Financial Flexibility, Firm Size and Capital Structure

Financial Flexibility, Firm Size and Capital Structure by Soku Byoun Hankamer School of Business Baylor University One Bear Place 98004 Waco, TX 76798 Tel: (254) 710 7849 Fax: (710) 710-1092 Email: Soku Byoun@Baylor.edu October 2007 (Preliminary results. Do not quote without author s permission. Comments are welcome.) We would like to thank the GAMF. We appreciate the support for this project that was provided by the Hankamer School of Business at Baylor University.

Financial Flexibility, Firm Size and Capital Structure Abstract We examine financial flexibility as an alternative explanation to existing capital structure theories. The financial flexibility hypothesis suggest non-linear inverted V-shape relationship between leverage ratio and firm size. Consistent with this prediction, we find that small firms have lower leverage ratios, not because of internally generated funds (as implied by the pecking order theory) but because of additional equity financing (consistent with our financial flexibility hypothesis). On the other hand large firms have low leverage because they rely on internally generated fund. This finding can not be explained by either the pecking order theory or the tradeoff theory; the pecking order may be reversed for small firms that prefer external equity to debt financing, while the tradeoff theory may miss some important aspects of capital structure decisions. The financial flexibility hypothesis suggest that small firms maintain low leverage by issuing equity and building up cash holdings for financial flexibility. covenants often carry restrictions on financing and investment decisions that are especially cumbersome for small, growing firms. Equity financing allows small firms to raise cash without impeding financial flexibility. We find supporting evidence. We also find that the positive relationship between firm size and leverage ratio found in previous studies holds only for small firms, but there is a clear negative relationship for large firms, supporting the financial flexibility hypothesis. JEL Classification: G32 Keywords: Financial Flexibility; Trade-off theory; Pecking-order theory 2

I. Introduction Empirical studies on capital structure find a positive relationship between firm size and leverage. 1 Suggested explanations in the literature include: large firms tend to have more leverage perhaps because they are more transparent; have lower asset volatility; are more diversified; naturally sell large enough debt issues so that the fixed costs of public borrowing are not prohibitive; have lower probability of default and lower financial distress costs. On the other hand, small firms incur higher costs of issuing debt because they are more likely to be growing firms with volatile cash flows and hence have less access to debt capital than do large firms. Further, the costs of financial distress are likely to be particularly severe for small firms because much of their value comes from growth options whose value depreciates rapidly if the firm experiences financial distress. In addition, small firms have a large fraction of their assets that are firm specific or intangible, limiting their value as collateral. We suggest the desire for financial flexibility as an alternative explanation for small firms low leverage. By taking firms financial choices to be the outcome of the interplay between today s capital demand and future supply of capital, we are able to explain some of the unexplained empirical regularities: e.g., why the pecking order theory of financing appears to break down for small risky firms (Frank and Goyal, 2003), why firms issue too much equity (Frank and Goyal, 2003) at the wrong times (Fama and French 2004, and Leary and Roberts, 2005), and why there is a negative relationship between firm size and leverage ratio (Faulkender and Petersen (2006)). Given that small firms grow faster than large firms (Evans (1987)), there are two alternative explanations for small firms to have lower leverage; by financing their growth exclusively through retained earnings or through external equity. Most explanations for the positive relationship between firm size and leverage assume implicitly or explicitly that 1 For example, see Titman and Wessels (1988), Rajan and Zingales (1995), Graham, Lemmon, and Schallheim (1998), Hovakimain, Opler, and Titman, (2001), Booth et al. (2001) and Fama and French (2002). However, Faulkender and Petersen (2006) find a weak or negative relationship between the leverage and firm size. 1

external equity is prohibitively expensive for small firms and hence small firms financing should come exclusively from internal funds. 2 There is also evidence that firms prefer internal financing to external financing. 3 According to the pecking order theory (Myers and Majluf (1984)), adverse selection costs of external equity are much greater than those of debt. Issuance costs are also much greater for equity than for debt. 4 Facing such high adverse selection/transaction costs, small firms should avoid issuing equity by all means. Hence, the literature has paid little attention to the potential role of external equity in relaxing financial constraints of small firms in debt financing. We argue that small firms with lack of financial flexibility maintain low leverage by relying on external equity financing. Financial flexibility is desired in order to handle financial uncertainties in the future. Here the future orientation is critical; financial flexibility concern arises from a formal decision problem in which the financial choices in the future are affected by the financial choices made today. In other words, financial flexibility is a property of today s financial decisions that affect the options management will have in response to unforeseeable future changes. Thus, our study is distinguished from most previous studies which focus on how current or past shocks affect firms financial decisions. Even though financial flexibility is very important for a firm to respond optimally to future changes, it would be fundamentally inappropriate to maximize financial flexibility. Therefore, we define financial flexibility as the extent to which a firm can mobilize its financial resources to take preventive and exploitive actions in response to subsequent information in order to maximize the firm value. According to this view, the optimal capital structure from the static point of view may not be optimal once financial flexibility, i.e., the interplay between today s financial decisions and future financial decisions, is taken into 2 See, for example, Frank and Goyal (2003), Leary and Roberts (2005), Strebulaev (2006), and Kurshev and Strebulaev (2006) for such arguments. 3 See Hovakimian, Opler, and Titman (2001) and Hovakimian, Hovakimian, and Tehranian (2004). 4 Altinkilic and Hansen (2000) find that equity issuing costs are on average 5.38% of the issue proceeds while debt issuing costs are on average 1.09%. Leary and Roberts (2005) also report significantly larger equity issuance costs. 2

account. In order to examine the importance of financial flexibility hypothesis in financial decisions, we focus on firms decisions on cash holdings and debt/equity financing which are the main sources of financial flexibility (DeAngelo and DeAngelo (2006)). Firms lacking in financial flexibility would increase cash holdings while reducing leverage ratios. Even though we consider the dividend policy in relation to financial flexibility, we do not take it as a source of financial flexibility but rather as the decision following from having ample flexibility. Thus, the dividend policy is taken as an identifier of a firm s flexibility. We also consider firm size, retained earnings and long-term credit ratings as interrelated measures of the firm s financial flexibility. Small growing firms typically have high demand for financial flexibility and have accumulated less profits than large firms, causing them to concern for financial flexibility. For example, small firms with accumulated losses (measured by retained earnings) are likely to have little financial flexibility and debt capacity but ample need for additional cash. Our main hypothesis is that small firms with negative retained earnings are more likely to issue equity to build up cash holdings and preserve debt capacity for financial flexibility and hence will have low leverage. A corollary to this hypothesis is that cash holdings will be negatively associated with leverage. We find supporting evidence for these predictions. On the other hand, large mature firms with ample retained earnings are almost selffinancing and have low leverage. Thus, their large cash flows are subject to opportunity/agency costs of free cash flow (Jensen and Meckling (1976)). Increasing debt financing may address the free cash flow problem but it may reduce financial flexibility to cope with future uncertainties. Thus, they prefer paying dividends in order to preserve debt capacity while reducing free cash flow. This implies that large firms with ample retained earnings also have lower leverage. The important implication of the financial flexibility hypothesis is that there is a negative relationship between firm size and leverage ratio, conditional on negative retained earnings, whereas there a positive relationship between firm size and leverage ratio, conditional on positive retained earnings. 3

We find that low leverage for small firms results from external equity financing rather than internal funds. While small firms avoid debt financing, they are much more active in tapping into external equity capital. Firms with negative retained earnings have lower leverage ratios and issue several times more equity than firms with positive retained earnings. We also find that small firms with negative retained earnings have more cash holdings than other firms. Thus, our findings are consistent with the argument that small firms build financial flexibility by increasing cash holdings and equity financing to cope with their earnings shortfalls. We further show that firm size has an overall strong and significant positive association with leverage ratios. However, the positive relationship between firm size and leverage ratios is reversed for large firms when we re-estimate regressions for firms divided by size, consistent with our financial flexibility hypothesis. We also find that negative retained earnings are negatively associated with leverage ratio, suggesting that firms with negative retained earnings are issuing more equity. On the other hand, positive retained positive retained earnings are significantly and negatively associated with leverage ratios. Thus, our results show that firms with large negative retained earnings build up cash holdings through equity financing in order to preserve financial flexibility, whereas firms with more positive retained earnings maintain lower leverage ratios by relying on internally generated equity financing and financial flexibility. Overall, the results show strong support for the financial flexibility hypothesis. Our study brings new evidence to bear on an important issue in the capital structure literature. The literature has wrestled with the problem of sorting out the effects of adverse selection costs of asymmetric information on capital structure. 5 On the one hand, the literature finds that larger firms appear to provide a better fit for the pecking order theory (Shyam-Sunder and Myers (1999) and Frank and Goyal (2003)) despite the fact that large 5 For example, see Myers and Majluf (1984), Viswanath (1993), Chang and Dasgupta (2003), and Lemmon and Zender (2004) under the pecking order framework, and Frank and Goyal (2003), Fama and French (2002), Barclay and Smith (2005), Leary and Roberts (2005), Leary and Roberts (2005a), Strebulaev (2006) and Byoun (2007) under the tradeoff framework. 4

firms are less subject to information asymmetry than small firms. Our finding suggests that large firms prefer using internal funds to preserve financial flexibility. On the other hand, our results show that small firms issue equity and increase cash holdings despite having low leverage in order to cope with the lack of financial flexibility. Lemmon and Zender (2004) argue that equity issuers are prevented from issuing debt because of concerns over financial distress or financial slack for future investment. This justification is essentially in the same vein as our financial flexibility argument which does not require the presence of asymmetric information. Further, Fama and French (2002) and Leary and Roberts (2005a) show that firms are more likely to use external equity as investment increases and/or cash flow decreases but the majority of equity financings occur when firms still have sufficient debt capacity to fill their investment needs. Their findings are also consistent with the financial flexibility argument. As an explanation for violating the financing hierarchy specified by the pecking order theory, Fama and French (2005) and Leary and Roberts (2005a) suggest that firms are able to issue securities in a manner that avoids the adverse selection costs associated with information asymmetry. Alternatively, managers may time the market when information asymmetry and the corresponding costs are low so that deviations from the hierarchy do not incur a significant penalty. 6 Our findings suggest that the external financing hierarchy suggested by the pecking order theory is reversed for small firms due to the concern for financial flexibility. Consistent with our evidence, the findings in Byoun (2007a) suggest that small firms become debt-free by having raised much external equity while having reduced debt and paid large dividends. In conclusion, asymmetric information falls short of providing a plausible explanation for motivation behind firms external financing decisions. A more comprehensive explanation should include the benefits and costs of financial flexibility, which may require a substantial 6 Rajan and Zingales (1995), Jung, Kim, and Stulz (1996), Pagano, Panetta, and Zingales (1998), Hovakimian, Opler, and Titman (2001), Baker and Wurgler (2002), and Leary and Roberts (2005a)) examine managers market-timing attempts. The survey results in Graham and Harvey (2001) suggest that managers issue equity following an increase in stock price. 5

alteration to the traditional tradeoff argument. The rest of the paper proceeds as follows: Section II provides a literature review and discusses the concept of financial flexibility. Section III develops financial flexibility hypotheses. Section IV describes the data. Section V provides the estimation results and Section V concluding remarks. II. A Literature Review and the Concept of Financial Flexibility Despite managers contention that financial flexibility is an important factor in their decision-making process, 7 the capital structure literature has to date remained aloof to recognize financial flexibility as a potentially important factor in explaining firms capital structure decisions. Frank and Goyal (2005) reason, the stress on financial flexibility is interesting, but potentially open to a variety of interpretations. In our view the survey evidence is of interest, but it is best regarded as being interesting and suggestive, rather than providing definitive tests. In addition to considerable ambiguity in the use of the term, judgments about financial flexibility are subjective and informal and flexibility levels are rarely monitored or even measured. Accordingly, dealing with financial flexibility may be criticized as being less than practical and based on speculation on the ability of a firm to respond to hypothetical future events. It is therefore not surprising that there is relatively little systematic study of financial flexibility in the capital structure literature. 8 Financial flexibility is distinguished from financial slack. Firms facing an adverse selection problem desire financial slack in order to avoid the need for external funds in the future, whereas financial flexibility is desired in order to handle uncertainties and variations in both internal and external environments. Thus, finding that managers value financial flexibility is not sufficient to prove that the pecking-order model is the true description of 7 The survey results of Graham and Harvey (2001), Bancel and Mittoo (2004), and Brounen et al. (2004) show that corporate managers explicitly express that they are mostly concerned about financial flexibility in their capital structure decisions. 8 In contrast, a branch of real options literature has been developed to deal with investment flexibility. Gamba and Triantis (2005) note that most real options models are designed to measure the value of investment flexibility under the assumption of perfect financial flexibility. 6

capital structure choice (Opler et al., 1999). Graham and Harvey (2001) make this point explicit: The most important item affecting corporate debt decisions is management s desire for financial flexibility,... However, the importance of flexibility in the survey responses is not related to informational asymmetry (size or dividend payout) or growth options in the manner suggested by the pecking-order theory. In fact, flexibility is statistically more important for dividend-paying firms, opposite the theoretical prediction (if dividend-paying firms have relatively little informational asymmetry). Therefore, a deeper investigation indicates that the desire for financial flexibility is not driven by the factors behind the peckingorder theory. Graham Harvey (2001) see financial flexibility as preserving debt capacity to make future expansions and acquisitions or minimizing interest obligations, so that they do not need to shrink their business in case of an economic down turn. Gamba and Triantis (2005), in their attempt to model the value of financial flexibility, define, financial flexibility as the ability of a firm to access and restructure its financing with low transaction costs. They further elaborate by adding that financially flexible firms are able to avoid financial distress in the face of negative shocks, and to fund investment at low cost when profitable opportunities arise. Donaldson (1969, 1971) uses financial mobility to describe the capacity to redirect the use of financial resources in a manner consistent with the evolving goals of management as it responds to new information about the company and its environment. Donaldson particularly relates financial mobility to capital structure decisions where the goal is to find the optimal mix of financing sources. Heath (1978) describes a financially flexible firm as one that can take corrective action that will eliminate an excess of required cash payments over expected cash receipts quickly and with minor adverse effect on its present and future earnings or on the market value of its stock. The American Institute of Certified Public Accountants (AICPA, 1993) adopts Heath s view by defining financial flexibility as the ability to take action that will eliminate an excess of required and expected cash payments over expected resources. The Financial 7

Accounting Standards Board s (FASB) defines financial flexibility as the ability of an entity to take effective actions to alter amounts and timing of cash flows so it can respond to unexpected needs and opportunities. Most treatments of financial flexibility in the finance literature are more or less about the ability of a firm to meet its unexpected future needs through large cash holdings, large unused borrowing capabilities, or large liquid assets. There are other views from different functional areas of business (See Koornhof (1998) for a review). For example, the importance of flexibility in a firm is well recognized in management and organization literature. Bueno-Campos (1992), Ahmed et al. (1996), Albizu-Gallastegui (1997), Hitt et al. (1998) and Volberda (1998) define flexibility as the ability to deliver cost-efficient responses quickly to changes in the business environment and to adapt and anticipate changes that affect the goals of firms. Bernstein (1993) defines flexibility as the ability of an enterprise to take steps to counter unexpected interruptions in the flow of funds for reasons however unexpected. Koornhof (1998) defines flexibility as an ability to take actions in order to reposition the resources and functions of the organization to new information and environment in a manner consistent with the evolving vision, strategies and goals of management. In this view, financial flexibility means the ability to borrow from a variety of sources, to raise equity capital, to sell and redeploy assets, and to adjust the level and the direction of operations in order to meet changing circumstances. The various definitions of flexibility as addressed in the literatures recognizes the reactive or preventive nature of flexibility, while failing to include the exploitive nature of flexibility for uncertain competitiveness or opportunities. The combination of preventive and exploitive nature of flexibility is more evident in Volberda (1998) who views flexibility in two different perspectives: internal flexibility as the firm s capacity to adapt to the demands of the environment, while external flexibility as the firm s capacity to influence their environment and thereby reduce their vulnerability. Actions initiated ahead of time are typically taken not only in anticipation of certain events but also in an attempt to change the rules of the game. When expectations are not met, or when events occur that have not been predicted, a firm may require flexibility after the fact. In these cases, flexibility is needed to deal with uncertain obstacles or to capitalize 8

on an unexpected opportunity. Actions taken ahead of time, even in the absence of a specific need, can create options that can be used at a later stage. When a new product unexpectedly becomes an industry standard (e.g., Apple s ipod), resulting in a rapid expansion of the market demand, exploitive maneuvers are important to focus financial resources in order to rapidly capitalize on the new opportunity. The most basic form of financial flexibility may be described in terms of the amount or the number of financial resources available in the future. However, many of the actions taken today for future financial flexibility can be very costly. Thus, it would be fundamentally inappropriate for a firm to maximize financial flexibility. A more intuitive measure of financial flexibility may be described as a firm s financial ability to respond optimally to future changes at lower costs. In many situations, however, we may not meaningfully separate the cost related to such a decision. Eventually, maximizing the firm value should be the ultimate goal of optimizing financial flexibility. Accordingly, we define financial flexibility as the extent to which a firm can mobilize its financial resources to take preventive and exploitive actions in response to subsequent information in order to maximize the firm value. It is apparent that certain aspects of financial flexibility have been noted and addressed in the capital structure literature. For example, Donaldson (1969) observes that managers do not follow the optimizing principle proposed by the tradeoff theory of optimal capital structure. Instead he finds that managers concentrate on the amount of debt not in use. Goldstein, Ju, and Leland (2001) note that a firm with low leverage today preserves the subsequent option to increase leverage. Byoun (2007) finds evidence that firms preserve borrowing capacity to finance future investment or growth opportunities. Graham (2000) shows that firms preserve debt capacity to make future expansions and acquisitions. DeAngelo, DeAngelo and Stulz (2007) document evidence that firms issue stock to cope with liquidity squeeze, a near-term need for cash. McLean (2007) shows that firms keep equity issuance proceeds as cash as precautionary motives. Motyka, Leuca, and Fawson (2005) also find that financial institutions hold excess liquidity to cope with the unpredictable nature of loss (infrequent but high impact risk) in order to achieve a competitive advantage for aggressive 9

pricing and better margins. Kale and Shahrur (2007) show that firms have lower leverage in consideration of relationship-specific investments by the firm s stakeholders. Parsons and Titman (2007) provides a review on how a firm s financial condition influences strategic decisions by both the firm and its stakeholders. DeAngelo and DeAngelo (2006) argues that firms maintain low leverage and high dividend payouts in normal periods to preserve the firm s option to borrow or issue equity in the future opportunities. III. Firms Financial Flexibility: Hypothesis Development The need for financial flexibility arises to deal with anticipated yet unforeseeable future financial constraints. Developing testable hypotheses in relation to financial flexibility is complicated since we need to consider not only the interactions of firms financial choices but also the interplay of current decisions with future ones. DeAngelo and DeAngelo (2006) argue that firms can develop potential sources of financial flexibility through cash accumulation, the preservation of debt capacity, and dividend payouts. We expect that firms with a greater need for financial flexibility hold more cash, less debt (issue more equity) and pay less dividend. Large cash holdings increase financial flexibility for the future but incur higher costs. In addition to the opportunity costs foregone, they are exposed to a risk of inefficient uses when the firm does not have enough investment opportunities. Financially flexible firms can pay out its free cash flow as dividend or repurchase securities to minimize its opportunity/agency costs. financing provides the benefits of tax savings and reduced agency costs of free cash flow but reduces future financial flexibility. External equity financing incurs higher issuing costs but provides greater flexibility than debt financing. Thus, growing firms with lack of financial flexibility prefer equity financing. Given these financial choices available for the firm to address financial flexibility concerns, we develop hypotheses that predict their impacts on capital structure decisions. The interactions among these financial choices and with investment decision are interesting but we limit our focus to capital structure decisions in order to have a manageable scope of the study. Thus, our objective is to examine how the various degree of financial flexibility 10

affects capital structure decisions by identifying firms that are likely or not likely to face future financing constraints. In practice, however, the dividend and investment policies may also be altered in the long run to mitigate financial constraint. For example, a firm may invest in safe and liquid investments rather than in risk ones (Almeida et. al. (2006)) or alter the dividend policy, when facing financial constraints and running out of other financial resources. However, firms are less likely to alter their investment and dividend policies when they have other sources of financial flexibility such as cash holdings and debt capacity. Accordingly, for the purpose of investigating the effect of financial flexibility on the capital structure decision, we assume that firms make financing decisions in order to accommodate future investment opportunities. Similarly, we consider a firm s dividend policy not as in itself the immediate source of financial flexibility but as a policy to reflect the firm s financial flexibility condition. In order to understand how a financial flexibility concern affects a firm s financial decisions, suppose a firm with a constant optimal debt ratio (debt/assets) of 0.5 currently has $200 financial surplus with uncertain financing needs in the next few years. In order to preserve financial flexibility for the unforeseeable future financing needs, the firm should make financial decisions today. Considering costs and benefits, suppose the firm set the optimal financing plan of raising $300 by next period (shown as negative numbers at t = 1) as follows: t = 0 1 Surplus = $200 = $400 Assets = $1000 -$300 The firm s current debt ratio is 0.4. The static target adjustment model predicts that the firm will use the current financial surplus to repurchase equity and then issue $150 debt and $150 equity in the next period in order to maintain the optimal debt ratio of 0.5. However, if the transaction costs of repurchasing and reissuing equity are greater than those of debt, the firm will reduce debt rather than equity with the surplus cash and preserve the debt capacity (as financial flexibility) for the future financing need despite the below-target debt ratio. Or the firm may hold all or some of the surplus in order to use it in the next period, as 11

long as the opportunity/agency costs of cash holdings are lower than the costs of reducing and reissuing debt or equity. Only firms with no financing needs in the foreseeable future may repurchase equity to adjust toward the target debt ratio. In the above example, assuming it is cost efficient for the firm to keep its surplus cash for the financing plan, it still requires $100 additional equity in order to obtain the optimal debt ratio with the plan. Knowing that they will need to issue additional equity, managers may attempt to issue equity today if they believe today s market condition is better than what is expected in the future. If the firm issues equity in the above example, then the equity issuance proceeds will be kept as cash (see Kim and Weisbach (2007) for evidence of firms such behavior). This will make the firm look inconsiderate about its capital structure as well as cash holdings since it is still issuing equity with a below-target debt ratio and ample cash balance. Indeed, Leary and Roberts (2004) find that most equity issuances occur when firms have sufficient debt capacity, without any apparent risk of entering financial distress from issuing debt. They conclude that the decision rule that firms use to access external capital markets is unclear. Welch (2004) also observe that firms do not use their issuing activities to counteract the external and large influences of stock returns on their capital structures. These puzzles may be explained not through the firm s reaction to the current financial information but through its financial flexibility concern to cope with future financial constraints and uncertainties. We consider firm size, dividend payout, retained earnings and long-term credit ratings as measures of firms various degree of financial flexibility. We put these financial flexibility measures together to make consistent predictions in the framework of financial flexibility. Small firms are more likely to be growing firms with more future investment opportunities, less credit history and less internal funds available. Thus, small firms are expected to be in need for more financial flexibility than large firms. Retained earnings are the accumulation of a firm s reinvested profits over time. DeAngelo, DeAngelo, and Stulz (2005, 2007) argue that retained earnings measure a firm s lifecycle stage; firms with low retained earnings relative to total assets (or total equity) tend to be in the capital infusion stage, whereas firms with greater retained earnings tend to be 12

more mature with ample cumulative profits that make them largely self-financing. Large cash holdings for a mature firm are exposed to a risk of inefficient uses when the firm does not have enough investment opportunities. Issuing debt reduces the agency costs, but it reduces financial flexibility as well. On the other hand, dividend payouts reduce the agency costs of free cash flow (Easterbrook (1984)) without reducing financial flexibility. Thus, large firms with greater retained earnings tend to be more mature with ample cumulative profits and are likely to pay dividends (DeAngelo, DeAngelo, and Stulz (2005)) and preserve financial flexibility through low leverage while limiting agency costs on cash balances (DeAngelo and DeAngelo (2006)). With positive earnings, small growth firms still lacking in their financial resources are likely to invest them in their growth options or hold them as cash for future financing needs since they are less subject to the agency costs of free cash flow. Overall, we propose that large mature firms with positive retained earnings maintain lower leverage, lower cash holdings and high dividend payouts than small firms with positive retained earnings (proposition 1). Negative retained earnings reflect firms earnings shortfalls over time. To the extent that a firm s retained earnings are the proxy for its life-cycle stage, we expect small firms with negative retained earnings are in the greatest need of financial flexibility. For small growing firms with little financial flexibility, facing shortfalls in cash flow over time, even a little debt may cause them to be in financial distress. The limitation on debt issuance that results from the risk of asset substitution (Jensen and Meckling (1976)) are more important for such firms. Firms with accumulated losses without financial flexibility lack investible funds for their profitable investments and hence sources of free cash flow tend to be relatively low for them, and thus reducing the benefit of debt that limits the scope of overinvestment and perquisites by managers (Jensen (1986), Stulz (1990) and DeAngelo and DeAngelo (2006)). Hence debt is less helpful both in providing capital and in reducing the costs of free cash flow for these firms. Further, debt financing renders firms with little financial flexibility vulnerable to predatory strategies such as price wars by established firms to exhaust lessflexible firms financially (Poitevin (1989)), thus further deteriorating financial flexibility. In addition, debt covenants often carry restrictions on financing and investment decisions 13

that are especially cumbersome for firms with lack of financial flexibility. In short, growing firms with negative retained earnings are constrained in borrowing. At the same time, they face little concern for agency costs of free cash flow. Thus, they can accumulate cash holdings through equity financing in order to increase financial flexibility. Equity issues neither require collateral or restrictive covenants, nor accentuate moral hazard problems that are associated with leverage, nor raise the probability of financial distress. Thus, equity financing provides greater financial flexibility than debt financing for less flexible firms such as small firms with negative retained earnings. Accordingly, we suggest that small growing firms with large negative retained earnings issue more equity, maintain lower leverage, have higher cash holdings and pay no dividend(proposition 2). Thus, propositions 1 and 2 imply lower leverage ratios for both small growth firms and large mature firms, but for different reasons; while large mature firms have been able to finance most their investments with the greatest financial flexibility through internal equity funds, small growing firms have been relying on external equity to preserve financial flexibility for the future. Between these flexibility-sufficient and flexibility-deprived firms are medium firms with mediocre retained earning that are still in much need of external financing. Medium firms generating some positive earnings have better financial flexibility than firms struggling with negative earnings and are able to use debt financing. Also, firms with built-up financial flexibility will use up their financial resources such as debt capacity and accumulated cash when they face abnormal periods characterized by earnings shortfalls over time ((DeAngelo and DeAngelo (2006)). Only when these firms run out of cash and debt capacity, they will turn to equity financing. Based on these arguments, we propose the following financial flexibility hypothesis. The Financial Flexibility Hypothesis Small growing firms and firms with large negative retained earnings are in the most need of financial flexibility and they issue more equity, maintain lower leverage ratios, have higher cash holdings and pay no dividend. Medium firms and firms with mediocre retained earnings have some flexibility to issue debt and meet their capital needs. Large firms and firms with large retained earnings are the most 14

financially flexible as they maintain lower leverage ratios, lower cash holdings and high dividend payouts. We summarize specific predictions based on the implications of the financial flexibility hypotheses in Table I. The hypothesis states that firms with large positive retained earnings are more likely to pay dividend (consistent with the findings in DeAngelo, DeAngelo, and Stulz (2005)), while limiting debt financing. This implies a negative relationship between leverage ratio and dividend payout, consistent with evidence documented in previous studies (Graham (2000), Frank and Goyal (2003), Minton and Wruck (2001) and Byoun (2007)). However, non-dividend-paying firms are most likely firms with low or negative retained earnings and we expect lower leverage for such firms. Thus, the relationship between leverage ratio and dividend payout is not clear within non-paying firms. The novelty of the financial flexibility hypothesis is its prediction of non-linear (inverted V-shaped) relationships between firm size and leverage ratio as well as between retained earnings and leverage ratio. The hypothesis implies that there is a positive relationship between retained earnings and leverage ratio conditional on negative retained earnings, while there is a negative relationship between retained earnings and leverage ratio conditional on positive retained earnings. Also, the hypothesis states that small firms or firms with negative retained earnings are more likely to issue equity, implying a positive relationship between firm size and leverage ratio, conditional on negative retained earnings. However, conditional on positive retained earnings, there is a negative relationships between firm size and leverage ratio. This prediction is contrary to the general finding in the literature, but consistent with the finding in Faulkender and Petersen (2006) who report a negative relationship between firm size and leverage ratio when they consider a sample of firms only with credit ratings. According to the financial flexibility hypothesis, large firms with positive retained earnings are most likely to use only safe debt and hence have better long-term credit ratings. Thus, the contradictory results regarding the relationship between firm size and leverage can be explained by the financial flexibility hypothesis. Faulkender and Petersen (2006) argue that market frictions may cause firms to be rationed by their lenders, leading some firms to appear under-levered relative to unconstrained 15

firms. They use firms long-term credit ratings as a measure of accessibility to the public debt markets (see also Lemmon and Zender (2004) and Byoun (2007a). The market friction argument may imply that firms with better credit ratings should use more leverage than those with worse credit ratings. However, according to our hypotheses, large mature firms with large retained earnings use only safe debt and as a result have better credit ratings. We expect that good credit ratings are associated with large mature firms with lower leverage. Thus, contrary to the implication of the market friction argument, we expect a negative relationship between leverage and credit ratings, conditional on having credit ratings. On the other hand, given that small firms with negative retained earnings are least likely to have credit ratings, we expect a lower leverage for firms with no credit ratings than for those with ratings and hence a positive relationship between leverage ratio and firms size, conditional on no credit ratings. The financial flexibility hypothesis also implies a negative relationship between leverage ratio and cash holdings as firms with lack of financial flexibility increase cash reserves through equity issues. Firms preference for financial flexibility generally depends on the uncertainty around expected earnings and investment opportunities. Issuing risky debt implies the firm loses its financial flexibility. Firms with greater needs for financial flexibility should issue safe debt whose payments are covered by earnings next period. Firms with less expected earnings or much uncertainty about the earnings will be better off issuing equity and/or maintaining high cash balances since issuing equity in the event of a low-profit outcome in the future will bear prohibitively high issuing costs for the firm. Thus, if the probability of low earnings is high, the firm with lack of financial flexibility will abstain from issuing risky debt. Consistent with this argument, firms appear issuing equity before declining earnings (McLaughlin, Safieddine, and Vasudevan (1996), Loughran and Ritter (1997) and Hansen and Crutchley (1990)). The demand for financial flexibility will also increase with expected investment opportunities and their uncertainties since future capital needs are more likely to create financial constraints. Only firms with high probability of great earnings and less expected investment opportunities can safely borrow more against the anticipated profits. In that case, large cash holdings incur high opportunity/agency costs without obtaining greater 16

financial flexibility. Accordingly, we expect that expected earnings (expected investment opportunities) will be associated positively (negatively) with leverage and negatively (positively) with cash holdings, consistent with the documented findings in the literature. We examine these predictions as well. IV. Data The initial sample consists of all available U.S. firms for the period of 1971 2005 from the annual Compustat files. Following previous studies, we exclude financial firms and regulated utilities from the sample. 9 We also require firms to have positive total assets, book and market value of equity and net sales. These variables are used to deflate other variables and it is difficult to interpret the results when they have non-positive values. We also delete observations with missing or non-positive values for the number of shares outstanding and stock price at the end of the fiscal year. Accordingly, we drop bout 8 % of firm-year observations in the sample that does not meet these requirements. After these initial requirements are applied, the sample consists of 179,418 firm-year observations. While Shyam-Sunder and Myers (1999) and Myers (1984) argue that there are rational reasons for managers to specify debt targets in terms of book values, Titman and Wessels (1988) and Welch (2004) are inclined toward the use of debt level measured at market value. Accordingly, we test our hypotheses using total and long-term debt ratios measured with both book and market value of total assets. The definitions of all the variables used in this study are provided in the Appendix. 9 Financial firms are represented by SIC codes 6000-6799 and utilities by SIC codes 4800-4999. These firms have very different capital structures and their financing decisions may not convey the same information as non-financial and non-regulated firms. For example, a relatively high leverage ratio is normal for financial firms, but the same high leverage ratio for non-financial firms may indicate possible financial distress. 17

V. Univariate Analyses A. Firm Size, Credit Ratings and Leverage In order to examine the relationship between firm size and leverage closely, we divide the sample into size deciles each year and report the leverage ratios measured in long-term and total debt to book/market value of assets. 10 We delete all observations with leverage ratios less than zero or greater than one. 11 We define size in three different ways based on book value of total assets, market value of total assets and net sales, but the results are similar and we report only those based on the book value of total assets. Panel A of Table II shows that regardless of the various definitions of leverage ratios, there is a positive monotonic relationship between firm size and leverage ratio especially for smaller size deciles. We also report the percentage of zero-debt firms in each size decile. Small firms are associated with much more zero-debt firms than large firms. Byoun (2007a) suggests that zero-debt firms are constrained by debt market while unconstrained by equity market. This implies that zero-debt firms lack financial flexibility. In order to examine whether the negative relationship between firm size and leverage ratio is driven by these zero-debt firms, we report the results excluding zero-debt firms in Panel B of Table II. Even though there still exists a fairly positive relationship between firm size and leverage ratio without the zero-debt firms, the positive relationship between size and leverage ratio became weaker for the largest three size deciles. Thus, our results confirm that there exists strong positive relationship between firm size and leverage except for firms in the largest three deciles in which the positive relationship is weakened or reversed. 10 The results does not change when we exclude deferred taxes and investment tax credit or include convertible debt (item 79) in the definition of book equity as in Alti (2006) and Kayhan and Titman (2006). 11 Without this requirement, the average book leverage ratio of the sample firms in the first size decile are greater (but market leverage ratios are smaller) than firms in larger size deciles since there are a few firms with book leverage ratios greater than one in the first size decile. When we winsorize leverage ratios at 99 percentile, there still exist firms with leverage ratios greater than one. 18

Table II In Table III, we divide the sample into prior- and post-1985 subperiods. Since S&P credit rating information is available only from 1985, on the one hand, this allows us to examine the distribution of rated firm across size deciles for this period. On the other hand, it allows us to examine if there is any discernible change in the relationship between firm size and leverage. For the period of 1971-1984, the relationship between firm size and leverage is positive and monotonic except for market total debt ratio. On the other hand, for the period of 1985-2005, the positive relationship holds only up to 7th deciles and the relationship becomes weakly negative in the largest three deciles. The results also show that small firms rarely have long-term credit ratings and most ratings are concentrated in the largest three deciles. The sample in Faulkender and Petersen (2006) includes firms with credit ratings that are mainly from the largest three size deciles for the period after 1985, which may explain their finding of a negative association between firm size and leverage. Table III In order to explicitly examine the implication of the financial flexibility hypothesis that there is a negative relationship between leverage and credit ratings, conditional on credit ratings. We examine firms leverage ratios based on long-term credit ratings (Rating) in Table IV. Consistent with the prediction of our hypothesis, the results show that firms with higher credit ratings are larger firms and have lower leverage ratios, confirming a negative relationships between credit rating and leverage ratio, conditional on having credit ratings. There is also a negative association between firm size and leverage ratio for rated firms, but unrated firms tend to be smaller and have lower leverage ratios on average than firms with credit ratings. Table IV 19

B. Firm Size, Cash Holdings, Retained Earnings, Dividends and External Financing Activities In order to examine weather the lower leverage for small firms results from accumulated internal equity (as suggested by the pecking order theory) or external equity (as suggested by the financial flexibility hypothesis), we report retained earnings, net long-term debt issue, net total debt issue and net new equity issue as proportions of total assets. We also examine the ratio of cash and marketable securities to total assets. 12 We drop observations with missing values in any of the reported variables. Table V reports the results. The results in Panel A show that small firms tend to have more cash holdings while having less retained earnings than large firms. In fact, the average retained earnings are negative for firms in smaller size deciles. Thus, small firms growth is not likely to come mainly from internally generated equity. Small firms long-term or total debt financing is minuscule compared to that of large firms, while their equity financing is phenomenal. On average firms in the first and second size deciles issue equity for 25% and 12% of total assets per year, respectively. There are strong positive and monotonic relationship between firm size and dividend payouts except for the first size decile firms whose dividend payouts are greater than firms in the second size decile. Table V The results in Table V can be driven by IPO firms that are more likely to be in small size deciles. In order to examine the IPO effect, we identify the IPO date from Compustat and designate the first fiscal year ending after the IPO date as a IPO year. We also identify the first year appearing in the Compustat for those that do not have IPO dates but the Compustat begins its coverage during our sample period and treat it like the IPO year. The results excluding these IPO years are reported in Panel B. They show that the magnitude of external equity raised by small firms becomes smaller after excluding IPO years, but it 12 Including accounts receivable in addition to cash and marketable securities produces almost identical results. 20

is still significantly greater than that raised by larger firms. There is another possibility that the results could be driven by a few outliers especially in small size deciles. To address this concern we reproduce results with exclusion of the outliers of net equity financing variable at 1st and 99th percentiles. Again the results in Panel C show that small firms heavily rely on external equity with little debt. The pattern remains intact but only with less magnitudes when we exclude observations with greater cutoff percentiles of the equity financing variable. The dividend payouts of the first decile firms are greater than those in the second through fifth size deciles. Byoun (2007a) find that small zero-debt firms pay higher cash dividends and issue much more equity than other firms. Since zero debt firms are concentrated in the smaller size decile, the higher average dividend payout for the first decile can be driven by zero-debt firms. Accordingly, we estimate the average dividend payouts for non-zero debt firms without zero-debt firms and find that the higher dividend payouts for the first decile firms are indeed due to high dividend payouts of zero-debt firms. Overall, firm size is negatively associated with cash holdings and equity financing and positively associated with retained earnings, debt issue and dividend payout. Thus, small firms have lower leverage ratios, not because of internally generated funds but because of additional equity financing. Small firms also build up cash holdings through external equity in order to preserve financial flexibility, consistent with our financial flexibility hypothesis. Firm size and retained earnings are highly correlated, making it difficult to examine the effect of retained earnings. Accordingly, we address this issue below. C. Firm Size, Retained Earnings and Leverage In order to disentangle the relationship between retained earning and leverage ratios while accounting for the strong association of firm size with retained earnings, we first examine the leverage ratios for firms divided into negative and positive retained earnings groups within each size decile. Panel A of Table VI shows cash holdings and leverage ratios for each group. The results show that the smaller firms (in size deciles below 6) with negative retained earnings hold 21

more cash balances than similar size firms with positive retained earnings as a means of preserving financial flexibility. Most differences are statistically significant except for the first two deciles (not reported in the table). On the other hand, large firms with negative retained earnings tend to carry less cash balances with higher leverage ratios than large firms with positive retained earnings. Large firms appear to use preserved debt capacity and cash balances to counteract to negative earnings shocks, consistent with our financial flexibility hypothesis. The market value leverage ratios for firms with negative retained earnings are always smaller than book value leverage ratios because negative retained earnings tend to decrease the book value of total assets. Since the portion of negative retained earnings relative to total assets are significantly greater for small firms, smaller firms (in 1 to 4 size deciles) with negative retained earnings have higher book leverage ratios whereas lower marketvalue leverage ratios than firms with positive retained earnings in the same size deciles. Accordingly, the relationship between firm size and leverage within smaller size deciles (lower than decile 5) can be affected by whether the leverage is measured in book value or market value because of the significant number of small firms with negative retained earnings. For this reason, firms with negative retained earnings have higher book leverage ratios despite their heavy reliance on equity financing as shown below. Thus, the higher leverage ratios for small firms with large negative retained earning result not from debt issues but from negative retained earnings. Panel B of Table VI shows that within the same size decile firms with negative retained earnings pay less dividend and issue substantially more equity than those with positive retained earnings. For example, for deciles 2 through 7, firms with negative retained earnings issue four to five time as much equity as the firms with positive retained earnings do. The results suggest that larger equity issues of small firms are driven by firms with negative retained earnings as they issue equity to raise cash as a means of of preserving financial flexibility. Firms with negative retained earnings issue less debt than firms with positive retained earnings in most size deciles except for the largest size decile in which the difference is not significant. Clearly, firms with negative retained earnings issue both debt and equity, 22

but their equity issues, especially for small firms, are significantly greater than those issued by firms with positive retained earnings. This finding is consistent with hypotheses 2 that firms with negative retained earnings pay less dividend and issue more equity to preserve financial flexibility. Consistent with the financial flexibility hypothesis, the results in Table V show that small firms with negative retained earnings hold more cash balances and pay less dividends while issuing more equity than those with positive retained earnings as a means of preserving financial flexibility. However, large firms with negative retained earnings, while paying less dividends and issuing more equity, have less cash balances and higher leverage ratios than those with positive retained earnings, again consistent with the financial flexibility hypothesis. Table VI D. Retained Earnings, Firm Size, Credit Ratings, Dividend and Leverage The financial flexibility hypothesis states that firms with large negative retained earnings use equity financing, firms with mediocre retained earnings use debt financing, and firms with large positive retained earning use internally generated equity financing. In order to test this, we divide the sample into deciles based on retained earnings and examine firms financing activities, dividend payouts and leverage ratios in Tables VII. Table VII We report retained earnings, total assets, and various leverage ratios for each decile in Panel A. An examination of leverage ratio across retained earnings deciles reveals that there is indeed a inverted V-shape relationship between retained earnings and leverage ratios, consistent with the implication of the financial flexibility hypothesis. There is also a similar inverted V-shape relationship between firm size and leverage, that is, a negative relationship between leverage ratio and firm size across the upper five retained earnings 23

deciles and a positive relationship between leverage ratio and firm size across the other lower deciles. The results in Panel B provide further supporting evidence for our hypotheses. Firms with large negative earnings in lower deciles hold the largest cash balances by issuing much more equity than debt. Largest firms in the middle deciles (see Panel B in Table VII) hold lowest cash balances, issuing more debt than firms in other deciles. As we move from decile 7 to upward, we can observe that firms reduce both debt and equity financing while increasing dividend. Firms in the largest retained earnings decile actually reduce both debt and equity. Overall, the results in Tables VII are very intriguing and surprisingly consistent with our financial flexibility hypotheses. V. Multivariate Analyses E. Estimation Results from Regression Model In this section, we directly test the implications of the financial flexibility hypotheses that are provided in Table I with the following regression models: LEV = α 0 + α 1 RE D +Re + α 2 RE D Re + α 3 Cash + α 4 LnA + α 5 Div +α 6 MB + α 7 RND + α 8 D 0RD + α 9 OI +α 10 MOI + Other V ariables + ε (1) LEV = α 0 + α 1RE D +Re + α 2RE D Re + α 3D No + α 4Rating +α 5Cash + α 6LnA + α 7Div + α 8MB + α 9RND + α 10D 0RD + α 11OI +α 12 MOI + Other V ariables + ε (2) The detailed definition of each variable is provided in the Appendix. Equation (2) include additional variables, a dummy variable for firms with no long-term credit ratings (D No ) and long-term credit ratings (Rating) numbered 0 for non-rated firms and from 1 for the lowest rating (below B) to 7 for the highest rating (AAA) (See Table IV). Since information on credit ratings is available only from 1985, the estimation of equation (1) is restricted 24

to the sample of 1985-2005 period. The variables that are of main interest in testing the implications of the financial flexibility hypotheses given in Section III are firm size (LnA), the interaction variables between retained earnings (RE) and dummy variables for positive (D +Re ) and negative retained earnings (D Re ),. We expect, positive coefficient estimates for D Re and negative estimates for D +Re, D No and Rating based on the signs given in Table I. Given the inverted V-shape relationship between firm size and leverage, we expect positive coefficient estimates for LnA within financially constrained groups such as small size firms, firms with negative retained earnings, non-dividend paying firms, and non-rated firms, whereas we expect negative coefficient estimates for the other non-constrained groups. We also include dividend Div even though the as Table I also provides a negative prediction for the relationship between dividend and leverage ratio, this prediction is consistent with existing evidence. Also, expected earnings and investment opportunities need to be proxied by some measures. Market-to-book assets ratio (M B), research and development expenditures (RND) and a dummy variable equal to one for missing value for RND variable (D 0RD ), are typically used in previous studies as measures of a firm s investment opportunities and are known to be negatively associated with leverage ratio, consistent with our prediction. For expected earnings, we use operating income (OI) and industry median operating income (M OI) as proxies for expected earnings. Other variables include a dummy variable equal to one for the zero debt year (Zero), dummy variable for IPO year (IP O), Fixed assets (F A), Altman s Z-score (AZ), depreciation and amortization (Dep), marginal tax rate (T ax), operating income (OI) and industry median debt ratio (M ed). Following the common practice in handling the outliers in the literature, we winsorize variables at 1st and 99th percentiles except for those that are bounded by 0 from below, in which case we winsorize only at 99th percentile. Table VIII Panel A in Table VIII reports two sets of estimation results for each dependent variable for the entire time period and for the period of 1985-2005 which also includes credit rating 25

variables. The coefficient estimates on firm size (LnA) are highly significant and positive in all regressions. Thus, there is a fairly strong positive relationship between leverage and firm size even after controlling for the additional variables. The coefficient estimates on positive retained earnings (RE D +Re ) are all significant and negative as predicted by the financial flexibility hypothesis. On the other hand, the coefficient estimates on negative retained earnings (REḊ Re ) are mixed: they are significant and positive with total debt ratios but either insignificant or negative with market value long-term debt ratio. This result could be driven by the fact that debt ratios for firms with negative retained earnings may have been inflated with market value debt ratio. But the coefficient estimates on REḊ Re are economically insignificant compared to those on REḊ+Re. This result suggests that firms with large positive retained earnings have significantly lower leverage. Also, firms holding more cash balances, consistent with the financial flexibility hypothesis. IPO firms and zerodebt firms tend to have lower leverage as expected. All other variable are as expected and consistent with previous results. In order to further examine the non-linear relationship between firm size and leverage ratio, conditional on those proxies for financial flexibility, we divide firm into two groups based on each variable and estimate separate regressions. We also do this for samples divided into before and after 1985 with and with credit rating variables. Since the results are very similar, report to different sets of results between small and large firms in Panel B and between rated and non-rated firms for the period of 1985-2005. For small and large groups, we divide the sample by deciles greater than and less than or equal to decile 5. By estimating the regressions for small and large firms separately, we can also examine the non-linear relationship between retained earnings and leverage ratio. The estimation results in Panel B show that the coefficient estimates on size (LnA) are all significant and positive for small firms but negative and significant for large firms. Thus, the positive relationship between firm size and leverage holds true only for small firms. This finding strongly support our financial flexibility hypothesis. The coefficient estimates on positive retained earnings (RE D +Re ) are all significant and negative, while the coefficient estimates on negative retained earnings (REḊ Re ) are all positive and significant for large 26

firms as predicted by the financial flexibility hypothesis. It appears that firm size better represents the financial flexibility constraint for small firms while retained earnings and firms size together seems catching large firms super financial flexibility. The coefficient estimates on dividend (Div) are highly significant and negative for large firms but insignificant for small firms. Large firms with large internal funds seem paying more dividend while maintaining low leverage, consistent with the financial flexibility hypothesis. We report the results for groups divided by rated and unrated firms for the period of 1985-2005. Again, the results for rated firms are very consistent with our financial flexibility hypothesis, while the results for non-rated firms give some mixed results. Overall, the regression results show support evidence for our financial flexibility hypothesis that the positive relationship between leverage ratio and firm size hold only for small firms but there is a strong negative relationship between leverage ratio and firms size for large firms. We find similar non-linear relationship between leverage ratio and firm size when we divide firms based on credit rating and non-rating, dividend paying and nonpaying, or positive and negative retained earnings. This finding is very significant because previous studies overlooked such non-linear relationship between leverage ratio and firm size. The non-linear relationship between leverage and firm size suggest that small firms lower leverage ratios result from their concern for financial flexibility (issuing equity and building up cash holdings) and large firms with positive retained earnings also have lower leverage as they mostly rely on internal funds while preserving debt capacity. VI. Robustness Check IN PROGRESS VI. Summary and Conclusions We examine the relationship between firm size and leverage in the view of financial flexibility. We define financial flexibility as the extent to which a firm can mobilize its financial resources to take preventive and exploitive actions in response to subsequent information in order to maximize the firm value. 27

We hypothesize that small growing firms and firms with large negative retained earnings are in the most need of financial flexibility and they issue more equity, maintain lower leverage ratios, have higher cash holdings and pay no dividend. Medium firms and firms with mediocre retained earnings have some flexibility to issue debt and meet their capital needs. Large firms and firms with large retained earnings are the most financially flexible as they maintain lower leverage ratios, lower cash holdings and high dividend payouts. Consistent with our hypothesis, firms with negative retained earnings issue several times more equity than firms with positive retained earnings. While small firms avoid debt financing, they are much more active in tapping into external equity capital. We also find that small firms often have negative retained earnings with no less cash holdings than other firms and that small firms with negative retained earnings have lower leverage than firms with positive retained earnings. We further show that firm size has an overall strong and significant positive association with leverage. However, the positive relationship between firm size and leverage are reversed for large firms when we divide the firms into small and large group, credit rating and non-rating, dividend paying and non-paying, or positive and negative retained earnings. Our regression results, coupled with univariate results, suggest that small firms with negative retained earnings build up cash holdings through equity financing, lowering leverage ratios, whereas large firms with positive retained earnings rely on internal equity for their capital needs and still pay dividends resulting in lower leverage ratios. Thus, the relationship between leverage and firm size is not linear this is what is predicted by the financial flexibility hypothesis. Overall, small firms have lower leverage ratios, not because of internally generated funds or additional debt financing but because of additional equity financing. Small firms also build up cash holdings in order to preserve financial flexibility through external equity. This finding can be explained by neither of the pecking order theory and the tradeoff theory the pecking order may be reversed for small firms that prefer external equity to debt financing while the tradeoff theory may miss out some important aspects of capital structure decisions. On the other hand, our study brings new insights into several unresolved 28

issues in the capital structure literature. For example, why do larger firms appear to provide a better fit for the pecking order theory (Shyam-Sunder and Myers (1999) and Frank and Goyal (2003)) despite the fact that large firms are less subject to information asymmetry than small firms? Our finding suggests that large firms prefer using internal funds to preserve financial flexibility. On the other hand, small firms issue equity and increase cash holdings despite having low leverage in order to cope with the lack of financial flexibility, thus reversing the external financing hierarchy suggested by the pecking order theory. This also answers why most equity issuances occur when firms have sufficient debt capacity (Fama and French (2002) and Leary and Roberts (2005a)), without any apparent risk of entering financial distress from issuing debt. (Leary and Roberts (2004)) Our findings may also have bearings on Welch (2004) who observe that firms do not use their issuing activities to counteract the external and large influences of stock returns on their capital structures. In conclusion, asymmetric information falls short of providing a plausible explanation for motivation behind firms external financing decisions. The benefits and costs associated with financial flexibility influence firms capital structure decisions but not in the manner hypothesized by the traditional trade-off theory. Thus, a substantial alteration may be required to the tradeoff argument. Future study should address the crux of the financial flexibility, how uncertainty affects a firm s financial decisions. 29

Appendix. Variable Definitions Total assets = Compustat item 6; Net sales = item 12; The number of shares outstanding = item 25; Stock price at the end of the fiscal year =item 199; Accounts receivable =item 2; Net long-term debt issue = item 111 - item 114; Net total debt issue = item 111 item 114 item 301 if item 318 = 1 and item 111 item 114 + item 301, otherwise. Changes in current debt (item 301) represent an increase in working capital for format code 1 but a decrease in working capital for format codes; 13 1 to 3, OCF equals item 123 + item 124 + item 125 + item 126 + item 106 + item 213 + item 217 + item 218. For firms reporting format code 7, OCF equals iitem 123 + item 124 + item 125 + item 126 + item 106 + item 213 + item 217 + item 314; 14 13 When we evaluate the Statement of Cash Flows or Changes in Financial Position for any company in Compustat, we first consider Format Code (item 318). This is important because the format code directs us to the data that are available for a particular company. Prior to the adoption of the Statement of Financial Accounting Standards (SFAS) #95 for U.S. companies and currently for foreign companies, the format code may have changed from one year to the next depending on the manner in which a company reported its data. Effective for fiscal years ending July 15, 1988, the SFAS #95 required U.S. companies to report the Statement of Cash Flows (format code = 7). Prior to the adoption of SFAS #95, companies had the option of reporting any one of the following: 1) Working Capital Statement; 2) Cash Statement by Source and Use of Funds; or 3) Cash Statement by Activity. These formats were specified beginning in 1971, which is the reason our sample period begins with this year. (See Compustat manual.) 14 Following Frank and Goyal (2003), we treat missing values that are not reported or combined with other data items in the definition of OCF as zero. 30

Net equity issues = item 108 item 115; Ratio of cash and marketable securities to total assets =[item 162 + item 193] / item 6; AZ = Altman s Z-score modified by MacKie-Mason (1990): (3.3EBIT (item 178) + sales (item 12) + 1.4 retained earnings (item 36) + 1.2 working capital (item 4 - item 5)) divided by total assets (item 6). Altman s Z-score measures the ex ante probability of distress (Graham (1996, 2000)); Cash = Cash and short-term investments (item 1) ; Dep = depreciation and amortization (item 14) as a proportion of total assets. Firms with more depreciation expenses have less need for the interest deductions associated with debt financing; DIV = common stock dividends (item 127) divided by total assets. DIV controls for possible trade-off between debt and dividend in reducing agency costs of free cash flow (Fama and French (2002)); D 0Div = dummy variable equal to one if the firm has missing values for common stock dividends (item 127) and zero otherwise; E it = net equity issues for firm i from time t 1 to t: item 108 item 115; F A = fixed assets (item 8) divided by total assets. Firms operating with greater tangible assets have a higher debt capacity; IP O = dummy variable equal to one for IPO year and zero otherwise; LnA = log of total assets (item 6) as a measure of firm size. Larger firms tend to: have more leverage (perhaps because they are more transparent); have lower asset volatility; or naturally sell large enough debt issues so that the fixed costs of public borrowing are not prohibitive; 15 15 The results are not affected whether the size is defined in terms of market value of assets or of net sales 31

MB = market-to-book ratio of assets. The market value of assets (MV ) equals total assets (item 6) minus total equity (item 216) minus balance sheet deferred taxes and investment tax credit (item 35) plus the market value of common equity (price (item 199) times shares outstanding (item 54)) plus preferred stock liquidating value (item 10, replaced by the redemption value of preferred stock (item 56) when missing). 16 A higher MB is generally taken as a sign of more attractive future growth options, which a firm tends to protect by limiting its leverage; Med = industry median debt ratio based on two-digit SIC (or Fama and French (2002) industry groupings). According to Frank and Goyal (2004), the industry median leverage is an important determinant of a firm s leverage ratio, acting as a proxy for several factors, including intangibility, regulation, stock variance, uniqueness, purchasing manager s sentiment index, etc.; OI = operating income (item 13) divided by total assets (item 6). A firm with higher earnings could prefer to operate with either lower or higher leverage. Lower leverage might occur, as higher retained earnings mechanically reduce leverage, or if the firm limits leverage to protect the franchise responsible for producing these high earnings. Higher leverage might reflect the firm s ability to meet debt payments out of its relatively high earnings cash flow; M OI = Median industry operating income (item 13) divided by total assets (item 6) based on two-digit SIC code; Rating = numeritized long-term credit rating (item 280). If item 280 =2 (AAA), then Rating = 7. If item 280 = 4 (AA+), 5 (AA) or 6 (AA ), then Rating = 6. If item 280 = 7 (A+), 8 (A) or 9 (9A ), then Rating = 5. If item 280 = 10 (BBB+), 11 (BBB) or 12 (BBB ), then Rating = 4. If item 280 = 13 (BB+), 14 (BB) or 15 (item 12). 16 The results do not change when we exclude deferred taxes and investment tax credit or include convertible debt (item 79) in the definition of book equity (as in Alti (2006) and Kayhan and Titman (2006)). 32

(BB ), then Rating = 3. If item 280 = 16 (B+), 17 (B) or 18 (B ), then Rating = 2. For all other ratings, Rating = 1. For unrated firms, Rating = 0. D No dummy variable equal to one if the firm has missing values long-term credit ratings (item 280) and zero otherwise; Following Faulkender and Mitchell (2003) we use firms long-term credit ratings (item 280) as a measure of accessibility to the public debt markets. Fualkender and Mitchell (2003) and Lemmon and Zender (2004) find that leverage ratios of firms with credit ratings are significantly higher than firms without ratings. RE = retained earnings (item 36) divided by total assets. DeAngelo, DeAngelo, and Stulz (2005) argue that firms with low RE/TE (RE/TA) tend to be in the capital infusion stage, whereas firms with high RE/TE (RE/TA) tend to be more mature with ample cumulative profits that make them largely self-financing, hence good candidates to pay dividends. It is also a better measure of a firm s lifecycle stage (hence suitability to pay dividends) than its cash balances, because the source of the cash impacts the dividend decision. For example, high cash holdings can reflect the proceeds of a recent equity offering for a firm whose low RE/TE and RE/TA show it to be in the infusion rather than the distribution stage; D +Re = dummy variable equal to one for the firm with positive retained earnings (item 36) and zero otherwise; D Re = dummy variable equal to one for the firm with negative retained earnings (item 36) and zero otherwise; RN D = research and development expenditures (item 46) divided by net sales (item 12). RND can be taken as a proxy for future expected investment (Fama and French (2002)). They also serve as an additional proxy for non-debt tax shields. Following previous studies (e.g., Opler, Pinkowitz, Stulz, and Williamson (1991), Loughran and Ritter (1997), Fama and French (2002), Kayhan and Titman (2003) and Leary and Roberts (2004)), we assume that missing values for the R&D variable are zero.; 33

D 0RD = dummy variable that equals one for firms with missing RND and zero otherwise; T ax = marginal tax rate equal to the statutory tax rate if the firm reports no net operating loss carryforwards (item 52) with positive pretax return (item 170) and zero otherwise. The statutory taxes are 48% from 1971 to 1978, 46% from 1979 to 1986, 40% in 1987, 34% from 1988 to 1992, and 35% from 1993 to 2003. Plesko (2003) shows that this binary measure captures the marginal tax effects; Zero = dummy variable equal to one for the year with zero debt and zero otherwise; 34

References Almeida, Heitor and Murillo Campello, 2005, Firm financing-investment interactions: Evidence from debt and equity issues, New York University and University of Illinois Working paper. Alti, Aydogan, 2006, How persistent is the impact of market timing on capital structure, Journal of Finance 61, 1681-1710. Altinkilic O. and R.S. Hansen, 2000, Are there economies of scale in underwriter fees? Evidence of rising external financing costs, Review of Financial Studies 13(1), 191 218. Baker, Malcolm and Jeffrey Wurgler, 2002, Market timing and capital structure, Journal of Finance 57, 1 32. Barclay, Michael J. and Clifford W. Smith, 2005, The capital structure puzzle: The evidence revisited, Journal of Applied Corporate Finance 17(1), 8 17. Byoun, Soku, 2007, How and When Do Firms Adjust Their Capital Structures toward Targets? Journal of Finance, Forthcoming. Byoun, Soku, 2006, Why do some firms go debt-free? Baylor University Working Paper. DeAngelo H. and L. DeAngelo, 2006, Capital Structure, Payout Policy, and Financial Flexibility, University of Southern California working paper. DeAngelo, Harry, Linda DeAngelo, and Ren M. Stulz, 2005, Dividend Policy and the Earned/Contributed Capital Mix: A Test of the Lifecycle Theory, Journal of Financial Economics. DeAngelo, Harry, Linda DeAngelo and Rene M. Stulz, 2007, Fundamentals, market timing, and seasoned equity offering, National Bureau of Economics Research working paper 13285. 35

Evans, D. 1987, Tests of alternative theories of firm growth, Journal of Political Economy 95, 657-674. Fama, E. F. and French, K. R., 2002, Testing trade-off and pecking order predictions about dividends and debt, Review of Financial Studies 15(1), 1 33. Fama, Eugene F. and Kenneth R. French, 2005, Financing decisions: Who Issues Stock?, Journal of Financial Economics 76, 549 582. Faulkender, Michael, and Mitchell A. Petersen, 2006, Does the source of capital affect capital structure?, Review of Financial Studies 19, 45 79. Frank, M. Z. and Vidhan K. Goyal, 2003, Testing the pecking order theory of capital structure, Journal of Financial Economics 67(2), 217 248. Frank, M. Z. and Vidhan K. Goyal, 2005, Trade-off and Pecking Order Theories of, B. Espen Eckbo(ed.), Handbook of Corporate Finance: Empirical Corporate Finance (Handbooks in Finance Series, Elsevier/North-Holland), Chapter 7. Goldstein, R. N. Ju, and H. Leland, 2001, An EBIT-based Model of Dynamic Capital Structure, Journal of Business 74, 483 512. Graham, John R. and Campbell R. Harvey, 2001, The theory and practice of corporate finance: evidence from the field, Journal of Financial Economics 61, 187 243. Hansen, R. and C. Crutchley, 1990. Corporate earnings and financings: An empirical analysis. Journal of Business 63, 347-371. Hovakimian, Armen, Tim C. Opler, and Sheridan Titman, 2001, The debt-equity choice: An analysis of issuing firms, Journal of Financial and Quantitative Analysis 36(1), 1 24. Hovakimian, Armen, Gayane Hovakimian and Hassan Tehranian, 2004, Determinants of target capital structure: The case of dual debt and equity issues, Journal of Financial Economics 71, 517 540. 36

Jensen, M.C., 1986, Agency costs of free cash flow, corporate finance and takeovers, American Economic Review 76, 323-339. Jensen, M.C. and W. Meckling, 1976, Theory of the firm: Managerial behavior, agency costs, and capital ttructure, Journal of Financial Economics 3, 305-360. Jung, Kooyul, Yong-Cheol Kim, and Rene M. Stulz, 1996, Timing, investment opportunities, managerial discretion, and the security issue decision, Journal of Financial Economics 42, 159 185. Kale, Jayant and Husayn Shahrur, 2007, Corporate capital structure and the characteristics of suppliers and customers, Journal of Financial Economics 83, 321-365. Kayhan, Ayla and Sheridan Titman, 2006, Firms histories and their capital structures, Journal of Financial Economics, forthcoming. Kim, Woojin and Michael Weisbach, 2007, Motivations for public equity offers: An international perspective, Journal of Financial Economics, Forthcoming. Kisgen, Darren J., 2006, Credit Ratings and Capital Structure, Journal of Finance 61 (3), 1035-1072. Leary, Mark T., Micheal R. Roberts, 2004, Financial slack and tests of the pecking order s financing hierarchy, Duke University working paper. Leary, Mark T. and Michael R. Roberts, 2005, Do firms rebalance their capital structures? Journal of Finance 60, 2575 2619. Leary, Mark T. and Michael R. Roberts, 2005a, The Pecking Order, debt capacity, and information asymmetry, Duke University and University of Pennsylvania working paper. Lemmon, Michael L. and Jaime F. Zender, 2004, capacity and tests of capital structure theories. University of Utah and University of Colorado working paper. 37

Loughran, T. and J. Ritter, 1997. The operating performance of firms conducting seasoned equity offerings. Journal of Finance 52, 1823-1850. McLaughlin, R., Safieddine, A., and G. Vasudevan, 1996. The operating performance of seasoned equity issuers: Free cash flow and post-issue performance, Financial Management 25, 41-53. McLean, R. David, 2007, Share Issuance and Cash Holdings: Evidence of Market Timing or Precautionary Motives? University Alberta working paper. Myers, S.C., 1984, The Capital structure puzzle, Journal of Finance 39, 575 592. Myers, S.C. and N.S. Majluf, 1984, Corporate financing and investment decisions when firms have information that investors do not have, Journal of Financial Economics 13, 187 221. Pagano, Marco, Fabio Panetta, and Luigi Zingales, 1998, Why do companies go public? An empirical analysis, Journal of Finance 53, 27 64. Parsons, Chris and Sheridan Titman, 2007, Capital Structure and Corporate Strategy, University of Texas working paper. Poitevin, M., 1989, Financial signalling and the deep-pocket argument, Rand Journal of Economics 20, 26-40. Rajan, R.G. and L. Zingales, 1995, What do we know about capital structure? Some evidence from international data, Journal of Finance 50, 1421 1460. Shyam-Sunder, L. and S.C. Myers, 1999, Testing static trade-off against pecking order models of capital structure, Journal of Financial Economics 51, 219 244. Titman, S. and R. Wessels, 1988, The determinants of capital structure choice, Journal of Finance 43, 1 19. Welch, Ivo, 2004, Capital structures and stock returns, Journal of Political Economy 112, 106 131. 38

Table I. Financial Flexibility Considerations and Leverage Decisions Financial Decisions Firm size conditional on positive retained earnings, long-term credit rating or dividend paying. (Theses firms are likely to use them to pay off existing debt for financial flexibility, while small firms invest in their growth option) Firm size conditional on negative retained earnings, no credit rating, or no dividend paying (small firms issue equity to increase cash and lower the leverage, while large firms utilize built-up cash and debt capacity) Positive retained earnings (As a firm accumulates more positive earnings, it will increase dividend and reduce leverage) Negative retained earnings (As a firm accumulates large negative earnings, it will reduce dividend, issue equity and hold more cash) No Long-term credit rating (Firms with no credit rating are most likely to have negative retained earnings without access to bond market and have lower leverage than firms with credit ratings) Long-term credit ratings (Firms with higher credit ratings maintain low leverage than firms with lower ratings) Leverage - + - + - - Cash Holdings (Firms with little flexibility hold high cash balances with low leverage) - Dividend (Dividend paying firms maintain low leverage) -/+ Expected earnings (Add flexibility to firm s financing resources allowing to hold less cash, issue more debt and pay dividend) + Expected investment opportunities (Increase the need for financial flexibility) -

Table II. Firm Size Deciles and Leverage Ratios The data consists of 179,418 firm-year observations for the period 1971-2005. Observations with missing values in any of the reported variables are deleted. Size is size deciles based on total assets. Book (Market) Long-term is long-term debt over book (market) value of total assets and Book (Market) Total is total debt over book (market) value of total assets. The market value of assets equals total assets minus total equity minus balance sheet deferred taxes and investment tax credit plus the market value of common equity plus preferred stock liquidating value. % of Firms with Zero is the percentage of firms relative to the total number of firms in each size decile. A zero-debt firm is a firm with no debt. A. All firm-year observations (179,418) Size Decile % of Firms with Zero Total Assets Book Longterm Book Total Market Longterm Market Total 1 3.62 0.0974 0.2074 0.0590 0.1164 0.2035 2 10.39 0.1241 0.2184 0.0943 0.1647 0.1578 3 21.66 0.1356 0.2186 0.1137 0.1828 0.1487 4 40.62 0.1452 0.2170 0.1295 0.1936 0.1480 5 71.75 0.1651 0.2286 0.1482 0.2062 0.1293 6 128.62 0.1916 0.2493 0.1687 0.2217 0.0994 7 240.64 0.2188 0.2725 0.1890 0.2378 0.0744 8 496.17 0.2437 0.2942 0.2089 0.2537 0.0445 9 1354.40 0.2476 0.2968 0.2125 0.2552 0.0299 10 15673.95 0.2387 0.3017 0.2138 0.2679 0.0063 B. Non-zero-debt firm-year observations (155,435) Size Decile Total Assets Book Longterm Book Total Market Longterm Market Total 1 3.67 0.1223 0.2604 0.0741 0.1462 2 10.26 0.1474 0.2593 0.1120 0.1955 3 21.17 0.1593 0.2568 0.1335 0.2147 4 39.21 0.1704 0.2547 0.1520 0.2273 5 68.78 0.1896 0.2626 0.1702 0.2368 6 123.99 0.2127 0.2768 0.1873 0.2461 7 237.02 0.2364 0.2944 0.2042 0.2569 8 490.26 0.2550 0.3079 0.2187 0.2655 9 1347.74 0.2553 0.3060 0.2190 0.2630 10 15694.31 0.2402 0.3036 0.2151 0.2696

Table III. Firm Size Deciles and Leverage Ratios for Sub-periods Divided into Before and After 1985 The data consist of firm-year observations for the period 1971-2005. Observations with missing values in any of the reported variables are deleted. Size is size deciles based on total assets. Book /Market Long-term/Total is longterm/total debt over book/market value of total assets. The market value of assets equals total assets minus total equity minus balance sheet deferred taxes and investment tax credit plus the market value of common equity plus preferred stock liquidating value. % of Firms with Zero is the percentage of firms with no debt relative to the total number of firms in each size decile. % of Firms with Bond Rating is the percentage of firms with long-term credit ratings relative to the total number of firms in each size decile. A. For 1971 1984 Period (53,702 Obs) Size Decile Market Long-term % of Firms with Zero Total Assets Book Longterm Book Total Market Total 1 4.19 0.1132 0.2046 0.0873 0.1525 0.1734 2 9.75 0.1481 0.2350 0.1345 0.2123 0.1034 3 17.06 0.1727 0.2552 0.1658 0.2453 0.0812 4 27.68 0.1923 0.2695 0.1939 0.2719 0.0594 5 44.14 0.2035 0.2750 0.2031 0.2754 0.0540 6 71.33 0.2193 0.2873 0.2189 0.2880 0.0352 7 122.58 0.2315 0.2941 0.2293 0.2926 0.0374 8 239.36 0.2326 0.2893 0.2322 0.2884 0.0239 9 631.22 0.2470 0.3008 0.2488 0.3016 0.0121 10 4276.30 0.2505 0.3044 0.2638 0.3172 0.0019 B. For 1985 2005 Period (119,813) Size Decile Market Long-term % of Firms with Zero % of Firms with Credit Rating Total Assets Book Longterm Book Total Market Total 1 3.36 0.0903 0.2087 0.0463 0.1003 0.2170 0.0000 2 10.67 0.1134 0.2110 0.0763 0.1433 0.1822 0.0004 3 23.72 0.1190 0.2023 0.0903 0.1548 0.1790 0.0006 4 46.41 0.1241 0.1934 0.1006 0.1586 0.1877 0.0022 5 84.14 0.1479 0.2078 0.1236 0.1751 0.1631 0.0128 6 154.28 0.1792 0.2323 0.1461 0.1920 0.1282 0.0551 7 293.57 0.2131 0.2628 0.1709 0.2133 0.0910 0.1533 8 611.28 0.2486 0.2963 0.1985 0.2382 0.0538 0.3279 9 1678.72 0.2479 0.2950 0.1962 0.2343 0.0379 0.5758 10 20778.29 0.2334 0.3005 0.1914 0.2458 0.0083 0.7800

Table IV. Credit Ratings, Firm Size, and Leverage Ratios The data consist of firm-year observations for the period 19851-2005. Observations with missing values in any of the reported variables are deleted. Credit ratings are Standard and Poor s long-term credit ratings. Book /Market Longterm/Total is long-term/total debt over book/market value of total assets. The market value of assets equals total assets minus total equity minus balance sheet deferred taxes and investment tax credit plus the market value of common equity plus preferred stock liquidating value. Dividend is cash dividend divided by total assets. N is the number of observations. Credit Ratings N Total Assets Book Long-term Book Total Market Long-term Market Total Dividend Before 1985 50844 4.4917 0.2668 0.1996 0.1953 0.2592 0.0142 0 (Unrated) 80476 4.0763 0.2195 0.1452 0.1150 0.1714 0.0081 1 (Below B) 610 6.2307 0.4756 0.3422 0.2983 0.4147 0.0022 2 (B) 3560 6.2826 0.4883 0.4413 0.3648 0.4064 0.0035 3 (BB) 4510 7.0817 0.4009 0.3635 0.2957 0.3277 0.0058 4 (BBB) 5393 8.0320 0.3133 0.2677 0.2182 0.2569 0.0135 5 (A) 4799 8.4704 0.2784 0.2122 0.1597 0.2113 0.0219 6 (AA) 1611 8.8645 0.2250 0.1608 0.1175 0.1609 0.0293 7 (A) 359 9.9890 0.1788 0.1017 0.0632 0.1084 0.0371

Table V. Firm Size Deciles, Cash Holdings, Retained Earnings, Dividend and External Financing Activities The data consist of firm-year observations for the period 1971-2005. Observations with missing values in any of the reported variables are deleted. Size deciles are based on total assets. All the variables are reported as a proportion of total assets. A. All firm-year observations (173,515) Size Decile Cash and Equivalents Retained Earnings Net Longterm Issue Net Total Issue Net New Equity Issue Dividend 1 0.4141-3.8807 0.0023-0.0004 0.2527 0.0077 2 0.4012-0.8725 0.0024 0.0027 0.1164 0.0068 3 0.3989-0.4103 0.0007 0.0025 0.0852 0.0078 4 0.3958-0.1390 0.0013 0.0024 0.0777 0.0083 5 0.3725 0.0244 0.0044 0.0059 0.0604 0.0093 6 0.3475 0.0961 0.0103 0.0113 0.0453 0.0097 7 0.3157 0.1656 0.0174 0.0183 0.0273 0.0117 8 0.2784 0.1940 0.0219 0.0229 0.0158 0.0136 9 0.2469 0.1995 0.0219 0.0227 0.0082 0.0170 10 0.2195 0.2102 0.0163 0.0172 0.0017 0.0193 B. Non-IPO firm-year observations (154,156) Size Decile Cash and Equivalents Retained Earnings Net Longterm Issue Net Total Issue Net New Equity Issue Dividend 1 0.4033-4.4562 0.0010-0.0016 0.1744 0.0083 2 0.3945-0.9800 0.0015 0.0019 0.0771 0.0069 3 0.3910-0.4492 0.0007 0.0030 0.0495 0.0074 4 0.3797-0.1444 0.0038 0.0056 0.0357 0.0073 5 0.3615 0.0281 0.0066 0.0084 0.0283 0.0082 6 0.3426 0.1029 0.0125 0.0140 0.0230 0.0092 7 0.3146 0.1742 0.0174 0.0183 0.0153 0.0111 8 0.2791 0.2003 0.0218 0.0228 0.0096 0.0132 9 0.2473 0.2047 0.0212 0.0219 0.0046 0.0164 10 0.2187 0.2120 0.0160 0.0168 0.0004 0.0193

C. Non-IPO firm-year observations within 1st and 99th percentiles of net new equity issue (151,058) Size Decile Cash and Equivalents Retained Earnings Net Longterm Issue Net Total Issue Net New Equity Issue Dividend 1 0.3976-3.4280 0.0025 0.0017 0.0743 0.0086 2 0.3880-0.8900 0.0019 0.0027 0.0541 0.0067 3 0.3870-0.4281 0.0010 0.0040 0.0432 0.0073 4 0.3776-0.1393 0.0037 0.0057 0.0346 0.0072 5 0.3599 0.0276 0.0064 0.0084 0.0289 0.0081 6 0.3411 0.1017 0.0119 0.0134 0.0248 0.0091 7 0.3140 0.1723 0.0168 0.0177 0.0177 0.0110 8 0.2784 0.1979 0.0214 0.0224 0.0122 0.0131 9 0.2469 0.2032 0.0206 0.0212 0.0073 0.0163 10 0.2181 0.2111 0.0156 0.0164 0.0021 0.0193 D. Non-IPO firm-year observations with zero-debt firms excluded (123,667) Size Decile Cash and Equivalents Retained Earnings Net Longterm Issue Net Total Issue Net New Equity Issue Dividend 1 0.3558-2.8477 0.0041 0.0063 0.0194 0.0028 2 0.3474-0.6983 0.0050 0.0076 0.0164 0.0039 3 0.3444-0.3126 0.0025 0.0068 0.0135 0.0054 4 0.3357-0.0751 0.0074 0.0106 0.0117 0.0059 5 0.3225 0.0388 0.0097 0.0133 0.0106 0.0077 6 0.3127 0.1059 0.0156 0.0178 0.0107 0.0086 7 0.2940 0.1580 0.0197 0.0208 0.0115 0.0104 8 0.2676 0.1839 0.0231 0.0243 0.0116 0.0126 9 0.2398 0.1940 0.0214 0.0221 0.0102 0.0159 10 0.2162 0.1999 0.0158 0.0163 0.0065 0.0187

Table VI. Leverage, Dividend and Financing Activities across Firm Size Deciles, Negative/Positive Retained Earnings The data consist of firm-year observations for the period 1971-2005. Observations with missing values in any of the reported variables are deleted. Size deciles are based on total assets. Firms are divided into positive (PosRE) and negative (NegRE) retained earnings groups within each size decile. All the variables are reported as a proportion of total assets. A. Cash Holdings, Dividend and Leverage Size Decile Retained Earnings Cash and Equivalents Book Longterm Book Total Market Longterm Market Total 1(NegRE) -5.2296 0.4144 0.1014 0.2276 0.0521 0.1098 1(PosRE) 0.3082 0.4132 0.0850 0.1447 0.0806 0.1370 2(NegRE) -1.7479 0.4062 0.1365 0.2508 0.0872 0.1619 2(PosRE) 0.3014 0.3944 0.1074 0.1748 0.1039 0.1684 3(NegRE) -1.2568 0.4267 0.1439 0.2463 0.1021 0.1770 3(PosRE) 0.2991 0.3748 0.1286 0.1952 0.1235 0.1877 4(NegRE) -0.8691 0.4397 0.1518 0.2370 0.1162 0.1832 4(PosRE) 0.2892 0.3688 0.1412 0.2051 0.1374 0.1998 5(NegRE) -0.6213 0.4221 0.1872 0.2702 0.1468 0.2152 5(PosRE) 0.2934 0.3510 0.1558 0.2111 0.1488 0.2023 6(NegRE) -0.5084 0.3784 0.2442 0.3198 0.1915 0.2548 6(PosRE) 0.2899 0.3371 0.1745 0.2264 0.1612 0.2109 7(NegRE) -0.4025 0.3146 0.3214 0.3939 0.2471 0.3074 7(PosRE) 0.2968 0.3159 0.1944 0.2437 0.1752 0.2213 8(NegRE) -0.3103 0.2600 0.3723 0.4357 0.2898 0.3442 8(PosRE) 0.2903 0.2823 0.2177 0.2656 0.1926 0.2354 9(NegRE) -0.3077 0.2327 0.3787 0.4394 0.2920 0.3434 9(PosRE) 0.2747 0.2492 0.2274 0.2748 0.2002 0.2415 10(NegRE) -0.1789 0.1896 0.3590 0.4179 0.2878 0.3378 10(PosRE) 0.2459 0.2226 0.2271 0.2905 0.2067 0.2612 B. External Financing Activities Size Decile Retained Earnings Net Longterm Issue Net Total Issue Net New Equity Issue Dividend Number of Observations 1(NegRE) -5.2296 0.0029 0.0008 0.3227 0.0039 12324 1(PosRE) 0.3082 0.0004-0.0041 0.0348 0.0195 3989 2(NegRE) -1.7479 0.0027 0.0038 0.1766 0.0051 9318 2(PosRE) 0.3014 0.0020 0.0012 0.0353 0.0091 6891 3(NegRE) -1.2568-0.0039-0.0008 0.1446 0.0065 7426 3(PosRE) 0.2991 0.0045 0.0052 0.0348 0.0089 8575 4(NegRE) -0.8691-0.0051-0.0045 0.1465 0.0043 6013 4(PosRE) 0.2892 0.0050 0.0065 0.0370 0.0106 9764 5(NegRE) -0.6213-0.0049-0.0033 0.1255 0.0053 4744 5(PosRE) 0.2934 0.0083 0.0098 0.0330 0.0111 10931 6(NegRE) -0.5084 0.0039 0.0029 0.1051 0.0039 3937 6(PosRE) 0.2899 0.0124 0.0141 0.0259 0.0115 11657 7(NegRE) -0.4025 0.0160 0.0168 0.0706 0.0071 3059 7(PosRE) 0.2968 0.0177 0.0186 0.0170 0.0127 12270 8(NegRE) -0.3103 0.0196 0.0185 0.0483 0.0087 2665 8(PosRE) 0.2903 0.0224 0.0239 0.0092 0.0146 12551 9(NegRE) -0.3077 0.0356 0.0344 0.0310 0.0105 2102 9(PosRE) 0.2747 0.0197 0.0209 0.0047 0.0180 12897 10(NegRE) -0.1789 0.0209 0.0172 0.0165 0.0097 1397 10(PosRE) 0.2459 0.0159 0.0172 0.0003 0.0202 13324 Total 155834

Table VII. Retained Earnings Deciles, Firm Size, Leverage Ratios, Cash Holdings, Dividend and External Financing Activities The data consist of 172,765 firm-year observations for the period 1971-2005. Observations with missing values in any of the reported variables are deleted. Retained earnings represent retained earnings divided by total assets. Book /Market Long-term/Total is long-term/total debt over book/market value of total assets. The market value of assets equals total assets minus total equity minus balance sheet deferred taxes and investment tax credit plus the market value of common equity plus preferred stock liquidating value. All the variables are reported as a proportion of total assets except for Total Assets. A. Firm Size and Leverage Ratios across Retained Earnings Deciles Retained Earnings Decile Retained Earnings Total Assets Book Longterm Book Total Market Long-term Market Total 1-5.1484 35.03 0.1545 0.2788 0.0956 0.1606 2-0.8008 209.05 0.1912 0.2947 0.1506 0.2271 3-0.2533 470.45 0.2281 0.3173 0.1930 0.2690 4-0.0299 1194.38 0.2429 0.3221 0.2161 0.2882 5 0.0797 2968.16 0.2322 0.3053 0.2126 0.2809 6 0.1532 4977.87 0.2140 0.2786 0.1952 0.2549 7 0.2239 2727.63 0.1917 0.2464 0.1733 0.2237 8 0.3058 2137.97 0.1649 0.2117 0.1472 0.1900 9 0.4152 2052.23 0.1231 0.1617 0.1041 0.1378 10 0.6316 1242.15 0.0606 0.0841 0.0470 0.0656 B. Cash Holdings, financing Activities and Dividend Payouts across Retained Earnings Deciles Retained Earnings Decile Cash and Equivalents Net Longterm Issue Net Total Issue Net New Equity Issue Dividend 1 0.4114 0.0004-0.0028 0.2742 0.0033 2 0.3856 0.0067 0.0077 0.1427 0.0041 3 0.3517 0.0129 0.0156 0.0981 0.0057 4 0.3255 0.0207 0.0225 0.0680 0.0078 5 0.3075 0.0184 0.0204 0.0510 0.0094 6 0.3069 0.0186 0.0197 0.0367 0.0097 7 0.3089 0.0120 0.0128 0.0240 0.0104 8 0.3150 0.0084 0.0092 0.0122 0.0119 9 0.3271 0.0027 0.0026 0.0014 0.0163 10 0.3715-0.0020-0.0024-0.0155 0.0272

Table VIII OLS Estimation Results for Panel Regressions on Determinants of Leverage Ratio The sample consists of 132,770 (92,172) firm-year observations with relevant Compustat data from 1971 to 2005 (1985 to 2005). The dependent variable is the total/long-term debt (TD/LD) divided by book/market value of assets (BA/MA). The independent variables are as follows: dummy variable equal to one if the firm has negative retained earnings and zero otherwise (NegRet); dummy variable equal to one if the firm has zero debt and zero otherwise (Zero); dummy variable equal to one for IPO year and zero otherwise (IPO); cash and equivalents divided by total assets (Cash); industry median debt ratio (Med); marginal tax rate equal to the statutory tax rate if the firm reports no net operating loss carryforwards with positive pretax return and zero otherwise (Tax); operating Income divided by total assets (OI); market-to-book ratio of assets (MB); log of book value of total assets (LnA); depreciation and amortization divided by total assets (DEP); fixed assets divided by total assets (FA); research and development expenditures divided by total assets (RND); a dummy variable for missing values in RND (D_RND); and common stock dividends divided by total assets (DIV). T-statistics are in the parentheses.

A. All firms TD / BA TD /MA LD /BA LD / MA Independent Variable Constant 0.2878 0.5224 0.2149 0.4208 0.1384 0.4257 0.1034 0.3450 (120.69) (109.11) 97.45 102.17 66.89 102.15 56.27 98.58 RE. +Re -0.2850-0.2163-0.2445-0.1742-0.2366-0.1745-0.1962-0.1377 (-108.89) (-65.95) -98.49-60.86-100.01-60 -93.24-56.32 RE.D -Re 0.0113 0.0103 0.0024 0.0006 0.0008-0.0001-0.0038-0.0048 (17.72) (12.56) 3.9 0.89 1.41-0.2-7.51-7.88 D No -0.2305-0.2136-0.2895-0.2477 (-59.74) -63.56-84.69-86.2 Rating -0.0445-0.0496-0.0592-0.0557 (-46.05) -58.82-69.11-77.28 Cash -0.1958-0.1838-0.1760-0.1604-0.1331-0.1220-0.1219-0.1123 (-72.73) (-58.35) -68.81-58.38-54.89-43.84-56.57-48.03 LnA 0.0032-0.0013 0.0026 0.0035 0.0132 0.0097 0.0104 0.0095 (14.94) (-3.79) 12.31 11.91 67.09 32.83 59.87 38.5 Div -0.0110 0.0271-0.0740-0.0432-0.0317 0.0141-0.0806-0.0343 (-1.02) (1.95) -7.25-3.58-3.25 1.14-9.31-3.33 MB -0.0010-0.0005-0.0080-0.0061-0.0010-0.0004-0.0045-0.0034 (-7.59) (-3.01) -62.75-46.79-7.83-3.36-41.29-30.65 RND -0.0512-0.0423-0.0520-0.0437-0.0117-0.0079-0.0225-0.0193 (-16.27) (-12.5) -17.42-14.8-4.12-2.63-8.93-7.63 D 0RD 0.0174 0.0199 0.0235 0.0259 0.0127 0.0144 0.0113 0.0151 (18.47) (16.58) 26.05 24.54 14.86 13.4 14.97 16.79 OI 0.0143 0.0140-0.0155-0.0136 0.0060 0.0023-0.0175-0.0185 (7.42) (5.88) -8.43-6.51 3.42 1.07-11.33-10.39 MOI 0.1546 0.0006 0.3373 0.0397 0.0850 0.0024 0.1059 0.0044 (13.6) (0.04) 31.56 3.08 8.1 0.18 11.44 0.4 Zero -0.1676-0.1679-0.1200-0.1101-0.0980-0.0973-0.0664-0.0614 (-109.61) (-93.36) -82.82-70.27-71.01-61.04-54.07-45.85 IPO -0.0171-0.0200-0.0291-0.0340-0.0141-0.0146-0.0216-0.0232 (-12.15) (-10.61) -21.74-20.71-11.06-8.75-19.06-16.55 FA 0.0079 0.0207 0.0387 0.0532 0.0766 0.0802 0.0671 0.0800 (3.17) (6.65) 16.3 19.55 32.82 28.24 32.48 33.42 AZ -0.0085-0.0077-0.0013-0.0004-0.0009-0.0002 0.0028 0.0034 (-18.92) (-13.28) -3.08-0.71-2.31-0.42 7.77 7.73 Depr -0.0349-0.0323-0.0180-0.0487-0.0113-0.0380-0.0015-0.0434 (-12.91) (-5.05) -7.02-8.72-4.64-6.69-0.67-9.09 Tax -0.0104-0.0167 0.0300-0.0304 0.0218 0.0225 0.0215 0.0006 (-3.98) (-4.36) 12.14-9.11 9.24 6.63 10.24 0.2 Med 0.3703 0.3279 0.3612 0.3044 0.3132 0.2540 0.4610 0.3372 (71.48) (50.94) 71.21 50.25 58.44 38.33 102.13 54.35 Adjusted-R 2 0.4066 0.4101 0.3978 0.4155 0.3635 0.3905 0.4086 0.4178

B. Small versus Large Firms (1971-2005) Small Firms TD / BA Large Firms Small Firms TD /MA LD /BA LD / MA Large Firms Small Firms Large Firms Small Firms Large Firms Independent Variable Constant 0.2623 0.4341 0.1496 0.4454 0.0831 0.3736 0.0375 0.3418 77.06 94.77 50.23 102.56 30.75 86.96 16.22 88.48 RE.D +Re -0.2436-0.3201-0.1829-0.2745-0.1770-0.2965-0.1328-0.2414-61.46-89.81-51.27-79.11-53.62-85.42-46.9-77.04 RE.D -Re 0.0049 0.0181-0.0054 0.0317-0.0045 0.0086-0.0083 0.0182 5.71 11.52-7.02 20.66-6.34 5.62-13.68 13.19 Cash -0.1807-0.2084-0.1662-0.1919-0.0997-0.2010-0.0947-0.1794-51.85-45.77-52.74-43.25-34.41-45.35-38.05-44.94 LnA 0.0018-0.0070 0.0136-0.0131 0.0160-0.0055 0.0193-0.0077 3.13-17.25 26.49-33.34 33.62-13.91 47.32-21.7 Div 0.0375-0.2229-0.0021-0.2808 0.0306-0.2914-0.0022-0.3517 2.95-9.17-0.19-11.87 2.9-12.31-0.24-16.44 MB -0.0004-0.0027-0.0054-0.0170-0.0002-0.0015-0.0024-0.0122-2.82-6.75-38.59-44.52-1.69-3.86-21.56-35.24 RND -0.0371-0.3165-0.0370-0.3501-0.0063-0.2219-0.0152-0.2205-10.64-20.5-11.77-23.32-2.16-14.78-6.11-16.35 D 0RD 0.0226-0.0013 0.0290-0.0002 0.0066 0.0039 0.0078 0.0001 16.06-1.04 22.83-0.2 5.61 3.11 7.74 0.06 OI 0.0001 0.0704-0.0267-0.1737-0.0079 0.1073-0.0241-0.0798 0.03 8.71-12.19-22.04-3.92 13.63-13.86-11.23 MOI 0.1961 0.0647 0.3773 0.2628 0.1403-0.0365 0.1239 0.0697 12.23 4.03 26.49 16.85 10.21-2.3 10.59 4.9 Zero -0.1748-0.1453-0.1283-0.0932-0.0961-0.1091-0.0690-0.0626-90.39-54.43-73.79-35.88-59.72-41.99-49.98-26.69 IPO -0.0185-0.0171-0.0294-0.0264-0.0140-0.0176-0.0203-0.0237-9.87-7.87-17.45-12.52-8.99-8.36-15.18-12.45 FA 0.0413-0.0268 0.0474 0.0497 0.1249 0.0191 0.0951 0.0477 11.15-7.73 14.31 14.48 39.66 5.42 35.6 14.84 AZ -0.0038-0.0207 0.0040-0.0147 0.0029-0.0134 0.0058-0.0075-6.32-30.64 7.26-22.32 5.68-20.3 13.42-12.7 Depr -0.0181-0.1560 0.0070-0.2867 0.0028-0.1466 0.0130-0.1717-5.74-9.07 2.47-17.01 1.06-8.73 5.78-11.23 Tax -0.0243-0.0135 0.0181 0.0205 0.0269-0.0118 0.0287-0.0054-5.86-4.16 4.88 6.48 7.79-3.71 9.64-1.9 Med 0.3151 0.3819 0.3087 0.2978 0.2385 0.3105 0.3985 0.3992 40.7 55.09 42.07 43.19 31.33 40.97 61.21 63.44 Adjusted-R 2 0.3629 0.4652 0.3584 0.4857 0.2779 0.4187 0.3329 0.4796

C. Rated versus Non-rated Firms (1971-2005) TD / BA TD /MA LD /BA LD / MA Non- Non- Non- Non- Independe Rated Rated Rated Rated Rated Rated Rated Rated Firms Firms Firms Firms nt Variable Firms Firms Firms Firms Constant 0.6679 0.2758 0.6138 0.1934 0.6435 0.1183 0.5617 0.0833 75.07 91.07 76.41 76.24 73.77 46.91 73.41 40.1 RE.D +Re -0.2868-0.2233-0.2583-0.1751-0.2998-0.1808-0.2536-0.1385-38.61-62.81-37.57-57.47-39.92-58.77-38.37-54.61 RE.D -Re 0.0094 0.0065 0.0482-0.0030 0.0074-0.0029 0.0372-0.0069 2.37 7.65 13.19-4.08 1.85-3.95 10.58-11.33 Cash -0.1835-0.1791-0.2262-0.1545-0.1908-0.1130-0.2205-0.1038-18.47-53.98-24.66-54.22-19.01-39.49-25.02-43.99 LnA -0.0293 0.0003-0.0256 0.0038-0.0313 0.0122-0.0260 0.0110-37.44 0.77-35.24 12.83-39.57 41.17-37.31 44.91 Div -0.0689 0.0126-0.1861-0.0454-0.2246 0.0130-0.2519-0.0302-2.04 0.91-5.95-3.8-6.57 1.07-8.37-3.03 MB 0.0008-0.0004-0.0119-0.0055 0.0013-0.0004-0.0098-0.0029 1.36-2.74-21.64-42.01 2.09-3.24-18.52-26.7 RND -0.3364-0.0379-0.3399-0.0364-0.2904-0.0065-0.2471-0.0159-10.28-11.00-11.25-12.32-8.77-2.2-8.51-6.47 D 0RD -0.0056 0.0242 0.0005 0.0291 0.0002 0.0145 0.0017 0.0153-2.25 17.94 0.22 25.14 0.07 12.42 0.75 15.86 OI 0.1837 0.0049-0.2164-0.0200 0.1873-0.0047-0.1350-0.0218 10.73 2.02-13.64-9.56 10.83-2.24-8.84-12.54 MOI -0.0531-0.0185 0.0622 0.0221-0.0560-0.0205 0.0233-0.0102-1.65-1.15 2.08 1.61-1.69-1.46 0.81-0.89 Zero -0.3054-0.1644-0.1881-0.1088-0.2602-0.0951-0.1566-0.0611-21.3-90.17-14.18-69.78-17.94-60.39-12.27-47.1 IPO 0.0206-0.0227-0.0152-0.0340 0.0235-0.0172-0.0107-0.0232 3.59-11.42-2.85-19.96 4.04-10.01-2.09-16.39 FA -0.0818 0.0391-0.0115 0.0621-0.0561 0.1003-0.0120 0.0904-13.29 11.12-1.96 20.68-8.69 32.18-2.07 35.2 AZ -0.0329-0.0051-0.0195 0.0023-0.0233 0.0017-0.0121 0.0048-21.13-8.38-13.55 4.34-14.79 3.26-8.69 11.16 Depr -0.0179-0.0320-0.0282-0.0347-0.0216-0.0371-0.0301-0.0333-0.59-4.79-0.99-6.05-0.7-6.43-1.1-6.98 Tax -0.0188-0.0313-0.0286-0.0337-0.0077 0.0119-0.0256-0.0023-2.71-7.1-4.47-8.93-1.1 3.12-4.15-0.73 Med 0.3139 0.3347 0.2270 0.3034 0.2629 0.2537 0.3000 0.3239 23.48 47.27 17.82 45.06 17.55 34.94 21.63 47.33 Adjusted- R 2 0.3884 0.3642 0.4516 0.3622 0.3658 0.2939 0.4342 0.3256