Download Capital Structure of Multinational Firms: Evidence from Developed and Emerging Markets and more Lecture notes International Finance and Trade in PDF only on Docsity!
The Corporate Finance of Multinational Firms
Isil Erel Ohio State University, NBER, and ECGI Yeejin Jang University of New South Wales Michael S. Weisbach Ohio State University, NBER, and ECGI November 10, 2019
Abstract An increasing fraction of firms worldwide operate in multiple countries. We study the costs andbenefits of being multinational in firms’ corporate financial decisions and survey the related academic evidence. We document that, among U.S. publicly traded firms, the prevalence of multinationals is approximately the same as domestic firms, using classification schemes relying on both income-based and a sales-based metrics. Outside the U.S., the fraction is lower but has been growing. Multinational firms are exposed to additional risks beyond those facing domesticfirms coming from political factors and exchange rates. However, they are likely to benefit from diversification of cash flows and flexibility in capital sources. We show that multinational firms, indeed, have a better access to foreign capital markets and a lower cost of debt than otherwise identical domestic firms, but the evidence on the cost of equity is mixed.
- We thank Fritz Foley, Jim Hines, David Wessel and participants in a presentation at Brookings for very helpful suggestions. Greg Allen, Hyeik Kim, Rick Ogden provided excellent research assistance.
1. Introduction As the world economy has become more integrated, there has been an increase in the number of multinational firms. As of 2017, about half of the publicly-traded firms in the U.S. are multinationals. For the average multinational firm, foreign income (sales) represent about 40% of aggregate income (sales). The extent of international operations of multinational firms is similar for international firms in MSCI World developed countries. As the global economy becomes more integrated, the fraction of firms with foreign sales also rose rapidly in emerging markets. Given that multinational firms are such significant players in the world economy, understanding their financial policies is an important task. Operating in more than one country can affect a firm’s financial decisions in a number of ways. Most importantly, being multinational appears to affect both firms’ cost of finance and their access to capital during poor economic times. An important reason for financing advantages of multinational firms is that they have more flexibility in their potential sources of financing than domestic firms. In principle, any firm could borrow from any bank in the world or issue public equity or debt in any country. However, for a number of reasons, it is usually much more cost-effective for firms to raise capital in locations where they have operations (see Jang (2017)). Financing international activities from local capital provides a natural hedge against currency risks. Furthermore, additional choices of where to raise capital can allow a firm to better optimize over rates, and also to diversify its sources of financing, which can be valuable when financing becomes scarce in one part of the world. In addition, being multinational diversifies a firm’s cash flows across countries and minimizes the impact of country-specific shocks. Therefore, multinational firms have lower cash-
Multinational firms also face exchange-rate risk. They receive revenues in a mix of currencies and have liabilities, both in terms of production costs and interest payments, which are likely in a different mix of currencies. Therefore, movements in exchange rates create a mismatch between the income the company receives and its liabilities, creating a demand for hedging foreign exchange risk. This paper surveys the academic literature on the costs and benefits of multinational firms relative to domestic ones with respect to their corporate financial decisions. Section 2 characterizes multinational firms and presents detailed characteristics of multinationals in the U.S. and overseas. In Section 3, we provide detailed summary statistics on the capital structure of multinational firms and compare their capital structure with the capital structure of domestic firms. Section 4 discusses the way in which firms diversify their sources of financing and the impact of this diversification plays on multinational firms’ costs of financing. Section 5 discusses the additional risks faced by multinational firms in comparison to the domestic counterparts. Section 6 provides a short summary.
2. Characteristics of Multinational Firms There are a number of ways in which one could define a multinational firm. Since none of these definitions are perfect, we utilize two different approaches. First, we use foreign pretax income to characterize multinational firms. Following Jang (2017), we create a dummy variable that takes on a value of one when a firm reports non-zero foreign income in the previous three years as a measure of whether a firm is multinational. 1 However, this approach could
(^1) We use pre-tax foreign income (PIFO) in Compustat and international operating income (WC07126) in Worldscope. Although Jang (2017) complements this definition using information on the location of subsidiaries from firms’ 10Kfilings, we concentrate on this definition based on non-zero foreign income definition due to data constraints.
mischaracterize firms that do not report foreign income when the percentage of foreign income is relatively low compared to overall income or when it is negative. Therefore, as a second measure, following Denis, Denis, and Yost (2002), we use information on foreign sales from the Compustat Geographic Segment database in the U.S. or Worldscope for foreign firms to define internationally diversified firms.^2 Specifically, a firm is defined as multinational if at least 5% of its sales are from outside of its home country. According to this definition, however, a firm that exports goods to other countries would be defined as a multinational firm even if it does not have any assets outside its home country.^3 Panel A of Table 1 presents statistics on the number of multinational and domestic public firms in the U.S. between 1986 and 2017 using both definitions. The percentage of multinational firms defined using foreign income (sales) increased from 20% (18%) to 48% (40%) over the last 32 years. Interestingly, Figure 1 documents the increase in the ratio of multinational firms despite the significant decline in the number of U.S. public firms over time (see Doidge, Karolyi and Stulz (2017)). The number of public firms in the U.S. decreased almost by half over the 32 years in our sample, with surviving firms being more than twice as likely to be multinational. There are a number of reasons for the sharp decline in the number of public firms in the early 2000s, but a particularly important one is the increasing role of private capital markets during this period. Many public firms had leveraged buyouts and went private, IPOs were rare despite the strong economy, and private capital markets allowed startups to remain private for a very long
(^2) We use international sales (WC07101) in Worldscope. (^3) The definitions for multinational firms in both approaches might not be comparable across countries as reporting requirements for foreign income or foreign sales might vary by country. In case of U.S., all publicly traded firms arerequired to disclose foreign income or foreign sales separately for material foreign operations or sales. For example, SEC Regulation §210.4-08(h) requires any U.S. public firms to separately disclose pre-tax income and income taxexpenses for domestic and foreign operations, if any of these measures for non-U.S. operations exceed 5% of the consolidated total. Under SFAS No. 131, any U.S. public firms must report separately information about an operatingsegment if its reported revenue is 10% or more of the combined revenue of all reported operating segments.
multinational likely reflect differences in the nature of firms in different countries and also the likelihood that firms in different countries go public. Using foreign income to identify multinational firms in countries other than the U.S. could be problematic due to variations in reporting requirements across countries. In addition, as we noted above, it is possible for firms not to report foreign income when the percentage of foreign income is relatively low compared to overall income or when it is negative. According to this definition, percent of firms that were multinationals was 17% (8%) in 2000 and 5% (3%) in 2017 in MSCI developed (emerging) markets. However, we know that, among firms that report foreign income in the past three years, on average, 38% (26%) of their income was foreign in 2017. Panel C presents the same numbers across individual countries averaged over 2000-2017. While U.S., Ireland, U.K, Hong Kong and Netherlands have the highest ratio of multinational firms (classified using the foreign income definition), China, Columbia, Russia, India, and Brazil have the lowest ratios of such firms. Panel C also reports debt-to-assets and cash-to-assets ratios of multinational and domestic firms in these countries. In Figure 2, we visualize the differences in these ratios between multinational firms and domestic firms. The reference line identifies the zero difference in leverage or cash ratios for multinational and domestic firms. Consistent with the idea that multinational activity increases a firm’s debt capacity, many countries are located above the reference line, showing that the leverage ratio of multinational firms tends to be higher than that of domestic firms, especially in emerging markets. We also find that, with a few exceptions of emerging countries such as Czech Republic, India, and Brazil, multinational firms have much lower cash holdings than domestic firms. One exceptional case is the U.S., where multinational firms hold more cash than domestic firms on average, consistent with the evidence documented by Pinkowitz, Stulz and Williamson (2016).
Next, we characterize the industry distribution of multinational firms. Table 2 presents the percentage of multinational firms across industries in the U.S. (Panel A), and in MSCI developed and emerging countries (Panels B and C) across twelve Fama-French industries.^6 As also shown in Figure 3, some industries stand out in terms of the fraction of multinational firms in the U.S.: Chemical and Allied Products (61%), Manufacturing (49%), Business Equipment (48%), and Consumer Durables (48%). The other U.S. industries have much lower percentages, with 33% or much less of their firms being multinational. Differences across industries are less evident in both other developed countries and emerging countries. While Finance and Utilities have the lowest fraction of multinational firms in the U.S. (1-5%), firms in those industries are much more globalized in other developed countries (about 12-13%). In addition, compared with multinational firms in emerging countries, the ones in developed countries rely more on foreign income, in particular, in Oil, Gas, and Coal Extraction and Chemical and Allied Products industries (40-44% in the U.S. and 31-39% in other developed countries). While the fraction of multinational firms in the U.S. is relatively low, multinational firms in Utilities and Finance generate 17-29% of income overseas. Multinational firms are likely to differ from domestic firms in a number of ways. In Table 3, we present firm-level characteristics of multinational firms and domestic ones. Panel A contains statistics for the U.S. sample, covering 1986-2017. This panel indicates that multinational firms tend to be larger and more stable than domestic firms. Average total assets are almost twice as high for multinationals as for domestic firms, and multinationals’ average market capitalization is almost four times as high. Cash flows are also much higher, and importantly, the standard deviation of cash flows is statistically significantly lower for multinationals, possibly because of
(^6) We rely on the foreign-income-based definition of multinational firms here, and throughout the rest of the paper. Similar tables created based on foreign-sales-based definition are available upon request from the authors.
3. The Capital Structure of Multinational Firms 3.1. Capital Structure of U.S. Multinationals In Table 4, we provide detailed summary statistics for capital structures of U.S. multinational firms, first for the entire universe of public firms and then for firms of differing credit quality. We start by reporting detailed characteristics for all firms, where the majority of firms are unrated. To compare similarly rated firms, we analyze investment-grade rated (Panel B) and speculative-grade rated (Panel C) separately. Table 4 indicates that U.S. multinational firms are more likely to hold cash, as a percentage of their assets, but less likely to borrow in the debt market. The average cash-to-assets ratio is 16.7% for multinationals while it is only 13.5% for domestic firms in the U.S. However, the mean total debt-to-asset ratio is 23.7% for multinationals and 26.5% for domestic firms (although median ratios are both 21%). Larger cash holdings likely reflect tax effects as multinationals have incentives to hold cash earned overseas rather than pay repatriation taxes for cash returned to domestic shareholders. The differences in debt ratios would occur if these firms followed the “pecking order” theory of capital (Myers and Majluf (1984)). Since multinationals are likely to have been historically successful in generating cash flows and diversifying internationally, they would not have had to go down the pecking order as often as similar domestic firms would. Thus, compared with domestic firms, multinational firms would carry less debt on their balance sheets. When we examine the details of the debt structure (i.e., senior bonds, subordinated bonds, commercial paper or bank debt), we see that the lower total debt ratio for multinational firms occurs because of multinationals’ bank loans being lower as a ratio of total debt. Term bank loans, on average, form 20% of total debt for domestic firms but only 14% for multinational firms, with the difference being statistically significant at the 1% level. Compared with domestic firms, an
average multinational firm relies more on senior bonds or notes (30% of total debt vs. 19.3% of total debt). They are also more likely to issue convertible debt and commercial paper. On the other hand, domestic firms, since they tend to be less credit-worthy, are 11% more likely to borrow with collateral (38% vs. 27%). Next, we compare capital structures of U.S. multinational and domestic firms grouped by credit ratings. We first compare investment-grade rated firms (Panel B of Table 4) and then speculative-grade rated ones (Panel C of Table 4). Rated multinational firms, investment grade or speculative grade, hold more cash but have less total debt, especially long-term debt, compared with respective domestic firms. As expected, investment-grade firms rely more on senior bonds or notes and less on bank debt. The ratio of senior bonds or notes, as a fraction of debt, increases to 58.4% for investment-grade-rated multinational firms from 30% overall and the ratio of bank term loans, as a fraction of debt, drops to 3% for investment-grade-rated multinational firms from 14% overall. For speculative-grade firms, we see different patterns in the details of debt structure for multinationals in comparison to domestic firms: the average multinational firm has more term loans, as a fraction of its total debt, than the average domestic firm (21% vs 16%) while the ratio of senior bonds and notes is weakly different from each other for speculative-grade firms. Holding credit ratings constant is one way to limit the sample to comparable multinational and domestic firms. Another way is to focus on firms with similar asset sizes. Next, in Table 5, we analyze multinationals’ capital structures relative to those of domestic firms within different size buckets. We first compare large multinational firms with similarly-sized domestic firms (Panel A). These large domestic firms have total assets larger than or equal to the median total assets of multinational firms, with the median estimated each year. The match creates relatively similar firms in terms of total assets, with mean total assets of $19.64 billion for multinational firms and
foreign multinational firms in developed countries, compared with domestic firms in their respective regions, hold less cash as a fraction of their total assets (14.6% vs. 16.6%). Also, even though their leverage ratio is similar to U.S. counterparts, non-U.S. multinational firms rely less on the senior bond market for financing, but more on the bank debt market. For example, as discussed above, bank debt is on average 26% of total debt for U.S multinational firms (and 20% of total debt for large U.S. multinationals, which are similar in size to the multinationals from other developed countries). However, for multinationals from MSCI developed or emerging countries, this ratio is over 50%.
4. Capital Raising by Multinational Firms 4.1. Where do Multinationals Get Financing? An important consideration when a firm raises capital is the location of the capital provider. Being geographically closer to the capital provider reduces information asymmetry between the lenders and the borrowers (see, e.g., Sufi (2007)). Given that they have assets in multiple countries, multinationals can more easily raise capital from multiple countries. The location of where to raise capital represents an important corporate financial decision faced by multinational firms.
4.1.1. Debt Financing The two most important forms of debt finance are publicly traded bonds and bank debt. Henderson, Jegadeesh, and Weisbach (2006) document that in their sample period (1990-2001), about 20% of all capital raised through bond issues comes from outside the issuing firm’s home
country. The most common places of issue of international bonds are the U.S. and Europe, and many issuers are multinationals from countries with less liquid capital markets.^7 Most large bank loans are syndicated across multiple banks, and the participating banks often come from different countries. Therefore, loans are often made up of capital from multiple countries. Table 7 summarizes these patterns for syndicated loans made to U.S. firms during the period 1990 to 2018. For multinationals, 33% of loans had at least one participating bank from Canada and 32% had at least one from the U.K. Domestic firms also have syndicated loans with foreign banks participating, but to a lesser degree: 23% of syndicated loans to domestic firms had at least one lender from Canada and 16% had at least one lender from the U.K. Next, we examine the difference in bank loan sources between multinational firms and domestic firms in a regression setting. The sample includes syndicated loans issued to U.S. public firms during the period of 1990-2018 obtained from Dealscan , aggregated at the loan package level. We estimate OLS regressions, where the dependent variable is the indicator for including at least one foreign lender in its syndicate in columns (1) and (2), and the percentage of foreign lender as the total number of syndicate members in columns (3) and (4). The main independent variables are the multinational indicator based on non-zero foreign income in past three years and the percentage of foreign income. The regressions include controls of loan features and borrower characteristics and year, industry, borrower rating, and deal purpose fixed effects. Table 8 presents the results that examine the differences in the sources of loans between multinational and domestic firms, with controls for a number of variables that could potentially affect the structure of the loan. This table illustrates that multinationals are 5% more likely to have at least one foreign lender in the syndicate. In addition, they have 3.5% higher fraction of foreign
(^7) See Henderson, Jegadeesh and Weisbach (2006), Appendix C, for details on the amount of different types of capital that firms from a number of countries receive from each other country.
between multinational and domestic firms still remains significant. Like debt, the decision to raise equity overseas is a financial decision that is usually made by multinational firms rather than domestic ones.
4.2. Why Multinational Firms Have a Financing Advantage Jang (2017) addresses one way in which multinational firms have a financing advantage over domestic firms. She considers the issue of whether the presence of an operation can facilitate capital-raising in the country where the operation is located. Presumably, the foreign operations can lower information asymmetry and monitoring costs to a capital provider. Jang (2017) documents that firms are more likely to receive bank loans from foreign lenders in countries where foreign subsidiaries are located. In addition, this better access to foreign capital markets can help multinationals raise capital if there are financial market disruptions in one part of the world. For example, during the Financial Crisis of 2008, capital raising in the U.S. and Europe became extremely difficult. Firms with Asian operations, where capital markets continued to function more or less normally, consequently had an advantage in raising capital so were better able to weather the Crisis. This funding advantage makes multinational firms’ income more stable than domestic firms’ income. Consistent with this idea, Figure 4 documents that U.S. multinationals’ foreign income as a percentage of sales has been growing irrespective of business cycles, while their domestic income declines during down cycles. Another way in which a multinational firm can take advantage of its foreign presence is by taking advantage of interest rate differences across countries. If rates differ in countries and firms do not perceive that the differences are offset by changes in expected currency movements, then firms can benefit by issuing debt in countries with lower rates. In principle, domestic firms could
raise debt in any country. However, in practice it is much more likely that multinational firms raise capital outside their own country, and as Jang (2017) documents, they are especially likely to raise debt in countries where their foreign operations are located. Consistent with this idea, Henderson, Jegadeesh, and Weisbach (2006) and McBrady, Mortal, and Schill (2010) find that firms are more likely to issue bonds in countries with lower rates, and Keloharju and Niskanen (2001) find that a sample of Finnish companies are more likely to issue foreign bank debt when interest rates in Finland are relatively high. Allen (2019) links these decisions directly to monetary policy; he finds that when the central bank lowers rates in one country, multinational firms in other countries become more likely to raise debt in that country’s currency. Multinational firms, even if they focus on one industry, are much like a multidivisional firm in that each country’s operation usually operates separately from the others and thus have different exposures to country-specific shocks. Consequently, much of the analysis in the internal capital markets literature is relevant to multinational firms as well. Diversifying operations and using internal capital markets could in principle lead to cross-subsidization that lowers financial constraints (see e.g., Stein (1997, 2003)). Since a firm’s cash flows in different countries are exposed to different shocks, international diversification could reduce the overall volatility of cash flows and lower default risk through cross-subsidization across countries. The diversification of cash flows across countries provides multinationals with a financing advantage over domestic firms even if their marginal source of finance is from domestic sources. To examine the importance of the cross-country internal capital markets generated by multinational firms, Desai, Foley, and Forbes (2008) compare the way in which large currency depreciations affect multinational and domestic firms. Consistent with the notion that being multinational helps diversify sources of capital, these authors find that U.S. multinational
otherwise identical domestic firms. In addition, the evidence in Reeb, Mansi and Allee (2011) suggests that the lower cost of debt goes beyond what is conveyed by debt ratings, so that the market incorporates the international aspects of firms’ operations even more than analysts. Houston, Itzkowitz, and Naranjo (2007) analyze the time series of syndicated loans in ten countries. These authors find that loans to European firms carry significantly lower spreads than loans to North American firms. They find that multinationals enjoy this cheaper bank funding only by using lead arrangers in Europe. They also find that larger firms on average pay lower spreads on their loans when they borrow from a foreign lender. This evidence suggests that global competition has helped reduce borrowing costs for multinational firms with greater access to lending markets around the world. We present our own analysis of whether multinationals have a lower cost of debt than domestic firms in Table 10. This analysis uses data from the Dealscan database, and presents estimates of equations predicting loan spreads as a function of whether the firm is multinational, as well as a host of other firm-level and loan-level factors that potentially affect spreads. In Column (1), the coefficient on the multinational dummy is -4.4 and is statistically significantly different from zero. This estimate implies that multinational firms pay about 4.4 basis points less on their debt than an otherwise identical domestic firm. In Column (2), we replace the multinational dummy with the fraction of foreign income; the coefficient is also negative (-.8.2) and statistically significantly different from zero. Overall, these equations suggest that when firms are more diverse geographically, they pay lower lending rates. An important issue in interpreting these results is that of causality. The literature finds that multinational firms pay lower lending rates than domestic firms, holding other things constant. However, it is possible that there are other factors that are related both to a firm being multinational
and also to its lending rates that could potentially lead to a spurious correlation between multinationality and lower rates. For example, if more profitable firms with more stable cash flows are more likely to diversify internationally, then these firms would likely borrow at lower rates regardless of the marginal impact of their being multinational. While the correlation between being multinational and lending rates appears to be robust, we cannot definitively say that the causal interpretation is appropriate. Identifying this relation more precisely would be an excellent topic for future research.
4.3.2. The Cost of Equity Measuring the cost of equity is typically more difficult than measuring the cost of debt since equity does not have an observable promised yield that equals the amount of interest the firm will pay if it does not default. Perhaps for this reason, the literature’s conclusion on the impact of multinationality on the cost of equity is not as clear as it is for the cost of debt, with the results varying depending on the method used. Baker, Foley, and Wurgler (2009) study the effect of domestic stock market valuations on firms’ foreign direct investment (FDI) decisions. The idea is that when valuations are unusually high, the firm faces a particularly low cost of equity, so they have an incentive to use this low cost of equity to make investments. The authors find that when multinationals appear to be overvalued in their home country, they take advantage of their valuation and act as cross-border arbitragers by investing overseas in Foreign Direct Investment. At least some of the time, Baker, Foley, and Wurgler’s results suggest that multinationals have a relatively low cost of equity. A more direct way to estimate the impact of multinationality on the cost of equity is by considering the returns that are earned by multinationals relative to the returns on otherwise
