Pecking Order Model and Determinants of Capital Structure:
A Comparison between Firms in the
United States and Developing Countries
Master Thesis
Name: Natasha Hick Student number: 6067719
University of Amsterdam Amsterdam Business School
MSc Business Economics, Finance track
Thesis supervisor: Dr. Vladimir Vladimirov Date: July 2014
ABSTRACT
This study investigates the impact of firm-specific and country-specific factors on the capital structure choice of listed firms in the United States and five developing
countries: China, India, Indonesia, Malaysia and Thailand. The sample period is from 1997-2012. Besides investigating the determinants of capital structure, this study also tests the pecking order model in developing countries.
Overall, the results show that the impact of the firm-specific factors on leverage is the same for firms in the United States and the developing countries. The firm-specific factors include: growth, firm size, tangibility and profitability. On the other hand, the impact of the country-specific factors on leverage is not the same across the countries. Macroeconomic factors like inflation, GDP growth and interest have a significant impact on the capital structure choice of firms. The results for the pecking order model show little support for the pecking order model in the developing countries. The hypothesis of a pecking order coefficient of 1 is rejected for each country. The sample is also divided into subsamples of small and large firms based on total assets. The results indicate that there is less support for the pecking order
expectation for small firms.
Keywords: Capital structure; Developing countries; Firm-specific factors;
Table of Contents
I. Introduction 1
II. Literature Review 4
A. Theories of Capital Structure 4
A.1. The Modigliani-Miller Theorem 4
A.2. The Static Tradeoff Theory 5
A.3. The Pecking Order Theory 5
A.4. Empirical Evidence on Static Tradeoff and Pecking Order Theory 6
B. Empirical Evidence on Determinants of Capital Structure 7
B.1. Empirical Evidence in Developed Countries 7
B.2. Empirical Evidence in Developing Countries 8
C. Macroeconomic Conditions 10
III. Research Method 10
A. Hypotheses 11
A.1. Firm-Specific Factors 11
A.2. Country-Specific Factors 12
A.3. The Pecking Order Hypothesis 13
B. Data 14
C. Methodology 16
C.1. Testing the Determinants of Capital Structure 16
C.2. Testing the Pecking Order Model 17
D. Hausman Specification Test 17
IV. Descriptive Statistics 18
A. Summary Statistics for the Determinants of Capital Structure 18
V. Results 25
A. The Determinants of Capital Structure 25
A.1. Firm-Specific Factors 25
A.2. Country-Specific Factors 27
A.3. Cross-Country Analysis 29
B. The Pecking Order Model 31
B.1. Financing Behavior of Developing Countries 32
B.2. Financing Behavior of Small and Large Firms 32
VI. Robustness Checks 37
VII. Conclusion 42
Appendix 45
I. Introduction
Although there has been research for over fifty years on capital structure choice, this is still one of the most controversial issues in corporate finance. It all started with the capital structure irrelevance proposition of Modigliani and Miller in 1958. The majority of the research has focused on the United States and increasingly also on other developed countries with similar institutional structures (Booth et al., 2001). Rajan and Zingales (1995) performed an important cross-country analysis and found that the capital structure choice of the G7-countries is related to the same firm-specific factors that are relevant in the US setting.
On the other hand, there are only a few studies on developing countries that have different institutional structures (e.g. Booth et al., 2001; De Jong et al., 2008). It is important to know whether the factors that seem to determine the capital structure of firms in the US and other developed countries also apply to developing countries (Booth et al., 2001).
This study broadens the research in several respects. First of all, this study will cover a much longer time period. Further, the focus will be on fast growing
developing countries in comparison to the US. This study also contributes to the empirical evidence on capital structure determinants by including the interest rate as a country-specific variable.
The aim is to test whether the capital structure choice of firms in developing countries is related to the same factors that determine the capital structure choice of firms in the US. These factors can be firm-specific and/or country-specific factors. Therefore, the research question is:
What is the impact of firm-specific and country-specific factors on the capital structure choice of firms in the United States and developing countries?
The on going debate on determinants of capital structure shows the relevance of this research area. De Jong et al. (2008) performed an important study and showed that the impact of firm-specific factors on leverage is not the same across countries, however this is often assumed. This indicates that even questions remain on the impact of firm-specific factors across countries.
This investigation is based on the United States and five developing
countries: China, India, Indonesia, Malaysia and Thailand. The sample period is from 1997-2012. Following Rajan and Zingales (1995), the included firm-specific factors are growth, firm size, tangibility and profitability. In order to determine the impact of country-specific factors on the capital structure choice of firms, three macroeconomic factors are included: inflation, GDP growth and interest. In order to test the
determinants of capital an appropriate research method is required. Since the sample consists of data across firms and over time, the panel data method is used. A fixed effects regression is performed to control for omitted variables that vary across firms but do not change over time (Stock and Watson, 2012).
The found relationship between the determinants of capital structure and
leverage is often explained by two important theories on capital structure: the static tradeoff theory and the pecking order theory. According to Myers (1984), the optimal leverage ratio of a firm is determined by the tradeoff between the costs and benefits of borrowing. In the static tradeoff theory, the value of the interest tax shield is weighted against the costs of financial distress (Myers, 1984). Firms move towards a target debt-to-value ratio (Myers, 1984).
An alternative view on capital structure is the pecking order theory, proposed by Myers (1984) and Myers and Majluf (1984). However, in the pecking order theory there is no clear target debt-to-value ratio (Shyam-Sunder and Myers, 1999). Firms are expected to follow a financing hierarchy because of information asymmetries (Myers, 1984). Firms prefer internal financing to external financing and debt is preferred to equity if external financing is necessary (Myers, 1984).
Shyam-Sunder and Myers (1999) tested the pecking order model and found that it is a great first-order descriptor of the financing behavior of public firms in the US from 1971-1989. According to Shyam-Sunder and Myers (1999), in a simple pecking order model firms will only issue debt when the internal cash flows are not sufficient for the firm’s investments and dividend obligations. Equity is only issued as a last resort (Shyam-Sunder and Myers, 1999). Therefore, the net debt issued will be equal to the financial deficit of a firm. Frank and Goyal (2003) also investigated the pecking order model for US firms and point out that it is important to know whether the pecking order model is broadly applicable. Empirical evidence on the pecking order model has been limited to the developed world.
Therefore, besides investigating the determinants of capital structure, this study also investigates the pecking order model. First, this study contributes to the empirical evidence by filling the gap in the empirical research on the pecking order model in developing countries. By doing so, our understanding of the pecking order model increases. Secondly, this study also contributes to the growing literature and empirical evidence on the pecking order model by investigating different firm sizes. Therefore, the research question related to the pecking order model is twofold:
- Does the pecking order model provide an explanation for the financing
behavior in developing countries?
- Does the pecking order model receive more support for small firms?
In order to test the pecking order model of capital structure in each developing country, the paper of Shyam-Sunder and Myers (1999) is followed. They also test the static tradeoff model in the US, however it is out of the scope of this research to also test the static tradeoff model. The simple pecking order model predicts that the amount of external debt financing is equal to the financial deficit of a firm (Shyam-Sunder and Myers, 1999). Therefore, the pecking order hypothesis is that the pecking order coefficient is equal to one and the constant is equal to zero. In order to test the pecking order model for different firm sizes, firms are divided into four subsamples based on total assets. According to the pecking order theory, large firms face less adverse selection problems and therefore issuing equity is easier (Frank and Goyal, 2003). Following Frank and Goyal (2003), it is expected that there is more support for the pecking order expectation for small firms.
The rest of the paper is organized as follows: Section II presents the literature review. Section III describes the research method, containing the hypotheses, data and methodology. Section IV presents descriptive statistics. Section V presents the
empirical results for the determinants of capital structure and the pecking order model, respectively. Section VI presents robustness checks. Conclusions and suggestions for future research are presented in Section VII.
II. Literature Review
This section describes the main theories of capital structure and their predictions. In the next part the empirical evidence on determinants of capital structure in developed and developing countries is discussed. The last part discusses the importance of macroeconomic conditions in determining capital structure.
A. Theories of Capital Structure
Many theories have been developed about capital structure. Capital structure is defined as the amount of debt, equity and other securities that a firm has outstanding (Berk and DeMarzo, 2007). It all started with the well-known paper of Modigliani and Miller in 1958. First, the most important results of the paper of Modigliani and Miller (1958) are discussed. Next, the two main theories on capital structure called the static tradeoff theory and the pecking order theory are discussed. There are many more theories developed in corporate finance about capital structure, however it is
impossible to take into account all the theories for this study. The last part discusses how the static tradeoff theory and pecking order theory fare empirically.
A.1. The Modigliani-Miller Theorem
An important topic in corporate finance is to determine the best capital structure choice of a firm (Berk and DeMarzo, 2007). It all started with two famous
researchers: Franco Modigliani and Merton Miller. They wrote a well-known paper about the cost of capital, corporation finance and the theory of investment in 1958. They stated that in perfect capital markets the total value of a firm does not depend on the choice of capital structure; this is the so-called irrelevance proposition (Modigliani and Miller, 1958). A perfect capital market means that all securities are fairly priced, there are no taxes and the cash flows generated by investments do not depend on the capital structure choice of a firm (Berk and DeMarzo, 2007).
The answer of Modigliani and Miller that the choice of capital structure is irrelevant for the value of firm surprised many researchers (Berk and DeMarzo, 2007). Since the paper of Modigliani and Miller many researchers have tried to find theories that show that the capital structure choice is relevant for the value of a firm. Myers (1984) explained the capital structure choice in a static tradeoff framework and a pecking order framework.
A.2. The Static Tradeoff Theory
According to Myers (1984), the optimal leverage ratio of a firm is determined by the tradeoff between the costs and benefits of borrowing. Firms move towards a target debt-to-value ratio (Myers, 1984). Firms have to pay taxes on their profits. The interest paid to debt holders is tax deductible, this creates an incentive for firms to use debt (Berk and DeMarzo, 2007). This tax benefit is called the interest tax shield (Berk and DeMarzo, 2007).
On the other hand, debt increases the costs of financial distress (Myers, 1984). Financial distress means that firms have difficulties meeting their debt obligations (Berk and DeMarzo, 2007). According to Myers (1984), the costs of financial distress contain bankruptcy, monitoring, moral hazard, agency and contracting costs. Firms that have high leverage levels may not be able to meet their debt obligations (Berk and DeMarzo, 2007). In the static tradeoff theory, the value of the interest tax shield is weighted against the costs of financial distress (Myers, 1984).
A.3. The Pecking Order Theory
An alternative view on capital structure is the pecking order theory, proposed by Myers (1984) and Myers and Majluf (1984). In the pecking order framework firms follow a financing hierarchy because of information asymmetries (Myers, 1984). It is assumed that managers have more information about the firm value than investors (Myers and Majluf, 1984). Firms prefer internal financing to external financing and debt is preferred to equity if external financing is necessary (Myers, 1984).
Myers and Majluf (1984) explain that there are costs associated with issuing equity, transaction costs and cost of asymmetric information, which leads to this hierarchy. They showed that it is better to issue save securities than risky securities (Myers and Majluf, 1984). When managers have private information about the value of a firm, investors will discount the price that they are willing to pay for equity due to adverse selection (Berk and DeMarzo, 2007). Managers who think that outside investors underprice the equity of the firm will follow this hierarchy (Berk and DeMarzo, 2007). Therefore, managers prefer to fund investments by using internal finance like retained earnings, then debt and as a last resort issue equity (Myers, 1984). In contrast with the static tradeoff theory, there is no clear target debt-to-value ratio (Myers, 1984). The costs and benefits of borrowing are assumed second-order (Shyam-Sunder and Myers, 1999).
A.4. Empirical Evidence on Static Tradeoff and Pecking Order Theory
An exciting discussion in the literature is about which theory is better at predicting the capital structure choice of firms (Lemmon and Zender, 2010). Shyam-Sunder and Myers (1999) found that the pecking order model is a great first-order descriptor of the financing behavior of firms in the US. According to Shyam-Sunder and Myers (1999), in a simple pecking order model firms will issue debt when the internal cash flows are not sufficient for the firm’s investments and dividend obligations.
Therefore, the amount of net debt issued by a firm is equal to the financial deficit and equity is only issued as a last resort (Shyam-Sunder and Myers (1999). Their results showed that there is more confidence in the pecking order theory than in the static tradeoff theory (Shyam-Sunder and Myers, 1999).
Frank and Goyal (2003) also tested the pecking order theory for publicly traded firms in the US. However, they found that the pecking order theory in no longer supported when a larger sample or a longer time period is studied. Fama and French (2005) found that only half (46%) of the capital structure decisions of the firms in their sample can be explained by the pecking order theory. They argue that both the static tradeoff theory and the pecking other theory have problems. Both theories have important elements that explain a part of the capital structure choice of firms but they should not be seen as stand-alone theories (Fama and French, 2005).
Lemmon and Zender (2010) modified the model of Shyam-Sunder and Myers (1999) and showed that it is important to control for debt capacity. They found that firms prefer debt to equity when external financing is necessary, however this is only the case when there are no debt capacity concerns (Lemmon and Zender, 2010). Their results suggest that often the concerns of firms over debt capacity explain the use of equity (Lemmon and Zender, 2010). To summaries, the publicly traded firms in their sample follow a pecking order when they control for debt capacity (Lemmon and Zender, 2010).
The previous studies show that researchers sometimes come to different conclusions. According to Leary and Roberts (2010), a common problem of the existing tests on the pecking order theory is the statistical power. They found that there is more support for the pecking order theory when the debt capacities of firms are allowed to vary with factors attributed to other theories (Leary and Roberts, 2010). They stated that this is line with the conjecture of Fama and French (2005). The static tradeoff and the pecking other theory are both important but should not be seen as
stand-alone theories (Fama and French, 2005). Shyam-Sunder and Myers (1999) stated that more empirical research is necessary to test the explanatory power of capital structure models. This research contributes to empirical evidence on capital structure models by filling the gap in the empirical research on the pecking order model in developing countries.
B. Empirical Evidence on Determinants of Capital Structure
Many researchers have tried to find empirical evidence on determinants of capital structure. This section presents the empirical evidence on the determinants of capital structure in developed and developing countries. The last part discusses the
importance of macroeconomic conditions in determining capital structure.
B.1. Empirical Evidence in Developed Countries
The empirical research on determinants of capital structure of firms started in the eighties (Chakraborty, 2010). The majority of the research has focused on the United States and other developed countries with similar institutional structures (Booth et al., 2001). Two important studies in the United States are of Titman and Wessels (1988) and Harris and Raviv (1991). Titman and Wessels (1988) found that leverage
decreases with uniqueness and profitability. They do not find an impact of tangibility, growth, volatility and non-debt tax shields on leverage (Titman and Wessels, 1988). However, Harris and Raviv (1991) did a survey and found different results. They found that profitability, advertising expenditures, uniqueness, R&D expenditures and volatility are negatively related to leverage (Harris and Raviv, 1991). They found that firm size, tangibility, non-debt tax shields and growth are positively related to
leverage (Harris and Raviv, 1991).
Rajan and Zingales (1995) performed an important cross-country analysis. They investigated the capital structure choice of the G7-countries (United States, United Kingdom, Canada, Japan, Germany, France and Italy) from 1987-1991. Their objective was to see whether the capital structure choice of the G7-countries is related to the same factors that are relevant in the US setting. They found that the
G7-countries are leveraged in a similar way, only the United Kingdom and Germany are less leveraged. Rajan and Zingales (1995) point out that further research must focus on a better understanding of the determinants of capital structure and investigate institutional differences. They were one of the first researchers that highlighted the
importance of country-specific factors in determining capital structure (Alves and Ferreira, 2011).
Frank and Goyal (2009) investigated public firms in the United States from 1950 to 2003, to see which factors are important for the capital structure choice of firms. They stated that even after so much research on capital structure decisions the important factors remain elusive (Frank and Goyal, 2009). They identified six core factors for market leverage: firm size, growth, profitability, tangibility, inflation and industry median (Frank and Goyal, 2009). They found evidence supporting both the static tradeoff theory and the pecking order theory. Their results show that the impact of profitability is explained by the pecking order theory (Frank and Goyal, 2009). The static tradeoff theory provides an explanation for the impact of firm size, tangibility and growth. They stated that their evidence shows weaknesses in each theory (Frank and Goyal, 2009). This suits with the view of Fama and French (2005) that both theories explain a part of the capital structure choice but should not be seen as stand-alone theories.
B.2. Empirical Evidence in Developing Countries
There are only a few studies on developing countries that have different institutional structures (Booth et al., 2001). Research is necessary to understand the impact of different institutional structures on capital structure decisions (Booth et al., 2001). Four studies are here discussed: Booth et al. (2001), Chen (2004), De Jong et al. (2008) and Alves and Ferreira (2011).
Booth et al. (2001) studied 10 developing countries (Pakistan, Thailand, Turkey, Malaysia, India, Zimbabwe, Jordan, Brazil, Mexico and Korea) from 1980 to 1991 and found that they are affected by the same firm-specific factors as in
developed countries. However, country-specific factors matter for developing
countries. It seems that factors like GDP growth, inflation and development of capital markets affect the debt ratios differently (Booth et al., 2001). They point out that a lot of empirical research has to be done and international databases have to be
investigated.
Chen (2004) studied the determinants of capital structure of 88 listed firms in China from 1995 to 2000. The results show that the firm-specific factors that are important in developed countries are also relevant for China (Chen, 2004). Leverage increases with firm size, growth opportunities and tangibility. The results showed that
profitability is negatively related to leverage. However, both the static tradeoff theory and the pecking order theory do not give a convincing explanation for the capital structure choice of firms in China (Chen, 2004). Chen (2004) found that firms first use retained earnings, then equity and as a last resort issue long-term debt. According to Chen (2004), the results show a “new Pecking Order”. The different institutional and macroeconomic environment in China may explain the different capital structure choice (Chen, 2004). However, Chen (2004) did not study this for China and points out that future research should take into account macroeconomic conditions when studying capital structure choices.
De Jong et al. (2008) studied the capital structure choice of 21 developing countries (as well as 21 developed countries) from 1997 to 2001. Their main objective was to study the importance of firm-specific and country-specific factors for capital structure choice (De Jong et al., 2008). They found that the impact of firm-specific factors on leverage is not the same across countries, however this is often assumed in studies (De Jong et al., 2008). They also found that country-specific factors have a direct and indirect impact on capital structure. Their results show that creditor right protection, GDP growth and development of bond markets have a significant impact on leverage (De Jong et al., 2008). Their conclusion is that country-specific factors are very important for determining the capital structure of firms worldwide (De Jong et al., 2008).
Alves and Ferreira (2011) investigated legal systems, the development of capital markets and investor protection of 31 countries from 1996 to 2001. Their results show that shareholder rights are an important determinant of leverage (Alves and Ferreira, 2011). They point out that most studies focus on firm-specific factors, however empirical evidence suggests that the institutional environment of countries also determines capital structure (Alves and Ferreira, 2011). They found evidence that the impact of both firm-specific and country-specific factors is not always equal between countries (Alves and Ferreira, 2011). However, it is often assumed that firm-specific factors have the same impact on leverage across countries (De Jong et al., 2008). Profitability seems to be the firm-specific factor that is often the same across countries in the study of Alves and Ferreira (2011).
C. Macroeconomic Conditions
The previous literature review showed the importance of country-specific factors for determining the capital structure choice of firms. Every country has a unique setting with its own regulations, legal system, banking system, language, culture and
macroeconomic conditions. The different leverage levels between countries might be explained by differences in these factors.
Booth et al. (2001) stated that it seems that factors like GDP growth, inflation and development of capital markets affect the leverage ratios differently. Alves and Ferreira (2011) stated that shareholder rights are an important factor to take into account for determining leverage. However, studies on capital structure determinants have not focused often on the impact of macroeconomic conditions. The
macroeconomic environment of a country consists of aggregate economic quantities, like the GDP growth, unemployment, inflation and interest rates (Pindyck and Rubinfeld, 2009). In particular two countries in my study are experiencing enormous growth: China and India. It is expected that the combined GDP of China and India will exceed that of the major seven OECD economies in 2025 (Johansson et al., 2012). This study wants to get further insight in the impact of macroeconomic conditions of countries on the capital structure choice of firms. The next section presents the research method.
III. Research Method
Considering the literature discussed in the previous section, I develop my hypotheses for this research. The hypotheses are related to four firm-specific and three country-specific factors, to see the influence of these factors on the capital structure choice of firms. Rajan and Zingales (1995) and Frank and Goyal (2009) found that growth, firm size, tangibility and profitability are the most important firm-specific factors.
Therefore, I will include these firm-specific factors in my study. In order to determine the impact of country-specific factors on the capital structure choice of firms, three macroeconomic factors are included: inflation, GDP growth and interest. This study also tests the pecking order model in developing countries. Therefore, the pecking order hypothesis is presented. Next the data and the methodology are presented. The last part discusses the Hausman specification test.
A. Hypotheses
A.1. Firm-Specific Factors
Growth
The empirical evidence, presented in the literature review, showed that growth
opportunities are an important determinant of capital structure. According to the static tradeoff theory, growth firms have lower leverage ratios (Frank and Goyal, 2009). Growth firms are expected to decrease more in value when they have difficulties meeting their debt obligations (Frank and Goyal, 2009). On the other hand, the pecking order theory predicts that firms that have many growth opportunities should have more leverage (Frank and Goyal, 2009). However, most studies found empirical evidence that supports the static tradeoff theory. The following hypothesis is derived:
H1 - There is negative relationship between growth and leverage
Firm size
According to the static tradeoff theory, large firms are more diversified and therefore have a lower probability of bankruptcy (Frank and Goyal, 2009). The static tradeoff theory predicts that large firms can increase their leverage and lower their costs of capital (Berk and DeMarzo, 2007). However, the pecking order theory predicts a negative relationship between firm size and leverage (Frank and Goyal, 2009). Large firms face less adverse selection problems and therefore issuing equity is easier (Frank and Goyal, 2003). Consistent with Rajan and Zingales (1995) and Frank and Goyal (2009), the following hypothesis is derived:
H2 - There is positive relationship between firm size and leverage
Tangibility
Tangible assets are long-term assets that have a physical substance like land, plant, buildings and equipment (Needles et al., 2008). Tangible assets can be used as
collateral and this gives investors security because in case of default debt holders have rights to the assets of the firm (Berk and DeMarzo, 2007). According to the static tradeoff theory, tangible assets have an important impact on the costs of financial distress (Frank and Goyal 2009). Tangible assets can be liquidated for close to their full value and therefore the costs of financial distress tend to be low (Berk and DeMarzo, 2007). Therefore, firms with tangible assets are expected to have higher leverage. According to the pecking order theory, tangible assets have less asymmetric
information problems and therefore it is less costly to issue equity for firms (Frank and Goyal, 2009). Firms with tangible assets are expected to have lower leverage. Based on most empirical evidence the following hypothesis is derived:
H3 - There is positive relationship between tangibility and leverage
Profitability
Profitable firms are expected to have lower costs of bankruptcy (Frank and Goyal, 2009). If a firm decides to use debt, the interest tax shield provides a tax benefit for the firm. The interest tax shield is expected to be more valuable for profitable firms (Frank and Goyal, 2009). According to the static tradeoff theory, profitable firms are expected to have higher debt levels (Frank and Goyal, 2009). On the other hand, the pecking order theory predicts a different relationship. Firms are expected to first finance their investments with retained earnings, then debt and as a last resort equity (Myers, 1984). Therefore, profitable firms will have higher retained earnings and are less leveraged (Frank and Goyal, 2009). Consistent with Rajan and Zingales (1995), Booth et al. (2001) and De Jong et al. (2008), the following hypothesis is derived:
H4 - There is a negative relationship between profitability and leverage
Based on the literature review, it is expected that the relationship between each firm-specific factor and leverage is the same for firms in the United States and the
developing countries. However, De Jong et al. (2008) and Alves and Ferreira (2011) showed that the assumption that firm-specific factors have the same impact on leverage across countries is not correct. It is interesting to know whether this will be detected in this study as well.
A.2. Country-Specific Factors
Inflation
A country is experiencing inflation when its price level is increasing (Krugman and Obstfeld, 2009). When a country has high inflation this may increase the riskiness of leverage (Demirgüç-Kunt and Maksimovic, 1996). Therefore, it is expected that when a country has high inflation firms will have lower leverage levels. Inflation has not been analyzed often as a determinant of capital structure. Booth et al. (2001) found that higher inflation leads to lower debt levels. The following hypothesis is derived:
GDP growth rate
The Gross Domestic Product (GDP) shows the value of all goods and services produced in a country (Krugman and Obstfeld, 2009). When an economy has a high growth rate this is associated with healthy firms and growth opportunities (Bartholdy and Mateus, 2008). Firms are willing to use more debt to finance new investments in countries that have high economic growth (De Jong et al., 2008).
The following hypothesis is derived:
H6 - There is positive relationship between GDP growth rate and leverage
Interest rate
The interest rate is the rate at which a firm can borrow money (Pindyck and Rubinfel, 2009). When interest rates are high borrowing is expensive for firms. Therefore, it is expected that when a country has a high interest rate firms will have lower leverage levels. Bartholdy and Mateus (2008) also stated that high interest rates are associated to increase the cost of borrowing. The following hypothesis is derived:
H7 - There is a negative relationship between interest rate and leverage
It is expected that macroeconomic factors have an impact on the capital structure choice of firms. It is expected that country-specific factors are important to take into account to explain possible differences in leverage levels between developed and developing countries.
A.3. The Pecking Order Hypothesis
Besides investigating determinants of capital structure, this study also tests the
pecking order model in developing countries. According to Shyam-Sunder and Myers (1999), firms only issue debt when the internal cash flows are not sufficient for the firm’s investments and dividend obligations. Equity is only issued as a last resort (Shyam-Sunder and Myers, 1999). Therefore, the net debt issued will be equal to the financial deficit of a firm. It is assumed that all the components of the financial deficit are exogenous (Shyam-Sunder and Myers, 1999). The methodology section will go further into the estimation of the financial deficit.
Following Shyam-Sunder and Myers (1999), the pecking order hypothesis is that the constant (!) is equal to zero and the pecking order coefficient (!!") is equal to
one. This indicates that when a firm does not face a financial deficit they will not issue debt. On the other hand, if a firm does face a financial deficit the amount of debt issued is equal to the financial deficit. Firms have no incentive to issue equity
(Shyam-Sunder and Myers. 1999). The simple pecking order model predicts that external debt financing is based on the firm’s financial deficit (Shyam-Sunder and Myers, 1999).
This study also investigates the pecking order model for different firm sizes. Following Frank and Goyal (2003), firms are divided based on total assets. In the pecking order theory firms follow a financing hierarchy because of information asymmetries (Myers, 1984). According to the pecking order theory, large firms face less adverse selection problems and therefore issuing equity is easier (Frank and Goyal, 2003). Therefore, large firms are expected to have lower leverage. On the other hand, small firms are expected to have more adverse selection problems (Frank and Goyal, 2003). Following Frank and Goyal (2003), it is expected that there is more support for the pecking order expectation for small firms.
In the next section the data and methodology are presented. The results will either confirm or reject the hypotheses. The hypotheses are tested at a significance level of 1%, 5% and 10%.
B. Data
This study uses annual financial data of non-financial firms listed in the United States and five developing countries: China, India, Indonesia, Malaysia and Thailand.1 The sample period is from 1997-2012. A more recent time period is not chosen because the country-specific factors were not available in World Development Indicators. I will make subsets; first the regression is done for the United States and then for each developing country. This study takes into account firm-specific and country-specific factors to analyze the relationship with leverage.
The firm-specific factors are collected from Compustat North America and Compustat Global (through WRDS). In these databases you can select all the firms that report in a Financial Services and/or Industrial format. This study only uses
1 According to the International Monetary Fund (2011), these countries are considered developing countries. Initially, the Philippines were also included, however because of a relatively small sample size it was excluded from the study.
financial firms (Industrial format) because the capital structure of financial firms is very different. Therefore, financial firms with a SIC code between 6000 and 6999 are excluded and also utilities with a SIC code between 4900 and 4999. Firms are also excluded when they have missing book value of assets, missing years and book leverage above one.2 The sample also excludes firm-observations with assets (sales) below one, because the log of assets (sales) then gives a negative firm size. Leverage and the other firm-specific variables, except firm size, are winsorized in both tails of the distribution at 1.00%. Winsorizing means that outliers are replaced with the value of the firm-specific variables at the cutoffs (Frank and Goyal, 2009).
In Compustat Global are the country codes for the developing countries selected and then the variables to create the firm-specific factors. The selected country codes are reported in Table A1 in the Appendix. The variable definitions and
Compustat annual data items are reported in Table A2 in the Appendix. The country-specific factors are obtained from World Development Indicators and are reported in Table A3 in the Appendix. In order to perform the regressions to test the determinants of capital structure, the data from Compustat and Word Development Indicators are merged in Stata.
The data to test the pecking order model are also obtained from Compustat (Industrial format) for each developing country from 1997-2012. Again financial firms and utilities are excluded from the sample. Shyam-Sunder and Myers (1999) stated that continuous data is not required in order to test the pecking order model.3 Therefore, firms are only excluded when they have missing book value of assets, long-term debt above one and missing years. The variables are scaled by the book value of net assets and winsorized in both tails of the distribution at 1.00%. The variable definitions and Compustat annual data items for the pecking order model are also reported in Table A2 in the Appendix.
2 Following Baker and Wurgler (2002), the number of excluded firms depends on the country sample being analyzed. Market leverage is also not above one.
3 It was not possible to test the pecking order model with continuous data on all the balance sheet and cash flow variables because only for a few firms in each developing country continuous data was available.
C. Methodology
C.1. Testing the Determinants of Capital Structure
In order to determine the impact of the firm-specific and country-specific factors on leverage, an appropriate research method is required. Since the sample consists of data across firms and over time, the panel data method is used. Following Booth et al. (2001), this study uses the fixed effects model to control for omitted variables that vary across firms but do not change over time.The empirical model is as follows:
!!" = !! + !’!!"+ !’!!+ !!" i = 1,…, N and t = 1,…, 16, (1)
where the first subscript, i, refers to the firm being observed and the second subscript,
t, refers to the year at which the firm is observed. The dependent variable (!!") is
leverage (!"#!"), !! is the intercept coefficient of firm i, β represents the coefficients, the firm-specific variables (!!") are growth, size, tangibility and profitability for firm i
at year t, the country-specific variables (!!) are inflation, GDP growth rate and interest rate at year t and !!"!represents the residual error term of firm i at year t. This leads to the following regression equation:
!"#!" = !! + !!!"#$%&!"+ !!!"#$!" + !!!"#$!"+ !!!"#$!"!
+!!!!"#$%&!'"!+ !!!"#!+ !!!"#$%$&#!+ !!". (2)
The variable definitions and Compustat annual data items are reported in Table A2 in the Appendix. The random effects model is also used. The Hausman specification test is performed to test the fixed effects model against the random effects model. After the regression is performed for each country, the results will show a coefficient for each variable. The found coefficients of the firm-specific and country-specific variables are used to test the hypotheses. The results will either confirm or reject the hypotheses. The hypotheses are tested at significance level of 1%, 5% and 10%. It is expected that the firm-specific factors have the following impact on leverage: growth (-), firm size (+), tangibility (+) and profitability (-). It is expected that the country-specific factors have the following impact on leverage: inflation (-), GDP (+) and interest (-).
C.2. Testing the Pecking Order Model
This study also tests the pecking order model of capital structure in developing countries. Following Shyam-Sunder and Myers (1999), the funds flow deficit is estimated using the following equation:
!"#!= !"#!+ !!+ ∆!!+ !!− !!, (3)
where !"#! is the funds flow deficit, !"#! are the dividend payments, !! are the
capital expenditures, ∆!! is the net increase in working capital, !!!is the current portion of long-term debt due in year t and!!! are the operating cash flows after interest and taxes. The variables are estimated at the end of period t.
Following Shyam-Sunder and Myers (1999), the pecking order hypothesis is tested using the following regression equation:
∆!!" = ! + !!"!"#!"+ !!", (4)
where the dependent variable (∆!!") is the change in the debt-to-asset ratio.4 !
!"
represents the pecking order coefficient. The independent variable (!"#!") is the funds flow deficit estimated with equation (3). !!" represents the error term. It is
expected that!! is equal to zero and !!" is equal to one.
D. Hausman Specification Test
The Hausman specification test is performed to test the fixed effects model against the random effects model for testing the determinants of capital structure. The random effects model assumes that the firm-specific intercept is a random variable (Booth et al., 2001). The fixed effects model assumes that the firm-specific intercept is
correlated with the other independent variables in the model (Hausman, 1978). When the null hypothesis is rejected the random effects estimator is biased (Hausman, 1978).
4 The net amount of debt issued could not be estimated as a dependent variable for the developing countries because long-term debt issuance (Item 111) and long-term debt reduction (Item 114) were not available in Compustat.
The results of the Hausman specification test are presented in Table A4 in the Appendix. The results indicate that in almost all the countries the null hypothesis is rejected.The results are significant at a 5% or 1% level. Therefore, I will use the fixed effects model to test the determinants of capital structure. The next section presents the descriptive statistics.
IV. Descriptive Statistics
This section provides summary statistics. Tables I through III provide summary statistics for leverage, the firm-specific and the country-specific variables,
respectively. A comparison is made between the United States and five developing countries: China, India, Indonesia, Malaysia and Thailand. The summary statistics include the mean, median, standard deviation, minimum, maximum and number of observations. Table IV provides summary statistics for the pecking order model across the developing countries.
A. Summary Statistics for the Determinants of Capital Structure
The average leverage ratio, reported in Table I, varies between 19.4 percent in the United States and 31.8 percent in Indonesia. All the developing countries show an average leverage ratio above the United States. Demirgüç-Kunt and Maksimovic (1996) stated that in countries where the stock market is developed, such as in the United States, firms substitute equity for leverage. On the other hand, in developing countries where the stock market has not been fully developed firms rely more on debt financing (Demirgüç-Kunt and Maksimovic, 1996). This could be an explanation why the developing countries, reported in Table I, show higher leverage ratios.
Table I also shows that the median leverage is below the mean leverage. All the countries show a large cross-sectional difference. For example, the minimum book leverage ratio for the United States is 0 percent and the maximum book leverage ratio is 83.4 percent. This finding is consistent with Frank and Goyal (2009), who also found a large cross-sectional difference for US firms in their sample. For the United States the market value of leverage is also computed, reported in the last row of Table I. Due to data limitations for the developing countries (the stock price was missing in Compustat) only the book value of leverage was computed. The United States shows an average book leverage of 19.4 percent, reported in Table I, and an average market leverage of 19.6 percent. This finding is consistent with Frank and Goyal (2009), who
also found a small difference between the average book leverage and the average market leverage for US firms.
Table I also provides information on the number of firms for each country. The United States has the largest sample with 1328 firms. Countries with also a relatively large sample of firms are China, India, Malaysia and Thailand. Indonesia has a relatively small sample of firms.
Table I
Summary Statistics of Leverage
This table presents summary statistics of leverage. The data are obtained from Compustat North America and Compustat Global for each country from 1997-2012. The dependent variable Leverage is defined in two ways, in book value and in market value. Book Leverage is defined as the book value of debt divided by total assets.
Market Leverage is defined as the book value of debt divided by the market value of
assets, which is estimated by the book value of assets minus the book value of equity plus the market value of equity. Market Leverage could only be estimated for the United States and is presented between parentheses. For leverage, the mean, median, standard deviation, minimum and maximum are presented. N is the number of firms for each country. Leverage is expressed in percentage terms. The Compustat annual data items are reported in Table A2 in the Appendix.
Leverage (%)
Country N Mean Median S.D. Min Max
China 800 23.8 23.1 15.9 0.0 63.6 India 436 29.3 28.7 19.8 0.0 77.6 Indonesia 74 31.8 30.6 23.3 0.0 91.7 Malaysia 251 21.8 20.3 17.6 0.0 73.3 Thailand 139 27.2 23.8 23.7 0.0 86.3 United States 1328 19.4 (19.6) 16.4 (13.0) 18.1 (21.3) 0.0 (0.0) 83.4 (89.9)
Table II provides summary statistics for the independent firm-specific variables across the United States and the five developing countries. This study includes four firm-specific variables to analyze the impact on the capital structure choice of firms. The firm-specific variables are growth, firm size, tangibility and profitability.
The average growth, reported in Table II, is around 5.5 percent across the countries. Malaysia shows the lowest average growth of 4.1 percent. The majority of the developing countries show an average growth above the United States. The average growth for the United States is 5.5 percent. The median growth is below the mean growth for every country. There is also quite some variation in growth rates, as can be seen from the standard deviation, reported in Table II, which is around 6 percent for each country.
The average firm size, reported in Table II, varies between 5911.4 million in Malaysia and 14165.9 million in Indonesia. Large firms are often more diversified and therefore have a lower probability of bankruptcy (Frank and Goyal, 2009). Large firms are expected to have higher leverage. The median firm size is below the mean firm size for every country. There is also quite some variation in the average firm size, as can be seen from the standard deviation, reported in Table II, which is around 1500 million for each country.
The average tangibility, reported in Table II, is around 40 percent across the developing countries. All the developing countries show an average tangibility above the United States, which shows a tangibility of 28.4 percent. The rationale behind tangibility is that tangible assets are easier to collateralize (Rajan and Zingales, 1995). Demirgüç-Kunt and Maksimovic (1996) stated that in developing countries where the stock market has not been fully developed firms rely more on debt financing. The United States has a developed stock market and therefore the need to have many tangible assets might be lower because firms are less dependent on debt financing. There is also quite some variation in tangibility, as can be seen from the standard deviation, reported in Table II, which is around 20 percent for each country.
The average profitability, reported in Table II, varies between 6.6 percent in China and 14.6 percent in Indonesia. The majority of the developing countries show an average profitability above the United States. Countries with the highest average profitability are India, Indonesia and Thailand. Profitable firms will have higher retained earnings and are less leveraged (Frank and Goyal, 2009). All the countries show a negative minimum value of growth. This finding is consistent with earlier studies of Chen (2004) and Frank and Goyal (2009).
Tab le I I S u mmar y S tati sti cs of I n d ep en d en t F ir m -S p ec ifi c V ar iab le s T hi s t abl e pre se nt s s um m ary s ta ti st ic s of the inde pe nde nt f irm -s pe ci fi c va ri abl es . T he da ta a re obt ai ne d from Com pus ta t N ort h A m eri ca a nd Com pus ta t G loba l f or e ac h c ount ry f rom 1997 -2012 . G rowt h is de fi ne d a s c api ta l e xpe ndi ture s di vi de d by t ot al a ss et s. F ir m s iz e is de fi ne d a log of tot al a ss et s. T angi bi li ty is de fi ne d a s ne t prope rt y, pl ant a nd e qui pm ent di vi de d by t ot al a ss et s. P rof it abi li ty is de fi ne d a s ope ra ti ng i nc be fore de pre ci at ion di vi de d by t ot al a ss et s. F or e ac h f irm -s pe ci fi c va ri abl e, t he m ea n, m edi an, s ta nda rd de vi at ion, m ini m um a nd m axi m um a pre se nt ed. N is the num be r of f irm s f or e ac h c ount ry. T he f irm -s pe ci fi c va ri abl es , e xc ept f irm s iz e, a re e xpre ss ed i n pe rc ent age te rm s. F irm s is e xpre ss ed i n m il li ons . T he Com pus ta t a nnua l da ta it em s a re re port ed i n T abl e A 2 in t he A ppe ndi x. Chi na Indi a Indone si a M al ays ia T ha il and U ni te d S ta te s G row th Mean 5.6 6.1 5.9 4.1 5.4 5.5 M edi an 3.8 3.4 4.4 2.4 3.8 3.6 S .D . 5.8 7.4 5.4 4.8 5.3 6.0 Min 0.0 0.0 0.0 0.0 0.0 0.0 Max 30.4 43.7 36.1 28.5 30.4 37.9 F irm s iz e Mean 7468.6 6262.6 14165.9 5911.4 8181.0 6022.3 M edi an 7351.8 6098.3 14090.2 5693.3 7919.9 5981.8 S .D . 1173.7 1496.2 1549.4 1501.2 1495.5 2415.3 Min 292.7 430.5 10293.8 802.0 3943.9 83.4 Max 12671.8 13586.9 19021.0 11546.5 12888.1 13589.6 T angi bi li ty Mean 34.7 35.1 38.2 35.5 42.2 28.4 M edi an 32.0 34.2 33.6 34.2 40.9 21.8 S .D . 19.3 19.5 21.2 20.3 21.7 23.1 Min 1.0 0. 0 0.3 0.3 0.7 0.0 Max 82.5 84.7 90.4 86.5 89.9 91.4 P rof it abi li ty Mean 6.6 9.8 14.6 7.6 11.4 8.6 M edi an 6.5 9.7 13.1 7.4 11.4 12.0 S .D . 6.8 8.5 11.8 8.5 9.8 20. 9 Min -16.1 -14.1 -25.1 -32.3 -24.0 -181.9 Max 33.5 38.9 52.6 35.9 43.9 41.5 N 800 436 74 251 139 1328
Table III provides summary statistics for the country-specific variables across the United States and the five developing countries. This study includes three country-specific variables to see the impact of macroeconomic determinants on the capital structure choice of firms. The macroeconomic variables are the inflation rate, the GDP growth rate and the interest rate.
The average inflation rate, reported in Table III, varies between 1.9 percent in China and 11.5 percent in Indonesia. All the developing countries show an average inflation rate equal or above the United States, except China. The average inflation rate for the United States is 2.4 percent. High inflation in countries may increase the riskiness of leverage (Demirgüç-Kunt and Maksimovic, 1996). Therefore, firms in countries with high inflation rates are expected to have lower leverage levels. Table III also shows that some countries report an average inflation rate that differs from the median inflation rate. The inflation rate shows the lowest variation in the United States, reported in Table III, a standard deviation of 1.0 percent.
Table III also provides information on the GDP growth rates across the countries. The average GDP growth rate varies between 2.4 percent in the United States and 9.7 percent in China. All the developing countries show an average GDP growth above the United States. China experienced the highest GDP growth over the sample period, as reported in Table III. The GDP growth rate is always positive for China and India during the sample period, while the other countries also experienced negative GDP growth. The median GDP growth rate is above the average GDP growth rate for every country, except China. The GDP growth rate shows the lowest variation in China, reported in Table III, a standard deviation of 1.7 percent.
The average interest rate, reported in Table III, varies between 5.9 percent in the United States and 17.4 percent in Indonesia. All the developing countries show an average interest rate above the United States. The interest rate is the rate at which a firm can borrow money (Pindyck and Rubinfel, 2009). When interest rates are high borrowing is expensive for firms. On the other hand, China and the United States show a relatively low interest rate. The summary statistics for the interest rate indicate that borrowing is more expensive in the developing countries than in the United States. All the countries show a median interest below the mean interest, except India. The interest rate shows the lowest variation in China, reported in Table III, a standard deviation of 0.9 percent.
Tab le I II S u mmar y S tati sti cs of C ou n tr y-S p ec ifi c V ar iab le s T hi s t abl e pre se nt s s um m ary s ta ti st ic s of the c ount ry -s pe ci fi c va ri abl es . T he da ta a re obt ai ne d f rom W orl d D eve lopm ent Indi ca tors f or e ac h count ry from 1997 -2012 . Inf lat ion is de fi ne d a s t he a nnua l c ha nge in t he c ons um er pri ce inde x. GDP is de fi ne d a s t he a nnua l pe rc ent age grow ra te of G D P . Int er es t i s de fi ne d a s t he le ndi ng i nt ere st ra te . F or e ac h c ount ry -s pe ci fi c va ri abl e, t he m ea n, m edi an, s ta nda rd de vi at ion, m ini m and m axi m um a re pre se nt ed. O bs . i s t he num be r of obs erva ti ons f or e ac h c ount ry . T he c ount ry -s pe ci fi c va ri abl es a re e xpre ss ed i n pe rc ent age te rm s. T he obt ai ne d count ry -s pe ci fi c va ri abl es a re re port ed i n T abl e A 3 in t he A ppe ndi x. Chi na Indi a Indone si a M al ays ia T ha il and U ni te d S ta te s Inf la ti on Mean 1.9 6.9 11.5 2.4 3.0 2.4 M edi an 1.7 6.3 6.5 1.9 2.9 2.5 S .D . 2.2 3.1 12.8 1.4 2.2 1.0 Min -1.4 3.7 3.7 0.6 -0.8 -0.4 Max 5.9 13.2 58.4 5.4 8.0 3.8 O bs . 16 16 16 16 16 16 GDP Mean 9.7 6.9 3.9 4.5 3.0 2.4 M edi an 9.3 7.3 5.0 5.6 4.7 2.8 S .D . 1.7 2.3 4.6 3.9 4.5 1.9 Min 7.6 3.8 -13.1 -7.4 -10.5 -2.8 Max 14.2 10.3 6.5 8.9 7.8 4.8 O bs . 16 16 16 16 16 16 Int ere st Mean 6.1 11.8 17.4 6.9 7.7 5.9 M edi an 5.9 12.0 15.3 6.4 7.1 5.7 S .D . 0.9 1.4 5.5 2.0 2.5 2.1 Min 5.3 8.3 11.8 4.8 5.5 3.3 Max 8.6 13.8 32.2 12.1 14.4 9.2 O bs . 16 16 16 16 16 16
B. Summary Statistics for the Pecking Order Model
Table IV provides summary statistics for the pecking order model across the five developing countries. Following Shyam-Sunder and Myers (1999), the book debt ratio and the return on assets are reported.
The average book debt ratio, reported in Table IV, varies between 10.1 percent in China and 23.8 percent in India. The median book debt ratio is below the average book debt ratio for every country. All the countries show a large cross-sectional difference. For example, the minimum book debt ratio for Indonesia is 0 percent and the maximum book debt ratio is 84.7 percent. There is also quite some variation in the book debt ratios, as can be seen from the standard deviation, reported in Table IV, which is around 20 percent for each country.
Table IV
Summary Statistics of the Pecking Order Model
This table presents summary statistics of the book debt ratio and the return on assets. The data are obtained from Compustat Global for each country from 1997-2012. Book
debt ratio is defined as long-term debt divided by the book value of net assets. Return on assets is defined as after-tax operating earnings divided by the book value of net
assets. These definitions follow Shyam-Sunder and Myers (1999). The mean, median, standard deviation, minimum and maximum are presented. N is the number of firms for each country. The values are expressed in percentage terms. The Compustat annual data items are reported in Table A2 in the Appendix.
Book debt ratio (%)
Country N Mean Median S.D. Min Max
China 742 10.1 3.6 13.7 0.0 57.1 India 420 23.8 18.8 21.6 0.0 78.2 Indonesia 64 22.4 17.1 23.4 0.0 84.7 Malaysia 241 12.5 6.6 15.3 0.0 69.3 Thailand 119 15.0 4.7 20.3 0.0 82.1 Return on assets (%) China 742 9.2 9.7 11.6 -56.0 36.3 India 420 10.2 9.8 11.9 -29.4 46.7 Indonesia 64 15.7 14.3 19.1 -65.5 84.3 Malaysia 241 8.5 8.7 12.9 -46.9 53.0 Thailand 119 14.8 14.9 14.3 -45.4 56.6
The average return on assets, reported in Table IV, varies between 8.5 percent in Malaysia and 15.7 percent in Indonesia. Table IV also shows that the average return on assets is often above the median return on assets. All the countries show a negative minimum value of return on assets. This finding is consistent with Shyam-Sunder and Myers (1999), who investigated US firms. There is also quite some variation in the return on assets, as can be seen from the standard deviation, reported in Table IV, which is around 15 percent for each country.
Table IV also provides information on the number of firms for each country. Countries with a relatively large sample of firms are China, India, Malaysia and Thailand. Indonesia has a relatively small sample of firms. The next section presents the results.
V. Results
This section presents the results on the determinants of capital structure and the pecking order model, respectively. The regression results for leverage on the
determinants of capital structure are reported in Table V. The results for the pecking order model are reported in Table VI and VII. The pecking order model is also tested for different firm sizes based on total assets.
A. The Determinants of Capital Structure
Following Rajan and Zingales (1995), this section describes broad patterns found across the countries for each firm-specific variable. This study also takes into account three country-specific factors to see the impact of macroeconomic conditions on the capital structure choice of non-financial firms. Each firm-specific and country-specific factor is analyzed. The next part presents a cross-country analysis, where the differences and similarities between the United States and the developing countries are discussed. Table V presents the regression results.
A.1. Firm-Specific Factors
Growth
Table V shows that growth has the expected negative sign and is highly significant for each country. The only exception is India that shows an insignificant negative result. The results indicate that growth is negatively related to leverage for all the developing countries and the United States. The results, in Table V, indicate that growth firm
have lower leverage ratios. This result is consistent with Booth et al. (2001), De Jong et al. (2008) and Alves and Ferreira (2011).
For the majority of the countries, the results for growth, reported in Table V, are consistent with the static tradeoff theory proposed by Myers (1984).Firms with growth opportunities are expected to decrease more in value when they have difficulties meeting their debt obligations (Frank and Goyal, 2009). The results for growth, reported in Table V, are consistent with the hypothesis (H1) that there is a negative relationship between growth and leverage for each country. Only for India the result is insignificant.
Firm size
Table V shows that firm size has the expected positive sign and is significant at the 1 percent level for each country. The results indicate that firm size is positively related to leverage for all the developing countries and the United States. This finding is consistent with De Jong et al. (2008) and Alves and Ferreira (2011). The results indicate that large firms have higher leverage ratios.
The results for firm size are consistent with the static tradeoff theory proposed by Myers (1984).The static tradeoff theory predicts that large firms are more
diversified and therefore have a lower probability of bankruptcy (Frank and Goyal, 2009). The rationale behind firm size is that large firms can increase their leverage and lower their cost of capital (Berk and DeMarzo, 2007). The results for firm size, reported in Table V, are consistent with the hypothesis (H2) that there is a positive relationship between firm size and leverage.
Tangibility
Table V shows that tangibility has the expected positive sign and is significant at the 1 percent level for each country. The results indicate that tangibility is positively related to leverage for all the developing countries and the United States. The results indicate that firms with more tangible assets have higher leverage ratios.
The results for tangibility are consistent with the static tradeoff theory proposed by Myers (1984). The rationale behind tangibility is that tangible assets, such as buildings and equipment, are easier to collateralize (Rajan and Zingales, 1995). This gives investors security because in case of default debt holdershave rights to the assets of the firm (Berk and DeMarzo, 2007). Tangible assets can be liquidated
for close to their full value and therefore the costs of financial distress of leverage tend to be low (Berk and DeMarzo, 2007). Therefore, firms with tangible assets have higher leverage ratios. The results for tangibility, reported in Table V, are consistent with the hypothesis (H3) that there is a positive relationship between tangibility and leverage.
Profitability
Table V shows that profitability has the expected negative sign and is significant at the 1 percent level for each country. This finding is consistent with Booth et al. (2001). The results indicate that profitability is negatively related to leverage for all the developing countries and the United States. The results indicate that profitable firms have lower leverage ratios.
The results for profitability are consistent with the pecking order theory proposed by Myers (1984) and Myers and Majluf (1984). The pecking order theory assumes that firms first finance their investments with retained earnings, then debt and as a last resort equity (Myers, 1984). Therefore, profitable firms will have higher retained earnings and are less leveraged (Frank and Goyal, 2009). The results for profitability, reported in Table V, are consistent with the hypothesis (H4) that there is a negative relationship between profitability and leverage.
A.2. Country-Specific Factors
Inflation
Table V shows that inflation has the expected negative sign and is highly significant for each country. The results indicate that inflation is negatively related to book leverage for all the developing countries and the United States. High inflation in countries may increase the riskiness of leverage (Demirgüç-Kunt and Maskimovic, 1996). Therefore, it is expected that when a country has a high inflation firms will have lower leverage ratios.
On the other hand, for the United States a positive relationship is found between inflation and market leverage, reported in the last column of Table V. Frank and Goyal (2009) also found a positive impact of inflation on market leverage for US firms. A possible explanation for this finding is that the interest paid to debt holders is tax deductible and this creates an incentive for firms to use debt (Berk and DeMarzo, 2007). Taggart (1985) stated that the value of the tax benefit is higher in countries that
are expected to have a high inflation. The positive relationship between market leverage and inflation is consistent with the static tradeoff theory (Frank and Goyal, 2009). The results for inflation, reported in Table V, are consistent with the
hypothesis (H5) that there is a negative relationship between inflation and (book) leverage. The only exception is for the United States when leverage is measured in market value. Unfortunately leverage could not be estimated in market value for the developing countries.
GDP growth rate
Table V shows that the GDP growth rate has the expected positive sign and is highly significant for each country, except for India and Indonesia. When an economy has a high growth rate this is associated with healthy firms and growth opportunities (Bartholdy and Mateus, 2008). Therefore, firms are expected to have higher leverage ratios when the GDP growth is high in a country.
On the other hand, Table V also shows a negative relationship between the GDP growth rate and leverage, contrary to my expectations. It seems that even though a country has a high GDP growth rate firms have lower leverage ratios. Frank and Goyal (2009) provide a possible explanation and stated that in times of expansions internal funds increase, ceteris paribus. Therefore, firms decrease their leverage when a country has a high GDP growth rate. The results for the GDP growth rate, reported in Table V, are consistent with the hypothesis (H6) that there is a positive relationship between GDP growth rate and (book) leverage, except for India and Indonesia. The results indicate that the relationship between GDP growth and leverage is not the same across the developing countries and the United States. This finding is consistent with Booth et al. (2001). On the other hand, De Jong et al. (2008) show a consistently significant positive relationship between GDP growth and leverage. Notable is that for the United States a significant positive result is found when leverage is measured in book value and an insignificant negative result when leverage is measured in market value. Unfortunately leverage could not be estimated in market value for the
developing countries.
Interest rate
Table V shows that the interest rate has the expected negative sign in only one country. The result for the interest rate is significantly negative at the 5 percent level
for China. When a country has a high interest rate borrowing is expensive for firms. Bartholdy and Mateus (2008) also stated that high interest rates are associated to increase the cost of borrowing.
On the other hand, for the other developing countries and the United States a positive relationship is found between the interest rate and leverage, contrary to my expectations. It seems that firms increase their leverage when the interest rate increases in a country. The results for the interest rate, reported in Table V, are only for China consistent with the hypothesis (H7) that there is a negative relationship between interest rate and leverage. The results indicate that the relationship between the interest rate and leverage is not the same across the developing countries and the United States. The findings on the interest rate in this study contribute to the existing literature on determinants of capital structure. Booth et al. (2001), De Jong et al. (2008) and Alves and Ferreira (2011) did not investigate interest as a determinant of capital structure.
A.3. Cross-Country Analysis
This section presents the most interesting part of the results: The cross-country analysis. This study investigates the impact of firm-specific and country-specific factors on leverage. The aim is to test whether the developing countries are leveraged in a similar way as firms in the United States. The explanation for the found
relationships between the independent variables and leverage has been described in the previous sections and is therefore not repeated here.
The United States is the only developed country in this study. Table V shows that all the firm-specific variables for the United States have the expected sign and are significant at the 1 percent level. These findings are consistent with earlier studies of Harris and Raviv (1991), Rajan and Zingales (1995) and Frank and Goyal (2009). Table V also shows that each firm-specific variable for the United States has the same sign, whether leverage is measured in book values or market values. This finding is consistent with Rajan and Zingales (1995).
Table V also shows the results for the developing countries in this study. The results indicate that the relationship between each firm-specific variable and leverage is the same for the developing countries and the United States. This finding is
consistent with Booth et al. (2001) and Alves and Ferreira (2011). The results for the firm-specific variables indicate that the static tradeoff theory and the pecking order
