Capital Structure Determinants: The Direct and Indirect Effect of Country Characteristics.

Transcription

1 TILBURG UNIVERSITY Capital Structure Determinants: The Direct and Indirect Effect of Country Characteristics. Cross country analysis of corporate capital structures P.P.J. Betzel 10/10/2014

2 Capital Structure Determinants: The Direct and Indirect Effect of Country Characteristics. Cross country analysis of corporate capital structures Abstract This research investigates the determinants of corporate leverage ratios based on a sample of twelve countries. The countries are chosen regarding their different legal origin. First I model the leverage ratio as a function of firm-specific as well as country-specific characteristics. In the end I expand the current literature by testing for indirect effects of the country-specific characteristics on the relation between the firm-specific determinants and corporate leverage ratios. The results of the firm-specific variables are in line the existing literature, providing support for the negative effect of profitability and the positive effect of size and tangibility on leverage. For macro-level variables I find negative effects of shareholder rights and positive effects for creditor rights and bank oriented countries on leverage ratios. The results of the indirect hypotheses are less obvious but clearly indicate that institutional and legal variables affect the importance of some firm-specific variables on corporate capital structures. Master Thesis Finance Pascal Betzel Supervisor: Prof. dr. L.D.R. Renneboog Date: 11th November

3 Table of Contents Chapter 1: Introduction <<<<<<<<<<<<<<<<<<<<<<<< 4 Chapter 2: Literature review<<<<<<<<<<<<<<<<<<<<<< 5 2.1: Capital structure theories : Firm-specific determinants of capital structure : Country-specific determinants of capital structure : Measures of capital structure Hypotheses. 32 Chapter 3: Data and variable description <<...<<<<<<<<<<<<<< : Data : Variables : Methodology.. 39 Chapter 4: Results <<<<<<<<<<<<<<<<<<<<<<<<<< : Firm-specific effects : Country-specific direct effects : Country-specific indirect effects Chapter 5: Conclusion and discussion <<<<.<<<<<<<<<<<<< : Conclusion : Contribution and recommendation. 53 References <<<<<<<<<<<<<<<<<<<<<<<<<<<<<. 55 Appendix <<<<<<<<<<<<<<<<<<<<<<<<<<<<<. 59 3

4 1. Introduction The total indebtedness in the world, including all parts of the public and private sectors amounts for 313% of global gross domestic product (S.Reddy, 2013). As this number shows, debt, and hereby capital structure is of much importance in the current economy. Although corporate capital structure research is one of the most widely investigated topics in corporate finance, studies comparing differences in the capital structure between countries started to appear much later in time and became especially hot in the last decade. One of the early studies on the country-specific influences on leverage was performed by Rajan and Zingales (1995) in which they conclude that indeed country-specific factors play an important role in determining corporate capital structures. A more recent study of de Jong, Kabir and Nguyen (2007) elaborates on this finding by suggesting that these country-specific factors influence leverage also indirectly, by influencing the way in which firm specific factors affect leverage. This study will analyze both influences of country-specific factors, direct and indirect, as well as the main firm-specific determinants. Hereby providing a clear defined picture of all the possible capital structure determinants. This research will start with a summary of past literature done in the field of capital structure. In this part I will discuss several capital structure theories and introduce the potential firm- and country-specific determinants of corporate leverage, I finish this chapter by developing the main hypotheses of this research. Next I will discuss the data collection and explain the variables used in this study by showing some basic descriptive statistics and definitions. After the data is clearly presented I divide my study into three sections; firm-specific effects, country-specific direct effects and country-specific indirect effects. I will provide methodology for each section and continue by testing the different hypotheses. The research will be conducted using an 4

5 OLS panel regression. To end this study, I will conclude my findings and give some recommendations for further research in the last part of this paper. 2. Literature review 2.1 Capital structure theories Explaining differences in capital structure between firms has been one of the most popular research topics in corporate finance. In 1958, Modigliani and Miller developed the capital structure irrelevance proposition, before that time there was no general accepted theory of capital structure. Modigliani and Miller hypothesized that in perfect markets it does not matter what capital structure a firm uses to finance its operations, the core determinants of the company s market value are its earning power and the risk of the underlying assets. Financial leverage, the proportion of debt financing, is irrelevant for the total firm value. This leverage-irrelevance result generalizes to any mix of issued securities by the company; the total firm value is not affected by the proportion of short-term or long-term debt, callable or call protected debt, straight or convertible debt, or the currency of the debt. Any mixture of these or other securities would results in the same total firm value. An arbitrage argument used by Miller and Modigliani in there research was that if a firm can change its market value by a pure financial operation, the investors in the firm can take actions that exactly replicate the resulting debt position of the firm. In the end, these transactions would solely change the weights of a portfolio and should result in zero profit, under the consideration that it is a perfect capital market. If the market were efficient enough to eliminate the profits for the investors, any profit made by the company would be eliminated too. 5

6 The basic Modigliani and Miller proposition is based on the following assumptions: - No transaction costs - No bankruptcy costs - No taxes - Borrowing costs for both firm and investor are equal - Symmetry of market information (firms and investors have the same information) - Debt has no effect on a firm s earnings before interest and taxes These assumptions are only valid in a perfect capital market and are all hard to justify in real world examples. But in order to understand how for example taxes affect the capital structure, you first have to understand this basic proposition and therefore it is a very useful model. The Modigliani and Miller model and its theoretical extensions have stimulated serious research on capital structure. After this study lots of researchers tried to prove that the irrelevance proposition was wrong or tried to propose a better model. And because of the many assumptions made in the original study, there was enough room for improvement. This Modigliani and Miller model is for sure intuitive, but is it credible? It is really hard to argue that capital markets are really sufficiently perfect, since we see constantly new type of securities arise. This continues innovation of securities provides theoretically evidence that finance does matter. There would be no incentive to innovate, if those new type of securities or financing tactics never added value. But testing the Modigliani and Miller theory in this way has been proven to be difficult. The costs of creating new securities or financing schemes are low, and the costs of imitating those securities are trivial. Many companies might use these new products, but only the first users will experience an increase in value. 6

7 Although many of the outstanding literature is empirically tested, there is still little consensus on how companies choose their capital structure. Up to now, most empirical research on capital structure examines one of the two traditional theories; the trade-off theory and the pecking order theory. In the following I will discuss the main findings of those two studies and of two other important theories on capital structure; the agency theory and the market timing theory. Trade-off theory Most countries tax corporate income, but interest paid on debt is a tax-deductible expense. Therefore financing operations with debt instead of equity increases the total after-tax return to investors, and should in the end, ceteris paribus, increase the value of the firm. The trade-off model is on the heart of this and says that a company will borrow just the amount of debt so that the marginal value of the tax shields on additional debt is just offset by the increase in the costs of financial distress. The trade-off theory is used to describe a number of related theories; in general it suggests that firms choose a leverage target that optimally balances between costs and benefits of borrowing, holding the company s assets and investment plans constant. The company substitutes equity for debt, or debt for equity until the company s value is maximized. The original trade-off model was a result of the discussion around the Modigliani and Miller theorem. When corporate taxes and bankruptcy costs were incorporated in the model there was an advantage for holding debt, the tax-shelter effect as explained above. But holding on for too much debt creates the cost of potential financial distress. Therefore this original trade-off theory is a sub theory of the general theory of capital structure because there are only two assumptions broken, the no tax and the no bankruptcy cost assumption. 7

8 The amount of debt should be inversely related to the bankruptcy or financial distress costs (Miglo, 2010). If bankruptcy costs would not exist, then a firm facing bankruptcy would not see any change in the total firm value. The value of a security is the same, regardless of whether a bankruptcy would occur or not under some specific conditions (Stiglitz, 1969). However, financial distress costs do exist and can be divided into direct and indirect costs. The direct costs include transferring the assets to the new owners, auditors fees, management fees and legal fees. Indirect costs arise when employees make the decision to continue their work elsewhere because the firm faces bankruptcy. As a result the company faces some difficulties, and customers and suppliers that rely on continuous transactions may lose confidence in the company. An increase in tax rate will increase the firm s value of the tax shield. A company is allowed to reduce income by deducting the paid interest on debt, so in the end it reduces their tax liabilities. Therefore an increase in tax rates should lead to an increase in leverage. The empirical evidence is however mixed; Graham and Harvey (2001) show that almost half of the CFOs agree that tax considerations play an important role in determining their leverage ratios. But on the other hand Wright (2004) provides evidence that the corporate debt ratios hardly do not change over time, although the period ( ) covers lots of different tax rate levels. Do firms change their debt ratios towards a pre-determined target debt ratio? The trade-off theory suggests a target leverage ratio, and most of the evidence finds support that this so called mean reversion (the process of adjusting capital structure towards target debt ratios) indeed exists. Although firms histories strongly influence their capital structures, over time their capital structures tend to move towards target debt ratios that are consistent with the trade-off theories of capital structure (Kayhan and Titman, 2007). Other studies support this finding of Kayhan and Titman, but there is still discussion about the speed of mean reversion, which is due to Fama and French 8

9 opinion too slow. Inconsistent with the static trade-off theory, firms adjust very slowly toward target leverage, and past securities issues have strong and long-lasting effects on capital structure even after controlling for target leverage (Huang and Ritter, 2004). The trade-off theory, more specifically, the trade-off between tax shield benefits and financial distress costs, create in theory a crucial relation with profitability. Expected is that bankruptcy costs are lower for more profitable companies and that the tax shields are more valuable for these companies since they produce more profit. So in the end, the trade-off theory expects these companies to incur higher debt ratios. This relationship has been widely tested and, in contrast to the theory, most studies find a negative correlation between profitability and debt ratios (Titman and Wessels, 1988). Although the trade-off theory cannot explain this relationship, it is consistent with the empirical findings of other important factors on leverage. It supports for example that firms with more tangible assets (more collateral), borrow more than companies with more risky, intangible assets. The trade-off model is the most dominant theory of capital structure used in corporate finance textbooks, but the overall empirical relevance of the trade-off model has often been questioned. According to Miller, the balancing between tax benefits and financial distress is not correct, as taxes are large and sure, bankruptcy is rare. To strength his point, he suggested that if the model were true, then companies ought to have more leverage than we observe in reality. Although the theory wrongly predicts the relation between profitability and debt, and the evidence on the influence of the tax rate on capital structure is somewhat ambiguous, most of the evidence does support the main findings of the trade-off model. More dynamic versions of the theory generally seem to be difficult to reject empirically. 9

10 The pecking order theory In a study on corporate financing decisions of Chinese listed companies, Tong and Green (2005) test the trade-off and pecking order theory. They set up three models based on the determinants of leverage, the relationship between leverage and dividends, and the determinants of corporate investments. The results of the first two models support the pecking order model over the trade-off model. The results in the last model are inconclusive. Therefore they conclude that the pecking order theory better explains financing behavior of Chinese firms. So what includes the pecking order theory? The pecking order theory suggests that firms follow a financing hierarchy designed to minimize adverse selection costs of security insurance. The theory was first suggested by Donaldson in 1961, but was popularized by Myers and Majluf in The traditional pecking order model explained the preference for internal financing over external financing by transaction and issuing costs. Retained earnings (internal financing) involve very few transaction costs, whereas equity issues involve more transaction costs then issuing debt. But in the modified theory, Myers argues that equity is a less preferred source of financing not because of the transaction costs, but because of a form information asymmetry. If managers issue new equity, investors believe that managers (who are assumed to know more about the firms true conditions), overvalue the firm and are trying to take advantage of this over-valuation, which results in a lower valuation of the new equity issuance among the investors. This theory is based on the fact that the cost of financing increases with asymmetric information. Companies prefer internal financing over debt and debt over equity issuing; resulting in the financing hierarchy where internal financing is used first, and when this source is exhausted, debt will be issued, and when it is no longer sensible to issue any more debt, the firm will issue equity. 10

11 In the pecking order theory well performing firms had to use internal funds to avoid adverse selection problem and in the end loss of value. Signaling is a key concept in the pecking order model. The signaling theory states that corporate financial decisions, like capital structure decisions, are signals sent by the company s managers to investors in order to shake up these asymmetries. The signaling model stated that only low profit firms would issue equity and that rational investors would foresee this. Therefore these investors will demand a discount, which costs will be borne by the internal shareholders. These costs increase with the level of information asymmetry. In practice, the pecking order theory cannot explain why certain financing tactics to avoid the financing consequences of the asymmetric information between managers and investors are not developed or widely implemented. Consider the following example, suppose that some special information available for the manager today reaches investors next year. Then the company could issue deferred equity securities. Since the manager has now idea today whether the future stock is too high or too low, these types of securities convey no information. In this way is the company able to issue equity with no adverse signals to investors. The empirical findings on the pecking order model are rather ambiguous. Shyam- Sunder and Myers (1999) find that the pecking order theory has much greater timeseries explanatory power than a trade-off model and support the model. But Leary and Roberts (2010) conclude differently; while the classificatory ability of the pecking order varies significantly depending on whether one interprets the hypothesis in a strict or liberal manner, the pecking order is never able to accurately classify more than half of the observed financing decisions, therefore providing evidence for the failure of the pecking order theory. 11

12 An argument in favor of the pecking order model with respect to the trade-off model is that this theory can explain the negative correlation between profitability and debt. One can imagine that high profitable firms can finance their new investments with their own retained earnings, simply because they earn more than less profitable firms. Therefore these less profitable firms are more obliged to use external funds, like debt, in order to make an investment. The previous discussed theories, i.e. the trade-off theory and the pecking order theory, both rely on the assumption that the supply of capital is perfectly elastic, therefore implying that the capital structure is solely determined by a company s demand for debt. However, there are theoretical arguments which suggest that supply-side factors may be relevant for corporate leverage ratios. Recently Antzoulopos et al. (2014) provided evidence that, along with firm- and industry-specific factors, indeed economy-wide supply factors influence corporate capital structure. Agency theory Besides the research on the trade-off and pecking order theory, there is also a wide variety of studies which examine agency theories in capital structure decision making. These models suggest that agency problems can affect capital structure from two different perspectives, either the relationship between shareholder and manager or the relationship between shareholder and bondholder influences capital structure. Agency costs arise when the principal (manager) does not necessarily act in the best interest of the agent (shareholder). This principal-agent relation also holds for the interaction between shareholders and bondholders, however in this case are the shareholders the agent. In the end this conflict of interest makes debt less valuable and therefore reduces the incentive to issue debt. 12

13 Suppose that managers act in the interest of the equity holders and that the risk of default is significant. The managers will be seduced in taking actions that transfer wealth from the bondholders to its equity holders. In order to do this there exist several strategies. First, a phenomenon that could arise is the asset substitution effect. It might be that firms implement projects with a negative net present value. This problem is based on the fact that shareholders benefit if the project undertaken yields high returns and the bondholders bear the consequences of the failure of the project. So the asset substitution effect (Jensen and Meckling, 1976) means that shareholders might benefit from high risk investments, even if those investments are value-decreasing since bondholder bear the negative payoff. Second, managers can try to borrow additional debt and pay out cash to the shareholders. This results in the same overall shareholder value, but a lower value of the existing debt. In the end, the decrease in the value of the stocks is much less than the cash received for the shareholders. Finally, firms may reject positive NPV projects on purpose. Debt overhang is the condition of a firm when it has already such huge debt levels that it hardly can borrow any more money, even when the new borrowing is for sure a good investment that would repay more than its initial cost. The consequences of excessive debt are that any earnings generated by new project are mostly assigned to existing debt holders. Furthermore the firm is not able to issue new junior debt because default is likely. And on the other hand the shareholders do not want to issue new shares since they will bear part of the loss that otherwise would have been borne by the junior debt holders. So in the end the company does not implement positive NPV projects since it is not able to fund them. Bondholders are obviously aware of these agency conflicts and try to control these. To protect against agency problems between bondholders and shareholders, several strategies exists. For example when using bank debt, the monitoring quality of the lender increase, since banks debt is in general better monitored, therefore firms have 13

14 less chance to implement value decreasing projects. Another possibility is to use convertible debt, which gives the lender the option to convert the debt towards shares. Finally you can make use of debt with some financing, asset, dividend or binding covenants. If firms violate these covenants, bondholders are allowed to intervene in the firm. The advantage of all these strategies is that it makes debt much safer to issue. Agency conflicts between manager and shareholder also arise. Separation of ownership and control in public corporations leads to several adverse consequences. Managers, will act in their own interest, and try to achieve better contract terms like, higher-thanmarket salaries, job security and perquisites. Further, they seek to increase their bargaining power by investing in projects where their managerial skills and knowledge are particularly useful. The shareholders will take some measures in order to discourage the managers in applying those value transferring actions. These measures include several monitoring and controlling actions like supervision by independent directors and the threat of takeover. However perfect monitoring is extremely costly and leads to decreasing returns. As a solution, shareholders try to align the goals of the manager with the goals of the firm by composing a certain compensation contract. But also this measure is not a perfect solution, since the manager never bears the full negative consequences of managerial actions. Moreover, we do not have a good, observable measure of the managers performance, since the actions undertaken by a manager may only account for a small proportion of the firm value. This makes it very difficult for shareholders to reward managers on specific characteristics like commitment and good decision making. To sum up, agency theory starts with the assumption that stakeholders (shareholders, bondholders and managers) have objectives and interests that are not necessarily in line with each other. This results in arising conflict between the different stakeholders. The 14

15 optimal leverage ratio results from a compromise between several funding options that in the end leads to the lowest level of conflicts. The market timing theory Recently, Baker and Wurgler (2002) have implemented a new theory of capital structure. The market timing theory of capital structure states that the current capital structure is the cumulative outcome of past attempts to time the equity market. Attempting to time the market means for example that a company issues new shares when he believes that the shares are overvalued and that the firm repurchases own shares when he believes the opposite is true; shares are undervalued. The authors claim that market timing is the first order determinant of a firms capital structure; companies in fact do not care whether their financing source is debt or equity, they just prefer the source of financing which, at that moment, seems to be more valued by financial markets. Most of the empirical evidence supports this theory. Baker and Wurgler, support their theory by showing an index of financing that reveals how much of the financing was done during the so called hot equity periods and how much during the hot debt periods. More prove comes from Huang and Ritter (2004), they find that publicly traded U.S. firms fund a much larger proportion of their financing deficit with net external equity when the expected equity risk premium is lower and when the first-day returns of initial public offerings are higher. However testing the persistence of the impact of the market timing revealed the following; Hot-market initial public offering (IPO) firms issue substantially more equity, and lower their debtto-asset ratios more than cold market firms do. But, immediately after going public, hotmarket firms increase their debt-to-asset ratios by issuing less equity and more debt 15

16 relative to cold-market firms. At the end of the second year following the IPO, the impact of market timing on leverage completely vanishes (Alti, 2006). The idea seems to be quite plausible. But unfortunately, the theory has almost nothing to say about the factors which are traditionally considered to impact the capital structure. It does not explain the negative relationship between profitability and leverage for example. It only suggests that firms issue equity when the equity market is found relatively favorable or that firms reduce their leverage ratio when the debt market conditions are relatively unfavorable. Concluding, none of these theories can explain capital structure by themselves. The agency theory contains plenty of interesting and intuitive ideas but it is hard to really quantify the overall effect of these agency problems on the capital structure. And although the trade-off theory can explain a lot of phenomena in capital structure, it fails in describing the important negative relation between profitability and debt. The pecking order model however provides a clear answer for this relationship, but there is no consistent opinion regarding the pecking order theory itself given the ambiguous empirical evidence. 2.2 Firm-specific determinants of capital structure Several studies examine the effect of firm-specific factors on the capital structure decision. Capital structures are usually explained by many variables arising out of static trade-off, agency and information asymmetry considerations (Booth et al., 2001). Regressions show that between-variation is much bigger than within-variation of capital structure (Lemmon et al. 2008). Thus most of the total variation of leverage ratios can be explained by cross-sectional differences rather than time-series variation. 16

17 For choosing the correct firm-specific variables I started with the study of Harris and Raviv (1991) which compared several studies on capital structure as a starting point; Bradley et al. (1984), Chaplinsky and Niehaus (1990), Friend and Hasbrouck (1988), Gonedes et al. (1988), Long and Malitz (1985), Kester (1986), Kim and Sorensen (1986), Marsh (1982), and Titman and Wessels (1988). Comparisons suffer from the fact that these studies used different measures of the firm characteristics, different time periods, different leverage measures, and different methodologies. Although these studies are more than 20 years old, they still are very useful, since they give a good set of starting control variables. Combining these findings with the current state of the literature gave me the most important factors of capital structure. Profitability The pecking order theory expects a negative correlation between profitability and leverage. More profitable firms can finance their next investment with their own retained earnings and are not obliged to issue any debt. But less profitable have no other option than using debt; therefore it is likely that profitability has a negative effect on the leverage ratio. Most of the literature confirms the negative relationship between profitability and leverage (Bevan and DanBolt, 2002; Huang and Song, 2005; Titman and Wessels, 1988). Considering the trade-off theory, one would expect a positive relation between profitability and leverage. The tax benefits for profitable firms are more valuable and the bankruptcy costs are lower, therefore expecting higher debt levels. This argument can somewhat be mitigated by considering the effect of personal taxation. Moreover interest tax shields may be unimportant to companies with other tax shields, such as depreciation (DeAngelo and Masulis, 1980). But again, most of the findings do not support this positive relation and are consistent with the pecking order model by confirming the negative relationship (Rajan and Zingales, 1995). 17

18 Size The trade-off theory suggests that the optimal leverage ratio balances the benefits of the tax shield of debt and the increasing agency and financial distress costs. From a financial distress perspective, as larger firms are more diversified they are expected to go bankrupt less often than smaller ones, therefore they can handle higher levels of leverage before facing higher bankruptcy costs. In addition, as firms grow in size the likelihood to have a credit rating increases. A credit rating gives a company access to non-bank debt financing, which is usually unavailable to smaller companies. Both arguments predict a positive relation between size and leverage. However, size might also be a proxy for the information available for outside investors, which should increase their preference for equity relative to debt. But empirical evidence supports the positive correlation (Bevan and DanBolt, 2002; Psillaki and Daskalakis, 2009; Huang and Song, 2005). Tangibility Agency theory suggests that highly levered firms tend to underinvest, or invest suboptimal, and thus reducing the bondholder s wealth. These cause lenders to require collateral because the use of secured debts can help alleviate this problem. Tangibility can be seen as some kind of collateral. As the tangibility of assets increases, the liquidation value of the company increases and the probability of mispricing in the case of bankruptcy decreases. Therefore it is likely that firms which are unable to provide collaterals get debt at less favorable terms, or will be forced to issue equity. In trade-off theory one would see tangibility as an important driver of financial distress costs. More tangible assets shift the point of incurring bankruptcy costs upwards. So a positive relation between tangibility and the leverage ratio is expected and confirmed by empirical evidence (Rajan and Zingales, 1995; Bevan and DanBolt, 2002). The above described variables are in most papers seen as one of the most influential capital structure determinants. But there are many other factors influencing capital 18

19 structure. Although profitability, size and tangibility have my main interest in the analysis, I will apply more determinants in my regression model to control. Growth opportunities Higher growth opportunities provide incentives to invest sub optimally, or to accept risky projects that transfer wealth from debt to shareholders. This raises the costs of borrowing and therefore firms with more growth opportunities tend to prefer internal resources or equity over debt. This agency problem can be mitigated if the firm issues short-term rather than long-term debt. Therefore expected future growth should be negatively related to long-term debt levels, whereas it should be positively related to short-term debt ratios. The empirical evidence regarding growth opportunities and (long-term) debt ratios, mostly confirms the negative correlation (Titman and Wessels, 1988; Rajan and Zingales, 1995; Bevan and DanBolt, 2002). Tax There exists consistency in the fact that nowadays almost all researchers believe that taxes play an important role to companies capital structures, but significant tax effects are found much less. In theory, firms with a higher effective marginal tax rate would use more debt to increase the tax-shield gain. Mackie-Mason (1990) comments that leverage ratios are the cumulative result of years of separate decisions and most tax-shields have a negligible effect on the marginal tax rate for most companies, thereby explaining the failure of finding significant empirical evidence. But Mackie-Mason finds some results which confirm that tax rates influence capital structure. He studies contradictory to other researchers, the incremental financing decisions. In the end his findings are consistent with the Modigliani and Miller theorem, the benefits of debt financing at the margin are positively correlated with the effective marginal tax rate. Non-debt tax shields Non-debt tax shields are the tax deduction for depreciation and investment tax credits. They are expected to be substitutes for the tax benefits of debt financing, which makes, holding other factors constant, the likelihood to use debt lower 19

20 for companies with larger non-debt tax shields (DeAngelo and Masulis, 1980). So nondebt tax shields are expected to be negatively correlated with leverage. Empirical evidence generally supports this relationship (Wald, 1999). An indicator for non-debt tax shields is the ratio of depreciation over total assets. Risk Volatility or firm risk is a proxy for the probability of financial distress. Expected is that it has a negative impact on leverage. When a company is seen as risky, the costs of debt will increase for that company since lenders will demand a risk premium. Therefore the debt levels are expected to be lower for more risky firms. Lots of studies confirm this negative relationship (Booth et al. 2001; Titman and Wessels, 1988; Wald, 1999). Uniqueness The costs that firms can potentially impose on their suppliers, customers and workers in case of bankruptcy are relevant to their leverage ratio. Suppliers, customers and workers of companies that produce unique or very specialized products are more likely to suffer high costs in the case that the firm faces financial distress or bankruptcy. The customers may experience difficulties to find an alternative firm that serves those unique products. And suppliers and workers probably have job specific skills and capital, which makes them less flexible. Because of these high costs in case of liquidation, uniqueness is expected to have a negative effect on leverage ratios. Empirical evidence supports this relationship (Titman and Wessels, 1988). Proxies for uniqueness are research and development expenditures over sales and selling expenses over sales. It is expected that this ratio measures uniqueness because firms that sell products with close substitutes faces higher chances that their innovations will be duplicated, which makes them therefore less likely to do lots of research and development. Additionally, successful R&D projects potentially lead to new and different products. 20

21 Besides these firm-specific factors, I will also control for time and industry effects. Firms in the same industry are exposed to similar economic conditions and regulatory influences, therefore one would expect more similar leverage ratios within industries. It is known that some industries, like the capital-intensive manufacturing industries and utilities, are characterized by high leverage. And for example, high-tech and mineral extraction industries are known to have low leverage. Overall firms within industries are more the same than those in different industries and industries tend to retain their leverage rankings over time (Harris and Raviv, 1991). So in order to control for these industry effects, I will add industry dummies where appropriate. 2.3 Country-specific determinants of capital structure Our knowledge of capital structures has mostly been derived from datasets containing US firms, or firms of one specific developed country. Therefore our knowledge on the effects of country-specific influences on capital structure is still rather undeveloped. However, prior research does confirm that a firm s capital structure is not only influenced by firm-specific factors but also by country-specific factors. Antzoulatos et al. (2014) find that institutional environment matters in capital structure decision. They conclude with; the extension to a cross-country setting would provide an excellent venue to explore the relative importance of institutional variables affecting leverage. This could be a significant contribution to a recent and growing literature that examines institutional determinants of corporate leverage. When focusing only on the Asia Pacific region, similar results have been found. These results suggest that the capital structure decision of firms is influenced by its operating environment, as well as firm-specific factors identified in the extant literature (Deesomsak et al., 2004), therefore providing evidence that country-specific variables 21

22 matter. A comparable study of Psillaki and Daskalakis (2008) on the other hand found that small and medium-sized enterprises in different countries determine the capital structure in similar ways, therefore mitigating the effect of country-specific factors. They conclude that firm-specific rather than country-specific facts explain differences in capital structure choices. However they used only Greek, French, Italian, and Portuguese companies, which are most of all south European countries. The similarities across these countries are much higher than between other countries in the world. This is why I compose my sample with different countries across the world based on different legal origins. Countries can be grouped based on their legal origin, which depends on their historical law system. I will discuss these groups, the so called legal families, in the following. Legal origin One can divide law systems around the world into two main laws; common law and civil law. Civil law is the oldest legal system and finds its origin in Europe, in Roman law. This law is nowadays outstretched over the globe due to a combination of conquest, imperialism, outright borrowing or other influential events. Today it covers about 150 countries. In civil law origin countries, codes and statutes are developed to cover all eventualities and the role of the judge is somehow mitigated. Within this civil law tradition, a subdivision can be made between French, German, and Scandinavian law. The French law was established under Napoleon, and has influence in the western parts of Europe as well in parts of Africa and Oceania. The German law was written several decades later, which could explain why it was not that widely adopted as the French law. It covers some of the middle and eastern parts of Europe as well as the eastern part of Asia. The last subdivision of the civil law is the Scandinavian law. This 22

23 legal family is much smaller than the French or German law and has fewer similarities with the Roman law than the other two civil laws have. Common law is a typical English development and can be found in 80 countries around the world that are former English colonies or have been influenced by the Anglo-Saxon tradition (for example; Australia, India, Canada and the United States). Although civil law heavily relies on codes and statutes, common law is known to be a system in which judges have a very active role in developing rules. Common law countries make use of statutes, but regard judicial cases as the most important source of law. But to be sure the power of the judges remains consistent; courts are restricted by precedents set by a higher court. The division of countries to a legal family is defined according to the following criteria: historical background and development of the legal system, theories and hierarchies of sources of law, the working methodology of jurists within the legal systems, the characteristics of legal concepts employed by the system, the legal institutions of the system, and the divisions of law employed within a system. As I mentioned before, instead of the two traditional legal systems, a subdivision of civil law is made, so that in the end there are four different legal families; British law (Common law), French law (Civil Law), German law (Civil Law) and Scandinavian law (Civil Law). La Porta, Lopez-de-Silanes, Schleifer and Vishny (hereafter referred to as LLSV), use these differences in legal origin and test the differences of shareholder protection, creditor protection and law enforcement between those systems, hereby providing some first insights in the concrete differences between the legal families. In the following I will discuss the possible effects of these different law implications on the capital structure as well as the measures and results of LLSV. Afterwards I will discuss some other country-specific variables that may affect capital structure decisions. 23

24 Shareholder protection Shareholder rights that increase shareholder protection and which are used to measure shareholder protection in the LSSV paper are: - One share-one vote - Antidirector rights (favor minority shareholders), which include; proxy by mail allowed or showing up in person/sending an authorized representative; allowing for cumulative voting for directors and mechanisms of proportional representation on the board; required share deposit prior to a shareholder meeting; preemptive rights to buy new issues of stock; possession of legal mechanisms against perceived oppression by directors; percentage of share capital needed to call an extraordinary shareholders meeting. - Right to mandatory dividend (remedial shareholder rights measure, only present in French Civil Law) La Porta et al. use anti-director rights (adding Proxy by mail, Shares not blocked, Cumulative voting, Oppressed minority, Pre-emptive rights plus one if Extraordinary meeting is less than or equal to 10%) as a measure for shareholder rights, which ranges from 0 to 6. They find that common Law provides overall the strongest shareholder protection and French Civil Law the worst, with Scandinavian and German Civil Law in between. It is expected that shareholder protection has a negative correlation with agency costs. In companies where shareholder rights are lower, it is more likely to experience a wider divergence of ownership and control, which increases the likelihood of agency problems. This relationship is used by Jirapron and Gleason (2007) to discuss the relationship between shareholder protection and leverage. Because leverage is related to agency costs, and agency costs in turn are associated with shareholder rights, they 24

25 hypothesize that the strength shareholder protection has an impact on a firm s capital structure. Firms with less shareholder protection should adopt higher leverage ratios in order to mitigate the higher agency costs between shareholders and management. In the end Jirapron and Gleason provide empirical evidence for this inverse relationship. But they do not find evidence for this negative relation in regulated firms. They explain this by the fact that regulation already helps to alleviate agency problems and therefore mitigates the role of debt in reducing agency costs. Further evidence on this inverse relationship between shareholder protection and leverage is provided by Cheng and Shiu (2007), in their study they conclude that countries with better shareholder protection use more equity funding, which ceteris paribus results in a lower leverage ratio. Creditor protection Measuring creditor rights is more difficult than the measurement of shareholder rights. This is due to the fact that there are many different kinds of creditors with different interests, hence protecting rights for one creditor may be unfavorable for another. And there are two general creditor strategies of dealing with a defaulting firm: liquidation and reorganization. Each one of them requires different creditor rights to be effective. Creditor protection rights exist of the following: - No automatic stay of assets in the reorganization procedure - Secured creditors gets paid first - Restrictions for going into reorganization (creditor consent) - Management does not stay in charge during the reorganization procedure - Existence of a legal reserve requirement (remedial creditor rights measure, more common in civil Law) 25

26 The creditor rights index used by LLSV is measured by adding the four binary variables ( Reorganization, No automatic stay, Secured first, and No management stay ) and therefore ranges from 0 to 4. Again, Common Law provides the strongest protection for creditors and French Civil Law the weakest, with German and Scandinavian Civil Law in between. Surprisingly, it is not the case that one family protects the shareholders and another protects the creditors. This is also confirmed by the correlations of creditor and shareholder rights, which are generally positive (La Porta et al., 1998). Ghoul et al. (2012) consider two views regarding the effect of creditor protection on capital structure. The first view suggests that creditor protection has a positive effect on the debt levels of the firm. This view focuses on the supply side of the financial market. The idea is that stronger creditor protection induces lenders to provide credit at more favorable terms, resulting in more use of debt by firms (expecting that the costs and benefits of equity issuing remain the same). This can be confirmed by the agency theory; agency cost decrease with creditor protection. If creditor protection is low (less creditor rights), debt holders face more asymmetric information costs. As discussed in the agency theory before, there exists a conflict between shareholder and bondholder. Shareholder might reject positive net present value projects or accept negative NPV projects. If a creditor is more protected, the control on the firm will be better and it will be more difficult for a firm to implement negative NPV projects. So in countries with higher creditor protection, debt is more valuable and the incentive to issue debt will be greater. This hypothesis is supported by the findings of Cheng and Shiu (2007), where they find a strong positive influence of creditor protection on leverage ratios. The second view, based on the demand side of the financial market, expects the opposite; creditor protection has a negative effect on the debt levels of the firm. It believes that strong creditor protection discourages managers and shareholders form 26

27 using large amounts of debt because they want to avoid losing control in the case of financial distress. Law enforcement One could expect that countries which adapt a less investor protective legal system might compensate this by improved law enforcement to outweigh those disadvantages. La Porta et al. find that French civil law has been proven to provide the lowest law enforcement. And although, common Law is not the leading origin for these factors (most of the times it is Scandinavian civil Law), it is well-ahead of the French Law. Also when considering the quality of accounting, French civil Law has the weakest performance. Thus these results suggest that enforcement is not a decent compensation to weak shareholder and creditor protection. LLSV measure law enforcement due to the following criteria: - Efficiency of the judicial system - Rule of law - Corruption - Risk of expropriation (i.e. outright confiscation or forced nationalization) by the government - Likelihood of contract repudiation by the government - Additionally, the rules of disclosure/quality of accounting standards are included Research done in field of capital structure with respect to corruption by Baxamusa and Jalal (2013) reveals that corruption increases the costs of debt and equity. Furthermore it is interested to note that as the level of corruption increases, the costs of debt and equity become more sensitive to changes in corruption. The researchers recommend that for 27

28 most corrupt countries, it is necessary to take some measures in order to pass a minimum threshold before they actually can influence the company s behavior. Additionally, countries with less influence of corruption need to remain vigilant against increases in corruption. Market- versus bank-oriented A capital market is a market for securities, where governments and firms can raise longterm funds. It includes both the stock and bond market. Capital markets channel savings and investment between suppliers of capital and users of capital. The size of a country s capital market is directly related to the size of its economy (e.g. the world s largest economy; the United States, has the biggest capital market in the world). In bank-oriented systems, like Germany and Japan, banks play a leading role in mobilizing savings, allocating capital, monitoring investment decisions, and providing risk management vehicles. Advocates of bank-oriented capital markets claim that their bank-centered structure fosters patient capital markets and long-term planning, while a stock market-centered is said to encourage short-term opportunities by managers. Advocates of stock market-centered systems emphasize the adaptive characteristics of a market for corporate control which are lacking in bank-oriented systems, further they mention the lack of empirical evidence towards their short-termism. Law is found to be an important determinant of the banking development. Countries where the legal system emphasizes creditor rights and enforces contracts have better developed banks than countries where laws do not give a high priority to creditors and where enforcement is lax (Levine, 1998). Furthermore countries with a common Law tradition, strong protection of shareholder rights, good accounting regulations, low levels of corruption, and no explicit deposit insurance tend to be more market based 28