Capital Structure and International Debt Shifting
Author:
Mr. Luc Laeven
Search for other papers by Mr. Luc Laeven in
Current site
Google Scholar Close
,
Harry Huizinga https://isni.org/isni/0000000404811396 International Monetary Fund

Search for other papers by Harry Huizinga in
Current site
Google Scholar Close
, and
Gaetan Nicodeme https://isni.org/isni/0000000404811396 International Monetary Fund

Search for other papers by Gaetan Nicodeme in
Current site
Google Scholar Close

Contributor Notes

Authors’ E-Mail Addresses: H.P.Huizinga@uvt.nl, llaeven@imf.org, and Gaetan.Nicodeme@cec.eu.int
Publication Date:
01 Feb 2007
eISBN:
9781451866032
Language:
English
Keywords:
WP; capital structure; multinational firm
Download PDF (243.5 KB)

This paper presents a model of a multinational firm's optimal debt policy that incorporates international taxation factors. The model yields the prediction that a multinational firm's indebtedness in a country depends on a weighted average of national tax rates and differences between national and foreign tax rates. These differences matter because multinationals have an incentive to shift debt to high-tax countries. The predictions of the model are tested using a novel firm-level dataset for European multinationals and their subsidiaries, combined with newly collected data on the international tax treatment of dividend and interest streams. Our empirical results show that corporate debt policy indeed not only reflects domestic corporate tax rates but also differences in international tax systems. These findings contribute to our understanding of how corporate debt policy is set in an international context.

Abstract

This paper presents a model of a multinational firm's optimal debt policy that incorporates international taxation factors. The model yields the prediction that a multinational firm's indebtedness in a country depends on a weighted average of national tax rates and differences between national and foreign tax rates. These differences matter because multinationals have an incentive to shift debt to high-tax countries. The predictions of the model are tested using a novel firm-level dataset for European multinationals and their subsidiaries, combined with newly collected data on the international tax treatment of dividend and interest streams. Our empirical results show that corporate debt policy indeed not only reflects domestic corporate tax rates but also differences in international tax systems. These findings contribute to our understanding of how corporate debt policy is set in an international context.

I. Introduction

In most countries, interest expenses are deductible for corporate tax purposes, while dividends have to be paid out of net-of-tax corporate income. Most tax systems thus favor debt finance over equity finance, but to different degrees given the dispersion in top corporate tax rates. In determining their financial structure, purely domestic firms have to deal only with the domestic tax system. Multinational firms, however, face the more complicated choice of determining their overall indebtedness and the allocation of their debts to the parent firm and the subsidiaries across all countries in which the multinational operates. As a consequence, the financial structure of a multinational firm is expected to reflect the tax systems of all the countries where it operates.

In an international setting, the tax costs of debt and equity finance depend on the combined tax systems of the subsidiary and parent countries of the multinational firm. Dividends, as indicated, have to be paid out of the subsidiary’s income after subsidiary-country corporate tax and in addition may be subject to a nonresident dividend withholding tax in the subsidiary country. In the parent country, the dividend income may again be subject to corporate income tax. If so, double tax relief may or may not be provided for the previously paid corporate income and nonresident withholding tax. The tax costs of equity finance thus reflect tax rates as well as the double-tax-relief convention used by the parent country. This paper collects detailed information on all of these aspects of the international tax system for European multinationals.

A firm’s financial policies are affected by tax as well as nontax considerations. A nontax consideration is that indebtedness of the overall multinational firm should not be too high to keep the probability of costly bankruptcy low. In contrast, an advantage of debt finance is that it reduces the free cash flow within the firm and hence can act as a disciplining device for otherwise overspending managers. The disciplining properties of debt finance can explain generally positive debt levels at each of a multinational’s individual establishments (i.e., its parent company and its foreign subsidiaries). These various considerations give rise to an optimal overall capital structure for the overall multinational firm for nontax reasons.

This paper first presents a model of the optimal overall capital structure of the multinational firm reflecting tax and nontax factors. Generally, the tax advantages of debt finance lead the firm to choose a higher leverage than would be desirable for purely nontax reasons. At the same time, a change in tax policy optimally causes the firm to rebalance its capital structure in all the countries where it operates. Specifically, stronger incentives for debt finance in one country encourage debt finance in that country but at the same time discourage debt finance in other countries to keep the overall indebtedness of the multinational in check. The model yields the result that the optimal debt-to-assets ratio at any establishment of the multinational is positively related to the national tax rate and to differences between the national and foreign tax rates. The relevant tax rates in this regard are the effective tax rates that take into account any double taxation and double taxation relief. International tax rate differences matter, since they determine the incentives to shift debt internationally within a multinational firm.

Next, the paper presents evidence on the impact of taxation on firm indebtedness for a sample of 33 European countries over the 1994–2003 period using a unique firm-level database on the financial structure of domestic and multinational firms, including their parent companies and their subsidiaries. For stand-alone domestic firms, we estimate that a 10 percent increase in the overall tax rate (reflecting corporate income taxes and nonresident dividend withholding taxes) increases the ratio of liabilities to assets by 1.84 percent. For multinational firms, the leverage ratio is found to be more sensitive to taxation on account of international debt-shifting. As an example, we can consider a multinational with two equal-sized establishments in two separate countries. A 10 percent overall tax increase in one country is found to increase the leverage ratio in that country by 2.44 percent, while the leverage ratio in the other country decreases by 0.6 percent. Corporate debt policy appears to reflect local, source-level taxes rather than residence-level taxes levied on a multinational’s worldwide income, perhaps because these latter taxes can often be deferred. Similarly, debt policy appears to reflect corporate income taxation rather than bilateral nonresident dividend withholding. In practice, multinationals may be able to avoid bilateral withholding taxes through triangular arbitrage involving a conduit company in a third country.

Several authors consider the relationship between firm leverage and taxation with U.S. data. Among these, MacKie-Mason (1990) and Gordon and Lee (2001) identify a tax effect by exploiting the different effective taxation faced by previously loss-making firms and firms of different sizes, respectively. Graham (2000) calculates the value of the tax benefits of debt finance for the U.S. case. Studies that use cross-country data have the advantage that they allow for international variation in tax rates. Examples are Rajan and Zingales (1995) and Booth, Aivazian, Demirgüc-Kunt, and Maksimovic (2001). The latter set of authors finds a weak effect on leverage for a tax variable that measures the tax shield of debt finance. Next, there is a set of papers that consider the debt finance of multinationals with either parent companies or subsidiaries in the United States. Specifically, Hines and Hubbard (1990), Collins and Shackelford (1992), Froot and Hines (1992), and Grubert (1998) provide evidence that U.S. multinational financial structure and the pattern of intrafirm interest and other income flows are consistent with tax minimization objectives. Newberry and Dhaliwal (2001) find that the debt issuance location of U.S. multinationals is affected by these firms’ jurisdiction-specific tax-loss carry-forwards and binding foreign tax credit limitations on the value of debt tax shields. Desai, Foley, and Hines (2004) find that both the internal and external financing of outward U.S. Foreign Direct Investment (FDI) is sensitive to foreign tax rates. Mills and Newberry (2004) analogously find that non-U.S. multinationals from countries with relatively low tax rates use relatively intensive debt finance of their foreign-controlled corporations in the United States.

Jog and Tang (2001) consider the leverage of firms in Canada that may or may not be part of U.S.-based or Canada-based multinationals. The debt-to-assets ratios of Canadian corporations without foreign affiliates are found to be more sensitive to Canadian tax rates than the debt-to-assets ratios of U.S.-controlled corporations located in Canada. Using data for member countries of the European Union, Moore and Ruane (2005) examine the leverage of 8,500 foreign subsidiaries. They find that leverage ratios of these subsidiaries are sensitive to the local corporate tax rate, unless the parent country operates a foreign tax credit system. This paper nests the approaches of the latter two papers by considering how both multinational firm structure and the international tax system affect leverage in Europe. Hence, we take into account whether a firm is a parent or a subsidiary of a multinational or a domestic firm. At the same time, we account for the tax systems of all the countries where the multinational operates. Thus, unlike previous research, our modeling and our empirical work take a fully multilateral approach. The main contribution of our paper is to explore in an international context the possibility that multinationals set the capital structure of individual subsidiaries by taking into account the tax rate faced by all other subsidiaries of the firm. Our finding that subsidiary leverage within a multinational firm responds to bilateral tax rate differences vis-à-vis both the parent firm and other foreign subsidiaries provides direct support for this multilateral approach.

In the remainder, Section II describes the international tax treatment of the debt and equity finance of multinational firms. Section III presents the model. Section IV discusses the company-level data. Section V presents the empirical results, and Section VI offers conclusions.

II. The International Tax System

This section describes the main features of the corporate income tax system applicable to a multinational firm with subsidiaries in one or more foreign countries.2 To fix ideas, let us consider a multinational firm that operates a foreign subsidiary in country i and has the parent firm in country p. The deductibility of interest from corporate income implies that there is no corporate taxation of interest to external debt holders. Dividends paid by the subsidiary to the parent firm, in contrast, are generally subject to corporate taxation in at least one country.

The subsidiary’s income in county i is first subject to the corporate income tax ti in this country. The first column of Table 1, indicates the statutory corporate tax rate on corporate profit for a sample of 33 European countries in 2003. These tax rates include regional and local taxes as well as specific surcharges. Germany has the highest tax rate at 39.6 percent, while Ireland is at the bottom with a tax rate of 12.5 percent. This and all other tax system information in this paper has been collected from the International Bureau of Fiscal Documentation and various websites of national ministries of finance. We have collected the tax system data reported in Tables 1–5 for all years in our sample period but only report the figures for the year 2003 due to space constraints.3

Table 1.

Corporate Taxation and Double-Tax-Relief Systems for Dividend Received in Selected European Countries in 200315

article image
article image
Note: Dividends are assumed to be paid by fully owned subsidiaries. Source: International Bureau of Fiscal Documentation.
Table 2.

Bilateral Withholding Tax on Dividend Payments Between Fully Owned Foreign Subsidiary and Parent on January 1, 2003

article image
Notes: (a) The Parent-Subsidiary directive is binding between EU Member States and provides exemption from withholding tax when holding is at least 25%. (b) Ireland: companies located in EU or treaty countries are exempt from withholding tax provided that they are not under the control of persons not resident in such countries, (c) Estonia: general exemption from withholding tax if holding in foreign company is at least 25%. (d) Italy: if the recipient can prove a tax is paid in its country on the dividend, the Italian authorities can provide a refund equal to the tax claimed limited to 4/9 of the Italian withholding tax. (e) Lithuania: general exemption from withholding tax if holding in foreign company is at least 25%. (f) Luxembourg: exemption from withholding tax for EU and treaty partners if holding in foreign company is at least 10%. (g) Sweden: no withholding tax if holding is 25% and there is normal corporate taxation in the foreign country and if the shares are held for business-related reasons. Source: International Bureau of Fiscal Documentation.
Table 3.

Existence of a Bilateral Tax Treaty on January 1, 2003

article image
Source: International Bureau of Fiscal Documentation, various ministries’ websites.
Table 4.

Bilateral Withholding Tax on Interest Payments Between Fully Owned Foreign Subsidiary and Parent on January 1, 2003

article image
Notes: (a) Ireland: interest paid to a 75% nonresident parent is deemed to be a dividend, (b) Spain: interest is generally exempt from tax in Spain provided the direct beneficiary is a resident in another EU Member State, (c) Estonia: 0% if rate in recipient country is not lower man 2/3 of the Estonia tax rate on interest of 26%. (d) Switzerland: no withholding tax on ordinary loans, 35% on bonds and deposits. (e)Turkey: exemption for government bonds and debentures, as well as loans obtained from foreign companies and institutions, (f) Italy: zero withholding tax with treaty countries for public bonds, private bonds, and deposits; 27% with non-treaty countries for deposits and private bonds of maturity of less than 18 months; 12.5% otherwise. Source: International Bureau of Fiscal Documentation.
Table 5.

Double-Tax-Relief Systems for Interest Payments Received in Selected European Countries in 2003

article image
Source: International Bureau of Fiscal Documentation

The subsidiary pays out its after-tax corporate income as a dividend to the parent company. The subsidiary country may levy on nonresidents a withholding tax wie on this outgoing dividend income. Bilateral dividend withholding taxes in Europe for 2003 are presented in Table 2. These rates are zero in the majority of cases. Specifically, they are zero among longstanding EU member states on account of the parent-subsidiary directive. New EU member states such as the Czech Republic, Hungary, Poland, and Slovenia still maintain nonzero dividend withholding taxes vis-à-vis considerable numbers of European countries. Non-EU member states such as Bulgaria, Romania, Russia, and Turkey similarly maintain nonzero dividend withholding taxes in a considerable number of cases. The combined corporate and withholding tax rate in the subsidiary country is seen to be 1−(1−ti)(1−wie) or  ti+wie− ti wie.

The parent country subsequently may or may not use its right to tax the income generated abroad. In case the parent country operates a territorial or source-based tax system, it effectively exempts foreign-source income from taxation. The effective marginal tax on income reported in country i, denoted τi, in this instance, equals combined corporate and withholding tax ti+wie −  ti wie in country i. 4Alternatively, the parent country operates a worldwide or residence-based tax system. In this instance, the parent country subjects income reported in country i to taxation, but it generally provides a foreign tax credit for taxes already paid in country i to reduce the potential for double taxation. The OECD model treaty, which summarizes recommended practice, in fact gives countries an option between an exemption and a foreign tax credit as the only two ways to relieve double taxation (see OECD, 1997). The foreign tax credit reduces domestic taxes on foreign source income one-for-one with the taxes already paid abroad. The foreign tax credit can be indirect in the sense that it applies to both the dividend withholding tax and the underlying subsidiary country corporate income tax. Alternatively, the foreign tax credit is direct and applies only to the withholding tax. In either case, foreign tax credits in practice are limited to prevent the domestic tax liability on foreign source income from becoming negative.

In the indirect credit regime, the multinational will effectively pay no additional tax in the parent country, if the parent tax rate tp is less than ti+wie− tiwie. The multinational then has unused foreign tax credits and is said to be in an excess credit position. Alternatively, tp exceeds ti+wie− tiwie. In that instance, the firm pays tax in the parent country at a rate equal to the difference between tp and ti+wie− tiwie. The effective, combined tax rate on the dividend income, τi;, then equals the parent country tax rate, tp. To summarize, with the indirect credit system the effective rate on income generated in country i, τi, is given by max [tp,ti+wie− tiwie]. In case of a direct foreign tax credit, the multinational analogously pays no additional tax in the parent country, if the parent tax rate tp is less than wie. In the more common case where tp exceeds wie, the firm instead pays tax in the parent country at a rate equal to (1−ti)(tp−wie). The effective, two-country tax rate, τi;, with the direct credit system is now given by ti+(1−ti)max⁡[tp,wie]. A few countries with worldwide taxation do not provide foreign tax credits, but instead allow foreign taxes to be deducted from the multinational’s taxable income. Under this deduction method, foreign taxes are essentially seen as a tax-deductible cost of seeing business at par with other business costs. In the scenario, the effective rate of taxation on dividends, τi, is given by 1−(1−ti)(1−wie)(1−tp).

The second and third columns of Table 1 provide information on the double taxation rules applied to incoming dividend. As reflected in the table, several countries are seen to discriminate between international tax treaty partners and nontreaty countries. Finland and Spain, for instance, exempt dividend income from treaty partners, while they provide a direct and indirect foreign tax credit in case of nontreaty counties, respectively. Note that signing a tax treaty makes the granted double tax relief more generous in these instances. The tendency to discriminate double tax relief on the basis of the existence of a tax treaty makes it necessary to know whether a bilateral tax treaty is indeed effective. Table 3 indicates with a binary variable whether any two countries had a taxtreaty in force by 2003.5 Across the categories of treaty and nontreaty countries, the exemption system is seen to be the most common method of double tax relief, followed by foreign tax credits. At the same time, indirect foreign tax credits regimes are somewhat more common than direct foreign tax credits. As an exceptional case, the Czech Republic is seen to apply the deduction method to foreign dividends from non-treaty countries, while Russia and the Slovak Republic provide no double tax relief at all to such income.

In practice, multinationals use equity as well as internal debt to provide own resources to their foreign subsidiaries. Thus, leverage is likely to be affected by the taxation of dividends, as considered so far, and by the taxation of interest on internal debt. To reflect this in our empirical work, we use a variable φi to denote the effective tax rate on cross-border dividends, i.e. τi, minus an analogous effective rate of tax on interest. Interest on internal debt is generally deductible from taxable corporate income in the subsidiary country. Such interest income thus escapes corporate income tax in the subsidiary country. As in the case of dividends, cross-border interest flows within the multinational firm may generally be subject to a nonresident withholding tax in the subsidiary country. Let wid denote the bilateral nonresident interest withholding tax. As seen in Table 4, these tax rates are mostly zero on a bilateral basis for the countries in our sample, even if Belgium, Estonia, Latvia, Portugal, and Romania continue to impose positive interest withholding taxes vis-à-vis almost all countries in our sample. As applied to internal interest flows, the parent country has three main options regarding double tax relief: (i) an exemption, (ii) a foreign tax credit, or (iii) a deduction. Table 5 provides information on the double taxation rules applicable to incoming interest from treaty and nontreaty signatory countries, respectively. The signing of a tax treaty, if anything, makes the double tax relief in case of interest flows more generous. Foreign-source interest flows are seen to benefit from a foreign tax credit in most countries, particularly in the case of interest payments originating from treaty partners. Clearly, the taxation of dividend income relative to interest income, φi, depends on the possibly different tax relief granted for dividends and for interest. Expressions for φi in the various possible combinations of double tax relief granted for dividend and interest income are provided in Table 6.

Table 6.

Expressions for φi or the Relative Taxation of Equity and Internal Debt of Subsidiaries

article image

III. The Model

The model considers a multinational firm that generally operates in n countries. The multinational is domiciled in country p, while it has foreign subsidiaries in one or more countries i with assets A i. The subsidiary is financed with debt Li, which for now we take to be external debt, and equity Ei. Hence, the balance sheet identity of a subsidiary implies Ai, =Li +Ei. The parent firm fully owns each subsidiary’s equity Ei. In addition, the parent firm owns “outside” assets Ap. The parent firm in turn can be financed through either debt Lp or equity Ep. Thus, the balance sheet identity for the parent can be stated as Ap+∑i≠pnEi=Lp+Ep. Let λi be the ratio of debt-to-assets or each establishment o the multinational—i.e., λi=LiAi.

Analogously, let λf be the debt-to-assets ratio for the entire firm, i.e. λf=∑i=1nLi∑i=1nAi.

Alternatively, we can write λf as the asset-weighted average of the establishment-specific debt ratios λi as ∑i=1nλiρi, where ρi=Ai∑i=1nAi are the assets o establishment/as a share o the firm s total assets. Throughout, we will assume that the assets Ai of subsidiary i and the parent firm’s outside assets Ap are given.6

In deciding its financial structure, the multinational firm takes taxation as well as nontax factors into account.7 To start with the latter, the multinational recognizes that higher leverage increases the chance of bankruptcy. We will assume that the parent firm provides credit guarantees for the debts of all its subsidiaries. This implies that the chance of bankruptcy of the overall multinational firm depends on the firm-wide leverage ratio λf. Specifically, we will assume that expected bankruptcy costs, Cf, of the firm are quadratic in the overall leverage ratio λf and proportional to the firm s overall outside assets as follows8

C f = γ 2 ( λ f ) 2 ( ∑ i = 1 n A i ) . ( 1 )

Next, it is recognized that leverage may bring benefits in that it disciplines local managers and aligns their incentives more closely to those of the firm. High leverage at a subsidiary may, for instance, serve to prevent local managers from overspending on perks for themselves to prevent de jure bankruptcy of the subsidiary. On the other hand, high leverage may have the disadvantage of making local managers too risk-averse, to the point where they do not make appropriate local investment decisions. In either case, the incentive effects of leverage are assumed to stem from the local leverage ratio λi for establishment i. 9 On the basis of these incentive considerations alone, let λ* be the optimal leverage ratio at each of the multinational’s establishments. Deviations of the leverage ratio at any establishment from the level λ* are assumed to imply incentive-related costs to the firm. These costs are assumed to be quadratic in λi,and now they are proportional to the outside assets Ai at establishment i as follows:

C i = μ 2 ( λ i − λ * ) 2 A i − μ 2 λ * 2 A i i = 1 , … ,   n ( 2 )

Note that these cost functions are scaled to equal zero if the debt ratios λi are zero, which implies that Ci can be of either sign. Next, let Vl and Vu be the values of the leveraged and completely unleveraged multinational firm, respectively. The two are different on account of the tax benefits of debt finance and of the (net) nontax costs associated with debt finance. Specifically, VL and Vu are related as follows

V L = V u + ∑ i = 1 n τ i L i − C f − ∑ i = 1 n C i , ( 3 )

where τi. again is the rate of taxation of dividend income relative to interest income in locale i taking into account the overall international tax system.

The multinational firm’s objective is to maximize its overall firm value VL in the leveraged state. Its instruments are the debt levels Li at each establishment.10 The first-order conditions w.r.t. Li—written in terms of leverage ratios—are given by

τ i − γ λ f − μ ( λ i − λ * ) = 0 i = 1 , … ,   n   . ( 4 )

The first-order conditions jointly allow us to solve for the optimal value of λi as follows:

λ i = β 0 λ * + β 1 τ i + β 2 [ ∑ j ≠ i n ( τ i − τ j ) ρ j ] i = 1 , … , n , ( 5 )

where  β0=(μγ+μ), β1=1γ+μ,and β2=(γμ(γ+μ)).

In expression (5), the term β0λ* is the optimal leverage ratio at all establishments on the basis of all nontax considerations, or equivalently if all the τ ’s are equal to zero. The term β0λ* can be seen to balance the expected costs of bankruptcy (with a value of λ above zero) against the costs of deviating from the optimal value of the leverage ratio λ* on the basis of incentive considerations. Expression (5) further contains two tax-related terms. First, the term β1τi reflects the impact of taxation on the optimal leverage ratio that would obtain for a purely domestic firm located in country i. For this reason, this term is dubbed the “domestic” effect of taxation on leverage. Second, the term β2[∑j≠in(τi−τj)ρj] reflects the impact of international tax rate differences on the optimal leverage in country/on account of international debt-shifting. Interestingly, this term weight the international tax differences τi. - τj. by the asset shares ρj. This reflects that the cost function Ci implies that it is relatively painless to shift (absolute) debt into or out of country j, if the assets in this country are relatively large. This second effect of taxation on leverage in country i is named the “international” or “debt-shifting” effect. Note that leverage λi in country i is negatively related to τj on account of the debt-shifting effect.

The theoretical equation (5) gives rise to the following regression equation

λ i = α i + β 1 τ i + β 2 [ ∑ j ≠ i n ( τ i − τ j ) ρ j ] + ε i i = 1 , … , n ( 6 )

where αi is a country-specific fixed effect and εi is an error term. In the benchmark case, the sample will consist of observations for all subsidiaries to the exclusion of parent firms.11 In practice, a range of firm-level and country-level control variables is included in the estimation.

IV. The Data

The data on multinational firms are taken from the Amadeus database compiled by Bureau Van Dijk.12 This database provides accounting data on private and publicly owned European firms as well as on their ownership relationships. These ownership data allow us to match European firms with their domestic subsidiaries and subsidiaries located in other European firms. A firm is defined to be a subsidiary if at least 50 percent of the shares are owned by another single firm. A multinational firm has at least one foreign subsidiary. Multinational firms tend to provide consolidated and unconsolidated accounting statements. Consolidated statements reflect the activities within the parent companies themselves and of all domestic and foreign subsidiaries. Nonconsolidated statements in contrast reflect the activities directly within the parent firm and in each of its subsidiaries. The data we use on parent firms and subsidiaries are based on nonconsolidated statements.

Information on the number of parent companies and subsidiaries—domestic and foreign in our data set—is provided in Panel A of Table 7. The total number of parent companies is 5,791, while the total number, of subsidiaries is 13,307. We have up to 10 years of data for each parent company and subsidiary, for a total of 38,736 parent-year observations and 90,599 subsidiary-year observations. Note that Amadeus only provides information on subsidiaries located in one of the European countries listed in the table.13 France, Spain, and the United Kingdom each are home to at least 4,000 parent companies in the data set. Each subsidiary has a home country (i.e., the country of its parent company) and a host country where the subsidiary is located (therefore, for domestic subsidiaries, home and host countries are the same). For each country, the table lists the number of subsidiaries by home country and by host country. The table reveals that, for instance, Germany and the Netherlands are the home country to relatively many subsidiaries. Hence, there are relatively many subsidiaries with a parent firm in one of these countries. Croatia, the Czech Republic, and Romania instead are host countries to relatively many subsidiaries.

Table 7.

Descriptive Statistics for Subsidiaries of European Multinationals

Panel A lists the number of parent companies and subsidiaries (by home and host country) in the sample. Intermediary companies, that are both parent firms and subsidiaries, are counted as subsidiaries only. Domestic subsidiaries are counted as subsidiaries by home country only. Panel B presents the sample averages of financial leverage and tax variables. Panel C presents the summary statistics for the financial leverage, tax, and other variables for subsidiaries only. Financial leverage is the ratio of total liabilities to total assets. Adjusted financial leverage is the ratio of total liabilities minus accounts payables minus cash to total assets minus accounts payables minus cash. Effective marginal tax rate (τ) is the statutory tax rate on dividend income generated in the subsidiary country, taking withholding taxes and the tax system for foreign source income into account. Tax incentive to shift debt is the sum of international tax differences between subsidiary countries weighted by subsidiary asset shares, taking withholding taxes and the international tax system into account. Tangibility is the ratio of subsidiary fixed assets to subsidiary total assets. Log of sales is the logarithm of subsidiary sales. Profitability is the ratio of subsidiary earnings before interest, taxes, depreciation and amortization to subsidiary total assets. Creditor rights is the annual index of credit rights in the country from Djankov et al. (2005). Political risk is the annual (December) index of political risk from the International Country Risk Guide. We inverted the scale from 0-100 with higher scores indicating greater risk. Inflation is the annual percentage change in the CPI of the subsidiary’s host country from the World Development Indicators. Growth opportunities is the median of the annual growth rate of subsidiary sales in a subsidiary’s country and industry. Sample consists of parent companies and subsidiaries of European firms in Amadeus. We have up to 10 years of data for each parent company and subsidiary. The total number of parent-year observations is 38,736, and the total number of subsidiary-year observations is 90,599.

article image
article image
article image
article image
article image

Panel B of Table 7 provides information on financial leverage and applicable tax rates. First, financial leverage is defined as the ratio of total liabilities to total assets (see the Appendix for variable definitions and data sources). Adjusted financial leverage, instead, is the ratio of, in the numerator, total liabilities minus accounts payable minus cash to, in the denominator, total assets minus accounts payable minus cash. These adjustments reflect that accounts payable are liabilities that reflect current operations rather than efforts to optimize the firm’s capital structure. Similarly, the subtraction of cash reflects that cash may be on hand to pay off existing debts. In Panel B of Table 7, we see that the average parent company financial leverage of 0.62 indeed exceeds the average adjusted financial leverage of 0.49. Average financial leverage ranges from 0.36 for Russia and Slovenia to 0.80 for Romania. Interestingly, subsidiaries by host country have average financial leverage and adjusted financial leverage of 0.62 and 0.49, respectively—exactly equal to the averages for parent firms. Hence, there is no tendency for subsidiaries to be either more or less leveraged than parent firms. Next, the effective tax rate for subsidiaries by host country is seen to be highest for Germany at 0.49, and lowest for Estonia at 0.14. As discussed before, the effective tax reflects the taxation of dividends in the host country itself as well as the tax treatment of this income in any foreign parent country. The tax incentive to shift debt for subsidiaries by host country is the asset-weighted difference of the effective tax rate in the host country and the effective tax rates applicable to other establishments of the same multinational firm. A positive value of this variable indicates that multinationals on average have an incentive to shift debt out of a particular host country. By this measure, subsidiaries hosted in Iceland and Germany have the largest incentive to attract debt, while subsidiaries located in Estonia and Hungary have the largest incentive to shift debt away.

Panel C of Table 7 provides summary statistics of our leverage and tax variables as well as of control variables included in the subsequent estimation. The control variables are several firm-level variables derived from the firm’s balance sheet or income statement as well as some country variables. Among the firm-level variables, tangibility is defined as the ratio of fixed assets to total assets. This variable captures that it may be relatively easy to borrow against fixed assets. In addition, depreciable assets may act as a nondebt tax shield and is therefore a substitute for debt in taxable profit minimization strategies. Next, log of sales is the logarithm of sales. This is a scaling variable to reflect that larger firms may have easier access to credit. Next, profitability is the ratio of earnings before interest, taxes, depreciation, and amortization to total assets. Profitability may affect leverage in several ways. Profitable firms may be perceived to be relatively riskless, which would facilitate their access to credit. On the other hand, profitable firms may use their profits to pay down their debts or, alternatively, to finance investments through retained earnings. In either way, high profitability may lead to low leverage. Among the country variables, creditor rights is an annual index of creditor rights in a country. Well-protected creditor rights are expected to encourage leverage. Next, political risk is an annual index of political risks. High political risks may encourage borrowing from local creditors, since this is a way to reduce a multinational’s value at risk in a country. Inflation is the annual percentage change in the consumer price index. High inflation increases the value of the tax deductibility of interest to the extent that inflation leads to higher nominal interest rates. At the same time, an inflationary environment may also lead to a higher risk premium as part of the nominal interest rate, which discourages debt finance. Finally, the growth opportunities variable measures the median annual growth rate of sales in an industry in a particular country. Growth opportunities signal future profitability and possibly an ability to borrow.

V. Empirical Results

Table 8 presents our basic regressions. The sample consists of all European subsidiaries in Amadeus. For each observation, an effective tax rate and a debt-shifting incentive variable can be constructed. All regressions in the table provide for parent, industry, and year fixed effects. Regression (1) includes the effective tax rate to the exclusion of the international debt-shifting incentive variable. The pertinent coefficient is estimated to be 0.259 and statistically significant.14 The tangibility variable has a negative coefficient, which suggests that debt and tangible, depreciable assets are substitutes. The log of sales enters positively. Profitability, in turn, obtains a negative coefficient, which suggests that the overall effect of higher profitability is to reduce leverage. Note that the 71,355 observations in the sample are associated with a total of 5,566 parent companies. Yearly observations of the same subsidiary are counted separately.

Table 8.

The Impact of Domestic and International Taxes on the Financial Leverage of Subsidiaries of Multinational Firms

The dependent variable in columns (1) to (3) is the ratio of subsidiary total liabilities to subsidiary total assets. The dependent variable in column (4) is the ratio of total liabilities minus accounts payables minus cash to total assets minus accounts payables minus cash. Effective marginal tax rate (τ) is the statutory tax rate on dividend income generated in the subsidiary country, taking withholding taxes and the tax system for foreign source income into account. Tax incentive to shift debt is the sum of international tax differences between subsidiary countries weighted by subsidiary asset shares, taking withholding taxes and the international tax system into account. Tangibility is the ratio of subsidiary fixed assets to subsidiary total assets. Log of sales is the logarithm of subsidiary sales. Profitability is the ratio of subsidiary earnings before interest, taxes, depreciation and amortization to subsidiary total assets. Creditor rights is the annual index of credit rights in the country from Djankov et al. (2005). Political risk is the annual (December) index of political risk from International Country Risk Guide. We inverted the scale from 0-100 with higher scores indicating greater risk. Inflation is the annual percentage change in CPI of the subsidiary’s host country from World Development Indicators. Growth opportunities is the median of the annual growth rate of subsidiary sales in a subsidiary’s country and industry. Sample consists of subsidiaries of European companies in Amadeus. All regressions are estimated using OLS and include parent, industry, and year fixed effects. We report White (1980)’s heteroskedasticity-consistent standard errors between brackets. * denotes significance at 10%; ** significance at 5%; and *** significance at 1%.

article image
article image

Regression (2) includes the debt-shifting incentive variable. The estimated coefficient for this variable is positive and statistically significant, which confirms that leverage at any subsidiary of a multinational reflects the overall international tax system faced by the multinational. Next, regression (3) includes a set of additional, country-level controls. The creditor rights variable enters the regression positively and significantly. The political risk variable is equally positively and significantly related to leverage. As indicated, this may reflect that political risks lead a multinational to increase local borrowing in order to reduce its own capital at risk. Next, inflation has a negative and significant impact on leverage. This could reflect that in an inflationary environment there is more uncertainty about the ex post real interest rate to be paid on nominal debt denominated in the local currency. Finally, the growth opportunities variable enters the regression positively and significantly. High growth at the industry and country level may facilitate debt finance of the affected subsidiaries. Finally, in regression (4) adjusted financial leverage is taken to be the dependent variable. In other respects, regression (4) mimics regression (3). The effective tax rate and debt-shifting incentive variables continue to obtain positive and significant coefficients, albeit somewhat larger than before. Hence the adjustment of financial leverage for accounts payable and cash has little impact on the estimated impact of taxation on leverage. In regression (4), however, the tangibility variable enters with a positive coefficient to suggest that debt and tangible assets are complements (since firms can relatively easily borrow against tangible assets), while the political risk and growth opportunities variables cease to obtain significant coefficients.

The estimated coefficients of regression (3) can serve to evaluate the size of the effect of taxation on leverage. First, the estimated size of β1, 0.18, indicates the full effect of domestic taxation on the leverage of firms. Specifically, the “domestic” effect of an increase in the effective tax rate by 0.06 (or one standard deviation) on leverage is 1.1 percent. Next, the estimated size of β2, 0.12, captures the “international debt-shifting” effect of taxation on leverage. As an example, we can take a hypothetical multinational firm that has a single foreign subsidiary with assets of equal size to those of the parent firm. In this instance, an increase of the effective tax rate by 0.06 in the subsidiary country has a positive “international” effect on leverage in the subsidiary country of 0.4 percent. The total effect of an increase of the effective tax rate by 0.06 on subsidiary leverage is now calculated to be 1.5 percent. In contrast, an increase in the effective tax in the parent country of 0.06 has a negative “international” effect on leverage in the subsidiary of -1.5 percent.

Next, Table 9 presents some robustness checks, taking regression (3) in Table 8 as a starting point. In regression (1), we correct standard errors for clustering across country-industry observations. The estimated coefficients for the two tax variables are virtually unchanged from those of the benchmark regression. Regression (2) in turn limits the sample to subsidiaries in the manufacturing sector. In this regression, the estimated sizes of β1 and β2 are somewhat smaller, and much larger, respectively. The relatively large size of β2 in regression (2) may reflect that manufacturing firms are relatively transparent. Hence, for these firms it may be relatively easy to borrow in one country against the assets located in other countries to explain that leverage in one country is relatively sensitive to international tax rate differences. Next, regression (3) limits the sample to foreign subsidiaries. This reduces the sample size to 23,296 subsidiaries rather than 49,248 in regression (3) in Table 8. Relative to the benchmark regression, the value of β2 is very similar in magnitude at 0.138. In regression (4) we restrict the sample to subsidiaries of multinationals, i.e. of firms that have at least one foreign subsidiary. The estimated size of β1 and β2 are very similar to the benchmark results. In regression (5) we exclude loss-making subsidiaries by dropping firms that have negative earnings before interest, taxes, depreciation and amortization. The reason for excluding loss-making firms is that these firms may be close to financial distress, which could alter their debt policy. The results are not qualitatively affected. In regression (6), we exclude Eastern European countries from the sample, as the coverage of subsidiaries in Eastern Europe is quite poor. This reduces the sample size to 48,444 subsidiaries, but does not alter our main results. Financial leverage can also be affected by firm-specific risk. In particular, riskier firms tend to be higher leveraged. In regression (7), we use the standard deviation of the firm’s ratio of earnings before interest, taxes, depreciation, and amortization (EBITDA) to total assets over the period 1994–2003 as a proxy for the riskiness of the firm. Consistent with capital structure theory, we find that financial leverage is positively correlated with risk. Controlling for risk, however, does not much alter the effect of our tax variables on leverage. We continue to find positive and statistically significant coefficients for β1 and β2, of about equal size as before. In regression (8), we control for financial development using the ratio of private credit to GDP rather than the index of creditor rights. While private credit to GDP does not enter significantly, unlike creditor rights did in previous regressions, our main results are not affected.

Table 9.

The Impact of Taxes on the Financial Leverage of Subsidiaries of Multinational Firms: Robustness Checks

The dependent variable is the ratio of subsidiary total liabilities to subsidiary total assets. Effective marginal tax rate (τ) is the statutory tax rate on dividend income generated in the subsidiary country, taking withholding taxes and the tax system for foreign source income into account. Tax incentive to shift debt is the sum of international tax differences between subsidiary countries weighted by subsidiary asset shares, taking withholding taxes and the international tax system into account. In column (1), we correct standard errors correct for clustering across country-industry observations. In column (2), we only include subsidiaries operating in the manufacturing industries. In column (3), we only include the subset of foreign subsidiaries. In column (4), we only include subsidiaries of multinational companies. In column (5), we exclude subsidiaries with negative earnings before interest, taxes, depreciation and amortization. In column (6), we exclude Eastern European countries from the sample. In column (7), we include the standard deviation of the firm’s EBITDA over the period 1994-2003 as a measure of firm-specific risk. In column (8), we control for financial development using the ratio of private credit to GDP rather than the index of creditor rights. Tangibility is the ratio of subsidiary fixed assets to subsidiary total assets. Log of sales is the logarithm of subsidiary sales. Profitability is the ratio of subsidiary earnings before interest, taxes, depreciation and amortization to subsidiary total assets. Creditor rights is the annual index of credit rights in the country from Djankov et al. (2005). Political risk is the annual (December) index of political risk from the International Country Risk Guide. We inverted the scale from 0-100 with higher scores indicating greater risk. Inflation is the annual percentage change in the CPI of the subsidiary’s host country from the World Development Indicators. Growth opportunities is the median of the annual growth rate of subsidiary sales in a subsidiary’s country and industry. Volatility of profits is the standard deviation of the firm’s ratio of earnings before interest, taxes, depreciation and amortization (EBITDA) to total assets over the period 1994-2003. Sample consists of subsidiaries of European companies in Amadeus. All regressions are estimated using OLS and include parent, industry, and year fixed effects. We report White (1980)’s heteroskedasticity-consistent standard errors between brackets. * denotes significance at 10%; ** significance at 5%; and *** significance at 1%.

article image
article image

Finally, Table 10 reports several robustness checks where we alter the taxation variables. In regression (1) of Table 10, we control for the relative taxation of equity and internal debt of subsidiaries, or φi. We construct φi using information on corporate tax rates in the parent and subsidiary countries, withholding taxes on dividend and interest payments in the subsidiary country, and double tax relief conventions applied by the parent country to incoming dividend and interest payments. We find that φi does not enter significantly, and that our main results are not affected after controlling for the relative taxation of equity and internal debt of subsidiaries.

Table 10.

The Impact of Taxes on the Financial Leverage of Subsidiaries of Multinational Firms: Additional Robustness Checks

The dependent variable is the ratio of subsidiary total liabilities to subsidiary total assets. Effective marginal tax rate (τ) is the statutory tax rate on dividend income generated in the subsidiary country, taking withholding taxes and the tax system for foreign source income into account. Tax incentive to shift debt is the sum of international tax differences between subsidiary countries weighted by subsidiary asset shares, taking withholding taxes and the international tax system into account. In column (1), we control for the relative taxation of equity and internal debt of subsidiaries (φ). In column (2), we include “intermediate” companies, i.e., subsidiaries that are also parent companies of other subsidiaries, in the sample. In column (3), we assess whether there is a differential effect of our tax variables on leverage for intermediate companies and pure subsidiaries. In column (4), we split the tax incentive to shift debt variable in one component that captures the incentive to shift debt to the parent country and another component that captures the incentive to shift debt to subsidiaries in other countries than the host and parent country. In column (5) we split the effective marginal tax rate variable in one component that captures the taxation in the source country (i.e., by setting the tax rate of the parent to zero) and one component that captures the taxation in the resident country. In column (6), we split both tax variables in one component that does not depend on withholding taxes and another component that depends on withholding taxes. In column (7), we use the population of the subsidiary country and the population of the parent country as instruments for the effective marginal tax rate. Taking the tax system into account and assuming that withholding taxes are zero. Tangibility is the ratio of subsidiary fixed assets to subsidiary total assets. Log of sales is the logarithm of subsidiary sales. Profitability is the ratio of subsidiary earnings before interest, taxes, depreciation and amortization to subsidiary total assets. Creditor rights is the annual index of credit rights in the country from Djankov et al. (2005). Political risk is the annual (December) index of political risk from the International Country Risk Guide. We inverted the scale from 0-100 with higher scores indicating greater risk. Inflation is the annual percentage change in the CPI of the subsidiary’s host country from the World Development Indicators. Growth opportunities is the median of the annual growth rate of subsidiary sales in a subsidiary’s country and industry. Sample consists of subsidiaries of European companies in Amadeus. All regressions are estimated using OLS and include parent, industry, and year fixed effects. We report White (1980)’s heteroskedasticity-consistent standard errors between brackets. * denotes significance at 10%; ** significance at 5%; and *** significance at 1%

article image
article image
article image

In regression (2), we include “intermediate” companies, i.e., subsidiaries that are also parent companies of other subsidiaries, in the sample. This increases the sample from 49,248 to 57,409 observations. Our main results on the effect of taxation on financial leverage are unaffected, but we no longer find an effect of political risk on financial leverage.

In regression (3), we assess whether there is a differential effect of our tax variables on leverage for intermediate companies and pure subsidiaries (i.e., subsidiaries that are not themselves parent companies). We find a negative coefficient for an included intermediate firm dummy variable, while the leverage of intermediate firms tends to respond relatively strongly to changes in the effective marginal tax rate. A heightened role for taxation to affect the leverage of intermediate firms is to be expected, if these firms are important in the overall tax planning of the firm. The tax incentive to shift debt abroad, however, is found to affect the leverage of intermediate and pure subsidiaries similarly.

In regression (4), we split the tax incentive to shift debt variable in one component that captures the incentive to shift debt to the parent country and another component that captures the incentive to shift debt to subsidiaries in other countries than the host and parent countries. Interestingly, we find that on average the incentives to shift debt to the parent country and to other countries both matter, although leverage appears to be more sensitive to the tax incentive to shift debt to other countries. These results imply that multinational firms not only consider tax-motivated debt-shifting opportunities between a foreign subsidiary and the parent country, but also among the various foreign subsidiaries. This finding supports our thesis that multilateral—rather than bilateral—differences in tax rates determine the financial structure of multinational firms.

In regression (5), we split the effective tax rate variable into one part that captures the taxation in the source country (to be found by setting the tax rate of the parent country to zero) and the complement that captures the taxation in the resident country. Parent country taxes should matter relatively little to the extent that multinationals are able to defer parent country taxes on foreign-source income unless this income is repatriated to the parent country. We find that the source-country part of the effective tax rate has a positive and statistically significant impact on leverage, while the parent-country part obtains a negative and insignificant coefficient, perhaps reflecting the option of deferral.

In regression (6), we split the two tax variables into parts that exclude and are specifically due to nonresident dividend withholding taxes. The first part is obtained by setting all withholding taxes to zero. The second part is obtained as the difference between our regular tax rate variables and the tax variables excluding withholding taxes. Interestingly, only the tax variables exclusive of withholding taxes are estimated with positive and significant coefficients. This suggests that withholding taxes are not seen as part of the effective tax burden, possibly because they can be avoided by triangular arbitrage involving a conduit company in a tax haven.

We are concerned about potential endogeneity that arises if countries respond to pervasive debt-shifting by changing their tax regimes (although tax regimes reflect a host of other factors as well). Since larger countries tend to have higher tax rates, we use the populations of the subsidiary and parent countries as instruments in the construction of the effective tax rate variable. Specifically, we recompute the effective tax rate using the populations of the subsidiary and the parent countries instead of these countries’ tax rates, taking into account possible double tax relief and assuming withholding taxes are zero. Because we do not have separate instruments for the tax incentive to shift debt variable, we only include the effective tax rate in this robustness check. The results for this instrumental variables regression are very similar. The coefficient on the effective tax rate variable is statistically significant and of similar magnitude as in previous regressions. A F-test of the excluded instruments supports the choice of our instruments. The first-stage results (nor reported) indicate that effective tax rates tend to be higher in more populous countries.

VI. Conclusions

This paper has considered the sensitivity of the capital structure of multinational firms to taxation. Generally this capital structure depends on the national or international structure of the firm and on the tax systems of all the countries where a firm operates. On the basis of a large sample of European firms over the 1994–2003 period, we find that a firm’s leverage depends on national tax rates as well as international tax rate differences. The relationship between leverage and international tax rate differences reflects the presence of international debt-shifting. While statistically highly significant, both the “domestic” and “international” effects of taxation on leverage are rather small.

International debt-shifting is shown to reflect a subsidiary’s tax rate differences vis-à-vis the parent firm as well as vis-à-vis other foreign-subsidiaries. This finding confirms our premise that international debt-shifting reflects the tax regimes of all the countries where the multinational operates rather then just bilateral tax rate differences vis-à-vis the parent firm. In practice, source-level taxation appears to be more important in affecting leverage than the residence-level taxation levied by a multinational’s parent country. This finding may reflect that parent-country taxes on a multinational’s foreign-source income in practice can be deferred, in some cases indefinitely. At the same time, corporate tax rates rather than nonresident dividend withholding tax rates appear to matter for leverage. This could reflect that multinationals are able to avoid bilateral nonresident dividend withholding taxes by using conduit companies in third countries.

International debt-shifting serves to lower average levels of corporate income taxation in high-tax countries. Countries with relatively low rates of taxation may benefit from international debt-shifting, as local establishments of multinational firms will be less highly leveraged than they would otherwise be—resulting in a higher corporate income tax revenues. Overall, international debt-shifting may introduce some dead-weight losses in the form of implementation costs for the multinational firms and also costs inherent in deviations from the firm’s optimal financial structure on the basis of nontax considerations. An obvious way to eliminate international debt-shifting is to harmonize top corporate income tax rates internationally. Alternatively, international debt-shifting is moot in case countries introduce a common, consolidated tax base for multinational firms. Tax coordination of either kind is not very likely in the near future so that international debt-shifting will remain an important policy for multinationals worldwide.

Appendix Table 1.

Variable Definitions and Data Sources

article image

References

  • Booth, Laurence, Varouj Aivazian, Asli Demirgüç-Kunt, and Vojislav Maksimovic, 2001, “Capital Structures in Developing Countries,” Journal of Finance, Vol. 54, pp. 87– 130.

    • Search Google Scholar
    • Export Citation

  • Chowdry, Bhagwan, and Josh Coval, 1998, “Internal Financing of Multinational Subsidiaries: Debt vs. Equity,” Journal of Corporate Finance, Vol. 4, pp. 87–106.

    • Search Google Scholar
    • Export Citation

  • Collins, J. H., and D. A. Shackelford, 1992, “Foreign Tax Credit Limitations and Preferred Stock Issuances,” Journal of Accounting Research, Vol. 30, pp. 103–24.

    • Search Google Scholar
    • Export Citation

  • Desai, Mihir A., C. Fritz Foley, and James R. Hines Jr., 2004, “A Multinational Perspective on Capital Structure Choice and Internal Capital Markets, Journal of Finance, Vol. 59, pp. 2451–88.

    • Search Google Scholar
    • Export Citation

  • Djankov, Simeon, Caralee McLieshand and Andrei Shleifer , 2005, “Private Credit in 129 Countries,” Journal of Financial Economics, forthcoming.

    • Search Google Scholar
    • Export Citation

  • Fama, Eugene F., and Kenneth R. French, 1998, “Taxes, Financing Decisions, and Firm Value,” Journal of Finance, Vol. 53, pp. 819–43.

    • Search Google Scholar
    • Export Citation

  • Froot, Kenneth A., and James R. Hines, 1995, “Interest Allocation Rules, Financing Patterns, and the Operation of U.S. Multinationals,” in Martin Feldstein, James R. Hines, and R. Glenn Hubbard, eds., The Effect of Taxation on Multinational Corporations (University of Chicago Press, Chicago) pp. 277–307.

    • Search Google Scholar
    • Export Citation

  • Gordon, Roger H., and Young Lee, 2001, “Do Taxes Affect Corporate Debt Policy? Evidence from U.S. Corporate Tax Return Data,” Journal of Public Economics, Vol. 82, pp. 195–224.

    • Search Google Scholar
    • Export Citation

  • Graham, John R., 1996, “Debt and the Marginal Tax Rate,” Journal of Financial Economics, Vol. 41, pp. 41–73.

  • Graham, John R., 2000, “How Big Are the Tax Benefits of Debt?” Journal of Finance, Vol. 55, pp. 1901–41.

  • Grubert, Harry, 1998, “Taxes and the Division of Foreign Operating Income Among Royalties, Interest, Dividends, and Retained Earnings,” Journal of Public Economics, Vol. 68, pp. 269–90.

    • Search Google Scholar
    • Export Citation

  • Hines, James R., Jr., 1994, “Credit and Deferral as International Investment Incentives,” Journal of Public Economics, Vol. 55, pp. 323–47.

    • Search Google Scholar
    • Export Citation

  • Hines, James R., Jr., and R. Glenn Hubbard, 1990, “Coming Home to America: Dividend Repatriations by U.S. Multinationals,” in Assaf Razin and Joel Slemrod, eds., Taxation and the Global Economy (University of Chicago, Chicago), pp. 161–200.

    • Search Google Scholar
    • Export Citation

  • Hines, James R., Jr., and Eric M. Rice, 1994, “Fiscal Paradise: Foreign Tax Havens and American Business,” Quarterly Journal of Economics, Vol. 109, pp. 149–82.

    • Search Google Scholar
    • Export Citation

  • Hodder, James, and Lemma Senbet, 1990, “International Capital Structure Equilibrium,” Journal of Finance, Vol. 45, pp. 1495–517

  • Hovakimian, Armen, Gayane Hovakimian, and Hassan Theranian, 2004, “Determinants of Target Capital Structures: The Case of Dual Debt and Equity Issues,” Journal of Financial Economics, Vol. 71, pp. 517–40.

    • Search Google Scholar
    • Export Citation

  • Job, Vijay, and Jianmin Tang, 2001, “Tax Reforms, Debt-shifting, and Tax Revenues: Multinational Corporations in Canada,” International Tax and Finance, Vol. 8, pp. 5–25.

    • Search Google Scholar
    • Export Citation

  • La Porta, Rafael, Florencio Lopez-de-Silanes, Andrei Shleifer, and Robert Vishny, 1997, “Legal Determinants of External Finance,” Journal of Finance, Vol. 52, pp. 1131–55.

    • Search Google Scholar
    • Export Citation

  • MacKie-Mason, Jeffrey K., 1990, “Do Taxes Affect Corporate Financing Decisions?” Journal of Finance, Vol. 45, pp. 1471–93.

  • Miller, Merton H., 1977, “Debt and Taxes,” Journal of Finance, Vol. 32, pp. 261–75.

  • Mills, Lillian F., and Kaye J. Newberry, 2004, “Do Foreign Multinationals’ Tax Incentives Influence their U.S. Income Reporting and Debt Policy?” National Tax Journal, Vol. 57, pp. 89–107.

    • Search Google Scholar
    • Export Citation

  • Moore, Padraig J., and Frances P. Ruane, 2005, “Taxation and the Financial Structure of Foreign Direct Investment,” IIIS Discussion Paper No. 88.

    • Search Google Scholar
    • Export Citation

  • Newberry, Kaye J., and Dan S. Dhaliwal, 2001, “Cross-Jurisdictional Income Shifting by U.S. Multinationals: Evidence from International Bond Offerings,” Journal of Accounting Research, Vol. 39, pp. 643–62.

    • Search Google Scholar
    • Export Citation

  • Organization for Economic Cooperation and Development, 1997, Model Tax Convention on Income and on Capital (Paris).

  • Rajan, Raghuram G., and Luigi Zingales, 1995, “What Do We Know About Capital Structure? Some Evidence from International Data,” Journal of Finance, Vol. 50, pp. 1421–60.

    • Search Google Scholar
    • Export Citation

  • Titman, Sheridan, and Roberto Wessels, 1998, “The Determinants of Capital Structure Choice,” Journal of Finance, Vol. 43, pp. 1–19.

    • Search Google Scholar
    • Export Citation
1

Harry Huizinga is affiliated with Tilburg University and the Center for Economic Policy Research (CEPR), Luc Laeven is affiliated with the International Monetary Fund (IMF) and the Center for Economic Policy Research (CEPR), and Gaëtan Nicodème is affiliated with the European Commission and Solvay Business School (ULB). The authors thank Julian Alworth, Michael Devereux, James Hines, Matthias Mors, Johannes Voget, Alfons Wiechenrieder, and seminar participants at the Institute for Fiscal Studies in London, Tilburg University, the University of Cologne, The University of Leuven, and the General Directorate of Taxation and Customs Union of the European Commission for their valuable comments.

2

It is reasonable to assume that multinationals do not take into account the taxation of dividend, interest, and capital gains at the investor level. First, important institutional investors such as pension funds may not be subject to taxation of the investor level. Second, private investors generally are subject to such taxation, but the internationally dispersed ownership of the shares of a multinational firm makes it difficult for these firms to take taxation at the personal level into account when deciding on their financing.

3

The complete dataset is available upon request from the authors.

4

Note that for the parent firm we have that the effective tax rate on corporate income equals the statutory rate, or τp =tp.

5

Most of the 33 countries in our sample had such treaties with each other. However, the treaty network of some countries—in particular some of the new EU member states and some non-EU countries—are far from complete. In contrast, France, Germany, Norway, Poland, Sweden, and the United Kingdom have a double-tax treaty in force with all other countries. Note that the table is not exactly symmetric because the entry into force may slightly differ in each of the two treaty partners.

6

In response to a change in Ei, the parent firm thus will change either Lp or Ep rather than Ap.

7

See Hovakimian, Hovakimian, and Tehranian (2004) for a recent discussion of the theoretical and empirical literature on target capital structures reflecting various costs and benefits of debt and equity.

8

Bankruptcy costs are incurred by loss-making firms and hence are assumed not to be deductible from taxable corporate income.

9

Higher local leverage may be disadvantageous if it increases the probability of losses that cannot be credited against profits made elsewhere in the firm. Losses that are not creditable per definition reduce the after-corporate-tax income of the firm one-for-one. For this reason, we assume that the costs associated with higher leverage at the establishment level are not deductible from taxable corporate income.

10

The firm recognizes all subsidiary and parent firm balance sheet identities, which means that the Ei are co-determined.

11

In this instance, the country fixed effect in part can serve to reflect so-called thin capitalization rules that may limit the tax benefits (in terms of interest deductibility) associated with subsidiary indebtedness.

12

The database is created by collecting standardized data received from 50 vendors across Europe. The local source for this data is generally the office of the Registrar of Companies.

13

The Amadeus database contains information only on European firms, and we therefore cover only the European operations of the multinationals in our sample. We can therefore not consider how tax differences between European countries and other parts of the world affect the capital structure of subsidiaries in Europe. While this is an important caveat to be mentioned, we do not see this as a major limitation of our analysis because European multinationals typically derive much of their revenues from operations in Europe rather than other parts of the world.

14

Desai, Foley, and Hines (2004) similarly find a coefficient of 0.2624 in their regression (1) in Table II, where they regress leverage ratios of U.S. outward FDI on the source country tax rate.

15

We only show the data for the year 2003 but have collected data on tax rates for each year in our sample period. The complete dataset on the international tax data collected is available upon request from the authors.

16

Zero percent on retained earnings. A distribution tax of 26% is applied on distributed profit.

17

Including a 3% social surcharge and a special 3.3% surcharge for large companies.

18

Including a solidarity surcharge of 5.5% and an average deductible trade tax of 16.14%. It also includes the exceptional 1.5% additional tax in 2003.

19

Including a deductible local business tax.

20

The rate is 25% for non-trading activities.

21

Including employment surcharge and local taxes.

22

Indirect tax credit if holding 75% for two years and treaty and where the EU Parent-subsidiary directive holds.

23

Exemption if EU Parent-Subsidiary directive applies (but foreign withholding tax is not creditable).

24

Including cantonal and local taxes in Zurich.

Same Series

Other IMF Content

Other Publishers

  • Collapse
  • Expand
Cover IMF Working Papers
Capital Structure and International Debt Shifting
Author:
Mr. Luc Laeven
,
Harry Huizinga
, and
Gaetan Nicodeme
Volume/Issue:
Volume 2007: Issue 039
Publisher:
International Monetary Fund
ISBN:
9781451866032
ISSN:
1018-5941
Pages:
37
DOI:
https://doi.org/10.5089/9781451866032.001