WPS7815 Policy Research Working Paper 7815 Corporate Borrowing and Debt Maturity The Effects of Market Access and Crises Juan J. Cortina Tatiana Didier Sergio L. Schmukler Development Research Group Macroeconomics and Growth Team September 2016 Policy Research Working Paper 7815 Abstract This paper studies how access to different markets and switched markets compensated for the financing shocks and crises impact debt financing and maturity. Using data on maintained, or increased, their borrowing maturity. Coun- worldwide corporate issuance activity in domestic and try-level maturities also remained stable or even lengthened. international bond and syndicated loan markets during However, firms that did not move across markets typically 1991–2014, the paper shows that these markets are affected experienced declining financing and shorter borrowing differently by crises, while providing financing to differ- maturities. Firm movements across markets are consistent ent firms at distinct maturities. During the global financial with credit tightening during crises due to supply-side shocks, crisis and domestic banking crises, large firms moved away significantly affecting debt composition, borrowing matu- from the crisis-hit markets toward less affected, longer-term rity, and credit redistribution across firms of different sizes. ones, switching their financing sources. Hence, firms that This paper is a product of the Macroeconomics and Growth Team, Development Research Group. It is part of a larger effort by the World Bank to provide open access to its research and make a contribution to development policy discussions around the world. Policy Research Working Papers are also posted on the Web at http://econ.worldbank.org. The authors may be contacted at jcortinalorente@worldbank.org, tdidier@worldbank.org, and sschmukler@worldbank.org. The Policy Research Working Paper Series disseminates the findings of work in progress to encourage the exchange of ideas about development issues. An objective of the series is to get the findings out quickly, even if the presentations are less than fully polished. The papers carry the names of the authors and should be cited accordingly. The findings, interpretations, and conclusions expressed in this paper are entirely those of the authors. They do not necessarily represent the views of the International Bank for Reconstruction and Development/World Bank and its affiliated organizations, or those of the Executive Directors of the World Bank or the governments they represent. Produced by the Research Support Team Corporate Borrowing and Debt Maturity: The Effects of Market Access and Crises Juan J. Cortina Tatiana Didier Sergio L. Schmukler * JEL Classification Codes: F65, G00, G10, G32 Keywords: borrowing maturity, capital raising, corporate bonds, debt markets, firm financing, global financial crisis (GFC), syndicated loans * We received very helpful comments from Tamon Asonuma, Max Bruche, Julián Caballero, Stijn Claessens, Asli Demirgüç-Kunt, Peter Eckley, Christian Eufinger, Catiana Garcia-Kilroy, Yuko Hashimoto, Aart Kraay, Olivier Jeanne, María Soledad Martínez Pería, Colin Mayer, Marco Pagano, Andrea Polo, Andrea Presbitero, Alexander Popov, Carmen Reinhart, Heinz Rudolph, Luis Servén, Nivikar Singh, Thierry Tressel, and participants at presentations held at the ASSA Annual Meetings, Bocconi University, Borsa Istanbul, Central Bank of Argentina, Central Bank of Chile, CSEF-University of Naples Federico II, EBRD, FIW Research Conference “International Economics,” Georgetown University, the International Monetary Fund, LACEA-LAMES Annual Meetings, Stanford University, University of Buenos Aires, University of Cape Town, the Washington Area International Finance Symposium, and the World Bank (Washington, DC and Malaysia Knowledge and Research Hub). This paper won the Best Paper Award at the 2nd International Borsa Istanbul Finance and Economics Conference. This paper has been developed as part of the CEPR’s Restarting European Long- Term Investment Finance (RELTIF) Programme (funded by Emittenti Titoli) and as background work for the World Bank Global Financial Development Report, which generously provided financial support through the World Bank Knowledge for Change Program and Strategic Research Program. An early version of this paper was circulated as World Bank Policy Research Working Paper No. 7815, under the title “How Long Is the Maturity of Corporate Borrowing? Evidence from Bond and Loan Issuances across Markets.” Email addresses: jcortinalorente@worldbank.org, tdidier@worldbank.org, sschmukler@worldbank.org. 1. Introduction During crises, financial conditions deteriorate. Financing dries up and maturities shorten. Companies experience difficulties repaying their debt and face increasing prospects of default. The existing literature typically characterizes these patterns by analyzing debt dynamics around crises within specific debt markets or countries. Though valuable, these analyses offer only a partial view of firm financing because corporations access different debt markets, domestically and internationally. In a world with perfectly integrated and frictionless markets, the specific market in which firms obtain financing would be irrelevant (Modigliani and Miller, 1958). But in practice various frictions can lead to market segmentation, with different debt markets attracting distinct types of firms and investors, delivering different volumes and terms of financing, even to the same firm. 1 In such scenario, when crises hit, some firms would change the amount and type of financing by moving across markets, while other firms would remain in specific markets. In this paper, we expand the literature by analyzing corporate debt financing and maturity across four different markets around the world: domestic and international bonds and syndicated loans. These markets became increasingly important sources of new financing to corporations, grew faster than equity and traditional bank financing, and expanded five times faster than gross domestic product (GDP) between 1991 and 2014. 2 By focusing on financing and maturity, we show that these markets are different from each other. They provide financing to different firms at different maturities, and have their own dynamics around crises. We also show that when negative shocks affect specific 1 For example, asymmetric information, taxes, fixed transaction costs, and regulations could inhibit some firms from using certain markets, and could result on certain investors and financial intermediaries (with specific preferences, time horizons, or abilities to diversify risk) dominating specific markets (La Porta et al.; 1997; Karolyi and Stulz, 2003; Pirinsky and Wang, 2006; Japelli and Pagano, 2008; Bekaert et al., 2011). 2 Traditional bank financing has grown at a much lower pace during the same period (Didier et al., 2015). Some studies estimate that syndicated loans are the main vehicle through which banks lend to large corporations, accounting for roughly one-third of the total outstanding cross-border loans (Ivashina and Scharfstein, 2010; Cerutti et al., 2015). In emerging economies, bond financing has increased particularly fast during the 2000s (Turner, 2014; Caballero et al., 2016). 1 debt markets, some firms appear to be constrained to the crisis-hit markets, while others counteract the shocks by seeking financing elsewhere. Consequently, the amount and maturity of financing varies significantly across firms depending on which markets they access, with important aggregate effects. The 2008-09 global financial crisis (GFC) and domestic banking crises allow us to show the effects of shocks to different markets and the reaction of different firms. To conduct the analysis, we assemble a new dataset using a comprehensive cross-section of 78 countries and 57,950 firms for 1991-2014. We collect publicly and privately placed issuances of domestic and cross-border bonds and syndicated loans with an original maturity of more than one year, encompassing issuances by listed and unlisted firms. The dataset includes 267,382 individual debt issuances conducted around the world by firms from 41 developed and 37 developing countries. Our main findings can be summarized as follows. First, we show that maturities at issuance vary significantly across the different debt markets. Although the overall debt maturity at issuance is on average 6.6 years, different markets provide financing at distinct maturities. Bonds have longer average maturity than syndicated loans. In developed countries, bonds are about 5 years longer than syndicated loans, no matter the market of issuance (domestic or international). Developing countries show a different pattern: international bonds and domestic syndicated loans have relatively longer maturities than domestic bonds and international syndicated loans. The differences in the maturity of debt issuances across markets exist even after controlling for firm-level and time-varying country fixed effects, currency of issuance, type of rate (fixed versus floating), and use of proceeds. Hence, the composition of financing across markets matters for debt maturity across firms and countries. Second, the relative importance of each market for firm financing varies over time, with significant compositional changes during market-specific crises. When the GFC suddenly hit the banking sector of major developed countries, firms from developed countries with demand for credit moved away from the impacted syndicated loan markets toward bond markets. In developing 2 countries, firms reacted by moving toward the less affected domestic markets, increasing their debt issuances of both domestic syndicated loans and bonds instead of international ones. Analogously, during domestic banking crises, firms mitigated the shock in the local banking system by moving away from domestic syndicated loans toward bond and international markets. These changes in debt financing composition during crises took place both at the country level and within firms. That is, some firms with demand for credit switched one type of debt for another during crises, evidencing a negative credit supply shock in the market they leave. Nonetheless, not all firms counteracted the negative shocks by raising capital in other markets. Firms that did not move from the markets in crisis tended to experience declining debt financing in the aftermath of the shocks. Third, the time-varying activity in each market and the ensuing changes in debt composition are reflected in borrowing maturities. During the GFC, the maturity of debt at issuance tended to decline in individual markets. However, the movement toward less affected markets translated into stable borrowing maturities, given that these markets were relatively longer term than the markets in crisis. The overall maturity at issuance during the GFC remained stable both at the country level and for those firms that switched markets. Firms issuing the same type of debt before and in the aftermath of the GFC tended to experience declining borrowing maturities. Similarly, during domestic banking crises, firms increased issuances in relatively longer-term markets, which lengthened their overall debt maturity despite shortening maturities in the shock-hit domestic syndicated loan markets. Thus, firm movements across debt markets triggered by supply-side shocks explain the stability (or even increase) in the borrowing maturity at the firm and aggregate levels during those events. Fourth, the composition of firms obtaining financing across markets changed during market- specific crises given that different firms responded differently to the negative shocks. The largest firms moved away from the shock-hit markets during the GFC and domestic banking crises, compensating for the declines in debt financing and maturity. In contrast, firms that did not change their financing 3 composition during crises (the non-switchers) typically experienced declining financing and, if able to raise capital, shorter borrowing maturities. During the GFC, non-switchers in domestic markets (comprising the smallest firms in the sample) experienced lower debt financing despite the increase in total lending in the markets not directly hit by the negative shock, namely, domestic bond and syndicated loan markets in developing countries and domestic bond markets in developed countries. This latter pattern is consistent with crowding out effects caused by the larger firms moving toward the less affected domestic markets, implying a redistribution of new credit to larger corporations. In sum, the results from this paper show that the market of issuance is relevant for firm- and country-level financing and maturity. The issuance place matters even after controlling for factors that the literature highlights as key determinants of the demand for short- and long-term debt. To the extent that individual debt markets specialize in particular types of financing, firms might select to issue in specific markets to obtain debt financing at certain maturities. Access to different debt markets also provides firms with diversification benefits, improving their resilience to shocks. Our findings indicate that access to several markets allows firms to use them as complements during good times (obtaining different types of financing in each market) and as substitutes when conditions deteriorate (cushioning the decline in financing in certain markets and in maturity across all individual markets). In contrast, when firms cannot choose markets, they are constrained to the financing and maturity dynamics of the markets where they issue. Our results show that smaller firms seem to be constrained by the funding conditions of the specific market they access. The joint analysis of a wider set of debt markets allows us to obtain a more complete perspective of the dynamics of firm- and country-level financing than analyses based on individual debt markets. The evidence in this paper contributes to different strands of the literature. One strand analyzes how firms use syndicated loan and bond markets to withstand credit supply shocks (Adrian 4 et al., 2013; Becker and Ivashina, 2014). 3 Though innovative, this analysis was conducted only for U.S. publicly listed firms, without an examination of the consequences of such substitutions on the new terms of financing. In this paper, we use a similar methodology to identify supply-side shocks and to study how firms change the composition of their financing at the time of these shocks. But we examine a wider and more heterogeneous set of firms and debt markets: listed and unlisted firms from around the world raising capital in domestic and international bond and syndicated loan markets. We provide evidence of firm-level substitution across markets at the global level during both domestic and external financial crises. Importantly, these market switches during crises are consequential for corporate debt maturity and financing of both firms that move across markets and those that do not. The overall evidence suggests that not all firms react to shocks in the same manner. Relatively smaller firms can be constrained to the markets in crisis, limiting their funding options. The different behavior of smaller and larger firms is consistent with findings based on balance sheet data showing that, following negative loan supply shocks, the leverage of smaller (more bank-dependent) firms falls relative to that of larger firms with access to different markets (Leary, 2009). Another strand of the literature emphasizes the risks associated with the use of short-term debt and the benefits related to long-term debt. Although the maturity structure is a risk-sharing outcome between debtors and creditors, long-term debt allows the former to finance large investments and reduce rollover risks associated with crises. 4 When firms rely more intensively on short-term debt, 3 Kashyap et al. (1993) use a similar a similar methodology to study the relative movements in bank loans and commercial paper after monetary shocks. However, the use of aggregate data makes it difficult to truly disentangle supply from demand effects (Oliner and Redebusch, 1995, 1996). Hale (2007) provides empirical evidence that default on sovereign debt has an impact on the composition of international debt instruments used by private and sovereign borrowers from emerging markets. 4 The literature that links “short-termism” to crises is quite large (e.g., Eichengreen and Hausmann, 1999; Rodrik and Velasco, 2000; Tirole, 2003; Brunnermeier, 2009; Jeanne, 2009; Raddatz, 2010; Beltratti and Stulz, 2012). Some papers argue that the bonds issued by corporations (particularly from emerging economies) in the aftermath of the GFC exposed them to shocks (Shin, 2013; Acharya et al., 2015; IMF, 2015; McCauley et al., 2015; Lo Duca et al., 2016; The Economist, 2016). Much of this discussion has centered on the currency of issuance and its associated risks (Bruno and Shin, 2016; Hale et al., 2016). But it is possible that there is a trade-off between currency and maturity (Shin, 2016) and that part of the increase in bond issuance is a response to the decline in bank loans. 5 they can experience higher financial constraints during credit crunches, leading to significant real effects (Almeida et al., 2011). Consequently, the discussions have focused on the need to understand the extent and causes of “short-termism” as well as on efforts aimed at lengthening debt maturity (G20, 2013; Giovannini et al., 2015; World Bank, 2015; Beck, 2016; Cortina et al., 2018). However, evidence on the actual debt maturity across countries and where long- and short-term debt originates is still scant. 5 This paper helps fill this void by analyzing the borrowing maturity around the world across different debt markets, as well as how they fluctuate around crises across different types of issuing firms. The results indicate that at least part of the changes in maturity over time and of the differences across firms and countries lie on the type of debt markets that firms access. Several studies examine the cross-sectional determinants of short- and long-term debt by focusing on the demand side of capital. Theories of agency cost, commitment, growth opportunities, asymmetric information, liquidation risk, and taxation show how firm attributes affect their debt maturity (Myers, 1977; Mauer and Lewellen, 1987; Flannery, 1986; Calomiris and Kahn, 1991; Diamond, 1991; Greenwood et al., 2010; Brunnermeier and Oehmke, 2013; Dangl and Zechner, 2016). 6 The empirical research examines these predictions for firms in the United States, mostly using firm balance sheets (Barclay and Smith, 1995; Guedes and Opler, 1996; Berger et al., 2006; Highfield, 2008; Custodio et al., 2013). However, these data only distinguish between debt with maturities below and above one year, missing most of the action within the so-called long-term range. Although others analyze debt issuance data, they usually focus on a single market (typically bonds), neglecting other important debt markets and their potential interactions. By using transaction-level data for more than 5 There is no widely accepted benchmark of what long term is. Most studies addressing issues associated with long-term finance focus on the share of debt with maturity exceeding one year (typically reported on national accounts and on firm and bank balance sheets) and, in some cases, on the share of debt with maturity exceeding five years. 6 Others highlight the role of different institutional environments across countries (Demirgüç-Kunt and Maksimovic, 1999; Bae and Goyal, 2009; Fan et al., 2012). 6 one debt market from around the world, we focus on the supply side by analyzing the borrowing maturity in different markets over time. We show how firms’ use of different markets during crises is reflected in aggregate and firm-level debt maturity at issuance, controlling for the demand side and other factors. Another strand of the literature studies the time-series behavior of short- and long-term debt and maturity, typically focusing on tranquil versus crisis times (Erel et al., 2012; Broner et al., 2013; Chen et al., 2012; He and Milbradt, 2016; Demirgüç-Kunt et al., 2015; Gonzalez, 2015; Bruche and Segura, 2017; Mian and Santos, 2018). On the theoretical side, research tends to highlight supply-side factors (such as risk aversion and liquidity) in explaining the behavior of both sovereign and corporate debt. On the empirical side, studies examine firm balance sheet information or data for a single debt market, finding that debt maturity shortens during crises. However, debt maturity from balance sheet data can shrink because either firms do not issue during crises (hence, the maturity structure of the existing debt shortens) or the new debt issued during crises is shorter term. Moreover, crises do not need to involve all markets. Firms might be able to withstand credit contractions in certain markets by moving to other markets, as we argue in this paper. Therefore, debt maturity will change depending on the relative importance of each market as crises evolve. We contribute to this discussion by analyzing how issuance activity and borrowing maturity react in different markets during crises, and how the main patterns vary across different types of firms and on aggregate. The remainder of the paper is organized as follows. Section 2 describes the data. Section 3 shows how firms have used the growth of primary bond and syndicated loan markets to obtain financing at different maturities. Section 4 investigates how different crises impact debt markets. Section 5 studies how different firms behave during crises. Section 6 concludes. 7 2. Data To conduct the analysis, we assemble a comprehensive transaction-level dataset on newly issued corporate bonds and syndicated loans in domestic and international markets from countries around the world from 1991 through 2014. Our data on firm capital raising activity come from Thomson Reuters’ Security Data Corporation (SDC) Platinum (June 2016 version), which provides information on the characteristics of issuances of publicly and privately placed bonds and syndicated loans with an original maturity of more than one year. 7 Given that SDC Platinum does not collect data on debt issuances with maturities shorter than one year, the dataset does not cover commercial paper. We exclude all financial sector issuances and conduct our analysis on financing to the non-financial sector, which is the primary focus of a large part of the literature. 8 We also exclude all public-sector issuances, comprising issuances by national, local, and regional governments, government agencies, regional agencies, and multilateral organizations. We exclude all mortgage-backed and other asset-backed issuances. Moreover, we exclude countries whose markets are considered offshore financial centers. 9 We end up with 57,950 listed and unlisted non-financial firms and 267,382 debt issuances: 108,184 bond issuances and 159,198 syndicated loan issuances. To study how firms borrow in different markets, we distinguish not only the instrument type (bonds or syndicated loans), but also the market location (domestic or international). For corporate bonds, we compare the market location in which bonds are issued to the issuing firm’s nationality. For offerings that take place simultaneously in more than one market, we consider tranches in each market as separate issuances. The dataset includes 82,249 bond issuances in domestic markets and 7 Around 60 and 38 percent of bonds and syndicated loans, respectively, were issued by unlisted firms. The database also contains information on issuances of common and preferred equity (157,578 equity issuances by 50,131 firms), which we use in Figure 1 to show the relative sizes of primary debt and equity markets. 8 We exclude from the analysis bond and syndicated loan issuers with a Standard Industry Classification (SIC) code between 6,000 and 6,800, which are classified as financial corporations. 9 We exclude offshore financial centers as defined by the IMF at https://www.imf.org/external/NP/ofca/OFCA.aspx. 8 25,935 bond issuances in international markets. For syndicated loans, we compare the nationality of the lead banks that arrange the deal with the issuing firm’s nationality to identify domestic and cross- border bank lending. Domestic loans are those in which only domestic banks lead the syndication, whereas international syndicated loans entail the participation of at least one foreign bank acting as a lead arranger. 10 The dataset includes 83,689 domestic syndicated loans and 75,509 international syndicated loans. We focus on firm size to characterize which firms issue debt at each point in time. We proxy firm size by the average debt transaction size, measured over all debt issuances per firm during the entire sample period. For robustness, we explore two alternative measures of firm size with less coverage over the sample. First, we examine the total assets of the issuers at the time of issuance, which are available in the SDC dataset. However, these data on firm size cover only 42 percent of the corporate bond issuances and 38 percent of the syndicated loan issuances. Second, we restrict the analysis to listed firms and study the total assets reported in their end-of-year balance sheets (from Bureau van Dijk’s Orbis database). Both sets of results are qualitatively similar to the ones reported in the paper. We classify countries in the sample as either developed or developing following the World Bank classification as of 2012. 11 Developed countries are those with a gross national income (GNI) per capita in 2011 above $12,476. All other countries are classified as developing countries. The final dataset comprises 78 countries, 41 of them are developed countries and 37 developing countries. All reported statistics are in U.S. dollars at 2011 constant prices. Appendix Table 1 reports the list of 10 For robustness, we consider two alternative definitions of international syndicated lending: (i) when only foreign banks participate in the deal; or (ii) when more than half of the banks that participate in the deal are foreign. The results are qualitatively similar to the ones reported here. 11 Strictly speaking, as noted in Appendix Table 1, a couple of the “countries” analyzed are economies rather than sovereign states. For ease of exposition, we use throughout the paper the generic term countries. 9 countries as well as the number of bond and syndicated loan issuances and the number of issuers by country. The reported results are qualitatively robust to the exclusion of the largest developed and developing countries from the sample (the United States and China). The paper also analyzes how firms react during market-specific crises. In particular, we examine changes in firm financing and borrowing maturity around the GFC and domestic banking crises. For domestic banking crises, we match our data on corporate debt issuances with data from the Reinhart and Rogoff’s financial crises database (Reinhart and Rogoff, 2011). 12 Due to data limitations, we restrict the analysis on domestic banking crises to the period from 1991 through 2010. After matching the datasets, we obtain a sample that comprises 111,836 debt issuances conducted by 40,561 firms from 43 countries. 3. Debt Issuance and Maturity across Markets The primary debt markets of corporate bonds and syndicated loans have rapidly expanded worldwide since the early 1990s (Figure 1). The total amount of debt issued using these instruments increased more than seven-fold in developed countries and almost twenty-seven-fold in developing countries between 1991 and 2014, reaching $4.7 trillion and $0.83 trillion in 2014, respectively. As a share of GDP, the total amount raised by developed and developing countries in 2014 equaled about 10 and 3.4 percent, respectively, up from 2 and 0.7 percent in 1991. Moreover, the growth in corporate bonds and syndicated loans was substantial when contrasted with that of primary equity markets. Although the growth in capital raising activity through new equity issues was significant, it was relatively subdued when contrasted with the growth in primary debt markets in both developed and developing countries over the sample period (from $171 billion in 1991 to $414 billion in 2014 in developed countries and 12For robustness, we also use the Leaven and Valencia (2012) banking crisis indicators. The results are qualitatively similar to the ones reported here. 10 from around $21 billion to $178 billion in developing countries). Therefore, the ratio of the total amount raised through debt over equity grew from 4 to 11 in developed countries and from 1.5 to 4.6 in developing countries during 1991-2014. Exploiting the rapid expansion in debt markets, the rest of this section describes the extent to which firms issue bonds and syndicated loans in domestic and international markets at different maturities, highlighting how different markets provide different types of financing. 3.1. Corporate Bonds versus Syndicated Loans The (value) weighted average maturity of the debt issued over the entire sample period is 6.6 years. But there is significant heterogeneity across debt instruments and countries. Specifically, corporate bonds provide significantly longer maturities than syndicated loans in developed countries: the weighted average maturity is 10.1 years for corporate bonds and 4.9 years for syndicated loans (Table 1). A similar pattern emerges from the cumulative distribution functions (CDFs) of the issued debt. For every given maturity, the accumulated share of syndicated loans issued with shorter maturities is greater than the accumulated share of bonds issued (Figure 2, Panel A). For example, about 37 percent of the bonds issued have a maturity of 6 years or shorter, in sharp contrast to the 80 percent of the syndicated loans. In developing countries, the differences in the weighted average maturities between bonds and syndicated loans are not as marked as those in developed countries: 7.2 years for corporate bonds versus 7.6 years for syndicated loans. Thus, the CDFs of bonds and syndicated loans issued by firms in these countries are not as different from each other (Figure 2, Panel B). To test whether the differences in the maturity at which firms issue bonds and syndicated loans are statistically significant, we regress the maturity of debt issuances (in years) on a dummy variable that equals one when the debt issuance is a bond and zero otherwise (a syndicated loan). All regressions include either year dummies to control for differences in maturities that reflect changing global market conditions or country-year dummies to control for time-varying country-specific 11 factors. Alternatively, we include firm fixed effects in some specifications to assess whether the maturity of debt issuances also varies consistently across markets when issued by the same firm. Moreover, there is significant variance in the use proceeds from debt issuances, especially in the case of syndicated loans. For example, firms in developing countries have a more intensive use of syndicated loans for project finance (longer-term debt typically funding infrastructure projects) than those in developed countries. 13 Hence, as controls, the regressions also include dummy variables for the different uses of proceeds. The regression estimates confirm that corporate bonds have longer maturity than syndicated loans (Table 2). The differences are statistically significant not only for developed countries, but also for developing countries once country-year and/or firm fixed effects are included in the regressions. For example, the estimates indicate that bond issuances are, on average, 4 to 5 years longer than syndicated loans in developed countries and between 1 and 2 years longer in developing countries, depending on the regression specification. 3.2. Domestic versus International Markets With increasingly integrated financial markets, the distinction between domestic and international debt markets is important. Because foreign capital raising activity is often located in the world’s most developed financial centers, firms (especially those from developing countries) might be able to obtain financing in international markets at terms not available domestically. In fact, most of the international debt issuances take place in a few developed countries. International bonds are typically issued in the euro area (61 percent of all international issuances), the United States (16 percent), or the United Kingdom (8 percent). The largest volumes of loans are also originated in syndicates led by banks from 13Syndicated loans for project finance have an average maturity of 12 years and account for about 35 percent of all syndicated loans contracted by developing countries, but only 5 percent for developed countries. 12 a few developed countries, mainly the United States (31 percent of international issuances) and the economies of Western Europe (36 percent). International markets account for a significant share of debt financing. Developed and developing countries issued internationally 38 percent and 30 percent of the total amount raised through bonds between 1991 and 2014 (Figure 2, Panel C). In syndicated loan markets, most of the financing entails a foreign bank as a lead arranger: over 66 (76) percent of the total amount raised by developed (developing) countries over the same period. Substantial differences in maturity arise when distinguishing the market location where bonds and syndicated loans are issued. Firms from developing countries issue bonds in foreign markets at longer terms than in domestic markets (9.9 years versus 6.1 years on average) (Figure 2, Panel C). The opposite occurs in syndicated loan markets, where firms from developing countries borrow at longer maturities domestically than internationally (11.6 years versus 6.3 years on average). These patters reverse for developed countries, but the differences in maturity across markets are significantly smaller. Bonds issued by firms from developed countries are on average 1.8 years shorter in international than in domestic markets, whereas syndicated loan issuances are slightly longer (by almost 1 year) in international markets compared to domestic ones. Overall, when contrasting the average maturity at issuance in the four markets (domestic and international bonds and syndicated loans), the following pattern emerges. In developed countries, domestic and international bonds have significantly longer maturities (about 4 years or more) than domestic and international syndicated loans. On the other hand, in developing countries, international bonds and domestic syndicated loans are longer-term instruments than domestic bonds and international syndicated loans (about 4 years or more). The CDFs of the debt issuances across the different maturities show a similar pecking order across the different markets (Figure 2, Panels A and B). 13 To assess more formally the statistical significance of the differences in debt maturity across the four different markets, we regress the maturity of debt issuances (in years) on: a dummy variable that equals one for international syndicated loans, a dummy variable that equals one for domestic bonds, and dummy variable that equals one for international bonds. The omitted market is domestic syndicated loans. Akin to the analysis in the previous subsection, we control for country-year and/or firm fixed effects, as well as for the use of proceeds. Because the type of rate (floating or fixed) and the currency denomination of the issuances could explain part of the observed differences in maturities between domestic and international issuances, we also include dummy variables for floating rate debt and for debt issuances denominated in foreign currency. By including these controls, we ensure (among other things) that the finding that developing country firms issue bonds abroad at relatively longer maturity than domestically does not simply reflect that issuances abroad are denominated in foreign currency. This is important because bonds denominated in hard currencies tend to have longer maturities as they do not embed the local currency risk. The estimations confirm that debt issuances have different maturities depending on the market where they take place (Table 3). The regressions for developed countries show that domestic and international bonds have longer maturities (up to 4 years) than domestic and international syndicated loans. The regressions for developing countries consistently show that international bonds and domestic syndicated loans are longer term than domestic bonds and international syndicated loans. The exact pecking order of the different markets varies slightly across specifications. Once we control for the use of proceeds, the terms of the debt issued (rate and currency), and firm fixed effects, international syndicated loans appear as the shortest-term type of instrument, followed by domestic bonds (about 0.8 years longer), domestic syndicated loans (about 1.4 years longer), and international bonds (3 years longer). The specifications with firm fixed effects indicate that the differences in 14 maturity across different markets are not solely driven by the composition of firms, as the differences in maturities are significant within firms across markets. 4. Debt Financing and Maturity during Crises To the extent that different markets provide financing at different maturities, the time-varying capital raising activity in each market and the ensuing changes in debt composition would be reflected in firm- and country-level maturities. In this section, we examine how the relative importance of each market in terms of financing varies around market-specific shocks, as well as its effects on the borrowing maturity at the country level (all firms and all markets), at the market level (domestic and international bonds and syndicated loans), and at the firm level. As shocks, we use the 2008-09 GFC and domestic banking crises. We focus on three sub-periods around the GFC: the “pre-crisis” (2003- 07), the “crisis” (2008-09), and the “post-crisis” (2010-14). We use the 1991-2010 period to analyze domestic banking crises. The assessment of the overall effects of the GFC on developed and developing countries provides a coarse, but conservative set of average effects across these two groups of countries as not all countries within these groups were necessarily hit across markets in the same manner. Domestic banking crises, on the other hand, consist of a more homogenous set of shocks, thus offering a more precise identification of where crises occurred and how firms reacted. 4.1. Composition of Debt Financing during the GFC The GFC suddenly hit hard the banking sector of major developed countries and rapidly spread throughout the global financial system. Although international banks were at the core of the crisis, this unique shock was exogenous to most countries and firms. It affected the volume and composition of new debt financing. In particular, the expansion in debt issuance activity temporarily halted during 2008-09. Whereas the total amount of debt issued grew at an average annual rate of about 18 percent 15 in developed economies during 2003-07, it fell by 38 percent between 2007 and 2008-09. Developing country debt issuance, which grew at a rate of 41 percent per year during 2003-07, declined 20 percent between 2007 and 2008-09. This contraction in primary debt markets reflects two important changes in the composition of the issued debt. 14 First, corporate debt shifted away from bank debt toward bonds. In fact, the standstill in capital raising activity during the GFC was a consequence of the collapse in the syndicated loan markets. The issuance of corporate bonds by non-financial firms increased during this period, partially compensating for the overall decline in syndicated loan financing (Figure 1). For example, total borrowing in syndicated loan markets declined by around 52 and 47 percent between 2007 and 2008- 09 in developed and developing countries, respectively, whereas the issuance of corporate bonds expanded by 20 and 49 percent over the same period. Such a negative correlation between the growth rate in the number of issuances of corporate bonds versus those of syndicated loans is part of a more general pattern. The former typically increased when the latter declined during the 2003-2014 period (Figure 3). Second, in developing countries the composition of corporate debt also shifted away from international toward domestic markets during the GFC. Cross-border syndicated loans to developing countries declined by 61 percent between 2007 and 2008-09, whereas the issuance of domestic syndicated loans increased by 80 percent over the same period. Domestic syndicated loans did not fully compensate for the decline in cross-border lending, partly because they constituted only a small fraction of the total financing volume. Analogous patterns occurred in corporate bond markets. 14For robustness, following Adrian et al. (2013), we alternatively restrict the sample of debt transactions to the financing of “real activities.” That is, we exclude bonds and syndicated loans for “acquisition financing and LBO,” “refinancing and capital structure management,” and other issuances whose purposes cannot be categorized as real investment. Overall, of the 52,473 corporate bonds and 118,944 syndicated loans issued during 2003-14, about 83 percent of the bond issuances and 56 percent of the syndicated loan issuances were for real investment purposes. The results based on this restricted sample are qualitatively similar to the reported ones. 16 Domestic bond issuances increased by 105 percent between 2007 and 2008-09, compensating for a decline of 35 percent in foreign bond markets over the same period. That is, a more dynamic domestic bond market activity underpinned the overall increase of bond issuances by developing countries during the GFC. Overall, as international markets were hit by the shock to global banking, debt issuance activity for developing countries moved toward domestic bond and syndicated loan markets. In contrast, developed countries witnessed a decline in the issuance of both domestic and international syndicated loans, in parallel with an increase in both domestic and international bond issuances. From these aggregate patterns, it is difficult to ascertain whether the changes in debt composition during the GFC were driven by within-firm market substitutions, compositional changes in the set of firms raising new debt, or both. It is also challenging to determine whether such market switches were caused by shifts in the supply- or demand-side of capital. 15 To address these issues, we analyze whether the reported shifts in capital raising activity across different debt markets also occurred at the firm level. Our methodology is similar to that in Adrian et al. (2013) and Becker and Ivashina (2014). Specifically, we focus on the set of firms with a revealed demand for debt during the GFC and study their choice of debt instrument. A move away from the crisis-hit market to demanding capital in a different market would be consistent with a negative credit supply shock in the former. To conduct the analysis, we estimate discrete choice logit models focusing on the firms’ decisions to issue bonds versus syndicated loans in domestic or international markets in the aftermath of the GFC vis-à-vis the pre-crisis period (2003-2007). The alternative dependent variables are: (i) a debt choice indicator that equals one if a firm issues a bond and zero if it issues a syndicated loan in a given quarter; or (ii) a debt choice indicator that equals one if a firm issues debt domestically and zero 15 Several papers document the contraction in syndicated loans in developed countries during the GFC. The research points to a combination of demand- and supply-side shocks: firms scaled back expansion plans during the recession that followed the GFC, reducing the demand for syndicated loans, while banks dealt with deleveraging pressures and tightened regulations that pushed down the supply curve (Ivashina and Scharfstein, 2010; Chui et al., 2010). 17 if it issues debt abroad in a given quarter for a given debt instrument. 16 For quarters in which a firm has issuances of both bonds and syndicated loans, we set the debt choice indicator equal to one if the amount raised through bonds exceeds that raised through syndicated loans. A similar rule is applied for the domestic versus international issuances within bond and syndicated loan markets. Logit regressions with country fixed effects estimate changes in the value of debt transactions within each country, whereas those with firm fixed effects provide evidence on whether individual firms switched between bonds and syndicated loans (or between domestic and international markets) in the aftermath of the GFC. The use of firm fixed effects addresses the concern about compositional changes in the set of firms raising debt financing. The regression results indicate that the issuance of corporate bonds relative to syndicated loans increased during the GFC in both developed and developing countries (Table 4, Panel A). For example, controlling for country fixed effects, the propensity to issue bonds relative to syndicated loans increased by around 3 percentage points (p.p.) in developed countries and by 5 p.p. in developing countries during 2008-09 relative to the pre-crisis average. The substitution away from syndicated loans toward bonds also took place within firms, indicating that the patterns observed are not solely driven by different firms issuing different types of debt at different points in time. For example, conditional on firms’ actual debt issuance, the probability of issuing bonds to obtain new financing (relative to syndicated loans) increased by 5 p.p. for firms in developed countries and by 10 p.p. for firms in developing countries during the GFC period relative to the pre-crisis. These changes in the probabilities of issuance imply a 21 percent increase in the bond-syndicated loan debt ratio during the crisis vis-à-vis the pre-crisis average for firms in both developed and developing countries. 16 For these estimations, we aggregate the transaction-level data to the firm-quarter level. Firm-quarter observations without capital raising activity are excluded from the sample by construction. For robustness, we also conduct the analysis by aggregating transaction-level data at the firm-monthly level and at the firm-daily level. The results are qualitatively similar to the ones reported here. 18 Furthermore, there is some persistence in these effects as the propensity to issue bonds with respect to syndicated loans in the post-crisis period (2010-2014) also increased compared to the pre-crisis period. 17 The estimations also provide evidence of an increase in the use of domestic markets relative to international markets during the GFC, especially in developing countries (Table 4, Panels B and C). For these countries, the probability of issuance in domestic bond and syndicated loan markets relative to international ones increased by 9 p.p. and 31 p.p. during the GFC compared to the pre-crisis period. The within-firm estimations for developing country firms also show market movements away from international markets toward domestic ones, indicating that the estimations at the country level are not driven entirely by shifts in the composition of issuing firms. For example, conditional on firms’ issuing corporate bonds during the GFC, the probability of issuing domestically (relative to internationally) increased by 25 p.p., a 35 percent increase over the pre-crisis probability. Similarly, the within-firm estimations show a 22 p.p. increase in the probability of firms issuing syndicated loans domestically rather than internationally during the GFC relative to the pre-crisis period. For developing countries, most of these changes in issuance probabilities persisted through the post-crisis period of 2010-14. For developed countries, the regression results also indicate an overall increase in the propensity to issue in domestic debt markets, but this movement did not take place within firms. Overall, these results are consistent with a negative shock in the supply of credit from global banks active in international syndicated loan markets during the GFC, prompting firms with positive 17 During 2011-13, the shock caused by the GFC could have intensified because of the financial stress experienced by European banks due to the sovereign debt crisis affecting several European countries (Feyen and Gonzalez del Mazo, 2013; Laeven and Tressel, 2014). Moreover, non-conventional monetary policies seem to have prompted the issuance of corporate bonds worldwide during the post-crisis period (Shin, 2013; Lo Duca et al., 2016). Both types of events might be behind the increasing propensity to issue bonds with respect to syndicated loans during the post-crisis period. 19 funding needs to move away from the affected banks toward alternative markets. 18 The ensuing new debt financing would thus move away from syndicated loans toward bonds in developed countries, and from international syndicated loans toward both bonds and domestic syndicated loans in developing countries. These movements across markets occurred both at the aggregate and firm level. 4.2. Debt Maturity during the GFC Building on the finding that the relative importance of each market for firm financing changed during the GFC, we examine how these compositional changes in debt financing are reflected in firm- and country-level maturities. Starting with aggregate patterns, the weighted average debt maturity remained relatively stable (even increased in developing countries) during the GFC (Figure 4, Panel A). This aggregate debt maturity reflects over-time variations in both maturity within each individual debt market and debt composition across markets (within and across firms). Hence, underlying these trends are several distinct patterns in debt markets in developed and developing countries. In developed countries, the average maturity of debt at issuance in both bond and syndicated loans markets (domestically and abroad) declined during the GFC (Figure 4, Panels B and C). However, the share of debt financing raised with bonds (which have significantly longer maturity than syndicated loans) increased. The two effects compensated for each other, yielding a fairly constant average debt maturity when comparing the pre-crisis and crisis periods. Moreover, these effects were short lived and reversed in the post-crisis period to their pre-crisis levels. 18 Erel et al. (2012) propose an alternative explanation for the counter-cyclical pattern of bond issuance. They argue that, in times of credit tightening, bond financing becomes more attractive to some high-quality firms as investors display “flight to quality.” However, Becker and Ivashina (2014) argue that the increasing bond financing during periods of credit tightening occurs without a decline in the price of bonds relative to syndicated loans, which would be necessary for this alternative explanation. Therefore, the increase in bond issuances relative to syndicated loans during the crisis seems more consistent with firms substituting one type of debt financing for another, rather than just taking advantage of cheaper bond markets. The new evidence of firms moving away from international markets toward domestic markets in developing countries can also be interpreted as a retrenchment by global banks. 20 In developing countries, the average maturity of the total debt issued increased in the aftermath of the GFC. Whereas the maturity of bonds declined (by about 2 years), the maturity of syndicated loans increased (by about 3 years) during the GFC. The decline in bond maturity occurred in both domestic and international markets (Figure 4, Panel B). For syndicated loans, however, the greater propensity to issue in the longer-term domestic markets explains the increase in maturity at issuance (Figure 4, Panel C). Again, the movement toward the markets with longer maturity during the GFC explains the increase in aggregate borrowing maturity. To formally disentangle the effects on debt maturity at issuance driven by changes in market composition from within-market effects, we estimate panel regressions of debt maturity during 2003- 14 on a dummy variable that equals one for the crisis period of 2007-08 (and zero otherwise) and another dummy variable that equals one for the post-crisis period of 2009-14 (and zero otherwise). We estimate these regressions both across and within debt markets, including either country fixed effects (to study within-country patterns) or firm fixed effects (to study within-firm variations). The estimations for developed countries are consistent with the aggregate patterns displayed in the figures. The results show that, even though the maturity of issuances of bonds and syndicated loans shortened during the GFC relative to the pre-crisis period, the overall debt maturity did not significantly change (Table 5, Panel A). 19 For example, the estimates with country fixed effects show that corporate bond and syndicated loan maturities declined by around 15 and 10 months during the GFC, respectively. However, the composition of debt shifted toward the longer-term bond financing away from syndicated loans. The share of bonds in total debt increased from 22 percent in the pre- crisis period to 37 percent during the GFC period (Table 5, Panel B). Consequently, the overall debt 19For simplicity, the results reported in Table 5 show the estimations for total debt, bonds, and syndicated loans. Appendix Table 2 shows the estimations when distinguishing between domestic and international markets. The results show that the maturity of bond and syndicated loan issuances shortened in both domestic and international markets during the GFC relative to the pre-crisis period. 21 maturity did not significantly shorten over this period. The regressions with firm fixed effects show similar patterns. Debt maturity remained stable within firms in developed countries, even though the maturity of bond and syndicated loan issuances declined 10 and 8 months between 2003-07 and 2008- 09. The regression estimates for developing countries also show that borrowing maturities remained stable during the GFC vis-à-vis the pre-crisis period (Table 5). The stability in the overall debt maturity is explained by changes in the issuance composition between bonds and syndicated loans and between domestic and international markets. As firms moved away from syndicated loans toward bonds, which have higher average maturity, overall debt maturity (based on the total amount of debt issued) remained stable during the GFC, even though the maturity within bond markets declined by an average of 13 months. Furthermore, the estimates with country fixed effects show that syndicated loan maturities increased during the GFC. This effect is the result of a higher propensity to use domestic (rather than international) markets, which are of longer maturity. 20 Indeed, the share of domestic syndicated loans in total debt issued increased from 6 percent during the pre-crisis years to 17 percent during the GFC, whereas the share of international syndicated loans declined from 63 to 32 percent over the same period. The within-firm estimations also show a stable overall average debt maturity amid significant declines in corporate bond maturities. 4.3. Domestic Banking Crises Expanding the previous analysis, we examine the time-varying capital raising activity in each debt market and the ensuing changes in both debt composition and maturity during domestic banking crises. To study the firms’ decisions to issue bonds versus syndicated loans in domestic and international markets following domestic banking crises, we estimate discrete choice logit models for 20 Appendix Table 3 reports estimations distinguishing between domestic and international issuances. 22 the 1991-2010 period using a methodology analogous to that employed for the analysis of the GFC. 21 The dependent variables are: (i) a debt choice indicator that equals one if a firm issues a bond and zero if it issues a syndicated loan in a given quarter; or (ii) a debt choice indicator that equals one if a firm issues debt domestically and zero if it issues debt abroad in a given quarter for a given debt instrument. 22 We estimate these logit regressions with year dummies and either country fixed effects or firm fixed effects. Most of the domestic banking crises in developing countries occurred during the 1990s, when the capital raising activity by this group of countries was much scarcer. Hence, we pool together developed and developing countries in the estimations. 23 The regression results show that the issuance of corporate bonds relative to syndicated loans increased during domestic banking crises vis-à-vis non-crisis periods (Table 6, Panel A). For example, controlling for country fixed effects, the propensity to issue bonds relative to syndicated loans during domestic banking crises increased by 14 p.p. relative to non-crisis periods, implying a 34 percent increase in the bond-loan debt ratio. This substitution also takes place within firms, indicating that these patterns are not entirely driven by changes in the composition of issuing firms. For example, the within-firm probability of issuing bonds (relative to syndicated loans) to obtain new debt financing increased by 15 p.p. during domestic banking crises (a 37 percent increase in the bond-syndicated loan debt ratio). The estimations also show an increase in the propensity to raise capital through bonds and syndicated loans in international markets rather than domestically during domestic banking crises. For example, the probability of issuing syndicated loans in domestic markets relative to international 21 In this case, non-crisis periods comprise all country-years without domestic banking crises. 22 Following the same methodology of the previous estimations, for quarters in which a firm has issuances of both bonds and syndicated loans, we set the debt choice indicator equal to one if the amount raised through bonds exceeds that raised through syndicated loans. A similar rule is applied for the domestic versus international issuances within bond and syndicated loan markets. 23 The sample for the two sets of countries together entails 111,836 debt issuances by 40,561 firms from 43 countries. 23 markets declined 18 p.p. during domestic banking crisis periods relative to non-crisis periods. Once more, the within-firm estimations show market movements away from domestic bonds and syndicated loans toward international ones. Some countries experienced domestic banking crises during the GFC period (2008-09). 24 For robustness, we conduct two alternative estimations: (i) we restrict the sample to the 1991-2007 period, thus excluding the GFC and post-GFC periods; and (ii) we exclude from the analysis countries that experienced a domestic banking crisis during the GFC. Both analyses provide qualitatively similar estimates on the switches from syndicated loans to bonds and from domestic to international syndicated loans (Table 6, Panels B and C). Overall, these results are consistent with a negative shock in the supply of domestic credit, causing firms to change the composition of their new debt financing away from the affected domestic banking system. The ensuing new debt financing composition thus reflects movements away from domestic syndicated loans toward international ones and toward bond markets. Importantly, these switches during domestic banking crises went in the opposite direction than the ones shown for the GFC. During domestic banking crises, firms moved away from the domestic banks, whereas during the GFC, firms moved away from international banks in crisis. As in the case of the GFC, the compositional effects we observe in debt financing around domestic banking crises are reflected in firm- and country-level debt maturities. To disentangle the effects on debt maturity at issuance driven by changes in the debt market composition from the within- market effects, we conduct an analysis analogous to that in Section 4.2. We estimate panel regressions of debt maturity during 1991-2010 on a dummy variable for periods of domestic banking crises (equal 24According to Reinhart and Rogoff (2009), nine countries experienced domestic banking crises during the GFC: Iceland, the Netherlands, Portugal, the Republic of Ireland, the Russian Federation, Spain, Switzerland, the United Kingdom, and the United States. 24 to zero during tranquil times), year dummies, and either country or firm fixed effects. The results show that overall debt maturity increased slightly (about 4-5 months) during domestic banking crises relative to non-crisis periods, even though the maturity of syndicated loans declined over the same period (Table 7). 25 5. Differences across Firms The previous section shows that some firms actively change their debt financing composition when crises hit specific markets, mitigating the negative shock. In this section, we examine these “switchers” versus “non-switchers” (the rest of the firms). 26 In particular, we study their size and the evolution of their market share. We then compare differences in their borrowing behavior and financing conditions. Lastly, we analyze the extent to which firm movements across markets have implications for the composition of firms obtaining financing during crises. Switching issuers are large corporations capturing a substantial fraction of the capital raising activity. These firms issued about 66 percent of the total debt raised in developed countries and 46 percent in developing countries before the GFC (during 2003-07). But they were relatively few as they accounted for only 26 percent and 19 percent of the number of issuing firms in developed and developing countries over the same period (Table 8, Panels A and B). Similarly, switching issuers in 25 Appendix Table 4 reports estimations distinguishing across the four debt markets. The results show that the maturity of domestic syndicated loans declined during banking crises, while domestic and international bond maturities remained stable. 26 For the GFC, we define switching issuers as the firms that changed their debt issuance composition in 2008-2014 relative to 2003-2007 (as identified by the logit regressions with firm fixed effects in Table 4). Non-switching issuers are the rest of the firms in the sample, comprising firms that only issued debt up to 2008 and firms that issued the same type of debt before and after the GFC. For domestic banking crises, we adopt an analogous definition. Switching issuers are the firms that changed their debt issuance composition during crisis years relative to non-crisis years (as identified by the logit regressions with firm fixed effects in Table 6). Non-switching issuers are the rest of the firms in the sample and encompass two types of firms: those that issued the same kind of debt before and during domestic banking crises and those that only issued debt in non-crisis periods. By construction, switching issuers issued in more than one debt market over the sample period, whereas the vast majority of non-switchers (86 and 96 percent of the non-switching firms for the GFC and domestic banking crises) issued in only one debt market over the full sample period. 25 domestic banking crises captured 75 percent of the total debt issued, but accounted for 24 percent of total issuers during non-crisis periods (Table 8, Panel C). The firm size distribution (FSD) of switching issuers during the GFC is to the right of the FSD of non-switchers (Figure 5, Panel A). 27 The same pattern applies to the FSD of switching and non-switching firms during domestic banking crises. In fact, the median size of switching issuers ($287 million) is more than three times larger than that of non-switchers (Figure 5, Panel D). Overall, the largest firms were the ones that switched markets, tapping different sources during the GFC and domestic banking crises. The differences in size between switching and non-switching issuers seem to come, to a large extent, from the fact that firms with access to international markets are substantially larger than those accessing domestic markets. Non-switching firms accessing domestic syndicated loan or domestic bond markets have a median size of $55 and $59 million, whereas the size of the median international syndicated loan and bond issuer is $190 and $160 million, respectively. Indeed, most switching firms have access to international markets: 88 percent of these firms issued debt internationally at least once over the sample period. Thus, it is not surprising that the FSD of switching firms is similar to the FSD of non-switching issuers in international markets (Figure 5, Panel B). In contrast, switching issuers comprise substantially larger firms than non-switching domestic issuers. The FSD of switching firms during crises is to the right of the size distribution of non-switchers in domestic bond markets and domestic syndicated loan markets, both of which have fatter left tails (Figure 5, Panel C). Differences between switching and non-switching issuers are also observed in their borrowing behavior and financing conditions. Switching issuers were capable of partly compensating for the decline in financing during crises by moving across markets, which also affected their borrowing 27 Similar patterns are observed when splitting the sample between developed and developing countries. 26 maturity. Non-switching issuers often experienced significant declines in financing and, if able to obtain credit, shorter maturities. Regarding the dynamics of borrowing maturities, the patterns for switching issuers are in line with the overall results previously presented. That is, the debt maturity at issuance for switching issuers remained relatively stable during crises. In contrast, the borrowing maturity declined on average for non-switchers (Figure 6). Panel regressions with year and firm fixed effects confirm these patterns. 28 The estimations for developed countries show that the average maturity of total debt issued by switching firms remained stable during the GFC vis-à-vis the pre-crisis period, despite declines in borrowing maturities in both bond and syndicated loan markets (Table 9, Panel A). Similarly, the overall debt maturity for switching firms remained stable during the GFC in developing countries relative to the pre-crisis period, even though corporate bond maturities largely shortened (Table 9, Panel B). Non-switching issuers, on the other hand, tended to experience declining debt maturities during the GFC period, akin to the patterns for each individual debt market. During domestic banking crises, the overall debt maturity at issuance increased for switching firms, while it declined for non- switching issuers of syndicated loans (Table 9, Panel C). Regarding the use of debt markets for new financing, non-switching issuers were hit harder than switching issuers by the overall credit decline during crises. For example, the average number of debt issuances per year by non-switching firms declined by 57 percent and 49 percent in developed and developing countries during the GFC (Figure 6, Panels A and B). In contrast, the number of debt issuances by switchers (which moved away from shock-hit markets toward alternative markets) declined only by 12 percent in developed countries and increased by 7 percent in developing countries during the GFC. A similar, though less marked, pattern takes place during domestic banking crises. 28 Regressions with country fixed effects yield results that are qualitatively similar to the ones reported here. 27 Debt issuances by switchers and non-switchers declined by 15 percent and 22 percent during domestic banking crises relative to non-crisis periods (Figure 6, Panel C). The reason why new financing for non-switching firms was relatively less affected during domestic banking crises than during the GFC is that, during the former, only non-switchers in domestic markets experienced declining financing (Table 10). Namely, non-switching international issuers increased their issuance activity abroad, possibly avoiding the traditional banking system (Table 10, Panel C). During the GFC, in contrast, non-switching firms experienced lower debt financing not only in the international markets hit by the crisis, but also in the domestic markets where the total amount of credit went up (domestic bond markets in developed countries and domestic bond and syndicated loan markets in developing countries, Table 10, Panels A and B). In fact, the largest decline in financing for non-switchers in developing countries occurred within domestic syndicated loan markets. Meanwhile, the total volume of domestic syndicated loans increased, driven by new financing directed toward the switching issuers moving into that market (Table 10, Panel B). 29 The patterns of declining financing for non-switching issuers during the GFC are consistent with crowding out effects caused by the move of switching issuers toward domestic markets. Increasing funding for the relatively larger switching firms was accompanied by less funding for the relatively smaller non-switching firms borrowing in domestic debt markets. In fact, the financing share of non-switching firms had its largest decline precisely in those domestic markets that expanded during the crisis (Figure 7). The financing share of non-switching issuers in domestic syndicated loan (bond) markets in developing countries fell from 58 (63) percent during 2003-07 to 9 (45) percent during the GFC. For developed countries, the financing share of non-switching firms in domestic bond markets fell from 21 to 4 percent over the same period. We do not observe analogous effects during domestic 29This is the case even when focusing only on the firms that showed an actual demand for credit by issuing domestic syndicated loans before and during the GFC. 28 banking crises, when switching firms moved toward international markets, dominated in general by large firms. Although the evidence does not rule out the possibility of reduced demand for credit by non-switching domestic firms during the GFC, this had to occur precisely when domestic banking systems in developing countries expanded and directed their funds toward large international firms demanding domestic debt. Namely, non-switchers in domestic markets with a positive demand for credit during the GFC had to compete with larger switching issuers coming back home. 6. Conclusions This paper provides an assessment of how the use of different debt markets impacts corporate debt financing and maturity. We analyze corporate debt dynamics across four different debt markets (domestic and international bonds and syndicated loans), during both tranquil and financial crisis periods at home and abroad. We focus on (i) the differences in financing and maturity across markets over time, (ii) how firms switch markets when faced with negative shocks, and (iii) the effects of those switches on firm financing and debt maturity. We find that different markets provide financing at distinct maturities, even for the same firms. Corporate bonds have longer average maturity than syndicated loans, and for developing countries international bonds and domestic syndicated loans are longer term than other debt markets. Differences in the type of instruments issued and market location partly explain differences in debt maturity at the firm and country level. The relative importance of each market over time also has significant effects on debt composition and maturity. During the GFC and domestic banking crises, some firms switched away from the most affected markets (those with shorter maturity in our analysis), compensating for the decline in financing and for the shortening maturity within markets. Thus, debt maturity remained stable during crises at the country level and for those switching firms. But firms that did not switch markets during crises (the non-switchers) typically experienced declining 29 debt financing and maturity. In fact, non-switchers during the GFC obtained declining debt financing in domestic markets, even when total domestic financing increased. These patterns are consistent with crowding out effects caused by the larger switching firms returning home. Overall, the movements across markets of firms with demand for credit are consistent with credit contractions during crises due to supply-side shocks, with significant effects on debt composition, borrowing maturity, and credit redistribution across firms of different sizes. The findings in this paper have implications for different discussions related to long- and short-term borrowing and debt markets more generally. First, our paper shows that firms often obtain financing from different sources, highlighting the importance of analyzing the different types of financing jointly. Whereas the study of domestic and foreign bond and syndicated loan markets is by no means complete (firms have other financing options), this paper shows that firms could use these markets as complements to the extent that they provide financing at different terms. Moreover, having access to different markets might allow firms to compensate for fluctuations in particular markets by raising funds elsewhere, as observed during banking crises when bond markets provide some of the “spare tire” function advocated for capital markets (Greenspan, 1999; Levine, et al., 2016). This paper is the first to provide evidence of firm-level debt substitution during crises at the global level across four different debt markets, showing its impact on the financing and maturity of firms and countries. These effects are difficult to gauge when studying debt dynamics within a single market or when analyzing balance sheet information. Second, understanding the degree of long-term financing is also relevant for the discussions on the possible risks related to the recent expansion in debt markets. The findings in this paper suggest that long-term issuances might have mitigated the increasing leverage and exposure to foreign currency risks discussed in the literature post GFC, at least for the firms that could issue at those longer maturities. But the trade-offs of different types of debt deserve more work going forward. 30 Third, the policy options on access to finance and crisis management might also need some reconsideration. By analyzing a wider set of debt markets, we find some compensation in the volume of financing across markets and stable borrowing maturities at the firm and country level during crises, even though the maturity of new financing in individual markets typically declines. But the aggregate results must be treated with caution because there is substantial heterogeneity across firms. On the one hand, relatively larger firms seem more capable of cushioning the consequences of crises as they seem able to move abroad when there is a domestic shock or move back home when the crisis is abroad. On the other hand, firms issuing in domestic markets only (typically smaller firms) face declining financing and maturities, and can be crowded out by larger firms returning to domestic markets during shocks in foreign markets. In fact, these crowding out effects could be larger than the ones suggested by the evidence in this paper if we considered a wider set of markets and firms. Hence, policy makers might want to focus on the constrained firms when implementing policies aimed at mitigating the effects of shocks. These policies might be focused not only on promoting more complete markets, but also on providing liquidity to the specific markets and firms under stress. 31 References Acharya, V., Cecchetti, S., De Gregorio, J., Kalemli-Ozcan, Ş., Lane, P., Panizza, U., 2015. “Corporate Debt in Emerging Economies: A Threat to Financial Stability?” Committee for International Policy Reform, Brookings Institution. Adrian, T., Colla, P., Shin, H.S., 2013. “Which Financial Frictions? Parsing Evidence from the Financial Crisis of 2007–09.” NBER Macroeconomics Annual 2012 27:159–214. Edited by D. Acemoglu, J. Parker, and M. Woodford. Almeida, H., Campello, M., Laranjeira, B., Weisbenner, S., 2011: “Corporate Debt Maturity and the Real Effects of the 2007 Credit Crisis.” Critical Finance Review 1(1): 3–58. Bae, K.H., Goyal, V.K., 2009. “Creditor Rights, Enforcement, and Bank Loans.” Journal of Finance 64(2): 823–860. Barclay, M., Smith, C.W., 1995. “The Maturity Structure of Corporate Debt.” Journal of Finance 50(2): 609–31. Beck, T., 2016. Long-term Finance in Latin America. A Scoreboard Model. Discussion Paper IDB- DP-476. Becker, B., Ivashina, V., 2014. “Cyclicality of Credit Supply: Firm-level Evidence.” Journal of Monetary Economics 62(C): 76–93. Bekaert, G., Harvey C.R., Lundblad, C.T., Siegel, S., 2011. “What Segments Equity Markets?” Review of Financial Studies 24(12): 3842–3890. Beltratti, A., Stulz, R., 2012. “The Credit Crisis around the Globe: Why Did Some Banks Perform Better?” Journal of Financial Economics 105(1): 1–17. Berger, A.N., Espinosa-Vega, M.A., Scott Frame, W., Miller, H.N., 2006. “Debt Maturity, Risk, and Asymmetric Information.” Journal of Finance 60(6): 2895–2923. Broner, F., Lorenzoni, G., Schmukler, S., 2013. “Why Do Emerging Economies Borrow Short Term?” Journal of the European Economic Association 11(1): 67–100. Bruche, M., Segura, A., 2017. “Debt Maturity and the Liquidity of Secondary Markets.” Journal of Financial Economics 124(3): 599–613. Brunnermeier, M., 2009. “Deciphering the Liquidity and Credit Crunch 2007–2008.” Journal of Economic Perspectives 23(1): 77–100. Brunnermeier, M., Oehmke, M, 2013. “The Maturity Rat Race.” Journal of Finance 68(2): 483–521. Bruno, V., Shin, H.S., 2016. “Global Dollar Credit and Carry Trades: A Firm-level Analysis.” Review of Financial Studies, 30(3): 704–749. Caballero, J., Fernandez, A., Park, J., 2016. “On Corporate Borrowing, Credit Spreads and Economic Activity in Emerging Economies: An Empirical Investigation.” IADB Working Paper Series No. IDB-WP-719. Calomiris, C.W., Kahn, C.M., 1991. “The Role of Demandable Debt in Structuring Optimal Banking Arrangements.” American Economic Review 81(3): 497–513. Cerutti, E., Hale, G., Minoiu, C., 2015. “Financial Crises and the Composition of Cross-Border Lending.” Journal of International Money and Finance 52: 60–81. Chen, H., Xu, Y., Yang, J., 2012. “Systematic Risk, Debt Maturity, and the Term Structure of Credit Spreads.” NBER Working Paper 18367. 32 Chui, M., Domanski, D., Kugler, P., Shek, J., 2010. “The Collapse of International Bank Finance during the Crisis: Evidence from Syndicated Loan Markets.” BIS Quarterly Review: 39–49. Cortina, J.J., Didier, T., Schmukler, S., 2018. “Corporate Debt Maturity in Developing Countries: Sources of Long- and Short-termism.” The World Economy, forthcoming. Custodio, C., Ferreira, M.A., Laureano, L., 2013. “Why Are U.S. Firms Using More Short-Term Debt?” Journal of Financial Economics 108(1): 182–212. Dangl, T., Zechner, J., 2016. “Debt Maturity and the Dynamics of Leverage.” CFS Working Paper Series 547. Demirgüç-Kunt, A., Maksimovic, V., 1999. “Institutions, Financial Markets, and Firm Debt Maturity.” Journal of Financial Economics 54(1): 295–336. Demirgüç-Kunt, A., Martinez-Peria, M.S., Tressel, T., 2015. “The Impact of the Global Financial Crisis on Firms’ Capital Structure.” World Bank Policy Research Working Paper 7522. Diamond, D., 1991. “Debt Maturity and Liquidity Risk.” Quarterly Journal of Economics 106(1): 709–37. Didier, T., Levine, R., Schmukler, S., 2015. “Capital Market Financing, Firm Growth, and Firm Size Distribution.” NBER Working Paper 20336. Eichengreen, B., Hausmann, R., 1999. “Exchange Rates and Financial Fragility.” Federal Reserve Bank of Kansas City Proceedings: 329–368. Erel, I., Julio, B., Kim, W., Weisbach, M.S., 2012. “Macroeconomic Conditions and Capital Raising.” Review of Financial Studies 25(2): 341–376. Fan, J., Titman, S., Twite, G., 2012. “An International Comparison of Capital Structure and Debt Maturity Choices.” Journal of Financial and Quantitative Analysis 44(1): 23–56. Feyen, E., Gonzalez del Mazo, I., 2013. “European Bank Deleveraging and Global Credit Conditions.” World Bank Policy Research Working Paper 6388. Flannery, M.L., 1986. “Asymmetric Information and Risky Debt Maturity Choice.” Journal of Finance 41(1): 18–38. G20 (Group of 20), 2013. “Long-term Investment Financing for Growth and Development;” Umbrella Paper. Giovannini, A., Mayer, C., Micossi, S., di Noia, C., Onando, M., Pagano, M., Polo, A., 2015. “Restarting European Long-Term Investment Finance: A Green Paper Discussion Document.” Centre for Economic Policy Research. Gonzalez, V.M., 2015. “The Financial Crisis and Corporate Debt Maturity: The Role of Banking Structure.” Journal of Corporate Finance 35(1): 310–328. Greenspan, A., 1999. “Do Efficient Financial Markets Mitigate Financial Crises?” Financial Markets Conference of the Federal Reserve Bank of Atlanta. Greenwood, R., Hanson, S., Stein, J., 2010. “A Gap-Filling Theory of Corporate Debt Maturity.” Journal of Finance 65(3): 993–1028. Guedes, J., Opler, T., 1996. “The Determinants of the Maturity of Corporate Debt Issues.” Journal of Finance 51(5): 1809–1833. Hale, G., 2007. “Bonds or Loans? The Effect of Macroeconomic Fundamentals.” The Economic Journal 117(516): 196–215. Hale, G., Jones, P., Spiegel, M., 2016. “The Rise in Home Currency Issuance.” Federal Reserve Bank of San Francisco Working Paper Series 2014-19. 33 He, Z., Milbradt, K., 2016. “Dynamic Debt Maturity.” Review of Financial Studies 29(10): 2677–2736. Highfield, M.J., 2008. “On the Maturity of Incremental Corporate Debt Issues.” Quarterly Journal of Finance and Accounting 47(2): 45–67. International Monetary Fund (IMF), 2015. “Corporate Leverage in Emerging Markets, A Concern?” Global Financial Stability Report. Ivashina, V., Scharfstein, D., 2010. “Bank Lending during the Financial Crisis of 2008.” Journal of Financial Economics 97(1): 319–38. Japelli, T., Pagano, M., 2008. “Financial Market Integration under EMU.” European Economy Economic Papers 312. Jeanne, O., 2009. “Debt Maturity and the International Financial Architecture.” American Economic Review 99(5): 2135–48. Kashyap, A., Stein, J., Wilcox, D., 1993. “Monetary Policy and Credit Conditions: Evidence from the Composition of External Finance.” American Economic Review 83(1): 78–98. Karolyi, G.A., Stulz, R.M., 2003. “Are Financial Assets Priced Locally or Globally? In Handbook of the Economics of Finance. Edited by G. Constantinides, M. Harris, and R.M., Stulz. Laeven, L., Tressel, T., 2014. “Fragmentation of the Financial System.” In From Fragmentation to Financial Integration in Europe. Edited by C. Enoch, L. Everaert, T. Tressel, and J. Zhou. Leaven, L., Valencia, F., 2012. “Systemic Banking Crisis Database: An Update.” IMF Working paper 08/224. La Porta, R., Lopez-De-Silanes, F., Shleifer, A., Vishny, R.W., 1997. “Legal Determinants of External Finance.” Journal of Finance 52(3): 1131–1150. Leary, M., 2009. “Bank Loan Supply, Lender Choice, and Corporate Capital Structure.” Journal of Finance 64(3): 1143–1185. Levine, R., Lin, C., Xie, W., 2016. “Spare Tire? Stock Markets, Banking Crises, and Economic Recoveries.” Journal of Financial Economics 120(1): 81–101. Lo Duca, M., Nicoletti, G., Martinez-Vidal, A., 2016. “Global Corporate Bond Issuance: What Role for U.S. Quantitative Easing?” Journal of International Money and Finance 60(1): 114–150. Mauer, D., Lawellen, W., 1987. “Debt Management under Corporate and Personal Taxation.” Journal of Finance 42(5): 1275–1291. McCauley, R., McGuire, P., Sushko, V., 2015. “Global Dollar Credit: Links to U.S. Monetary Policy and Leverage.” Economic Policy 30(82): 187–229. Mian, A., Santos, J.A., 2018. “Liquidity Risk and Maturity Management over the Credit Cycle.” Journal of Financial Economics 127(2): 264–284. Myers, S., 1977. “Determinants of Corporate Borrowing,” Journal of Financial Economics 5(1): 147–76. Oliner, S., Rudebusch, G.D., 1995. “Is There a Bank Lending Channel for Monetary Policy?” Federal Reserve Bank of San Francisco Economic Review, 1995(2): 290–303. Oliner, S., Rudebusch, G.D., 1996. “Monetary Policy and Credit Conditions: Evidence from the Composition of External Finance: Comment.” American Economic Review 86(1): 300–309. Pirinsky, C., Wang, Q., 2006. “Does Corporate Headquarters Location Matter for Stock Returns?” Journal of Finance 61(4): 1991-2015. Raddatz, C., 2010. “When the Rivers Run Dry: Liquidity and the Use of Wholesale Funds in the Transmission of the U.S. Subprime Crisis.” World Bank Policy Research Working Paper 5203. 34 Reinhart, C., Rogoff, K., 2011. This Time It’s Different: Eight Centuries of Financial Folly. Princeton University Press. Rodrik, D., Velasco, A., 2000. “Short-Term Capital Flows.” In Annual World Bank Conference on Development Economics 200. Edited by B. Pleskovic and J. Stiglitz. Shin, H.S., 2013. “The Second Phase of Global Liquidity and its Impact on Emerging Economies.” Keynote Address at Federal Reserve Bank of San Francisco Asia Economic Policy Conference: 3– 5. Shin, H.S., 2016. “Global Liquidity and Procyclicality.” Presentation at the World Bank Conference: The State of Economics, the State of the World. Tirole, J., 2003. “Inefficient Foreign Borrowing: A Dual- and Common-Agency Perspective.” American Economic Review 93(5): 1678–1702. The Economist, 2016. “The Well Runs Dry.” March 5. Turner, P., 2014. “The Global Long-term Interest Rate, Financial Risks, and Policy Choices in EMEs.” BIS Working Paper No. 441. World Bank, 2015. “The Use of Markets for Long-term Finance.” Global Financial Development Report. 35 Figure 1. Total Amount Raised in Equity, Corporate Bond, and Syndicated Loan Markets This figure shows the total amount raised per year in equity, corporate bond, and syndicated loan markets by developed countries (Panel A) and developing countries (Panel B). Total debt is the sum of the amount raised through corporate bonds and syndicated loans. A. Developed Countries 6,000 12 5,000 10 4,000 8 Share of GDP, % 2011 US$ billion 3,000 6 2,000 4 1,000 2 0 0 1993 1996 2003 2006 2010 2013 1991 1992 1994 1995 1997 1998 1999 2000 2001 2002 2004 2005 2007 2008 2009 2011 2012 2014 B. Developing Countries 1,200 4 900 3 Share of GDP, % 2011 US$ billion 600 2 300 1 0 0 1999 2001 2003 2005 2007 1991 1992 1993 1994 1995 1996 1997 1998 2000 2002 2004 2006 2008 2009 2010 2011 2012 2013 2014 Equity Corporate bonds Syndicated loans Equity, % of GDP Total debt, % of GDP Figure 2. Cumulative Distribution Functions This figure shows the cumulative distribution functions (CDF) of the maturity of newly issued debt during 1991-2014 for developed countries (Panel A) and developing countries (Panel B). The figure shows separately the CDF for domestic syndicated loans, international syndicated loans, domestic bonds, and international bonds. Panel C shows the (value) weighted average maturity (in years) of debt issued in different markets. A. Developed Countries 100 100 90 90 80 80 Cumulative weights, % Cumulative weights, % 70 70 60 60 50 50 40 40 30 30 20 20 10 10 0 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 Maturity, in years Maturity, in years B. Developing Countries 100 100 90 90 80 80 Cumulative weights, % Cumulative weights, % 70 70 60 60 50 50 40 40 30 30 Developing countries, domestic bonds 20 20 Developing countries, international bonds Corporate bonds Developing countries, domestic syndicated loans 10 10 Syndicated loans Developing countries, international syndicated loans 0 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 Maturity, in years Maturity, in years C. Average Maturity Total debt Corporate bonds Syndicated loans Corporate Syndicated Share raised with Share raised Share raised Country type bonds loans syndicated loans Domestic International internationally Domestic International internationally Developed countries 10.1 4.9 69% 10.8 9.0 38% 4.4 5.2 66% Developing countries 7.2 7.6 52% 6.1 9.9 30% 11.6 6.3 76% Figure 3. Changes in Debt Issuance over Time This figure shows the percent change of the number of issuances per year in corporate bond and syndicated loan markets for developed countries (Panel A) and developing countries (Panel B). A. Developed Countries 50 40 30 Percent change, % 20 10 0 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 -10 -20 -30 B. Developing Countries 100 80 60 Percent change, % 40 20 0 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 -20 -40 Corporate bonds Syndicated loans Figure 4. Debt Maturity and Issuance Composition over Time 13 maturity at issuance per year (left axis) and the share of the total amount raised in different debt markets (right axis) for developed This figure shows the weighted average debt 100 countries and developing countries. 12 A. Total Debt: Bonds and Syndicated Loans 90 Developed Countries Developing Countries Share raised with bonds, % of total debt 13 11 100 13 80 100 Share raised with bonds, % of total debt Share raised with bonds, % of total debt 12 90 12 90 10 70 11 80 11 80 Maturity, in years 9 60 10 70 10 70 Maturity, in years Maturity, in years 9 8 60 9 50 60 8 7 50 8 40 50 7 40 7 40 6 30 6 30 6 30 5 20 5 20 5 20 4 4 10 4 10 10 3 3 0 3 0 0 2003 2004 2005 2006 2007 2003 2008 2004 2010 2009 20052011 2006 2012 2007 2013 2008 2014 2009 2010 2003 2011 2004 2012 20062013 2005 2007 2014 2008 2009 2010 2011 2012 2013 2014 Share of bonds Bonds Syndicated loans Total debt 15 100 B. Domestic and International Bonds 14 90 Share raised domestically, % of total bonds Developed Countries Developing Countries 13 13 100 15 80 100 Share raised domestically, % of total bonds 12 90 14 90 Share raised domestically, % of total bonds 12 70 11 80 13 80 Maturity, in years 11 60 10 70 12 70 Maturity, in years Maturity, in years 9 10 60 11 50 60 8 9 50 10 40 50 7 8 40 9 30 40 6 30 8 30 7 20 5 20 7 20 6 10 4 10 6 10 3 5 0 5 0 0 2003 2004 2005 2006 2007 2003 2008 2004 2010 2009 2005 20112006 2012 2007 2013 2008 2014 2009 2010 2003 2011 2004 20132007 2014 2012 2006 2005 2008 2009 2010 2011 2012 2013 2014 Share of domestic bonds International bonds Domestic bonds Total bonds 13 C. Domestic and International Syndicated Loans 100 Share raised domestically, % of total syndicated loans 12 Developed Countries Developing Countries 90 7.0 100 13 100 11 80 6.5 90 12 90 Share raised domestically, % of total s.loans Share raised domestically, % of total s.loans 10 70 6.0 80 11 80 Maturity, in years 9 60 5.5 70 10 70 8 50 Maturity, in years Maturity, in years 5.0 60 9 60 7 40 4.5 50 8 50 6 30 4.0 40 7 40 5 20 3.5 30 6 30 4 10 3.0 20 5 20 3 0 2.5 2003 2004 2005 2006 2007 2008 10 2009 4 2010 2011 2012 2013 2014 10 2.0 0 3 0 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 Share of domestic syndicated loans International syndicated loans Domestic syndicated loans Total syndicated loans Figure 5. Firm Size Distributions This figure shows the firm size distributions for switching and non-switching debt issuers. The left-side (righ-side) graphs show the firm size distributions for switchers and non- switchers during the GFC (domestic banking crises). Firm size is measured during non-crisis periods. Panel A shows the distributions for switchers against non-switchers in all markets. Panel B compares switchers with non-switchers in international markets. Panel C compares switchers with non-switchers in domestic markets. Panel D shows the median size of switchers and non-switchers. Densities are estimated using the Epanechnikov kernel function. Firm size is proxied by the average debt transaction measured over all debt issuances during the entire sample period (in 2011 US$ million). A. Non-switchers versus Switchers GFC Domestic Banking Crises B. Non-switchers in International Markets versus Switchers GFC Domestic Banking Crises C. Non-switchers in Domestic Markets versus Switchers GFC Domestic Banking Crises D. Median Firm Size (2011 US$ Million) Non-switchers in different markets Switchers Non-switchers International Domestic International Domestic bonds bonds syndicated loans syndicated loans $287 $81 $59 $160 $55 $190 Figure 6. Debt Financing and Maturity Dynamics for Different Types of Issuers This figure shows the average number of debt issuances per year (left axis) and the weighted average debt maturity (in years) at issuance per year (right axis) for switching and non-switching firms around the GFC (Panels A and B) and domestic banking crises (Panel C). Averages across countries are reported. A. GFC, Developed Countries 120 8 100 7 80 6 Number of issuances Maturity, in years 60 5 40 4 20 3 0 2 2003-07 2008-09 2003-07 2008-09 Switchers Non-switchers B. GFC, Developing Countries 16 7 14 12 6 Number of issuances 10 Maturity, in years 8 5 6 4 4 2 0 3 2003-07 2008-09 2003-07 2008-09 Switchers Non-switchers C. Domestic Banking Crises 80 7 70 60 6 Number of issuances Maturity, in years 50 40 5 30 20 4 10 0 3 Non-crisis periods Domestic banking Non-crisis periods Domestic banking crises crises Switchers Non-switchers Financing (left axis) Maturity (right axis) Figure 7. Total Debt Financing by Market This figure shows, on the left axis, the average yearly amount raised in each market and, on the right axis, the amount raised by non-switching issuers as a share of the total amount raised in each market. The gray areas show the markets hit by the GFC. A. Developed Countries 2,000 100 1,800 90 Share of non-switchers, % of total 1,600 80 1,400 70 2011 US$ billion 1,200 60 1,000 50 800 40 600 30 400 20 200 10 0 0 2003-07 2008-09 2003-07 2008-09 2003-07 2008-09 2003-07 2008-09 Domestic bonds International bonds Domestic syndicated International loans syndicated loans B. Developing Countries 160 100 90 140 Share of non-switchers, % of total 80 120 70 2011 US$ billion 100 60 80 50 40 60 30 40 20 20 10 0 0 2003-07 2008-09 2003-07 2008-09 2003-07 2008-09 2003-07 2008-09 Domestic bonds International bonds Domestic syndicated International loans syndicated loans Markets in crisis Total lending (left axis) Share raised by non-switchers (right axis) Table 1. Summary Statistics This table shows the total number of firms issuing debt, the number of debt issuances, the total amount raised, and the (value) weighted average maturity of debt at issuance (in years) for the 1991-2014 period. Data on the total amount raised are in 2011 US$ billion. A. All Countries Type of debt No. of firms No. of issuances Amount raised Average maturity Total debt 57,950 267,382 $68,037 6.6 Corporate bonds 25,213 108,184 $22,311 9.8 Syndicated loans 41,746 159,198 $45,727 5.1 B. Developed Countries Type of debt No. of firms No. of issuances Amount raised Average maturity Corporate bonds 19,993 91,166 $19,503 10.1 Syndicated loans 36,172 145,940 $42,694 4.9 C. Developing Countries Type of debt No. of firms No. of issuances Amount raised Average maturity Corporate bonds 5,220 17,018 $2,807 7.2 Syndicated loans 5,574 13,258 $3,033 7.6 Table 2. Debt Maturity of Corporate Bonds and Syndicated Loans This table shows regression estimates characterizing the differences in maturity between corporate bonds and syndicated loans during 1991-2014. The maturity of debt issuances (in years) is regressed on a dummy that equals one for bond issuances and zero otherwise (syndicated loan issuances), and dummies for the different uses of proceeds raised ("acquisition financing/leveraged buyout operations" is the omitted category). The regressions also include different combinations of year dummies, country-year dummies, and/or firm fixed effects. The regressions are estimated using ordinary least squares with standard errors (reported in brackets) clustered at the firm level. *, **, and *** denote statistical significance at 10%, 5%, and 1%, respectively. Dependent variable: maturity of debt issuances A. Developed Countries Firm fixed effects Omitted category: syndicated loans Country-year Firm fixed effects and country-year Year dummies dummies and year dummies dummies Corporate bonds 2.88 *** 4.48 *** 5.16 *** 5.26 *** [0.09] [0.11] [0.08] [0.08] Debt financing for: general corporate purposes -0.85 *** -0.50 *** -0.44 *** [0.04] [0.04] [0.04] Debt financing for: project finance and long-term investments 4.29 *** 1.10 *** 1.15 *** [0.11] [0.15] [0.15] Debt financing for: refinancing and capital structure management -0.25 *** -0.28 *** -0.23 *** [0.04] [0.04] [0.04] Debt financing for: other expenses -0.80 *** -1.09 *** -1.08 *** [0.11] [0.09] [0.09] No. of observations 237,106 237,106 237,106 237,106 No. of clusters 48,183 48,183 48,183 48,183 R-squared 0.077 0.223 0.146 0.157 B. Developing Countries Firm fixed effects Omitted category: syndicated loans Country-year Firm fixed effects and country-year Year dummies dummies and year dummies dummies Corporate bonds -0.80 *** 1.19 *** 1.84 *** 1.97 *** [0.13] [0.12] [0.19] [0.20] Debt financing for: general corporate purposes 0.93 *** 0.94 *** 0.81 *** [0.18] [0.24] [0.23] Debt financing for: project finance and long-term investments 5.55 *** 3.12 *** 3.07 *** [0.22] [0.33] [0.33] Debt financing for: refinancing and capital structure management 0.66 *** 0.93 *** 0.77 *** [0.20] [0.26] [0.25] Debt financing for: other expenses 2.15 *** 1.48 *** 1.62 *** [0.26] [0.32] [0.31] No. of observations 30,276 30,276 30,276 30,276 No. of clusters 9,767 9,767 9,767 9,767 R-squared 0.036 0.300 0.041 0.120 Table 3. Debt Maturity across Different Debt Markets This table shows regression estimates characterizing differences in maturity between corporate bonds and syndicated loans issued domestically and internationally during 1991-2014. The maturity of debt issuances (in years) is regressed on a dummy that equals one for international syndicated loans, a dummy that equals one for domestic bonds, and a dummy that equals one for international bonds. The omitted category is domestic syndicated loan issuances. The regressions also include different combinations of dummies for the uses of proceeds, a foreign-currency dummy, a floating rate dummy, year dummies, country-year dummies, and/or firm fixed effects. Each panel in the table also reports Wald tests for differences in coefficients. The regressions are estimated using ordinary least squares with standard errors (reported in brackets) clustered at the firm level. *, **, and *** denote statistical significance at 10%, 5%, and 1%, respectively. Dependent variable: maturity of debt issuances A. Developed Countries Firm fixed effects Omitted category: domestic syndicated loans Country-year Firm fixed effects and country-year Year dummies dummies and year dummies dummies International syndicated loans 1.16 *** 0.28 *** -0.07 0.03 [0.04] [0.03] [0.05] [0.05] Domestic bonds 2.23 *** 3.32 *** 4.13 *** 4.08 *** [0.19] [0.25] [0.16] [0.18] International bonds 2.07 *** 1.81 *** 2.87 *** 2.93 *** [0.18] [0.20] [0.16] [0.18] Use of proceeds dummies No Yes Yes Yes Foreign currency dummy Yes Yes Yes Yes Floating rate dummy Yes Yes Yes Yes No. of observations 234,389 234,389 234,389 234,389 No. of clusters 47,347 47,347 47,347 47,347 R-squared 0.088 0.240 0.148 0.160 Differences in coefficients International bonds vs. domestic bonds -0.16 -1.51 *** -1.26 *** -1.15 *** International bonds vs. international syndicated loans 0.91 *** 1.53 *** 2.95 *** 2.90 *** Domestic bonds vs. international syndicated loans 1.07 *** 3.04 *** 4.21 *** 4.05 *** B. Developing Countries Firm fixed effects Omitted category: domestic syndicated loans Country-year Firm fixed effects and country-year Year dummies dummies and year dummies dummies International syndicated loans -2.35 *** -2.11 *** -1.40 *** -1.38 *** [0.19] [0.19] [0.29] [0.30] Domestic bonds -3.80 *** -1.41 *** -0.76 ** -0.62 * [0.19] [0.22] [0.33] [0.33] International bonds -1.31 *** 0.39 1.54 *** 1.57 *** [0.32] [0.28] [0.39] [0.38] Use of proceeds dummies No Yes Yes Yes Foreign currency dummy Yes Yes Yes Yes Floating rate dummy Yes Yes Yes Yes No. of observations 30,214 30,214 30,214 30,214 No. of clusters 9,736 9,736 9,736 9,736 R-squared 0.080 0.322 0.060 0.137 Differences in coefficients International bonds vs. domestic bonds 2.49 *** 1.80 *** 2.30 *** 2.19 *** International bonds vs. international syndicated loans 1.04 *** 2.50 *** 2.95 *** 2.95 *** Domestic bonds vs. international syndicated loans -1.45 *** 0.70 *** 0.64 *** 0.77 *** Table 4. Debt Market Choice during the Global Financial Crisis This table shows logit regressions analyzing firms' debt market choice in the aftermath of the GFC vis-à-vis the pre-crisis period (2003- 2007). Panel A estimates the choice between bonds and syndicated loans. The dependent variable is a dummy that equals one (zero) if a firm issues a bond (syndicated loan) in a given quarter. Panel B estimates the choice between domestic and international bond financing. The dependent variable is a dummy that equals one (zero) if a firm issues a domestic (international) bond in a given quarter. Panel C estimates the choice between domestic and international syndicated loan financing. The dependent variable is a dummy that equals one (zero) if a firm issues a domestic (international) syndicated loan in a given quarter. The independent variables are dummy variables for the GFC period (2008-09) and the post-crisis period (2010-14). For quarters in which a firm has issuances of both bonds and syndicated loans, we set the debt choice indicator equal to one if the amount raised through bonds exceeds that raised through syndicated loans. A similar rule is applied for the domestic versus international issuances within bond and syndicated loan markets. The reported statistics are the marginal effects implied by the logit estimations. Standard errors (reported in brackets) are clustered at the firm level. *, **, and *** denote statistical significance at 10%, 5%, and 1%, respectively. A. Corporate Bonds versus Syndicated Loans Dependent variable: dummy dit =1 if the firm issues a bond in quarter t , dit =0 if the firm issues a syndicated loan in quarter t Developed countries Developing countries Pre-crisis mean dit : 0.24 Pre-crisis mean dit : 0.48 Country fixed effects Firm fixed effects Country fixed effects Firm fixed effects GFC (2008-09) 0.03 *** 0.05 *** 0.05 *** 0.10 *** [0.01] [0.01] [0.02] [0.03] Post crisis (2010-14) 0.02 *** 0.04 *** 0.08 *** 0.17 *** [0.00] [0.01] [0.01] [0.02] No. of observations 91,913 33,012 14,460 3,373 No. of clusters 34,716 4,554 7,486 623 B. Domestic versus International Corporate Bonds Dependent variable: dummy dit =1 if the firm issues a domestic bond in quarter t , dit =0 if the firm issues an international bond in quarter t Developed countries Developing countries Pre-crisis mean dit : 0.70 Pre-crisis mean dit : 0.71 Country fixed effects Firm fixed effects Country fixed effects Firm fixed effects GFC (2008-09) 0.05 *** -0.011 0.09 *** 0.25 *** [0.01] [0.02] [0.01] [0.03] Post crisis (2010-14) -0.11 *** -0.14 *** 0.08 *** 0.07 ** [0.01] [0.01] [0.01] [0.04] No. of observations 27,003 8,706 8,400 1,359 No. of clusters 10,655 1,490 3,941 250 C. Domestic versus International Syndicated Loans Dependent variable: dummy dit =1 if the firm issues a domestic syndicated loan in quarter t , dit =0 if the firm issues an international syndicated loan in quarter t Developed countries Developing countries Pre-crisis mean dit : 0.60 Pre-crisis mean dit : 0.12 Country fixed effects Firm fixed effects Country fixed effects Firm fixed effects GFC (2008-09) 0.04 *** 0.02 0.31 *** 0.22 *** [0.01] [0.01] [0.03] [0.02] Post crisis (2010-14) -0.05 *** -0.13 *** 0.34 *** 0.30 *** [0.01] [0.01] [0.02] [0.02] No. of observations 67,924 17,658 5,449 488 No. of clusters 28,841 488 3,649 107 Table 5. Debt Maturity during the Global Financial Crisis This table shows in Panel A regression estimates of the debt maturity at issuance (in years) for developed and developing countries around the GFC, and in Panel B, the share of each debt market in the total amount raised. In Panel A, the debt maturity is regressed on dummies for the GFC (2008-09) and the post-crisis period (2010-2014). The regressions are estimated using ordinary least squares on quarterly data during 2003-14. Debt maturity is aggregated from the daily transaction-level data to a quarterly frequency, weighted by the amount raised. Standard errors are clustered at the firm level. *, **, and *** denote statistical significance at 10%, 5%, and 1%, respectively. A. Maturity of Debt Issuances Country fixed effects Developed countries Developing countries Total debt Corporate bonds Syndicated loans Total debt Corporate bonds Syndicated loans GFC (2008-09) -0.16 -1.26 *** -0.80 *** 0.43 -1.08 *** 1.27 *** [0.12] [0.25] [0.11] [0.31] [0.40] [0.38] Post crisis (2010-14) 0.25 *** -0.62 *** 0.07 0.80 ** -0.56 1.29 * [0.08] [0.21] [0.07] [0.33] [0.54] [0.33] No. of observations 91,512 26,864 68,050 14,466 8,503 6,157 No. of clusters 34,639 10,627 29,011 7,500 4,017 4,142 Firm fixed effects Developed countries Developing countries Total debt Corporate bonds Syndicated loans Total debt Corporate bonds Syndicated loans GFC (2008-09) -0.15 -0.81 *** -0.69 *** -0.26 -1.25 ** 0.17 [0.19] [0.29] [0.07] [0.27] [0.59] [0.28] Post crisis (2010-14) 0.71 *** 0.44 0.55 *** 1.48 ** 0.68 0.79 *** [0.17] [0.29] [0.07] [0.67] [1.11] [0.27] No. of observations 91,512 26,864 68,050 14,466 8,503 6,157 No. of clusters 34,639 10,627 29,011 7,500 4,017 4,142 B. Debt Market Shares Developed countries Developing countries Total debt Corporate bonds Syndicated loans Total debt Corporate bonds Syndicated loans Pre crisis (2003-07) 100% 22% 78% 100% 31% 69% GFC (2008-09) 100% 37% 63% 100% 52% 48% Post crisis (2010-14) 100% 30% 70% 100% 57% 43% Table 6. Debt Market Choice during Domestic Banking Crises This table shows logit regressions analyzing the firms’ debt market choice around domestic banking crises. The first two columns estimate the choice between bonds and syndicated loans. The dependent variable is a dummy that equals one (zero) if a firm issues a bond (syndicated loan) in a given quarter. The two middle columns focus on the choice between domestic and international bond financing. The dependent variable is a dummy that equals one (zero) if a firm issues a domestic (international) bond in a given quarter. The last two columns estimate the choice between the domestic and international syndicated loan financing. The dependent variable is a dummy that equals one (zero) if a firm issues a domestic (international) syndicated loan in a given quarter. The independent variable is a dummy for the periods of domestic banking crises. For quarters in which a firm has issuances of both bonds and syndicated loans, the debt choice indicator equals one if the amount raised through bonds exceeds that raised through syndicated loans. A similar rule is applied for the domestic versus international issuances within bond and syndicated loan markets. Panel A shows the results for the 1991-2010 period. Panel B shows the results for the 1991-2007 period, excluding the GFC period and its aftermath. Panel C shows the results for the 1991-2010 period, excluding countries that experienced domestic banking crises during the GFC (2008-09). The reported statistics are the marginal effects implied by the logit estimations. Standard errors (reported in brackets) are clustered at the firm level. *, **, and *** denote statistical significance at 10%, 5%, and 1%, respectively. A. Sample Period: 1991 - 2010 Corporate bonds versus syndicated loans Domestic versus international bonds Domestic versus international syndicated loans Dependent variable: dummy dit =1 if the firm Dependent variable: dummy dit =1 if the firm Dependent variable: dummy dit =1 if the firm issues a domestic syndicated loan in quarter t , issues a bond in quarter t , dit =0 if the firm issues a domestic bond in quarter t , dit =0 if the dit =0 if the firm issues an international issues a syndicated loan in quarter t firm issues an international bond in quarter t syndicated loan in quarter t Non-crisis mean dit : 0.41 Non-crisis mean dit : 0.81 Non-crisis mean dit : 0.62 Country and year Firm fixed effects and Country and year Firm fixed effects and Country and year Firm fixed effects and dummies year dummies dummies year dummies dummies year dummies Domestic banking crises 0.14 *** 0.15 *** -0.07 *** -0.11 *** -0.18 *** -0.19 *** [0.01] [0.01] [0.00] [0.01] [0.01] [0.02] No. of observations 111,235 47,414 50,605 14,394 63,887 19,153 No. of clusters 40,295 6,263 17,696 1,924 29,100 3,714 B. Sample Period: 1991 - 2007 Corporate bonds versus syndicated loans Domestic versus international bonds Domestic versus international syndicated loans Non-crisis mean dit : 0.42 Non-crisis mean dit : 0.81 Non-crisis mean dit : 0.60 Country and year Firm fixed effects and Country and year Firm fixed effects and Country and year Firm fixed effects and dummies year dummies dummies year dummies dummies year dummies Domestic banking crises 0.14 *** 0.10 *** -0.01 * 0.02 -0.17 *** -0.16 *** [0.01] [0.01] [0.00] [0.02] [0.01] [0.03] No. of observations 90,594 37,978 43,654 11,012 49,385 14,605 No. of clusters 33,993 5,471 15,939 1,528 23,619 3,000 C. Sample Period: 1991-2010, excluding Countries with Domestic Banking Crises during 2008-09 Corporate bonds versus syndicated loans Domestic versus international bonds Domestic versus international syndicated loans Non-crisis mean dit : 0.46 Non-crisis mean dit : 0.82 Non-crisis mean dit : 0.70 Country and year Firm fixed effects and Country and year Firm fixed effects and Country and year Firm fixed effects and dummies year dummies dummies year dummies dummies year dummies Domestic banking crises 0.24 *** 0.22 *** 0.01 0.07 ** -0.12 *** -0.11 ** [0.02] [0.02] [0.01] [0.03] [0.02] [0.05] No. of observations 46,191 15,527 25,021 7,462 21,432 2,198 No. of clusters 16,575 2,003 8,174 989 10,192 477 Table 7. Debt Maturity during Domestic Banking Crises This table shows regression estimates of the debt maturity at issuance (in years) around domestic banking crises. The maturity of debt issuances is regressed on a dummy for domestic banking crises. The regressions are estimated using ordinary least squares on quarterly data during 1991- 2010. Debt maturity is aggregated from the daily transaction-level data to a quarterly frequency, weighted by the amount raised. Standard errors are clustered at the firm level. *, **, and *** denote statistical significance at 10%, 5%, and 1%, respectively. Dependent variable: maturity of debt issuances Country and year dummies Total debt Corporate bonds Syndicated loans Domestic banking crises 0.38 *** 0.03 -0.14 *** [0.05] [0.08] [0.04] No. of observations 111,257 50,638 64,365 No. of clusters 40,306 17,725 29,369 Firm fixed effects and year dummies Total debt Corporate bonds Syndicated loans Domestic banking crises 0.36 ** 0.32 -0.35 *** [0.16] [0.21] [0.10] No. of observations 111,257 50,638 64,365 No. of clusters 40,306 17,725 29,369 Table 8. Switchers and Non-switchers: Number of Firms and Amount Raised This table shows the number of switching and non-switching firms and their share in the total amount raised. A. GFC, Developed Countries Share of total debt raised Number of firms Pre crisis (2003-07) GFC (2008-09) Switching issuers 4,721 66% 87% Non-switching issuers 13,628 34% 13% B. GFC, Developing Countries Share of total debt raised Number of firms Pre crisis (2003-07) GFC (2008-09) Switching issuers 452 46% 70% Non-switching issuers 1,916 54% 30% C. Domestic Banking Crises Share of total debt raised Number of firms Non-crisis periods Domestic banking crises Switching issuers 8,172 75% 76% Non-switching issuers 25,501 25% 24% Table 9. Debt Maturity during Crises for Different Types of Issuers This table shows regression estimates of the debt maturity at issuance (in years) for switching and non-switching issuers around crises. In Panels A and B, focusing on the 2003-14 period, debt maturity is regressed on dummies for the GFC period (2008-09) and the post-crisis period (2010-2014). In Panel C, debt maturity is regressed on a dummy for domestic banking crises, covering the 1991-2010 period. All the regressions include firm fixed effects. The regressions are estimated using ordinary least squares on quarterly data. Debt maturity is aggregated from the daily transaction-level data to a quarterly frequency, weighted by the amount raised. Standard errors are clustered at the firm level. *, **, and *** denote statistical significance at 10%, 5%, and 1%, respectively. Dependent variable: maturity of debt issuances A. GFC, Developed Countries Switching issuers Non-switching issuers Corporate Syndicated Domestic International Domestic International Total debt bonds loans bonds bonds syndicated loans syndicated loans GFC (2008-09) -0.12 -0.87 *** -0.73 *** -1.32 ** -0.21 -0.18 * -1.17 *** [0.20] [0.30] [0.08] [0.51] [1.29] [0.10] [0.23] Post crisis (2010-14) 0.76 *** 0.46 0.62 *** -0.71 0.63 0.19 -0.24 [0.18] [0.30] [0.08] [0.69] [1.21] [0.15] [0.19] No. of observations 35,343 14,881 23,199 3,537 1,791 17,506 7,118 No. of clusters 4,722 3,561 4,572 1,844 911 7,677 4,309 B. GFC, Developing Countries Switching issuers Non-switching issuers Corporate Syndicated Domestic International Domestic International Total debt bonds loans bonds bonds syndicated loans syndicated loans GFC (2008-09) -0.46 -1.71 ** 0.22 -1.15 *** -1.19 * 0.27 -0.49 * [0.29] [0.67] [0.34] [0.27] [0.71] [1.21] [0.30] Post crisis (2010-14) 2.06 *** 1.76 0.91 *** -0.54 * 0.80 -0.93 0.18 [0.70] [1.17] [0.32] [0.31] [1.33] [0.65] [0.27] No. of observations 2,949 1,829 1,254 1,648 323 237 1,427 No. of clusters 452 431 393 618 250 184 943 C. Domestic Banking Crises Switching issuers Non-switching issuers Corporate Syndicated Domestic International Domestic International Total debt bonds loans bonds bonds syndicated loans syndicated loans Domestic banking crises 0.75 *** 0.35 -0.13 * 0.26 0.92 -0.20 ** -0.79 ** [0.18] [0.22] [0.08] [0.38] [1.40] [0.09] [0.37] No. of observations 56,814 28,453 31,754 18,501 3,870 21,301 11,624 No. of clusters 8,518 6,939 8,036 8,631 2,253 12,995 8,522 Table 10. Changes in Debt Issuance Composition during Crises This table shows the changes in debt issuance during crises. Panels A and B show the changes in debt issuance during the GFC (2008-2009) relative to 2007 (the peak in capital raising activity before the GFC for most countries). Column C shows the average change in debt issuance during domestic banking crises relative to non-crisis periods. Each column shows changes in the total amount raised within each market by switching and non-switching isuers. A. GFC, Developed Countries Corporate bonds Syndicated loans ∆% Domestic ∆% International ∆% Domestic ∆% International Type of issuer bonds bonds syndicated loans syndicated loans Switching issuers 8.1% 17.0% -48.1% -55.0% Non-switching issuers -72.5% 5.0% -72.4% -85.5% B. GFC, Developing Countries Corporate bonds Syndicated loans ∆% Domestic ∆% International ∆% Domestic ∆% International Type of issuer bonds bonds syndicated loans syndicated loans Switching issuers 100.4% -36.5% 25.4% -70.1% Non-switching issuers -15.1% -83.5% -90.4% -88.5% C. Domestic Banking Crises Corporate bonds Syndicated loans ∆% Domestic ∆% International ∆% Domestic ∆% International Type of issuer bonds bonds syndicated loans syndicated loans Switching issuers -13.7% 44.6% -49.7% 6.3% Non-switching issuers -26.6% 18.5% -43.0% 27.0% Appendix Table 1. Total Number of Issuances and Firms per Country This table shows, for each country in the sample, the total number of issuances in corporate bond and syndicated loan markets and the number of issuing firms per country during 1991-2014. Developed countriesNo. of Developing countries No. of No. of bond syndicated loan No. of bond syndicated loan Country issuances issuances No. of firms Country issuances issuances No. of firms Australia 2,024 3,744 1,452 Argentina 497 291 253 Austria 252 193 147 Azerbaijan 3 30 14 Bahrain 3 62 20 Bangladesh 4 46 23 Belgium 416 611 222 Bolivia 137 17 40 Canada 3,577 4,638 2,066 Brazil 2,247 790 1020 Croatia 12 117 42 Bulgaria 3 86 30 Cyprus 11 70 31 Chile 626 381 232 Czech Republic 46 243 104 China 4,366 1,878 2674 Denmark 182 225 98 Colombia 365 126 169 Finland 397 503 206 Costa Rica 197 21 30 France 2,297 5,533 1,509 Ecuador 118 14 74 Germany 1,093 4,244 1,167 Egypt, Arab Rep. 12 174 49 Greece 97 388 155 El Salvador 122 25 20 Hong Kong SAR, China 981 1,368 825 Ghana 2 78 24 Hungary 12 205 74 India 1,837 2,486 1336 Iceland 7 71 24 Indonesia 401 1,133 566 Ireland, Rep. 221 459 182 Jamaica 22 45 15 Israel 79 100 54 Jordan 3 34 20 Italy 486 2,399 897 Kazakhstan 28 82 57 Japan 10,789 21,122 6,328 Malaysia 1,774 641 484 Korea, Rep. 17,426 972 3,184 Mexico 1,147 802 500 Kuwait 2 87 32 Morocco 4 50 17 Luxembourg 466 347 196 Nigeria 5 108 48 Netherlands 2,455 2,212 798 Pakistan 29 164 78 New Zealand 306 953 203 Panama 78 159 132 Norway 431 902 371 Peru 536 138 145 Oman 7 131 50 Philippines 270 348 153 Poland 66 351 150 Romania 9 153 65 Portugal 636 530 552 Russian Federation 598 855 446 Qatar 7 148 49 South Africa 146 314 167 Saudi Arabia 20 335 112 Sri Lanka 3 21 12 Singapore 735 903 515 Thailand 1,182 918 456 Slovak Republic 10 92 35 Tunisia 0 37 11 Slovenia 2 52 20 Turkey 37 476 184 Spain 463 3,887 1,297 Ukraine 27 78 39 Sweden 571 874 260 Venezuela, R.B. 155 77 87 Switzerland 794 866 393 Vietnam 28 182 97 Taiwan, China 5,004 4,100 1,250 United Arab Emirates 56 464 151 United Kingdom 3,633 8,402 2,587 United States 35,094 73,037 20,444 Total 91,166 145,940 48,252 Total 17,018 13,258 9,767 Appendix Table 2. Debt Maturity during the GFC for Developed Countries This table considers separately all four debt markets, domestic and international bonds and syndicated loans. It shows in Panel A regression estimates of the debt maturity at issuance (in years) for developed countries around the GFC, and in Panel B, the share of each debt market in the total amount raised. In Panel A, debt maturity is regressed on dummies for the GFC (2008-09) and the post-crisis period (2010-2014). The regressions are estimated using ordinary least squares on quarterly data during 2003-14. Debt maturity is aggregated from the daily transaction-level data to a quarterly frequency, weighted by the amount raised. Standard errors are clustered at the firm level. *, **, and *** denote statistical significance at 10%, 5%, and 1%, respectively. A. Maturity of Debt Issuances Country fixed effects International Domestic International Domestic bonds bonds syndicated loans syndicated loans GFC (2008-09) -0.95 *** -1.62 *** -0.63 *** -0.86 *** [0.31] [0.33] [0.11] [0.15] Post crisis (2010-14) 0.02 -0.88 *** 0.18 * -0.01 [0.26] [0.32] [0.10] [0.08] No. of observations 17,817 9,446 40,786 27,950 No. of clusters 7,404 4,761 18,837 14,140 Firm fixed effects International Domestic International Domestic bonds bonds syndicated loans syndicated loans GFC (2008-09) -0.79 * -0.33 -0.02 -0.88 *** [0.41] [0.55] [0.09] [0.09] Post crisis (2010-14) 0.10 1.28 ** 1.17 *** 0.39 *** [0.31] [0.54] [0.23] [0.07] No. of observations 17,817 9,446 40,784 27,950 No. of clusters 7,404 4,761 18,837 14,140 B. Debt Market Shares International Domestic International Domestic bonds bonds syndicated loans syndicated loans Pre crisis (2003-07) 13% 9% 22% 56% GFC (2008-09) 20% 16% 22% 41% Post crisis (2010-14) 15% 15% 19% 51% Appendix Table 3. Debt Maturity during the GFC for Developing Countries This table considers separately all four debt markets, domestic and international bonds and syndicated loans. It shows in Panel A regression estimates of the debt maturity at issuance (in years) for developing countries around the GFC, and in Panel B, the share of each debt market in the total amount raised. In Panel A, the debt maturity is regressed on dummies for the GFC (2008-09) and the post-crisis period (2010-2014). The regressions are estimated using ordinary least squares on quarterly data during 2003-14. Debt maturity is aggregated from the daily transaction-level data to a quarterly frequency, weighted by the amount raised. Standard errors are clustered at the firm level. *, **, and *** denote statistical significance at 10%, 5%, and 1%, respectively. A. Maturity of Debt Issuances Country fixed effects International Domestic International Domestic bonds bonds syndicated loans syndicated loans GFC (2008-09) -1.53 *** -2.43 *** 0.64 0.45 [0.44] [0.85] [1.19] [0.32] Post crisis (2010-14) -1.76 *** 0.64 -0.11 0.76 ** [0.59] [0.81] [1.12] [0.30] No. of observations 7,175 1,407 2,000 4,208 No. of clusters 3,391 899 1,596 2,689 Firm fixed effects International Domestic International Domestic bonds bonds syndicated loans syndicated loans GFC (2008-09) -1.84 *** -4.25 *** 0.54 -0.17 [0.66] [1.04] [1.03] [0.26] Post crisis (2010-14) -1.20 * 0.50 -0.08 0.85 *** [0.69] [1.57] [1.19] [0.28] No. of observations 7,175 1,407 2,000 4,208 No. of clusters 3,391 899 1,596 2,689 B. Debt Market Shares International Domestic International Domestic bonds bonds syndicated loans syndicated loans Pre crisis (2003-07) 18% 13% 6% 63% GFC (2008-09) 43% 9% 17% 32% Post crisis (2010-14) 44% 14% 16% 26% Appendix Table 4. Debt Maturity during Domestic Banking Crises This table considers separately all four debt markets, domestic and international bonds and syndicated loans. It shows regression estimates of the debt maturity at issuance (in years) around domestic banking crises. The debt maturity is regressed on a dummy for domestic banking crises. The regressions are estimated using ordinary least squares on quarterly data during 1991-2010. Debt maturity is aggregated from the daily transaction-level data to a quarterly frequency, weighted by the amount raised. Standard errors are clustered at the firm level. *, **, and *** denote statistical significance at 10%, 5%, and 1%, respectively. Dependent variable: maturity of debt issuances Country and year dummies Domestic International Domestic bonds International bonds syndicated loans syndicated loans Domestic banking crises 0.24 0.19 -0.22 *** 0.06 [0.18] [0.16] [0.14] [0.07] No. of observations 40,452 10,903 38,823 26,135 No. of clusters 14,403 5,351 19,188 14,068 Firm fixed effects and year dummies Domestic International Domestic bonds International bonds syndicated loans syndicated loans Domestic banking crises 0.25 0.01 -0.38 * -0.42 *** [0.28] [0.37] [0.19] [0.14] No. of observations 40,452 10,903 38,823 26,135 No. of clusters 14,403 5,351 19,188 14,068