The Relationship between International Diversification and
the Capital Structure
New evidence from EU companies
Jing Liu
S1940473
Supervisor: Prof. Dr. L.J.R. Scholtens
Second assessor: Dr. P.P.M. Smid
Aug 29
th, 2011
University of Groningen
Faculty of Economics and Business
The Relationship between International Diversification and
the Capital Structure
New evidence from EU companies
Jing Liu
1Abstract: This paper aims to investigate the capital structure determinants and the relationship between firms’ international diversification and the capital structure for European Union companies. Using a sample of 4379 firm-year observations in 15 EU countries for multinational and domestic firms over the 2001-2010 period, I find that the asset tangibility, firm size, and future growth opportunities are significantly and positively related to the financial leverage. Profitability has negative effect, while tax rate has negative influence on capital structure when it is measured by firms’ statutory tax rate. Most of the prior literatures based on US data find a negative relationship between multinationality and capital structure. My study confirms their conclusion for the Europe context. Multinational companies in EU countries have less debt level than domestic companies.
JEL Classification: G15, G32
Keywords: capital structure, financial leverage, international diversification, panel data,
fixed effects model
1
I.
Introduction
Companies’ international diversification is an important determinant of capital structure (Bradley, Jarrell, and Kim, 1984; Michel and Shaked, 1986; Fatemi, 1988; Lee and Kwok, 1988; Burgman, 1996; Chen et al., 1997; Chkir and Cosset, 2001; Akhtar and Oliver, 2009). Theoretically, it is often asserted that the international diversification of earnings should enable multinational companies (MNCs) to sustain higher levels of debt than their domestic counterparts (DCs) (Akhtar and Oliver, 2009). However, the empirical studies show different results. Most researchers, such as Fatemi (1988), Lee and Kwok (1988), Burgman (1996), Chen et al. (1997), Homaifar et al. (1998), Chkir and Cosset (2001) and Doukas and Pantzalis (2003) investigate capital structures between MNCs and DCs and all report MNCs as having less debt than DCs. These researches all use US data. The study of Akhtar (2005) shows the level of internationalization has no significant effect on Australian firms. Akhtar and Oliver (2009) base sample from Japan reach the same results as those from US sample. However, no researches examine the relationship between international diversification and the capital structure for European companies. The main objective of this thesis is to filling this blank by providing empirical evidence from European Union (EU)2 countries on the role of international diversification in determining the capital structure, and to investigate the factors that influence the capital structure. My focus is on answering these two questions:
(1) Do the factors significantly influencing the capital structure decisions identified in previous literatures also determine the financial structure of EU companies?
(2) Does international diversification have a direct relationship with capital structure? The second question is the core of this study. By using observations from 15 countries, and panel data analysis, the main result confirms the commonly accepted conclusion that the foreign involvement has significant negative effect on the capital structure. In the
2EU is composed of 27 sovereign Member States: Austria, Belgium, Bulgaria, Cyprus, the Czech
determinants of the capital structure, asset tangibility, firm size and future growth opportunities are found to be significant and positive. Profitability has negative relationship with financial structure, and the tax rate has effect on the financial leverage when use specific measurement.
The remaining sections of the thesis are outlined as follows: the next section reviews prior literatures and the underlying hypotheses; section three describes the data composition and the methodology used; section four interprets the results of the research; the conclusion closes the thesis.
II.
Literature review:
2.1 Theory background
I focus on the commonly tested theories of the capital structure, namely, the tradeoff
theory, the agency cost theory, and the pecking order theory, to investigate the significance
of the proxies of these theories, including the effect of international diversification, in determining the financial structure for EU companies. These theories have been shown up most consistently as being correlated with leverage in previous studies (see Table 1, Table 2, Harris and Raviv (1991)).
Tradeoff theory:
Goyal, 2009), and they suggest that the tradeoff model performs reasonably well in predicting capital structures for firms with typical levels of debt. The tradeoff theory emphasizes that the corporate tax rate is a fundamental driver to determine the optimal capital structure. Theoretically, DeAngelo and Masulis (1980) hypothesize a negative relationship between the corporate tax rate and the amount of debt employed by corporations. Davis (1987) finds certain weak evidences in his empirical study. Graham (2008) concludes that, in general, taxes do affect corporate financial decisions, but the magnitude of the effect is “not large”. This conclusion is confirmed by Chen and Green (2008).
Agency cost theory:
Jensen and Meckling (1976) put forward the concept of agency costs, which are resulted when managers place their own interest before the interest of the company. The agency cost may lower the market value of the whole firm. How well the company can control the losses associated with the agency problem can have a dramatic impact on its capital structure (Jensen and Meckling 1976; Myers 1977; Ryen, Vasconcellos, and Kish 1997; Laurence, 2001). There are two major conflicts generated from agency costs: between managers and shareholders, and between shareholders and bondholders.
Conflicts between managers and shareholders arise because managers hold less than 100% of the residual claim. Managers’ compensation is measured by the performance of the company, but they do not get the entire profit enhancement of their activities, and they do bear the entire cost of these activities. Consequently, they tend to invest less effort in managing the resources of the company. When face high return, high risk projects, firms financed with risky debt may pass up value-creating investment opportunities since the added debt would increase the total risk that managers have to bear. Myer (1977) shows firms with higher future opportunity have higher agency cost of debt. Jensen (1986) also finds that firms with higher-valued portfolios of investment opportunities have lower preference with debt financing. Homaifar et.al (1994) and Rajan and Zingales (1995) find that future growth opportunities have negative effect on financial leverage.
(NPV) project could potentially reduce the share value. Hence, managers have incentives to reject positive NPV projects that would increase the value of the company if they believe the benefits of the project would accrue only to bondholders, or if they decide they need cash available to keep the firm solvent. In either situation, the value is lost, and the rejection leads to the problem of underinvestment. The existence of underinvestment problem causes lenders to require collaterally because the use of secured debts can help alleviate this problem (Deesomsak, Paudyal, and Pescetto, 2004). Tangible assets such as plant, property and equipment retain values even in bankruptcy. Tangible asset-intense companies can support higher levels of debt because there is little threat to bondholders that the assets they lay claims to will be worthless (Ryen, Vasconcellos, and Kish, 1997). Rajan and Zingales (1995) argue that if a large fraction of a firm's assets are tangible, then assets should serve as collateral, diminishing the risk of the lender suffering the agency costs of debt (like risk shifting).
Pecking order theory
influence only in part of their sample.3
Effect of international diversification:
There are several reasons that the capital structure of MNCs is deviated from that of DCs. The first issue relates to a political consideration. MNCs face social and political risks that are unique to the international market, which may be perceived by investors as a sign of weakness or insecurity (Aliber, 1984). In addition, MNCs are always exposed to the criticism that they siphon funds out of countries in which they do business. Governments are, therefore, liable to limit the company's freedom to repatriate any of its profits. These factors may increase MNCs’ cost of debt, which would lower the debt level. The second concern relates to the risks of international diversification. Lessard (1973) argues that international diversification may lead to a reduction in the general risks facing a firm. In this case, they may use more debt since MNCs face lower bankruptcy burden. On the other hand, MNCs faces more future investment opportunities compared to DCs, which may induce the overinvestment problem, the increased agency cost would reduce the debt level. Third, based on the pecking order theory, costs related to information asymmetries have a substantial effect on the firm’s choice of funding. MNCs may face higher degrees of information asymmetries due to institutional, legal, socio-cultural and political differences across nations (Burgman, 1996). For example, the differences in language and legal systems as well as greater information gaps across countries could result in higher monitoring costs for MNCs, so it is more difficult for bondholders to gather information and monitor MNCs' business operations relative to domestic firms. Lee and Kwok (1988) outline several different international environmental factors that make the capital structure of MNCs different from that of DCs: foreign exchange risk; political risk; international tax differentials; international market imperfections; corporate international diversification; international availability of capital; influence of local factors on foreign affiliate's capital structure.
Table 1 Summary of empirical works on general determinants of the capital structure:
Author Country
(size)
Period Variables estimated Methodology Key results
Ferry & Jones (1979)
U.S (233) 1969-1974 1971-1976
DEP4: total debt to total assets
IND: business risk, industrial class, size and operating leverage
Howard-Harris clustering algorithm
Business risk is negatively related to debt ratio (F-test=2.61, p-value<0.05); the operating leverage has negative, linear impact on financial structure (F-test=2.13, p-value<0.1).
Bradley, Jarrell, & Kim (1984)
U.S (821) 1962-1981 DEP: BV of debt to the sum of the MV of equity and the BV of debt
IND: industrial class, volatility of firm earnings, non-debt tax shields, agency cost
ANOVA, simulation analysis, linear regression
The debt ratio is positively related to the level of non-debt tax shields (t-test=7.61), inversely related to the variability of firm value (t-test=-12.3), strongly related to industry classification; negatively related to advertising and R&D expense percentage (t-test=-13.13).
Kim & Sorensen (1986)
U.S (168) 1978-1980 DEP: long-term debt to total capitalization IND: Growth, insider ownership, operating risk, size
ANOVA and linear regression
Firms with higher inside ownership have greater debt ratios (t-test=2.03, p-value<0.05); high-growth firms use less debt (t-test=-3); high operating-risk firms use more debt (t-test=2.42); and firm size is uncorrelated to the level of debt.
Davis (1987) Canada
(115)
1966-1982 DEP: long-term debt plus notes payable to total assets
IND: effective corporate tax rate
Spearman rank and Kendall W coefficient of concordance
Effective corporate tax rate is inversely related to debt ratio (p-value<0.001).
Titman & Wessel (1988)
U.S (469) 1974-1982 DEP: 6 measurements of leverage5 IND: non-debt tax shields, growth, uniqueness, industry classification, size, earnings volatility and profitability
Linear structural modeling
Firms with unique products have low debt ratios (t-test=21.6, 19.7). Smaller firms tend to use significantly more short-term debt (t-test=-0.3, -3.2,-2.3, -2.1). No evidence to support that debt ratios are related to a firm's expected growth, non-debt tax shields, volatility, or the collateral value of its assets.
4
IND means independent variable, DEP means dependent variable. MV refers to Market value, BV refers to Book value
5
Author Country (size)
Period Variables estimated Methodology Key results
Homaifar et.al (1994)
U.S (370) 1978-1988 DEP: current portion of interest bearing long-term debt plus long-term debt normalized by sums of MV of equity, BV of debt, and BV of preferred stock IND: tax rate, size, future growth opportunities, capital market conditions, inflation rate, earnings volatility
Autoregressive distributed lag model
Firm size (t-test=3.6, p-value<0.01) and future growth opportunities (t-test=-7.8, p-value<0.01) are important determinants of the capital structure. Rajan & Zingales (1995) G-76 (4557)
1978-1991 DEP: 3 measurements of leverage7; IND: assets tangibility, investment opportunities, size, profitability
Cross-sectional regression
Tangibility is always positively correlated with leverage in all countries (p-value<0.05); market-to-book ratio enters with a negative coefficient in all countries (p-value<0.1), Size is positively correlated with leverage except in Germany where it is negatively correlated (p-value<0.01).
Shyam-Sund er & Myers (1999)
NG (157) 1971-1989 DEP: net and gross debt issued to book assets
IND: Pecking order theory, tradeoff theory
Cross-sectional regression
Pecking order model(p-value<0.01) explains much more of the time-series variance in actual debt ratios than a target adjustment model based on the static tradeoff theory
Deesomsak, Paudyal, & Pescetto (2004) Asian Pacific region8 (1527)
1993-2000 DEP: total debt/(total debt + MV of equity+ BV of preference shares)
IND: tangibility, profitability, size, growth opportunities, non-debt tax shield, liquidity, earnings volatility, and share price
Regression method with panel data
Non-tax debt shield (p-value<0.05), liquidity (p-value<0.1) and share price performance (p-value<0.1) appear to significantly influence the leverage decision in all countries; size is also a very important factor except Singapore (t-test=0.618).
6
G-7 countries are: the United States, Japan, Germany, France, Italy, the United Kingdom, and Canada
7
The ratio of total liabilities to total assets; the ratio of debt (both short term and long term) to total assets; the ratio of total debt to net assets (total assets less accounts payable and other liabilities)
8
Author Country (size)
Period Variables estimated Methodology Key results
Chen & Green (2008)
European9 (129)
1993-2005 DEP: 6 measurements of leverage10 IND: Tax rate and non-tax shield
Panel generalized method of moments (GMM) procedure
Tax policy has a significant but small impact on firms’ debt ratios (t-test=4.12, 2.47,8.5, 5.7, 3.4,2.47 for the 6 measure of leverage respectively) and that non-debt tax shields are a substitute for debt in company activities
Frank & Goyal (2009)
U.S (NG) 1950-2003 DEP: Total debt to market value of asset; Total debt to book value of asset.
IND: tradeoff theory, pecking order theory, market timing
Regression method with panel data
For explaining market leverage are: median industry leverage (+) (t-test=51.8, p-value<0.01), market-to-book assets ratio (-) (t-test=-42.7, p-value<0.01), tangibility (+)(t=17, p-value<0.01), profits (-) (t-test=-34.9, p-value<0.01), log of assets (+) (t-test=29.5, p-value<0.01), and expected inflation (+)(t-test=48, p-value<0.01); for book leverage, the impact of firm size, the market-to-book ratio, and the effect of inflation are not reliable.
9 Referred to Belgium, France, Germany, Greece Ireland, Italy, Netherlands, Portugal, Spain, Sweden, UK 10
Table 2 Summary of empirical works on the relationship between foreign involvement and the capital structure
Author Country (size)
Period Variables estimated Methodology Key results Other remarks
Fatemi (1988)
U.S (136) 1978-1982 DEP: 9 measurements of leverage11;
IND: bankruptcy cost, firm size, foreign involvement
Kruskal-Wallis one-way analysis
MNCs have lower level leverage than DCs (The Kruskal- Wallis H Statistic=23.63 for leverage measurements R1-R7).
MNCs obtain a greater proportion of their borrowing needs from short-term sources.
Lee & Kwok (1988)
U.S (834) 1964-1983 DEP: LTdebt12;
IND: agency cost, bankruptcy cost, foreign involvement
Kruskal-Wallis test, two-way ANOVA test
MNCs tended to be less leveraged than DCs (multinationality effect: F-value=7.85, p-value<0.01)
MNCs have higher agency costs (p-value<0.01) and lower bankruptcy costs than DCs (p-value<0.01).
Burgman (1996)
U.S (487) 1987-1991 DEP: LTdebt;
IND: agency cost, business risk, foreign exchange risk, political risk, foreign involvement
T-test, Mann-Whitney U-test
MNCs have lower target debt ratios than purely DCs (t-test=2.8, p-value<0.01). MNCs have higher agency costs than purely DCs (F-test=83.47, p-value<0.01)
International factors are relevant to the MNCs capital structure decision. Chen et al. (1997) U.S (2219) 1984-1993 DEP: LTdebt
IND: size, agency cost, bankruptcy cost, profitability, foreign
involvement.
Multiple regression analysis
MNCs have lower debt ratios than DCs (t-test=-9.94). The debt ratio is negatively related to both bankruptcy costs (t-test=-9.02) and growth options (t-test=-11.92).
Within the MNCs, D/E ratio is positively related to the degree of internationalization.
11 Fatemi (1988)’s nine definitions of leverage: Rl = A broad measure of financial leverage defined as the ratio of all financing other than common equity (i.e., total liabilities plus preferred stock) to total assets. R2 = A narrower measure of financial leverage defined as the ratio of all financing other than common equity and preferred stock to total assets. R3 = the familiar debt ratio defined as the ratio of the sum of current liabilities and total long-term debt to total assets. R4 = the ratio of interest-bearing debt (short-term and long-term) to total assets. R5 = the ratio of long-term debt to total assets. R6 = the ratio of current liabilities to total debt. R7 = A debt-to-equity ratio defined as the ratio of all financing other than common equity and preferred to equity. R8 = an interest-coverage ratio defined as the ratio of before-tax operating cash flows to total interest expense. R9 = the current ratio defined as the ratio of current assets to current liabilities.
12
Chkir & Cosset (2001)
U.S (973) 1992-1996 DEP: LTdebt;
IND: size, agency cost, bankruptcy cost, profitability, exchange rate exposure,political risk exposure
Switching regressions model
Leverage increase with internationalization (t-test=1.41). profitability is negatively related to the debt ratio of MNCs (t-test=-5.04, p-value<0.01)
International diversification leads to lower levels of bankruptcy risk; the effect of some of the
explanatory variables varies with the diversification. Doukas & Pantzalis (2003) U.S (6517)
1988-1994 DEP: LTdebt, STdebt, TotalDebt13;
IND: agency cost, geography diversification
Fixed effect model of panel data analysis
Firm’s increasing foreign involvement exacerbates agency costs of debt leading to lower (greater) use of long-term (short-term) debt financing (t-test= 2.85, 4.21, 3.18 for three measures of leverage respectively)
MNCs have higher agency costs of debt than DCs because geographic diversity renders active monitoring more difficult and expensive. Mittoo & Zhang (2008) Canada (4824)
1998-2002 DEP: LTdebt, STdebt, TotalDebt; IND: international involvement, market to book ratio, tangibility of asset, non-debt tax shield, size, profitability, business risk, industry effect
Linear regression
MNCs have higher leverage than DCs (t-test=2.28, p-value<0.05). The agency costs of debt and business risk increase with internationalization level, the agency costs of debt have stronger impact on leverage (coefficient=-0.58, p-value<0.05).
the sensitivity of leverage to the independent variables is different between the U.S and Canada samples, especially for the MNCs. Akhtar & Oliver (2009) Japan (365) 1994-2003 DEP: LTdebt;
IND: business risks, age, collateral value of assets, free cash flows, foreign exchange risks, growth opportunities, non-debt tax shields, political risks, profitability, size and foreign involvement
Cross-sectional time series analysis
MNC have significantly less leverage than Japanese DCs, multinationality is an important aspect of leverage for Japanese firms(t-test=-1.68, p-value<0.1)
Business risks are not found to be significantly different between the two groups of organizations.
13
2.2 Analysis of literatures on determinants of capital structure:
There are hundreds of articles examine the determinants of capital structure. To highlight the current state of the art, I consider mainly papers written since 1980. The summary of literatures is listed in Table 1. In this table only the articles that provide empirical predictions of the determinants of the capital structure and particularly, the most frequently sited papers are listed.
Titman and Wessel (1988) use a factor-analytic technique for estimating the impact of unobservable attributes on the choice of corporate debt ratios. Their selection of variables can be seen the foundation of the later researches and they serve to document empirical regularities that are consistent with existing theory. Rajan and Zingales (1995) are the first testing the capital structure determinants in multiple countries. They find that the determinants of capital structure that have been reported for the USA (size, growth, profitability, and importance of tangible assets) are important in other countries as well. They show that a good understanding of the relevant institutional context (bankruptcy law, fiscal treatment, ownership concentration, and accounting standards) is required when identifying the fundamental determinants of capital structure. By using international data, Deesomsak, Paudyal, and Pescetto (2004) investigate the determinants of corporate capital structure in the Asia Pacific region, including Malaysia, Thailand, Singapore, and Australia, covering the period of 1993-2001. Their conclusions support existing evidence with respect to firm-specific determinants. The positive effect of firm size and the negative effect of growth opportunities, non-debt tax shield, liquidity and share price performance on leverage lend support to major capital structure theories. However, they also show some significant differences between the sample countries. It can be concluded from their research that the capital structure decision is not only the product of the firm’s own characteristics but also the result of the corporate governance, legal framework and institutional environment of the countries in which the firm operates.
(Lee and Kwok, 1988; Chen et al., 1997; Doukas and Pantzalis; 2003). In a benchmark study, Lee and Kwok (1988) investigate the difference in capital structure between U.S.-based MNCs and DCs by relying on the difference in bankruptcy and agency costs of debt between MNCs and DCs. They use the foreign tax ratio (as measured by foreign pre-tax income to total pre-tax income) to classify companies as either MNCs or DCs, and find that the agency cost of debt is the dominant reason for MNCs having lower debt- to- equity ratios.
As an extension of Lee and Kwok (1988), the present study employs multivariate analyses to control for traditional capital structure. Burgman (1996) estimates the effect of foreign exchange risk and political risk on the capital structure of MNCs. Using the foreign tax ratio to classify firms as either MNCs or DCs and controlling for industry and size effects, Burgman arrives at the same conclusion with Lee and Kwok (1988). Chen et al. (1997) conduct regression analyses to investigate the effect of international activities on capital structure. Although their result is in conformance with previous findings, they find that within their sample of MNCs, debt ratios increase with the level of international activities. The result of Chkir and Cosset (2001) is contradicting to the prior researches, who find that MNCs have higher leverage level by using switching regression with sample from U.S. Akhtar (2005) tests ten variables for the sample from Australia14, finding that leverage level does not differ significantly between MNCs and DCs. However, Akhtar and Oliver (2009) test the same explanatory variables for the sample of Japan; they find MNCs have significantly less leverage than Japanese DCs, and they document that significantly positive leverage effects of foreign exchange risks and sizes are subsumed by the negative effect of business risks to explain the lower leverage experienced by Japanese multinationals relative to Japanese DCs. Doukas and Pantzalis (2003) focus on the effects of the agency costs of debt on leverage for multinational firms and non-multinational firms, their general conclusion shows that the negative effect of agency costs of debt on long-term leverage is significantly greater for MNCs than DCs.
These papers provide foundation of the independent variables selection and model build
14
for this thesis, especially Lee and Kwok (1988), who select independent variables generated from capital structure theories, and Doukas and Pantzalis (2003), who use panel data analysis. These earlier studies, however, are confined to exploring the effects of international diversification on capital structure within a single country. This thesis adds value to current literature in the following ways. It provides evidence on the relationship between the international diversification and the capital structure from sample of EU countries which are not solely tested in prior papers. EU countries are selected because it is a unique economic and political partnership, around two thirds of EU countries’ total trade is done with other EU countries. In addition, I use explanatory variables that are commonly agreed to be significantly related to the capital structure in previous studies but with a better estimation method (explanation of using panel data analysis is in section III). In order to find out that if these variables are also shown significant in determining EU firms’ capital structure decisions. Finally, I use the most recent time period that offer latest evidence on the determinants of financial leverage.
III.
Data and methodology
All EU public listed firms in Orbis dataset are initially selected. Orbis is a global database which has information on more than 73 million companies worldwide. It provides financial information, including ratios for both private and public companies. The limitation of the database is the short time period included (only most recent 8-10 years). So in order to minimize the bias due to short estimation window, the longest time period offered by Orbis is selected, i.e. 10 year period from 2001 to 2010. The market capitalization data is searched through Datastream since it is not available on Orbis in the full estimation window15. According to Rajan and Zingale (1995), financial firms such as banks and insurance companies should be eliminated from the sample because their leverage is strongly influenced by explicit (or implicit) investor insurance schemes such as deposit insurance. Furthermore, their debt-like liabilities are not strictly comparable to the debt issued by nonfinancial firms.
Therefore, financial firms, insurance firms are excluded. Then, those companies with no data reported are deleted. This process results total 4379 companies from 15 countries, including 1119 MNCs, and 3260 DCs. These firms are classified into 8 main industries according to their first digit of the SIC codes.
3.1. Model and methodology:
The first objective of the study is to investigate the general determinants of financial leverage. The model identifies five general attributes that determine the firm's leverage ratio (DR or LEV). These are the corporate tax rate (TAXPAID), future growth opportunities (M/B), the asset tangibility (AT), profitability (EBITTA), firm size (SIZE). Previous scholars prove that these independent variables have linear relationship with financial leverage (see Table 1 and 2). I want to test if the linearity exists between these variables and leverage in EU context. Accordingly, the theoretical model can be written as:
Model 1:
DRi,t (or LEVi,t) =α + β1 TAXPAIDi,t+ β2 ATi,t + β3 SIZEi,t + β4 EBITTAi,t + β5 M/Bi,t +εi,t (1)
For the second objective, the international diversification (MULT) is added, which is a variable measures whether a firm is involved in international activities or not. If MULT significantly influences the capital structure (the variable has significant positive or negative coefficient on the dependent variable), the relationship between foreign involvement and the capital structure exists. The country effect and the industry effect are controlled when the effect of MULT is tested. Then, model (1) becomes:
Model 2:
DRi,t(orLEVi,t)=α+β1TAXPAIDi,t+β2ATi,t+β3SIZEi,t+β4EBITTAi,t+β5M/Bi,t+β6MULTi,t+
+ + εi,t (2)
Both model 1 and 2 are estimated with panel data analysis, with full 4379 sample from period 2001-2010. The methodology I use are based on the prior researchers such as Doukas and Pantzalis (2003), who perform panel data analysis to investigate the relationship between the multinationality and the leverage level. Panel data regression analysis is used since the data has both cross-sectional and time series dimensions. The use of panel data provides a more satisfactory basis for my purpose. It is argued that, by providing a large number of data points and blending characteristics of the cross-sectional and time series data, panel data improves the efficiency of econometric estimates (Hsiao, 1985). Moreover, it enables to effectively choose more efficient instruments to control for endogeneity (Ozkan, 2001).
The rationale for each of the variables and their expected signs are discussed next. The hypothesized correlation and their measurements are summarized in Appendix A.
3.2. Variables and Hypothesizes:
Leverage
Among the previous literatures, there is no consensus definition of firm leverage. Fatemi (1988) tests the effect of foreign diversification on leverage. The author uses nine definitions of leverage. Generally, he reaches the same conclusion with different definitions, which is MNCs have lower level of leverage level. Rajan and Zingales (1995) use three approaches to define leverage, and generate the similar conclusions with these three measurements. Due to limited data availability16, I use two proxies for capital structure, which are the common measurements used by both papers mentioned above (see Table 1 and Table 2). The first proxy is the broad definition of leverage (LEV), is defined as:
Leverage (LEV) = Total Liability/ Total asset
The second one is called debt ratio (DR), which is defined as:
Debt Ratio (DR) = Long-Term Debt/ (Book value of Equity + Long-Term Debt)
This definition is used by most of the researchers (Bradley, Jarrell, and Kim, 1984; Lee and Kwok, 1988; Burgman, 1996; Chen et al., 1997; Chkir and Cosset, 2001; Akhtar, 2005) as a measure of firms’ financial structure. Some analysts argue the book value maybe not
appropriate to reflect the real debt and equity level, the market value should be a better proxy, Bowmen (1980) indicates that the accounting measurement may be a very good surrogate for market value of debt in the leverage variable. The book value will be used in order to compare my results to those reported by prior studies (Fatemi 1988; Lee and Kwok 1988; Titman and Wessels 1988; Burgman, 1996; Chen et al., 1997; Frank and Goyal, 2009).
Tax rate
According to the tradeoff theory, the tax rate has positive effect on the financial leverage since the tax savings benefit of debt increases with the tax rate. With higher tax rate, firms use more debt to enlarge the tax advantages. The ratio of taxes paid to pre-tax earnings will be used, which is suggested by Davis (1987) and Homaifar et.al (1994). In addition, I use statutory tax rates of corporations as an alternative measurement, on the grounds that statutory rates correspond most closely to the instruments of tax policy (Chen and Green, 2008).
H1: the tax rate (TAXPAID) is positively related to the debt ratio.
Future growth opportunities:
Higher growth opportunities provide incentives for equity holders to invest sub-optimally, or to accept risky projects that expropriate wealth from debt holders. This raises the cost of debt, and thus growth firms tend to adopt less debt. A finding of a negative and significant relationship between leverage and the present value of future growth opportunities is likely to be consistent with this inference. Frank and Goyal (2009) show that the market-to-book ratio of common equity is the most reliable proxy for growth opportunities. They argue that if market timing drives capital structure decisions, a higher market-to-book ratio should reduce leverage as firms exploit equity mispricing through equity issuances. Accordingly, I measure future growth opportunity as the ratio of the market capitalization to book value of equity, which is used the same as Homaifar, et.al (1994), Chen et al. (1997), Deesomsak, Paudyal, and Pescetto (2004), and Akhtar and Oliver (2009).
H2: future growth opportunity (M/B) is negatively related to the debt ratio
Asset tangibility
ratio tends to confirm with this argument. The asset tangibility is used to measure the agency cost of debt, it is defined as the ratio of fixed asset to the total assets, which is used commonly in previous literatures (Deesomsak, Paudyal, and Pescetto, 2004; Frank and Goyal, 2009)
H3: the asset tangibility (AT) is positively related to the debt ratio.
Profitability
In a simple pecking order world, debt typically grows when investment exceeds retained earnings and falls when investment is less than retained earnings. Thus, if profitability and investment outlays are persistent, the simple version of the model predicts that, holding investment fixed, leverage is lower for more profitable firms. Highly profitable firms might be able to finance their growth by using retained earnings and by maintaining a constant debt ratio. In contrast, less profitable firms will be forced to resort to debt financing. This conclusion is theoretically supported by Myers and Majluf (1984), and empirically supported by Chkir and Cosset (2001). The earnings before interest and tax (EBIT) scaled by total assets (EBITTA) will be used as proxy of profitability in this work. It is a frequently used ratio by current papers including Titman and Wessel (1988) and Deesomsak, Paudyal, and Pescetto (2004). The ratio of return on assets (ROA) (followed by Chen et.al, 1997), will be used for robustness checks.
H4: profitability (EBITTA) is negatively related to the debt ratio.
Firm size
The effect of foreign involvement
The empirical evidence so far suggests that MNCs have lower target debt levels than DCs (Fatemi, 1988; Burgman, 1996; Chen et al., 1997;Doukas and Pantzalis, 2003; Akhtar and Oliver, 2009). One explanation is that international diversification can lower the volatility of earnings; reduction in cash flow volatility reduces the probability of bankruptcy and therefore bankruptcy cost (Shapiro, 1978). However, MNCs are also exposed to exchange rate risk and political risk, and it may be difficult to adequately diversify these risks away. If the additional exchange rate risk and political risk incurred is more than offset by the reduction in business risk due to reduced earnings volatility, the net effect could be a higher optimal debt ratio for MNCs. On the other hand, Fatemi (1988) and Lee and Kwok (1988) argue that there are additional agency costs of debt as well as equity because of the more than proportionately larger monitoring costs associated with larger firms, greater costs of international dispute resolution, wider informational gaps, and jurisdictional and sovereignty uncertainties. Since the operations of MNCs are geographically dispersed, difficulties in gathering and processing information make monitoring more costly than the cost of monitoring domestic firms. Hence, it is expected that the inherent agency problem between shareholders and debt holders will be exacerbated with the diverse geographic structure of MNCs and, therefore, bondholders will require higher interest payments on loans to firms that are more susceptible to information asymmetries and greater monitoring costs (Doukas and Pantzalis, 2003). This reduces MNCs’ level of leverage. The conflicting prediction of this relationship in different theory make it is difficult to draw a conclusion as to whether MNCs are more leveraged than DCs. It would be interesting to examine the empirical findings. To formulate a tentative hypothesis for empirical tests, the conventional wisdom is followed:
H6: MNCs have lower debt ratio than DCs.
been suggested in the literature, including the foreign tax ratio (Lee and Kwok, 1988; Burgman, 1996; Chen et al. 1997; Chkir and Cosset, 2001), foreign sales ratio (Michel, Allen and Shaked, 1986; Fatemi, 1988; Doukas and Pantzalis, 2003). However, Lee and Kwok (1988) argue that foreign sales figures include both sales by foreign subsidiaries and sales due to exports from the parent company. Using this measure may lead to the disadvantages of mixing international trade with international investment. In addition, Michel, Allen and Shaked (1986), Akhtar (2005) use the number of countries in which the firm operates as a criterion to distinguish MNCs and DCs. This is consistent with the definition of MNC: a corporation that has its management headquarters in one country, known as the home country, and operates in several other countries17. However, there is no one single definition that is generally accepted. Due to the limited availability of data18, the number of countries in which the firm operates will be used to differentiate MNCs and DCs. If a company reports geographic segment in non-home countries, it is recognized as a MNC, if not, it is a DC. A dummy variable MULT is added to measure a firm’s degree of internationalization, for MNCs, MULT = 1; for DCs, MULT = 0.
Two control variables are added when the effect of MULT is tested in equation (2): country dummies and industry dummies. It is argued that country-specific factors (such as GDP growth and capital market development) can affect corporate leverage (Rajan and Zingales, 1995; Low and Chen, 2000). In order to control for possible country effects on leverage, fifteen country dummy variables are included in the model, they equal to 1 for Austria, Belgium, Denmark, Finland, France, Germany, Greece, Ireland, Italy, Luxemburg, Netherland, Portugal, Spain, Sweden, United Kingdom, respectively, and 0 otherwise. In addition, it is argued that the examination of capital structure should control for the industry effect. Firms in the same industry experience similar amounts of business risk, since these firms produce similar products, face similar costs of material and skilled labor, and rely on similar technology. It is reasonable to believe that financial structure is closely related to the firm’s industry classification (Ferry and Jones, 1979). The 3-digit SIC codes (a measure
17 Defined by International Labor Organization (ILO)
defined by a firm’s dominant product line) are used to classify firms into different industries. The industry dummies are added to control the industry effect, which includes 8 industries from INDUSTRY0 to INDUSTRY7, according to the first digit of firms’ 3-digit SIC codes.
3.3. Data description:
The description of data is summarized in Table 3. In Table 3, Panel A compares the debt ratio and the leverage rate, between MNCs and DCs, across different countries, for the ten-year period (2001-2010). Panel A also shows the number of observations in different countries belonging to the two groups (MNCs and DCs). Largest number of firms comes from UK (29.5%), while sample from Luxemburg contains largest percentage of MNCs (80%). The overall MNCs show higher debt ratio (1.643 compared to 0.223) and lower leverage ratio (0.564 compared to 0.72) than DCs. The last two columns compare the means of the DRs and LEVs between MNCs and DCs for each country. It is shown that MNCs and DCs have notable dissimilar debt ratios and leverage ratios at 10% significance level.
Panel B of Table 3 provides a comparison of the means of LEVs and the DRs in different industries for MNCs and DCs. About 67% of the sample is concentrated on the Manufacturing (41%) and service (26%) sector. From the comparison of the means of the DRs and LEVs between MNCs and DCs for each industry (listed in the last two rows), the two groups (MNCs and DCs) have similar debt ratios except constructions and public administration industry (p-value<0.1), and the leverage ratios are only significantly different in TCEGS19 industry (p-value=0.079).
Panel C shows the means, median, and standard deviations of the independent variables measuring the firms’ capital structure. The MNCs differ from DCs on several dimensions but the most striking difference is in the M/B ratio. M/Bs of DCs are much larger than that of MNCs (255.86 compared to 73.736). In addition, MNCs have larger size than DCs, but are significantly less profitable than DCs. The median values of these variables within these two groups are quite similar. In addition, the large standard deviation of M/B ratio, EBITTA ratio, and the firm size may indicate that the transformation process is necessary to reduce the
standard errors. The last column in Panel C provides comparison of the mean value of every independent variable between the MNCs and DCs. These two groups only differ significantly in the asset tangibility (p-value<0.01), while for the other four variables, the difference is not significant.
Panel D reports correlations between the independent variables and dependent variables, and among independent variables. It indicates that multicollinearity is unlikely to be a major problem with all correlations less than 50% (Akhtar and Oliver, 2009). On a univariate basis on the full sample the debt ratio is positively related to asset tangibility, and multinationality, negatively correlated with the firm size, but the correlations are all very small (less than 0.01) and insignificant (p-value>0.1). The correlations between debt ratio and asset tangibility is consistent with the prior hypothesis. The leverage ratio is negatively correlated to all the variables, except that firm size has insignificantly positive correlation with LEV.
There are two biases in the sample selection process. The first one relates to the use of two datasets, they may be inconsistent in variable definition or data collection process. Another selection bias arises from the fact that only listed companies are reported20. The fraction of listed firms differs widely across different countries, and so does the average size of companies listed.
Table 3- Panel A: Mean and median values of leverage ratio (LEV), debt ratio (DR) by country sector for MNCs and DCs sample for 2001-2010
Country All MNCs DCs Comparison
No. obs DR LEV No. obs DR LEV No. obs DR LEV MNC-DC
(DR) MNC-DC (LEV) 1 Austria 69 0.031 0.55 39 -0.070 0.590 30 0.165 0.498 -1.324(0.190) -0.144(0.886) 2 Belgium 116 0.291 0.525 42 0.322 0.629 74 0.247 0.519 0.475(0.636) -0.489(0.626) 3 Denmark 116 0.209 0.510 40 0.214 0.508 76 0.206 0.533 0.031(0.976) -0.846(0.399) 4 Finland 109 0.258 0.566 34 0.203 0.534 75 0.282 0.581 1.994(0.048) -0.237(0.813) 5 France 719 0.228 0.661 135 0.282 0.746 584 0.215 0.640 0.323(0.746) -1.242(0.214) 6 Germany 730 2.474 0.880 298 5.543 0.573 432 0.345 1.093 0.545(0.586) -0.559(0.576) 7 Greece 229 0.240 0.605 6 0.251 0.618 223 0.239 0.604 0.33(0.742) -1.316(0.189) 8 Ireland 58 0.217 0.475 38 0.219 0.497 20 0.215 0.432 -0.846(0.401) -0.409(0.684) 9 Italy 217 0.241 0.633 6 0.398 0.747 211 0.231 0.627 -0.016(0.987) 0.555(0.579) 10 Luxemburg 31 0.276 0.790 24 0.331 0.902 7 0.088 0.408 -0.413(0.682) -2.224(0.034) 11 Netherland 113 0.239 0.590 44 0.314 0.583 69 0.192 0.596 -2.636(0.000) -1.114(0.268) 12 Portugal 51 0.531 0.749 0 -- -- 51 0.539 0.750 -- -- 13 Spain 122 0.301 0.578 3 0.317 0.703 119 0.301 0.575 0.597(0.551) -0.497(0.620) 14 Sweden 407 0.192 0.564 78 0.209 0.488 329 0.188 0.582 1.369(0.172) -0.072(0.942) 15 UK 1292 0.163 0.694 333 0.203 0.471 959 0.148 0.772 -0.479(0.632) -0.244(0.807) Total 4379 0.586 0.680 1119 1.643 0.564 3260 0.223 0.720 -1.697(0.089) 1.873(0.062)
Table 3- Panel B: Mean and median values of leverage ratio (LEV), debt ratio (DR) by industry sector for MNCs and DCs sample for 2001-2010
Industry All firms MNCs DCs Comparison
No. obs DR LEV No. obs DR LEV No. obs DR LEV MNC-DC
(DR) MNC-DC (LEV) 0 Agriculture 47 0.196 0.466 20 0.152 0.378 27 0.229 0.531 1.311(0.197) 2.034(0.047) 1 mining 300 0.104 0.455 138 0.099 0.547 162 0.108 0.378 0.373(0.709) -0.967(0.334) 2 constructions 120 0.200 0.655 18 -0.075 0.708 102 0.248 0.646 1.714(0.089) -0.832(0.407) 3 manufacturing 1803 1.130 0.610 537 3.337 0.573 1266 0.236 0.625 -1.547(0.122) 1.318(0.317) 4 TCEGS 486 0.299 0.653 109 0.319 0.576 377 0.293 0.676 -0.434(0.665) 1.579(0.079) 5 trade 435 0.327 0.814 82 0.399 0.569 353 0.31 0.871 -0.397(0.692) 0.734(0.464) 6 service 1147 0.151 0.766 214 0.088 0.545 933 0.166 0.817 1.342(0.179) 1.029(0.304) 7 public administration 10 0.271 0.586 2 0.511 0.757 8 0.211 0.543 -1.967(0.085) -1.807(0.309) Total 4379 0.586 0.680 1119 1.643 0.564 3260 0.223 0.720 -1.697(0.089) 1.873(0.062)
Table 3- Panel C: Firm characteristics comparison between MNCs and DCs
Indep. variables All MNCs DCs Comparison
mean median std mean median std mean median std MNC-DC(△t-test)
TAXPAID -0.321 -0.225 11.888 -0.195 -0.234 4.007 -0.368 -0.221 13.721 1.188(0.235)
M/B 206.523 1.278 24730.960 73.736 1.409 6431.480 255.860 1.227 28696.770 0.8180.413)
AT 0.469 0.467 0.244 0.482 0.482 0.226 0.465 0.460 0.251 5.869(0.000)
EBITTA 5.258 0.015 966.130 -0.089 0.017 2.479 7.081 0.014 1118.717 0.579(0.562)
SIZE 2030231 81826 10936201 3278780.847 1449930.331 14077043.902 1445619.459 63353.892 9331699.983 -0.289(0.773)
MNCs are selected if the firm report geographic segment(s) other than the home country in Orbis database. DCs are those report geographic segment the same as the home country and no geographic segments. Tax rate is calculated as tax paid divided by pre-tax earnings. M/B ratio is the ratio of market value of equity to book value of equity. Asset tangibility is the ratio of fixed assets to total assets. Profitability is calculated by dividing EBIT to total assets. Firm size is the total assets. MNC-DC (△t-test) is the comparison of the mean value of each independent variable of MNCs and DCs. p-value is in parentheses.
Table 3- Panel D correlations between variables
DR LEV MUL AT EBITTA TA TAXPAID
MUL 0.009(0.101) -0.011(0.048) AT 0.004(0.489) -0.030 (0.000) 0.029(0.000) EBITTA 0.000( 0.994) -0.002( 0.691) -0.003( 0.548) 0.003(0.523) SIZE -0.001( 0.908) 0.002(0.675) 0.079(0.000) 0.136(0.000) -0.001( 0.858) TAXPAID 0.000(0.973) -0.007( 0.204) 0.007(0.229) 0.009(0.108) 0.000(0.986) 0.000(0.958) M/B 0.000(0.993) -0.001( 0.907) -0.003( 0.548) 0.005(0.325) 0.965(0.000) -0.002( 0.778) 0.000(0.997)
IV. Empirical results
4.1. Regression results
The results of model 1 and model 2 estimated with fixed effects approach are summarized in Table 4. The reason of selecting fixed effects approach of panel data analysis is that the Hausman tests reject the validation of the results from random effects model (the results of Hausman tests are shown in Appendix B). The results of these two models display similar conclusions on the general factors determine the capital structure. The asset tangibility, firm size, market to book ratio have significant and positive relationship with the debt ratio, and the alternative measurement of the capital structure, the leverage ratio. Profitability has negative influence on financial structure, and the tax rate has no significant effect.
Table 4 Results of Model 1 and Model 2
Model 1 Model 2
DR LEV DR LEV
coefficient t-value coefficient t-value coefficient t-value coefficient t-value
TR 0.000 -0.499 0.000 0.508 0.000 -0.357 0.000 0.439 AT 0.043 21.318* 0.028 16.659* 0.059 37.022* 0.004 2.672* SIZE 0.132 31.058* 0.076 21.632* 0.088 67.377* 0.093 69.061* EBITTA -0.003 -5.822* -0.005 -11.887* -0.005 -7.635* -0.004 -11.629* M/B 0.041 34.784* 0.053 54.301* 0.040 34.335* 0.052 55.102* MULT -0.099 -4.588* -0.071 -3.219* Adjusted R2 0.729 0.782 0.750 0.837 F-test 16.439* 26.555* 8.533* 4.469* Durbin-Watson stat 1.164 1.084 1.991 1.812
Residual normality test 16345.000* 16573.000* 13871.690* 13945.078*
Zero mean of error test 0.000 0.000 0.000 0.000
This table reports fixed effect model of panel data analysis for 4379 firms in the sample from period 2001-2010. Reported are the coefficients, and t-values for the following regressions:
Model 1: DRi,t (or LEVi,t) =α + β1 TAXPAIDi,t+ β2 ATi,t + β3 SIZEi,t + β4 EBITTAi,t + β5 M/Bi,t +εi,t
Model 2: DRi,t (or LEVi,t) = α + β1 TAXPAIDi,t + β2 ATi,t + β3 SIZEi,t + β4 EBITTAi,t+β5 M/Bi,t +β6 MULTi,t
+ + + εi,t
dummies and industry dummies in model 2 are not shown here. The regression includes an intercept whose coefficient is not reported. *means p-value<0.01.
The zero coefficients of tax rate and insignificant t-value prove that the tax rate has no effect on financial leverage. This result is different from that of Davis (1987) and Chen and Green (2008), who find the tax rate has significant relationship with capital structure. Although the tax benefit increases with tax rate, it is not crucial in determining firms’ capital structure. MacKie-Mason (1990) comments that the reason why some studies fail to find plausible or significant tax effects on financing behaviors, which is implied by Modigliani and Miller theorem, is because the debt to equity ratios are the cumulative result of years’ of separate decisions and most tax shields have a negligible effect on the marginal tax rate for most firms.
The significant and positive effect of asset tangibility confirms my hypothesis 3 that the asset tangibility is positively related to capital structure, and it is consistent with the result of most researchers, e.g. Rajan and Zingales (1995) and Frank and Goyal (2009). However, the coefficient is small (about 0.045 in the regression with DR, 0.02 in the regression with LEV), suggesting the collateral value of companies has minor effect on the capital structure. This result illustrates that a higher percentage of tangible assets increases in collateral value of the firm, which diminishing the lender’s risk of suffering such agency costs of debt. The result is higher leverage level. The agency cost theory is significant in explaining the capital structure determinants.
The positive sign of firm size in both models for both measurements of the financial leverage support the conclusion of Homaifar et.al (1994), Rajan and Zingales (1995), Frank and Goyal (2009) and partly support the findings of Deesomsak, Paudyal, and Pescetto (2004), who find significant and positive relationship between firm size and the capital structure. Hypothesis 5 (firm size is positively correlated with the financial structure) is verified, larger firms tend to sustain higher leverage level. One explanation behind the outcome is that larger firms have more diversified cash flow; they are less vulnerable to financial distress, so they employ more debt in their capital structure (Rajan and Zingales, 1995).
are consistent with that of Titman and Wessel (1988) and Deesomsak, Paudyal, and Pescetto (2004). The hypothesis that leverage is lower for more profitable companies is supported. The reason behind is that from the perspective of pecking order theory, more profitable firms tend to issue more equity with the retained earnings and then move to bonds and new equity only if necessary. The result is lower financial leverage ratio (Jensen, 1986). Since the profitability and firm size are variables generated from the pecking order theory, and both of them are shown significant on the capital structure, it can be concluded that the pecking order theory is also decisive in explaining the capital structure.
The market to book ratio, which is the measurement of the future growth opportunities, is significantly and positively correlated to the financial leverage in both models. This result is contradicting to hypothesis 2, that firms with higher future growth opportunity has lower financial structure, and it is also opposite to most researches, such as the result of Chen et al. (1997), Rajan and Zingales (1995) and Homaifar et.al (1994) in Table 1 and 2. One possible explanation behind the outcome may be more growth opportunities tend to increase firms’ confidence about future, so they increase the leverage level.
indicate that the explanatory variables can explain around 75% of the variation in the capital structure. In order to test the general assumptions of regression, Durbin-Watson (DW) test is performed to detect the existence of autocorrelation, The DW statistic shows there is no autocorrelation between errors for both model 1 and 2. Jarque-Bera (JB) test is conduced to test the normality of residuals; the null hypothesis that the residuals are normally distributed is rejected. In order to deal with this problem, I remove the outliers that do not fit in with the pattern of the remainder of the data. However, after this step, the residuals are still non-normally distributed (the JB test results are shown in Appendix C). Simple t-test is performed to test if error terms have zero mean, the assumption is also rejected. The violations of these assumptions demonstrate that the coefficient estimates may be inefficient, the distribution assumed for the test statistics may be inappropriate. However, I can only ignore them since no better remedies can be applied here.
In the previous section, Table 3-Panel A, B, and C show that MNCs and DCs are significantly differ on asset tangibility, and both debt ratio and leverage ratio are confirmed by the result of regression. Table 3-Panel D shows only MULT and asset tangibility have significant correlation with leverage ratio. However, from the result of regression, asset tangibility, firm size, market to book ratio, profitability and multinational diversification all have significant effect on the capital structure. The reason that the other four independent variables are shown significance here may be the result of the logarithm transformation.
4.2 Robustness checks
For robustness checks, I use alternative measurements for tax rate, firm size, and profitability to reduce the bias arises from the measurements selection of explanatory variables, these alternative measurements are also frequently used by previous researchers (as specified in Section 3.2), while the same measurements are used for the remaining two independent variables: asset tangibility and market-to-book ratio since almost all the scholars use the same definitions for them. The result is listed in Table 5.
(2008), who have positive coefficients on the tax rate for all the measurements of the capital structure (See Table 1). The results of asset tangibility, size, profitability and market to book ratio are similar with the original definitions. The most important variable I am estimating, MULT, is still significant and negative related to the two measurements of the financial structure. My hypothesis on the relationship between multinationality the capital structure is accepted. However, the violation of regression assumptions are still exists in these models21.
Table 5 Regression results with alternative definition of independent variables
Model 1 Model 2
DR LEV DR LEV
coefficient t-value coefficient t-value coefficient t-value coefficient t-value
TR -0.269 -8.412* -0.160 -6.403* -0.212 -6.729* -0.112 -4.539* AT 0.082 46.95* 0.017 10.233* 0.082 46.159* 0.021 12.979* SIZE 0.062 56.075* 0.093 83.119* 0.061 48.461* 0.092 78.812* PRO -0.012 -6.881* -0.015 -9.399* -0.010 -7.44* -0.016 -10.32* M/B 0.009 13.609* 0.016 25.696* 0.009 13.817* 0.019 29.862* MULT -0.013 -4.523* -0.028 -4.496* Adjusted R2 0.719 0.808 0.215 0.282 F-test 16.245* 26.744* 106.255* 199.436* Durbin-Watson stat 1.135 1.084 0.432 0.358
Residual normality test 4419.016* 16573.080* 4405.67* 34.115*
Zero mean of error test 0.000 0.000 -2.638* -3.690*
This table reports fixed effects model of panel data analysis for 4379 firms in the sample from period 2001-2010. Reported are the coefficients and t-values for the following variables:
Model 1: DRi,t (or LEVi,t) =α + β1 TRi,t+ β2 ATi,t + β3 SIZEi,t + β4 PROi,t + β5 M/Bi,t +εi,t
Model 2: DRi,t (or LEVi,t) = α + β1 TRi,t + β2 ATi,t + β3 SIZEi,t + β4 PROi,t+β5 M/Bi,t
+β6MULTi,t+ + + εi,t
Both models are estimated with cross-sectional fixed effects approach. LEV is calculated as total liabilities to total assets; DR is calculated as long-term debt divided by the sum of long-term debt and market value of equity. TR is the statutory tax rate of the companies. M/B is the logarithm of market to book value of equity. AT is the logarithm of the ratio of fixed assets to total assets. PRO is calculated by dividing net income to total assets. SIZE is logarithm the sales. MNCs are selected if the firm report geographic segment(s) other than the home country in Orbis database. DCs are those report geographic segment the same as the home country and no geographic segments. The coefficients of country dummies and industry dummies in model 2 are not shown here. The regression includes an intercept whose coefficient is not reported. * means p-value<0.001.
V. Conclusion
The literature on firms’ capital structure has established a series of stylized facts relating corporate financial decisions to a variety of independent variables. In this thesis, I have extended the analysis by using panel data analysis by taking account of the tradeoff theory, agency costs, pecking order theory, and the effect of international diversification with a panel of 4379 non-financial listed firms from 15 European Union countries for the period 2001 to 2010.
The main findings of this study is that the asset tangibility, firm size, and future growth opportunities have significant and positive relationship with the capital structure; profitability is negatively related to financial leverage; the tax rate, if measured by statutory tax rate, shows negative effect on capital structure, while if it is measured by the ratio of tax-paid to the pre-tax earnings, the tax rate has no impact on capital structure. In addition, leverage is negatively related to the level of multinationality, which means multinational corporations have lower leverage than domestic firms. The results have three implications. Firstly, different measurements of the variable would lead to different results. Secondly, the explanatory variables influencing the capital structure of US firms also have effects on the financial structure of EU companies. Thirdly, agency cost theory and the pecking order theory have stronger explanation power of the capital structure than tradeoff theory.
Reference list:
Aliber, R. Z. (1984), International Banking: A Survey, Journal of Money, Credit and Banking, Vol. 16(4), pp. 661-78
Akhtar, S. (2005), The Determinants of Capital Structure for Australian Multinational and Domestic Corporations, Australian Journal of Management, Vol. 30, pp. 321–4.
Akhtar S., B. Oliver (2009). Capital structure for Japanese multinational and domestic corporations, International Review of Finance, Vol. 9(1-2); pp. 1-26.
Barclay, M. J., C. W. Smith and R. L. Watts (1995), The Determinants of Corporate Leverage and Dividend Policies, Journal of Applied Corporate Finance, Vol. 7, pp. 4-19.
Bradley, M., G. Jarrell, and E .H. Kim (1984), On the Existence of an Optimal Capital Structure: Theory and Evidence, Journal of Finance, Vol. 39, pp. 857-877.
Bowman, R. G. (1980), The Importance of a Market-Value Measurement of Debt in Assessing Leverage, Journal of Accounting Research, Vol. 18, pp. 242-254
Burgman, Todd A. (1996), An empirical examination of multinational corporate capital structure, Journal of International Business Studies, Vol. 27(3), pp. 553–570
Chkir, I., and J. C. Cosset (2001), Diversification Strategy and Capital Structure of Multinational Corporations, Journal of Multinational Financial Management, Vol. 11, pp. 17–37.
Davis, A. (1987), Effective Tax Rates as Determinants of Canadian Capital Structure, Financial Management, Vol. 16, pp. 22-28.
De Angelo, H. and R.W. Masulis (1980), Leverage and Dividend Irrelevancy under Corporate and Personal Taxation, Journal of Finance, Vol. 35, pp. 453-464.
Doukas, J. A., C. Pantzalis (2003), Geographic diversification and agency costs of debt of multinational firms,
Journal of Corporate Finance, Vol. 9, pp.59–92.
Fama, E. and K. R. French (2002), Testing Trade-Off and Pecking Order Predictions about Dividends and debt, Review of Financial Studies, Vol.15, pp. 1-33.
Fatemi, A. M. (1988), The effect of international diversification on corporate financing policy, Journal of Business Research, Vol. 16, pp., 17-30.
Frank. M. Z. and V. K. Goyal (2003), Testing the Pecking Order Theory of Capital Structure, Journal of Financial Economics, Vol.67, pp.217-248.
Graham. J. R. (1996), Proxies for the Corporate Marginal Tax Rate, Journal of Financial Economics, Vol.42, pp. 187-221.
Harris, M. and A. Raviv (1991), The Theory of Capital Structure, Journal of Finance, Vol.46, pp. 297-356. Homaifar, G., J. Zietz and O. Benkato (1994), An empirical model of capital structure: Some new evidence, Journal of Business Finance and Accounting, Vol. 21, pp, 1-14
Hsiao, C. (1985), Benefits and limitations of panel data, Econometric Reviews, Vol. 4, pp. 121-174.
Jensen, M.C. and W. H. Meckling (1976), Theory of firm managerial behavior, agency costs and ownership structure, Journal of Financial Economics, Vol. 3, pp. 305-60.
Jensen, M.C. (1986), Agency Costs of Free Cash Flow, Corporate Finance, and Takeovers, American Economic Review, Vol.76, pp.323-329.
Kim, H. (1978), A Mean-Variance Theory of Optimal Capital Structure and Corporate Debt Capacity, Journal of Finance, Vol. 33, pp. 45-64.
Finance, Vol.28, pp. 911-922.
Laurance B, V. Aivazian, A. Demirguc-kunt, and V. Maksimovic (2001), Capital structures in developing countries, Journal of Finance, Vol. 1, pp.87-128
Lee, K. C. and C. C.Y. Kwok (1988), Multinational corporations vs. domestic corporations: International environmental factors and determinants of capital structure, Journal of International Business Studies, Vol. 19, pp. 195-217.
Lessard, D. R (1974), World, National, and Industry Factors in Equity Returns, Journal of Finance, Vol.29, pp. 379-91.
MacKie, M. and K. Jeffrey (1990), Do taxes affect corporate financing decisions? Journal of Finance, Vol. 45, pp. 1471-1493.
Michel, A. and I. Shaked (1986), Multinational corporations vs. domestic corporations: Financial performance and characteristics, Journal of International Business Studies, Vol. 17, pp. 89-100.
Modigliani, F. and M. Miller (1958), The cost of capital, corporation finance and the theory of investments, American Economic Review, Vol. 48, pp. 261-97.
Myers, S. C. (1977), Determinants of Corporate Borrowing, Journal of Financial Economics, Vol. 5, pp, 147-175. Myers, S. C. (1984), The capital structure puzzles, Journal of Finance, Vol. 39, pp.575–592.
Ozkan, A. (2001), Determinants of Capital Structure and Adjustment to Long Run Target: Evidence from UK Company Panel Data. Journal of Business Finance & Accounting, Vol.28, pp. 175–198.
Rajan, R. G., and L. Zingales (1995), What do we know about capital structure? Some evidence from international data, Journal of Finance, Vol. 50(5), pp.1421–1460.
Reeb, D., C. Kwok, and H. Baek (1998), Systematic Risk of the Multinational Corporation, Journal of International Business Studies, Vol.29, pp.263–79.
Ryen, G. T., Vasconcelos, G. M. and Kish, R. J. (1997), Capital structure decisions: What have we learned? Business Horizons, Vol. 40(5), pp. 41-50.
Shapiro, A. C. (1978), Financial structure and cost of capital in the multinational corporation, Journal of Financial and Quantitative Analysis, Vol. 13, pp. 211-26.
Shyam-Sunder, L, and S.C. Myers (1999), Testing Static Tradeoff against Pecking Order Models of Capital Structure, Journal of Financial Economics, Vol. 51, pp.219-244.
