The differences in capital structure between the G-7 countries and the E-7 countries

Transcription

1 The differences in capital structure between the G-7 countries and the E-7 countries How the determinants of the capital structure influence the differences in capital structure between the G-7 and the E-7 countries Supervisor: Author: Erik von Schedvin Rachel van Esch 1

2 The differences in capital structure between the G-7 countries and the E-7 countries How the determinants of the capital structure influence the differences in capital structure between the G-7 and the E-7 countries Supervisor: Department: Author: ANR: Study Program: Faculty: Erik von Schedvin Finance Rachel van Esch s International Business Administration Faculty of Economics and Business Number of words: 6686 Date: May 27,

3 Introduction In January 2011 the company PricewaterhouseCoopers came with a groundbreaking report. They stated that by the year of 2050, the E-7 countries would have larger economies than the G-7 countries. The G-7 countries are; the United States, Canada, France, the United Kingdom, Germany, Italy and Japan. The E-7 countries are; China, India, Brazil, Mexico, Russia, Indonesia and Turkey. Off course, these E-7 countries would not be able to have larger economies without the firms based in these countries. This made me wonder whether there are differences in the capital structure between the largest firms in those countries. From the G-7 countries it is known that the determinants of the capital structure are fairly equal for every country (Rajan & Zingales, 1995). Rajan & Zingales (1995) argue that these determinants are; the tangibility of assets, the market-to-book ratio, firms size and profitability. More and more investors are investing in the emerging markets (E-7 countries). Therefore it seems logical that these markets are growing. Booth, Aivazian, Demirguc-Kunt & Maksimovic (2001) and Kayo & Kimura (2011) both found that the determinants of the capital structure of de developed countries are the same as those for the developing countries. Athough Kayo & Kimura (2001) used a slightly different sample of countries than Booth et al. (2001), the results remained the same. In this thesis I explore whether there is a difference between those two groups of countries with regard to their capital structure and which factors cause this difference. This leads to my research question: Which factors cause the difference in the capital structure choice between the G-7 countries and the E-7 countries? For the empirical analysis, data from 13,360 companies is used. All G-7 and E-7 countries are present in this dataset. According to the descriptive statistics, there are not many differences between the G-7 countries and the E-7 countries with regard to both leverage ratios, total leverage and long-term 3

4 leverage. While the G-7 countries show a slightly higher total leverage ratio, the E-7 countries show a slightly higher long-term leverage ratio. The E-7 countries are more profitable than the G-7 countries. Next to that, the E-7 countries are on average larger than the G-7 countries and the tangible assets ratio was also larger for the E-7 countries compared to the G-7 countries. There is a negative relationship between profitability and both leverage ratios, while the relationships between tangible assets and size with these leverage ratios is positive. The relationships between tangible assets and both leverage ratios turn out to be the strongest. However, profitability has also a relatively strong relationship with the total leverage ratio. 4

5 Contents Literature review... 6 Pecking order theory... 6 Static trade-off theory... 8 Agency theory... 9 Behavioral approach Empirical findings theory Data overview Empirical setup Empirical results Dependent variable total leverage ratio Dependent variable long-term leverage ratio Economic interpretation Conclusions and recommendations Conclusions Limitations Recommendations References Appendix Table 1: Descriptive statistics of total leverage ratio Table 2: Descriptive statistics of long-term leverage ratio Table 3: Descriptive statistics of profitability Table 4: Descriptive statistics of tangible assets ratio Table 5: Descriptive statistics of company size Table 6: Regression analysis for total leverage ratio Table 7: Regression analysis for the long-term leverage

6 Literature review The proposition Modigliani & Miller holds that the value of the levered firm should not be different from the value of the unlevered firm, if we live in perfect capital markets (Modigliani & Miller, 1958). Thus, the choice between equity and capital has no effect on firm value or on the cost of capital. The insights of Modigliani & Miller are now widely accepted. However, in markets that are not perfect, the capital structure does matter. That is because taxes are included, just like information and agency costs. The pecking order theory and the static trade-off theory have different views on how to choose between equity and capital when these factors play a role. Pecking order theory The pecking order theory states that firms prefer internal financing over external financing. If for some reason external financing is required, debt is preferred over equity. This theory exists due to asymmetric information costs, transaction costs and the incentives of the owner/manager not to lose control over his company. When managers have information from the inside that other investors do not have, we speak of asymmetric information. The costs associated with this are called the asymmetric information costs. These costs can increase opportunistic behavior, since managers can issue new shares when the current shares are overpriced. This situation arises because the investors have less information and do not always know when equity is mispriced (Myers & Majluf, 1984). However, most investors are aware of the fact that firms tend to issue new shares when the firm believes that the existing shares are overpriced. If the investors then adjust the price they are willing to pay downwards, the new shares can become underpriced. If the new shares are underpriced too much, the firm has to reject positive NPV projects in order to protect the existing shareholders (Swinnen, Voordeckers and Vandemaele, n.d.). In order to avoid the asymmetric information costs, firms could use other forms of financing. 6

7 Transaction costs are the largest for equity financing, followed by debt financing and then internal financing. In order to avoid these transaction costs, firms try to fund their positive NPV projects by internal resources first. Another reason to prefer internal financing over external financing has to do with control. The owner / manager of a company does not like to lose his control over the firm (Holmes & Kent, 1991; Hamilton & Fox, 1998). Therefore, the owner / manager tries to keep control by not bringing on board new shareholders and thus tries to finance his positive NPV projects by internal funds (Swinnen et al., n.d.). If these internal funds are not available, debt financing is preferred over equity financing since the owner/manager looses less control with debt financing as opposed to equity financing. The pecking order theory does not claim that there is an optimal capital structure, contrarily to the static trade-off theory. There is no optimal capital structure since there are two kinds of equity namely retained earnings at the top of the pecking order and shares at the bottom of the pecking order (Myers, 1984). Myers (1984) stated that according to the simple pecking order theory, a company s debt-equity ratio varies over time, dependent on the need for external financing. If firms have a negative cash-flow they are expected to increase their debt ratio, while when they have a positive cash flow it is expected that firms decrease their debt ratio (Myers, 1984). However, according to the complex view of the pecking order model, firms are also concerned about their future financing costs (Myers, 1984). If firms are concerned about the future finance costs and future growth opportunities, they favor a lower debt ratio (Swinnen et al., n.d.). Myers (1984) argues that reserve-borrowing capacity is valuable for a firm. A firm may issue new common stock if it has not yet reached its borrowing capacity, since this give the firm the possibility to go for future profitable investments (Myers, 1984). 7

8 Static trade-off theory According to the static trade-off theory, there does exist an optimal capital structure. The firm should choose its capital structure by balancing the value of the interest tax shields against financial distress costs (Myers, 1984). Therefore, the trade-off theory predicts that tax-paying firms will borrow moderately (Myers, 2001). Those firms borrow up to the point that the benefits of the tax shields are offset by the increase of financial distress costs (Myers, 2001). If firms are not at their optimal capital structure, they should move towards it. As for example the costs of financial distress are higher than the gains of the tax shields, the firm should decrease its debt-level. Contrary, firms with a debt-ratio below the optimal debt-ratio should increase their level of debt. If this reasoning is followed, firms with low financial distress costs are expected to have a high level of leverage and thus high debt-ratios. This however is not always true. As Myers (2001) pointed out, firms like Microsoft and other pharmaceutical companies have low debt-ratios while they have superior credit ratings, and thus are nowhere near financial distress. In 1984, Myers already pointed out that the most profitable firms tend to have the lowest debt ratios. These findings seem to be in contrast to the static trade-off theory, since this theory suggests that the most profitable firms borrow more due to the low probability on financial distress. They would have more income and thus gain more by exploiting the tax shields (Myers, 2001). The static trade-off theory cannot explain this phenomenon. Miller (1977) did have an answer to this. Miller (1977) argued that the low effective tax gains on capital gains diminish the value of the corporate tax shields. The investors pay more taxes on debt interest than on equity income and therefore cause the company to favor equity. The static trade-off theory seems to be consistent with some obvious facts about financing. It seems for example very logical that firms with risky assets tend to borrow less, due to the higher probability of financial distress. 8

9 According to the study of Flannery & Rangan (2004), there is strong evidence that nonfinancial firms pursued target ratios and thus an optimal capital structure. Those targets depend on the specific firm characteristics (Flannery & Rangan, 2004). Agency theory The agency theory is involved with resolving two problems that occur in the relationships between the agent and the principal (Eisenhardt, 1989). The first problem can arise when the desires or goals of the principal, differ from those of the agent (Eisenhardt, 1989). Eisenhardt (1989) also states that it can be hard for the principal to verify exactly what the agent is doing. The second problem is that the agent may have a different attitude than the principal towards risk and therefore they may prefer different actions (Eisenhardt, 1989). The agency theory has two streams; the positivist agency theory and the principal-agent theory (Jensen, 1983). The positivist agency theory is aiming to find solutions to the agency problems in business. It has become widely accepted since the 1980 s, since the positivist agency theory does not only look at the relation between the principal and the agent, but it expands its view to all the different relationships between different principals and different agents within an organization (Eisenhardt, 1989). The principal-agent theory is the more general view. This theory is concerned about the relationships between one principal and one agent (Harris & Raviv, 1978). This theory involves a careful specification of assumptions, which have to be followed by logical deduction and then mathematical proof (Eisenhardt, 1989). The simple model of the principal-agent theory assumes that that there exist a goal-conflict between the principal and the agent, an easily measured outcome and an agent who is more risk averse than the principal (Eisenhardt, 1989). According to Leland (1998), the agency costs are modest since they are less than two percent of the tax benefits associated with debt. The agency costs of firms without risk flexibility are even lower, only one fifth percent (Leland, 1998). The presence of agency costs decreases the level of leverage and 9

10 thus changes the capital structure (Leland, 1998). According to Leland (1998), the expected maturity falls, which was also what Myers (1977) had predicted. Behavioral approach The assumption in the static trade-off theory, as well as in the pecking order theory, is that the decision-making process is purely rational, which leads to wealth maximization (Swinnen et al., n.d.). However, according to Simon (1997), decision-makers cannot make a fully rational decision and the limitations of the decision-maker must be taken into account. Bounded rationality is due to the incompleteness and uncertainty of information that decision makers face (Simon, 1979). Therefore, according to Simon (1979), the traditional paradigm of maximization should be substituted by a more realistic approach. Therefore, he came up with the idea of satisficing. A decision maker who chooses the best available alternative according to some criterion is said to optimize; one who chooses an alternative that meets or exceeds specified criteria, but is not guaranteed to be either unique or in any sense the best, is said to satisfice (Simon, 1997). Individuals simplify the decision making process, because they have a hard time considering all the alternatives and all the relevant information (March, 1978). Next to that, biases and heuristics do influence the decision-making process. By using biases and heuristics, the decision-maker is able to deal with complex problems (Simon, 1957). Although it seems that these biases and heuristics would hardly enhance the decision-making effectiveness, most of the times these simplifications lead to acceptable solutions for people faced with an uncertain and complex situation (Busenitz & Barney, 1997). The people large organizations that face such an environment are the managers. However, managers of larger organizations tend to have more resources available to make the optimal decision (Vermeulen & Curşue, 2008). If a manager or owner does not have enough knowledge to make the optimal decision, he could look at other similar firms for guidance. This is called the behavioral principle (Emery, Finnerty & Stowe, 2004). According to this principle, industry averages can be used as financial targets, because managers / owners of small firms do have a limited understanding on the subject of capital structure (Emery, Finnerty & Stowe, 2004). 10

11 Empirical findings theory Rajan & Zingales (1995) argue that firms are fairly similar levered across the G-7 countries. Only the United Kingdom and Germany are relatively less levered, although the differences are not easily explained by institutional differences (Rajan & Zingales, 1995). According to Rajan & Zingales (1995), the factors that previous studies have identified as important for leverage in the United States, seem similarly related in other countries as well. However, Rajan & Zingales (1995) have focused on four factors; tangibility of assets, market-to-book ratio, firm size and profitability. Those factors are chosen because they have shown up most consistently in correlation with leverage (Rajan & Zingales, 1995). The study described above, suggests how these four factors might be related to leverage. If a firm has many tangible assets, they can be used as security. The lender can take these assets if the firm is unable to make a payment. Thus the greater the proportion of tangible assets in a company, where the proportion is defined as fixed assets divided by total assets, the more willing the lenders should be to supply loans and this in turn should increase leverage (Rajan & Zingales, 1995). According to Myers (1984), this is due to the fact that tangible assets can be reallocated in case of bankruptcy or when firms are not able to make a payment. This is however not true for firms with intangible assets, since these assets are more firm-specific (Myers, 1984). As stated above, highly levered companies are more likely to pass up profitable investment opportunities (Myers, 1977). Therefore, according to the static-trade off theory, firms who expect a high future growth should use a greater amount of equity finance (Rajan & Zingales, 1995). There are conflicting arguments about the effect of size on leverage, according to the capital structure theories. Larger firms have a more diversified portfolio and fail less often, thus the larger firms are expected to have a lower probability of bankruptcy (Rajan & Zingales, 1995). However, larger companies typically have more outside investors. Those outside investors are expected to have more information about larger firms, which increases the preference for equity relative to debt (Rajan & Zingales, 1995). 11

12 According to Myers & Majluf (1984) and the pecking order theory, there should be a negative relationship between profitability and leverage. Myers & Majluf (1984) argued that firms prefer to finance with internal funds rather than with debt. The results of Rajan & Zingales (1995) indicate that; tangibility is always positively related to leverage (in all countries), the idiosyncratic component of the market-to-book ratio for a firm is as important as the industry market to book; size is positively correlated with leverage except in Germany and profitability is negatively correlated with leverage in all countries except for Germany. Overall, the factors that were said to be correlated with leverage in the United States tend to be similarly correlated in the other G-7 countries (Rajan & Zingales, 1995). Titman & Wessels (1988) found that smaller firms use more short-term debt compared to larger firms. No evidence was found that debt ratios can be predicted by expected growth, non-debt tax shields, volatility or the collateral value of its assets (Titman & Wessels, 1988). Titman & Wessels (1988) found some evidence that profitable firms have less debt. The studies of Rajan & Zingelas (1995) and Titman & Wessels (1988) show fairly different results. This can be explained by the fact that Titman & Wessels did their study several years before Rajan & Zingales did. Therefore, it seems logical that Rajan & Zingales did have more knowledge and that there was more information available. Kayo & Kimura (2011) argued that a significant part of the leverage variance is due to intrinsic firm characteristics. Kayo & Kimura (2011) agreed with Rajan & Zingales (1995) about the determinants of the capital structure. The significant variables are; growth opportunities, profitability, size and tangibility (Kayo & Kimura, 2011). The variables tangibility and size have a positive relation with leverage, while size had a positive relation with debt (Kayo & Kimura, 2011). Growth opportunities show a negative relation with leverage (Kayo & Kimura, 2011), which is also shown in the studies of Rajan & Zingales (1995) and Booth, Aivazian, Demirguc-Kunt & Maksimovic (2001). Profitability has a negative relationship with leverage (Kayo & Kimura, 2011), Booth et al. (2001) found the same result for the emerging markets. 12

13 Booth et al. (2001) found that the variables which are relevant for explaining the capital structure of firms in the United States and Europe are also relevant for the developing countries. It should be noted however, that Booth et al. (2001) did use some other countries than those of the E-7 countries. India, Brazil, Turkey and Mexico are the four countries of the E-7 used in their sample. Next to those, they used; Pakistan, Thailand, Malaysia, Zimbabwe, Jordan and Korea. In general, the debt ratios of the developing countries seem to be affected by the same determinants and in the same way as the determinants of the developed countries (Booth et al. 2001). However, there are some differences in the way these ratios are affected by country factors, for example; inflation rates, the development of capital markets and GDP growth rates (Booth et al. 2001). Booth et al. (2001) also note that that knowing the country is important, it is as least as important as knowing the size of the independent variables. This is size is important for both the total and the longterm book-debt ratio s, it should be noted that this is not true for the market-to-book ratios (Booth et al. 2001). Kayo & Kimura (2011) investigated whether the G7 countries had the same determinants, with their accompanying relation towards the capital structure choice, as the emerging countries. For the emerging markets, they choose to use the countries; Brazil, Mexico, South-Korea, India, Malaysia, Pakistan, Thailand and Turkey. Their sample is based on the sample of Booth et al. (2001), but they left Zimbabwe and Jordan out. For the determinants; growth opportunities, profitability, size and tangibility, the same signs and statistical significances are shown for the developed countries as for the emerging countries (Kayo & Kimura, 2011). 13

14 Data overview For this research, data was needed from different countries. In total, 13,360 companies provided data. They were divided over the following countries; the United States (3165), Canada (681), Germany (603), France (570), the United Kingdom (1492), Italy (213), Japan (3030), Brazil (194), China (1544), Indonesia (208), India (1389), Mexico (87), Russia (88) and Turkey (87). These are the number of companies that have passed the selection criteria. First, the companies with a standard industrial classification (SIC) of between 6000 and 6999 were excluded, since those codes indicate financial firms. Next to that, the companies with a SIC code between 4900 and 4999 were excluded, since those codes indicate utility firms. Utility and financial firms are excluded, since their capital structure is strongly influenced by explicit (or implicit) investor insurance schemes, as for example deposit insurance (Rajan & Zingales, 1995). Next to that, the debt-like liabilities of utility and financial firms are not comparable to the debt issued by non-financial companies (Rajan & Zingales, 1995). Thereafter, some boundaries were set to the different variables in order to exclude outliers. For variable 1, total leverage ratio, only the companies with a value between zero and one are included. This is the same for variable 2, the long-term leverage ratio. For variable 3, return on assets, only the firms with a value between minus one and one are included. For variable 5, size, all firms with a negative size value were excluded, since size cannot be negative. [insert table 1 here] Table 1 shows the descriptive statistics for the ratio of total debt divided by total assets, hereafter called total leverage ratio. From this figure indicates that there are not many differences between the G-7 countries and the E-7 countries when it comes to the total leverage ratio. If the average means of those leverage ratios is taken, it can be seen that the G-7 countries are slightly higher levered (0,514 on average for the G-7 countries, compared to 0,507 on average for the E-7 countries). Next to that, the standard 14

15 deviations and therefore the 95% confidence interval are small for all countries. This means that most firms in each country have around the same total leverage ratio. [insert table 2 here] The descriptive statistics of the leverage ratio of long-term debt divided by total assets, hereafter called long-term leverage ratio, can be seen in table 2. When it comes to the long-term leverage ratio, it seems that also here are not many differences between the G-7 and the E-7 countries. The average longterm leverage ratio for the G-7 countries (0,123) is slightly lower than the average for the E-7 countries (0,132). Outstanding is the ratio for China, which is really small. The standard deviations are also small for the long-term leverage ratio, and thus the 95% confidence interval is also small. This indicates that most firms in each country have around the same long-term leverage ratio. [insert table 3 here] Table3 shows the descriptive statistics for the return on assets (earnings before interest divided by total assets). This figure shows that there is a difference between the return on assets of the G-7 countries compared to the E-7 countries. It is evident that the return on assets is significantly larger for the E-7 countries. The average return on assets for the G-7 countries equals 0,096 while the average for the E-7 countries equals 0,127. This difference is mostly due to the extremely low return on assets of the United Kingdom and the really large return on assets of Russia. Also here the standard deviations are very small and thus the 95% confidence interval is also small. 15

16 [insert table 4 here] The descriptive statistics for the tangibility of assets (fixed assets divided by total assets) can be seen in table 4. Also here a difference between the G-7 countries and the E-7 countries can be seen. The E-7 countries have an average tangibility of assets ratio of 0,386 while the G-7 countries have an average of 0,247. This is mostly due to the fact that most of the G-7 countries have a ratio of around 0.20 (except for Canada, which has an extremely high ratio compared to the other G-7 countries). The standard deviations for this variable are also higher compared to the other variables, which lead to a larger 95% confidence interval. Therefore, it can be said that the data for this variable is more divers than for the other variables. [insert table 5 here] Table 5 shows the descriptive statistics for the size (the natural logarithm of total assets). Here too seems that there is a difference between the G-7 countries and the E-7 countries. The E-7 countries seem to be much larger (based on total assets) than the G-7 countries, since the E-7 countries show an average size of 8,601 and the G-7 countries show an average size of 5,958. Next to that, the standard deviations for the G-7 countries are larger than for the E-7 countries. Outstanding here is the average size of the firm from Indonesia, which is substantially larger than the average of all other countries. 16

17 Empirical setup The dependent variable should be the capital structure of the countries. This dependent variable will be determined by the use of two ratio s namely; long-term leverage and total leverage. Long-term leverage will be determined by dividing long term debt (DLTT) by total assets (AT). The total leverage variable will thus also include the short-term debt. This ratio will be calculated by dividing total liabilities (LT) by total assets (AT). The independent variables are; tangible assets, firm size and profitability of each company in the year The tangible assets ratio is determined by dividing fixed assets (PPENT) by total assets (AT). The total assets (AT) is a measure of the firm size. However, since the measure total assets gives high values, I choose to use the logarithm of total assets. Otherwise, it is hard to compare the size of companies. The profitability of each company is determined by the return on assets, thus net income divided by total assets (AT). For net income I have chosen to use earnings before interests, taxes, depreciation and amortization (EBITDA). I have chosen EBITDA because it gives a clearer overview, since differences in regulation between countries are avoided as extensive as possible. In total, eight regression analyses are conducted. The first analysis has the following equation:. This means that one dummy is made in order to compare the E-7 countries to the G-7 countries, the G-7 countries are thus taken as a base level. The is the constant in the formula. Thereafter, a second regression with the following equation is conducted:. In this analysis, the E-7 countries are again compared to the G-7 countries, but the independent variables are included here. Return on assets (profitability) is denoted by tangible assets is denoted by and size is denoted by equation:. The third regression includes the country dummies and has the following. In total there are 13 country dummies, all of them are compared to the United States. These dummies are used to compare the E-7 countries and the G-7 countries on a country level, whereas the E-7 dummy compares the two groups on a group level. In the fourth model, all 17

18 variables and country dummies are included, which leads to the following equation:. Those four regression analyses are made for both dependent variables, which brings the total to eight models. 18

19 Empirical results In order to state whether there are differences in the capital structure between the G-7 countries and the E-7 countries, regression analyses are conducted. In order to compare those two country groups, several dummies are made. There is one dummy for all the E-7 countries together. This dummy is used to compare the G-7 countries and the E-7 countries as a group, thus not on a country level. In order to see the differences on a country level, 13 other dummies are made. There are 14 countries in the sample and all countries are compared to the United States, which is taken as base level. In this thesis, eight models are made in order to assess the differences. In the first model, only the E-7 country component is included to predict the dependent variable. In the second model, the E-7 country component together with the independent variables; profitability, tangible assets ratio and size were included to predict the dependent variable. In the third model, only the country dummies are included to predict the dependent variable. In the fourth model, the country dummies, together with the independent variables; profitability, tangible assets ratio and size, are included to predict the independent variable. These four models are made for both the total leverage ratio and the long-term leverage ratio, therefore there are in total eight models. Dependent variable total leverage ratio [insert table 6 here] From model 1 it can be derived that the E-7 countries have a lower total leverage ratio than the G- 7 countries. This is because the E-7 country component is negative (-0.066). This is confirmed by the second model, since the E-7 country component here equals From the second model it can also be derived that profitability has a negative influence on the total leverage ratio, while the tangible assets ratio and size have a positive influence on the dependent variable. 19

20 The results of model 1 and model 2 seem to go against the results of model 3. In model 3, all the significant country components for the E-7 countries are positive, while there are negative components for the G-7 countries. Especially the f or the United Kingdom has a large negative value. However, the fourth model gives the same idea as the first and second model. According to the fourth model, profitability has again a negative influence on the total leverage ratio, while the tangible assets ratio and size have a positive influence on the dependent variable. Next to that, most country components of the E- 7 countries have a negative beta. Therefore, it can be said that the E-7 countries have a lower total leverage ratio compared to the G-7 countries. Next to that, it can be derived from both model 2 and model 4 that the tangible assets ratio has the greatest positive influence on the total leverage ratio. However, profitability has a strong negative relationship with the total leverage ratio. Model 4 shows the largest of all the four models and therefore explains the most of the variance in the dependent variable. Therefore, this model is the best model out of those four models. Dependent variable long-term leverage ratio [insert table 7 here] From model 5 it can be derived that the E-7 countries have a lower long-term leverage ratio than the G-7 countries. This is because the E-7 country component is negative (-0.026). However, it should be noted that this relationship is not as strong as the negative relationship with the total leverage ratio, since the beta is much lower. Next to that, this beta is only significant on a 10 percent level. The sixth model gives the same result as the fifth model. Again, it shows a negative relationship (-0.033) between the E-7 country component and the long-term leverage ratio. Next to that, profitability has a negative relationship with long-term leverage ratio, while this relationship for the tangible assets ratio and size is positive. The 20

21 strongest determinant for the long-term leverage ratio in this model is the tangible assets ratio, since it has a high beta. The seventh model indicates that all country components, except for India, are negative. This indicates that all countries, except India, are less levered than the United States. It seems that those betas for the E-7 countries are less negative, except for Turkey, than the betas of the G-7 countries. This implies that the E-7 countries are closer to the United States in terms of long-term leverage. In the eighth model, there is again a negative relationship between profitability and long-term leverage, while this relationship is positive for the tangible assets ratio and size. Again, all the country components show a negative beta. However, contrary to the seventh model, the betas of the E-7 countries are more negative than those of the G-7 countries. Next to that, both model 6 and model 8 show that tangible has the greatest influence on the longterm leverage ratio. However, model 8 shows that profitability has also a fairly large impact on the longterm leverage ratio. Since model 8 has the highest, it could be concluded that this model explains the most of the variance in the dependent variable. Therefore, this model is the best model out of those four models. Economic interpretation It could be that highly profitable firms have a lower tendency to borrow, because the internal funds are sufficient to finance their projects. This could explain the negative relationship between profitability and leverage. This is consistent with the pecking order theory, since those firms will prefer internal financing over external financing. If a firm chooses to finance his project internally, it has more funds left if external financing is needed. A high level of tangible assets leads to more possibilities for the firm to leverage. The tangible assets can be used as a security, since the lender can take them if the firm is unable to make a payment. Next to that, tangible assets are mostly not firm specific and could thus be reallocated in times of financial 21

22 distress, which is not true for intangible assets (Myers, 1984).Thus, if firms have a high level of tangible assets, the lenders should be more willing to supply loans to those firms (Rajan & Zingales, 1995). This would imply a positive relationship between the tangible assets ratio and the leverage ratios, which is indeed a finding in this thesis. Small firms are expected to use more short-term financing because of the transaction costs involved with long-term financing. Larger firms do have a larger borrowing capacity, since the risk of bankruptcy is smaller. This can for example be due to diversification strategies. However, larger firms typically have more outside investors, which causes a preference for equity relative to debt (Rajan & Zingales, 1995). Therefore, the theory of the relationship between size and leverage is not consistent. This could explain the low beta of the size variable, which indicates that size has only a small influence on the leverage ratios. 22

23 Conclusions and recommendations Conclusions In this thesis, the differences between the capital structure and their determinants between the G-7 countries and the E-7 countries are tested. From the data overview it can be concluded that the differences between the G-7 countries and the E-7 countries are very small with regard to the total leverage ratio, with the notion that the G-7 countries are slightly higher levered. For the long-term leverage ratio it can also be stated that the differences between the G-7 countries and the E-7 countries are fairly small. The E-7 countries have a slightly higher long-term leverage ratio compared to the G-7 countries. According to Rajan & Zingales (1995) the firms across the G-7 countries are fairly similarly leveraged, with the United Kingdom and Germany as less levered countries. In the sample of this research, the two less levered countries, based on total leverage ratio, would be Canada and the United Kingdom. On the basis of long-term leverage ratio, the two less levered countries would be Japan and the United Kingdom. Therefore it seems safe to argue that the United Kingdom is less levered than the other G-7 countries, which is in accordance with the findings of Rajan & Zingales (1995). However, in this thesis Germany does not come forward as a less levered country. For the E-7 countries, Turkey has a lower leverage on both leverage variables than the other countries. This is not really unexpected since their tangible of assets ratio and their average company size is also fairly small compared to the other E-7 countries. Profitability shows a negative relationship with both leverage ratios. However, it should be noted that this relationship is only significant when all country dummies are included, thus in model 4 and model 8. This negative correlation is in accordance with Myers & Majluf (1984), who argue that firms prefer internal financing over external financing. It is also in accordance with Meyers (1984), who pointed out that the most profitable firms tend to have to lowest debt ratios. These findings are in contrast 23

24 with the static trade-off theory, since this theory suggests that firms with a low possibility of financial distress, and thus a high profitability ratio, are expected to have higher debt levels and thus higher leverage ratios. However, the most profitable firms borrow less due to the low probability of financial distress. This negative relationship is also in accordance with the findings of Rajan & Zingales (1995). That this relationship is negative is also true for the emerging markets, which is in accordance with the findings of Booth et al. (2001) and Kayo & Kimura (2011). According to Rajan & Zingales (1995), a higher the tangibility of assets should lead to a higher the debt ratio. This is because tangible assets can be used as a security, so that when a borrower is not able to make a payment, the lender could take these tangible assets. This thesis shows a positive relationship between tangible assets and both leverage ratios, which is in accordance with the findings of Rajan & Zingales (1995). According to Booth et al. (2001) and Kayo & Kimura (2001), these results would also be true for the emerging countries, which is also the case in this thesis. In this thesis, the relationship between the tangibility ratio and the leverage ratio turned out to be the strongest, since this variable showed the highest beta. Size is positively related with both leverage ratios, but this relationship is not a strong one, due to the low beta. The study of Rajan & Zingales (1995) also showed a positive relationship between size and leverage for the G-7 countries. The same relationship between size and leverage was shown for the E-7 countries by both Booth et al. (2001) and Kayo & Kimura (2011). Limitations The use of consistent accounting practices is not taken into account. It is highly likely that accounting practices differ across countries and even across firms, therefore this should be taken into account, which is not done in this thesis. Furthermore, not all countries do have a large sample size, which especially true for some of the E-7 countries (Mexico, Russia and Turkey). If a country has a small sample size, it means that one individual firm has more influence on the average outcome. This is unfavorable, since one firm could than influence the outcomes. This could then influence the 24

25 comparisons that are made. Also, only the listed firms are taken into account here. However, the fraction of listed firms is not the same in every country, just as the average size of listed companies (Rajan & Zingales, 1995). Another aspect that limited this thesis was the use of only one measurement per variable. The use of more measurements per variable lead to outcomes that are more reliable. Next to that, firms in different countries are exposed to different regulations and taxes. For the variable profitability, the measurement earnings before interest, taxes and amortization is used in order to exclude as much of this different regulations as possible. However, it is highly likely that other variables are influenced by different regulations and taxations in different countries. This may have affected the outcomes of this thesis. Recommendations Further research should more extensively explore the differences between the G-7 countries and the E-7 countries with regard to the determinants of the capital structure. The differences in accounting practices in different countries should then be taken into account in order to make a better comparison between the leverage ratios of the G-7 and the E-7 countries. The different regulations in the different countries should be taken into account by complementary research, since it is likely that the variables used in this thesis are affected by different regulations. As financial and utility firms are excluded from this research, supplementary research could investigate the differences in the capital structure of those firms and their determinants for different countries. 25

26 References Booth, l., Aivazian, V., Demirguc-Kunt, A., & Maksimovic, V. (2001). Capital Structures in Developing Countries. The Journal of Finance, Vol. 56., No. 1., pp Busenitz, L.W. & Barney, J. (1997). Differences between entrepreneurs and managers in large organizations: biases and heuristics in strategic decision-making. The journal of Business Venturing, Vol. 12, No. 1., pp Eisenhardt, K.M. (1989). Agency Theory: an Assessment and Review. The academy of management review, Vol. 14., No. 1., pp Emery, D.R., Finnerty, J.D. & Stowe, J.D. (2004). Corporate Financial Management. Upper Saddle River: Prentice Hall. Flannery, M.J., Rangan, K.P. (2006). Partial adjustment towards target capital structures. Journal of Financial Times, Vol. 79., pp Hamilton, R.T. & Fox, M.A. (1998). The financing preferences of small firm owners. International Journal of Entrepreneurial Behavior & Research, Vol. 4, No. 3., pp Harris, M.& Raviv, A. (1978). Some results on incentive contracts with application to education and employment, health insurance and law enforcement. The American Economic Review, Vol. 68., pp

27 Holmes, S. & Kent, P. (1991). An empirical Analysis of the Financial Structure of Small and Large Australian Manufacturing Enterprises. Journal of Small Business Finance, Vol. 1, No. 2., pp Jensen, M. (1983). Organization theory and methodology. The Accounting Review, Vol. 56., pp Kayo, E.K. & Kimura, H. (2011). Hierarchical determinants of capital structure. The Journal of Banking and Finance, Vol. 35., No. 2., pp March, J.G. (1978). Bounded rationality, ambiguity, and the engineering of choice. The Bell Journal of Business Finance and Accounting, Vol. 9, No. 2., pp Miller, M. (1977). Debt and Taxes. The Journal of Finance, Vol. 32, No. 2., pp Modigliani, F. & Miller, M.H. (1958). The Cost of Capital, Corporation Finance and the Theory of Investment. The American Economic Review, Vol. 48, No. 3., pp ). Myers, S.C. (1984). The capital structure puzzle. Journal of Finance, Vol. 34. No. 3., pp Myers, S.C. (2001). Capital structure. The Journal of Economic Perspectives, Vol. 15, No. 2., pp Myers, S.C. & Majluf, N. (1984). Corporate Financing and Investment Decisions when Firms have Information Investors do not have. Journal of Financial Economics, Vol. 13, No.2., pp Leland, H.E. (1998) Agency Costs, Risk Management and Capital Structure. The Journal of Finance, Vol. 53., No. 4., pp

28 Rajan, R.G. & Zingales, L. (1995). What Do We Know about Capital Structure? Some Evidence from International Data. The journal of Finance, Vol. 50., No. 5., pp Sheifler, A. & Vishney, R. (1992). Liquidation values and debt capacity: A Market Equilibrium Approach. The Journal of Finance, Vol. 47, pp Simon, H.A. (1997). Models of bounded rationality, empirically grounded economic reason. Massachusetts Institute of Techonology Simon, H.A. (1957). Administrative Behavior: A Studyof Decision-making Processes in Administrative Organization. London and New York: Macmillan Swinnen, S., Voordeckers, W. & Vandemaele, S. (n.d.). Capital structure in SMEs: Pecking order versus static trade-off, bounded rationality and the behavioural principle. Titman, S. & Wessels, R. (1988). The Determinants of Capital Structure Choice. The Journal of Finance, Vol. 43., No. 1., pp Vermeulen, P.A.M. & Curşue, P.L. (2008). Entrepreneurial strategic decision-making, a cognitive perspective. Edward Elgar Publishing Limited. 28

29 Appendix Table 1: Descriptive statistics of total leverage ratio Table 1 shows the descriptive statistics of the total leverage ratio, which is defined as total leverage divided by total assets, for each country. Also, an average per group is shown. Total leverage ratio (var1 = LT/AT) G-7 the United States Canada Germany France the United Kingdom Italy Japan Average G-7 E-7 Brazil China Indonesia India Mexico Russia Turkey Average E-7 Mean Median Lower bound 95% confidenc e interval 0,477 0,426 0,536 0,582 0,452 0,600 0,522 0,514 0,545 0,511 0,522 0,541 0,481 0,495 0,452 0,507 0,480 0,412 0,573 0,602 0,471 0,599 0,534 0,524 0,551 0,527 0,526 0,577 0,476 0,504 0,455 0,516 0,469 0,412 0,519 0,567 0,440 0,575 0,515 0,499 0,516 0,502 0,494 0,530 0,438 0,454 0,408 0,478 Upper bound 95% confidenc e interval 0,484 0,441 0,553 0,598 0,465 0,625 0,530 0,528 0,574 0,520 0,551 0,552 0,523 0,536 0,497 0,536 Standar d deviatio n 0,219 0,192 0,217 0,189 0,250 0,187 0,204 0,208 0,205 0,185 0,211 0,202 0,202 0,196 0,211 0,202 Number of companie s

30 Table 2: Descriptive statistics of long-term leverage ratio Table 2 shows the descriptive statistics of the long-term leverage ratio, which is defined as long-term leverage divided by total assets, for each country. Also, an average per group is shown. Long-term leverage ratio (var2 = DLTT/AT) Mean Median Lower bound 95% confidence interval Upper bound 95% confidence interval Standard deviation Number of companies G-7 the United States Canada Germany France the United Kingdom Italy Japan Average G-7 E-7 Brazil China Indonesia India Mexico Russia Turkey Average E-7 0,168 0,133 0,113 0,118 0,101 0,136 0,092 0,123 0,143 0,052 0,147 0,188 0,155 0,158 0,082 0,132 0,117 0,089 0,059 0,077 0,025 0,108 0,058 0,076 0,116 0,011 0,091 0,150 0,124 0,133 0,050 0,096 0,162 0,122 0,102 0,107 0,093 0,120 0,088 0,114 0,124 0,048 0,125 0,179 0,126 0,130 0,061 0,113 0,175 0,144 0,125 0,128 0,108 0,151 0,095 0,132 0,163 0,057 0,169 0,198 0,184 0,186 0,103 0,151 0,182 0,147 0,143 0,129 0,144 0,118 0,105 0,138 0,139 0,083 0,164 0,172 0,138 0,133 0,101 0,

31 Table 3: Descriptive statistics of profitability Table 3 shows the descriptive statistics of the profitability per country. Profitability is measured by return on assets, which is defined as earnings before interest, taxes, depreciation and amortization divided by total assets. Also, an average per group is given. Return on assets (var3 = EBITDA/AT) G-7 the United States Canada Germany France the United Kingdom Italy Japan Average G-7 E-7 Brazil China Indonesia India Mexico Russia Turkey Average E-7 Mean Median Lower bound 95% confidence interval 0,141 0,146 0,086 0,107 0,019 0,088 0,084 0,096 0,126 0,088 0,107 0,131 0,143 0,184 0,111 0,127 0,126 0,127 0,098 0,110 0,076 0,086 0,077 0,100 0,112 0,079 0,102 0,122 0,130 0,178 0,100 0,118 0,138 0,138 0,074 0,098 0,008 0,075 0,082 0,088 0,111 0,084 0,092 0,126 0,122 0,154 0,094 0,112 Upper bound 95% confidence interval 0,144 0,154 0,097 0,116 0,030 0,101 0,087 0,104 0,141 0,092 0,122 0,137 0,163 0,214 0,127 0,142 Standard deviation 0,091 0,103 0,144 0,111 0,217 0,097 0,067 0,119 0,109 0,082 0,113 0,110 0,097 0,144 0,079 0,105 Number of companies

32 Table 4: Descriptive statistics of tangible assets ratio Table 4 shows the descriptive statistics of the tangible assets ratio, which is defined as property, plant and equipment (net total) divided by total assets. The descriptive statistics are given on a country level and a group average is given. Tangibility of assets (var4 = PPENT/AT) G-7 the United States Canada Germany France the United Kingdom Italy Japan Average G-7 E-7 Brazil China Indonesia India Mexico Russia Turkey Average E-7 Mean Median Lower bound 95% confidence interval 0,252 0,422 0,197 0,157 0,205 0,217 0,276 0,247 0,318 0,386 0,401 0,342 0,434 0,497 0,328 0,386 0,178 0,393 0,142 0,102 0,107 0,166 0,262 0,193 0,325 0,355 0,372 0,321 0,461 0,499 0,321 0,379 0,244 0,400 0,182 0,144 0,193 0,192 0,270 0,232 0,290 0,375 0,370 0,331 0,392 0,455 0,288 0,357 Upper bound 95% confidence interval 0,260 0,444 0,212 0,171 0,217 0,242 0,282 0,261 0,345 0,396 0,432 0,352 0,476 0,540 0,367 0,415 Standard deviation 0,230 0,294 0,187 0,162 0,233 0,184 0,174 0,209 0,197 0,206 0,228 0,205 0,201 0,203 0,188 0,204 Number of companies