Payout initiation by IPO firms: The choice between dividends and share repurchases

The Quarterly Review of Economics and Finance 49 (2009) 1275–1297

Contents lists available at ScienceDirect

The Quarterly Review ofEconomics and Finance

journa l homepage: www.e lsev ier .com/ locate /qre f

Payout initiation by IPO firms: The choice betweendividends and share repurchases

Bharat A. Jaina,∗, Chander Shekharb, Violet Torbeyc

a Department of Finance, Towson University, 800 York Road, Towson, MD 21252, United Statesb University of Melbourne, Carlton, VIC 3053, Australiac QSuper Limited, Brisbane, QLD 4000, Australia

a r t i c l e i n f o

Article history:Received 11 October 2008Received in revised form 13 July 2009Accepted 10 September 2009Available online 15 September 2009

JEL classification:G35G32G31

Keywords:Dividend initiationShare repurchasesIPOsHazard analysis

a b s t r a c t

This study evaluates the economics of the choice of form of pay-out initiation mechanism adopted by IPO firms. Our results suggestthat IPO firms demonstrate a preference for repurchases over div-idends as the specific form of payout initiation mechanism. Wehowever, find that while the market views post-IPO payout ini-tiations favorably, it is indifferent to the specific form of payoutmechanism adopted. Further, we find that dividends and repur-chases represent distinct payout mechanisms adopted by IPO firmswith fundamentally different characteristics and motivation to ini-tiate payouts during the post-IPO phase. Our results suggest thatwhile dividend initiations are primarily driven by life cycle andcatering theory considerations, signaling theory provides the morelikely explanation for payout initiations through share repurchases.

© 2009 The Board of Trustees of the University of Illinois.Published by Elsevier B.V. All rights reserved.

1. Introduction

Initial public offering (IPO) firms typically go public on the promise of growth thereby conditioninginvestors to expect capital gains rather than dividends or share repurchases during the post-IPO phase.In fact, the offering prospectuses of most IPO issuers usually indicate that the firm is unlikely to bein a position to pay dividends in the foreseeable future. Since IPO firms are typically young, growthoriented companies that are pursuing innovative products and technologies, they are expected to

∗ Corresponding author. Tel.: +1 410 704 3542.E-mail addresses: [email protected] (B.A. Jain), [email protected] (C. Shekhar), [email protected]

(V. Torbey).

1062-9769/$ – see front matter © 2009 The Board of Trustees of the University of Illinois. Published by Elsevier B.V. All rights reserved.doi:10.1016/j.qref.2009.09.003

http://www.sciencedirect.com/science/journal/10629769

http://www.elsevier.com/locate/qref

mailto:[email protected]



dx.doi.org/10.1016/j.qref.2009.09.003

1276 B.A. Jain et al. / The Quarterly Review of Economics and Finance 49 (2009) 1275–1297

invest substantially in areas such as R&D, capital expenditures, and advertising during the post-IPOphase in an effort to gain market share and achieve technological dominance in their rapidly evolvingproduct markets and industries. At the time of going public, these firms are often cash flow negativeand are likely to continue to need substantial external financing during the post-IPO phase to sustaintheir high growth rates as well as help finance acquisitions. Further, even if IPO issuing firms areprofitable at the time of going public, they are likely to favor earnings retention over distribution toensure availability of financing for their growth projects and thereby reduce their reliance on externalcapital markets, which are often unpredictable and can go through periods when financing constraintsare high.

Despite the expectation that IPO firms are unlikely to be in a position to initiate cash flow payouts,extant empirical evidence suggests that a small but economically significant proportion of IPO firmstends to pay dividends during the early post-IPO phase. For instance, Michaely and Shaw (1994) reportthat approximately 22% of IPO firms start paying dividends within the first 3 years of going public.Similarly, Fama and French (2001) report that 25% of new lists that survive eventually start payingdividends. One potential explanation for payout initiations by IPO firms is based on the life cycle the-ory of dividends, which suggests that a firm’s tradeoff between earnings retention versus distributionevolves over time and at a certain stage of development when profits accumulate and availabilityof profitable investment opportunities decline, the desirability of dividends over retention increases(DeAngelo, DeAngelo, & Stulz, 2006; Fama & French, 2001; Grullon, Michaely & Swaminathan, 2002).An alternative explanation for cash flow payouts by IPO firms can be inferred from the catering theorywhich argues that firms tend to pay dividends when investors desire them as reflected by the dividendpremiums offered (Baker & Wurgler, 2004). As such, catering theory implies that even if IPO firms havenot reached a stage in their life cycle where payouts become desirable, they might initiate distribu-tions to satisfy investor demand. Consistent with both the life cycle and catering theories, a surveyof financial executives revealed that managers attribute an increase in earnings and/or demand byinstitutional investors as the primary motivation for firms to initiate payouts (Brav, Graham, Harvey,& Michaely, 2005).

While an extensive body of corporate finance literature has focused on the payout policies of estab-lished firms, relatively little is known regarding payout behavior of newly public firms. Further, thelimited research on payout policies of IPO firms has largely focused either on the valuation effectsof dividend initiations or on the identification of factors influencing the occurrence and timing ofdividend initiations (Bulan, Subramanian, & Tanlu, 2007; Desmukh, 2003; Kale, Kini, & Payne, 2006;Lipson, Maquieira, & Megginson, 1998). For instance, Kale et al. (2006) identify the determinants ofdividend initiations as well as evaluate factors influencing the level and timing of payout initiationsby IPO firms. The authors conclude that the determinants of dividend initiation identified in theirstudy are consistent with the predictions of several major theories of dividends such as signaling,tax, transaction costs, clientele, agency, catering, and residual income. The above-described literatureon payout policies of IPO firms has however, largely focused on dividends and not evaluated sharerepurchases as an alternative form of payout initiation mechanism available to IPO firms.

In the context of established firms, researchers have provided compelling empirical evidence toindicate that over the past two decades a significant shift has occurred in the corporate payout behav-ior of U.S. firms with share repurchases emerging as an economically significant phenomenon andovertaking dividends as a percentage of total payouts (Brav et al., 2005; Fama & French, 2001; Grullon& Michaely, 2002; Jagannathan, Stephens, & Weisbach, 2000; Skinner, 2008). For instance, Grullon andMichaely (2002) point out that that during the 1985–2000 period, a majority of firms initiated cashpayouts through repurchases rather than dividends. The emergence of repurchases as an economi-cally significant phenomenon has generated a vigorous debate in the corporate finance literature asto whether dividends and repurchases are substitutes or alternatively whether they represent twodistinct forms of payout mechanisms adopted by different types of firms at different stages in theirevolution. For instance, support for the substitution hypothesis is based on evidence suggesting a directcorrespondence between the increasing propensities for the adoption of repurchases with a declin-ing trend to pay dividends, as well as on indications that both dividends and repurchases are linkedto earnings (Grullon & Michaely, 2002; Skinner, 2008). On the other hand, researchers have arguedthat managers tend to use dividends to pay out permanent cash flows, while repurchases are used to

B.A. Jain et al. / The Quarterly Review of Economics and Finance 49 (2009) 1275–1297 1277

pay out temporary cash flows relatively quickly and therefore represent two distinct forms of payoutmechanisms utilized by firms in fundamentally different circumstances (Jagannathan et al., 2000).Further support for the notion that dividends and repurchases represent distinct forms of payouts canbe inferred from research that indicates that while factors such as stock price performance, excesscash on the balance sheet, and dilutive effects of options drive repurchases, life cycle factors such asprofitability, maturity, size, leverage, and reduced investment opportunities tend to favor dividends(Brav et al., 2005; Grullon & Michaely, 2002; Jagannathan et al., 2000; Kahle, 2002; Skinner, 2008).

The corporate finance literature has argued that the dividend initiation decision represents oneof the most important payout decisions a firm must make (Kale et al., 2006). Therefore, in additionto issues such as “if and when” to initiate payouts and the size of the payout, managers are alsolikely to focus on the economics of the choice of the specific form of payout initiation mechanism(dividends versus repurchases) to be adopted by the firm. Since IPO firms face considerable challengesand uncertainties related to industry conditions, technological shifts, financing conditions, and productmarket competition during the post-IPO phase, it provides a useful setting to evaluate the choice ofform of payout initiation mechanism. On the one hand, in a highly uncertain and evolving competitiveenvironment, the inherent flexibility of repurchases relative to the rigidity of dividends is likely tofavor the adoption of repurchases as the payout initiation mechanism. Consistent with this conjecture,Skinner (2008) argues that newer firms with no history of dividends are likely to substitute repurchasesfor dividends. Similarly, Brav et al. (2005) report that two-thirds of managers of non-payout firmsindicated a preference for repurchases over dividends as the mechanism to initiate cash flow payouts.On the other hand, investor preference for the dependability of dividends relative to uncertaintyof repurchases and the potential for repurchase premia (Chowdhry & Nanda, 1994) tends to favordividends as the payout initiation mechanism. As such, research focused on examining the economicconsequences of the choice between repurchases versus dividends for IPO firms is likely to be a fruitfularea of inquiry.

Our study therefore attempts to extend the extant literature on payout initiation by IPO firms byevaluating both dividends and share repurchases as alternative mechanisms available to firms to ini-tiate post-IPO cash flow payouts. In order to evaluate the economics of the choice between dividendsversus repurchases as payout initiation mechanisms, our research addresses the following relatedquestions: (a) Are IPO firms more likely to initiate payouts through repurchases rather than dividends?(b) Does the market react differentially to the announcement of payout initiations via dividends ver-sus repurchases? (c) What factors influence the choice between dividends versus repurchases for IPOfirms? and (d) Is the timing of payout initiations via repurchases affected by the same set of economicfactors as those affecting the timing of dividend initiations? Overall, by extending research on pay-out initiations by IPO firms to include both dividends and share repurchases, our study is likely toprovide a more complete picture of the payout behavior of entrepreneurial firms than what can beobtained from studies focused solely on dividend initiations. Further, the above analysis is likely toprovide additional insights on the extent of substitutability between dividends and share repurchasesby extending research beyond established companies to include IPO firms.

2. Theoretical analysis of post-IPO payout initiation decision

The extant literature on payout policies of corporations has provided several theoretical modelssuch as signaling theory, residual dividend theory, life cycle theory, catering theory, clientele theory,agency theory of free cash flows, and tax theory to explain payout decisions by firms (Allen & Faulhaber,1989; Baker & Wurgler, 2004; Jensen, 1986; Kale & Noe, 1990; Miller & Rock, 1985). For instance, thecorporate payout literature has generally argued that dividends are influenced by life cycle relatedfactors such as profitability, maturity, investment opportunities, leverage, and size (DeAngelo et al.,2006; Fama & French, 2001; Grullon & Michaely, 2008; Kale et al., 2006). In addition to life cycle theory,we draw from several of the above cited theoretical models of dividends to identify factors that havethe potential to impact the choice of form of payout initiation by IPO firms. Specifically, we discussbelow the impact of venture capital financing, choice of financing strategy, extent of product marketcompetition, investor demand for dividends, and industry technological focus on the choice of formof payout initiation mechanism adopted by IPO firms.


2.1. Venture capital involvement

Venture capitalists (henceforth, VCs) are active investors who, in addition to providing capital, oftenparticipate in aspects such as strategic and operational planning, capital structure choices, design ofhuman resource policies, and CEO and investment banker selection decisions for their portfolio com-panies (Gorman & Sahlman, 1989; Hellman & Puri, 2002; Jain & Kini, 1995; Lerner, 1994; Sahlman,1990). Additionally, VCs tend to serve on the board of directors of their portfolio companies and there-fore play an influential role in guiding the firm through the IPO process as well as monitor managementduring the post-IPO phase (Jain & Kini, 1995; Lerner, 1994). As such, VCs are well positioned to play aninfluential role in shaping the post-IPO financial policy choices of IPO firms. While an extensive bodyof literature has focused on the role and impact of VC participation in the creation and developmentof public corporations, relatively little is known regarding the specific impact of their involvement onpayout behavior of IPO firms. We therefore attempt to gain insights into this unexplored area by exam-ining empirically the relationship between VC participation and choice of form of payout initiation byIPO firms.

The life cycle and dividend signaling theories provide the theoretical background to explore the linkbetween VC involvement and choice of post-IPO payout initiation mechanism adopted by IPO firms.The life cycle theory suggests that dividends are typically paid by mature, profitable, established firmswith low growth prospects while earnings retention is preferred by young, high growth firms with anabundance of investment opportunities and limited resources (DeAngelo et al., 2006). Since VC backedfirms are typically young, high growth firms that make the transition from private to public firms atan earlier stage in their growth cycle relative to similar non-VC backed firms (Lerner, 1994), the lifecycle theory would suggest that VC backed IPO firms are more likely to prefer retention to payoutsand therefore are not expected to initiate dividends during the post-IPO phase.

The signaling theory on the other hand, provides theoretical support for the argument that althoughVC backed firms are unlikely to initiate post-IPO dividends; they may find it beneficial to implementrepurchases during the post-IPO phase. Venture capitalist firms typically view the IPO as a mechanismto facilitate their eventual exit, since their investment objective is the realization of capital gains eitherat the IPO or shortly thereafter, rather than collecting a stream of dividends in the future. Additionally,since VCs seek to exit their IPO investments through a sale of their stake during the post-IPO period,they are likely to be highly sensitive to the market valuation of their IPO firms. As argued by Kale etal. (2006), firms are concerned about correct market valuation when shares have to be sold either byinsiders or by the firm for its investment needs. Further, the extant literature has suggested that sharerepurchases have the potential to serve as a signaling mechanism to correct undervaluation of thefirm’s stock (Grullon & Michaely, 2004). As such, in order to improve stock valuations prior to theirexit, VCs are more likely to encourage management to initiate payouts through repurchases ratherthan dividends. Additionally, as has been argued in the literature, employee options provide firmswith incentives to repurchase shares to avoid dilution (Kahle, 2002). Since VC backed IPO firms arequicker and more likely to adopt employee stock option programs relative to non-VC backed firms(Hellman & Puri, 2002), we would expect a higher likelihood of VC backed firms implementing sharerepurchases relative to non-VC backed firms.

In line with our above discussion, Allen, Bernardo, and Welch (2000) argue that in general, thepresence of any shareholders “who can mitigate the agency/information concerns but who do notvalue dividends, we would expect target companies to substitute re-investment, share repurchases, orother means of payouts for dividend payments.” Therefore, on the basis of the life cycle and signalingtheories, we expect VC participation to be negatively related to the probability of post-IPO payoutinitiation occurring via dividends rather than repurchases.

2.2. Post-IPO financing strategy

At the time of going public, IPO firms provide a window into their post-IPO financing strategyby providing information regarding the planned uses for the capital raised at the IPO. Commonlycited uses of IPO proceeds listed in IPO offering prospectuses include allocations for items such ascapital expenditures, working capital requirements, general corporate purposes, acquisition financ-


ing, marketing and sales, product development/R&D, and retiring/refinancing debt. Since informationregarding planned uses of IPO proceeds is likely to provide an idea of the extent of post-IPO demandfor additional external funds as well as the financial strategy of the IPO firm, there is likely to be a linkbetween variables associated with uses of IPO proceeds and the form of payout initiation mechanismadopted by IPO firms.

It is logical to infer that IPO firms indicating a larger number of uses for the IPO proceeds are likelyto have a greater demand for additional capital during the post-IPO phase compared to firms thatindicate a smaller number of uses for the IPO funds. Further, a more extensive list of uses of equitycapital is indicative of stronger growth prospects and fewer internal resources compared to similarfirms that indicate fewer uses for the proceeds raised at the IPO. As such, the proportion of earned tocontributed capital is likely to be lower for firms indicating a larger number of uses of IPO proceedsrelative to firms indicating fewer uses for the IPO proceeds. Since the life cycle theory of dividendspredicts a positive relationship between earned to contributed capital and dividend payout (DeAngeloet al., 2006), we would expect to find a negative relationship between number of uses of IPO proceedsand probability of dividend initiation by IPO firms.

On the other hand, firms with a larger number of uses for funds raised at the IPO may be more likelyto initiate share repurchases in an effort to signal the quality of their growth prospects in advance ofplanned equity issuances. As argued earlier, IPO firms that indicate a larger number of uses for IPOproceeds are more likely to seek additional equity financing during the post-IPO phase to continueto fuel their greater growth prospects. Firms seeking to issue equity are more likely to be concernedabout the market value of their stock and may therefore seek to use repurchase announcements asa signaling mechanism to improve their market valuation (Kale et al., 2006). As such, on the basis ofthe life cycle and signaling theory of dividends, we expect that the higher the number of uses of IPOproceeds, the more likely the firm will initiate payouts via repurchases rather than dividends.

In addition to indicating the number of planned uses for IPO proceeds, issuing firms also identifythe primary intended use of IPO proceeds. This can shed additional light on the strategic intent ofthe firm in areas such as capital structure choices, growth strategy, and product market strategy.For instance, issuing firms, indicating that the primary intended use for the IPO proceeds is to financeacquisitions, are signaling intent to pursue an aggressive acquisition strategy, which likely reduces theprobability that cash flows would be available for distributing dividends to shareholders. However,since company stock serves as a currency to finance acquisitions, acquisition oriented firms are likelyto adopt stock repurchases to enhance their stock valuations and thereby lower the cost of theiracquisitions. Therefore, we would expect acquisition oriented firms to initiate post-IPO payouts viarepurchases rather than dividends. Similarly, firms that plan to use the IPO proceeds to rebalance theircapital structure or reduce leverage are signaling the likelihood of higher future cash flows for equityholders and are therefore more likely to be in a position to initiate dividends. As such, we empiricallyevaluate whether the post-IPO choice of form of payout initiation is related to the primary intendeduse for the IPO funds such as acquisition financing, refinancing/retiring debt, or general corporatepurposes.

2.3. Extent of post-IPO product market competition

Recently, researchers have attempted to examine the link between product market competitionand payout policies of firms. For instance, Grullon and Michaely (2008) argue that product marketcompetition through its effect on agency conflicts can be an important factor influencing a firm’sdecision to payout excess cash flows to shareholders. Further, they argue that product market compe-tition has two potentially opposing effects on payout policy and consequently propose the outcomeversus substitution hypotheses to explain these alternative effects. The outcome model is based onthe premise that product market competition raises the risk and cost of overinvestment and thereforeputs pressure on management to pay out cash flows to shareholders thereby reducing their ability topursue unprofitable projects. Therefore, the outcome model suggests that competition affects corpo-rate payout policy in ways similar to a strong legal system by creating conditions forcing managersto behave efficiently by distributing cash flows rather than deploying them in unprofitable projects(Grullon & Michaely, 2008). On the other hand, the substitution model is based on the premise that


managers use payouts as an internal disciplinary device to substitute for the lack of external disci-plinary mechanisms such as product market competition in order to establish a good reputation inthe capital markets thereby enhancing their ability to raise capital on attractive terms. As such, whilethe outcome model predicts a positive relationship between extent of product market competitionand cash flow payouts, the substitution model points to a negative relationship. Using industry con-centration as a proxy for product market competition, Grullon and Michaely (2008) find that firms inless concentrated industries (and therefore facing stronger product market competition) have signif-icantly higher payout ratios and exhibit a higher probability for distributing cash flows, and concludethat this evidence provides support for their outcome model.

However, while Grullon and Michaely (2008) provide competing explanations and evidence regard-ing the link between product market competition and payout policies of firms, they do not directlyaddress the question of whether the extent of product market competition influences the choice ofform of cash flow payouts. Drawing on their outcome and substitution hypotheses, we attempt toexamine the link between the extent of post-IPO product market competition and the choice of formof payout initiation by IPO firms. In the outcome model, strong product market competition raises therisk of overinvestment and undertaking unprofitable projects, forcing managers to be more efficientby distributing cash flows to shareholders thereby lowering the likelihood of the firm pursuing neg-ative net present value projects. However, in order to retain the ability to increase or decrease thelevel of investment in response to changes in product market conditions or the availability of attrac-tive investment opportunities, managers are more likely to prefer payout mechanisms that permitflexibility on aspects such as level and timing of payouts and further does not implicitly commit thefirm to continue to make the payouts in the foreseeable future. Since repurchases represent a moreflexible payout mechanism on some of the above aspects, we would expect strong product marketcompetition to result in a higher likelihood of post-IPO payout initiations occurring in the form ofrepurchases rather than dividends.

On the other hand, on the basis of the substitution model, which predicts a higher likelihood of pay-outs when product market competition is low, we argue that the preferred form of payout is more likelyto be dividends rather than repurchases. Since in the substitution model payouts are used as an internaldisciplinary device to substitute for the lack of external disciplinary mechanisms such as competition,managers are likely to prefer conveying an intent to maintain regular payouts to effectively signalthat internal control mechanisms are consistently in place thereby ensuring that managers will likelypursue value-maximizing strategies. Therefore, under the substitution hypothesis, the predictabilityof dividends is likely to be preferred to the uncertainty of repurchases. As such, we expect weak prod-uct market competition to result in a higher likelihood of post-IPO payout initiations occurring viadividends rather than share repurchases.

Firms that are likely to face strong product market competition are more likely to invest in R&Dand capital expenditure compared to IPO firms facing weaker product market competition. As such, inorder to evaluate the above-described link between product market competition and form of payoutinitiation mechanism adopted by IPO firms, we utilize investments in R&D and capital expendituresat the time of going public as proxies for the extent of product market competition faced by IPOfirms. Therefore, drawing from our argument that the Grullon and Michaely’s (2008) outcome modelimplies management preference for the flexibility afforded by repurchases over the inherent rigidityof dividends, we would expect a positive relationship between the extent of investments in R&D andcapital expenditures and the probability of payout initiation via repurchases rather than dividends.Similarly, on the basis of our argument that the Grullon and Michaely (2008) substitution modelimplies management preference for the stability of dividends, we would expect a negative relationshipbetween investments in R&D and capital expenditures on the probability of payout initiation viarepurchases rather than dividends by IPO firms.

2.4. Industry technological focus

IPO firms typically belong to a diverse set of industries that can be broadly classified as high tech-nology, emerging industries versus mature stable industries. Issuers in emerging, high technologyindustries face an innovation-oriented environment, with no dominant technologies, shorter product


life cycles, evolving product markets, and criticality of knowledge capital (Liu, 2000). Further, firmsin emerging industries are subject to substantial uncertainties with regard to aspects such as demandconditions, product life cycles, extent of competition from new entrants, and changes in product mar-ket conditions that tend to exacerbate the inherent information asymmetry between managementand shareholders of IPO firms. Since high technology industries are more susceptible to shocks as aresult of changes in technology, government policy or supply and demand conditions, industry firmsare forced to invest additional resources as they strive to adapt to the structural changes occurring intheir industries thereby increasing their demand for external funds. As such, IPO firms in high technol-ogy emerging industries, even if profitable, are more likely to retain all their earnings and in additionseek substantial additional funding from external capital markets to remain competitive.

On the other hand, firms in mature industries face a more stable economic environment, withwell-developed product markets and established industry leaders. Further, IPO firms in more matureindustries are likely to be presented with relatively fewer investment opportunities, face lower likeli-hood of demand or technological shocks, and are unlikely to face many new entrants. Therefore, suchfirms are likely to have more stable and predictable cash flows and higher earned to contributed cap-ital. As such, on the basis of life cycle theory of dividends (DeAngelo et al., 2006), we would expect alower probability of post-IPO dividend initiation by IPO firms belonging to high technology, emergingindustries compared to IPO firms in mature industries.

We however, additionally argue that while IPO firms in high technology industries are not wellpositioned to commit to a regular stream of dividend payments during the post-IPO phase, they arelikely to find it advantageous to adopt repurchases from time to time. This is due to the greater degreeof information asymmetry surrounding the prospects of high technology firms, and that managementmay need to signal their favorable future prospects through repurchases in an effort to increase thevaluation of their stock in advance of equity issuance. Further, firms in emerging, high technologyindustries exhibit a greater propensity relative to firms in mature industries to issue executive andemployee options to recruit, reward, and retain key employees. Since repurchases help offset stockoption dilution (and unlike dividends do not lower option values), management is likely to maximizethe value of their options by substituting repurchases for dividends (Kahle, 2002). Therefore, we wouldexpect that the probability of payout initiation through repurchases rather than dividends is higherfor IPO firms in high technology industries.

2.5. Extend of demand for dividends

Recently, Baker and Wurgler (2004) proposed the catering theory of dividends, which is based onthe premise that managers tend to cater to investor demand for dividends by paying dividends wheninvestors put a stock price premium on payers and avoiding dividends when investors prefer non-payers. Their argument is based on the assumption that investor demand for dividend paying stocks istime varying causing the relative price of dividend and non-dividend payers to vary under the assump-tion of arbitrage limits (Li & Lie, 2006). Consistent with their catering theory, Baker and Wurgler (2004)find evidence indicating a positive relationship between their measure of dividend premiums and theprobability of dividend initiations. They do not, however, find a statistically significant relationshipbetween dividend premiums and announcement returns.

Subsequently, Li and Lie (2006) extended the catering model beyond dividend initiations to includedividend changes (increases and decreases). Consistent with catering theory, they find that the decisionto change dividends as well as the magnitude of the change is dependent on the dividend premiums.Further, unlike Baker and Wurgler (2004), they find that the stock market reaction to dividend changesdepends on the dividend premiums. Li and Lie (2006) also extended the catering theory to includerepurchases by evaluating whether dividend premiums influence the probability of repurchases. Theyargue that firms intending to payout cash flows are likely to adopt repurchases rather than dividendswhen dividend premiums are low because dividends are both less desirable and are effectively sub-stitutes for repurchases (Grullon & Michaely, 2002). Consistent with their conjecture, they find thatthe probability of repurchases declines with the dividend premium.

On the basis of the above discussion, we argue that in the context of the IPO market, the choiceof form of payout initiation mechanism can be explained by the catering theory. In periods where


investor appetite for dividends is strong, as indicated by high dividend premiums, managers of IPOfirms are likely to cater to investor demand and initiate payouts via dividends. On the other hand, asargued by Li and Lie (2006), during periods of low dividend premiums, IPO firms seeking to payoutcash flows are likely to select repurchases as the payout initiation mechanism. Therefore, we wouldexpect a negative relationship between dividend premiums and the probability of initiating payoutsvia repurchases instead of dividends.

2.6. Life cycle factors and financing constraints

The corporate payout policy literature has suggested that factors such as profitability, maturity,growth, leverage, investment opportunities, risk, and size influence the ability of firms to pay dividends(DeAngelo et al., 2006; Fama & French, 2001; Grullon & Michaely, 2008; Kahle, 2002; Kale et al., 2006).In a nutshell, the extant literature indicates that profitability, maturity, and size are positively relatedto the likelihood of dividends while leverage, and growth opportunities are negatively related to theprobability of dividends. Additionally, the relationship between the above variables with dividendshas been found to hold for dividend changes as well as dividend initiations. Consistent with thesestudies, we include life cycle factors such as profitability, maturity, size, growth opportunities, andleverage as control variables in our analysis of factors that influence the choice of form of post-IPOpayout initiation as well as the timing of dividend initiations and share repurchases.

In addition to life cycle factors, we also control for the impact of extent of financing constraintsfaced by IPO firms on their choice of form of payout initiation mechanism as well as on the timing ofpost-IPO dividend and repurchase initiations. IPO firms facing lower levels of financing constraints arelikely to be in a better position to initiate post-IPO payouts via dividends and thereby implicitly com-mit to maintaining a regular stream of payouts in the foreseeable future. On the other hand, IPO firmsfacing higher financing constraints are likely to prefer announcing payout initiations via repurchases,thereby retaining the flexibility to exercise the option to discontinue or alter the level and timing offuture payouts if required. In their study evaluating whether investment cash flow sensitivities provideuseful measures of financing constraints, Kaplan and Zingales (1997) adopt variables such as leverage(debt over assets), interest coverage, and cash flows deflated by investments as quantitative measuresof financing constraints. Consistent with their approach, in addition to leverage as described above,we include interest coverage and cash flows deflated by assets as proxies for the extent of financingconstraints facing the IPO firms and include them as control variables in our analyses. Finally, we alsoinclude investment bank prestige as a control variable as extant research on dividend initiations byIPO firms has argued that participation by reputable investment bankers reduces information asym-metry and consequently the need to signal with payouts, thereby suggestive of a negative relationshipbetween investment bank prestige and the probability of payout initiations (Kale et al., 2006).

3. Sample description, variable definitions and summary statistics

3.1. Sample description

Our initial sample consists of all IPOs issued during the period 1990–2000 identified from Secu-rities Data Corporation’s (SDC) New Issues database. Consistent with most IPO studies, we excludefinancial firms, utilities, Real Estate Investment Trusts (REITs), equity carve-outs, and unit offeringssince their institutional structure and payout policy decisions are likely to be fundamentally differentfrom that of a typical IPO. Additionally, we exclude Regulation A securities which represent smallerofferings raising less than $5 million at the IPO and are subject to less stringent disclosure require-ments. Next, we require that IPO firms be listed on CRSP and Compustat immediately after the IPO. Inorder to confine our dividend group to firms that initiated ordinary regular cash dividends, we retainonly dividend firms with CRSP distribution codes 1212, 1232, 1242 and 1252 that are also listed asdividend firms on Compustat. All the above restrictions result in a final sample of 1886 IPO firms. Eachfirm is tracked on CRSP from the IPO date until the end of year 2005 or the delisting date (whicheveris earlier) to determine whether the firm initiated an ordinary regular cash dividend or repurchasedshares. Consistent with Kahle (2002), we identify repurchases from the SDC Mergers and Acquisition


database. Information on the announcement date and transaction details of the initial and any subse-quent repurchase by each firm in the sample was also obtained from the SDC Mergers and Acquisitiondatabase. Our sample of firms was initially classified into four groups: (a) Non-Initiator group con-sisting of 1445 firms that did not initiate any form of post-IPO payouts, (b) dividend initiator groupconsisting of 75 IPO firms that initiated dividends but no repurchases, (c) repurchase group consistingof 307 IPO firms that repurchased shares but did not pay dividends, and (d) 59 IPO firms that paidboth dividends as well as repurchased shares during the post-IPO phase. Since the focus of our studyis on the choice of form of payout initiation mechanism, the sub-sample of 59 firms that paid bothdividends and repurchased shares were assigned to either the dividend initiation group or repurchasegroup depending on whether their first post-IPO payout occurred via dividends or repurchases. There-fore, for the purpose of our analyses, our overall sample of 1886 IPO firms was segmented into threegroups consisting of 100 dividend firms, 341 repurchase firms and 1445 non-payout firms.

3.2. Variable definitions

This section describes the main variables used in this study. The variables defined below are eitherobtained from CRSP, COMPUSTAT, SDC New Issues Database, SDC Mergers and Acquisition Databaseor a combination of these sources. Variables associated with IPO firm and offering characteristicssuch as proceeds raised at the IPO, number of uses of IPO proceeds, specific use of IPO proceeds,identity of lead underwriter, venture capitalist involvement, and post-IPO ownership retention byinsiders are obtained from the SDC New Issues database. The variable VC is a dichotomous variablethat takes on the value 1 if the IPO firm is VC backed and 0 otherwise. The various aspects of post-IPOfinancing strategy are proxied by the variables NUMUSES, ACQ, REF, and GEN respectively. The variableNUMUSES represents the number of planned uses for the IPO proceeds listed by the firm in its IPOoffering prospectus. ACQ is a dichotomous variable that takes on the value 1 if the IPO firm indicatesits primary intended use of the IPO proceeds is for acquisition financing and 0 otherwise. Similarly, thevariable REF (GEN) is a dichotomous variable that take on the value 1 if the IPO firm indicates its primaryintended use of the IPO proceeds is for refinancing/repayment of debt (general corporate purposes)and 0 otherwise. The age of the firm is measured as the number of years between incorporation of thefirm and its IPO. The variable MATURITY is defined as log(1 + age of the firm) and proxies for the stageof development of the firm. The variable RANK represents the prestige of the lead investment bankerand is measured by the widely used Carter, Dark and Singh (1998) nine-point underwriter reputationscale. Consistent with the extant IPO literature, we measure the size of the IPO firm as the log of theproceeds raised at the IPO (LAMT). The demand for dividends is proxied by the variable BWDPREMIUMwhich represents the Baker and Wurgler (2004) dividend premium measure and is computed as thedifference between the logarithms of the market to book ratios of dividend payers and non-payers.

To identify IPO firms in high technology industries, we adopt the classification scheme developedby Loughran and Ritter (2004) who define technology stocks as those belonging to the following indus-try sectors: computer hardware (SIC codes 3572, 3575, 3577, 3578), communication equipment (SICcodes 3661, 3663, 3669), electronics (SIC codes 3671, 3672, 3674, 3675, 3677, 3678, 3679), navigationequipment (SIC code 3812), measuring and controlling devices (SIC codes 3823, 3825, 3826, 3827,3829), medical instruments (SIC codes 3841, 3845), telephone equipment (SIC codes 4812, 4813),communication services (SIC code 4899) and software (SIC codes 7371, 7372, 7373, 7374, 7375, 7378,7379). The variable TECH takes on the value 1 if the IPO firm belongs to a high technology industry asidentified above and 0 otherwise.

Financial information regarding each IPO firm is obtained from COMPUSTAT and is measured bothin the fiscal year of the IPO as well as the fiscal year of payout initiation. The financial informationobtained from COMPUSTAT is used to construct the following variables:

1. OPINCOME. Defined as earnings before interest taxes, depreciation, and amortization (Compustatdata item 13) divided by book value of total assets (Compustat data item 6).

2. LEVERAGE. Defined as long-term debt (Compustat data item 9) divided by total assets (Compustatdata item 6).


3. CAPEX. Defined as capital expenditure intensity (Compustat data item 128) divided by total assets(Compustat data item 6).

4. RDA. Defined as R&D intensity (Compustat data item 46) divided by total assets (Compustat dataitem 6).

5. GROWTH. Measures the market to book ratio and is defined as {book value of assets (Compustatdata item 6) − book value of equity (Compustat data item 216) + market value of equity (Compustatdata item 24 × data item 25)} divided by book value of assets (Compustat data item 6).

6. CASHFLOW. Defined as cash flow divided by total assets {(earnings before extraordinary items(Compustat data item 18) + depreciation (Compustat data item 14)) divided by book value of assets(Compustat data item 6)}.

7. TIE. Defined as interest coverage {earnings before interest, taxes, depreciation and amortization(Compustat data item 13) divided by interest expense (Compustat data item 15)}.

3.3. Summary statistics

Table 1 provides a yearly distribution of IPO firms for the overall sample as well as for each of thethree subgroups consisting of dividend, repurchase, and non-payout firms respectively. In addition,the dividend and repurchase sub-samples are further segmented on the basis of the time intervalbetween the IPO and dividend or repurchase initiation. Therefore, dividend and repurchase firms aresegmented into four subgroups depending on whether the IPO firm initiated dividends (repurchases):(a) within 1 year of IPO, (b) between 1 and 3 years after the IPO, (c) between 3 and 5 years after theIPO, or (d) after more than 5 years from the IPO.

The results in Table 1 indicate a greater propensity of IPO firms to initiate post-IPO payouts throughrepurchases rather than dividends. For instance, approximately 18% of IPO firms initiate post-IPOpayouts by repurchasing shares compared to 5.30% of firms initiating dividends. The results in Table 1also indicate differences in the timing of initiation of dividends versus repurchases by IPO firms.Dividend firms tend to initiate payouts either immediately in the first year after the IPO or after atleast 5 years of going public. On the other hand, the timing of repurchase initiations by IPO firms isspread out more evenly over time. For instance, the results indicate that 49% of dividend firms initiatedividends within 1 year of going public while only 16% of repurchase firms initiate repurchases duringthe first year after the IPO. Similarly, while 33% of dividend firms wait at least 5 years after the IPOto initiate dividends, only 23% of repurchase firms initiate repurchases after 5 years of going public.Further, while only 18% of dividend firms initiate payouts between 1 and 5 years of going public, 61%of repurchase firms initiate repurchases within 1–5 years of the IPO.

In results not reported here but available with the authors we also evaluate the distribution ofdividend, repurchase, and non-payout firms segmented by industry. The industry distribution of pro-portion of dividend initiating IPO firms varies from a low of 2.3% for Transportation and Shipping to ahigh of 27.8% for Metal Products. Similarly, the proportion of repurchase firms by industry varies froma low of 0% for Food and Kindred Products to a high of 29% for Transportation and Shipping. Further,the proportion of firms preferring initiation through repurchases is lower than the proportion of firmspreferring dividends in only 3 of the 16 industry groups represented in our sample of IPO firms.

Overall, the summary statistics indicate differences in the timing pattern of dividend versus repur-chase initiation by IPO firms. Further, the results indicate a propensity of IPO firms to prefer initiatingpayouts via repurchases rather than dividends across most industry groups. As such, our results aresuggestive of IPO firms demonstrating a preference for share repurchases over dividends as the payoutinitiation mechanism.

4. Results

4.1. Valuation effects of post-IPO payout initiation announcements

The market reaction to the announcement of post-IPO payout initiations through dividends as wellas repurchases is assessed using standard event study methodology. The cumulative abnormal returns(henceforth, CARs) are estimated over several time windows using both the market model as well as

B.A.Jain

etal./The

Quarterly

Review

ofEconomics

andFinance

49(2009)

1275–12971285

Table 1Yearly distribution of payout and non-payout IPO firms.

Year of IPO All firms Non-initiating IPO firms Dividend IPO firms Repurchase IPO firms

Time to dividend initiation (years) 100 firms Time to repurchase initiation (years) 341 firms

t ≤ 1 1 < t ≤ 3 3 < ≤ 5 5 < t t ≤ 1 1 < t ≤ 3 3 < t ≤ 5 5 < t

1990 65 40 (61.5%) 3 (33%) 3 (33%) 2 (22%) 1 (11%) 3 (19%) 3 (19%) 2 (13%) 8 (50%)1991 152 97 (63.8%) 9 (50%) 0 (0%) 2 (11%) 7 (39%) 9 (24%) 6 (16%) 8 (22%) 14 (38%)1992 212 139 (65.6%) 9 (47%) 1 (5%) 3 (16%) 6 (32%) 7 (13%) 14 (26%) 15 (28%) 18 (33%)1993 242 178 (73.6%) 5 (45%) 1 (9%) 0 (0%) 5 (45%) 7 (13%) 19 (36%) 12 (23%) 15 (28%)1994 204 149 (73.0%) 7 (54%) 1 (8%) 0 (0%) 5 (38%) 6 (14%) 15 (36%) 9 (21%) 12 (29%)1995 184 143 (77.7%) 8 (80%) 0 (0%) 1 (10%) 1 (10%) 2 (6%) 10 (32%) 14 (45%) 5 (16%)1996 269 215 (79.9%) 3 (50%) 1 (17%) 0 (0%) 2 (33%) 7 (15%) 26 (54%) 13 (27%) 2 (4%)1997 164 137 (83.5%) 2 (33%) 0 (0%) 0 (0%) 4 (67%) 5 (24%) 13 (62%) 1 (5%) 2 (10%)1998 109 98 (89.9%) 0 (0%) 1 (50%) 0 (0%) 1 (50%) 4 (44%) 2 (22%) 0 (0%) 3 (33%)1999 170 149 (87.7%) 3 (60%) 1 (20%) 0 (0%) 1 (20%) 3 (19%) 11 (69%) 2 (13%) 0 (0%)2000 115 100 (87.0%) 0 (0%) 0 (0%) 1 (100%) 0 (0%) 2 (14%) 7 (50%) 5 (36%) 0 (0%)1990–2000 1886 (100%) 1445 (76.6%) 49 (49%) 9 (9%) 9 (9%) 33 (33%) 55 (16%) 126 (37%) 81 (24%) 79 (23%)

The sample consists of 1886 IPOs issued during the period 1990–2000. The IPO sample is segmented into three groups on the basis of their payout initiation decision as follows: (a)dividend firms, (b) regular repurchase firms and (c) non-payout firms. The distribution of these three types of payout firms segmented by the year of IPO issuance is reported below. Inaddition, the number (%) of firms that initiated dividends or repurchases is reported on the basis of the time interval between the IPO and the dividend (repurchase) initiation decisionfor the overall sample as well as segmented on the basis of IPO issuance year.


Table 2Valuation effects of payout initiation announcements by IPO firms.

Estimation model Event window abnormal returns (Patell’s Z statistic)

−1 to +1 0 to +1 −3 to +3

Panel A: repurchase initiation announcementsMarket model equally weighted index 1.72% (5.20)a 1.91% (7.83)a 2.56% (3.43)a

Market model value weighted index 1.35% (4.43)a 1.57% (6.92)a 1.76% (2.21)b

Market adjusted returns equally weighted 1.25% (3.29)a 1.63% (5.64)a 1.61% (2.35)b

Market adjusted returns value weighted 1.24% (2.99)a 1.58% (5.13)a 1.68% (2.38)a

Panel B: dividend initiation announcementsMarket model equally weighted index 1.55% (3.44)a 1.42% (3.83)a 1.85% (2.35)b

Market model value weighted index 1.53% (3.17)a 1.46% (3.64)a 1.72% (2.00)b

Market adjusted returns equally weighted 1.62% (3.21)a 1.39% (3.42)a 2.04% (2.27)b

Market adjusted returns value weighted 1.89% (3.37)a 1.57% (3.69)a 2.34% (2.52)a

The market reaction to the announcement of the first post-IPO payout initiation via stock repurchases and dividends arereported below. In panel A, the valuation effects of share repurchase initiation announcements are reported while in panelB, the valuation effects of dividend initiation announcements are reported. Standard event study methodology is applied toestimate the cumulative abnormal returns (CAR) for several event windows. The CARs are reported for three event windows,−1 to +1, 0 to +1, and −3 to +3. Results are reported both for the market model and market adjusted returns model (both equallyweighted as well as value weighted). The test of whether the abnormal returns are significantly different from zero is based onPatell’s Z statistic for a two-tailed test, which is reported in parenthesis. Significant at the 0.10 level.

a Significant at the 0.01 level.b Significant at the 0.05 level.

market adjusted return methods to compute value weighted as well as equally weighted returns. Theresults of the event study are reported in Table 2 where CARs are reported for event time windows −1to +1, 0 to +1, and −3 to +3 where day 0 represents the payout initiation announcement date. PanelA of Table 2 reports the valuation effects of post-IPO payout initiation via repurchases while panel Breports the valuation effects of payout initiation via dividends.

As can be seen from panel A of Table 2, IPO firms experience a significant positive market reactionto their first post-IPO repurchase announcement. For instance, depending on the benchmark modeland weighting approach, the 2-day (0 to +1) CAR for the first repurchase announcement by IPO firms’ranges from a low of 1.58% to a high of 1.91%. Similarly, for the time window −3 to +3, the CAR forrepurchase announcements ranges from a low of 1.61% to a high of 2.56%. From panel B of Table 2,it can be seen that the market reaction to post-IPO payout initiations via dividends is similar bothin direction as well as magnitude to the market reaction to payout initiations via repurchases. Forinstance, the 2-day (0 to +1) market reaction to post-IPO dividend initiations ranges from a low of1.39% to a high of 1.57% which is essentially similar in magnitude to the valuation effects over thesame time window for repurchase firms as reported in panel A. Additionally, the valuation effects ofpost-IPO dividend initiation reported in panel B of Table 2 are similar in direction and magnitude tothose documented by Lipson et al. (1998). As such, the results in Table 2 indicate that while the marketreacts positively to both forms of post-IPO payout initiations, it does not appear to value one form ofpayout over the other.

In order to gain additional insights into factors that influence the market reaction to post-IPOpayout initiations, we conduct a cross-sectional regression analysis. The dependent variable is theCAR measured over the time window −3 to +3 around the announcement date. We therefore estimateregression models of the following form:

CARit = ˇ0 + ˇ1FORM + ˇ2VC + ˇ3NUMUSES + ˇ4ACQ + ˇ5REF + ˇ6GEN + ˇ7RDA + ˇ8CAPEX

+ ˇ9OPINCOME + ˇ10TECH + ˇ11BWDPREMIUM + ˇ12CASHFLOW + ˇ13LEVERAGE

+ ˇ14TIE + ˇ15TIME (1)

where CARit is the 7-day cumulative abnormal returns around the announcement of post-IPO pay-out initiations by IPO firms. The variable FORM is a dichotomous variable that takes on the value 1if the IPO firm initiated its first post-IPO payout via dividends and 0 if initiation occurred through


share repurchases. A significant positive or negative coefficient for the variable FORM would provideevidence to suggest that the market reaction to payout initiation depends on the form of the payoutafter controlling for other factors that potentially influence the market reaction to payout initiations.The variables VC, TECH, NUMUSES, ACQ, REF, and GEN proxy for VC participation, industry technologyfocus, and various aspects of financing strategy, respectively. The variable BWDPREMIUM is the Bakerand Wurgler (2004) dividend premium variable measured in the fiscal year prior to payout initiationand is included on the basis of catering theory, which predicts a positive relationship between dividendpremiums and the market reaction to payout announcements. The extant evidence on the impact ofdividend premiums on the market reaction to payout announcements, however, is mixed with Bakerand Wurgler (2004) finding no effect while Li and Lie (2006) reporting a significant positive effect.The variables CAPEX and RDA represent capital intensity and R&D intensity of the IPO firm at the timeof payout initiation and proxy for the extent of product market competition facing the firm. All theabove-described independent variables are as defined earlier in the variables section and allow usto examine the impact of VC participation, industry technological status, financing strategy, investordemand for dividends, and extent of product market competition on the valuation effects of payoutinitiations by IPO firms. Additionally, we include the variables OPINCOME, LEVERAGE, CASHFLOW, andTIE measured in the fiscal year of payout initiation to control for life cycle and financing constraint fac-tors that might potentially influence the market reaction to the announcement of payout initiations.Finally, the variable TIME is included as a control variable and is measured as the time in monthsbetween the IPO date and the announcement of the first post-IPO payout initiation. Including thisvariable allows us to assess whether the market differentiates between early post-IPO payouts versusthose made several years after going public. In estimating the cross-sectional regression models, weconduct the usual tests to ensure that none of the key assumptions of OLS such as heteroskedasticityand multicollinearity are violated.

The results of the determinants of the valuation effects of announcement of payout initiations byIPO firms are reported in Table 3. In model 1, the independent variables include FORM, TECH, VC,financing strategy variables and the five control variables (OPINCOME, LEVERAGE, CASHFLOW, TIE,and TIME). In model 2, the Baker and Wurgler (2004) dividend premium measure BWDPREMIUM isincluded as an additional independent variable. Finally, in model 3, the product market variables RDAand CAPEX are included in addition to all the other independent variables included in model 2.

The results in Table 3 indicate that the coefficient of FORM in all three models is positive but notstatistically significant. As such, our results suggest that while the market reacts positively to payoutinitiations by IPO firms, it does not appear to demonstrate a preference regarding the specific formof post-IPO payout initiation. Further, while the coefficient of VC is positive in all three regressions,it is not consistently significant indicating weak evidence of the market reacting positively to payoutinitiations by VC backed firms. Industry technological status does not appear to have any effect on themarket reaction to a post-IPO payout initiation announcement. However, the results in Table 3 consis-tently indicate a negative and significant coefficient for the financing strategy variable ACQ. As such,we find that the market reacts negatively to post-IPO payout initiations by acquisition oriented IPOfirms. Our results are suggestive of the market viewing the payout initiation announcement by acquisi-tion oriented IPO firms as a signal of unavailability of attractive acquisition opportunities and therebyreacting negatively. Additionally, the results in Table 3 indicate that the coefficient of RDA is positiveand significant indicating a positive market reaction to payout initiation by IPO firms facing strongproduct market competition. The positive reaction to payout initiation by IPO firms facing higher prod-uct market competition is consistent with the Grullon and Michaely (2008) outcome payout model andsuggestive of investors viewing the payout decision as a signal of managerial discipline and efficiencyby avoiding overinvestment or pursuing risky projects. Finally, the results in Table 3 also indicate thatamong the control variables, the coefficient of CASHFLOW is consistently negative and significant sug-gesting that the market reacts negatively to payout initiation by IPO firms facing financing constraints.

4.2. The choice between post-IPO payout initiation via dividends versus repurchases

In this section, we attempt to identify the differences in characteristics of IPO firms that initiatepayouts via dividends versus share repurchases. Additionally, we attempt to estimate the factors that


Table 3Multivariate analysis estimating determinants of valuation effects of payout initiation announcements by IPO firms.

Variables Model 1 Model 2 Model 3

Intercept 0.0779 (1.51) 0.0817 (1.50) 0.0696 (1.27)FORM 0.0144 (0.60) 0.0159 (0.63) 0.0158 (0.61)TECH −0.0113 (−0.55) −0.0107 (−0.51) −0.0201 (−0.92)VC 0.0326 (1.68)c 0.0318 (1.61) 0.0233 (1.14)NUMUSES 0.0004 (0.04) 0.0001 (0.01) 0.0001 (0.01)ACQ −0.1298 (−2.18)b −0.1289 (−2.16)b −0.1381 (−2.24)b

REF −0.0669 (−1.64)c −0.0665 (−1.63) −0.0662 (−1.61)GEN −0.0410 (−1.03) −0.0411 (−1.03) −0.0527 (−1.30)OPINCOME 0.1578 (1.24) 0.1564 (1.22) 0.1677 (1.13)LEVERAGE −0.0155 (−0.29) −0.0153 (−0.29) −0.0185 (−0.33)CASHFLOW −0.2703 (−2.62)a −0.2711 (−2.62)a −0.2640 (−2.54)a

TIE 0.0000 (0.08) 0.0000 (0.11) 0.0000 (0.07)TIME −0.0002 (−0.84) −0.0002 (−0.87) −0.0001 (−0.59)BWDPREMIUM 0.0003 (0.23) 0.0002 (0.18)RDA 0.3075 (1.92)c

CAPEX 0.0159 (0.14)R2 0.1024 0.1026 0.1174

The dependent variable is the cumulative abnormal return (CAR) measured over a 7-day event window extending from day −3to day +3 relative to the announcement date of the first post-IPO dividend or repurchase announcement for the sub-sampleof 441 IPO firms that initiated post-IPO cash flow payouts. FORM is a dichotomous variable that takes on the value 1 if theIPO firm initiated its first post-IPO payout via dividends and 0 if the payout initiation occurred via a repurchase. TECH takeson the value 1 if the IPO issuer belongs to a high technology industry and 0 otherwise. VC takes on the value 1 if the issuingfirm is venture capital backed and 0 otherwise. NUMUSES represents the number of uses for the IPO proceeds identified bythe issuing firm in its offering prospectus. ACQ is a dummy variable that takes on the value 1 if the primary use of the IPOproceeds is for financing acquisitions and 0 otherwise. REF takes on the value 1 if the primary intended use of the IPO proceedsis for retiring/refinancing debt and 0 otherwise. GEN takes in the value 1 if the primary intended use of the IPO proceeds is forgeneral corporate purposes and 0 otherwise. OPINCOME represents operating income and is measured by the ratio of earningsbefore interest, taxes, depreciation and amortization (EBITDA, Compustat item #13) divided by total assets (Compustat item#6) measured in the fiscal year of payout initiation. LEVERAGE represents firm leverage at the time of the IPO and is measuredby the long-term debt (Compustat item #9) divided by total assets (Compustat item #6). The variable CASHFLOW representscash flow deflated by total assets and is measured as the sum of the earnings before extraordinary items (Compustat item#18) and depreciation (Compustat item #14) divided by total assets (Compustat item #6) measured in the fiscal year of payoutinitiation. TIE is the interest coverage and is measured as EBITDA (Compustat item #13) divided by interest expense (Compustatitem #15) measured in the fiscal year of payout initiation. TIME represents the time period in months between the IPO andthe announcement of the first post-IPO dividend or share repurchase decision. BWDPREMIUM is the value weighted Baker andWurgler (2004) dividend premium measure, which is computed as the difference between the logarithms of the market to bookratios of dividend payers and non-payers and is computed in the year prior to payout initiation. RDA is the R&D intensity and ismeasured as the ratio of R&D expenditures (Compustat item #46) divided by total assets (Compustat item #6) measured in thefiscal year of the payout initiation announcement. CAPEX is the capital intensity defined as capital expenditures (Compustatitem #128) divided by the total assets (Compustat item #6) in the fiscal year of the payout initiation announcement. Thet-statistics are reported in parentheses and significance levels are indicated.

a Significant at the 0.01 level.b Significant at the 0.05 level.c Significant at the 0.10 level.

influence the choice of form of payout initiation by IPO firms. We begin by reporting a univariatecomparison of the characteristics of dividend, repurchase, and non-payout firms at the time of theIPO. In addition, significance tests of differences in the mean (median) values of various variables arereported for (a) dividend versus non-payout firms, (b) repurchase versus non-payout firms, and (c)dividend versus repurchase firms.

The results of the comparison of the characteristics of dividend versus non-payout firms in Table 4indicate several significant differences between these two groups of firms. For instance, the mean(median) values of leverage, operating income, interest coverage, cash flow over assets, age of firm,sales, and total assets at the time of going public are significantly higher for dividend firms relative tonon-payout firms. On the other hand, the mean (median) values of initial returns, risk, market to bookratio, and R&D intensity are significantly lower for dividend payout firms relative to non-payout firms.

B.A.Jain

etal./The

Quarterly

Review

ofEconomics

andFinance

49(2009)

1275–12971289

Table 4Descriptive statistics of payout and non-payout firms.

Description of variables Dividend IPO firmsmean (median)

Repurchase IPOfirms mean(median)

Non-initiating IPOfirms mean(median)

Difference betweendividend versusnon-initiating firms t(z)

Difference betweenrepurchase versusnon-initiating firm t(z)

Difference betweendividend versusrepurchase firms t(z)

IPO proceeds ($M) 58.8 (35.1) 38.1 (27.6) 37.4 (27.6) 2.82a (1.29) 0.28 (0.33) 2.63a (1.45)Initial returns (%) 9.3 (4.1) 17.6 (10.0) 21.9 (9.1) −6.57a (−2.59a) −1.69c (0.81) −3.85a (−2.82a)Inside ownership retention (%) 42.2 (31.7) 34.3 (30.2) 36.3 (30.1) 0.68 (0.25) −0.44 (0.15) 0.92 (0.30)# Uses of proceeds 1.5 (1.0) 1.6 (1.0) 1.8 (1.0) −3.38a (−0.26) −1.85c (−0.82) −1.47 (0.19)Underwriter prestige 7.3 (8.1) 7.1 (8.1) 6.7 (8.1) 2.79b (0.93) 2.95a (0.53) 0.75 (0.65)Risk (%) 3.3 (3.1) 7.6 (3.7) 7.3 (4.1) −2.81b (−5.99a) 0.08 (−4.75)a −1.23 (−2.93a)Leverage (%) 15.7 (8.3) 9.5 (1.4) 8.4 (1.0) 4.02a (4.35a) 1.17 (0.66) 3.11a (3.43a)Interest coverage 10.49 (8.12) 20.99 (9.65) −8.70 (2.38) 2.16b (6.18)a 2.73a (7.83)a −0.85 (−0.94)Operating income (%) 16.0 (16.7) 10.6 (13.3) −3.7 (4.3) 15.94a (8.08a) 9.48a (9.67)a 4.11a (2.42b)Market to book ratio 2.1 (1.7) 3.0 (2.5) 3.2 (2.5) −7.03a (−4.55a) −1.23 (0.00) −5.03a (−5.00a)Cash flow/TA (%) 10.67 (10.64) 5.65(9.71) −7.46 (1.84) 6.73a (8.35)a 8.70a (8.45)a 3.17a (0.71)Firm age (years) 2.8 (2.9) 2.2 (2.2) 2.0 (1.9) 6.76a (4.97a) 3.93a (5.18)a 4.82a (3.86a)High technology industry 0.3 (0.0) 0.5 (0.0) 0.5 (1.0) −4.39a (−4.17a) 2.54a (2.53)a −2.56b (−2.46b)VC 0.3 (0.0) 0.5 (1.0) 0.5 (1.0) −6.21a (−5.46a) −0.83 (−0.82) −5.02a (−4.54a)Sales ($M) 264.9 (102.5) 140.7 (44.6) 68.1 (27.9) 4.10b (6.63a) 3.98a (4.67)a 2.15b (4.65a)Total assets ($M) 256.8 (103.2) 124.4 (46.9) 84.0 (41.4) 3.00b (4.76a) 2.33b (2.46)a 2.10b (3.66a)Capital intensity (%) 7.5 (4.8) 7.7 (4.4) 7.7 (4.2) −0.33 (1.04) −0.10 (0.55) −0.22 (1.17)R&D intensity (%) 1.9 (0.0) 5.6 (1.1) 8.1 (2.6) −10.88a (−5.37a) −3.22a (−1.26) −5.91a (−2.70a)

Descriptive statistics for sub-samples of IPO firms segmented on the basis of post-IPO payout status. The IPO proceeds are the gross amounts raised at the IPO in millions of dollars. Theinitial returns are measured as the difference of the first day market price minus the offer price divided by the offer price expressed as a percentage. The inside ownership retention is thepercentage ownership retained in the post-IPO firm by insiders and directors. The number of uses of the proceeds represents all the planned uses of the IPO proceeds listed by the issuingfirms in its offering prospectus. Underwriter prestige is a measure of investment-banking reputation on the basis of a nine-point ranking scale developed by Carter et al. (1998). The riskis measured as the standard deviation of the first 30 days of aftermarket returns of the IPO firms. Leverage is measured as the long-term debt (Compustat item #9) divided by total assets(Compustat item #6) measured in the fiscal year of the IPO. Interest coverage is measured as EBITDA (Compustat item #13) divided by interest expense (Compustat item #15) measuredin the fiscal year of the IPO. Operating income is measured by the earnings before interest, taxes, depreciation and amortization (EBITDA, Compustat item #13) divided by total assets(Compustat item #6) measured in fiscal year of the IPO. Growth opportunities are measured by the market to book ratio [(book value of assets − book value of equity + market value ofequity)/book value of assets where market value of equity = Compustat item #24 × item #25 and book value of equity = Compustat item #216]. Cash flow over total assets is measured asthe sum of the earnings before extraordinary items (Compustat item #18) and depreciation (Compustat item #14) divided by total assets (Compustat item #6) measured in the fiscal yearof the IPO. Firm age is the time between incorporation and the IPO date measured in number of years. High technology is a dummy variable that takes on the value 1 if the firm belongsto a high technology industry and 0 otherwise. VC is a dummy variable that takes on the value 1 if the IPO firm received venture capital backing and 0 otherwise. Sales (Compustat item#12) and total assets (Compustat item #6) represent the sales and total assets respectively in millions of dollars measured in the fiscal year of the IPO. Capital intensity is measured asthe capital expenditures (Compustat item #128) divided by total assets (Compustat item #6) in the fiscal year of the IPO. R&D intensity is measured as the ratio of R&D expenditures(Compustat item #46) divided by total assets (Compustat item #6). Results of significance tests for differences in the above-described variables are reported for the following subgroupsof firms; (a) dividend versus non-initiating firms, (b) repurchase versus non-initiating firms, and (c) dividend versus repurchase IPO firms.



In addition, dividend firms are less likely to belong to high technology industries or receive venturecapital financing compared to non-payout firms.

A similar comparison of repurchase versus non-payout IPO firms however, indicates that unlikedividend firms, there are fewer significant differences in the characteristics of repurchase firms versusnon-payout firms. For instance, similar to dividend firms, repurchase firms are older, and have signif-icantly higher mean (median) operating income, interest coverage, cash flow over assets, sales, andtotal assets compared to non-payout firms. However, unlike dividend firms, repurchase firms have agreater propensity to belong to high technology industries relative to non-payout firms and are notsignificantly different from non-payout firms on aspects such as initial returns, risk, market to bookratio, leverage, and ability to attract VC participation.

A direct comparison of dividend versus repurchase firms is provided in the last column of Table 4.The results indicate several significant differences between these two types of payout firms. Forinstance, dividend firms are characterized by significantly lower values of initial returns, venturecapital participation, membership in high technology industries, and R&D intensity and significantlyhigher values of sales and total assets relative to repurchase firms. In terms of life cycle based factors,dividend firms have significantly higher leverage, operating income and firm age, but significantlylower market to book ratios compared to repurchase IPO firms. Overall, the comparisons in Table 4provide evidence to indicate that dividend IPO firms are fundamentally different from repurchase IPOfirms and are at a more mature life cycle stage at the time of going public relative to repurchase IPOfirms.

To identify factors that influence the choice of payout initiation via dividends versus repurchases,we conduct a logistic regression analysis. Since our objective is to identify the determinants of thechoice of specific form of payout initiation, our analysis is confined to the sub-sample of IPO firms thatinitiated post-IPO payouts either through dividends or repurchases. The dependent variable equals 1if the IPO firm initiated dividends and 0 if it initiated repurchases. The basic form of the logit model isas follows:

Pi(dividend = 1|Xi) = exp(Xiˇ)1 + exp(Xiˇ)

(2)

where ˇ is the vector of unknown parameters to be estimated and Xi is a vector of firm specific variablessuch as TECH, VC, NUMUSES, ACQ, REF, GEN, CAPEX, RDA, BWDPREMIUM, LEVERAGE, OPINCOME,MATURITY, GROWTH, TIE, CASHFLOW, LAMT, and RANK, and are as previously defined in Section 3.The variable VC is included to evaluate the impact of VC participation on the choice of form of post-IPO payout mechanism while the variables NUMUSES, ACQ, REF, and GEN are included to evaluate theeffect of post-IPO financial strategy on the choice of form of payout. The variables RDA and CAPEXare included to evaluate the impact of extent of product market competition while the variable TECHis included to evaluate the impact of industry technological status on choice of form of post-IPOpayout. The variable BWDPREMIUM is included to evaluate whether the catering theory can providean explanation for the choice of dividends versus repurchases as the payout initiation mechanismadopted by IPO firms and is measured in the fiscal year prior to initiation. The life cycle and financingconstraint variables LEVERAGE, OPINCOME, MATURITY, GROWTH, TIE and CASHFLOW are measuredin the fiscal year of payout initiation and are included as control variables. Additionally, consistent withthe IPO dividend payout literature, additional control variables such as LAMT and RANK are includedto control for the effect of size of the IPO firm and investment bank prestige on the choice of form ofpayout initiation mechanism. The results of the logistic regression analysis are reported in Table 5.

In the first model, the independent variables include the two product market competition variables(CAPEX and RDA) and the Baker and Wurgler (2004) dividend premium variable (BWDPREMIUM)along with the eight control variables (LEVERAGE, OPINCOME, MATURITY, GROWTH, TIE, CASHFLOW,LAMT, and RANK) and allows us to evaluate the impact of extent of product market competition andinvestor demand for dividends on the choice of form of payout initiation mechanism. In model 2, inaddition to the independent variables included in model 1, variables proxying for various aspects ofpost-IPO financial strategy (NUMUSES, ACQ, REF, GEN) are included which allows us to additionallyevaluate the impact of financial strategy on the choice of initiating payouts via dividends versus sharerepurchases. In model 3, the variables VC and TECH are included in addition to all the independent


Table 5Logit models of choice of payout initiation via dividends versus repurchases by IPO firms.

Variables Model 1 Model 2 Model 3

Intercept −2.4911 (0.0072)b −1.6153 (0.2200) −1.6027 (0.2367)TECH 0.3572 (0.3696)VC −1.1286 (0.0073)a

NUMUSES −0.6010 (0.0285)b −0.4804 (0.0791)c

ACQ −0.1292 (0.9003) −0.0230 (0.9821)REF 0.2590 (0.7140) 0.3672 (0.6100)GEN 0.0058 (0.9936) 0.0033 (0.9964)CAPEX −4.8109 (0.0411)b −6.1955 (0.0288)b −6.7620 (0.0203)b

RDA −11.9901 (0.0115)b −10.9053 (0.0292)b −8.6624 (0.0999)c

BWDPREMIUM 0.0426 (0.0308)b 0.0451 (0.0337)b 0.0544 (0.0127)b

LEVERAGE 1.2774 (0.1795) 2.0327 (0.0549)c 1.8248 (0.0885)OPINCOME 3.0170 (0.2546) 1.8242 (0.4985) 2.6607 (0.3554)MATURITY 0.3456 (0.0156)b 0.2969 (0.0550)c 0.2969 (0.0509)c

GROWTH −0.4485 (0.0363)b −0.5380 (0.0451)b −0.5131 (0.0685)c

TIE −0.0014 (0.2029) −0.0015 (0.1846) −0.0018 (0.1738)CASHFLOW 0.7921 (0.7584) 0.4494 (0.8433) −0.0586 (0.9805)LAMT 0.1260 (0.6016) 0.1079 (0.6800) 0.0397 (0.8825)RANK −0.0104 (0.9129) 0.0071 (0.9452) 0.0531 (0.6226)Likelihood ratio 51.8326 (0.0001)a 54.6152 (0.0001)a 63.3873 (0.0001)a

The results of the logistic regression analysis to identify the determinants of the choice between dividends versus repurchasesto initiate post-IPO payouts are reported below. The analysis is based on a sub-sample of 441 IPO firms that initiated payoutsvia either dividends or repurchases after going public. The dependent variable CHOICE takes on the value 1 if the IPO firminitiated payouts via dividends and 0 if firm initiated repurchases. TECH takes on the value 1 if the IPO issuer belongs to a hightechnology industry and 0 otherwise. VC takes on the value 1 if the issuing firm is venture capital backed and 0 otherwise.NUMUSES represents the number of uses for the IPO proceeds identified by the issuing firm in its offering prospectus. ACQ is adummy variable that takes on the value 1 if the primary use of the IPO proceeds is for acquisitions and 0 otherwise. REF takes onthe value 1 if the primary purpose of the IPO proceeds is for retiring/refinancing debt and 0 otherwise. GEN takes in the value 1if the primary purpose of the IPO proceeds is for general corporate purposes and 0 otherwise. CAPEX is the capital intensity inthe IPO year and is defined as capital expenditures (Compustat item #128) divided by the total assets (Compustat item #6) inthe fiscal year of payout initiation. RDA is the R&D intensity and is measured as the ratio of R&D expenditures (Compustat item#46) divided by total assets (Compustat item #6) measured in the fiscal year of payout initiation. BWDPREMIUM is the valueweighted Baker and Wurgler (2004) dividend premium measure, which is computed as the difference between the logarithmsof the market to book ratios of dividend payers and non-payers and is computed in the fiscal year prior to payout initiation.LEVERAGE is measured by the long-term debt (Compustat item #9) divided by total assets (Compustat item #6) measured inthe fiscal year of payout initiation. OPINCOME represents profitability and is measured by the ratio of earnings before interest,taxes, depreciation, and amortization (EBITDA, Compustat item #13) divided by total assets (Compustat item #6) measured inthe fiscal year of payout initiation. MATURITY is measured by the log of 1 + age of firm at the IPO. GROWTH is the market to bookratio in the fiscal year of payout initiation [(book value of assets − book value of equity + market value of equity)/book value ofassets where market value of equity = Compustat item #24 × item #25 and book value of equity = Compustat item #216]. TIEis the interest coverage and is measured as EBITDA (Compustat item #13) divided by Interest expense (Compustat item #15)measured in the fiscal year of payout initiation. CASHFLOW represents cash flow deflated by total assets and is measured as thesum of the earnings before extraordinary items (Compustat item #18) and depreciation (Compustat item #14) divided by totalassets (Compustat item #6) measured in the fiscal year of payout initiation. LAMT represents the log of IPO proceeds. RANK isinvestment-banking prestige measured using the Carter et al. (1998) nine-point reputation scale. The p-values are reported inparentheses and significance levels are as indicated.


variables used in model 2, thereby allowing us to evaluate the impact of VC participation and firmtechnological focus on the choice of form of payout initiation mechanism adopted by IPO firms.

The results in Table 5 indicate that the coefficient of both product market related variables, CAPEXand RDA are consistently negative and significant in all three models. Therefore, the results in Table 5indicate that the probability of payout initiation via dividends is lower for firms facing high productmarket competition. The results are supportive of our hypothesis that IPO firms facing strong productmarket competition prefer the flexibility of repurchases to the inherent rigidity of dividends as theircash flow payout mechanism. Further, consistent with the life cycle and signaling theory of dividends,the coefficient of NUMUSES is negative and significant in all three models indicating that the prob-


ability of payout initiation via repurchases is higher for firms with a greater number of uses for IPOproceeds. As argued earlier, such firms are likely to have lower earned to contributed capital and aremore likely to need to raise additional capital during the post-IPO phase and therefore, not yet ata stage of development where dividends become feasible. However, payouts via repurchases allowthem to signal their favorable future prospects and therefore help position the firm to raise equitycapital on favorable terms.

The results in Table 5 also indicate that the coefficient of BWDPREMIUM is positive and significantindicating that the probability of payout initiation is more likely via dividends rather than repurchaseswhen investor demand for dividends is strong and therefore supportive of the catering theory ofdividends. Further, the results indicate that the coefficient of VC is negative and significant therebysupporting our conjecture that VC backed firms are more likely to initiate post-IPO payouts throughrepurchases rather than dividends in an effort to signal the value of the firm and enhance valuationsin advance of equity sales that help facilitate their eventual exit. Finally, the results indicate thatthe coefficient of MATURITY is positive and significant while GROWTH is negative and significantsuggesting that life cycle factors influence the choice of form of cash flow payouts by IPO firms.

4.3. Timing of post-IPO payout initiation decision

Hazard analysis methodology is applied to estimate the factors influencing the timing of payoutinitiation decisions by IPO firms. We separately estimate Cox Proportional Hazard (henceforth, CPH)models for the timing of payout initiation through dividends as well as repurchases. In the CPH modelfor payout initiations via dividends, the event is defined as the first post-IPO dividend announcement.The non-event group therefore includes both non-payout firms and repurchases firms. In the post-IPOperiod, issuing firms will initiate dividends, leave the sample as a result of an acquisition or failure,or continue to operate until the end of the tracking period without initiating dividends. Censoredfirms are defined as IPO firms that do not initiate dividends until the end of the tracking period.The event time is measured as the number of months from the IPO date until the dividend initiationannouncement or the end of the tracking period whichever comes first. Similarly, in the CPH modelfor payout initiation through repurchases, we define the event as the first post-IPO share repurchaseannouncement. The non-event group in the repurchase hazard models includes non-payout firms anddividend firms. During the post-IPO phase, IPO firms will either initiate a repurchase, leave the sampleas a result of failure or acquisition, or continue to operate without repurchasing shares. The dependentvariable for both types of CPH models is the log of the hazard rate and a common set of 17 independentvariables are included in the analysis.

Since the dependent variable is the log of the hazard rate, a positive coefficient indicates thanan increase in the value of the independent variable results in an increase in the likelihood of theevent (payout initiation) and consequently a decrease in the time between IPO and payout initiation.Similarly, a negative coefficient indicates that an increase in the value of the independent variableleads to a lower likelihood of payout initiation and consequently longer time between IPO and payoutinitiation. In addition to the direction of relationship and significance of the coefficients of the inde-pendent variables, risk ratios generated by CPH models are also of interest. The hazard or risk ratioshelp increase our understanding of the interpretation of the coefficients in hazard analysis (LeClere,2000) and provide an idea of the economic significance of each independent variable. For indicatorvariables the risk ratio of an independent variable is interpreted as the estimated hazard of the eventof interest for firms with a value of one relative to the hazard of the event of interest for firms with avalue of 0, after controlling for other covariates. Similarly, for continuous variables, the risk ratio rep-resents the percentage change in the hazard of the event of interest for a unit increase in the covariate,after controlling for other factors (Allison, 1984, 2000). The results of the CPH model for the timing ofdividend initiations are reported in Table 6.

The overall model Chi-square, individual variable coefficients and their associated p-values as wellas risk ratios are reported. The results in Table 6 indicate that several variables are significantly relatedto the timing of the dividend initiation decision. The estimated coefficients of VC, and NUMUSESare negative and significant while the coefficients for OPINCOME, and MATURITY are positive andsignificant. As such, the results indicate that the time to dividend initiation is longer for IPO firms


Table 6Cox proportional hazard models of timing of dividend initiations by IPO firms.

Variables Coefficient (p-value) Risk ratio

TECH 0.1625 (0.5963) 1.177VC −0.7782 (0.0262)b 0.459NUMUSES −0.4115 (0.0448)b 0.663ACQ 0.4809 (0.5379) 1.618REF −0.2371 (0.6693) 0.789GEN −0.2914 (0.6268) 0.747CAPEX −3.3368 (0.1564) 0.036RDA −6.1462 (0.1558) 0.002BWDPREMIUM 0.0165 (0.2128) 1.017LEVERAGE 0.1932 (0.7965) 0.213OPINCOME 6.4783 (0.0187)b 650.87MATURITY 0.3534 (0.0024)a 1.424GROWTH −0.2124 (0.2687) 0.809TIE −0.0017 (0.3337) 0.998CASHFLOW 3.4391 (0.2058) 31.16LAMT 0.0967 (0.6123) 1.102RANK 0.1001 (0.2422) 1.105Likelihood ratio 110.3036 (0.0001)a

The factors influencing the timing of the dividend initiation decision by IPO firms are estimated by theCox Proportional Hazard models. The time to dividend initiation is measured as the number of monthsfrom the IPO to dividend initiation announcement. The dependent variable is the log of the hazard rate.TECH takes on the value 1 if the IPO issuer belongs to a high technology industry and 0 otherwise. Thevariable VC takes on the value 1 if the IPO firm is venture capital backed and 0 otherwise. NUMUSES rep-resents the number of uses for the IPO proceeds identified by the issuing firm in its offering prospectus.ACQ is a dummy variable that takes on the value 1 if the primary use of the IPO proceeds is for financingacquisitions and 0 otherwise. REF takes on the value 1 if the primary intended use of the IPO proceedsis for retiring/refinancing debt and 0 otherwise. GEN takes in the value 1 if the primary intended useof the IPO proceeds is for general corporate purposes and 0 otherwise. CAPEX is the capital intensitydefined as capital expenditures (Compustat item #128) divided by the total assets (Compustat item#6) in the fiscal year of dividend initiation. RDA is the R&D intensity computed as R&D expenditure(Compustat item #46) divided by total assets (Compustat item #6) and is measured in the fiscal year ofdividend initiation. BWDPREMIUM is the value weighted Baker and Wurgler (2004) dividend premiummeasure, which is computed as the difference between the logarithms of the market to book ratios ofdividend payers and non-payers and is measured in the year prior to the dividend initiation. LEVERAGErepresents firm leverage and is measured by the long-term debt (Compustat item #9) divided by totalassets (Compustat item #6) in the fiscal year of dividend initiation. OPINCOME is measured as EBITDA(Compustat item #13) divided by total assets (Compustat item #6) measured in the fiscal year of divi-dend initiation. MATURITY is measured as the log of 1 + age of the firm at the IPO. GROWTH is the marketto book ratio [(book value of assets − book value of equity + market value of equity)/book value of assetswhere market value of equity = Compustat item #24 × item #25 and book value of equity = Compustatitem #216] and is measured in the fiscal year of the dividend initiation. TIE is the interest coverageand is measured as EBITDA (Compustat item #13) divided by Interest expense (Compustat item #15)measured in the fiscal year of dividend initiation. CASHFLOW represents cash flow deflated by totalassets and is measured as the sum of the earnings before extraordinary items (Compustat item #18)and depreciation (Compustat item #14) divided by total assets (Compustat item #6) measured in thefiscal year of dividend initiation. LAMT is a proxy for the size of the firm and is measured as the logarithmof the gross proceeds raised at the IPO. RANK is the investment-banking prestige and is measured usingthe Carter et al. (1998) nine-point reputation scale. The estimated coefficient, p-value (in parenthesis),significance level, and risk ratios for each independent variable are reported. Significant at the 0.10 level.


that receive VC financing and indicate a larger number of uses for the proceeds raised at the IPO. Onthe other hand, the time lag between the IPO and dividend initiation is shorter for more mature andprofitable firms. For instance, the risk ratio of 0.459 for the variable VC, indicates that the hazard (risk)of dividend initiation for VC backed IPO firms is 45.9% that of a non-VC backed IPO firms. Alternatively,non-VC backed are 2.18 times more likely to initiate post-IPO dividends compared to VC backed firms.Similarly, the risk ratio of NUMUSES indicates that every additional use of IPO proceeds indicated in


the IPO prospectus lowers the hazard of dividend initiation by 33.70%. Overall, the results suggest thatthe timing of the post-IPO dividend initiation decision depends on aspects such as VC participation,extent of demand for funds, and life cycle factors such as maturity and profitability.

The results of the CPH model for payout initiation through repurchases are reported inTable 7.

Table 7Cox proportional hazard models of timing of repurchase announcements by IPO firms.

Variables Coefficient (p-value) Risk ratio

TECH −0.0834 (0.6056) 0.920VC 0.1573 (0.3176) 1.170NUMUSES 0.0428 (0.5752) 1.044ACQ −0.1679 (0.7078) 0.845REF −0.3453 (0.2456) 0.708GEN −0.2872 (0.3296) 0.750CAPEX 3.6525 (0.0001)a 38.57RDA 1.6225 (0.1260) 5.066BWDPREMIUM −0.0104 (0.1022) 0.990LEVERAGE −1.4317 (0.0016)a 0.750OPINCOME 1.9700 (0.0599)b 7.171MATURITY 0.0699 (0.3381) 1.072GROWTH −0.4920 (0.0001)a 0.611TIE 0.0004 (0.2862) 1.00CASHFLOW 1.7041 (0.0567)b 5.497LAMT 0.1075 (0.3532) 1.114RANK 0.0306 (0.4927) 1.031Likelihood ratio 114.4429 (0.0001)a

The factors influencing the timing of the repurchase initiation decision by IPO firms are estimated bythe Cox Proportional Hazard models. The time to repurchase initiation is measured as the numberof months from the IPO date to the repurchase initiation announcement. The dependent variable isthe log of the hazard rate. TECH takes on the value 1 if the IPO issuer belongs to a high technologyindustry and 0 otherwise. VC takes on the value 1 if the IPO firm is venture capital backed and zerootherwise. NUMUSES represents the number of uses for the IPO proceeds identified by the issuingfirm in its offering prospectus. ACQ is a dummy variable that takes on the value 1 if the primaryuse of the IPO proceeds is for financing acquisitions and 0 otherwise. REF takes on the value 1 if theprimary intended use of the IPO proceeds is for retiring/refinancing debt and 0 otherwise. GEN takesin the value 1 if the primary intended use of the IPO proceeds is for general corporate purposes and0 otherwise. CAPEX is the capital intensity defined as capital expenditures (Compustat item #128)divided by the total assets (Compustat item #6) in the fiscal year of repurchase initiation. RDAis the R&D intensity computed as R&D expenditure (Compustat item #46) divided by total assets(Compustat item #6) and is measured in the fiscal year of repurchase initiation. BWDPREMIUM isthe value weighted Baker and Wurgler (2004) dividend premium measure, which is computed as thedifference between the logarithms of the market to book ratios of dividend payers and non-payersand is measured in the year prior to repurchase initiation. LEVERAGE represents firm leverage andis measured by the long-term debt (Compustat item #9) divided by total assets (Compustat item#6) in the fiscal year of repurchase initiation. OPINCOME is measured as EBITDA (Compustat item#13) divided by total assets (Compustat item #6) measured in the fiscal year of repurchase initiation.MATURITY is measured as the log of 1 + age of the firm at the IPO. GROWTH is the market to book ratio[(book value of assets–book value of equity + market value of equity)/book value of assets wheremarket value of equity = Compustat item #24 × item #25 and book value of equity = Compustat item#216] and is measured in the fiscal year of the repurchase initiation. TIE is the interest coverage andis measured as EBITDA (Compustat item #13) divided by Interest expense (Compustat item #15)measured in the fiscal year of repurchase initiation. CASHFLOW represents cash flow deflated bytotal assets and is measured as the sum of the earnings before extraordinary items (Compustat item#18) and depreciation (Compustat item #14) divided by total assets (Compustat item #6) measuredin the fiscal year of repurchase initiation. LAMT is a proxy for the size of the firm and is measured asthe logarithm of the gross proceeds raised at the IPO. RANK is the investment-banking prestige andis measured using the Carter et al. (1998) nine-point reputation scale. The estimated coefficient, p-value (in parenthesis), significance level, and risk ratios for each independent variable are reported.Significant at the 0.05 level.



With the exception of profitability, the factors influencing the timing of share repurchase initiationsare different from those affecting dividend initiations. The results indicate that the coefficients ofCAPEX, OPINCOME and CASHFLOW are positive and significant indicating that the time to repurchaseinitiation is shorter for IPO firms with higher capital expenditures, higher profitability, and strongercash flows. On the other hand, the coefficients of LEVERAGE, and GROWTH are negative and significantindicating that the time to repurchase initiation is longer for IPO firms that are more highly leveredand firms with stronger growth prospects.

5. Discussion and conclusions

Extant research on payout initiations by IPO firms has largely focused on dividends without con-sidering repurchases as an alternative form of payout initiation mechanism available to newly publicfirms. Recent empirical research on established companies however, suggests that repurchases havegrown at a faster rate than dividends and increasingly become the preferred and more economicallysignificant form of distributing cash flow payouts to shareholders. In the context of established firms,there is ongoing debate in the corporate payout literature regarding the extent to which repurchasessubstitute for dividends. While some researchers have argued in favor of the substitution hypoth-esis suggesting that dividends and repurchases are essentially interchangeable, others have arguedthat these two forms of payout mechanisms are adopted by fundamentally different types of firmsat different stages in their evolution. Since the comparative economic analysis of dividends versusrepurchases has largely been confined to established firms, relatively little is known regarding theextent to which newly public firms utilize dividends versus share repurchases to initiate payouts oron the economics of the choice between these two forms of payout initiation mechanisms availableto IPO firms.

This study attempts to add to the extant literature on payout initiations by IPO firms by extendingthe analysis beyond dividends to include share repurchases and evaluating the economic impact ofinitiating post-IPO payouts via dividends versus repurchases. Specifically, this study evaluates theextent to which dividends or repurchases represent the preferred form of initiating post-IPO payouts,whether investors demonstrate a preference for one form of payout over the other, and identify factorsthat influence a firm’s choice between these two forms of payout initiation mechanisms. Finally, weinvestigate whether the timing of these two forms of payout initiation mechanisms are affected by asimilar or different set of economic factors.

Our results indicate that consistent with recent trends documented for established companies, thepreferred form of payout initiation by IPO firms is through share repurchases. For instance, we findthat payout initiating IPO firms are approximately three times more likely to select repurchases overdividends as the specific payout mechanism. Additionally, we find that while dividend firms tend toinitiate payouts either immediately after the IPO or after many years of going public, repurchase firmstend to initiate payouts more uniformly over time during the post-IPO phase. An analysis of the char-acteristics of dividend versus repurchase IPO firms reveals several significant differences between thetwo groups. For instance, dividend firms are characterized by higher leverage, profitability, maturity,sales, and total assets, but have significantly lower growth prospects, initial returns, and R&D inten-sity. They are also less likely to have been backed by VCs. Further, dividend firms are significantly lesslikely to be in high technology industries compared to repurchase firms. As such, dividend IPO firmsare characterized by the typical life cycle characteristics that favor distribution over retention andtherefore appear to be motivated to initiate dividends as a result of transitioning to a more maturestage of their development. On the other hand, our analysis reveals that repurchase IPO firms havefundamentally different characteristics from dividend firms and do not appear to follow the life cycleapproach to distribution, but rather tend to initiate payouts during the early post-IPO phase while stillrelatively young and in a high growth mode.

We find that the market reacts positively to both dividend as well as share repurchase initiationsby IPO firms. Further, the difference in the market reaction to the announcement of these two formsof payout initiation mechanisms is not economically significant. Analyzing the determinants of themarket reaction to announcements of post-IPO payout initiations, we find that while the specificform of payout mechanism is not significant, factors such as extent of product market competition,


acquisition growth strategy adopted, and extent of financing constraints influence investor reaction.Overall, our results are suggestive of the fact that while the market views payout initiations positively,it remains indifferent to the specific form of payout mechanism selected to distribute cash flows toshareholders. As such, we find no evidence to indicate that the market reaction should be a factor inconsidering whether to initiate payouts via dividends versus repurchases for IPO firms.

Our analysis identifies several factors that influence the choice between dividends versus repur-chases to initiate post-IPO payouts. We find strong evidence to indicate that IPO firms are more likelyto choose repurchases over dividends when they receive VC backing, have larger number of uses forcapital raised, and are subjected to strong competition in their product markets. On the other hand,IPO firms prefer dividends to repurchases when investor demand for dividends is high as indicated bytheir willingness to offer higher dividend premiums and when the firm has reached a certain stage ofmaturity and growth. The propensity to initiate post-IPO payouts through repurchases by firms thatare characterized by a strong demand for external funds, receipt of VC backing and strong productmarket competition is consistent with the argument in the literature that repurchases can serve ascredible signals to potentially correct equity undervaluation. Signaling with stock repurchases is par-ticularly relevant for firms that intend to raise equity capital to finance their growth or when influentialinsiders such as VCs need to exit through equity sales during the post-IPO phase and are thereforeconcerned about market valuations. On the other hand, life cycle factors and investor appetite fordividends appear to be the driving force behind dividend initiations by IPO firms.

Overall, our results indicate that in the context of the IPO market, repurchases and dividendsrepresent distinct forms of payout mechanisms used by different types of firms under different cir-cumstances. Further, IPO firms demonstrate a preference for repurchases over dividends as the specificform of payout initiation mechanism. While the market views post-IPO payout initiations favorably, itis indifferent to the specific form of payout mechanism adopted by the firm. Rather, while the choiceof dividend initiations are more in tune with the life cycle and catering theory, repurchase decisionsappear to be driven by signaling theory. Our study provides additional insights into the economics ofthe choice between dividends versus repurchases as the form of payout initiation mechanism adoptedby entrepreneurial firms as they make the transition from private to public corporations.

References

Allen, F., Bernardo, A. E., & Welch, I. (2000). A theory of dividends based on tax clienteles. Journal of Finance, 55, 2499–2536.Allen, F., & Faulhaber, G. (1989). Signaling by underpricing in the IPO market. Journal of Financial Economics, 23, 303–323.Allison, P. D. (1984). Event history analysis. Beverly Hills: Sage Publications.Allison, P. D. (2000). Survival analysis using the SAS system: A practical guide. SAS Institute Publishing.Baker, M., & Wurgler, J. (2004). A catering theory of dividends. Journal of Finance, 59, 1125–1165.Brav, A., Graham, J., Harvey, C., & Michaely, R. (2005). Payout policy in the 21st century. Journal of Financial Economics, 77,

483–527.Bulan, L., Subramanian, N., & Tanlu, L. (2007). On the timing of dividend initiations. Financial Management, 36, 1–35.Carter, R., Dark, F. H., & Singh, A. K. (1998). Underwriter reputation, initial returns, and the long-run performance of IPO stocks.

Journal of Finance, 53, 285–311.Chowdhry, B., & Nanda, V. (1994). Repurchase premia as a reason for dividends: A dynamic model of corporate payout policies.

Review of Financial Studies, 7(2), 321–350.DeAngelo, H., DeAngelo, L., & Stulz, R. (2006). Dividend policy and the earned/contributed capital mix: A test of the lifecycle

theory. Journal of Financial Economics, 81, 227–254.Desmukh, S. (2003). Dividend initiations and asymmetric information: A hazard model. The Financial Review, 38, 351–368.Fama, E. F., & French, K. M. (2001). Disappearing dividends: Changing firm characteristics or lower propensity to pay? Journal

of Financial Economics, 60, 3–43.Gorman, M., & Sahlman, W. A. (1989). What do venture capitalists do? Journal of Business Venturing, 4(4), 231–248.Grullon, G., & Michaely, R. (2002). Dividends, share repurchases, and the substitution hypothesis. Journal of Finance, 57,

1649–1684.Grullon, G., Michaely, R., & Swaminathan, B. (2002). Are dividend changes a sign of firm maturity? Journal of Business, 75,

387–424.Grullon, G., & Michaely, R. (2004). The information content of share repurchase programs. Journal of Finance, 59(2), 651–680.Grullon, G., & Michaely, R. (2008). Corporate payout policy and product market competition. AFA 2008 New Orleans meetings

paper.Hellman, T., & Puri, M. (2002). Venture capital and the professionalization of start-up firms: empirical evidence. Journal of

Finance, 57, 169–197.Jagannathan, M., Stephens, C. P., & Weisbach, M. S. (2000). Financial flexibility and the choice between dividends and stock

repurchases. Journal of Financial Economics, 57, 355–384.


Jain, B. A., & Kini, O. (1995). Venture capital participation and the post-issue operating performance of IPO firms. Managerialand Decision Economics, 16, 593–606.

Jensen, M. (1986). Agency costs of free cash flow, corporate finance and takeovers. American Economic Review, 76, 323–329.Kahle, K. M. (2002). When a buyback isn’t a buyback: Open market repurchases and employee options. Journal of Financial

Economics, 63, 235–261.Kale, J. R., Kini, O., & Payne, J. D. (2006). On the dividend initiation decisions of newly public firms. Available at SSRN:

http://ssrn.com/abstract=913909Kale, J., & Noe, T. (1990). Dividends, uncertainty, and underwriting costs under asymmetric information. Journal of Financial

Research, 4, 265–277.Kaplan, S. N., & Zingales, L. (1997). Do investment cash flow sensitivities provide useful measures of financing constraints?

Quarterly Journal of Economics, 112(1), 169–215.LeClere, M. J. (2000). The occurrence and timing of events: Survival analysis applied to the study of financial distress. Journal of

Accounting Literature, 19, 158–189.Lerner, J. (1994). Venture capitalists and the decision to go public. Journal of Financial Economics, 35, 293–316.Li, W., & Lie, E. (2006). Dividend changes and catering incentives. Journal of Financial Economics, 80, 293–308.Lipson, M. L., Maquieira, C. P., & Megginson, W. (1998). Dividend initiations and earnings surprises. Financial Management, 27,

36–45.Liu, Q. (2000). Growth opportunities, knowledge capital, and leverage: Evidence from U.S. biotech firms. Working paper, Uni-

versity of Hong Kong.Loughran, T., & Ritter, J. (2004). Why has IPO underpricing changed over time? Financial Management, 5–37.Michaely, R., & Shaw, W. H. (1994). The pricing of initial public offerings: Tests of adverse-selection and signaling theories.

Review of Financial Studies, 7, 279–319.Miller, M., & Rock, K. (1985). Dividend policy under asymmetric information. Journal of Finance, 40, 1021–1051.Sahlman, W. A. (1990). The structure and governance of venture-capital organizations. Journal of Financial Economics, 27(2),

473–521.Skinner, D. J. (2008). The evolving relation between earnings, dividends, and stock repurchases. Journal of Financial Economics,

87, 582–609.

http://ssrn.com/abstract=913909

Documents

Payout initiation by IPO firms: The choice between dividends and share repurchases