Zurich Minikurs on Behavioral Economics and Finance · Zurich Minikurs on Behavioral Economics and Finance Ulrike Malmendier June 10-16, 2009. ... production decisions, tailored to

Zurich Minikurs on Behavioral

Economics and Finance

Ulrike Malmendier

June 10-16, 2009

Outline of Lectures 2 and 3

1. Behavioral Economics in the Market

2. Behavioral Corporate Finance: Managerial Biases

3. Behavioral IO: Sorting, Contracts, Prices, Products

1 Behavioral Economics in the Mar-

ket

• Core concept in applied Behavioral Economics/Finance:

— Interaction betweenbiased (“behavioral”) agents andunbiased (“rational”) agentsin the market.

— Effects on trade, prices, other economic outcomes.

• Why?

— Individual biases are interesting in and of them-selves, but as economists we would like to explainmarket outcomes.

— Helps to address criticism about the real-worldrelevance of behavioral economics / behavioralfinance.

• Critiques

1. “Agents may display biases in the laboratory, butnot in the real world that is relevant to econo-mists.”

Example: Read Bob Aumann’s (very interesting)interview, conducted by Sergiu Hart, onhttp://www.ma.huji.ac.il/~hart/papers/md-publ-aumann.pdf?,especially his “grave doubts” about behavioraleconomics (p. 711): “behavioral economics isabout people looking at flashing green lights”(probability matching experiment) and is the op-posite of “empirical economics.”

→ Provides empirical evidence on biased con-sumer behavior

— directly: consumption decisions

— indirectly: production decisions,tailored to biases

2. “Even if agents do display biases outside the lab-oratory, they do not affect market outcomes be-cause (a) agents sort or learn (experience) and(b) the biases are not systematic.”

Example: Becker (1962), “Irrational Behavioraland Economic Theory”, JPE 70, 1-13: The de-mand curve is downward-sloping even if consumersare irrational. Argument: budget constraint.

→ Provides evidence on biases mattering for prices,traded quantities etc.

3. “Even if agents do display biases outside the lab-oratory and even if they do not “disappear” dueto sorting or experience, they are not of first-order importance.”

→ Provides evidence that fims react to biasesin the design of contracts or products; thus offirst-order importance.

• Applications

— Behavioral Finance I (Asset Pricing): Biasedtraders (“noise traders”) and unbiased traders(“arbitrageurs”).

Example: What explains “bubbles” in the stockmarket? That is, why do stock prices deviatefrom the fundamental value of the firm?

— Behavioral Finance II (Corporate Finance): Bi-ased investors (with “investor sentiment”) andrational firms.

Example: What explains the decision of a CEOto issue stock rather than bonds?

— Behavioral Finance III (Corporate Finance): Bi-ased CEOs and rational financiers (and other in-vestors).

Example: Why are merger announcements oftenaccompanied by a strongly negative stock marketreaction?

— Industrial Organization: Biased consumers andrational, profit-maximizing firms.

Example: Why do credit card companies offerteaser rates? Why are health-club membershipsautomatically renewed, even if the member doesnot attend?

• “Promise” of research in Behavioral Industrial Or-ganization and Behavioral Corporate Finance

— potential to tackle important unsolved puzzles;

— rich set of empirical applications;

— important welfare implications;

— may convince even the critic.

Behavioral Corporate Finance

Systematic deviations from our standard model of rational decision-making

Relevant to corporate finance from two perspectives:

Perspective 1: Investor biasesManagerial response = Non-standard corporate finance policies

Perspective 2: Managerial biasesNon-standard corporate finance policies Market response

Perspective 1: Biased Investors

Examples

• Investors sentiment Timing of security issuances Baker and Wurgler, 2000; 2002

• Employee sentiment Stock-based compensation to lower-level employees Oyer, 2004; Bergman and Jenter, 2005

Perspective 2: Biased Managers

Examples

• Overconfidence of CEOs “Urge to merge” / to overinvestMalmendier and Tate, 2005a; 2005b

• Sunk-cost fallacy of directors and CEOs Reluctance to disinvest

Kuhnen and Malmendier, 2005

Corporate Finance• Investment

• Mergers & Acquisitions

• Capital Structure

Three Puzzles• Investment




– Should depend only on investment opportunities (Tobin’s Q)– Depends on internal cash flow (Investment-cash flow sensitivity)– Even for large firms with access to debt and equity markets





• Mergers & Acquisitions– U.S. firms spent more than $3.4 trillion on over 12,000 mergers

during the last two decades.– Acquiring shareholders lost over $220 billion (net) at the

announcement of merger bids from 1980 to 2001 (negative announcement effects).

– Long-term returns appear to deepen the puzzle.• Capital Structure



• Mergers & Acquisitions– U.S. firms spent more than $3.4 trillion on over 12,000 mergers

during the last two decades.– Acquiring shareholders lost over $220 billion (net) at the

announcement of merger bids from 1980 to 2001 (negative announcement effects).

– Long-term returns appear to deepen the puzzle.• Capital Structure

– Pecking order of financing: Cash < Safe Debt < Risky Debt < Equity

– Debt conservatism: Companies issue too little debt given the taxadvantages (tax deductibility of interest payments)

– …

Research Agenda

• Traditional approach:– Market characteristics (taxes, bankruptcy costs)– Industry characteristics (asset tangibility)– Firm characteristics (information asymmetry)e. g. Trade-Off Theory, Pecking-Order Theory

• Alternative approach:– Managerial characteristics

Can individual differences between managers (managerial fixed effects) help to explain CF puzzles?Can we measure managerial beliefs / preferences and evaluate their impact on corporate decisions?

Research Agenda

• Managerial overconfidence(= overestimation of future cash flows generated by the manager)

provides a unifying framework for these three puzzles.• Investment: Overconfident managers perceive their firm

to be undervalued; thus reluctant to invest if they have to issue debt and, in particular, equity.

• Mergers: Overconfident managers overestimate returns to mergers; thus too prone to undertake (bad) mergers.

• Capital structure: Overconfident managers overestimate the value of risky securities; thus reluctant to issue debt and, in particular, equity.

Managerial Overconfidence

• DefinitionOverestimating own abilities / skills / knowledge Overestimating outcomes of own actions

• Stylized Fact: “Better-than-Average” Effect– Driving skills, IQ, … (Larwood and Whittaker, 1977; Svenson,

1981; Alicke, 1985)– Market entry decision (Camerer and Lovallo, 1999)

• Executives appear particularly prone to display overconfidence in experiments.– Three main factors (Weinstein, 1980; Alicke et al., 1995)

• Being in control (incl. illusion of control)• High commitment to good outcomes• Reference point not concrete

Evidence from Psychology on Overconfidence

1. “Better-than-average effect”• Abilities and Skills (IQ, driving skills)• Personal Situation (no severe illness, no divorce)

2. Overconfidence when• Noisy or Infrequent Feedback• (Illusion of) Control• Commitment

3. Other aspects of overconfidence(NOT in this paper)• overconfidence in precision (calibration)• time-variation

permanent,first moment

transitory,second moment

Evidence from Economics & Finance

• Overconfidence about abilities and self-control (Camerer-Lovallo 1999; O’Donoghue and Rabin 1999)

• Overconfidence of corporate decision-makers– Takeovers (Roll 1986)

– Risk-tolerance (Goel and Thakor 2000)

What Causes Mergers and Acquisitions?

Standard Stories

1. “Synergies”2. Market Power (1920s)3. Diversification (1960s)4. Market Discipline (1980s)5. Deregulation (1990s)

Alternatives: Departures from Rationality

Biased Market

Stock Price Bubbles(Shleifer and Vishny 2001)

Biased Managers

The Hubris Hypothesis(Roll 1986)

Efficiency-Driven

Many managements apparently were overexposed in impressionable childhood years to the story in which the imprisoned handsome prince is released from a toad’s body by a kiss from a beautiful princess. Consequently, they are certain their managerial kiss will do wonders for the profitability of Company T[arget]…

We’ve observed many kisses but very few miracles.

Nevertheless, many managerial princesses remain serenely confident about the future potency of their kisses—even after their corporate backyards are knee-deep in unresponsive toads.

-Warren Buffet

(Berkshire Hathaway Inc. Annual Report, 1981)

Overconfidence

Overconfident CEOs overestimate their ability to generate returns

• In their own company • In other companies

Implication for Mergers

Overvalue future returns of combined company Overvalue future returns of current company

Reluctant to issue equity

Too eager to merge

ModelAssumptions 1. CEO acts in interest of current shareholders.

(No agency problem.) 2. Efficient capital market.

(No asymmetric information.) Notation

AV = market value of the acquiring firm TV = market value of the target firm

V = market value of the combined firm (after cash payments)

AV̂ = acquiring CEO’s valuation of his firm V̂ = acquiring CEO’s valuation of the combined firm

c = cash used to finance the merger

Rational CEO

• Target shareholders demand share s of firm such that: cVsV T −= .

• CEO decides to merge if ( ) AT VcVV >−− (levels). ⇒ Merge if e > 0 (differences), where e is “synergies.”

⇒ First-best takeover decision.

• Post-acquisition value to current shareholders:

eVcVceVVcVVV ATTAT +=−−−++=−−= )()()(

⇒ 0=∂∂

cV (No financing prediction.)

Overconfident CEO (I)

• CEO overestimates future returns to own firm:

AA VV >ˆ CEO overestimates returns to merger:

AA VVVV −>− ˆˆ

• Target shareholders demand share s of firm such that: cVsV T −=

CEO believes he should have to sell s such that: cVVs T −=ˆ

Overconfident CEO (II)

• CEO decides to merge if

AT

T VV

cVVVcVV ˆ))(ˆ()(ˆ >⎥⎦

⎤⎢⎣

⎡ −−−−− (levels),

i.e. merges if

⎥⎦

⎤⎢⎣

⎡ −+−>+V

cVeVVee TAA ))(ˆˆ(ˆ (differences),

where ê are perceived “synergies.”

Empirical Predictions

Rational CEO Overconfident CEO

1. On average?2. Overconfident CEOs do more

mergers that are likely todestroy value

3. Overconfident CEOs do moremergers when they haveabundant internal resources

4. The announcement effect afteroverconfident CEOs make bidsis lower than for rational CEOs

Data

Data on private accounts1. Hall-Liebman (1998)

Yermack (1995)

Key: Panel data on stock andoption holdings of CEOs ofForbes 500 companies 1980-1994

2. Personal information aboutthese CEOs from

- Dun & Bradstreet- Who’s who in finance

Data on corporate accounts1. CRSP/COMPUSTAT

Cash flow, Q, stock price…

2. CRSP/SDC-merger databases

Acquisitions

Overconfidence

On private accounts

• Hold on to options.

Idea: Rational CEO who is- underdiversified- risk averse

should- exercise options early.

On corporate accounts

• Higher probability of acquiringanother company, particularlywhen:• Merger has low expected

value• Manager has lots of cash and

untapped debt capacity

Primary Measure of Overconfidence“Longholder”

(Malmendier and Tate 2003)

CEO holds an option until the year of expiration. CEO displays this behavior at least once during sample period.

minimizes impact of CEO wealth, risk aversion, diversification

Robustness Checks:1. Require option to be at least x% in the money at the beginning of

final year

2. Require CEO to always hold options to expiration

3. Compare “late exercisers” to “early exercisers”

Empirical Specification

Pr{Yit = 1 | X, Oit} = G(β1 + β2•Oit + X?γ) with i company O overconfidence

t year X controls Y acquisition (yes or no)

H0: β2 = 0 (overconfidence does not matter) H1: β2 > 0 (overconfidence does matter)







Identification Strategy

Case 1:Wayne Huizenga (Cook Data Services/Blockbuster)• CEO for all 14 years of sample• Longholder

M MM M M MH

1980 1981 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994

J Willard Marriott (Marriott International)• CEO for all 15 years of sample• Not a Longholder

1980 1981 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994

Logit & RandomEffects Logit

Fixed EffectsLogit

Case 2:Colgate Palmolive• Keith Crane CEO from 1980-1983 (Not a Longholder)• Reuben Mark CEO from 1984-1994 (Longholder)

M MM MH

1980 1981 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994

Keith Crane Reuben Mark

Yes No

Yes Yes

Table 4. Do Overconfident CEOs Complete More Mergers?

logit with controls random effects logit

logit with fixed effects

Size 0.8733 0.8600 0.6234(1.95)* (2.05)** (2.60)***

Qt-1 0.7296 0.7316 0.8291(2.97)*** (2.70)*** (1.11)

Cash Flow 2.0534 2.1816 2.6724(3.93)*** (3.68)*** (2.70)***

Ownership 1.2905 1.3482 0.8208(0.30) (0.28) (0.11)

Vested Options 1.5059 0.9217 0.2802(1.96)* (0.19) (2.36)**

Governance 0.6556 0.7192 1.0428(3.08)*** (2.17)** (0.21)

Longholder 1.5557 1.7006 2.5303(2.58)*** (3.09)*** (2.67)***

Year Fixed Effects yes yes yesObservations 3690 3690 2192Firms 327 327 184

Longholder = holds options until last year before expiration (at least once)Distribution: Logistic. Constant included.Dependent Variable: Acquistion (yes or no); Normalization: Capital.

Table 6. Are Overconfident CEOs Right to Hold Their Options? (I)

Percentile10th20th30th40th50th60th70th80th90th

MeanStandard Deviation

All exercises occur at the maximum stock price during the fiscal year

0.390.030.27

-0.030.030.10

Returns from exercising 1 year sooner and investing in the S&P 500 index

Return

0.19

-0.24-0.15-0.10-0.05

Table 6. Are Overconfident CEOs Right to Hold Their Options? (II)

logitrandom

effects logitfixed effects

logitSize 0.8471 0.8387 0.5703

(2.30)** (2.27)** (2.99)***Qt-1 0.7924 0.8038 0.8748

(2.27)** (2.00)** (0.80)Cash Flow 1.4647 1.4392 1.6066

(2.04)** (2.78)*** (1.72)*Stock Ownership 2.0954 1.9519 0.8058

(0.87) (0.61) (0.11)Vested Options 2.2233 1.8178 0.7333

(0.98) (0.62) (0.22)Corporate Governance 0.6053 0.6619 1.0221

(3.44)*** (2.58)** (0.10)Longholder: Did OK 1.1112 1.1339 1.0876

(0.44) (0.52) (0.15)Longholder: Should Have Exercised 1.7357 1.8857 4.6484

(1.67)* (2.17)** (2.29)**Year Fixed Effects yes yes yesObservations 3298 3298 1963Firms 312 312 167

Dependent Variable: Acquistion (yes or no) ; Normalization: Capital.

Longholder = holds options until last year before expiration (at least once)Do "Mistaken" Holders Drive the Acquisitiveness Result?

Distribution: Logistic. Constant included.

Alternative Explanations1. Inside Information or Signalling

• Mergers should “cluster” in final years of option term• Market should react favorably on merger announcement• CEOs should “win” by holding

2. Stock Price Bubbles• Year effects already removed• All cross-sectional firm variation already removed• Lagged stock returns should explain merger activity

3. Volatile Equity

4. Finance Training

Figure 10. Odds Ratios for different % in the money

0

0.5

1

1.5

2

2.5

3

3.50 10 20 30 40 50 60 70 80 90 100

% in the money

Odd

s R

atio

Fixed Effects LogitOdds RatioRandom Effects LogitOdds RatioLogit Odds Ratio

Robustness

• Do the results hold as we vary the percentage in themoney required for a holder to be overconfident?Yes.

• Do the results hold when we require overconfidentCEOs to always hold until expiration? Yes.

• Do the results hold when we limit the control groupto “timely exercisers?” Yes.

• Do the results hold when we control for otherobservable personal characteristics (Titles, Age, andTenure)? Yes.







Diversifying Mergers

1. Diversification discount(Lamont and Polk 2002; Servaes 1996; Berger and Ofek 1995; Lang and Stulz 1994)

2. Market understands ex ante (AT&T/NCR)(Morck, Shleifer, and Vishny 1990)

Table 14. Diversifying Mergers

logit logit with random effects


Longholder 1.6470 1.9132 3.6071(2.54)** (2.95)*** (2.66)***


Longholder 1.3596 1.4812 1.4791(1.26) (1.53) (0.71)

Year Fixed Effects yes yes yesObservations 3457 3457 1143Firms 320 320 96Regressions include Cash Flow, Qt-1, Size, Ownership, Vested Options, and Governance. Industries are Fama French industry groups.

Longholder = holds options until last year before expiration (at least once)

Dependent Variable: Diversifying merger (yes or no).

Dependent Variable: Intra-industry merger (yes or no).

Distribution: Logistic. Constant included; Normalization: Capital.







Empirical Specification

CARi = β1 + β2•Oi + X'γ + εi

with i company O overconfidenceX controls

[ ]( )∑ −=−=

1

1tititi rErCAR

where [ ]itrE is daily S&P 500 returns (α=0; β=1)

Table 14. Market Response

OLS OLS OLS(1) (2) (3)

Relatedness 0.0057 0.0050 0.0053(1.67)* (1.30) (1.56)

Corporate Governance 0.0079 0.0036 0.0073(2.18)** (0.64) (1.98)**

Cash Financing 0.014 0.0127 0.0145(3.91)*** (2.60)*** (3.99)***

Age -0.0005(1.46)

Boss 0.0001(0.04)

Longholder -0.0067 -0.0099 -0.0079(1.81)* (2.33)** (2.00)**

Year Fixed Effects yes yes yesIndustry Fixed Effects no yes noIndustry*Year Fixed Effects no yes noObservations 673 673 673R-squared 0.06 0.14 0.09Regressions include Ownership and Vested Options.

(at least once)Dependent Variable: Cumulative abnormal returns [-1,+1]

Longholder = holds options until last year before expiration

Do Outsiders Recognize CEO Overconfidence?

Portrayal in Business Press:

1. Articles in • New York Times • Business Week • Financial Times • The Economist • Wall Street Journal

2. Articles published 1980-1994 3. Articles which characterize CEO as

• Confident or optimistic • Not confident or not optimistic • Reliable, conservative, cautious, practical, steady or frugal

Measuring Press Portrayal

TOTALconfident =

1 if [“confident” + “optimistic”] > [“not confident” + “not optimistic + “reliable, conservative, cautious, practical, steady, frugal]

0 otherwise

Independent of the effects of coverage frequency

Market Perception versus CEO beliefs

• TOTALconfident positively and statistically significantly correlated with Longholder– Farrell and Mark are TOTALconfident– Marriott and Crane are not TOTALconfident

• TOTALconfident CEOs (like Longholders) are more acquisitive on average– Especially through diversifying mergers– Especially when they are financially unconstrained

Overconfidence – identified by CEO or market beliefs – leads to heightened acquisitiveness

Table 13. Press Coverage and Diversifying Mergers

logit logit with random effects


TOTALconfident 1.6971 1.7826 1.5077(2.95)*** (3.21)*** (1.48)


TOTALconfident 1.0424 1.0368 0.8856(0.20) (0.16) (0.31)

Year Fixed Effects yes yes yesObservations 3690 3690 1227Firms 326 326 100Regressions include Total Coverage, Cash Flow, Qt-1, Size, Ownership, Vested Options, and Governance. Industries are Fama French industry groups.

Dependent Variable: Diversifying merger (yes or no).

Dependent Variable: Intra-industry merger (yes or no).

Distribution: Logistic. Constant included; Normalization: Capital.

Conclusions

• Overconfident managers are more acquisitive.

• Much of this acquisitiveness is in the form ofdiversifying mergers.

• Overconfidence has largest impact if CEO hasabundant internal resources.

• The market reacts more negatively to the mergersof overconfident CEOs

Final Remarks on the Overconfidence Hypothesis

• Accounting for market reaction, we yield more subtle and testable predictions.

• Example: rational bankers (investors) overconfidence CEOs would like to do too many mergers, but may abstain if they have to tab outside financing.

Measurement

• Idea: Measure overconfidence in one setting and look at behavior in other setting.

Overconfidence

On private accounts

• Hold on to options.

Idea: Rational CEO who is- underdiversified- risk averse

should- exercise options early.

On corporate accounts

• Higher probability of acquiringanother company, particularlywhen:• Merger has low expected

value• Manager has lots of cash and

untapped debt capacity

Caveat

• Advantage of “identification in one setting and outcome in other setting”: paper does not rely on people being biased on average + proof of bias (rather than alternative explanations)

• Disadvantage: biases in one setting do not need to be present in other setting!– E.g. overconfident about making deadlines but

underconfident about driving abilities.– Thus: the two setting are ideally related, e.g. option

measure and investment/merger decisions are both about overestimating future stock price.

Other Applications

1. Corporate Investment2. Capital Structure

Two Patterns

1. Conditional on accessing external markets, preference for debt over equity (“pecking order”)

e.g. Shyam-Sunders and Myers (1999)

2. Debt conservatisme.g. Graham (2000)

Explanations and empirical evidence debated in capital structure literature.

e.g. Frank and Goyal (2005) survey

Two Patterns (II)

• Manager-level explanation: some pre-identified managers prefer internal financing over debt, debt over equity

Accommodates variation over time or across firms with identical market/industry/firm characteristicsPredicts patterns only in a subset of firms

• One framework: overconfidence

CEO Overconfidence

Overconfident managers overestimate their ability to generate value ↔Overestimate future (non-default) cash flows

perceive their companies' risky securities to be undervalued by the market reluctant to seek external financingconditional on external finance, prefer debt over equity Pecking Order

of Financing

Debt Conservativism

Figure 1. Stylized Example

33%

33%

33%

67%

22%

11%

0%

10%

20%

30%

40%

50%

60%

70%

80%

90%

100%


Equity Financing

Debt Financing

Cash Financing

Prediction 1

Conditional on accessing external financial markets, overconfident CEOs issue less equity than their non-overconfident peers.

Table 2. Public Security Issues

Years with a Security Issue

% of Issue Years with an Equity

Issue

% of Issue Years with a Debt Issue

% of Issue Years with a

Hybrid Security

IssueLongholder = 0 621 42% 57% 16%Longholder = 1 141 31% 63% 19%

Pre-Longholder = 1 91 31% 63% 23%Post-Longholder = 1 50 32% 64% 12%

Difference t (Longholder = 0 - Longholder = 1) 2.03** 0.85 0.85

Holder 67 = 0 95 39% 65% 21%Holder 67 = 1 182 23% 73% 16%Difference t 3.12*** 1.18 1.04

TOTALconfident = 0 452 48% 47% 18%TOTALconfident = 1 214 25% 79% 14%Difference t 5.37*** 6.77*** 1.43

Table 2. Logits (dependent variable: equity issue indicator)

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

Longholder -0.469 -0.592 -0.534 -0.46 -0.457(1.94)* (2.34)** (2.10)** (1.80)* (1.66)*

CEO Stock Ownership -0.266 -0.996 -1.279 -0.655(0.16) (0.59) (0.72) (0.34)

CEO Vested Options 6.766 4.669 4.234 7.328(3.43)*** (2.21)** (2.14)** (3.05)***

ln(Sales) -0.414 -0.437 -0.355(3.79)*** (3.70)*** (2.84)***

Q -0.088 -0.074 0.139(0.68) (0.56) (1.00)

Profitability -1.872 -1.493 -2.463(1.53) (1.21) (1.74)*

Tangibility 0.139 0.088 0.113(0.30) (0.19) (0.23)

Book Leverage 0.651 1.288(1.14) (2.07)**

Year Fixed Effects XObservations 762 644 627 617 617Number of Firms 330 174 171 171 171* significant at 10%; ** significant at 5%; *** significant at 1%

Alternative Framework

• Financing Deficit regressions– FDt = It + DIVt + ΔWt – CFt– FD captures amount of financing CEO has to

raise through debt or equity (accounting for bank loans, other private sources).

– Conditioning on “external financing.”• Additional advantages

– Larger sample; allows inclusion firm FE.– Comparability with capital structure literature.

Table 5. FD Framework (Dependent Variable: net debt issues)

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

Net Financing Deficit (FD) 0.729(9.90)***

Longholder -0.006 -0.005 -0.008 -0.008 -0.005(1.43) (1.37) (1.95)* (2.03)** (1.43)

Longholder * FD 0.350 0.348 0.332 0.322 0.334(1.78)* (1.77)* (1.77)* (1.69)* (1.90)*

CEO Stock Ownership 0.015 0.015 0.014 0.010(0.87) (0.90) (0.85) (0.76)

CEO Stock * FD 0.373 0.431 0.370 0.348(2.30)** (2.63)*** (2.14)** (2.17)**

CEO Vested Options -0.025 -0.021 0.000 0.011(1.49) (1.15) (0.00) (0.52)

CEO Vested Options * FD -0.088 -0.098 -0.135 -0.156(3.21)*** (3.59)*** (3.06)*** (3.76)***

Book Leverage -0.096(5.98)***

Book Leverage * FD -0.129(0.54)

FD Control Variables X XFD Control Variables * FD X XYear Fixed Effects X X XFirm Fixed Effects X X X X XFirm Fixed Effects * FD X X X X XObservations 2385 2385 2385 2385 2385 2346Number of Firms 263 263 263 263 263 262R-squared 0.75 0.93 0.93 0.94 0.94 0.94* significant at 10%; ** significant at 5%; *** significant at 1%

Prediction 2

Unconditionally, overconfident CEOs are more reluctant to access external capital markets than their non-overconfident peers.

May imply “debt conservativism.”

Measure of Debt Conservatism

• “Kink” proxy of Graham (2000):– Debt conservatism = firm can increase its interest payment

without lowering the marginal tax rate. – “Kink" = ratio of a hypothetical interest level at which the

marginal tax rate starts to fall (numerator) to the actual amount of interest (denominator) paid by the firm

• Mechanism: – When a firm is committed to low future interest payments, all

interest payments are likely to be deduced from future profits tax benefit = (interest payment)* (marginal corporate tax rate)marginal tax benefit flat

– When firm committed to pay high future interest, then increasingly possible that firm cannot generate enough profits to fully realize the interest tax shield

marginal tax benefit is decreasing

Table 7. Tobits (Kink censored above 8)

(1) (2) (3) (4)Longholder 1.122 1.256 0.605 0.852

(1.75)* (1.94)* (1.72)* (2.25)**CEO Stock Ownership 3.369 -1.049 -0.956

(1.01) (0.47) (0.43)CEO Vested Options -3.025 -3.170 -2.974

(0.70) (2.05)** (1.91)*Low Cash Status -0.123

(0.59)Longholder * (Low Cash Status) -0.654

(1.24)Kink Controls X X

Industry Fixed Effects X X

Observations 1726 1726 1726 1725Number of Firms 189 189 189 189* significant at 10%; ** significant at 5%; *** significant at 1%

Alternative Specification• Risky rather than riskless debt

– “Riskless debt = cash”– Results should not be driven by extremely-high kink firms.

• Check 1:– Sample split by measures of “debt riskiness” (S&P rating)– Possible endogeneity, but not much within firm variation– No result for “safe-debt firms”– No result for “financial-distress firms”– Strong result for “medium firms”

• Check 2:– Logit with 1 = “Firm debt conservative”

= “Kink bigger than 1”– Additional benefits of logit specification

• Inclusion of firm fixed effects• Clustering of standard errors

Table 8. LogitsLogit Logit Logit Logit Logit Logit(1) (2) (3) (4) (5) (6)

Longholder 0.606 0.721 0.552 0.836 1.116(1.59) (1.79)* (1.76)* (2.04)** (1.59)

Pre-Longholder 0.803(0.99)

Post-Longholder 1.657(1.98)**

Stock (%) 2.407 -0.443 -0.452 8.318 8.174(0.98) (0.33) (0.34) (1.64)* (1.65)*

Vested Option (%) -2.147 0.175 0.287 -4.591 -4.195(1.79)* (0.10) (0.17) (0.08) (0.78)

Graham (2000) controls X X X X

I(Bottom 40% of Cash Stockt-1 / Mean Industry Investmentt) (CS) -0.050 0.340 0.354(0.18) (0.97) (1.01)

Longholder * CS -0.659 -1.034(1.26) (1.58)

Pre-Longholder * CS 0.280(0.44)

Post-Longholder * CS -2.256(1.90)*

Firm Fixed Effects X XObservations 1726 1726 1726 1725 745 745Number of Firms 75 75* significant at 10%; ** significant at 5%; *** significant at 1%

Kink ≤ 1 1 < Kink ≤ 3 3 < Kink ≤ 7 Kink ≥ 810th percentile -0.00834 -0.02923 -0.02668 -0.0516225th percentile 0.0000 -0.00003 -0.01055 -0.0128650th percentile 0.00544 0.00180 0.0000 0.000075th percentile 0.04148 0.00629 0.00348 0.0079490th percentile 0.09536 0.01733 0.02928 0.01685

Observations 37 110 111 96Mean 0.02869 0.00600 0.00497 0.00352Standard Deviation 0.06086 0.05291 0.08199 0.09174

Table 9. Distribution of Longholder CEOs' Net Equity Issues by KinkThe sample is all firm years in which Longholder equals 1. Longholder is a binary variable where 1 signifiesthat the CEO at some point during his tenure held an option package until the last year before expiration,provided that the package was at least 40% in the money entering its last year. Kink (Graham (2000)) is theamount of interest at the point where the marginal benefit function becomes downward sloping, as aproportion of actual interest expense. Net equity issues are sales of common stock minus stock repurchasesand are normalized by beginning of the year assets.

Managerial Characteristics more broadly

• Consider variation in managers' personal histories which is likely to generate differences in beliefs.

• Two ideas:1. Depression Babies

CEOs who grew up during the Depression are likely to have less faith in capital markets

May lean excessively on internal financing.2. Serving in the military

Military service typically occurs in early adulthood and, particularly when it results in combat exposure, is likely to lead to more aggressive beliefs (or lower risk aversion)

May make less conservative capital structure choices.

Depression Babies

• Prediction: heightened preference for internal financing

• Note: different from overconfidence!– Depression Babies: conservative beliefs arising from

early life experience during a macroeconomic crisis – Does not predict a mis-assessment of expected

returns to investment No prediction for investment policy. No prediction for the choice between debt and equity conditional on accessing external markets or, ultimately, for the firm's leverage ratio.

Table 9. Depression Baby and Financing Choices

Panel A. Kink Regressions(1) (2)

Tobit TobitDepression Baby 0.8984 0.5053

(2.27)** (2.07)**Age -0.0260

(-1.38)Tenure -0.0159

(-0.88)Kink Controls XIndustry Fixed Effects X

Observations 1,717 1,717* significant at 10%; ** significant at 5%; *** significant at 1%

Notes:

• Depression Baby indicates CEOs born between 1920 and 1929.

• Kink is the amount of interest at which the marginal benefit function starts to slope down, as a proportion of actual interest expense.

Interpretation

• Depression Babies have significantly higher levels of the kink variable, even including the full set of controls from Graham (2000) (Panel A, Table 9).

• Economically, the 0.5053 increase in kink represents a 13% increase from the overall sample mean.

• Need to verify that higher kinks among Depression Babies do not come from substituting equity issuance

Table 9. Depression Baby and Financing Choices

Notes:

• Depression Baby indicates CEOs born between 1920 and 1929.

• Kink is the amount of interest at which the marginal benefit function starts to slope down, as a proportion of actual interest expense.

Panel B. Distribution of "Depression Baby" Net Equity Issues by KinkKink ≤ 1 1 < Kink ≤ 3 3 < Kink ≤ 7 Kink > 7

10th percentile -0.00846 -0.03568 -0.04293 -0.0625425th percentile 0 -0.00855 -0.01158 -0.0231550th percentile 0.00104 0.00047 0.00001 075th percentile 0.00800 0.00570 0.00575 0.0052390th percentile 0.05131 0.04080 0.01893 0.01646

Observations 74 270 240 175Mean 0.00950 0.00277 -0.00088 -0.01053Standard Deviation 0.03470 0.05085 0.07096 0.06885

Army Babies

• Prediction: CEOs with a military background pursue more aggressive financial policies.

Higher risk tolerance implies heightened use of debt financing conditional on accessing risky capital Ultimately, higher leverage.

• Note: The same mechanism is likely to lead to heightened investment. (We find evidence of heightened investment among military CEOs, too.)

Table 10. Military Service and Market LeverageMilitary Service Controls

WWII Veterans

Military, Longholder, Depression

Army 0.0353 0.0326 -0.0015 0.0299(1.69)* (1.92)* (-0.08) (1.89)*

World War II Veteran 0.0695(2.21)**

Depression Baby 0.0048(0.24)

CEO Stock Ownership 0.0291(0.73)

CEO Vested Options -0.0261(-1.60)

Longholder 0.0535(1.71)*

Usual controls X X XYear Effects X X X XFirm Effects X X X X

Observations 1,626 1,626 1,626 1,614Number of Firms 210 210 210 208Adjusted R-squared (withi 0.08 0.19 0.19 0.21* significant at 10%; ** significant at 5%; *** significant at 1%

Notes:

• Comparing leverage chosen by CEOs with prior military service to the leverage chosen by their predecessors and/or successors in their firms.

Army Babies

• Military CEOs choose higher significantly leverage. • Economically smaller effect than overconfidence efect, increasing

leverage 13% from its mean, or by roughly 17% of a standard deviation.

• Impact of military service on leverage choices comes primarily from World War II veterans: 25% higher leverage than the sample mean.– Consistent with the importance of combat exposure in shaping

individual risk attitudes and beliefs. It also suggests the importance of achieving a major victory in reinforcing those beliefs.

– Less consistent with a self-selection mechanism under which aggressive or risk-tolerant individuals choose to serve in the military (and then later take more aggressive decisions as CEOs), as non-voluntary service was common during World War II.

• Jointly, both Longholder and Army positive and significant, suggesting that they capture the behavior of different subsets of CEOs and/or different types of beliefs.

Conclusions

• Overconfidence induces pecking-order behavior & debt conservatism.– “Add-on” to traditional market/industry/firm

level determinants of capital structure (trade-off theory, pecking-order theory).

• Overall, the results indicate the importance of managerial beliefs and characteristics in determining firm-level financial choices.

Documents

Zurich Minikurs on Behavioral Economics and Finance · Zurich Minikurs on Behavioral Economics and Finance Ulrike Malmendier June 10-16, 2009. ... production decisions, tailored to