1
The influence of multinationality on the
capital structure of firms in the United
Kingdom
Marit van de Kerkhof S1905821
University of Groningen Faculty of Economics and Business
Supervisor Dr. J.H. von Eije MSc Finance
DD MSc International Financial Management June 2014
ABSTRACT
This paper studies the influence of multinationality on the capital structure of U.K. firms included in the FTSE All-share index. Motivated by the increased international activity of firms, I use fixed panel data analysis to find how multinationality influences the firms’ capital structure. Current findings in the U.S. suggest a negative relation between multinationality and the leverage ratio of the firm. This relation is explained by the agency costs of debt theory. However, I find no clear influence of multinationality on the capital structure. The absence of a strong and significant negative relation between multinationality and the capital structure of U.K. firms can be caused by the less pronounced conflicts of interests between creditors and shareholders. This might be either due to the new developments within corporate governance control and more prominence on transparency or other, yet to be explained, differences between U.K. and U.S. firms.
2
I.
Introduction
Over the past decades international activity has become the norm rather than the exception. The number of multinational corporations (MNC) is rising and the international activity has an impact on the firms’ optimal capital structure. Today there is no consensus on the net effect of this impact. Literature indicated higher debt ratios for MNCs. However, in earlier empirical studies, MNCs frequently displayed lower leverage ratios than domestic corporations (DCs).
Understanding the factors influencing the optimal capital structure of MNCs can be considered very relevant. I test if the results, found in previous empirical studies completed in the U.S., also hold for current U.K. firms.
3 pronounced. The increased costs create debt funding to be less attractive and MNCs will make less use of debt within their capital structure.
To date, most prominent research on MNCs capital structure is done using U.S. multinational firms as a sample (see, e.g., Lee and Kwok, 1988; Burgman, 1996; Doukas and Pantzalis, 2003; Aggarwala and Aung Kyaw, 2010). Results of these studies create the current espoused theories on the influence of multinationality on the capital structure of the firm. However, an updated verification for the adherence to these theories outside the U.S. might be needed. Moerland (1995) and Armour, Cheffins and Skeel (2002) indicate that differences in corporate governance systems have an influence on the significance of the agency problem that arises between shareholders and creditors of a firm. During the recent years there has been a significant emphasis on corporate governance control and more prominence on transparency from firms to their stakeholders. Corporate governance has become stricter and financial reporting more regulated. This results in a reduced information asymmetry between stakeholders. The information asymmetry causes the agency costs between shareholders and creditors, corporate governance and transparencies might mitigate these agency costs. In addition, the research of Mittoo and Zhang (2008) indicates that there is a difference in the agency problem of debt arising between U.S. MNCs and Canadian MNCs. They find that country-specific factors do play an important role in the determination of the optimal capital structure of an MNC. Therefore factors affecting U.S. firms’ leverage ratio, might not have the same effect outside the U.S.
4 Given these similarities, research in the U.K. would provide a logical way to test whether the current espoused agency theory on the capital structure of MNCs has explanatory power outside the U.S. (Armour, Cheffins and Skeel, 2002). The contribution of this study will be threefold. First, I will test if the current espoused agency theory between shareholders and creditors have explanatory power outside the U.S. According to the current espoused agency costs of debt theory, MNCs will have lower debt ratios than DCs (see, e.g., Lee and Kwok, 1988; Burgman, 1996; Doukas and Pantzalis, 2003; Aggarwala and Aung Kyaw, 2010). I will test if the lowered debt ratios of MNCs are also observable within U.K. firms. Second, I will update the timeframe since previous work by Doukas and Pantzalis (2003), Mittoo and Zhang (2008) and Aggarwala and Aung Kyaw (2010). They covered samples over the periods ranging from 1988 until 1994, 1998 until 2002, and 1996 until 2005, respectively. The updated timeframe provides more insights because the current status quo of corporate governance and transparency is much different from previous research. This might result in mitigated agency costs of debt. Third, I will measure the effect of multinationality on the leverage ratio of the firm both directly with a percentage of foreign sales and by comparing MNCs and DCs. Because there is no consensus in the definition of a MNC in literature, first, the direct effect of multinational activity on the leverage ratio of the firm is tested. To be able to compare results of the differences in leverage ratios between MNCs and DCs with previous literature, I will also define an MNC within this paper.
5
II. Literature review
There are several reasons motivating dissimilar leverage ratios between MNC and DCs. The explanations to what constitute the differential benefits and costs of debt within a MNC will be elaborated on in this section. First, I will discuss the influence of the traditional trade-off theory of debt between tax benefits and bankruptcy costs on the capital structure of an MNC. Within the traditional trade-off theory, the firm weighs the benefits of tax deductibility versus the higher bankruptcy costs when increasing the leverage within the firm. Firms will choose the form of financing that maximizes their value. Firms add debt in their capital structure until the marginal benefits that stem from debt equal the marginal cost of debt. Following, I will address the influence of agency costs of debt on the capital structure of a MNC. The agency costs of debt increase due to the influence of the asset substitution problem and the debt overhang problem. Afterwards, I will provide a short overview of empirical results in literature, followed by the hypotheses and underpinnings of the hypotheses of my research.
Taxes
6 These incentives are reduced, to some extent, by the existence of the non-debt tax shield. When firms can make use of a high non-debt tax shield, consisting out of depreciation and investment tax credits, they have lower incentives for using debt, because debt benefits evaporate. MNCs can employ several non-debt tax shields to reduce taxes. Because of this reduction the debt incentives of MNCs will decrease. Results by González and González (2012), Mittoo and Zhang (2008) and Ramirez and Kwok (2010) show that there is a significant impact of the non-debt tax shield on MNCs making their capital decisions. The deductibility of corporate interest payments pushes MNCs toward a higher debt level. This effect is offset by the higher non-debt tax shield, pushing MNCs towards a lower debt level. Consequently, the impact of taxes on the optimal leverage ratio of the MNC is unique and depends on its tax benefits and non-debt tax shield (see, e.g., DeAngelo and Masulis, 1980; Fama and French, 2002).
Bankruptcy costs
Bankruptcy costs of firms consist of two components, direct and indirect bankruptcy costs. Direct bankruptcy costs relate to the legal process involved in reorganizing a bankrupt firm, like the costs of legal expenses, administration fees, advisory fees and the costs incurred in liquidating assets. Indirect bankruptcy costs are not directly related to costs of bankruptcy itself. These costs are related to the costs that can arise among potential financially distressed firms. Creditors are aware most of the direct bankruptcy costs are borne by them in the event of bankruptcy; therefore they will demand a default premium on the interest rate (Hillier, Grinblat and Titman, 2002). Fama and French (2002) argue that the bankruptcy costs are higher if earnings within the firm are more volatile and profitability is low. Higher volatility of earnings increases the operating risk of the firm, causing the threat of bankruptcy to become more realistic. Higher volatility of cash flows makes the firm less attractive to creditors. The default premium for debt increases when a firm bears a higher debt ratio. Ceteris paribus the price of debt will rise. Therefore firms with higher operating risk will have less capacity to use debt.
7 of the impact of the direct bankruptcy costs within a firm decreases when firms become larger. MNCs are often larger than DCs and which implies bankruptcy costs are partially fixed. These fixed costs result in relative lower costs. Second, diversification within MNCs causes the bankruptcy costs to fall. MNCs operate in multiple, less than perfectly correlated, economies, and therefore make use of diversification. Diversification results in a lowered earnings volatility, and hence, a lower probability of bankruptcy. It is therefore expected a MNC bear lower bankruptcy costs. According to the trade-off theory, these characteristics should result in a lower cost of debt and therefore in a higher debt ratio (González and González, 2012).
Burgman (1996) searched for the trade-off between taxes and bankruptcy costs in his research on U.S. MNCs traded on the NYSE over the years of 1987 until 1991. His results indicate that there is no significant trade-off between the tax advantages of debt and the expected bankruptcy costs; he even finds leverage to be positively related to volatility of earnings. Lee and Kwok (1988) and Burgman (1996) argue that the real trade-off to be made is between the tax advantages and agency costs of debt. The agency costs of debt in this case refer to the increased agency costs arising between stockholders and creditors. They debate that this trade-off can explain why MNCs have a lower target debt ratio, than previous theory prescribes.
Agency costs of debt
8 the conflicts between shareholders and creditors, are essential in order to determine the optimal ratio of outside equity to debt. In this paper the focus will therefore be on assessing the importance of the agency costs of debt.
The agency costs of debt arise due to conflicts of interests between shareholders and creditors of a firm. This conflict of interests arises because the incentives of equity holders are to maximize the value of their shares. This incentive is not always consistent with the incentives to maximize firm value. The influence of shareholders on management decisions can lead to sub-optimal investment decisions for the creditors. Sub-optimal investment decisions for creditors arise when management engages in high-risk investments or faces a free-rider problem. The free-rider problem can cause difficulties in renegotiating financial claims in case of financial distress, resulting in the underinvestment problem. In literature the main two sources of agency costs of debt are defined as the asset substitution problem (Jensen and Meckling, 1976) and the underinvestment problem (Myers, 1977).
The asset substitution problem
9 The underinvestment problem
The underinvestment problem appears when firms have problems attracting funding for a profitable and incentive compatible project due to shareholder incentives and pre-existing debt. When a firm has high pre-existing debt which matures after a certain investment decision, shareholders tend to reject profitable and incentive compatible projects. Shareholders reject the projects because the returns of the projects will go to the pre-existing debt holders without increasing shareholders’ wealth. In this case the suboptimal investment decisions will be made for the firm at the costs of the creditors (Lee and Kwok, 1988). Next to the refusal of projects, the costs of debt of the new project increase when creditors are more diffused. In case of one current creditor, this creditor is willing to provide the capital for the new profitable project at the market price of debt for that certain investment, to be able to regain part of his earlier investments. Yet, new creditors only get paid after the pre-existing claimants recovered their part. New creditors therefore value the investment more risky and will demand higher interest rates on their borrowings. In case the current creditors are diffused the renegotiation of the debt and refinancing at the current market price will become difficult. The diffused owners should all be willing to provide capital at the market price or otherwise the price of debt will increase. Information and free-rider problem occur (Moerland, 1995). Part of the current creditors will benefit from the returns of the new investment without investing. This is at the costs of other current investors or new investors who bare the higher risk by investing in the new project.
The problems of the shareholders non-willing to accept new projects and current creditors non-willing to reinvest in the new project to bearing the higher risk, will lead to higher agency costs of debt (Tirole, 2006).
Increased agency costs of debt within MNCs
10 Madura and Wiant, 2002; Doukas and Pantzalis, 2003; Grinblatt and Titman, 2002). The geographical dispersion of MNCs increases the level of asymmetric information between the creditor outside the firm and managers and shareholders who influence decisions within the firm. Moreover, the level of asymmetric information further increases by the more complex organizational structures and the wider range of managerial issues within MNCs (Aggarwala and Aung Kyaw, 2010). The increased asymmetric information makes it harder for creditors to gather information on the firm and hereby increases both the risk and the monitoring costs for the creditor. Creditors will therefore require higher interest payments on their loans for firms.
The increased agency costs of debt within a MNC will boost the borrowing costs for MNCs. Therefore a negative relation between multinationality and the amount of debt is expected. This relation will be more pronounced for MNCs with more foreign involvement.
Empirical results
Empirical results by Lee and Kwok (1988) and Burgman (1996) suggested that MNCs have less debt in their capital structures than DCs. They argue that the real trade-off to be made is between the tax advantages and agency costs of debt. The lowered debt ratios can be explained by the higher agency costs of debt within MNCs. Lee and Kwok (1988) find, after controlling for size and industry effects, in their research on U.S.-based MNCs that the lower bankruptcy costs are offset by the higher agency costs and non-debt tax shields within MNCs. More recent research by Doukas and Pantzalis (2003) and Aggarwal and Kyaw (2010) also controlled for the non-debt tax shield, and their results likewise indicated that lowered debt ratios for MNCs compared to DCs. Doukas and Pantzalis (2003) concluded that due to their greater geographic diversity, active monitoring of managerial decisions is more difficult in MNCs and is more expensive in comparison to DCs.
Hypotheses
11 espoused theories on the capital structure decisions of MNCs. More specific, I study if the lowered debt ratios observed in prior empirical studies completed in the U.S. also hold for U.K. firms.
My study consists of six different hypotheses testing the influence of multinationality on the leverage ratio of the firm. First, the effects of multinationality, measured in foreign sales, on the leverage ratio of U.K. firms is analysed. This approach is chosen because in literature there is no consensus in the definition of a MNC. The degree of multinationality can therefore be assumed to be the best approach. After searching for the effects of multinationality on the leverage ratio of U.K. firms, I will search for the influence of size and growth opportunities in interaction with multinationality on the leverage ratio. Next, I perform tests to measure the differences in leverage ratios between MNCs and DCs. A study of the differences in leverage ratios between MNCs and DCs create a good opportunity to compare results with previous literature.
H10: Multinationality will not have an influence on the leverage ratio of U.K. firms.
H1a: Multinationality will have an influence on the leverage ratio of U.K. firms.
The null hypothesis H10 indicates that leverage ratios of U.K. firms are not
influenced by multinationality. The alternative hypothesis H1a implies that the
12 wider range of managerial issues. This will increase the costs of gathering information for creditors. Creditors will therefore require higher interest payments on their loans and the benefit of using debt within the capital structure will decrease.
A significant positive coefficient for the multinationality variable indicates higher leverage ratios for firms with more multinational activity. Now, it can be argued that taxes are more important in the decision of the leverage ratio when increasing multinationality or the agency costs arising between creditors and shareholders are less pronounced. If taxes play a more important role for firms with more multinational activity, the not fully integrated financial markets provide the opportunity to raise more capital through foreign debt and at more favourable terms then firm with less multinational activity. This results in more favourable terms of using debt and hence, higher debt ratios. Of course, also lowered agency costs arising between creditors and shareholders can also explain the higher debt ratios. Over the recent years, there has been more emphasis on corporate governance control and more prominence on transparency and therefore it becomes easier for creditors to gather information on a firm.
13 multinationality. Creditors will require lower interest payments. This leads to the formulation of the following hypotheses:
H20: The size of the firm does not have an influence on the relation between multinationality and the leverage ratio of the firm.
H2a: The size of the firm does have an influence on the relation between multinationality and the leverage ratio of the firm.
The null hypothesis H20 indicates that the relation between multinationality and
the leverage ratio of U.K. firms is not influenced by the size of the firm. A significant positive coefficient for the multinationality variable in interaction with the firm size is in line with the results of González and González (2012). The positive coefficient indicates that the conflicts between creditors and shareholders are less pronounced for larger firms. A significant negative coefficient for the multinationality variable in interaction with firm size indicates that multinationality increases the proportional agency costs within a firm.
In previous literature, Doukas and Pantzalis (2003) and Mittoo and Zhang (2008) argue that the conflicts between creditors and shareholders are more pronounced for firms with more growth opportunities. Growth opportunities provide the ability to a firm to transfer wealth from creditors to shareholders via asset substitution. Shareholders have the opportunity to increase the overall risk of the firm by engaging in new high risk projects. Creditors, who made agreements based on the earlier risk of the firm, will bear the costs of the higher risk, without being compensated. Therefore the agency costs of multinationality, arising from the increased level of asymmetric information between the shareholders and creditors, will be increased by the higher opportunities of expropriation. This indicates that the influence of multinationality should have a larger negative effect on the leverage ratio in firms with higher growth opportunities. This leads to the formulation of the following hypotheses:
14 H3a: The growth opportunities of the firm do have an influence on the relation between multinationality and the leverage ratio of the firm
The null hypothesis H30 indicates that the relation between multinationality and
the leverage ratio of U.K. firms is not influenced by the growth opportunities of the firm. Arguments provided by Doukas and Pantzalis (2003) and Mittoo and Zhang (2008) reason for a significant negative coefficient for the multinationality variable in interaction with growth opportunities. This negative coefficient shows that the agency costs of debt are more pronounced for firms with higher growth opportunities.
Next, I search for the differences in leverage ratios between MNCs and DCs.
H40: There is no difference in the leverage ratio of U.K. MNCs and U.K. DCs. H4a: There is a difference in the leverage ratio of U.K. MNCs and U.K. DCs. The null hypothesis H40 indicates that leverage ratios of MNCs are similar to
leverage ratios of U.K. DCs, while the alternative hypothesis H4a implies that
MNCs have altered leverage ratios compared to DCs. A rejection of the null hypothesis with a significant negative coefficient will indicate lower leverage ratios among U.K. MNCs. A significant negative coefficient will be in line with previous empirical findings on U.S. MNCs. Again, with the lowered leverage ratios, it can be argued that agency costs of debt do play an important role for the capital decisions of U.K. MNCs. A significant positive coefficient of the multinationality dummy will indicate higher leverage ratios among U.K. MNCs compared to U.K. DCs. This again can be explained by the more important role of taxes or the less pronounced agency costs.
15 H50: There is no difference between U.K. MNCs and U.K. DCs in how much size impacts the leverage ratio of the firm.
H5a: There is a difference between U.K. MNCs and U.K. DCs in how much size impacts the leverage ratio of the firm.
H60: There is no difference between U.K. MNCs and U.K. DCs in how much growth opportunities impact the leverage ratio of the firm.
H6a: There is a difference between U.K. MNCs and U.K. DCs in how much growth opportunities impact the leverage ratio of the firm.
The null hypothesis H50 indicates that there is no difference between MNCs and
DCs in how much size impacts the leverage ratio of the firm. A significant positive coefficient of the multinationality dummy in interaction with firm size indicates that MNCs are less affected by the impact of size on the leverage ratio of the firm compared to DCs. A significant negative coefficient would indicate that MNCs are more affected by the impact of size on the leverage ratio of the firm compared to DCs.
Finally, the null hypothesis H60 indicates the influence of growth opportunities is
equal for MNCs and DCs. A significant negative coefficient indicates MNCs react more strongly, adjusting their leverage ratio, on the agency costs of growth opportunities than DCs do. Although reasons were provided, both research of Doukas and Pantzalis (2003) and Mittoo and Zhang (2008) did not find any significant differences in their differences’ tests between the growth opportunities of MNCs and DCs. A significant positive coefficient would indicate that MNCs react less strongly, adjusting their leverage ratio, on the agency costs of growth opportunities than DCs do.
II.
Data selection and methodology
16 variable. An overview of the different possibilities is presented and a selection for the measurement criteria is made. Furthermore, I provide a description of the control variables, followed by the summary statistics. Subsequent, the research methodology is explained. As a final point, I present the robustness checks to strengthen the results of my study.
Sample and data collection
This study relies on collected data of all U.K. MNCs and pure U.K DCs included in the FTSE All-Share index 1 over the period ranging from 2000 - 2012. Originally, I
included data of all 622 firms in the All-Share index. Data of the included companies is obtained from DataStream and Orbis. In line with most previous studies on the capital structure, firms in the financial, I excluded the insurance and real estate sector and utility sector. European industrial activity classification codes (NACE Rev.2) are used to determine these sectors. Firms within the financial sector, real estate and utility sector are excluded because their leverage ratios are frequently influenced by governmental regulation or arise from other determinants than previously discussed (Ramirez and Kwok, 2010, Gonzales and Gonzales, 2012). In addition, I excluded firms within the residual category “other service activities”. A closer look at this category shows that it includes firms named “UK Commercial Property trust LTD.” and “Blackrock Commodities Income Investment Trust”. It is expected that these firms will show similar characteristics to financial firms and therefore this category is excluded from the analysis. Next, all firms for which data required for the analysis were unavailable are excluded. The final sample consists of 331 firms and 4279 firm year observations.
The leverage variable
The dependent variable in the study is corporate leverage. Corporate leverage is measured with two different proxies: the long-term debt ratio and the total debt ratio. In line with the analysis of Doukas and Pantzalis (2003) and Mittoo and Zhang (2008) the total debt ratio is measured by the amount of long-term debt
1 The FTSE All-Share index is the market value and free float-adjusted index representing
17 plus the short term debt and the current portion of long-term debt, divided by the sum of the total debt plus the value of equity. Long-term debt is measured by
the amount of long-term debt divided by the sum of the total debt plus the value of equity.In both proxies the equity variable is measured on market-basis and debt ratios are expressed in percentage terms.
The multinationality variable
In literature different measures are used to determine the degree of multinationality of a firm 234. In prior empirical research the percentage of
foreign assets and the percentage of foreign sales are commonly used to identify MNCs. I will follow the approach of Ramirez and Kwok (2010) and Aggarwal and Kyaw (2010) and use the percentage of foreign sales to measure multinationality. Because there is no consensus in the definition of a MNC in literature, first, I test the direct effect of multinational activity on the leverage ratio of the firm. To be able to compare results of the differences in leverage ratios between MNCs and DCs with previous literature, I will also define a MNC. A firm will be considered a MNC with a foreign sales ratio of more than 10, 20 and 50 percent respectively, consistent with measures used in previous literature (Doukas and Pantzalis, 2003; Mittoo and Zhang, 2008; Ramirez and Kwok, 2010; Aggarwala and Aung Kyaw, 2010). I use a dummy variable to distinguish between MNCs and DCs. The MNCs dummy variable will be 1 if the amount of foreign sales as a percentage of total sales exceeds the prescribed percentage; if the percentage of foreign sales does not exceed prescribed percentage the dummy will be 0. In line with the research of Doukas and Pantzalis (2003) I perform mean and median difference tests to get a closer comparison of aggregate debt ratios between MNCs and DCs with the different degrees of foreign involvement. Results of the tests for both long-term debt and total debt are provided in table 1.
2 Doukas and Pantzalis (2003) base their definition of multinationality on requirements of the
Statement of Financial Accounting Standard No. 14, identifying MNCs when they report ratios of foreign assets, foreign sales or foreign income of at least 10%.
3 Mittoo and Zhang (2008) measure the degree of multinationality only by the percentage of
foreign assets over total assets.
4 Ramirez and Kwok (2010) and Aggarwal and Kyaw (2010) use the percentage of the firm’s
18 The mean-difference test shows that both long-term debt and total debt ratios are significantly lower for MNCs. This result is observable for all different degrees of foreign involvement. My findings are in line with the findings of Doukas and Pantzalis (2003), Mittoo and Zhang (2008) and Aggarwal and Kyaw (2010). The Wilcoxon rank sum z-score test reports no significant median differences for both total debt and long-term debt for MNCs, independent of the foreign sales ratio used to determine the consideration of a MNC. MNCs included in the FTSE all-share index, on average, have lower long-term debt ratios compared to DCs. The differences remain similar for overall debt ratios. Doukas and Pantzalis (2003) explain the effects of multinationality on the capital
Table 1
Mean and median [in brackets] values of the different leverage measures between MNC and DCs constructed based on the degree of foreign involvement Foreign involvement is measured by the amount of foreign sales (foreign sales/ total sales). Total debt is measured by (long-term debt + short term debt & current portion of long-term debt) / (long-term debt + short term debt & current portion of long-term debt + value of equity), long-term debt is measured by long-term debt / (long-term debt + short term debt & current portion of long-term debt + value of equity), short-term debt is measured by short-term debt/ (long-term debt + short term debt & current portion of long-term debt + value of equity). Non-reported foreign-asset ratios are considered 0.
Foreign sales MNCs > 10% Foreign sales MNCs > 20% Foreign sales MNCs > 50%
MNC obs: 2317 DC obs: 814 MNC obs: 2111 DC obs: 1020 MNC obs: 1506 DC obs: 1625 LTD (%) 15.681 17.334 15.761 16.834 15.445 16.727 [12.720] [11.774] [12.904] [11.801] [13.008] [12.141]
Means [medians] difference test
MNC - DC t-test 2.506* (0.012) 1.737** (0.009) 2.212** (0.027) Wilcoxon rank sum z (0.827) 0.218 (0.608) 0.513 (0.820) 0.228 TD (%) 20.184 21.850 20.216 21.448 19.701 21.466 [17.318] [18.045] [17.490] [17.620] [17.147] [17.736]
Means [medians] difference test
MNC - DC t-test 2.273** (0.023) 1.796** (0.073) 2.744** (0.006) Wilcoxon rank sum z (0.754) 0.314 (0.835) 0.208 (0.242) 1.169
19 structure of the firm can be better measured by long-term debt ratio. Results of the total debt ratios are ambiguous because of short-term debt ratios. No remarkable differences are observed between the different selection criteria of a firm being a MNC. Therefore, for the remainder of my research, firm will be categorized as a MNC if the firm has more than 20% foreign sales. This is in line with the research of Aggarwal and Kyaw (2010).
Control variables
Firm size
The firm size is positively related to the leverage ratio of the firm according to Warner (1977) and González and González (2012). Firm size is suggested to be positively related due to the decrease of relative direct bankruptcy costs, their easier access to capital markets, the additional diversification within larger firms and their greater likelihood of using tax shields from interest payments, thereby increasing their tax benefits. Within larger firms the relative costs of direct bankruptcy decrease. The costs decrease because direct costs remain relatively stable, whereas the firm value increases. The relative costs of bankruptcy are becoming less significant within the capital structures decisions. Internal capital markets also become more relevant for larger firms. These markets become more relevant because large firms have more resources that can be used to finance new investments, without the need of attracting external capital. Additional diversification is possible because larger firm are assumed to stretch across more cultural, economic and political divisions and therefore are geographically more diversified than smaller firms. The wider geographical diversification also creates more possibilities of using tax shields from interest payments, increasing their tax benefits. I measure the size of the firm by the natural log of the book value of total assets (González and González, 2012; Titman and Wessels, 1988; Ferri and Jones, 1979).
Market to book ratio
20 the firm to have strong financing needs and this will lead to the firm trying to attract more debt (González and González, 2012). This will generate a higher leverage ratio within the firm. A positive relation is expected between the growth opportunities and the leverage ratio of the firm. Growth opportunities of the firm are measured by the market-to-book value of equity.
Non-debt tax shield
Firms make use of non-debt tax shields to reduce taxes, such as depreciation and investment tax credits. MNCs can employ several non-debt tax shields to reduce taxes further. When firms use non-debt tax shield they have a lower incentive to use debt. Results by González and González (2012) confirm this by reporting a negative effect of the non-debt tax shield on the target leverage of a firm. Results by Doukas and Pantzalis (2003) show that the non-debt tax shield has a significant negative impact on the long-term leverage ratio of the firm. Also Mittoo and Zang (2008) and Ramirez and Kwok (2010) find significant negative results for the influence of the non-debt tax shield. I will measure the non-debt tax shield by operating income before depreciation and amortization minus the interest expense and minus taxes paid divided by the corporate tax rate.
Operating risk
21 Profitability
According to Myers’ (1984) pecking order theory of capital more profitable firms are likely to hold less debt. According to the pecking order theory firms prefer the most cost efficient financing resource. If available, internal resources are the most cost efficient financing resource. Firms with a high profitability are able to finance new projects with internal resource and therefore do not need to acquire debt. Consequently, their debt ratios will be lower. Results of Doukas and Pantzalis (2003) and Akhtar (2005) show that MNCs, on average, are more profitable compared to DCs. Consequently a negative relation between multinationality and the amount of long-term debt within the firm is expected. The profitability within the firm is measured by the average of the cash flow divided by sales for the past three years.
Dividend payout ratio
Also the dividend payout ratio is included as a control variable. According to Doukas and Pantzalis (2003) firms who pay out a high portion of their net income have a lower internal capital available for new investment opportunities, regardless of the efficiency of the internal market. Thereby firms who pay out more dividends could signal expected future returns and this could make it possible for firms to attract more debt. The dividend payout ratio and debt are therefore expected to have a positive relation. The dividend payout ratio is measured by the annual dividends per ordinary share times common shares outstanding, divided by the net income.
Appendix A contains a more detailed overview of the definitions of the variables used within this paper and their predicted influence on the debt ratio of the firm.
Summary statistics
22 get a more detailed view of the coefficients’ deviations from zero in the regression results. An overview of the summary statistics of the variables when a firm is considered a MNC with a foreign sales ratio of more than 10 and 50 percent respectively is provided in appendix B.
Table 2
Descriptive Statistics: Means and standard deviations of the variables included in the regression models, split in a multinational (MNCs) a domestic (DCs) sample based on the multinationality dummy of 20%. The variables total debt (TD) and long-term debt (LTD) are measured in percentages, they contain thirteen yearly observations ranging from 2000-2012. The variables multinationality (MNC10, MNC20, MNC50, FRS), market-to-book value (MTB) and dividend payout ratio (DIVPO) are measured in ratios. Size (SIZE) is measured by taking a logarithm. Non-debt tax shield (NDTS) is measured by the formula provided on appendix B. All the previous independent variables are lagged for one year, and therefore contain thirteen yearly observations ranging from 1999-2011. Operating Risk (OPRISK) is measured by a ratio comprising a prior 5 year standard deviation. The observations range from 1995-2007. Profitability (PROF) is measured by a ratio comprising a prior 3 year average. The observations range from 1998-2009.
Total sample; N =
2889 MNCs > 20% Foreign sales; N = 1952 DCs < 20% Foreign sales; N = 937
Mean Std. Dev. Mean Std. Dev. Mean Std. Dev.
LTD (%) 16.110 16.210 15.761 15.197 16.834 18.113 TD (%) 20.617 17.997 20.212 17.051 21.448 19.815 FRS (ratio) 44.744 36.000 66.448 24.846 4.133 5.868 MNC20 (ratio) 0.652 0.477 1.000 0.000 0.000 0.000 SIZE (ln) 13.406 1.749 13.603 1.824 13.008 1.511 MTB (ratio) 3.162 111.140 3.303 134.860 2.872 20.375 NDTS (ratio) 0.081 0.359 0.087 0.409 0.071 0.235 OPRISK (ratio) 8.503 97.585 9.476 114.160 6.566 50.287 PROF (ratio) 0.049 0.672 0.034 0.581 0.081 0.832 DIVPO (ratio) 0.105 3.590 0.042 0.726 0.236 6.209
23 has been a significant drop in the exchange rate from 1,64 GBP/EU 5 in February
2002 towards 1,05 GBP/EU in December 2008. According to Rees and Unni (2005) 87% of large U.K. firms have significantly been exposed to exchange rate fluctuations with the Euro over the years. In addition, Olugbode, El-Masry, and Pointon (2014) explain these high exchange rate fluctuations can impact the cash flows and volatility of multinational firms. Due to the foreign sales of MNCs, their cash flows will be more vulnerable towards the exchange rate fluctuations. The decreasing value of the British pound results in lower foreign cash flows. Therefore the high exchange rate fluctuations and decreasing value of the British Pound over the years in the sample explain the lowered profitability and high cash flows fluctuations of the MNCs.
Next to refute multicollinearity, the correlation between the independent variables should not be higher than 80 percent (Brooks, 2008). The correlation matrix in appendix C indicates that the independent variables are below 80 percent correlated and therefore the standard errors of the coefficients will not be additionally increased by multicollinearity.
Methodology
Earlier research by Mittoo and Zhang (2008) and Aggarwala and Aung Kyaw (2010) used an independently pooled sample of MNC and DCs as their main research method. However, the coefficients of using an OLS regression using the pooled sample suffer from biases due to implicitly assuming that the average values of the variables and the relations between them are constant over time and across all of the cross-sectional units in the sample (Brooks, 2008). This bias leads towards autocorrelation of the residuals. The previous pooling procedure could cause problems as the correlation of error terms across years biases the regression coefficients. A way to overcome this problem is by using the fixed or random panel data. This provides the possibility of examining how variables, which change over time, influence a dependent variable. It embodies information across both time and space. Doukas and Pantzalis (2003) used an independently pooled sample and verified these results by fixed panel data.
24 The data set for my research consist of panel data with both a time series and a cross-section dimension. Therefore, within this research, I will directly use panel data. To statistically verify between the random effect model and fixed effect model a Hausman test is performed. Results of the Hausman test indicate that the 0-hypotheses, indicating no differences between the random and fixed effects model, should be rejected6. Therefore I will use the fixed effects panel data
model. This model emphasis on the within-subject variability, it is designed to study changes within the firm. Because the fixed effects model reduces bias by controlling for all stable characteristics of the firm it will provide the most useful information on the influence of multinationality on the leverage ratio of the firm. In addition, the fixed panel data model is chosen because this model is preferred with missing data in the sample. An unbalanced panel data structure can be used because the sample then constitutes all available information on the shares included in the FTSE All-share index. Consequently, an unbalanced panel data structure with both cross-section and time-invariant fixed effects is used.
My main model, testing the influence of multinationality on the leverage ratio of U.K. firms is provided below.
All independent variables are lagged values. The values are lagged to avoid endogeneity of all independent variables. In the equations i represents the firm index and t represents the time index. I will first test the model without including the interaction terms into the regression. The interactive terms consists out of the multinational variable times the explanatory variable of size and growth opportunities, respectively, and will be included after. Next, I test if the relation between multinationality and the leverage ratio of U.K. firms is a possible
6 Results of the Hausman test indicate that the 0-hypotheses, no differences between the random
and fixed effects model, should be rejected. The chi-square distribution with 7 degrees of freedom is 18.232 and can be rejected with a probability of 0.010.
25 quadratic relation. Consequently, additional tests are conducted to analyse whether the leverage ratios of MNCs significantly differ from the leverage ratios of DCs.
Robustness Test
The first robustness test is implicitly done within the former regressions. The results of the regressions measuring the effect of multinationality on the leverage ratio within the firm are verified by the differences in leverage ratios between MNCs and DCs. When coefficients are significantly positive related to the long-term leverage ratio, the dummy for MNCs is also expected to have a significantly positive coefficient. To further strengthen this study, a second robustness tests is conducted. In this robustness test it is tested whether the results of the regressions with long-term debt also hold when long-term debt is replaced by total debt.
III. Empirical results
26
Table 3
This table contains fixed panel data regression results with long-term debt (LTD) used as the dependent variable for the period ranging from 2000-2012. Column 1 indicates the first equation using the foreign sales ratio measuring multinationality. Column 2 represents the equations including the interaction variables for size and growth opportunities with the foreign sales ratio. Equations (3) and (4) examine if the relation between multinationality and the leverage ratio of U.K. firms is a possible quadratic relation. Column 5 indicates the equation including the dummy (MNC20) separating MNCs and DCs. Column 6, includes the interaction variables in the regression with the dummy separating MNCs and DCs.
1 2 3 4 5 6 FRS (ratio) -0.042** 0.086 -0.048 0.098 MNC20 (ratio) 0.389 1.964 SIZE (ln) 3.123** 3.577** 3.135** 3.808** 2.912** 2.992** MTB (ratio) 0.000 -0.006 0.000 -0.005 0.000 -0.004 NDTS (ratio) -0.499 -0.494 -0.499 -0.495 -0.518 -0.517 OPRISK (ratio) 0.004 0.005 0.004 0.005 0.005 0.006 PROF (ratio) -2.206** -2.171** -2.208** -2.166** -2.205** -2.204** DIVPO (ratio) -0.014 -0.014 -0.014 -0.014 -0.013 -0.013 (FRS*SIZE) -0.009 -0.013 (FRS*MTB) 0.000 0.000 FRS² (ratio) 0.000 0.000 (MNC20*SIZE) -0.122 (MNC20*MTB) 0.004 Intercept -23.684** -29.681** -23.710** -32.016** -23.095** -24.101** R-squared 0.734 0.735 0.734 0.735 0.733 0.733 Adjusted R-squared 0.695 0.695 0.695 0.695 0.694 0.694 F-statistic 18.863 18.767 18.795 18.712 18.762 18.628 N 2484 2484 2484 2484 2484 2484
27 My results show that in the main regression the leverage-ratio of the firm is significantly influenced by multinationality. The results of the first regression indicate that firms with more foreign sales have smaller leverage ratios. However, the change in leverage is small. As previously mentioned, I included the leverage in the regression as a percentage and the independent variables with ratios to create a more detailed overview of the coefficients’ deviations from zero. Interpretation of the foreign sales ratio indicates that if a firm increases its foreign sales by one percentage this would imply a 0.00042 percentage decrease in leverage. If the interaction terms with size and the growth opportunities of the firm are included, the coefficient of foreign sales becomes insignificant. The third and fourth regression show that there is no significant quadratic relation between multinationality and the leverage ratio of U.K. firms. Also there is no significant difference in the leverage ratios between MNCs and DCs. Although a small decrease in the leverage ratio is visible for the first regression if multinationality increases, the overall findings of this study are contrary to earlier results of Lee and Kwok (1988), Burgman (1996) and Doukas and Pantzalis (2003). They suggested that MNCs have significant less debt in their capital structures compared to DCs.
The size variable has a significant positive effect on the long-term debt levels of the firm. Size is significant in all regressions. The positive effect of firm size on the long-term debt level of the firm can be explained by the decrease of relative direct bankruptcy costs, an easier access to capital markets, the additional diversification within larger firms and their greater likelihood of using tax shields from interest payments, thereby increasing their tax benefits. Results also suggest that more profitable firms have a lower long-term leverage ratio. This is in line with the pecking order theory. More profitable firms will have higher internal resources and will be able to finance new projects with internal resources. The internal resources are preferred since these are the most cost efficient financing resource, and therefore firms do not need to acquire debt.
28 Therefore a positive relation was expected from the influence of size in interaction with multinationality on the leverage ratio of U.K. firms. The positive relation was expected because shareholders are more external and therefore have less ability of expropriating creditors at the time of conflicts of interests between shareholders and creditors. However, results of this research indicate that the conflicts between creditors and shareholders, arising from multinationality, are not influenced by the size of the firm. Results of the interaction variable of growth opportunities and multinationality on the leverage ratio of the firm are also insignificant. Doukas and Pantzalis (2003) and Mittoo and Zhang (2008) argued that the agency costs of multinationality, arising from the increased level of asymmetric information between the shareholders and creditors, will be increased by the higher opportunity of expropriation when the firm has a lot of growth opportunities. Therefore a negative relation was expected from the influence of growth opportunities in interaction with multinationality on the leverage ratio of U.K. firms. Results of my analyses indicate that there is no significant relation.
The use of panel data provides substantial insights on the relation between multinationality and the long-term debt ratio. The R squared of the results appears to be quite high; within this research around 70%7 approximates the
real data points. The earlier research by Doukas and Pantzalis (2003) reported a R squared of around 7% and Aggarwal and Kyaw (2010) reported values around 35%.
Next, the results of the robustness test are presented in table 4. Within the robustness test the long-term debt ratio is replaced with the total debt ratio. The robustness test measures the effect of multinationality on the total debt level of the firm. The robustness test shows a stronger effect of multinationality on the leverage ratio of the firm for my main regression. In addition, the coefficient of foreign sales is still significant when the quadratic variable of foreign sales is added. Size remains to be positively related to the leverage ratio of the firm. The profitability variable is only significantly different between MNCs and DCs.
7 The high approximates of the real data points is reached by using panel data. An OLS regression
29
Table 4
This table contains the robustness check. Total debt (TD) is used as the dependent variable for the period ranging from 2000-2012. Column 1 indicates the first equation using the foreign sales ratio measuring multinationality. Column 2 represents the equations including the interaction variables for size and growth opportunities with the foreign sales ratio. Equations (3) and (4) examine if the relation between multinationality and the leverage ratio of U.K. firms is a possible quadratic relation. Column 5 indicates the equation including the dummy (MNC20) separating MNCs and DCs. Column 6, includes the interaction variables in the regression with the dummy separating MNCs and DCs.
1 2 3 4 5 6 FRS (ratio) -0.065** -0.002 -0.089* 0.0131 MNC20 (ratio) 0.389 -5.747 SIZE (ln) 3.390** 3.619** 3.436** 3.904** 2.912** 2.751** MTB (ratio) 0.001 -0.006 0.001 -0.004 0.000 -0.005 NDTS (ratio) -0.526 -0.525 -0.528 -0.525 -0.518 -0.568 OPRISK (ratio) 0.000 0.000 0.000 0.000 0.005 0.002 PROF (ratio) -0.909 -0.895 -0.918 -0.890 -2.205** -0.919 DIVPO (ratio) -0.020 -0.020 -0.020 -0.020 -0.013 -0.019 (FRS*SIZE) -0.005 -0.009 (FRS*MTB) 0.000 0.000 FRS² (ratio) 0.000 0.000 (MNC20*SIZE) 0.484 (MNC20*MTB) 0.005 Intercept -21.802** -28.805** -21.908** -27.692** -23.095** -16.868** R-squared 0.747 0.747 0.747 0.748 0.733 0.745 Adjusted R-squared 0.710 0.710 0.710 0.710 0.694 0.708 F-statistic 22.204 20.073 20.138 20.018 18.762 19.847 N 2484 2484 2484 2484 2484 2484
30
IV.
Conclusions and discussion
Motivated by the increased international activity of firms and current theories mainly based on U.S. results, I used panel data analysis to find the influence of multinationality on the capital structure of U.K. firms included in the FTSE All-share index. The relation between multinationality and the amount of long-term debt within a company can be explained by the traditional trade-off theory and the agency costs of debt theory. The traditional trade-off theory predicts that MNCs should have higher debt ratios compared to DCs. Earlier research within the U.S., however, showed results indicating lowered debt ratios. These results were explained by the agency costs of debt. This theory assumes that in imperfect capital markets, agency costs of debt arise from conflicts of interests between firms’ creditors and shareholders. The agency costs are driven by the increased costs of monitoring and controlling operations, caused by the increased level of asymmetric information, due to the geographical dispersion, more complex organizational structures and a wider range of managerial issues MNCs face.
I did not find a consistent unambiguous relation between multinationality and the leverage ratio of the firm in the regressions. In addition, there were no significant differences between the leverage ratios of MNCs and DCs. When the foreign sales ratio is directly used as a variable measuring multinationality, I observed a marginal negative significant effect. Although significant, the alteration from zero is very small. When adding the interaction variables the significance is not consistent. Also I did not find any significant difference in the leverage ratios between MNCs and DCs. This leads to the overall conclusion of this research: the lowered debt ratios within MNCs, indicated by earlier research, are not that robust in the current sample of U.K. firms.
31 ratios within MNCs. Different arguments can be provided why U.K. MNCs do not differ from DCs, contrary to U.S. findings. It can be argued that the effect of agency costs of debt, on the capital decisions, is offset by the more important role of taxes within U.K. MNCs. Huizinga, Leaven and Nicodeme (2008) used an European sample to search for the sensitivity of the capital structure of MNCs to taxes. Their results indicate that parent-country corporate tax ratios have a relatively small effect on the capital structure of the MNCs. The effect is small because MNCs are able to defer the parent-country taxes on foreign-source income. Therefore the corporate tax rate within the parent-country has less effect on the capital structure of the firm. Results of Huizinga, Leaven and Nicodeme (2008) imply that the differences between the U.K. and U.S. corporate tax rates can most likely not fully justify why the results do not indicate capital structure differences for U.K. MNCs and DCs.
Another, more likely clarification is that the agency costs of debt, arising between creditors and shareholders, are less pronounced in the U.K. sample ranging from 2000-2012. Over the recent years, there has been significant emphasis on corporate governance control and more prominence on transparency. Continuing upon the foundations of the Cadbury (1992) recommendations, corporate governance within the U.K. has evolved significantly with series of far-reaching reports (e.g. Greensbury and Hampel), the adoption of internationally accepted accounting and auditing standards, and the revising of the combined Code (Chiu, 2014). Corporate governance and increased financial reporting completes two crucial roles in reducing creditor agency costs. First, it reduces the problems of debt overhang by reducing information asymmetries between existing and new creditors. Because the firm provides timely information concerning its financial conditions it becomes easier for managers to renegotiate debt contracts.
32 universal and therefore it becomes easier to use the reported numbers for input (Armstrong, Wayne and Weber, 2010).
A third reason for the less pronounced agency costs is provided by Armour, Cheffins and Skeel (2002). Armour, Cheffins and Skeel (2002) argue that debt financing is more concentrated within the U.K. compared to the U.S.. The U.K. is similar to the U.S., as they are both market-oriented system and shareholder economies, with the main goal of maximizing profits for investors. Yet, debt financing is more concentrated within the U.K., leading towards more conducive lender monitoring. More concentrated creditors will lead to lower monitoring costs. Therefore the potential conflicts between suppliers of equity and debt will be less pronounced.
The first limitation of this study is regarding the definition of multinationality. No uniform definition of multinationality is established in literature. Within this research a firm is defined a MNC when it has more than 20 percent foreign sales. Another percentage as cut-off point or the use of foreign assets within the definition might result in somewhat altered findings.
Second, within this research no distinction is made between the varieties of creditors. Results of the study by Doukas and Pantzalis (2003), presented individual or institutional investor block shareholdings did, minimal yet significantly, influence the capital decisions of firms. The altered results can be explained by the block holders serving as a monitor on the firm. Due to a lack of data the influence of block holders was not included in this research. It is advised for further research to include this variable.
33 firm concerning the capital structure might shed more light on the decisions-making process and the determinants affecting this decision.
Subsequently, since the current espoused theories on the influence of multinationality on the capital structure of the firm do not hold for U.K. MNCs, it would be interesting to test whether the espoused theory holds for other countries to the U.S. The U.K. should, according to theory, show similarities in the degree of agency conflicts and given these similarities show similar results. According to Moerland (1995), the agency problem between shareholders and creditors should be smaller, or at least be more easily solvable, in network-oriented systems. The agency problems of debt should be more easily solvable because there is more proximity between the management and other stakeholders. With more proximity between the management and creditors, it becomes easier for creditors to monitor the actions of the firm, also in MNCs. I would expect results on other countries in the world testing the espoused theory to be even further off. These results would verify that the lower information asymmetry between shareholders and creditors does lead to fewer changes in debt ratios when increasing multinationality. The lower information asymmetry could be reached by the proximity, or in my case, by the new developments within corporate governance control and more prominence on transparency.
V.
References
Aggarwala, R., & Aung Kyaw, N. (2010). Capital structure, dividend policy, and multinationality: Theory versus empirical evidence. International Review of Financial Analysis , 19 (2), 140-150.
Akhtar, S. (2005). The determinants of capital structure for australian multinational and domestic corporations. Australian Journal of Management, 321 - 341.
34 Armstrong, C. S., Wayne, G. R., & Weber, J. P. (2010). The role of information and financial reporting in corporate governance and debt contracting. Journal of Accounting and Economics , 50 (2-3), 179 – 234. Burgman, T. A. (1996). An empirical examination of multinational corporate capital structure. Journal of International Business Studies , 27 (3), 553-570.
Brooks, C. (2008). Introductory econometrics for finance. (2nd edition). Cambridge, U.K.: Cambridge University Press.
Chiu, I. (2014). Reviving sharholder stewardship: critically examining the impact of corporate transparance reforms in the UK. Delaware Journal of Corporate Law , 38 (3), 983-1023.
Easterbrook, F. H. (1984). Two agency-cost explanations of dividends. American Economic Review , 74 (4), 650.
DeAngelo, H., & DeAngelo, L. (2006). The irrelevance of the MM dividend relevance theorem. Journal of financial economics , 79, 293-315.
DeAngelo, H., & Masulis, R. W. (1980). Optimal capital structure under corporate and personal taxation. Journal of Financial Economics , 8 (1), 3-27.
Doukas, J. A., & Pantzalis, C. (2003). Geographic diversification and agency costs of debt of multinational firms. Journal of Corporate Finance , 9 (1), 59–92.
Egger, P., Eggert, W., & Winner, H. (2010). Corporate taxation, debt financing and foreign-plant ownership. European Economic Review 54 , 54, 96 - 107.
Fama, E. F., & French, K. R. (2002). Testing trade-off and pecking order predections about dividends and debt. The Review of Financial Studies , 15 (1), 1-33.
Ferri, M., & Jones, W. (1979). Determinants of financial structure: a new methodological approach. The Journal of Finance , 631 - 644.
35 González , V. M., & González, F. (2012). Firm size and capital structure: evidence using dynamic panel data. Applied Economics, 44 (36), 4745– 4754.
Harris, M., & Raviv, A. (1990). Capital Structure and the Informational Role of Debt. The journal of Finance , 45, 321-349.
Hillier, D., Grinblat, M., & Titman, S. (2002). Financial markets and corporate strategy (European ed of 2nd revised edition). New York: McGraw-Hill.
Huizinga, H., Laeven, L., & Nicodeme, G. (2008). Capital structure and international debt shifting. Journal of Financial Economics , 88, 80 - 118. Jensen, M. C. (1986). Agency costs of free cash flow, corporate finance and takeovers. American Economic Review , 323 - 329.
Jensen, M. C., & Meckling, W. H. (1976). Theory of the firm: Managingarial behavior, agency costs and ownership structure. Journal of Financial Economics , 3 (4), 305-360.
Kohler, U., & Kreuter, F. (2012). Data analysis using stata. (3th edition). College Station: Stata Press.
Lang, L., & Stulz, R. M. (1994). Tobin's q, corporate diversification, and firm performance. Journal of Political Economy , 1248 - 1280.
Lee, K., & Kwok, C. C. (1988). Multinational corporations vs. domestic corporations: International environmental factors and determinants of capital structure. Journal of International Business Studies , 19 (2), 195-217.
Maçãs, P. J., & Serrasqueiro, Z. M. (2007). Capital Structure of Portuguese Service Industries: A Panel Data Analysis. Service Industries Journal , 27 (5), 549-562.
Miller, M. H. (1977). Debt and taxes. Journal of Finance , 32 (2), 261-275. Mittoo, U., & Zhang, Z. (2008). The capital structure of multinational corporations: Canadian versus U.S. evidence. 14 (5), 706-720.
36 Myers, S. C. (1984). The capital structure puzzle. Journal of Finance, 575 - 592.
Myers, S. C. (1977). Determinants of corporate borrowing. Journal of Financial Economics , 5 (2), 147–175.
Olugbode, M., El-Masry, A., & Pointon, J. (2014). Exchange rate and interest rate exposure of UK industries using first-order autoregressive exponential GARCH-in-mean approach. The Manchester School , 82 (4), 409 - 464.
Ramirez, A., & Kwok, C. (2010). Settling the debate on multinational capital structure using the CEPR measure. Journal of Multinational Financial Management , 20 (4/5), 251-271.
Rees, W., & Unni, S. (2005). Exchange rate exposure among European firms: evidence from France, Germany and the UK. Accounting and Finance , 45 (3), 479 – 497.
Tirole, J. (2006). The theory of corporate finance. Princeton: Princeton (DeAngelo & DeAngelo, 2006) (DeAngelo & DeAngelo, 2006) (DeAngelo & DeAngelo, 2006) (DeAngelo & DeAngelo, 2006)University Press.
Titman, S., & Wessels, R. (1988). The determinants of capital structure choice. The Journal of Finance , 1 - 19.
Verheugen, G. (2003). SME user guide explaining the new SME definition. European Commission, Enterprise and Industry. Brussels: EU Publication Office.
Warner, J. B. (1977). Bankruptcy costs: some evidence. Journal of Finance , 32 (2), 337-347.
37
Appendix A
Variable definition and predicted signs with leverage
Variables Definition Predicted Sign
Long-term debt ratio (LTD) Long-term debt / (long-term debt + short term debt & current portion of long-term debt + value of equity)
Total debt ratio (TD) Long-term debt + short term debt & current portion of term debt) / (term debt + short term debt & current portion of
long-term debt + value of equity)
Foreign sales percentage (FRS) Percentage of foreign sales over total sales
Multinational Dummy (MNCD) MNC = 1, if foreign sales ratio is higher than 20% and MNC = 0, if foreign sales ratio is lower than 20%, NA is interpreted as 0
Firm size (SIZE) Log of book value of total assets +
Growth (MTB) Market value of equity / book value of equity +
Non-debt tax shield (NDTS) Operating income before depreciation and amortization - interest expense - (taxes paid/ tax rate) / sales), where tax rate is assumed
30% for 2000-2007, 28% for 2008, 2009 and 2010, 26% for 2011 and 24% for 2012 -
Operating Risk (OPRISK) Standard deviation of (Cash Flow / Sales) over the past 5 years -
Profitability (PROF) Average of (Net Income / Sales) over the past 3 years -
38
Appendix B
Descriptive Statistics: Means and standard deviations of the variables included in the regression models, split in a multinational (MNCs) a domestic (DCs) sample based on the multinationality dummies of 10%, 20% and 50%. The variables total debt (TD) and long-term debt (LTD) are measured in percentages, they contain thirteen yearly observations ranging from 2000-2012. The variables multinationality (MNC10, MNC20, MNC50, FRS), market-to-book value (MTB) and dividend payout ratio (DIVPO) are measured in ratios. Size (SIZE) is measured by taking a logarithm. Non-debt tax shield (NDTS) is measured by the formula provided on appendix B. All the previous independent variables are lagged for one year, and therefore contain thirteen yearly observations ranging from 1999-2011. Operating Risk (OPRISK) is measured by a ratio comprising a prior 5 year standard deviation. The observations range from 1995-2007. Profitability (PROF) is measured by a ratio comprising a prior 3 year average. The observations range from 1998-2009.
Total sample; N = 3143 MNCs > 10% Foreign sales; N = 2329 DCs < 10% Foreign sales; N = 814 MNCs > 20% Foreign sales; N = 2123 DCs < 20% Foreign sales; N = 1020 MNCs > 50% Foreign sales; N = 1518 DCs < 50% Foreign sales; N = 1625
Mean Std. Dev. Mean Std. Dev. Mean Std. Dev. Mean Std. Dev. Mean Std. Dev. Mean Std. Dev. Mean Std. Dev.
39
Appendix C
Correlation matrix checking for correlation between the independent variables. Bold numbers denote a correlation between variables higher than 80%.
LTD (%) (%) TD (ratio) FRS MNC20 (ratio) SIZE (ln) (ratio) MTB (ratio) NDTS OPRISK (ratio) (ratio) PROF DIVPO (ratio)
