of allother firms in the same industry The industry is defined at the level of the four-digit SICcode All specifications include the year fixed effects The specifications in column (2)

additionally include the firm fixed effects We report t-statistics where the standard errorsare clustered at the firm level in parentheses

aSignificant at the 1 level

Quantile regressions

(1) (2) (3) (4) (5)

Profitabilityt1 013a 013a 007a 011a 020a

(275) (262) (349) (425) (538)

IndLevt1 046a 020a 032a 058a 063a

(506) (239) (1299) (1630) (1042)MB

t1

003a 001a 001a 002a 004a

(509) (303) (743) (709) (534)

Tangibilityt1 013a 018a 013a 014a 010a

(191) (181) (762) (571) (232)

LnethAssetsTHORNt1 001a 005a 001a 001a lt001a

(112) (258) (682) (342) (29)

Constant 010a 011a 005a 005a 029a

(171) (98) (173) (118) (398)

Year FE Yes Yes Yes Yes Yes

Firm FE No Yes No No No

Regression type OLS FE 25th tile Median 75thtile

Clustered SE Yes Yes No No No

R2 030 067

Pseudo R2 016 023 0201

N 158578 158578 158578 158578 158578

PROFITSndashLEVERAGE PUZZLEREVISITED 1449

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

Table AII Profitability sorts for firms issuing large amounts

Percentage of firms

Iss Ret Iss Iss Rep Iss Iss IssD IssE

D D NetD E E NetE DampE RepE RetD

(1) (2) (3) (4) (5) (6) (7) (8) (9)

Panel A 10 cutoff to define issuing or retiring activity

Allfirm

s

Low 205 150 134 260 05 257 45 00 23

2 207 175 103 77 06 75 11 01 10

3 243 196 101 50 08 47 08 01 09

4 252 186 113 50 13 48 08 02 08

High 227 161 118 60 27 57 07 04 09

t Highfrac14Low 70 38 62 755 226 768 320 90 152

Smallfirm

s Low 230 132 162 380 05 375 82 00 32

2 228 143 165 334 04 333 65 00 26

3 211 146 147 216 03 214 38 00 23

4 190 154 120 126 04 124 15 01 20

High 194 162 114 99 09 97 12 01 15

t Highfrac14Low 42 40 68 342 20 340 169 07 131

Largefirm

s Low 231 200 82 31 06 29 08 01 04

2 231 187 67 17 06 14 03 01 02

3 234 177 77 17 09 14 03 02 03

4 218 158 75 16 17 14 02 03 02

High 218 135 99 19 41 15 03 07 02

t Highfrac14Low 19 106 36 46 135 57 39 58 23

Percentage of firms

Iss Ret Iss Iss Rep Iss Both Iss D Iss E

D D Net D E E Net E DampE Rep E Ret D

(1) (2) (3) (4) (5) (6) (7) (8) (9)

Panel B 15 cutoff to define issuing or retiring activity

Allfirm

s

Low 158 101 100 230 01 228 27 00 15

2 142 115 63 57 01 56 05 00 06

3 165 128 57 32 01 31 03 00 05

4 173 118 65 35 01 34 03 00 05

High 160 102 70 46 04 44 03 00 05

t Highfrac14Low 08 06 141 738 87 746 277 24 131

Smallfirm

s Low 195 95 132 354 01 350 56 00 24

2 185 96 131 303 01 301 44 00 18

3 160 96 104 179 00 176 18 00 15

4 134 99 78 99 01 98 08 00 12

High 133 95 73 82 01 80 06 00 10

t Highfrac14Low 80 00 95 345 12 344 149 12 52

Largefirm

s Low 145 123 42 14 01 12 03 00 01

2 137 114 32 07 01 06 01 00 01

3 142 104 35 06 01 05 00 00 01

4 130 91 36 08 03 07 01 01 01

High 134 81 48 09 06 07 01 01 01

t Highfrac14Low 19 85 17 26 54 33 34 17 11

1450 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

Table AIV Debt and equity changesmdashwithout fixed effects

The sample contains nonfinancial firms listed on the annual Compustat files for the period from 1971

to 2009 The table presents estimates from regressions of change in debt (D) change in the book

value of equity (BVE) change in the market value of equity (MVE) and net equity issuances

(EqyIss) The explanatory variables include current and lagged change in operating income before

depreciation (Profits and Profitst1) and lagged change in book value of assets (Assetst1) The

regressions include year indicator variables The reported t-statistics are corrected for clustering at the

firm level

aMean significant at the 1 level

bMean significant at the 5 level

cMean significant at the 10 level

D BVE MVE EqyIss

(1) (2) (3) (4) (5) (6) (7) (8)

Profitst1 025a 024a 023a 038a 007 066a 007a 009a

(40) (32) (36) (39) (06) (29) (42) (51)Profits 051b 082a 149a 004a

(25) (33) (49) (33)Assetst1 002 005c 024a 001b

(08) (17) (34) (23)Constant 93 176a 45 71 1552a 1704a 130a 140a

(12) (28) (07) (09) (49) (49) (45) (45)Year FE Yes Yes Yes Yes Yes Yes Yes YesR2 001 005 001 011 001 005 002 002N 162056 162056 162130 162130 157550 157550 162154 162154

Table AIII Median profitability sorts

The sample contains nonfinancial firms listed on the annual Compustat files for the period from 1971

to 2009 The table presents median debt and equity median changes in debt and equity and median

equity issuances (EqyIss) in millions of dollars for all firms sorted on profitability and for profitability

sorts within the smallest and largest firms The sorts are done annually

D D BVE BVE MVE MVE EqyIss Assets(1) (2) (3) (4) (5) (6) (7) (8)

Allfirm

s

Low 21 00 75 17 256 20 01 1922 186 02 449 18 598 19 00 11033 816 06 1274 03 1637 05 00 32244 787 02 1541 18 2488 03 00 3556High 273 00 1361 66 3300 14 00 2578

Smallfirm

s Low 05 00 13 07 140 14 06 402 06 00 25 08 114 12 01 563 09 00 32 05 85 07 00 674 12 00 39 01 68 03 00 79High 10 00 51 04 98 01 00 90

Largefirm

s Low 11485 381 10269 371 13227 202 03 333902 13051 393 12786 140 17571 169 05 392463 11554 287 13429 24 21623 305 00 370314 9963 209 15052 194 29424 308 00 36705High 7361 00 16380 613 49142 439 00 34976

PROFITSndashLEVERAGE PUZZLEREVISITED 1451

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

References

Auerbach A J (1985) Real determinants of corporate leverage in B M Friedman (ed)

Corporate Capital Structures in the United States University of Chicago Press Chicago

IllinoisBertrand M and Mullainathan S (2003) Enjoying the quiet life corporate governance and

managerial preferences Journal of Political Economy 111 1043ndash1075Booth L Varouj A Demirguc-Kunt A and Maksimovic V (2001) Capital structures in

developing countries Journal of Finance 56 87ndash130Brennan M J and Schwartz E S (1984) Optimal financial policy and firm valuation

Journal of Finance 39 593ndash607Cameron A C and Trivedi P K (2010) Microeconometrics using Stata Stata Press TexasConstantinides G M (1979) Multiperiod consumption and investment behavior with

convex transactions costs Management Science 25 1127ndash1137Constantinides G M and Richard S F (1978) Existence of optimal simple policies for

discounted-cost inventory and cash management in continuous time Operations Research

26 620ndash636DeAngelo H DeAngelo L and Whited T M (2011) Capital structure dynamics and

transitory debt Journal of Financial Economics 99 235ndash261Dixit A K (1993) The Art of Smooth Pasting Routledge London and New YorkFama E and French K R (2002) Testing trade-off and pecking order predictions about

dividends and debt Review of Financial Studies 15 1ndash33Faulkender M and Petersen M A (2006) Does the source of capital affect capital struc-

ture Review of Financial Studies 19 45ndash79Faulkender M Flannery M J Hankins K W and Smith J M (2012) Cash Flows and

Leverage Adjustments Journal of Financial Economics 103 632ndash646Fischer E Heinkel R and Zechner J (1989) Dynamic capital structure choice theory

and tests Journal of Finance 44 19ndash40Frank M Z and Goyal V K (2008) Trade-off and pecking order theories of

debt in B E Eckbo (ed) Handbook of Corporate Finance Empirical Corporate Finance

Vol 2 Handbook of Finance Series Chapter 12 ElsevierNorth-Holland Amsterdam pp

135ndash202

Frank M Z and Goyal V K (2009) Capital structure decisions which factors are reliably

important Financial Management 38 1ndash37

Garvey G T and Hanka G (1999) Capital structure and corporate control the effect of

antitakeover statutes on firm leverage Journal of Finance 54 519ndash546

Goldstein R Ju N and Leland H (2001) An EBIT-based model of dynamic capital

structure Journal of Business 74 483ndash512

Graham J R and Tucker A L (2006) Tax shelters and corporate debt policy Journal of

Financial Economics 81 563ndash594

Hackbarth D and Mauer D C (2012) Optimal priority structure capital structure and

investment Review of Financial Studies 25 747ndash796

Halling M Yu J and Zechner J (2012) Leverage dynamics over the business cycle

Working paper University of Utah University of New South Wales and Vienna

University of Economics and Business AdministrationHennessy C A and Whited T A (2005) Debt dynamics Journal of Finance 60 1129ndash1165

Huang R and Ritter J R (2009) Testing theories of capital structure and estimating the

speed of adjustment Journal of Financial and Qualitative Analysis 44 237ndash271

1452 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

Hovakimian A Opler T and Titman S (2001) The debt-equity choice Journal of

Financial and Quantitative Analysis 36 1ndash24Jung K Kim Y and Stulz R M (1996) Timing investment opportunities managerial

discretion and the security issue decisions Journal of Financial Economics 42 159ndash185Kane A Marcus A J and McDonald R L (1984) How big is the tax advantage to

debt Journal of Finance 39 841ndash853Kayhan A and Titman S (2007) Firmsrsquo histories and their capital structures Journal of

Financial Economics 83 1ndash32Leary M T (2009) Bank loan supply lender choice and corporate capital structure

Journal of Finance 64 1143ndash1185Leary M T and Roberts M R (2005) Do firms rebalance their capital structures

Journal of Finance 60 2575ndash2619Leland H E (1994) Corporate debt value bond covenants and optimal capital structure

Journal of finance 49 1213ndash1252Lemmon M L Roberts M R and Zender J F (2008) Back to the beginning persist-

ence and the cross-section of corporate capital structure Journal of Finance 63 1575ndash1608Long M S and Malitz I B (1985) The investment-financing nexus some empirical

evidence Midland Corporate Finance Journal 3 53ndash59Mackie-Mason J K (1990) Do taxes affect corporate financing decisions Journal of

Finance 45 1471ndash1493Morellec E Nikolov B and Schurhoff N (2012) Corporate governance and capital

structure dynamics Journal of Finance 67 803ndash848Myers S C (1984) The capital structure puzzle Journal of Finance 39 575ndash592Myers S C (1993) Still searching for optimal capital structure Journal of Applied

Corporate Finance 61 4ndash14Rajan R and Zingales L (1995) What do we know about capital structure some evidence

from international data Journal of Finance 50 1421ndash1460Strebulaev I A (2007) Do tests of capital structure theory mean what they say Journal of

Finance 62 1747ndash1787

Strebulaev I A and Whited T M (2012) Dynamic Models and Structural Estimation in

Corporate Finance Now Publisher Boston and Delft

Titman S and Wessels R (1988) The determinants of capital structure choice Journal of

Finance 43 1ndash21

Welch I (2004) Capital structure and stock returns Journal of Political Economy 112

106ndash131

Welch I (2011) Two common problems in capital structure research the financial-debt-

to-asset ratio and issuing activity versus leverage changes International Review of Finance

11 1ndash17

PROFITSndashLEVERAGE PUZZLEREVISITED 1453

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

inasmuch as short-term debt is probably more associated with day-to-dayoperations rather than longer term planning considerationsIn Table XIII we examine change in leverage ( D

DthornE) and change inleverage due to external financing (LEVEF) during good and bad timesseparately The traditional negative effect of a change in profits on achange in leverage is found both in good times and in bad timesHowever when we decompose the impact in order to isolate the externalcomponent as before the result disappears In good times the externalfinancing component of a change in profits is weak but still positiveDuring bad times it is not statistically significant This implies that distin-guishing between good and bad times does not fundamentally alter the basic

Table XII Short- and long-term debt issuances

The sample contains nonfinancial firms listed on the annual Compustat files for the period

from 1971 to 2009 The table presents estimates from regressions of short-term debtissuance to total issuance (in columns (1) and (2)) long-term debt issuance to totalissuance (in columns (3) and (4)) short-term debt issuance to total capitalization (in

columns (5) and (6)) and long-term debt to total capitalization (in columns (7) and (8))Some of the specifications include firm fixed effects The reported t-statistics are correctedfor clustering at the firm level

aMean significant at the 1 level

bMean significant at the 5 level

cMean significant at the 10 level

STDebtIssTotIss

LTDebtIssTotIss

STDebtIssethDthornETHORN

LTDebtIssethDthornETHORN

(1) (2) (3) (4) (5) (6) (7) (8)

Profitabilityt1 0002 0003 0042a 0008 0001 0001 0019a 0012a

(02) (04) (30) (05) (14) (10) (45) (28)

IndLevt1 0009 0007 0070b 0033 0001 0002c 0014a 0005

(06) (05) (26) (11) (08) (17) (26) (09)

M=Bt1 0003a 0003a 0003b 0010a 0000 0000 0003a 0002a

(48) (35) (22) (65) (10) (01) (84) (42)

Tangibilityt1 0020 0007 0011 0067 0005a 0006a 0097a 0081a

(12) (03) (03) (16) (34) (33) (108) (89)

Assetst1 0039a 0011b 0034a 0024b 0001a 0001 0046a 0020a

(89) (22) (38) (23) (36) (13) (190) (82)

Constant 0001b 0153a 0798a 0591a 0000c 0006a 0057a 0096a

(25) (144) (470) (288) (17) (124) (292) (234)

Year FE Yes Yes Yes Yes Yes Yes Yes Yes

Firm FE No Yes No Yes No Yes No Yes

R2 0025 0089 0018 0073 0030 0147 0016 0298

N 136146 136146 136140 136140 143031 143031 142847 142847

PROFITSndashLEVERAGE PUZZLEREVISITED 1445

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

Table XIII Externally financed leverage changes in good and bad times

The sample contains nonfinancial firms listed on the annual Compustat files for the period

from 1971 to 2009 The table presents estimates from regressions of change in debt tocapitalization ratio (in columns (1) (2) (5) and (6)) and change in leverage ratio consider-ing only external financing (in columns (3) (4) (7) and (8)) The change in leverage ratio

considering only external financing is defined as

LevEF frac14Dt1 thorn d

Dt1 thorn Et1 thorn dthorn e

Dt1

Dt1 thorn Et1

where D is debt E is book value of equity d is debt issues net of retirements and e is equityissues net of repurchases The regressions include lagged leverage factors and year indicator

variables The reported t-statistics are corrected for clustering at the firm level An industryis defined as having ldquogood timesrdquo if the median firm in that industry has a market-to-bookratio that is higher than the 67th percentile of the time-series distribution of industry

median market-to-book ratios Conversely an industry is defined as having ldquobad timesrdquoif the median firm in that industry has a market-to-book ratio that is lower than the 33rdpercentile of the time-series distribution of industry median market-to-book ratios

aMean significant at the 1 level

bMean significant at the 5 level

cMean significant at the 10 level

Good times Bad times

DDthornE LevEF D

DthornE LevEF

(1) (2) (3) (4) (5) (6) (7) (8)

Profitabiltyt1 0022c 0005 0016c 0025b 0073b 0041 0007 0007

(17) (03) (18) (24) (25) (11) (04) (03)

IndLevt1 0000 0011 0001 0009 0010 0018 0001 0000

(00) (09) (01) (13) (07) (10) (02) (00)

MB

t1

0004a 0003b 0000 0001 0008a 0001 0000 0000

(40) (24) (05) (08) (28) (04) (03) (03)

Tangt1 0097a 0041 0033a 0038b 0127a 0048 0059a 0033

(45) (16) (27) (25) (49) (14) (43) (16)

Assetst1 0014a 0013b 0024a 0019a 0037a 0023c 0036a 0025a

(30) (20) (84) (48) (45) (17) (88) (43)

Constant 0030b 0036 0000 0003 0028c 0014 0021 0001

(20) (16) (00) (07) (17) (06) (11) (01)

Year FE Yes Yes Yes Yes Yes Yes Yes Yes

Firm FE No Yes No Yes No Yes No Yes

R2 0006 0088 0009 0117 0026 0251 0023 0202

N 56820 56820 58200 58200 23462 23462 23511 23511

1446 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

conclusions The traditional leverage result on profits is driven by increasesin equity that are experienced and then partially but not completely offsetby firmsTo the extent that firms are engaging in active financing they are following

the prescriptions of the trade-off theory Thus we have a coherent picture ofhow firms are responding to profits Basically more profitable firms experiencean increase in the value of their equity They respond to this by issuing debtand repurchasing equity But the magnitude of the increase in equity is largerthan the magnitudes of the active steps taken in the debt and equity marketsThis means that the firmsrsquo actual actions are as predicted by the trade-offtheory But it leaves open the question of why they do not act more stronglyThe capital structure literature has long been interested in dynamics and

capital structure adjustment paths In recent years this interest has increas-ingly resulted in explicit dynamic optimizing models such as those ofGoldstein Ju and Leland (2001) Hennessy and Whited (2005) andStrebulaev (2007) These models commonly make specialized assumptionsleading some to question their generalityWe argue that both fixed transactions costs and variable transactions costs

are plausible and realistic in models that pin down firm size The simultaneouspresence of both is important Due to fixed costs firms only rebalance inter-mittently Due to variable adjustment costs the extent of the adjustment isreduced to the point at which an extra unit to adjustment is balanced againstthe extra unit transactions costs This means that both leverage increases andleverage reductions will only go part way toward a static optimumThe key point is that with both fixed and variable transactions costs we

should expect to only see intermittent leverage adjustments When leverage isadjusted it should ldquounder adjustrdquo when compared to a frictionless modelFurther discussion of transactions costs in a capital structure setting isprovided by Leary and Roberts (2005)

9 Conclusion

The connection between corporate profits and capital structure has beenvery influential in the assessment of the trade-off theory The conventionalinterpretation of the evidence has pushed the literature away from the trade-off theory and toward much more complex dynamic models and ideas As aresult it is important to make sure that the evidence has been correctlyinterpreted We show that the problem is with the interpretation of thetest results and a failure to take into account the costs associated with re-sponding to a profit shock

PROFITSndashLEVERAGE PUZZLEREVISITED 1447

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

We document that more profitable firms really do borrow more (not less)They tend to repurchase equity They experience an increase in both thebook value of equity and the market value of equity Less profitable firmsreally do tend to reduce their debt and to issue equityFirm size and market conditions also matter Larger firms tend to be more

active in the debt markets while smaller firms tend to be relatively moreactive in the equity markets During good times there is more use ofexternal financingThe usual profitsndashleverage puzzle result is primarily driven by the increase

in equity that is experienced by the more profitable firms Accordingly thepuzzle should be restated as asking why do firms not take sufficiently largeoffsetting actions to fully undo the change in equity What limits the mag-nitudes of the typical leverage response to profit shocksIn a frictionless model the partial response appears puzzling But there is

good empirical reason to believe that rebalancing entails both fixed andvariable costs and that firm size matters The rebalancing costs can befully avoided by doing nothing Accordingly the firm must decide whetherany given shock is big enough to be worth responding to If it is then thefirm must decide how big a response is called for These technical conditionsare known as ldquovalue matchingrdquo and ldquosmooth pastingrdquo Thus optimizationimplies that some shocks will be ignored Even if the shock is not ignoredthe optimal response will only partially undo the shock The magnitude ofthe leverage response must balance the marginal cost and the marginalbenefit of an extra unit of leverage Since the marginal cost of adjustingleverage is strictly positive the adjustment toward that static leverageoptimum will only go part way This is true both for leverage increasesand for leverage reductionsOne might conjecture that the variable costs would be time varying and

that such costs might have declined over time Since we do not have a fullaccounting of all of the variable costs it is of course impossible to be sureHowever there is some evidence suggesting that at least the directunderwriting fees are remarkably stable over time10

Surely nobody ought to be surprised that more profitable firms are moreinclined to repurchase equity while unprofitable firms tend not to do so Thefact that more profitable firms typically experience an increase in the value of

10 For IPOs Jay Ritter provides important summary evidence as httpbearwarringtonufleduritterIPOs2013Underwritingpdf He finds that for medium-sized deals a remark-

ably large fraction of the deals have a 7 fee year after year Whether this is due tobehavioral factors or some sort of a coordination device is not clear The extent towhich similar intertemporal stability carries over to other forms of external financemerits further study

1448 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

equity is equally natural Were it not for the contrary claims in the literature(ie the profitsndashleverage puzzle) we would argue that nobody ought to findit surprising that more profitable firms are more inclined to issue debt Thefact that leverage adjustments typically only partially offset profit shocksshould also sound quite natural These facts all fit together easily within thetrade-off theory once we take account of both fixed and variable costs ofactively adjusting capital structure

Appendix

Table AI Market leverage regressions

The table presents estimates of the regressions of market leverage on firm characteristics

The sample comes from the annual Compustat files during the period 1971ndash2009 Financialfirms are excluded The leverage is estimated as the ratio of debt over debt plus marketequity The explanatory variables Profitabilityt1

MB

t1

Tangibilityt1 and LnethAssetsTHORNt1are described in Table 1 IndMedianLevt1 is estimated as the median market leverage of allother firms in the same industry The industry is defined at the level of the four-digit SICcode All specifications include the year fixed effects The specifications in column (2)

additionally include the firm fixed effects We report t-statistics where the standard errorsare clustered at the firm level in parentheses

aSignificant at the 1 level

Quantile regressions

(1) (2) (3) (4) (5)

Profitabilityt1 013a 013a 007a 011a 020a

(275) (262) (349) (425) (538)

IndLevt1 046a 020a 032a 058a 063a

(506) (239) (1299) (1630) (1042)MB

t1

003a 001a 001a 002a 004a

(509) (303) (743) (709) (534)

Tangibilityt1 013a 018a 013a 014a 010a

(191) (181) (762) (571) (232)

LnethAssetsTHORNt1 001a 005a 001a 001a lt001a

(112) (258) (682) (342) (29)

Constant 010a 011a 005a 005a 029a

(171) (98) (173) (118) (398)

Year FE Yes Yes Yes Yes Yes

Firm FE No Yes No No No

Regression type OLS FE 25th tile Median 75thtile

Clustered SE Yes Yes No No No

R2 030 067

Pseudo R2 016 023 0201

N 158578 158578 158578 158578 158578

PROFITSndashLEVERAGE PUZZLEREVISITED 1449

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

Table AII Profitability sorts for firms issuing large amounts

Percentage of firms

Iss Ret Iss Iss Rep Iss Iss IssD IssE

D D NetD E E NetE DampE RepE RetD

(1) (2) (3) (4) (5) (6) (7) (8) (9)

Panel A 10 cutoff to define issuing or retiring activity

Allfirm

s

Low 205 150 134 260 05 257 45 00 23

2 207 175 103 77 06 75 11 01 10

3 243 196 101 50 08 47 08 01 09

4 252 186 113 50 13 48 08 02 08

High 227 161 118 60 27 57 07 04 09

t Highfrac14Low 70 38 62 755 226 768 320 90 152

Smallfirm

s Low 230 132 162 380 05 375 82 00 32

2 228 143 165 334 04 333 65 00 26

3 211 146 147 216 03 214 38 00 23

4 190 154 120 126 04 124 15 01 20

High 194 162 114 99 09 97 12 01 15

t Highfrac14Low 42 40 68 342 20 340 169 07 131

Largefirm

s Low 231 200 82 31 06 29 08 01 04

2 231 187 67 17 06 14 03 01 02

3 234 177 77 17 09 14 03 02 03

4 218 158 75 16 17 14 02 03 02

High 218 135 99 19 41 15 03 07 02

t Highfrac14Low 19 106 36 46 135 57 39 58 23

Percentage of firms

Iss Ret Iss Iss Rep Iss Both Iss D Iss E

D D Net D E E Net E DampE Rep E Ret D

(1) (2) (3) (4) (5) (6) (7) (8) (9)

Panel B 15 cutoff to define issuing or retiring activity

Allfirm

s

Low 158 101 100 230 01 228 27 00 15

2 142 115 63 57 01 56 05 00 06

3 165 128 57 32 01 31 03 00 05

4 173 118 65 35 01 34 03 00 05

High 160 102 70 46 04 44 03 00 05

t Highfrac14Low 08 06 141 738 87 746 277 24 131

Smallfirm

s Low 195 95 132 354 01 350 56 00 24

2 185 96 131 303 01 301 44 00 18

3 160 96 104 179 00 176 18 00 15

4 134 99 78 99 01 98 08 00 12

High 133 95 73 82 01 80 06 00 10

t Highfrac14Low 80 00 95 345 12 344 149 12 52

Largefirm

s Low 145 123 42 14 01 12 03 00 01

2 137 114 32 07 01 06 01 00 01

3 142 104 35 06 01 05 00 00 01

4 130 91 36 08 03 07 01 01 01

High 134 81 48 09 06 07 01 01 01

t Highfrac14Low 19 85 17 26 54 33 34 17 11

1450 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

Table AIV Debt and equity changesmdashwithout fixed effects

The sample contains nonfinancial firms listed on the annual Compustat files for the period from 1971

to 2009 The table presents estimates from regressions of change in debt (D) change in the book

value of equity (BVE) change in the market value of equity (MVE) and net equity issuances

(EqyIss) The explanatory variables include current and lagged change in operating income before

depreciation (Profits and Profitst1) and lagged change in book value of assets (Assetst1) The

regressions include year indicator variables The reported t-statistics are corrected for clustering at the

firm level

aMean significant at the 1 level

bMean significant at the 5 level

cMean significant at the 10 level

D BVE MVE EqyIss

(1) (2) (3) (4) (5) (6) (7) (8)

Profitst1 025a 024a 023a 038a 007 066a 007a 009a

(40) (32) (36) (39) (06) (29) (42) (51)Profits 051b 082a 149a 004a

(25) (33) (49) (33)Assetst1 002 005c 024a 001b

(08) (17) (34) (23)Constant 93 176a 45 71 1552a 1704a 130a 140a

(12) (28) (07) (09) (49) (49) (45) (45)Year FE Yes Yes Yes Yes Yes Yes Yes YesR2 001 005 001 011 001 005 002 002N 162056 162056 162130 162130 157550 157550 162154 162154

Table AIII Median profitability sorts

The sample contains nonfinancial firms listed on the annual Compustat files for the period from 1971

to 2009 The table presents median debt and equity median changes in debt and equity and median

equity issuances (EqyIss) in millions of dollars for all firms sorted on profitability and for profitability

sorts within the smallest and largest firms The sorts are done annually

D D BVE BVE MVE MVE EqyIss Assets(1) (2) (3) (4) (5) (6) (7) (8)

Allfirm

s

Low 21 00 75 17 256 20 01 1922 186 02 449 18 598 19 00 11033 816 06 1274 03 1637 05 00 32244 787 02 1541 18 2488 03 00 3556High 273 00 1361 66 3300 14 00 2578

Smallfirm

s Low 05 00 13 07 140 14 06 402 06 00 25 08 114 12 01 563 09 00 32 05 85 07 00 674 12 00 39 01 68 03 00 79High 10 00 51 04 98 01 00 90

Largefirm

s Low 11485 381 10269 371 13227 202 03 333902 13051 393 12786 140 17571 169 05 392463 11554 287 13429 24 21623 305 00 370314 9963 209 15052 194 29424 308 00 36705High 7361 00 16380 613 49142 439 00 34976

PROFITSndashLEVERAGE PUZZLEREVISITED 1451

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

References

Auerbach A J (1985) Real determinants of corporate leverage in B M Friedman (ed)

Corporate Capital Structures in the United States University of Chicago Press Chicago

IllinoisBertrand M and Mullainathan S (2003) Enjoying the quiet life corporate governance and

managerial preferences Journal of Political Economy 111 1043ndash1075Booth L Varouj A Demirguc-Kunt A and Maksimovic V (2001) Capital structures in

developing countries Journal of Finance 56 87ndash130Brennan M J and Schwartz E S (1984) Optimal financial policy and firm valuation

Journal of Finance 39 593ndash607Cameron A C and Trivedi P K (2010) Microeconometrics using Stata Stata Press TexasConstantinides G M (1979) Multiperiod consumption and investment behavior with

convex transactions costs Management Science 25 1127ndash1137Constantinides G M and Richard S F (1978) Existence of optimal simple policies for

discounted-cost inventory and cash management in continuous time Operations Research

26 620ndash636DeAngelo H DeAngelo L and Whited T M (2011) Capital structure dynamics and

transitory debt Journal of Financial Economics 99 235ndash261Dixit A K (1993) The Art of Smooth Pasting Routledge London and New YorkFama E and French K R (2002) Testing trade-off and pecking order predictions about

dividends and debt Review of Financial Studies 15 1ndash33Faulkender M and Petersen M A (2006) Does the source of capital affect capital struc-

ture Review of Financial Studies 19 45ndash79Faulkender M Flannery M J Hankins K W and Smith J M (2012) Cash Flows and

Leverage Adjustments Journal of Financial Economics 103 632ndash646Fischer E Heinkel R and Zechner J (1989) Dynamic capital structure choice theory

and tests Journal of Finance 44 19ndash40Frank M Z and Goyal V K (2008) Trade-off and pecking order theories of

debt in B E Eckbo (ed) Handbook of Corporate Finance Empirical Corporate Finance

Vol 2 Handbook of Finance Series Chapter 12 ElsevierNorth-Holland Amsterdam pp

135ndash202

Frank M Z and Goyal V K (2009) Capital structure decisions which factors are reliably

important Financial Management 38 1ndash37

Garvey G T and Hanka G (1999) Capital structure and corporate control the effect of

antitakeover statutes on firm leverage Journal of Finance 54 519ndash546

Goldstein R Ju N and Leland H (2001) An EBIT-based model of dynamic capital

structure Journal of Business 74 483ndash512

Graham J R and Tucker A L (2006) Tax shelters and corporate debt policy Journal of

Financial Economics 81 563ndash594

Hackbarth D and Mauer D C (2012) Optimal priority structure capital structure and

investment Review of Financial Studies 25 747ndash796

Halling M Yu J and Zechner J (2012) Leverage dynamics over the business cycle

Working paper University of Utah University of New South Wales and Vienna

University of Economics and Business AdministrationHennessy C A and Whited T A (2005) Debt dynamics Journal of Finance 60 1129ndash1165

Huang R and Ritter J R (2009) Testing theories of capital structure and estimating the

speed of adjustment Journal of Financial and Qualitative Analysis 44 237ndash271

1452 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

Hovakimian A Opler T and Titman S (2001) The debt-equity choice Journal of

Financial and Quantitative Analysis 36 1ndash24Jung K Kim Y and Stulz R M (1996) Timing investment opportunities managerial

discretion and the security issue decisions Journal of Financial Economics 42 159ndash185Kane A Marcus A J and McDonald R L (1984) How big is the tax advantage to

debt Journal of Finance 39 841ndash853Kayhan A and Titman S (2007) Firmsrsquo histories and their capital structures Journal of

Financial Economics 83 1ndash32Leary M T (2009) Bank loan supply lender choice and corporate capital structure

Journal of Finance 64 1143ndash1185Leary M T and Roberts M R (2005) Do firms rebalance their capital structures

Journal of Finance 60 2575ndash2619Leland H E (1994) Corporate debt value bond covenants and optimal capital structure

Journal of finance 49 1213ndash1252Lemmon M L Roberts M R and Zender J F (2008) Back to the beginning persist-

ence and the cross-section of corporate capital structure Journal of Finance 63 1575ndash1608Long M S and Malitz I B (1985) The investment-financing nexus some empirical

evidence Midland Corporate Finance Journal 3 53ndash59Mackie-Mason J K (1990) Do taxes affect corporate financing decisions Journal of

Finance 45 1471ndash1493Morellec E Nikolov B and Schurhoff N (2012) Corporate governance and capital

structure dynamics Journal of Finance 67 803ndash848Myers S C (1984) The capital structure puzzle Journal of Finance 39 575ndash592Myers S C (1993) Still searching for optimal capital structure Journal of Applied

Corporate Finance 61 4ndash14Rajan R and Zingales L (1995) What do we know about capital structure some evidence

from international data Journal of Finance 50 1421ndash1460Strebulaev I A (2007) Do tests of capital structure theory mean what they say Journal of

Finance 62 1747ndash1787

Strebulaev I A and Whited T M (2012) Dynamic Models and Structural Estimation in

Corporate Finance Now Publisher Boston and Delft

Titman S and Wessels R (1988) The determinants of capital structure choice Journal of

Finance 43 1ndash21

Welch I (2004) Capital structure and stock returns Journal of Political Economy 112

106ndash131

Welch I (2011) Two common problems in capital structure research the financial-debt-

to-asset ratio and issuing activity versus leverage changes International Review of Finance

11 1ndash17

PROFITSndashLEVERAGE PUZZLEREVISITED 1453

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

Table XIII Externally financed leverage changes in good and bad times

The sample contains nonfinancial firms listed on the annual Compustat files for the period

from 1971 to 2009 The table presents estimates from regressions of change in debt tocapitalization ratio (in columns (1) (2) (5) and (6)) and change in leverage ratio consider-ing only external financing (in columns (3) (4) (7) and (8)) The change in leverage ratio

considering only external financing is defined as

LevEF frac14Dt1 thorn d

Dt1 thorn Et1 thorn dthorn e

Dt1

Dt1 thorn Et1

where D is debt E is book value of equity d is debt issues net of retirements and e is equityissues net of repurchases The regressions include lagged leverage factors and year indicator

variables The reported t-statistics are corrected for clustering at the firm level An industryis defined as having ldquogood timesrdquo if the median firm in that industry has a market-to-bookratio that is higher than the 67th percentile of the time-series distribution of industry

median market-to-book ratios Conversely an industry is defined as having ldquobad timesrdquoif the median firm in that industry has a market-to-book ratio that is lower than the 33rdpercentile of the time-series distribution of industry median market-to-book ratios

aMean significant at the 1 level

bMean significant at the 5 level

cMean significant at the 10 level

Good times Bad times

DDthornE LevEF D

DthornE LevEF

(1) (2) (3) (4) (5) (6) (7) (8)

Profitabiltyt1 0022c 0005 0016c 0025b 0073b 0041 0007 0007

(17) (03) (18) (24) (25) (11) (04) (03)

IndLevt1 0000 0011 0001 0009 0010 0018 0001 0000

(00) (09) (01) (13) (07) (10) (02) (00)

MB

t1

0004a 0003b 0000 0001 0008a 0001 0000 0000

(40) (24) (05) (08) (28) (04) (03) (03)

Tangt1 0097a 0041 0033a 0038b 0127a 0048 0059a 0033

(45) (16) (27) (25) (49) (14) (43) (16)

Assetst1 0014a 0013b 0024a 0019a 0037a 0023c 0036a 0025a

(30) (20) (84) (48) (45) (17) (88) (43)

Constant 0030b 0036 0000 0003 0028c 0014 0021 0001

(20) (16) (00) (07) (17) (06) (11) (01)

Year FE Yes Yes Yes Yes Yes Yes Yes Yes

Firm FE No Yes No Yes No Yes No Yes

R2 0006 0088 0009 0117 0026 0251 0023 0202

N 56820 56820 58200 58200 23462 23462 23511 23511

1446 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

conclusions The traditional leverage result on profits is driven by increasesin equity that are experienced and then partially but not completely offsetby firmsTo the extent that firms are engaging in active financing they are following

the prescriptions of the trade-off theory Thus we have a coherent picture ofhow firms are responding to profits Basically more profitable firms experiencean increase in the value of their equity They respond to this by issuing debtand repurchasing equity But the magnitude of the increase in equity is largerthan the magnitudes of the active steps taken in the debt and equity marketsThis means that the firmsrsquo actual actions are as predicted by the trade-offtheory But it leaves open the question of why they do not act more stronglyThe capital structure literature has long been interested in dynamics and

capital structure adjustment paths In recent years this interest has increas-ingly resulted in explicit dynamic optimizing models such as those ofGoldstein Ju and Leland (2001) Hennessy and Whited (2005) andStrebulaev (2007) These models commonly make specialized assumptionsleading some to question their generalityWe argue that both fixed transactions costs and variable transactions costs

are plausible and realistic in models that pin down firm size The simultaneouspresence of both is important Due to fixed costs firms only rebalance inter-mittently Due to variable adjustment costs the extent of the adjustment isreduced to the point at which an extra unit to adjustment is balanced againstthe extra unit transactions costs This means that both leverage increases andleverage reductions will only go part way toward a static optimumThe key point is that with both fixed and variable transactions costs we

should expect to only see intermittent leverage adjustments When leverage isadjusted it should ldquounder adjustrdquo when compared to a frictionless modelFurther discussion of transactions costs in a capital structure setting isprovided by Leary and Roberts (2005)

9 Conclusion

The connection between corporate profits and capital structure has beenvery influential in the assessment of the trade-off theory The conventionalinterpretation of the evidence has pushed the literature away from the trade-off theory and toward much more complex dynamic models and ideas As aresult it is important to make sure that the evidence has been correctlyinterpreted We show that the problem is with the interpretation of thetest results and a failure to take into account the costs associated with re-sponding to a profit shock

PROFITSndashLEVERAGE PUZZLEREVISITED 1447

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

We document that more profitable firms really do borrow more (not less)They tend to repurchase equity They experience an increase in both thebook value of equity and the market value of equity Less profitable firmsreally do tend to reduce their debt and to issue equityFirm size and market conditions also matter Larger firms tend to be more

active in the debt markets while smaller firms tend to be relatively moreactive in the equity markets During good times there is more use ofexternal financingThe usual profitsndashleverage puzzle result is primarily driven by the increase

in equity that is experienced by the more profitable firms Accordingly thepuzzle should be restated as asking why do firms not take sufficiently largeoffsetting actions to fully undo the change in equity What limits the mag-nitudes of the typical leverage response to profit shocksIn a frictionless model the partial response appears puzzling But there is

good empirical reason to believe that rebalancing entails both fixed andvariable costs and that firm size matters The rebalancing costs can befully avoided by doing nothing Accordingly the firm must decide whetherany given shock is big enough to be worth responding to If it is then thefirm must decide how big a response is called for These technical conditionsare known as ldquovalue matchingrdquo and ldquosmooth pastingrdquo Thus optimizationimplies that some shocks will be ignored Even if the shock is not ignoredthe optimal response will only partially undo the shock The magnitude ofthe leverage response must balance the marginal cost and the marginalbenefit of an extra unit of leverage Since the marginal cost of adjustingleverage is strictly positive the adjustment toward that static leverageoptimum will only go part way This is true both for leverage increasesand for leverage reductionsOne might conjecture that the variable costs would be time varying and

that such costs might have declined over time Since we do not have a fullaccounting of all of the variable costs it is of course impossible to be sureHowever there is some evidence suggesting that at least the directunderwriting fees are remarkably stable over time10

Surely nobody ought to be surprised that more profitable firms are moreinclined to repurchase equity while unprofitable firms tend not to do so Thefact that more profitable firms typically experience an increase in the value of

10 For IPOs Jay Ritter provides important summary evidence as httpbearwarringtonufleduritterIPOs2013Underwritingpdf He finds that for medium-sized deals a remark-

ably large fraction of the deals have a 7 fee year after year Whether this is due tobehavioral factors or some sort of a coordination device is not clear The extent towhich similar intertemporal stability carries over to other forms of external financemerits further study

1448 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

equity is equally natural Were it not for the contrary claims in the literature(ie the profitsndashleverage puzzle) we would argue that nobody ought to findit surprising that more profitable firms are more inclined to issue debt Thefact that leverage adjustments typically only partially offset profit shocksshould also sound quite natural These facts all fit together easily within thetrade-off theory once we take account of both fixed and variable costs ofactively adjusting capital structure

Appendix

Table AI Market leverage regressions

The table presents estimates of the regressions of market leverage on firm characteristics

The sample comes from the annual Compustat files during the period 1971ndash2009 Financialfirms are excluded The leverage is estimated as the ratio of debt over debt plus marketequity The explanatory variables Profitabilityt1

MB

t1

Tangibilityt1 and LnethAssetsTHORNt1are described in Table 1 IndMedianLevt1 is estimated as the median market leverage of allother firms in the same industry The industry is defined at the level of the four-digit SICcode All specifications include the year fixed effects The specifications in column (2)

additionally include the firm fixed effects We report t-statistics where the standard errorsare clustered at the firm level in parentheses

aSignificant at the 1 level

Quantile regressions

(1) (2) (3) (4) (5)

Profitabilityt1 013a 013a 007a 011a 020a

(275) (262) (349) (425) (538)

IndLevt1 046a 020a 032a 058a 063a

(506) (239) (1299) (1630) (1042)MB

t1

003a 001a 001a 002a 004a

(509) (303) (743) (709) (534)

Tangibilityt1 013a 018a 013a 014a 010a

(191) (181) (762) (571) (232)

LnethAssetsTHORNt1 001a 005a 001a 001a lt001a

(112) (258) (682) (342) (29)

Constant 010a 011a 005a 005a 029a

(171) (98) (173) (118) (398)

Year FE Yes Yes Yes Yes Yes

Firm FE No Yes No No No

Regression type OLS FE 25th tile Median 75thtile

Clustered SE Yes Yes No No No

R2 030 067

Pseudo R2 016 023 0201

N 158578 158578 158578 158578 158578

PROFITSndashLEVERAGE PUZZLEREVISITED 1449

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

Table AII Profitability sorts for firms issuing large amounts

Percentage of firms

Iss Ret Iss Iss Rep Iss Iss IssD IssE

D D NetD E E NetE DampE RepE RetD

(1) (2) (3) (4) (5) (6) (7) (8) (9)

Panel A 10 cutoff to define issuing or retiring activity

Allfirm

s

Low 205 150 134 260 05 257 45 00 23

2 207 175 103 77 06 75 11 01 10

3 243 196 101 50 08 47 08 01 09

4 252 186 113 50 13 48 08 02 08

High 227 161 118 60 27 57 07 04 09

t Highfrac14Low 70 38 62 755 226 768 320 90 152

Smallfirm

s Low 230 132 162 380 05 375 82 00 32

2 228 143 165 334 04 333 65 00 26

3 211 146 147 216 03 214 38 00 23

4 190 154 120 126 04 124 15 01 20

High 194 162 114 99 09 97 12 01 15

t Highfrac14Low 42 40 68 342 20 340 169 07 131

Largefirm

s Low 231 200 82 31 06 29 08 01 04

2 231 187 67 17 06 14 03 01 02

3 234 177 77 17 09 14 03 02 03

4 218 158 75 16 17 14 02 03 02

High 218 135 99 19 41 15 03 07 02

t Highfrac14Low 19 106 36 46 135 57 39 58 23

Percentage of firms

Iss Ret Iss Iss Rep Iss Both Iss D Iss E

D D Net D E E Net E DampE Rep E Ret D

(1) (2) (3) (4) (5) (6) (7) (8) (9)

Panel B 15 cutoff to define issuing or retiring activity

Allfirm

s

Low 158 101 100 230 01 228 27 00 15

2 142 115 63 57 01 56 05 00 06

3 165 128 57 32 01 31 03 00 05

4 173 118 65 35 01 34 03 00 05

High 160 102 70 46 04 44 03 00 05

t Highfrac14Low 08 06 141 738 87 746 277 24 131

Smallfirm

s Low 195 95 132 354 01 350 56 00 24

2 185 96 131 303 01 301 44 00 18

3 160 96 104 179 00 176 18 00 15

4 134 99 78 99 01 98 08 00 12

High 133 95 73 82 01 80 06 00 10

t Highfrac14Low 80 00 95 345 12 344 149 12 52

Largefirm

s Low 145 123 42 14 01 12 03 00 01

2 137 114 32 07 01 06 01 00 01

3 142 104 35 06 01 05 00 00 01

4 130 91 36 08 03 07 01 01 01

High 134 81 48 09 06 07 01 01 01

t Highfrac14Low 19 85 17 26 54 33 34 17 11

1450 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

Table AIV Debt and equity changesmdashwithout fixed effects

The sample contains nonfinancial firms listed on the annual Compustat files for the period from 1971

to 2009 The table presents estimates from regressions of change in debt (D) change in the book

value of equity (BVE) change in the market value of equity (MVE) and net equity issuances

(EqyIss) The explanatory variables include current and lagged change in operating income before

depreciation (Profits and Profitst1) and lagged change in book value of assets (Assetst1) The

regressions include year indicator variables The reported t-statistics are corrected for clustering at the

firm level

aMean significant at the 1 level

bMean significant at the 5 level

cMean significant at the 10 level

D BVE MVE EqyIss

(1) (2) (3) (4) (5) (6) (7) (8)

Profitst1 025a 024a 023a 038a 007 066a 007a 009a

(40) (32) (36) (39) (06) (29) (42) (51)Profits 051b 082a 149a 004a

(25) (33) (49) (33)Assetst1 002 005c 024a 001b

(08) (17) (34) (23)Constant 93 176a 45 71 1552a 1704a 130a 140a

(12) (28) (07) (09) (49) (49) (45) (45)Year FE Yes Yes Yes Yes Yes Yes Yes YesR2 001 005 001 011 001 005 002 002N 162056 162056 162130 162130 157550 157550 162154 162154

Table AIII Median profitability sorts

The sample contains nonfinancial firms listed on the annual Compustat files for the period from 1971

to 2009 The table presents median debt and equity median changes in debt and equity and median

equity issuances (EqyIss) in millions of dollars for all firms sorted on profitability and for profitability

sorts within the smallest and largest firms The sorts are done annually

D D BVE BVE MVE MVE EqyIss Assets(1) (2) (3) (4) (5) (6) (7) (8)

Allfirm

s

Low 21 00 75 17 256 20 01 1922 186 02 449 18 598 19 00 11033 816 06 1274 03 1637 05 00 32244 787 02 1541 18 2488 03 00 3556High 273 00 1361 66 3300 14 00 2578

Smallfirm

s Low 05 00 13 07 140 14 06 402 06 00 25 08 114 12 01 563 09 00 32 05 85 07 00 674 12 00 39 01 68 03 00 79High 10 00 51 04 98 01 00 90

Largefirm

s Low 11485 381 10269 371 13227 202 03 333902 13051 393 12786 140 17571 169 05 392463 11554 287 13429 24 21623 305 00 370314 9963 209 15052 194 29424 308 00 36705High 7361 00 16380 613 49142 439 00 34976

PROFITSndashLEVERAGE PUZZLEREVISITED 1451

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

References

Auerbach A J (1985) Real determinants of corporate leverage in B M Friedman (ed)

Corporate Capital Structures in the United States University of Chicago Press Chicago

IllinoisBertrand M and Mullainathan S (2003) Enjoying the quiet life corporate governance and

managerial preferences Journal of Political Economy 111 1043ndash1075Booth L Varouj A Demirguc-Kunt A and Maksimovic V (2001) Capital structures in

developing countries Journal of Finance 56 87ndash130Brennan M J and Schwartz E S (1984) Optimal financial policy and firm valuation

Journal of Finance 39 593ndash607Cameron A C and Trivedi P K (2010) Microeconometrics using Stata Stata Press TexasConstantinides G M (1979) Multiperiod consumption and investment behavior with

convex transactions costs Management Science 25 1127ndash1137Constantinides G M and Richard S F (1978) Existence of optimal simple policies for

discounted-cost inventory and cash management in continuous time Operations Research

26 620ndash636DeAngelo H DeAngelo L and Whited T M (2011) Capital structure dynamics and

transitory debt Journal of Financial Economics 99 235ndash261Dixit A K (1993) The Art of Smooth Pasting Routledge London and New YorkFama E and French K R (2002) Testing trade-off and pecking order predictions about

dividends and debt Review of Financial Studies 15 1ndash33Faulkender M and Petersen M A (2006) Does the source of capital affect capital struc-

ture Review of Financial Studies 19 45ndash79Faulkender M Flannery M J Hankins K W and Smith J M (2012) Cash Flows and

Leverage Adjustments Journal of Financial Economics 103 632ndash646Fischer E Heinkel R and Zechner J (1989) Dynamic capital structure choice theory

and tests Journal of Finance 44 19ndash40Frank M Z and Goyal V K (2008) Trade-off and pecking order theories of

debt in B E Eckbo (ed) Handbook of Corporate Finance Empirical Corporate Finance

Vol 2 Handbook of Finance Series Chapter 12 ElsevierNorth-Holland Amsterdam pp

135ndash202

Frank M Z and Goyal V K (2009) Capital structure decisions which factors are reliably

important Financial Management 38 1ndash37

Garvey G T and Hanka G (1999) Capital structure and corporate control the effect of

antitakeover statutes on firm leverage Journal of Finance 54 519ndash546

Goldstein R Ju N and Leland H (2001) An EBIT-based model of dynamic capital

structure Journal of Business 74 483ndash512

Graham J R and Tucker A L (2006) Tax shelters and corporate debt policy Journal of

Financial Economics 81 563ndash594

Hackbarth D and Mauer D C (2012) Optimal priority structure capital structure and

investment Review of Financial Studies 25 747ndash796

Halling M Yu J and Zechner J (2012) Leverage dynamics over the business cycle

Working paper University of Utah University of New South Wales and Vienna

University of Economics and Business AdministrationHennessy C A and Whited T A (2005) Debt dynamics Journal of Finance 60 1129ndash1165

Huang R and Ritter J R (2009) Testing theories of capital structure and estimating the

speed of adjustment Journal of Financial and Qualitative Analysis 44 237ndash271

1452 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

Hovakimian A Opler T and Titman S (2001) The debt-equity choice Journal of

Financial and Quantitative Analysis 36 1ndash24Jung K Kim Y and Stulz R M (1996) Timing investment opportunities managerial

discretion and the security issue decisions Journal of Financial Economics 42 159ndash185Kane A Marcus A J and McDonald R L (1984) How big is the tax advantage to

debt Journal of Finance 39 841ndash853Kayhan A and Titman S (2007) Firmsrsquo histories and their capital structures Journal of

Financial Economics 83 1ndash32Leary M T (2009) Bank loan supply lender choice and corporate capital structure

Journal of Finance 64 1143ndash1185Leary M T and Roberts M R (2005) Do firms rebalance their capital structures

Journal of Finance 60 2575ndash2619Leland H E (1994) Corporate debt value bond covenants and optimal capital structure

Journal of finance 49 1213ndash1252Lemmon M L Roberts M R and Zender J F (2008) Back to the beginning persist-

ence and the cross-section of corporate capital structure Journal of Finance 63 1575ndash1608Long M S and Malitz I B (1985) The investment-financing nexus some empirical

evidence Midland Corporate Finance Journal 3 53ndash59Mackie-Mason J K (1990) Do taxes affect corporate financing decisions Journal of

Finance 45 1471ndash1493Morellec E Nikolov B and Schurhoff N (2012) Corporate governance and capital

structure dynamics Journal of Finance 67 803ndash848Myers S C (1984) The capital structure puzzle Journal of Finance 39 575ndash592Myers S C (1993) Still searching for optimal capital structure Journal of Applied

Corporate Finance 61 4ndash14Rajan R and Zingales L (1995) What do we know about capital structure some evidence

from international data Journal of Finance 50 1421ndash1460Strebulaev I A (2007) Do tests of capital structure theory mean what they say Journal of

Finance 62 1747ndash1787

Strebulaev I A and Whited T M (2012) Dynamic Models and Structural Estimation in

Corporate Finance Now Publisher Boston and Delft

Titman S and Wessels R (1988) The determinants of capital structure choice Journal of

Finance 43 1ndash21

Welch I (2004) Capital structure and stock returns Journal of Political Economy 112

106ndash131

Welch I (2011) Two common problems in capital structure research the financial-debt-

to-asset ratio and issuing activity versus leverage changes International Review of Finance

11 1ndash17

PROFITSndashLEVERAGE PUZZLEREVISITED 1453

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

conclusions The traditional leverage result on profits is driven by increasesin equity that are experienced and then partially but not completely offsetby firmsTo the extent that firms are engaging in active financing they are following

the prescriptions of the trade-off theory Thus we have a coherent picture ofhow firms are responding to profits Basically more profitable firms experiencean increase in the value of their equity They respond to this by issuing debtand repurchasing equity But the magnitude of the increase in equity is largerthan the magnitudes of the active steps taken in the debt and equity marketsThis means that the firmsrsquo actual actions are as predicted by the trade-offtheory But it leaves open the question of why they do not act more stronglyThe capital structure literature has long been interested in dynamics and

capital structure adjustment paths In recent years this interest has increas-ingly resulted in explicit dynamic optimizing models such as those ofGoldstein Ju and Leland (2001) Hennessy and Whited (2005) andStrebulaev (2007) These models commonly make specialized assumptionsleading some to question their generalityWe argue that both fixed transactions costs and variable transactions costs

are plausible and realistic in models that pin down firm size The simultaneouspresence of both is important Due to fixed costs firms only rebalance inter-mittently Due to variable adjustment costs the extent of the adjustment isreduced to the point at which an extra unit to adjustment is balanced againstthe extra unit transactions costs This means that both leverage increases andleverage reductions will only go part way toward a static optimumThe key point is that with both fixed and variable transactions costs we

should expect to only see intermittent leverage adjustments When leverage isadjusted it should ldquounder adjustrdquo when compared to a frictionless modelFurther discussion of transactions costs in a capital structure setting isprovided by Leary and Roberts (2005)

9 Conclusion

The connection between corporate profits and capital structure has beenvery influential in the assessment of the trade-off theory The conventionalinterpretation of the evidence has pushed the literature away from the trade-off theory and toward much more complex dynamic models and ideas As aresult it is important to make sure that the evidence has been correctlyinterpreted We show that the problem is with the interpretation of thetest results and a failure to take into account the costs associated with re-sponding to a profit shock

PROFITSndashLEVERAGE PUZZLEREVISITED 1447

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

We document that more profitable firms really do borrow more (not less)They tend to repurchase equity They experience an increase in both thebook value of equity and the market value of equity Less profitable firmsreally do tend to reduce their debt and to issue equityFirm size and market conditions also matter Larger firms tend to be more

active in the debt markets while smaller firms tend to be relatively moreactive in the equity markets During good times there is more use ofexternal financingThe usual profitsndashleverage puzzle result is primarily driven by the increase

in equity that is experienced by the more profitable firms Accordingly thepuzzle should be restated as asking why do firms not take sufficiently largeoffsetting actions to fully undo the change in equity What limits the mag-nitudes of the typical leverage response to profit shocksIn a frictionless model the partial response appears puzzling But there is

good empirical reason to believe that rebalancing entails both fixed andvariable costs and that firm size matters The rebalancing costs can befully avoided by doing nothing Accordingly the firm must decide whetherany given shock is big enough to be worth responding to If it is then thefirm must decide how big a response is called for These technical conditionsare known as ldquovalue matchingrdquo and ldquosmooth pastingrdquo Thus optimizationimplies that some shocks will be ignored Even if the shock is not ignoredthe optimal response will only partially undo the shock The magnitude ofthe leverage response must balance the marginal cost and the marginalbenefit of an extra unit of leverage Since the marginal cost of adjustingleverage is strictly positive the adjustment toward that static leverageoptimum will only go part way This is true both for leverage increasesand for leverage reductionsOne might conjecture that the variable costs would be time varying and

that such costs might have declined over time Since we do not have a fullaccounting of all of the variable costs it is of course impossible to be sureHowever there is some evidence suggesting that at least the directunderwriting fees are remarkably stable over time10

Surely nobody ought to be surprised that more profitable firms are moreinclined to repurchase equity while unprofitable firms tend not to do so Thefact that more profitable firms typically experience an increase in the value of

10 For IPOs Jay Ritter provides important summary evidence as httpbearwarringtonufleduritterIPOs2013Underwritingpdf He finds that for medium-sized deals a remark-

ably large fraction of the deals have a 7 fee year after year Whether this is due tobehavioral factors or some sort of a coordination device is not clear The extent towhich similar intertemporal stability carries over to other forms of external financemerits further study

1448 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

equity is equally natural Were it not for the contrary claims in the literature(ie the profitsndashleverage puzzle) we would argue that nobody ought to findit surprising that more profitable firms are more inclined to issue debt Thefact that leverage adjustments typically only partially offset profit shocksshould also sound quite natural These facts all fit together easily within thetrade-off theory once we take account of both fixed and variable costs ofactively adjusting capital structure

Appendix

Table AI Market leverage regressions

The table presents estimates of the regressions of market leverage on firm characteristics

The sample comes from the annual Compustat files during the period 1971ndash2009 Financialfirms are excluded The leverage is estimated as the ratio of debt over debt plus marketequity The explanatory variables Profitabilityt1

MB

t1

Tangibilityt1 and LnethAssetsTHORNt1are described in Table 1 IndMedianLevt1 is estimated as the median market leverage of allother firms in the same industry The industry is defined at the level of the four-digit SICcode All specifications include the year fixed effects The specifications in column (2)

additionally include the firm fixed effects We report t-statistics where the standard errorsare clustered at the firm level in parentheses

aSignificant at the 1 level

Quantile regressions

(1) (2) (3) (4) (5)

Profitabilityt1 013a 013a 007a 011a 020a

(275) (262) (349) (425) (538)

IndLevt1 046a 020a 032a 058a 063a

(506) (239) (1299) (1630) (1042)MB

t1

003a 001a 001a 002a 004a

(509) (303) (743) (709) (534)

Tangibilityt1 013a 018a 013a 014a 010a

(191) (181) (762) (571) (232)

LnethAssetsTHORNt1 001a 005a 001a 001a lt001a

(112) (258) (682) (342) (29)

Constant 010a 011a 005a 005a 029a

(171) (98) (173) (118) (398)

Year FE Yes Yes Yes Yes Yes

Firm FE No Yes No No No

Regression type OLS FE 25th tile Median 75thtile

Clustered SE Yes Yes No No No

R2 030 067

Pseudo R2 016 023 0201

N 158578 158578 158578 158578 158578

PROFITSndashLEVERAGE PUZZLEREVISITED 1449

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

Table AII Profitability sorts for firms issuing large amounts

Percentage of firms

Iss Ret Iss Iss Rep Iss Iss IssD IssE

D D NetD E E NetE DampE RepE RetD

(1) (2) (3) (4) (5) (6) (7) (8) (9)

Panel A 10 cutoff to define issuing or retiring activity

Allfirm

s

Low 205 150 134 260 05 257 45 00 23

2 207 175 103 77 06 75 11 01 10

3 243 196 101 50 08 47 08 01 09

4 252 186 113 50 13 48 08 02 08

High 227 161 118 60 27 57 07 04 09

t Highfrac14Low 70 38 62 755 226 768 320 90 152

Smallfirm

s Low 230 132 162 380 05 375 82 00 32

2 228 143 165 334 04 333 65 00 26

3 211 146 147 216 03 214 38 00 23

4 190 154 120 126 04 124 15 01 20

High 194 162 114 99 09 97 12 01 15

t Highfrac14Low 42 40 68 342 20 340 169 07 131

Largefirm

s Low 231 200 82 31 06 29 08 01 04

2 231 187 67 17 06 14 03 01 02

3 234 177 77 17 09 14 03 02 03

4 218 158 75 16 17 14 02 03 02

High 218 135 99 19 41 15 03 07 02

t Highfrac14Low 19 106 36 46 135 57 39 58 23

Percentage of firms

Iss Ret Iss Iss Rep Iss Both Iss D Iss E

D D Net D E E Net E DampE Rep E Ret D

(1) (2) (3) (4) (5) (6) (7) (8) (9)

Panel B 15 cutoff to define issuing or retiring activity

Allfirm

s

Low 158 101 100 230 01 228 27 00 15

2 142 115 63 57 01 56 05 00 06

3 165 128 57 32 01 31 03 00 05

4 173 118 65 35 01 34 03 00 05

High 160 102 70 46 04 44 03 00 05

t Highfrac14Low 08 06 141 738 87 746 277 24 131

Smallfirm

s Low 195 95 132 354 01 350 56 00 24

2 185 96 131 303 01 301 44 00 18

3 160 96 104 179 00 176 18 00 15

4 134 99 78 99 01 98 08 00 12

High 133 95 73 82 01 80 06 00 10

t Highfrac14Low 80 00 95 345 12 344 149 12 52

Largefirm

s Low 145 123 42 14 01 12 03 00 01

2 137 114 32 07 01 06 01 00 01

3 142 104 35 06 01 05 00 00 01

4 130 91 36 08 03 07 01 01 01

High 134 81 48 09 06 07 01 01 01

t Highfrac14Low 19 85 17 26 54 33 34 17 11

1450 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

Table AIV Debt and equity changesmdashwithout fixed effects

The sample contains nonfinancial firms listed on the annual Compustat files for the period from 1971

to 2009 The table presents estimates from regressions of change in debt (D) change in the book

value of equity (BVE) change in the market value of equity (MVE) and net equity issuances

(EqyIss) The explanatory variables include current and lagged change in operating income before

depreciation (Profits and Profitst1) and lagged change in book value of assets (Assetst1) The

regressions include year indicator variables The reported t-statistics are corrected for clustering at the

firm level

aMean significant at the 1 level

bMean significant at the 5 level

cMean significant at the 10 level

D BVE MVE EqyIss

(1) (2) (3) (4) (5) (6) (7) (8)

Profitst1 025a 024a 023a 038a 007 066a 007a 009a

(40) (32) (36) (39) (06) (29) (42) (51)Profits 051b 082a 149a 004a

(25) (33) (49) (33)Assetst1 002 005c 024a 001b

(08) (17) (34) (23)Constant 93 176a 45 71 1552a 1704a 130a 140a

(12) (28) (07) (09) (49) (49) (45) (45)Year FE Yes Yes Yes Yes Yes Yes Yes YesR2 001 005 001 011 001 005 002 002N 162056 162056 162130 162130 157550 157550 162154 162154

Table AIII Median profitability sorts

The sample contains nonfinancial firms listed on the annual Compustat files for the period from 1971

to 2009 The table presents median debt and equity median changes in debt and equity and median

equity issuances (EqyIss) in millions of dollars for all firms sorted on profitability and for profitability

sorts within the smallest and largest firms The sorts are done annually

D D BVE BVE MVE MVE EqyIss Assets(1) (2) (3) (4) (5) (6) (7) (8)

Allfirm

s

Low 21 00 75 17 256 20 01 1922 186 02 449 18 598 19 00 11033 816 06 1274 03 1637 05 00 32244 787 02 1541 18 2488 03 00 3556High 273 00 1361 66 3300 14 00 2578

Smallfirm

s Low 05 00 13 07 140 14 06 402 06 00 25 08 114 12 01 563 09 00 32 05 85 07 00 674 12 00 39 01 68 03 00 79High 10 00 51 04 98 01 00 90

Largefirm

s Low 11485 381 10269 371 13227 202 03 333902 13051 393 12786 140 17571 169 05 392463 11554 287 13429 24 21623 305 00 370314 9963 209 15052 194 29424 308 00 36705High 7361 00 16380 613 49142 439 00 34976

PROFITSndashLEVERAGE PUZZLEREVISITED 1451

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

References

Auerbach A J (1985) Real determinants of corporate leverage in B M Friedman (ed)

Corporate Capital Structures in the United States University of Chicago Press Chicago

IllinoisBertrand M and Mullainathan S (2003) Enjoying the quiet life corporate governance and

managerial preferences Journal of Political Economy 111 1043ndash1075Booth L Varouj A Demirguc-Kunt A and Maksimovic V (2001) Capital structures in

developing countries Journal of Finance 56 87ndash130Brennan M J and Schwartz E S (1984) Optimal financial policy and firm valuation

Journal of Finance 39 593ndash607Cameron A C and Trivedi P K (2010) Microeconometrics using Stata Stata Press TexasConstantinides G M (1979) Multiperiod consumption and investment behavior with

convex transactions costs Management Science 25 1127ndash1137Constantinides G M and Richard S F (1978) Existence of optimal simple policies for

discounted-cost inventory and cash management in continuous time Operations Research

26 620ndash636DeAngelo H DeAngelo L and Whited T M (2011) Capital structure dynamics and

transitory debt Journal of Financial Economics 99 235ndash261Dixit A K (1993) The Art of Smooth Pasting Routledge London and New YorkFama E and French K R (2002) Testing trade-off and pecking order predictions about

dividends and debt Review of Financial Studies 15 1ndash33Faulkender M and Petersen M A (2006) Does the source of capital affect capital struc-

ture Review of Financial Studies 19 45ndash79Faulkender M Flannery M J Hankins K W and Smith J M (2012) Cash Flows and

Leverage Adjustments Journal of Financial Economics 103 632ndash646Fischer E Heinkel R and Zechner J (1989) Dynamic capital structure choice theory

and tests Journal of Finance 44 19ndash40Frank M Z and Goyal V K (2008) Trade-off and pecking order theories of

debt in B E Eckbo (ed) Handbook of Corporate Finance Empirical Corporate Finance

Vol 2 Handbook of Finance Series Chapter 12 ElsevierNorth-Holland Amsterdam pp

135ndash202

Frank M Z and Goyal V K (2009) Capital structure decisions which factors are reliably

important Financial Management 38 1ndash37

Garvey G T and Hanka G (1999) Capital structure and corporate control the effect of

antitakeover statutes on firm leverage Journal of Finance 54 519ndash546

Goldstein R Ju N and Leland H (2001) An EBIT-based model of dynamic capital

structure Journal of Business 74 483ndash512

Graham J R and Tucker A L (2006) Tax shelters and corporate debt policy Journal of

Financial Economics 81 563ndash594

Hackbarth D and Mauer D C (2012) Optimal priority structure capital structure and

investment Review of Financial Studies 25 747ndash796

Halling M Yu J and Zechner J (2012) Leverage dynamics over the business cycle

Working paper University of Utah University of New South Wales and Vienna

University of Economics and Business AdministrationHennessy C A and Whited T A (2005) Debt dynamics Journal of Finance 60 1129ndash1165

Huang R and Ritter J R (2009) Testing theories of capital structure and estimating the

speed of adjustment Journal of Financial and Qualitative Analysis 44 237ndash271

1452 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

Hovakimian A Opler T and Titman S (2001) The debt-equity choice Journal of

Financial and Quantitative Analysis 36 1ndash24Jung K Kim Y and Stulz R M (1996) Timing investment opportunities managerial

discretion and the security issue decisions Journal of Financial Economics 42 159ndash185Kane A Marcus A J and McDonald R L (1984) How big is the tax advantage to

debt Journal of Finance 39 841ndash853Kayhan A and Titman S (2007) Firmsrsquo histories and their capital structures Journal of

Financial Economics 83 1ndash32Leary M T (2009) Bank loan supply lender choice and corporate capital structure

Journal of Finance 64 1143ndash1185Leary M T and Roberts M R (2005) Do firms rebalance their capital structures

Journal of Finance 60 2575ndash2619Leland H E (1994) Corporate debt value bond covenants and optimal capital structure

Journal of finance 49 1213ndash1252Lemmon M L Roberts M R and Zender J F (2008) Back to the beginning persist-

ence and the cross-section of corporate capital structure Journal of Finance 63 1575ndash1608Long M S and Malitz I B (1985) The investment-financing nexus some empirical

evidence Midland Corporate Finance Journal 3 53ndash59Mackie-Mason J K (1990) Do taxes affect corporate financing decisions Journal of

Finance 45 1471ndash1493Morellec E Nikolov B and Schurhoff N (2012) Corporate governance and capital

structure dynamics Journal of Finance 67 803ndash848Myers S C (1984) The capital structure puzzle Journal of Finance 39 575ndash592Myers S C (1993) Still searching for optimal capital structure Journal of Applied

Corporate Finance 61 4ndash14Rajan R and Zingales L (1995) What do we know about capital structure some evidence

from international data Journal of Finance 50 1421ndash1460Strebulaev I A (2007) Do tests of capital structure theory mean what they say Journal of

Finance 62 1747ndash1787

Strebulaev I A and Whited T M (2012) Dynamic Models and Structural Estimation in

Corporate Finance Now Publisher Boston and Delft

Titman S and Wessels R (1988) The determinants of capital structure choice Journal of

Finance 43 1ndash21

Welch I (2004) Capital structure and stock returns Journal of Political Economy 112

106ndash131

Welch I (2011) Two common problems in capital structure research the financial-debt-

to-asset ratio and issuing activity versus leverage changes International Review of Finance

11 1ndash17

PROFITSndashLEVERAGE PUZZLEREVISITED 1453

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

We document that more profitable firms really do borrow more (not less)They tend to repurchase equity They experience an increase in both thebook value of equity and the market value of equity Less profitable firmsreally do tend to reduce their debt and to issue equityFirm size and market conditions also matter Larger firms tend to be more

active in the debt markets while smaller firms tend to be relatively moreactive in the equity markets During good times there is more use ofexternal financingThe usual profitsndashleverage puzzle result is primarily driven by the increase

in equity that is experienced by the more profitable firms Accordingly thepuzzle should be restated as asking why do firms not take sufficiently largeoffsetting actions to fully undo the change in equity What limits the mag-nitudes of the typical leverage response to profit shocksIn a frictionless model the partial response appears puzzling But there is

good empirical reason to believe that rebalancing entails both fixed andvariable costs and that firm size matters The rebalancing costs can befully avoided by doing nothing Accordingly the firm must decide whetherany given shock is big enough to be worth responding to If it is then thefirm must decide how big a response is called for These technical conditionsare known as ldquovalue matchingrdquo and ldquosmooth pastingrdquo Thus optimizationimplies that some shocks will be ignored Even if the shock is not ignoredthe optimal response will only partially undo the shock The magnitude ofthe leverage response must balance the marginal cost and the marginalbenefit of an extra unit of leverage Since the marginal cost of adjustingleverage is strictly positive the adjustment toward that static leverageoptimum will only go part way This is true both for leverage increasesand for leverage reductionsOne might conjecture that the variable costs would be time varying and

that such costs might have declined over time Since we do not have a fullaccounting of all of the variable costs it is of course impossible to be sureHowever there is some evidence suggesting that at least the directunderwriting fees are remarkably stable over time10

Surely nobody ought to be surprised that more profitable firms are moreinclined to repurchase equity while unprofitable firms tend not to do so Thefact that more profitable firms typically experience an increase in the value of

10 For IPOs Jay Ritter provides important summary evidence as httpbearwarringtonufleduritterIPOs2013Underwritingpdf He finds that for medium-sized deals a remark-

ably large fraction of the deals have a 7 fee year after year Whether this is due tobehavioral factors or some sort of a coordination device is not clear The extent towhich similar intertemporal stability carries over to other forms of external financemerits further study

1448 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

equity is equally natural Were it not for the contrary claims in the literature(ie the profitsndashleverage puzzle) we would argue that nobody ought to findit surprising that more profitable firms are more inclined to issue debt Thefact that leverage adjustments typically only partially offset profit shocksshould also sound quite natural These facts all fit together easily within thetrade-off theory once we take account of both fixed and variable costs ofactively adjusting capital structure

Appendix

Table AI Market leverage regressions

The table presents estimates of the regressions of market leverage on firm characteristics

The sample comes from the annual Compustat files during the period 1971ndash2009 Financialfirms are excluded The leverage is estimated as the ratio of debt over debt plus marketequity The explanatory variables Profitabilityt1

MB

t1

Tangibilityt1 and LnethAssetsTHORNt1are described in Table 1 IndMedianLevt1 is estimated as the median market leverage of allother firms in the same industry The industry is defined at the level of the four-digit SICcode All specifications include the year fixed effects The specifications in column (2)

additionally include the firm fixed effects We report t-statistics where the standard errorsare clustered at the firm level in parentheses

aSignificant at the 1 level

Quantile regressions

(1) (2) (3) (4) (5)

Profitabilityt1 013a 013a 007a 011a 020a

(275) (262) (349) (425) (538)

IndLevt1 046a 020a 032a 058a 063a

(506) (239) (1299) (1630) (1042)MB

t1

003a 001a 001a 002a 004a

(509) (303) (743) (709) (534)

Tangibilityt1 013a 018a 013a 014a 010a

(191) (181) (762) (571) (232)

LnethAssetsTHORNt1 001a 005a 001a 001a lt001a

(112) (258) (682) (342) (29)

Constant 010a 011a 005a 005a 029a

(171) (98) (173) (118) (398)

Year FE Yes Yes Yes Yes Yes

Firm FE No Yes No No No

Regression type OLS FE 25th tile Median 75thtile

Clustered SE Yes Yes No No No

R2 030 067

Pseudo R2 016 023 0201

N 158578 158578 158578 158578 158578

PROFITSndashLEVERAGE PUZZLEREVISITED 1449

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

Table AII Profitability sorts for firms issuing large amounts

Percentage of firms

Iss Ret Iss Iss Rep Iss Iss IssD IssE

D D NetD E E NetE DampE RepE RetD

(1) (2) (3) (4) (5) (6) (7) (8) (9)

Panel A 10 cutoff to define issuing or retiring activity

Allfirm

s

Low 205 150 134 260 05 257 45 00 23

2 207 175 103 77 06 75 11 01 10

3 243 196 101 50 08 47 08 01 09

4 252 186 113 50 13 48 08 02 08

High 227 161 118 60 27 57 07 04 09

t Highfrac14Low 70 38 62 755 226 768 320 90 152

Smallfirm

s Low 230 132 162 380 05 375 82 00 32

2 228 143 165 334 04 333 65 00 26

3 211 146 147 216 03 214 38 00 23

4 190 154 120 126 04 124 15 01 20

High 194 162 114 99 09 97 12 01 15

t Highfrac14Low 42 40 68 342 20 340 169 07 131

Largefirm

s Low 231 200 82 31 06 29 08 01 04

2 231 187 67 17 06 14 03 01 02

3 234 177 77 17 09 14 03 02 03

4 218 158 75 16 17 14 02 03 02

High 218 135 99 19 41 15 03 07 02

t Highfrac14Low 19 106 36 46 135 57 39 58 23

Percentage of firms

Iss Ret Iss Iss Rep Iss Both Iss D Iss E

D D Net D E E Net E DampE Rep E Ret D

(1) (2) (3) (4) (5) (6) (7) (8) (9)

Panel B 15 cutoff to define issuing or retiring activity

Allfirm

s

Low 158 101 100 230 01 228 27 00 15

2 142 115 63 57 01 56 05 00 06

3 165 128 57 32 01 31 03 00 05

4 173 118 65 35 01 34 03 00 05

High 160 102 70 46 04 44 03 00 05

t Highfrac14Low 08 06 141 738 87 746 277 24 131

Smallfirm

s Low 195 95 132 354 01 350 56 00 24

2 185 96 131 303 01 301 44 00 18

3 160 96 104 179 00 176 18 00 15

4 134 99 78 99 01 98 08 00 12

High 133 95 73 82 01 80 06 00 10

t Highfrac14Low 80 00 95 345 12 344 149 12 52

Largefirm

s Low 145 123 42 14 01 12 03 00 01

2 137 114 32 07 01 06 01 00 01

3 142 104 35 06 01 05 00 00 01

4 130 91 36 08 03 07 01 01 01

High 134 81 48 09 06 07 01 01 01

t Highfrac14Low 19 85 17 26 54 33 34 17 11

1450 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

Table AIV Debt and equity changesmdashwithout fixed effects

The sample contains nonfinancial firms listed on the annual Compustat files for the period from 1971

to 2009 The table presents estimates from regressions of change in debt (D) change in the book

value of equity (BVE) change in the market value of equity (MVE) and net equity issuances

(EqyIss) The explanatory variables include current and lagged change in operating income before

depreciation (Profits and Profitst1) and lagged change in book value of assets (Assetst1) The

regressions include year indicator variables The reported t-statistics are corrected for clustering at the

firm level

aMean significant at the 1 level

bMean significant at the 5 level

cMean significant at the 10 level

D BVE MVE EqyIss

(1) (2) (3) (4) (5) (6) (7) (8)

Profitst1 025a 024a 023a 038a 007 066a 007a 009a

(40) (32) (36) (39) (06) (29) (42) (51)Profits 051b 082a 149a 004a

(25) (33) (49) (33)Assetst1 002 005c 024a 001b

(08) (17) (34) (23)Constant 93 176a 45 71 1552a 1704a 130a 140a

(12) (28) (07) (09) (49) (49) (45) (45)Year FE Yes Yes Yes Yes Yes Yes Yes YesR2 001 005 001 011 001 005 002 002N 162056 162056 162130 162130 157550 157550 162154 162154

Table AIII Median profitability sorts

The sample contains nonfinancial firms listed on the annual Compustat files for the period from 1971

to 2009 The table presents median debt and equity median changes in debt and equity and median

equity issuances (EqyIss) in millions of dollars for all firms sorted on profitability and for profitability

sorts within the smallest and largest firms The sorts are done annually

D D BVE BVE MVE MVE EqyIss Assets(1) (2) (3) (4) (5) (6) (7) (8)

Allfirm

s

Low 21 00 75 17 256 20 01 1922 186 02 449 18 598 19 00 11033 816 06 1274 03 1637 05 00 32244 787 02 1541 18 2488 03 00 3556High 273 00 1361 66 3300 14 00 2578

Smallfirm

s Low 05 00 13 07 140 14 06 402 06 00 25 08 114 12 01 563 09 00 32 05 85 07 00 674 12 00 39 01 68 03 00 79High 10 00 51 04 98 01 00 90

Largefirm

s Low 11485 381 10269 371 13227 202 03 333902 13051 393 12786 140 17571 169 05 392463 11554 287 13429 24 21623 305 00 370314 9963 209 15052 194 29424 308 00 36705High 7361 00 16380 613 49142 439 00 34976

PROFITSndashLEVERAGE PUZZLEREVISITED 1451

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

References

Auerbach A J (1985) Real determinants of corporate leverage in B M Friedman (ed)

Corporate Capital Structures in the United States University of Chicago Press Chicago

IllinoisBertrand M and Mullainathan S (2003) Enjoying the quiet life corporate governance and

managerial preferences Journal of Political Economy 111 1043ndash1075Booth L Varouj A Demirguc-Kunt A and Maksimovic V (2001) Capital structures in

developing countries Journal of Finance 56 87ndash130Brennan M J and Schwartz E S (1984) Optimal financial policy and firm valuation

Journal of Finance 39 593ndash607Cameron A C and Trivedi P K (2010) Microeconometrics using Stata Stata Press TexasConstantinides G M (1979) Multiperiod consumption and investment behavior with

convex transactions costs Management Science 25 1127ndash1137Constantinides G M and Richard S F (1978) Existence of optimal simple policies for

discounted-cost inventory and cash management in continuous time Operations Research

26 620ndash636DeAngelo H DeAngelo L and Whited T M (2011) Capital structure dynamics and

transitory debt Journal of Financial Economics 99 235ndash261Dixit A K (1993) The Art of Smooth Pasting Routledge London and New YorkFama E and French K R (2002) Testing trade-off and pecking order predictions about

dividends and debt Review of Financial Studies 15 1ndash33Faulkender M and Petersen M A (2006) Does the source of capital affect capital struc-

ture Review of Financial Studies 19 45ndash79Faulkender M Flannery M J Hankins K W and Smith J M (2012) Cash Flows and

Leverage Adjustments Journal of Financial Economics 103 632ndash646Fischer E Heinkel R and Zechner J (1989) Dynamic capital structure choice theory

and tests Journal of Finance 44 19ndash40Frank M Z and Goyal V K (2008) Trade-off and pecking order theories of

debt in B E Eckbo (ed) Handbook of Corporate Finance Empirical Corporate Finance

Vol 2 Handbook of Finance Series Chapter 12 ElsevierNorth-Holland Amsterdam pp

135ndash202

Frank M Z and Goyal V K (2009) Capital structure decisions which factors are reliably

important Financial Management 38 1ndash37

Garvey G T and Hanka G (1999) Capital structure and corporate control the effect of

antitakeover statutes on firm leverage Journal of Finance 54 519ndash546

Goldstein R Ju N and Leland H (2001) An EBIT-based model of dynamic capital

structure Journal of Business 74 483ndash512

Graham J R and Tucker A L (2006) Tax shelters and corporate debt policy Journal of

Financial Economics 81 563ndash594

Hackbarth D and Mauer D C (2012) Optimal priority structure capital structure and

investment Review of Financial Studies 25 747ndash796

Halling M Yu J and Zechner J (2012) Leverage dynamics over the business cycle

Working paper University of Utah University of New South Wales and Vienna

University of Economics and Business AdministrationHennessy C A and Whited T A (2005) Debt dynamics Journal of Finance 60 1129ndash1165

Huang R and Ritter J R (2009) Testing theories of capital structure and estimating the

speed of adjustment Journal of Financial and Qualitative Analysis 44 237ndash271

1452 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

Hovakimian A Opler T and Titman S (2001) The debt-equity choice Journal of

Financial and Quantitative Analysis 36 1ndash24Jung K Kim Y and Stulz R M (1996) Timing investment opportunities managerial

discretion and the security issue decisions Journal of Financial Economics 42 159ndash185Kane A Marcus A J and McDonald R L (1984) How big is the tax advantage to

debt Journal of Finance 39 841ndash853Kayhan A and Titman S (2007) Firmsrsquo histories and their capital structures Journal of

Financial Economics 83 1ndash32Leary M T (2009) Bank loan supply lender choice and corporate capital structure

Journal of Finance 64 1143ndash1185Leary M T and Roberts M R (2005) Do firms rebalance their capital structures

Journal of Finance 60 2575ndash2619Leland H E (1994) Corporate debt value bond covenants and optimal capital structure

Journal of finance 49 1213ndash1252Lemmon M L Roberts M R and Zender J F (2008) Back to the beginning persist-

ence and the cross-section of corporate capital structure Journal of Finance 63 1575ndash1608Long M S and Malitz I B (1985) The investment-financing nexus some empirical

evidence Midland Corporate Finance Journal 3 53ndash59Mackie-Mason J K (1990) Do taxes affect corporate financing decisions Journal of

Finance 45 1471ndash1493Morellec E Nikolov B and Schurhoff N (2012) Corporate governance and capital

structure dynamics Journal of Finance 67 803ndash848Myers S C (1984) The capital structure puzzle Journal of Finance 39 575ndash592Myers S C (1993) Still searching for optimal capital structure Journal of Applied

Corporate Finance 61 4ndash14Rajan R and Zingales L (1995) What do we know about capital structure some evidence

from international data Journal of Finance 50 1421ndash1460Strebulaev I A (2007) Do tests of capital structure theory mean what they say Journal of

Finance 62 1747ndash1787

Strebulaev I A and Whited T M (2012) Dynamic Models and Structural Estimation in

Corporate Finance Now Publisher Boston and Delft

Titman S and Wessels R (1988) The determinants of capital structure choice Journal of

Finance 43 1ndash21

Welch I (2004) Capital structure and stock returns Journal of Political Economy 112

106ndash131

Welch I (2011) Two common problems in capital structure research the financial-debt-

to-asset ratio and issuing activity versus leverage changes International Review of Finance

11 1ndash17

PROFITSndashLEVERAGE PUZZLEREVISITED 1453

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

equity is equally natural Were it not for the contrary claims in the literature(ie the profitsndashleverage puzzle) we would argue that nobody ought to findit surprising that more profitable firms are more inclined to issue debt Thefact that leverage adjustments typically only partially offset profit shocksshould also sound quite natural These facts all fit together easily within thetrade-off theory once we take account of both fixed and variable costs ofactively adjusting capital structure

Appendix

Table AI Market leverage regressions

The table presents estimates of the regressions of market leverage on firm characteristics

The sample comes from the annual Compustat files during the period 1971ndash2009 Financialfirms are excluded The leverage is estimated as the ratio of debt over debt plus marketequity The explanatory variables Profitabilityt1

MB

t1

Tangibilityt1 and LnethAssetsTHORNt1are described in Table 1 IndMedianLevt1 is estimated as the median market leverage of allother firms in the same industry The industry is defined at the level of the four-digit SICcode All specifications include the year fixed effects The specifications in column (2)

additionally include the firm fixed effects We report t-statistics where the standard errorsare clustered at the firm level in parentheses

aSignificant at the 1 level

Quantile regressions

(1) (2) (3) (4) (5)

Profitabilityt1 013a 013a 007a 011a 020a

(275) (262) (349) (425) (538)

IndLevt1 046a 020a 032a 058a 063a

(506) (239) (1299) (1630) (1042)MB

t1

003a 001a 001a 002a 004a

(509) (303) (743) (709) (534)

Tangibilityt1 013a 018a 013a 014a 010a

(191) (181) (762) (571) (232)

LnethAssetsTHORNt1 001a 005a 001a 001a lt001a

(112) (258) (682) (342) (29)

Constant 010a 011a 005a 005a 029a

(171) (98) (173) (118) (398)

Year FE Yes Yes Yes Yes Yes

Firm FE No Yes No No No

Regression type OLS FE 25th tile Median 75thtile

Clustered SE Yes Yes No No No

R2 030 067

Pseudo R2 016 023 0201

N 158578 158578 158578 158578 158578

PROFITSndashLEVERAGE PUZZLEREVISITED 1449

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

Table AII Profitability sorts for firms issuing large amounts

Percentage of firms

Iss Ret Iss Iss Rep Iss Iss IssD IssE

D D NetD E E NetE DampE RepE RetD

(1) (2) (3) (4) (5) (6) (7) (8) (9)

Panel A 10 cutoff to define issuing or retiring activity

Allfirm

s

Low 205 150 134 260 05 257 45 00 23

2 207 175 103 77 06 75 11 01 10

3 243 196 101 50 08 47 08 01 09

4 252 186 113 50 13 48 08 02 08

High 227 161 118 60 27 57 07 04 09

t Highfrac14Low 70 38 62 755 226 768 320 90 152

Smallfirm

s Low 230 132 162 380 05 375 82 00 32

2 228 143 165 334 04 333 65 00 26

3 211 146 147 216 03 214 38 00 23

4 190 154 120 126 04 124 15 01 20

High 194 162 114 99 09 97 12 01 15

t Highfrac14Low 42 40 68 342 20 340 169 07 131

Largefirm

s Low 231 200 82 31 06 29 08 01 04

2 231 187 67 17 06 14 03 01 02

3 234 177 77 17 09 14 03 02 03

4 218 158 75 16 17 14 02 03 02

High 218 135 99 19 41 15 03 07 02

t Highfrac14Low 19 106 36 46 135 57 39 58 23

Percentage of firms

Iss Ret Iss Iss Rep Iss Both Iss D Iss E

D D Net D E E Net E DampE Rep E Ret D

(1) (2) (3) (4) (5) (6) (7) (8) (9)

Panel B 15 cutoff to define issuing or retiring activity

Allfirm

s

Low 158 101 100 230 01 228 27 00 15

2 142 115 63 57 01 56 05 00 06

3 165 128 57 32 01 31 03 00 05

4 173 118 65 35 01 34 03 00 05

High 160 102 70 46 04 44 03 00 05

t Highfrac14Low 08 06 141 738 87 746 277 24 131

Smallfirm

s Low 195 95 132 354 01 350 56 00 24

2 185 96 131 303 01 301 44 00 18

3 160 96 104 179 00 176 18 00 15

4 134 99 78 99 01 98 08 00 12

High 133 95 73 82 01 80 06 00 10

t Highfrac14Low 80 00 95 345 12 344 149 12 52

Largefirm

s Low 145 123 42 14 01 12 03 00 01

2 137 114 32 07 01 06 01 00 01

3 142 104 35 06 01 05 00 00 01

4 130 91 36 08 03 07 01 01 01

High 134 81 48 09 06 07 01 01 01

t Highfrac14Low 19 85 17 26 54 33 34 17 11

1450 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

Table AIV Debt and equity changesmdashwithout fixed effects

The sample contains nonfinancial firms listed on the annual Compustat files for the period from 1971

to 2009 The table presents estimates from regressions of change in debt (D) change in the book

value of equity (BVE) change in the market value of equity (MVE) and net equity issuances

(EqyIss) The explanatory variables include current and lagged change in operating income before

depreciation (Profits and Profitst1) and lagged change in book value of assets (Assetst1) The

regressions include year indicator variables The reported t-statistics are corrected for clustering at the

firm level

aMean significant at the 1 level

bMean significant at the 5 level

cMean significant at the 10 level

D BVE MVE EqyIss

(1) (2) (3) (4) (5) (6) (7) (8)

Profitst1 025a 024a 023a 038a 007 066a 007a 009a

(40) (32) (36) (39) (06) (29) (42) (51)Profits 051b 082a 149a 004a

(25) (33) (49) (33)Assetst1 002 005c 024a 001b

(08) (17) (34) (23)Constant 93 176a 45 71 1552a 1704a 130a 140a

(12) (28) (07) (09) (49) (49) (45) (45)Year FE Yes Yes Yes Yes Yes Yes Yes YesR2 001 005 001 011 001 005 002 002N 162056 162056 162130 162130 157550 157550 162154 162154

Table AIII Median profitability sorts

The sample contains nonfinancial firms listed on the annual Compustat files for the period from 1971

to 2009 The table presents median debt and equity median changes in debt and equity and median

equity issuances (EqyIss) in millions of dollars for all firms sorted on profitability and for profitability

sorts within the smallest and largest firms The sorts are done annually

D D BVE BVE MVE MVE EqyIss Assets(1) (2) (3) (4) (5) (6) (7) (8)

Allfirm

s

Low 21 00 75 17 256 20 01 1922 186 02 449 18 598 19 00 11033 816 06 1274 03 1637 05 00 32244 787 02 1541 18 2488 03 00 3556High 273 00 1361 66 3300 14 00 2578

Smallfirm

s Low 05 00 13 07 140 14 06 402 06 00 25 08 114 12 01 563 09 00 32 05 85 07 00 674 12 00 39 01 68 03 00 79High 10 00 51 04 98 01 00 90

Largefirm

s Low 11485 381 10269 371 13227 202 03 333902 13051 393 12786 140 17571 169 05 392463 11554 287 13429 24 21623 305 00 370314 9963 209 15052 194 29424 308 00 36705High 7361 00 16380 613 49142 439 00 34976

PROFITSndashLEVERAGE PUZZLEREVISITED 1451

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

References

Auerbach A J (1985) Real determinants of corporate leverage in B M Friedman (ed)

Corporate Capital Structures in the United States University of Chicago Press Chicago

IllinoisBertrand M and Mullainathan S (2003) Enjoying the quiet life corporate governance and

managerial preferences Journal of Political Economy 111 1043ndash1075Booth L Varouj A Demirguc-Kunt A and Maksimovic V (2001) Capital structures in

developing countries Journal of Finance 56 87ndash130Brennan M J and Schwartz E S (1984) Optimal financial policy and firm valuation

Journal of Finance 39 593ndash607Cameron A C and Trivedi P K (2010) Microeconometrics using Stata Stata Press TexasConstantinides G M (1979) Multiperiod consumption and investment behavior with

convex transactions costs Management Science 25 1127ndash1137Constantinides G M and Richard S F (1978) Existence of optimal simple policies for

discounted-cost inventory and cash management in continuous time Operations Research

26 620ndash636DeAngelo H DeAngelo L and Whited T M (2011) Capital structure dynamics and

transitory debt Journal of Financial Economics 99 235ndash261Dixit A K (1993) The Art of Smooth Pasting Routledge London and New YorkFama E and French K R (2002) Testing trade-off and pecking order predictions about

dividends and debt Review of Financial Studies 15 1ndash33Faulkender M and Petersen M A (2006) Does the source of capital affect capital struc-

ture Review of Financial Studies 19 45ndash79Faulkender M Flannery M J Hankins K W and Smith J M (2012) Cash Flows and

Leverage Adjustments Journal of Financial Economics 103 632ndash646Fischer E Heinkel R and Zechner J (1989) Dynamic capital structure choice theory

and tests Journal of Finance 44 19ndash40Frank M Z and Goyal V K (2008) Trade-off and pecking order theories of

debt in B E Eckbo (ed) Handbook of Corporate Finance Empirical Corporate Finance

Vol 2 Handbook of Finance Series Chapter 12 ElsevierNorth-Holland Amsterdam pp

135ndash202

Frank M Z and Goyal V K (2009) Capital structure decisions which factors are reliably

important Financial Management 38 1ndash37

Garvey G T and Hanka G (1999) Capital structure and corporate control the effect of

antitakeover statutes on firm leverage Journal of Finance 54 519ndash546

Goldstein R Ju N and Leland H (2001) An EBIT-based model of dynamic capital

structure Journal of Business 74 483ndash512

Graham J R and Tucker A L (2006) Tax shelters and corporate debt policy Journal of

Financial Economics 81 563ndash594

Hackbarth D and Mauer D C (2012) Optimal priority structure capital structure and

investment Review of Financial Studies 25 747ndash796

Halling M Yu J and Zechner J (2012) Leverage dynamics over the business cycle

Working paper University of Utah University of New South Wales and Vienna

University of Economics and Business AdministrationHennessy C A and Whited T A (2005) Debt dynamics Journal of Finance 60 1129ndash1165

Huang R and Ritter J R (2009) Testing theories of capital structure and estimating the

speed of adjustment Journal of Financial and Qualitative Analysis 44 237ndash271

1452 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

Hovakimian A Opler T and Titman S (2001) The debt-equity choice Journal of

Financial and Quantitative Analysis 36 1ndash24Jung K Kim Y and Stulz R M (1996) Timing investment opportunities managerial

discretion and the security issue decisions Journal of Financial Economics 42 159ndash185Kane A Marcus A J and McDonald R L (1984) How big is the tax advantage to

debt Journal of Finance 39 841ndash853Kayhan A and Titman S (2007) Firmsrsquo histories and their capital structures Journal of

Financial Economics 83 1ndash32Leary M T (2009) Bank loan supply lender choice and corporate capital structure

Journal of Finance 64 1143ndash1185Leary M T and Roberts M R (2005) Do firms rebalance their capital structures

Journal of Finance 60 2575ndash2619Leland H E (1994) Corporate debt value bond covenants and optimal capital structure

Journal of finance 49 1213ndash1252Lemmon M L Roberts M R and Zender J F (2008) Back to the beginning persist-

ence and the cross-section of corporate capital structure Journal of Finance 63 1575ndash1608Long M S and Malitz I B (1985) The investment-financing nexus some empirical

evidence Midland Corporate Finance Journal 3 53ndash59Mackie-Mason J K (1990) Do taxes affect corporate financing decisions Journal of

Finance 45 1471ndash1493Morellec E Nikolov B and Schurhoff N (2012) Corporate governance and capital

structure dynamics Journal of Finance 67 803ndash848Myers S C (1984) The capital structure puzzle Journal of Finance 39 575ndash592Myers S C (1993) Still searching for optimal capital structure Journal of Applied

Corporate Finance 61 4ndash14Rajan R and Zingales L (1995) What do we know about capital structure some evidence

from international data Journal of Finance 50 1421ndash1460Strebulaev I A (2007) Do tests of capital structure theory mean what they say Journal of

Finance 62 1747ndash1787

Strebulaev I A and Whited T M (2012) Dynamic Models and Structural Estimation in

Corporate Finance Now Publisher Boston and Delft

Titman S and Wessels R (1988) The determinants of capital structure choice Journal of

Finance 43 1ndash21

Welch I (2004) Capital structure and stock returns Journal of Political Economy 112

106ndash131

Welch I (2011) Two common problems in capital structure research the financial-debt-

to-asset ratio and issuing activity versus leverage changes International Review of Finance

11 1ndash17

PROFITSndashLEVERAGE PUZZLEREVISITED 1453

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

Table AII Profitability sorts for firms issuing large amounts

Percentage of firms

Iss Ret Iss Iss Rep Iss Iss IssD IssE

D D NetD E E NetE DampE RepE RetD

(1) (2) (3) (4) (5) (6) (7) (8) (9)

Panel A 10 cutoff to define issuing or retiring activity

Allfirm

s

Low 205 150 134 260 05 257 45 00 23

2 207 175 103 77 06 75 11 01 10

3 243 196 101 50 08 47 08 01 09

4 252 186 113 50 13 48 08 02 08

High 227 161 118 60 27 57 07 04 09

t Highfrac14Low 70 38 62 755 226 768 320 90 152

Smallfirm

s Low 230 132 162 380 05 375 82 00 32

2 228 143 165 334 04 333 65 00 26

3 211 146 147 216 03 214 38 00 23

4 190 154 120 126 04 124 15 01 20

High 194 162 114 99 09 97 12 01 15

t Highfrac14Low 42 40 68 342 20 340 169 07 131

Largefirm

s Low 231 200 82 31 06 29 08 01 04

2 231 187 67 17 06 14 03 01 02

3 234 177 77 17 09 14 03 02 03

4 218 158 75 16 17 14 02 03 02

High 218 135 99 19 41 15 03 07 02

t Highfrac14Low 19 106 36 46 135 57 39 58 23

Percentage of firms

Iss Ret Iss Iss Rep Iss Both Iss D Iss E

D D Net D E E Net E DampE Rep E Ret D

(1) (2) (3) (4) (5) (6) (7) (8) (9)

Panel B 15 cutoff to define issuing or retiring activity

Allfirm

s

Low 158 101 100 230 01 228 27 00 15

2 142 115 63 57 01 56 05 00 06

3 165 128 57 32 01 31 03 00 05

4 173 118 65 35 01 34 03 00 05

High 160 102 70 46 04 44 03 00 05

t Highfrac14Low 08 06 141 738 87 746 277 24 131

Smallfirm

s Low 195 95 132 354 01 350 56 00 24

2 185 96 131 303 01 301 44 00 18

3 160 96 104 179 00 176 18 00 15

4 134 99 78 99 01 98 08 00 12

High 133 95 73 82 01 80 06 00 10

t Highfrac14Low 80 00 95 345 12 344 149 12 52

Largefirm

s Low 145 123 42 14 01 12 03 00 01

2 137 114 32 07 01 06 01 00 01

3 142 104 35 06 01 05 00 00 01

4 130 91 36 08 03 07 01 01 01

High 134 81 48 09 06 07 01 01 01

t Highfrac14Low 19 85 17 26 54 33 34 17 11

1450 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

Table AIV Debt and equity changesmdashwithout fixed effects

The sample contains nonfinancial firms listed on the annual Compustat files for the period from 1971

to 2009 The table presents estimates from regressions of change in debt (D) change in the book

value of equity (BVE) change in the market value of equity (MVE) and net equity issuances

(EqyIss) The explanatory variables include current and lagged change in operating income before

depreciation (Profits and Profitst1) and lagged change in book value of assets (Assetst1) The

regressions include year indicator variables The reported t-statistics are corrected for clustering at the

firm level

aMean significant at the 1 level

bMean significant at the 5 level

cMean significant at the 10 level

D BVE MVE EqyIss

(1) (2) (3) (4) (5) (6) (7) (8)

Profitst1 025a 024a 023a 038a 007 066a 007a 009a

(40) (32) (36) (39) (06) (29) (42) (51)Profits 051b 082a 149a 004a

(25) (33) (49) (33)Assetst1 002 005c 024a 001b

(08) (17) (34) (23)Constant 93 176a 45 71 1552a 1704a 130a 140a

(12) (28) (07) (09) (49) (49) (45) (45)Year FE Yes Yes Yes Yes Yes Yes Yes YesR2 001 005 001 011 001 005 002 002N 162056 162056 162130 162130 157550 157550 162154 162154

Table AIII Median profitability sorts

The sample contains nonfinancial firms listed on the annual Compustat files for the period from 1971

to 2009 The table presents median debt and equity median changes in debt and equity and median

equity issuances (EqyIss) in millions of dollars for all firms sorted on profitability and for profitability

sorts within the smallest and largest firms The sorts are done annually

D D BVE BVE MVE MVE EqyIss Assets(1) (2) (3) (4) (5) (6) (7) (8)

Allfirm

s

Low 21 00 75 17 256 20 01 1922 186 02 449 18 598 19 00 11033 816 06 1274 03 1637 05 00 32244 787 02 1541 18 2488 03 00 3556High 273 00 1361 66 3300 14 00 2578

Smallfirm

s Low 05 00 13 07 140 14 06 402 06 00 25 08 114 12 01 563 09 00 32 05 85 07 00 674 12 00 39 01 68 03 00 79High 10 00 51 04 98 01 00 90

Largefirm

s Low 11485 381 10269 371 13227 202 03 333902 13051 393 12786 140 17571 169 05 392463 11554 287 13429 24 21623 305 00 370314 9963 209 15052 194 29424 308 00 36705High 7361 00 16380 613 49142 439 00 34976

PROFITSndashLEVERAGE PUZZLEREVISITED 1451

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

References

Auerbach A J (1985) Real determinants of corporate leverage in B M Friedman (ed)

Corporate Capital Structures in the United States University of Chicago Press Chicago

IllinoisBertrand M and Mullainathan S (2003) Enjoying the quiet life corporate governance and

managerial preferences Journal of Political Economy 111 1043ndash1075Booth L Varouj A Demirguc-Kunt A and Maksimovic V (2001) Capital structures in

developing countries Journal of Finance 56 87ndash130Brennan M J and Schwartz E S (1984) Optimal financial policy and firm valuation

Journal of Finance 39 593ndash607Cameron A C and Trivedi P K (2010) Microeconometrics using Stata Stata Press TexasConstantinides G M (1979) Multiperiod consumption and investment behavior with

convex transactions costs Management Science 25 1127ndash1137Constantinides G M and Richard S F (1978) Existence of optimal simple policies for

discounted-cost inventory and cash management in continuous time Operations Research

26 620ndash636DeAngelo H DeAngelo L and Whited T M (2011) Capital structure dynamics and

transitory debt Journal of Financial Economics 99 235ndash261Dixit A K (1993) The Art of Smooth Pasting Routledge London and New YorkFama E and French K R (2002) Testing trade-off and pecking order predictions about

dividends and debt Review of Financial Studies 15 1ndash33Faulkender M and Petersen M A (2006) Does the source of capital affect capital struc-

ture Review of Financial Studies 19 45ndash79Faulkender M Flannery M J Hankins K W and Smith J M (2012) Cash Flows and

Leverage Adjustments Journal of Financial Economics 103 632ndash646Fischer E Heinkel R and Zechner J (1989) Dynamic capital structure choice theory

and tests Journal of Finance 44 19ndash40Frank M Z and Goyal V K (2008) Trade-off and pecking order theories of

debt in B E Eckbo (ed) Handbook of Corporate Finance Empirical Corporate Finance

Vol 2 Handbook of Finance Series Chapter 12 ElsevierNorth-Holland Amsterdam pp

135ndash202

Frank M Z and Goyal V K (2009) Capital structure decisions which factors are reliably

important Financial Management 38 1ndash37

Garvey G T and Hanka G (1999) Capital structure and corporate control the effect of

antitakeover statutes on firm leverage Journal of Finance 54 519ndash546

Goldstein R Ju N and Leland H (2001) An EBIT-based model of dynamic capital

structure Journal of Business 74 483ndash512

Graham J R and Tucker A L (2006) Tax shelters and corporate debt policy Journal of

Financial Economics 81 563ndash594

Hackbarth D and Mauer D C (2012) Optimal priority structure capital structure and

investment Review of Financial Studies 25 747ndash796

Halling M Yu J and Zechner J (2012) Leverage dynamics over the business cycle

Working paper University of Utah University of New South Wales and Vienna

University of Economics and Business AdministrationHennessy C A and Whited T A (2005) Debt dynamics Journal of Finance 60 1129ndash1165

Huang R and Ritter J R (2009) Testing theories of capital structure and estimating the

speed of adjustment Journal of Financial and Qualitative Analysis 44 237ndash271

1452 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

Hovakimian A Opler T and Titman S (2001) The debt-equity choice Journal of

Financial and Quantitative Analysis 36 1ndash24Jung K Kim Y and Stulz R M (1996) Timing investment opportunities managerial

discretion and the security issue decisions Journal of Financial Economics 42 159ndash185Kane A Marcus A J and McDonald R L (1984) How big is the tax advantage to

debt Journal of Finance 39 841ndash853Kayhan A and Titman S (2007) Firmsrsquo histories and their capital structures Journal of

Financial Economics 83 1ndash32Leary M T (2009) Bank loan supply lender choice and corporate capital structure

Journal of Finance 64 1143ndash1185Leary M T and Roberts M R (2005) Do firms rebalance their capital structures

Journal of Finance 60 2575ndash2619Leland H E (1994) Corporate debt value bond covenants and optimal capital structure

Journal of finance 49 1213ndash1252Lemmon M L Roberts M R and Zender J F (2008) Back to the beginning persist-

ence and the cross-section of corporate capital structure Journal of Finance 63 1575ndash1608Long M S and Malitz I B (1985) The investment-financing nexus some empirical

evidence Midland Corporate Finance Journal 3 53ndash59Mackie-Mason J K (1990) Do taxes affect corporate financing decisions Journal of

Finance 45 1471ndash1493Morellec E Nikolov B and Schurhoff N (2012) Corporate governance and capital

structure dynamics Journal of Finance 67 803ndash848Myers S C (1984) The capital structure puzzle Journal of Finance 39 575ndash592Myers S C (1993) Still searching for optimal capital structure Journal of Applied

Corporate Finance 61 4ndash14Rajan R and Zingales L (1995) What do we know about capital structure some evidence

from international data Journal of Finance 50 1421ndash1460Strebulaev I A (2007) Do tests of capital structure theory mean what they say Journal of

Finance 62 1747ndash1787

Strebulaev I A and Whited T M (2012) Dynamic Models and Structural Estimation in

Corporate Finance Now Publisher Boston and Delft

Titman S and Wessels R (1988) The determinants of capital structure choice Journal of

Finance 43 1ndash21

Welch I (2004) Capital structure and stock returns Journal of Political Economy 112

106ndash131

Welch I (2011) Two common problems in capital structure research the financial-debt-

to-asset ratio and issuing activity versus leverage changes International Review of Finance

11 1ndash17

PROFITSndashLEVERAGE PUZZLEREVISITED 1453

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

Table AIV Debt and equity changesmdashwithout fixed effects

The sample contains nonfinancial firms listed on the annual Compustat files for the period from 1971

to 2009 The table presents estimates from regressions of change in debt (D) change in the book

value of equity (BVE) change in the market value of equity (MVE) and net equity issuances

(EqyIss) The explanatory variables include current and lagged change in operating income before

depreciation (Profits and Profitst1) and lagged change in book value of assets (Assetst1) The

regressions include year indicator variables The reported t-statistics are corrected for clustering at the

firm level

aMean significant at the 1 level

bMean significant at the 5 level

cMean significant at the 10 level

D BVE MVE EqyIss

(1) (2) (3) (4) (5) (6) (7) (8)

Profitst1 025a 024a 023a 038a 007 066a 007a 009a

(40) (32) (36) (39) (06) (29) (42) (51)Profits 051b 082a 149a 004a

(25) (33) (49) (33)Assetst1 002 005c 024a 001b

(08) (17) (34) (23)Constant 93 176a 45 71 1552a 1704a 130a 140a

(12) (28) (07) (09) (49) (49) (45) (45)Year FE Yes Yes Yes Yes Yes Yes Yes YesR2 001 005 001 011 001 005 002 002N 162056 162056 162130 162130 157550 157550 162154 162154

Table AIII Median profitability sorts

The sample contains nonfinancial firms listed on the annual Compustat files for the period from 1971

to 2009 The table presents median debt and equity median changes in debt and equity and median

equity issuances (EqyIss) in millions of dollars for all firms sorted on profitability and for profitability

sorts within the smallest and largest firms The sorts are done annually

D D BVE BVE MVE MVE EqyIss Assets(1) (2) (3) (4) (5) (6) (7) (8)

Allfirm

s

Low 21 00 75 17 256 20 01 1922 186 02 449 18 598 19 00 11033 816 06 1274 03 1637 05 00 32244 787 02 1541 18 2488 03 00 3556High 273 00 1361 66 3300 14 00 2578

Smallfirm

s Low 05 00 13 07 140 14 06 402 06 00 25 08 114 12 01 563 09 00 32 05 85 07 00 674 12 00 39 01 68 03 00 79High 10 00 51 04 98 01 00 90

Largefirm

s Low 11485 381 10269 371 13227 202 03 333902 13051 393 12786 140 17571 169 05 392463 11554 287 13429 24 21623 305 00 370314 9963 209 15052 194 29424 308 00 36705High 7361 00 16380 613 49142 439 00 34976

PROFITSndashLEVERAGE PUZZLEREVISITED 1451

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

References

Auerbach A J (1985) Real determinants of corporate leverage in B M Friedman (ed)

Corporate Capital Structures in the United States University of Chicago Press Chicago

IllinoisBertrand M and Mullainathan S (2003) Enjoying the quiet life corporate governance and

managerial preferences Journal of Political Economy 111 1043ndash1075Booth L Varouj A Demirguc-Kunt A and Maksimovic V (2001) Capital structures in

developing countries Journal of Finance 56 87ndash130Brennan M J and Schwartz E S (1984) Optimal financial policy and firm valuation

Journal of Finance 39 593ndash607Cameron A C and Trivedi P K (2010) Microeconometrics using Stata Stata Press TexasConstantinides G M (1979) Multiperiod consumption and investment behavior with

convex transactions costs Management Science 25 1127ndash1137Constantinides G M and Richard S F (1978) Existence of optimal simple policies for

discounted-cost inventory and cash management in continuous time Operations Research

26 620ndash636DeAngelo H DeAngelo L and Whited T M (2011) Capital structure dynamics and

transitory debt Journal of Financial Economics 99 235ndash261Dixit A K (1993) The Art of Smooth Pasting Routledge London and New YorkFama E and French K R (2002) Testing trade-off and pecking order predictions about

dividends and debt Review of Financial Studies 15 1ndash33Faulkender M and Petersen M A (2006) Does the source of capital affect capital struc-

ture Review of Financial Studies 19 45ndash79Faulkender M Flannery M J Hankins K W and Smith J M (2012) Cash Flows and

Leverage Adjustments Journal of Financial Economics 103 632ndash646Fischer E Heinkel R and Zechner J (1989) Dynamic capital structure choice theory

and tests Journal of Finance 44 19ndash40Frank M Z and Goyal V K (2008) Trade-off and pecking order theories of

debt in B E Eckbo (ed) Handbook of Corporate Finance Empirical Corporate Finance

Vol 2 Handbook of Finance Series Chapter 12 ElsevierNorth-Holland Amsterdam pp

135ndash202

Frank M Z and Goyal V K (2009) Capital structure decisions which factors are reliably

important Financial Management 38 1ndash37

Garvey G T and Hanka G (1999) Capital structure and corporate control the effect of

antitakeover statutes on firm leverage Journal of Finance 54 519ndash546

Goldstein R Ju N and Leland H (2001) An EBIT-based model of dynamic capital

structure Journal of Business 74 483ndash512

Graham J R and Tucker A L (2006) Tax shelters and corporate debt policy Journal of

Financial Economics 81 563ndash594

Hackbarth D and Mauer D C (2012) Optimal priority structure capital structure and

investment Review of Financial Studies 25 747ndash796

Halling M Yu J and Zechner J (2012) Leverage dynamics over the business cycle

Working paper University of Utah University of New South Wales and Vienna

University of Economics and Business AdministrationHennessy C A and Whited T A (2005) Debt dynamics Journal of Finance 60 1129ndash1165

Huang R and Ritter J R (2009) Testing theories of capital structure and estimating the

speed of adjustment Journal of Financial and Qualitative Analysis 44 237ndash271

1452 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

Hovakimian A Opler T and Titman S (2001) The debt-equity choice Journal of

Financial and Quantitative Analysis 36 1ndash24Jung K Kim Y and Stulz R M (1996) Timing investment opportunities managerial

discretion and the security issue decisions Journal of Financial Economics 42 159ndash185Kane A Marcus A J and McDonald R L (1984) How big is the tax advantage to

debt Journal of Finance 39 841ndash853Kayhan A and Titman S (2007) Firmsrsquo histories and their capital structures Journal of

Financial Economics 83 1ndash32Leary M T (2009) Bank loan supply lender choice and corporate capital structure

Journal of Finance 64 1143ndash1185Leary M T and Roberts M R (2005) Do firms rebalance their capital structures

Journal of Finance 60 2575ndash2619Leland H E (1994) Corporate debt value bond covenants and optimal capital structure

Journal of finance 49 1213ndash1252Lemmon M L Roberts M R and Zender J F (2008) Back to the beginning persist-

ence and the cross-section of corporate capital structure Journal of Finance 63 1575ndash1608Long M S and Malitz I B (1985) The investment-financing nexus some empirical

evidence Midland Corporate Finance Journal 3 53ndash59Mackie-Mason J K (1990) Do taxes affect corporate financing decisions Journal of

Finance 45 1471ndash1493Morellec E Nikolov B and Schurhoff N (2012) Corporate governance and capital

structure dynamics Journal of Finance 67 803ndash848Myers S C (1984) The capital structure puzzle Journal of Finance 39 575ndash592Myers S C (1993) Still searching for optimal capital structure Journal of Applied

Corporate Finance 61 4ndash14Rajan R and Zingales L (1995) What do we know about capital structure some evidence

from international data Journal of Finance 50 1421ndash1460Strebulaev I A (2007) Do tests of capital structure theory mean what they say Journal of

Finance 62 1747ndash1787

Strebulaev I A and Whited T M (2012) Dynamic Models and Structural Estimation in

Corporate Finance Now Publisher Boston and Delft

Titman S and Wessels R (1988) The determinants of capital structure choice Journal of

Finance 43 1ndash21

Welch I (2004) Capital structure and stock returns Journal of Political Economy 112

106ndash131

Welch I (2011) Two common problems in capital structure research the financial-debt-

to-asset ratio and issuing activity versus leverage changes International Review of Finance

11 1ndash17

PROFITSndashLEVERAGE PUZZLEREVISITED 1453

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

References

Auerbach A J (1985) Real determinants of corporate leverage in B M Friedman (ed)

Corporate Capital Structures in the United States University of Chicago Press Chicago

IllinoisBertrand M and Mullainathan S (2003) Enjoying the quiet life corporate governance and

managerial preferences Journal of Political Economy 111 1043ndash1075Booth L Varouj A Demirguc-Kunt A and Maksimovic V (2001) Capital structures in

developing countries Journal of Finance 56 87ndash130Brennan M J and Schwartz E S (1984) Optimal financial policy and firm valuation

Journal of Finance 39 593ndash607Cameron A C and Trivedi P K (2010) Microeconometrics using Stata Stata Press TexasConstantinides G M (1979) Multiperiod consumption and investment behavior with

convex transactions costs Management Science 25 1127ndash1137Constantinides G M and Richard S F (1978) Existence of optimal simple policies for

discounted-cost inventory and cash management in continuous time Operations Research

26 620ndash636DeAngelo H DeAngelo L and Whited T M (2011) Capital structure dynamics and

transitory debt Journal of Financial Economics 99 235ndash261Dixit A K (1993) The Art of Smooth Pasting Routledge London and New YorkFama E and French K R (2002) Testing trade-off and pecking order predictions about

dividends and debt Review of Financial Studies 15 1ndash33Faulkender M and Petersen M A (2006) Does the source of capital affect capital struc-

ture Review of Financial Studies 19 45ndash79Faulkender M Flannery M J Hankins K W and Smith J M (2012) Cash Flows and

Leverage Adjustments Journal of Financial Economics 103 632ndash646Fischer E Heinkel R and Zechner J (1989) Dynamic capital structure choice theory

and tests Journal of Finance 44 19ndash40Frank M Z and Goyal V K (2008) Trade-off and pecking order theories of

debt in B E Eckbo (ed) Handbook of Corporate Finance Empirical Corporate Finance

Vol 2 Handbook of Finance Series Chapter 12 ElsevierNorth-Holland Amsterdam pp

135ndash202

Frank M Z and Goyal V K (2009) Capital structure decisions which factors are reliably

important Financial Management 38 1ndash37

Garvey G T and Hanka G (1999) Capital structure and corporate control the effect of

antitakeover statutes on firm leverage Journal of Finance 54 519ndash546

Goldstein R Ju N and Leland H (2001) An EBIT-based model of dynamic capital

structure Journal of Business 74 483ndash512

Graham J R and Tucker A L (2006) Tax shelters and corporate debt policy Journal of

Financial Economics 81 563ndash594

Hackbarth D and Mauer D C (2012) Optimal priority structure capital structure and

investment Review of Financial Studies 25 747ndash796

Halling M Yu J and Zechner J (2012) Leverage dynamics over the business cycle

Working paper University of Utah University of New South Wales and Vienna

University of Economics and Business AdministrationHennessy C A and Whited T A (2005) Debt dynamics Journal of Finance 60 1129ndash1165

Huang R and Ritter J R (2009) Testing theories of capital structure and estimating the

speed of adjustment Journal of Financial and Qualitative Analysis 44 237ndash271

1452 M Z FRANKANDVK GOYAL

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

Hovakimian A Opler T and Titman S (2001) The debt-equity choice Journal of

Financial and Quantitative Analysis 36 1ndash24Jung K Kim Y and Stulz R M (1996) Timing investment opportunities managerial

discretion and the security issue decisions Journal of Financial Economics 42 159ndash185Kane A Marcus A J and McDonald R L (1984) How big is the tax advantage to

debt Journal of Finance 39 841ndash853Kayhan A and Titman S (2007) Firmsrsquo histories and their capital structures Journal of

Financial Economics 83 1ndash32Leary M T (2009) Bank loan supply lender choice and corporate capital structure

Journal of Finance 64 1143ndash1185Leary M T and Roberts M R (2005) Do firms rebalance their capital structures

Journal of Finance 60 2575ndash2619Leland H E (1994) Corporate debt value bond covenants and optimal capital structure

Journal of finance 49 1213ndash1252Lemmon M L Roberts M R and Zender J F (2008) Back to the beginning persist-

ence and the cross-section of corporate capital structure Journal of Finance 63 1575ndash1608Long M S and Malitz I B (1985) The investment-financing nexus some empirical

evidence Midland Corporate Finance Journal 3 53ndash59Mackie-Mason J K (1990) Do taxes affect corporate financing decisions Journal of

Finance 45 1471ndash1493Morellec E Nikolov B and Schurhoff N (2012) Corporate governance and capital

structure dynamics Journal of Finance 67 803ndash848Myers S C (1984) The capital structure puzzle Journal of Finance 39 575ndash592Myers S C (1993) Still searching for optimal capital structure Journal of Applied

Corporate Finance 61 4ndash14Rajan R and Zingales L (1995) What do we know about capital structure some evidence

from international data Journal of Finance 50 1421ndash1460Strebulaev I A (2007) Do tests of capital structure theory mean what they say Journal of

Finance 62 1747ndash1787

Strebulaev I A and Whited T M (2012) Dynamic Models and Structural Estimation in

Corporate Finance Now Publisher Boston and Delft

Titman S and Wessels R (1988) The determinants of capital structure choice Journal of

Finance 43 1ndash21

Welch I (2004) Capital structure and stock returns Journal of Political Economy 112

106ndash131

Welch I (2011) Two common problems in capital structure research the financial-debt-

to-asset ratio and issuing activity versus leverage changes International Review of Finance

11 1ndash17

PROFITSndashLEVERAGE PUZZLEREVISITED 1453

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022

Hovakimian A Opler T and Titman S (2001) The debt-equity choice Journal of

Financial and Quantitative Analysis 36 1ndash24Jung K Kim Y and Stulz R M (1996) Timing investment opportunities managerial

discretion and the security issue decisions Journal of Financial Economics 42 159ndash185Kane A Marcus A J and McDonald R L (1984) How big is the tax advantage to

debt Journal of Finance 39 841ndash853Kayhan A and Titman S (2007) Firmsrsquo histories and their capital structures Journal of

Financial Economics 83 1ndash32Leary M T (2009) Bank loan supply lender choice and corporate capital structure

Journal of Finance 64 1143ndash1185Leary M T and Roberts M R (2005) Do firms rebalance their capital structures

Journal of Finance 60 2575ndash2619Leland H E (1994) Corporate debt value bond covenants and optimal capital structure

Journal of finance 49 1213ndash1252Lemmon M L Roberts M R and Zender J F (2008) Back to the beginning persist-

ence and the cross-section of corporate capital structure Journal of Finance 63 1575ndash1608Long M S and Malitz I B (1985) The investment-financing nexus some empirical

evidence Midland Corporate Finance Journal 3 53ndash59Mackie-Mason J K (1990) Do taxes affect corporate financing decisions Journal of

Finance 45 1471ndash1493Morellec E Nikolov B and Schurhoff N (2012) Corporate governance and capital

structure dynamics Journal of Finance 67 803ndash848Myers S C (1984) The capital structure puzzle Journal of Finance 39 575ndash592Myers S C (1993) Still searching for optimal capital structure Journal of Applied

Corporate Finance 61 4ndash14Rajan R and Zingales L (1995) What do we know about capital structure some evidence

from international data Journal of Finance 50 1421ndash1460Strebulaev I A (2007) Do tests of capital structure theory mean what they say Journal of

Finance 62 1747ndash1787

Strebulaev I A and Whited T M (2012) Dynamic Models and Structural Estimation in

Corporate Finance Now Publisher Boston and Delft

Titman S and Wessels R (1988) The determinants of capital structure choice Journal of

Finance 43 1ndash21

Welch I (2004) Capital structure and stock returns Journal of Political Economy 112

106ndash131

Welch I (2011) Two common problems in capital structure research the financial-debt-

to-asset ratio and issuing activity versus leverage changes International Review of Finance

11 1ndash17

PROFITSndashLEVERAGE PUZZLEREVISITED 1453

Dow

nloaded from httpsacadem

icoupcomrofarticle19414151568525 by guest on 02 August 2022