Firm and country determinants of debt maturity. International evidence * Víctor M. González Méndez University of Oviedo

Transcription

1 Firm and country determinants of debt maturity. International evidence * Abstract Víctor M. González Méndez University of Oviedo This paper analyses the effect of firm- and country-level determinants on debt maturity structure and how this effect varies across countries and across firm size. Results for 39 countries show that firm-level variables such as asset maturity, size, firm quality and leverage affect debt maturity structure. Institutions and banking structure also influence corporate debt maturity. While the efficiency of the legal system, protection of creditors rights and bank concentration show a positive relationship to debt maturity, the protection of property rights and the weight of banks in the economy have a negative effect on firm debt maturity. However, these firm- and country-level determinants vary according to firm size. The agency costs and signalling hypotheses are more relevant in explaining the debt maturity structure of large firms, while the asset maturity and tax hypotheses are more pertinent in the case of small firms. Most of the country-level determinants of debt maturity are size dependent; in particular, bank concentration has a positive influence on debt maturity only for the subsample of small firms, while the weight of banks in the economy has a negative influence for small firms. JEL classification: G18, G32. Keywords: Debt maturity; Firm-level determinants; Institutions; Banking structure. * I am grateful for the useful comments from the participants in the Finance Workshop at Cardiff Business School, especially Marc Goergen. Financial support from the Spanish Ministry of Economy and Competitiveness, Project ECO , is gratefully acknowledged. Department of Business Administration. University of Oviedo. Avda. Del Cristo s/n Oviedo. Spain. Tel.: ; fax: ; vmendez@uniovi.es 1

2 Firm and country determinants of debt maturity. International evidence I. Introduction Studies on firm capital structure have recently focused on the influences of institutional and legal features on the amount of debt financing. The literature on capital structure has analysed the influence of investor protection and institutions and has revealed that institutional factors play an important role in the choice of capital structure. Empirical studies show, for instance, that better protection of creditors increases the availability of debt by reducing adverse selection and moral hazard problems of debt (Levine, 1999; Giannetti, 2003; Antoniou et al., 2008; Bae and Goyal, 2009), while stronger protection of property rights favours increased use of equity over debt (González and González, 2008). Previous studies allow us to highlight the existence of differences in debt maturity among countries. For example, Barclay and Smith (1995) report a percentage of total long-term debt of around 70% for their sample of US listed firms for the period , while Antoniou et al. (2006) show that the maturity of debt the long-term debt ratio of French, German and UK listed firms is around 59%, 53% and 46%, respectively. Major differences have also been found between developed and developing countries. Demirgüç-Kunt and Maksimovic (1999) and Fan et al. (2012) reveal that firms in developing countries use less long-term debt as a proportion of total debt and that this difference cannot be explained by the maturity of assets. Figure 1 represents the average ratio of long-term debt to total debt among the countries included in our sample. Debt maturity can be seen to vary widely among countries. For example, the average percentage of long-term debt for firms in the USA or Norway is more than two times the percentage of long-term debt for firms in Thailand, South Korea, Taiwan or Turkey. These differences in financing patterns across countries may be addressed by differences in the legal and institutional environment across countries. INSERT FIGURE 1 ABOUT HERE Despite these differences in debt maturity among countries, the financial literature has focused less on the institutional determinants of debt maturity and the differences of firm-specific determinants among countries. Demirgüç-Kunt and Maksimovic (1999) examine how the differences in financial and legal institutions affect the use of debt and especially the choice of debt maturity by firms in a sample of 30 countries in the period Their paper highlights the importance of the efficiency of the legal system, the level of activity of the stock market, and the size of the banking sector as determinants of debt maturity, revealing the existence of differences in the influence of 2

3 these aspects depending on firm size. Fan et al. (2012) examine how the institutional environment influences the capital structure and debt maturity choices of firms in 39 countries in the period They find that corruption, the legal system and the size of the banking sector explain a significant portion of the variation in debt maturity ratios. Within this context, the present paper examines how institutional and banking environments affect the choice of debt maturity made by firms in a sample of 39 countries in the period Our paper makes several main contributions to the literature. First, we analyse the firm-level determinants of debt maturity and the way in which observed differences in institutional and legal environments and in banking structures across countries affect the debt maturity choice of firms within an international context. The influence of legal and institutional environments has also been analysed by Demirgüç-Kunt and Maksimovic (1999) and Fan et al. (2012). However, we not only exploit the variation over time of the firm and country variables 3, but also consider other aspects, namely protection of investors rights and bank concentration. Petersen and Rajan (1994, 1995) highlight the effect of bank concentration on the value of lending relationships. Hernández-Cánovas and Koëter-Kant (2008) examine the influence of relationship lending on bank debt maturity for a sample of SMEs from 19 European countries, finding that stronger firm-bank relationships lengthen the maturity of bank loans. Second, we consider how results vary depending on firm size, as reported evidence has highlighted the relevance of this variable. For instance, Giannetti (2003) has shown that the effect of institutional variables on the use of debt depends on firm size in a sample of listed and unlisted companies in eight European countries. Figure 2 shows the average ratio of longterm debt to total debt for our sample when considering large and small firms separately for each country. Large firms can be seen to report higher ratios of long-term debt to total liabilities than small firms. In several countries, debt maturity for large firms is twofold higher than that for small firms. Differences in financing patterns according to firm size may reflect differences in access to financial markets and institutions even within the same economy. Demirgüç-Kunt and Maksimovic (1999) also analyse how institutions can influence small and large firms differently. However, due to the fact that these authors consider aggregate country-level data, they do not obtain strong evidence related to the influence of institutions on debt maturity according to firm size. Finally, we provide evidence regarding the complementary or substitutive role of the country-level determinants of debt maturity and how the firm-level determinants of corporate debt maturity vary according to legal, institutional and banking structure variables. Demirgüç-Kunt and Maksimovic (1999) and Fan et al. (2012) do not address these issues in their papers. 3 Demirgüç-Kunt and Maksimovic (1999) rely on cross-sectional analysis across countries, taking the timeseries country means of each variable as observations. 3

4 INSERT FIGURE 2 ABOUT HERE Our results indicate that the matching, signalling and agency cost hypotheses explain the debt maturity structure of firms. As regards country determinants, legal enforcement, protection of investors rights, bank concentration and size of the banking system influence corporate debt maturity. Legal enforcement, protection of creditors rights and bank concentration lengthen the maturity of firm debt, while the protection of property rights and the size of the banking system favour the use of short-term debt. These results vary with firm size. The positive effect of bank concentration and the negative effect of the weight of banks in the economy are concentrated in smaller firms. The remainder of this paper is organized as follows. Section II discusses the influence of the legal and institutional environment and the structure of the banking system on firm debt maturity as well as the hypotheses tested in the paper. Section III describes the data sources and variables used in the study. Section IV presents the empirical results. Finally, Section V provides the conclusions. II. Theoretical background and hypotheses Access to external financing will depend partly on the legal and institutional features of the country, seeing as these provide the mechanisms for monitoring and enforcing financial contracts. The papers by Rajan and Zingales (1995) and Booth et al. (2001) suggest that institutional differences are unimportant in explaining firm capital structure in both developed and developing countries. Both papers show that factors identified as being cross-sectionally correlated with firm leverage in the United States are likewise similarly correlated in other developed and developing countries. However, empirical papers including legal and institutional variables in the analysis provide clear evidence regarding the relevance of these country variables on the use of debt. Giannetti (2003) explicitly introduces creditor right protection in the analysis of a sample of listed and unlisted companies in eight European countries. Her results suggest that the relevance of institutional variables depends on firm size. The corporate finance decisions of larger listed companies are less subject to institutional constraints in the domestic market due to the fact that they have easier access to international financial markets. However, institutional constraints in domestic markets have a greater impact on the corporate finance decisions of unlisted companies 4. Stronger creditors protection means that loans for investing in intangible assets are more available for unlisted companies and guarantees access to long-term debt for firms in sectors with highly volatile returns. 4 Beck et al. (2008) also show that small firms benefit the most from a high quality institutional environment because they are more dependent on domestic credit markets. 4

5 Fan et al. (2012) find that institutional factors are critical determinants of firm capital structure in a cross-section of 39 developed and developing countries. A country s legal and tax system, level of corruption and the availability of information intermediaries explain a significant portion of the cross-country variation in leverage. Bae and Goyal (2004) and Qian and Strahan (2007) show the relevance of institutional conditions for terms of bank loans. The former find lenders charge lower spreads on loans in countries with stronger protection of property rights, while the latter find that stronger creditors rights favour loan availability. Finally, González and González (2008) show that stronger protection of property rights reduces the use of debt and that bank concentration favours access to debt. In summary, we know that the legal and institutional environment influence corporate use of debt. We know less, however, about the effect of legal and institutional features on debt maturity, although important differences can be observed in debt maturity among countries. Legal and institutional features influence debt structure because lenders need to assess not only the credit quality of borrowers, but also the risks implicit in the legal and institutional system. Debt maturity may influence the possibility for firms to defraud creditors. Shorter maturities reduce the period during which an opportunistic firm can exploit its creditors without being in default (Diamond, 1991, 1993; Rajan, 1992). In countries in which the legal system does not provide proper protection or such protection is costly to use, firms will employ short-term debt more than longterm debt. In this context, a positive relationship can be expected between the efficiency of a legal system and the percentage of long-term debt. This relationship has already been analysed by Demirgüç-Kunt and Maksimovic (1999) and Fan et al. (2012). These authors show that long-term debt is clearly higher in countries with an effective legal system or less corruption and hence that there is a positive relationship between the enforceability of law and firm debt maturity. Several papers have shown that the protection of creditors and shareholders rights has a significant influence on the amount of debt (Giannetti, 2003; González and González, 2008). The ability of creditors to force repayment will exert an ex ante influence on the terms of the credit. When creditors rights are better protected, they will lend on more favourable terms; e.g. longer maturities (Qian and Strahan, 2007). From this point of view, we would expect a direct relation between the protection of creditor rights and corporate debt maturity 5. 5 Demirgüç-Kunt and Maksimovic (1999) do not find that the protection of creditor rights promotes debt maturity. 5

6 In countries with strong legal protection, investors can solve the conflicts associated with information asymmetries. In this respect, better shareholder protection is associated with higher valuation of corporate assets (La Porta et al., 2002), higher stock performance (Goergen et al., 2013) and lower leverage ratios (González and González, 2008). The use of financial instruments that allow insiders less discretion will be more common in poor contracting environments. Weaker legal systems should be associated with shorter maturity debt contracts seeing as short-term debt allows insiders less discretion (Fan et al., 2012). In this context, we would expect a positive relationship between the protection of property rights and corporate debt maturity. However, better protection of property rights may extend the credit market to low-grade borrowers, increasing the average risk of borrowers (Hernández-Cánovas and Koëter-Kant, 2011). Consequently, lenders could reduce this risk by promoting shorter debt, thereby generating an inverse relationship between the protection of property rights and debt maturity. As both types of relationships are possible, we make no a priori forecast for the relationship between the protection of property rights and debt maturity and treat it as an empirical issue. Financial intermediaries directly influence corporate financial structure. Financial intermediaries have advantages in collecting information (Diamond, 1984) and incentives to use this information to discipline borrowers due to the fact that they have economies of scale in obtaining information and do not suffer from free-rider problems. Fan et al. (2012) report a negative effect of the weight of banks in the economy on debt maturity as a consequence of bank preferences for short-term debt. Demirgüç-Kunt and Maksimovic (1999) also stress that short-term debt allows banks to use their advantages in monitoring borrowers. Short-term debt forces lenders to monitor corporate performance more frequently and enables the bank firm to change the terms of contract or not to renew the loan (Diamond, 1991; Rajan, 1992). Large firms have better access to domestic and international markets and are therefore usually less dependent on domestic bank credit. However, smaller firms will be more affected by bank preferences as they have more financing constraints. Consequently, we expect a negative relationship between the weight of banks in the economy and corporate debt maturity, particularly in the case of smaller firms. The banking literature suggests two potential effects of bank concentration on firm leverage. In a market without asymmetric information, there will be a negative relationship between bank concentration and firms leverage, since higher bank market power results in a higher price for debt and less credit availability. However, in markets with asymmetric information, higher bank market concentration may increase the incentives of banks to invest in the acquisition of soft information by establishing close relationships with borrowers over time. This will lead to a higher availability of 6

7 credit, thus reducing corporate financial constraints (Boot, 2000; Dell Ariccia and Marquez, 2004; González and González, 2008). The importance of bank concentration was argued by Petersen and Rajan (1994, 1995) and Berlin and Mester (1999). These authors show that US firms in less concentrated credit markets are subject to greater financial constraints. They offer evidence from small business data indicating that creditors are more likely to finance credit-constrained firms when credit markets are concentrated because it is easier for these creditors to internalize the benefits of assisting firms. The existence of a positive relationship between bank concentration and credit availability is in line with the fact that relationship banking serves to mitigate information asymmetries between creditors and debtors. Given that information asymmetries are greater in long-term debt and in small firms (Rajan and Zingales, 1995; Fama and French, 2002; Frank and Goyal, 2003; Gaud et al., 2005; Flannery and Rangan, 2006), the positive effect of bank concentration on leverage could be concentrated in long-term debt and in small firms. Hernández- Canovas and Koeter-Kant (2008) provide evidence along these lines, revealing that stronger firmbank relationships lengthen the maturity of bank loans. III. Databases and variables Our source for firm data is the Worldscope database, which contains financial statement data and stock prices from many countries in comparable form. We initially selected the 49 countries considered by La Porta et al. (1998), but eliminated 10 of them because of lack of data: Colombia, Ecuador, Egypt, Jordan, Kenya, Nigeria, Sri Lanka, Uruguay, Venezuela and Zimbabwe. Financial firms (SIC codes ) were likewise excluded. Finally, our sample comprises 30,727 firms and 171,892 observations for 39 countries. The sample includes countries with different institutional environments. For instance, 67 per cent of the countries are developed countries, 53 per cent of the economies have a market-oriented system and 36 per cent of the countries have a common law legal origin. The dependent variable is debt maturity (DEBTMAT), defined as the percentage of the firm s total debt that has a maturity of more than one year. The observed differences in debt maturities among countries depend partly on the characteristics of the firms in each economy. Thus, we control for the differences in the sample in firm characteristics among countries. To do so, we introduce firm-level variables suggested by theory and which have been used in previous studies analysing firm debt maturity (Myers, 1977; Barnea et al., 1980; Flannery, 1986; Barclay and Smith, 1995; Stohs and Mauer, 1996; Guedes and Opler, 1996; Ozkan, 2000; and Scherr and Hulburt, 2001). 7

8 Firms have an interest in matching their debt maturities to their asset maturities. If the maturity of debt is shorter than that of assets, the firm may not have sufficient cash available to pay its financial obligations when they are due. However, if debt has a longer maturity, debt payments remain due when cash flows from assets cease. Matching the maturities of assets and debt reduces these risks and hence we expect a positive relationship between the maturity of assets and firm debt maturity, like that found by Guedes and Opler (1996), Ozkan (2000) and Stohs and Mauer (1996). As a measure of the maturity of assets, we have considered the ratio between net fixed assets and total assets (ASSET_MAT). The agency costs of debt may influence corporate debt maturity given that outstanding debt may create incentive problems for shareholders, i.e. underinvestment problems. Myers (1977) argues that a firm may control this underinvestment incentive by shortening the effective maturity of its debt so that debt matures before growth options are exercised. This explanation of debt maturity based on agency costs suggests that firms whose value depends to a large extent on investment opportunities have an incentive to borrow short term. Several papers have provided favourable evidence for this relationship, such as Barclay and Smith (1995), Guedes and Opler (1996) and Ozkan (2000). The variable used as a proxy for the firm s investment opportunity set is the market-to-book ratio (GROWTH). Agency problems between shareholders and debtholders may be particularly severe for small firms as a consequence of underinvestment incentives and risk shifting (Smith and Warner, 1979). Barnea et al. (1980) suggest that these problems may be reduced by issuing shorter-term debt. These arguments thus suggest that debt maturity varies directly with firm size. Barclay and Smith (1995), Stohs and Mauer (1996) and Ozkan (2000) provide results in line with a positive relationship between size and debt maturity. We have measured firm size as the natural logarithm of sales (SIZE). The signalling hypothesis implies that a firm s choice of debt maturity structure can signal insider information about firm quality when insiders are better informed than outside investors (Flannery, 1986). In this context, high-quality firms signal their quality by issuing short-term debt. Following Antoniou et al. (2006), we use the ratio of net income plus depreciation to net debt as a proxy for firm quality (FIRM_QUALITY). Kane et al. (1985) develop a model in which the optimal debt maturity structure involves a trade-off between bankruptcy and debt issue flotation costs and the per-period tax advantage of debt financing. If firm value is highly volatile, the risk of insolvency increases. Firms with high volatility in their value have to change their capital structure frequently to reduce bankruptcy costs and hence 8

9 these firms will use more short-term debt (López-Gracia and Mestre-Barbera, 2011). In this context, the maturity of debt should rise if the volatility of firm value decreases. Following Antoniou et al. (2006), the firm s level of volatility is proxied by the absolute value of change in earnings before interest and taxes (VOL_EBIT). Diamond (1991) shows that liquidity risk increases with leverage and hence that highly leveraged firms may be expected to use more long-term debt. This positive relationship between leverage and debt maturity has been found by Stohs and Mauer (1996) and Scherr and Hulburt (2001). However, Barclay et al. (2003) argue that the relationship between leverage and maturity should be negative due to the fact that leverage and maturity are substitutes in mitigating under- and overinvestment problems. The opposing arguments and mixed empirical evidence mean that the influence leverage has on debt maturity is basically an empirical question. Leverage has been proxied in this paper as the ratio between total debt and firm market value (LEV). The market value of assets is defined as total assets minus the book value of equity plus the market value of equity. As regards country-level variables, we have used the rule of law component from the Worldwide Governance Indicators (WGI) compiled by Kaufmann et al. (2009) to proxy the efficiency of a country s legal system. Rule of law is one of the six dimensions of the WGI and captures perceptions of the extent to which agents have confidence in and abide by the rules of society, in particular the quality of contract enforcement, property rights, the police and the courts, as well as the likelihood of crime and violence (RULE_OF_LAW). The index ranges from -2.5 to 2.5, low levels of the index denoting less efficiency in the legal system. We have used this measure and not the index drawn up by the International Country Risk Guide, which was the proxy used by Demirgüç-Kunt and Maksimovic (1999), for the reason that we have annual data on RULE_OF_LAW during the period of our study and it allows us to exploit the variation over time of the variable. However, both indexes classify the countries under study quite similarly. Countries such as Indonesia, Pakistan, Philippines, Peru and Turkey present lower levels of legal enforcement according to both indexes, whereas the countries with the highest levels of legal efficiency in both indicators are Switzerland, Sweden, New Zealand, Finland, Denmark and Norway. Moreover, the correlation between these two indexes is high, presenting a value of , using the year 1996 for our proxy of the efficiency of the legal system. We use the index developed in Djankov et al. (2007) to measure the legal rights of creditors against defaulting debtors (C_RIGHTS). This index follows the creditors rights index proposed by La Porta et al. (1998), although in the former paper the creditors rights index is constructed in January each year. It measures four powers of secured lenders in bankruptcy: (1) whether there are restrictions, 9

10 such as creditor consent, when a debtor files for reorganization; (2) whether secured creditors are able to seize their collateral after the petition for reorganization is approved, i.e. whether there is no automatic stay or asset freeze imposed by the court; (3) whether secured creditors are paid first out of the proceeds of liquidating a bankrupt firm; and (4) whether an administrator, and not management, is responsible for running the business during the reorganization. A value of one is added to the index when a country s laws and regulations provide each one of these powers to secured lenders. It consequently varies between 0 and 4, with higher values indicating stronger creditors rights or stronger protection against borrower expropriation. We measure the protection of property rights by means of the index of private property rights (S_RIGHTS) published by the Heritage Foundation. This is an annual index of the degree to which private property rights are protected and the degree to which government enforces laws that protect private property. It also accounts for the possibility that private property may be expropriated, as well as analysing the independence of the judiciary, corruption within the judiciary and the ability of individuals and businesses to enforce contracts. This index ranges between 0 and 100, a high score indicating greater legal protection of property rights. We also use two variables proxying the banking structure in the country. First, the weight of banks in the economy, measured as the private credit by deposit money banks to GDP (BANK_CREDIT). The data are obtained from the Financial Structure and Economic Database (Beck et al., 2006). Second, we also use a measure of bank concentration in a country. Following Demirgüç-Kunt et al. (2004) and Beck et al. (2006), we measure bank concentration as the fraction of bank assets held by the three largest commercial banks in the country (BANK_CONC). Figures are obtained from the World Bank Database, whose main source is Fitch IBCA s Bankscope Database. We include three specific effects: industry-year, country-year and firm-specific effects. These specific effects try to control for most shocks affecting debt maturity. This approach has the advantage of being less likely to suffer from omitted variable bias or model specification than traditional regressions (Dell Ariccia et al., 2008). Table 1 provides descriptive statistics on the firm- and country-level variables used in this paper. Panel A describes all the firms included in the sample. Panels B and C show the descriptive statistics according to size. We split our overall sample into quartiles according to the amount of total assets. We define large firms as those in the largest quartile and small firms as those belonging to the smallest quartile. The mean (median) debt maturity of the sample is (48.19) per cent. Large firms present higher average maturity of debt (63.15 per cent) than small firms (35.75 per cent). 10

11 Large firms also show higher maturity of assets, growth opportunities and leverage than small firms. However, small firms have higher firm quality and a higher value of volatility of earnings. INSERT TABLE 1 ABOUT HERE Table 2 reports the descriptive statistics of firm- and country-level variables for each country. Debt maturity varies widely among countries, as we saw previously in Figure 1. Thailand has the lowest level of long-term debt (32.84 per cent), while the US has the highest percentage of long-term debt (73.10 per cent). There are also important differences according to the efficiency of the legal system, protection of investors rights, bank concentration or the weight of banks in the economy. Table 3 shows the correlation matrix. DEBT_MAT shows a positive correlation with asset maturity, size and leverage, but correlates negatively with the quality of the firm. Appendix A summaries the definitions of the variables used in the empirical analysis and their sources. INSERT TABLES 2 AND 3 ABOUT HERE IV. Results A. Determinants of firm debt maturity The estimations are carried out using panel data. Prior to testing, we use the Breusch-Pagan test (Breusch and Pagan, 1980) to identify the existence of individual effects. The null hypothesis of no unobserved heterogeneity is rejected. In this context, a model that captures individual heterogeneity, as the panel data methodology does, is appropriate. The panel data methodology corrects for unobserved firm-specific and time-specific effects. The panel data estimation was calculated using fixed effects, as the Hausman test (1978) rejects the null hypothesis of the lack of correlation between individual effects and observable variables in all regressions. Table 4 presents the results from the panel data estimation. Column (1) shows the results when only firm-level variables are considered. The relation between asset and debt maturities is positive. This is consistent with the matching hypothesis, according to which firms match assets and liabilities to reduce risk. The effect of size on debt maturity is positive, indicating that large firms have larger debt maturities. This positive relationship is in line with the idea that firms with more agency problems small firms may use shorter-term debt to reduce underinvestment and risk-shifting problems. The coefficient of VOL_EBIT is negative and significant. This result is in line with the tax hypothesis (Kane et al., 1985). FIRM_QUALITY has a negative influence on debt maturity, indicating that high-quality firms tend to issue short-term debt as a mechanism to signal private information, as per the signalling hypothesis (Flannery, 1986). Leverage has a positive relationship to debt maturity in a way 11

12 that is consistent with the arguments put forward by Diamond (1991), as liquidity risk increases with leverage and hence highly leveraged firms can be expected to use more long-term debt. Moreover, this effect dominates the use of leverage and debt maturity as substitutes in mitigating under- and overinvestment problems. The GROWTH variable has a positive and significant coefficient. This result is inconsistent with the agency cost hypothesis. Stohs and Mauer (1996) also obtain a positive relationship between the market-to-book ratio and debt maturity 6. The positive relationship could be a consequence of the liquidity risk argument, according to which firms with long-term investment opportunities prefer to hedge against liquidity risk by issuing long-term debt (Antoniou et al., 2006; Diamond, 1991; Guedes and Opler, 1996). The impacts of the proxies of asset maturity, size, firm quality and financial leverage on corporate debt maturity are economically important. Using the coefficient in column (1) in Table 4, a onestandard deviation increase in asset maturity, growth, size and leverage would respectively cause an increase in the mean value of the dependent variable of 5.59, 1.63, 6.25 and 4.11 per cent. This represents 7.75, 2.26, 8.67 and 5.71 per cent, respectively, of the standard deviation of debt maturity. On the other hand, a one-standard deviation increase in volatility of earnings and firm quality would cause a decrease in the mean value of the dependent variable of per cent and per cent, respectively. This represents and per cent, respectively, of the standard deviation of debt maturity. Consequently, as regards firm-level determinants, the results provide strong evidence in line with the matching and signalling hypotheses and only partial evidence supporting the agency cost hypothesis. We also provide support to the argument that highly leveraged firms use more long-term debt. Columns (2) to (5) include the institutional and legal variables as well as the banking structure variables considered in this study. All the results for firm-level variables discussed above are maintained when we include the country-level determinants of debt maturity in the estimations. INSERT TABLE 4 ABOUT HERE The RULE_OF_LAW variable has a positive and significant influence on debt maturity, indicating that firms in countries with strong legal enforcement have higher maturity of debt. This result is similar to that obtained by Demirgüç-Kunt and Maksimovic (1999) and means that the higher the quality of legal institutions, the greater the proportion of long-term financing. This finding is also consistent with the shorter debt maturity found by Fan et al. (2012) in countries with high levels of corruption. 6 Fan et al. (2012) find that the coefficient of the market-to-book ratio is also positive, though not significant in the overall sample. However, it is positive and significant in the developing economy subsample. 12

13 The level of protection of property rights (S_RIGHTS) has a negative influence on debt maturity, while the coefficient of the protection of creditors rights (C_RIGHTS) is positive. The negative coefficient of S_RIGHTS suggests that improvements in the institutional environment reduce corporate debt maturity. Firms in countries with strong protection of creditors rights tend to issue debt with a longer maturity. This result is in line with creditors lending on more favourable terms when their rights are strongly protected and is consistent with the evidence provided by Qian and Strahan (2007) for bank loans. Bank concentration has a positive effect on firm debt maturity. The maturity of debt increases in countries in which bank concentration is high. This result suggests that higher bank concentration increases bank incentives to establish close relationships with borrowers over time and thus reduces the financial constraints on firms. The weight of banks in the economy is seen to have a negative influence on debt maturity. This result is in line with the evidence provided by Fan et al. (2012) which is consistent with the preferences of suppliers of capital having an influence on debt maturity structures. These results for country-level determinants of debt maturity are maintained when all the variables are included jointly. Although the correlation between RULE_OF_LAW and S_RIGHTS is high, (see Table 3), it does not affect the results. The coefficients of both variables maintain their values, which are significant when the other variable is excluded from the estimations. Moreover, the results are similar when we use the index of economic freedom published by Heritage Foundation as a proxy of the protection of property rights. Column (5) shows the results when considering this other measure of S_RIGHTS. The index of economic freedom is a composite index of 10 categories of a country s economic freedom which ranges from 0 to 100, with high values implying more economic freedom. The impact of the country-level variables is economically significant. Using the coefficients in column (4), a one-standard deviation increase in RULE_OF_LAW, S_RIGHTS, C_RIGHTS, BANK_CONC and BANK_CREDIT would cause a variation in debt maturity of 6.62 per cent, per cent, per cent, 3.70 per cent and per cent, respectively. It represents 9.18, -3.03, 4.69, 5.13 and per cent, respectively, of the standard deviation of debt maturity. B. Firm-level determinants of debt maturity across firm size Table 5 analyses whether the influence of the firm-level determinants of debt maturity is different depending on firm size. To do so, the overall sample is divided into quartiles according to the 13

14 amount of total assets and two new variables are created. LARGE is a dummy variable that takes the value of 1 if the firm belongs to the largest quartile and 0 otherwise. SMALL is a dummy variable that takes the value of 1 if the firm belongs to the smallest quartile and 0 otherwise. To analyse the influence of firm size on the firm-level determinants, these two variables are multiplied by each one of the firm-level determinants. Columns (1) and (3) show the results when the overall sample is used in the estimations, whereas medium-sized firms are excluded from Columns (2) and (4). In column (1), ASSET_MAT, GROWTH, SIZE, VOL_EBIT, FIRM_QUALITY and LEV show the effect of firm-level determinants of debt maturity for small and medium-sized firms, while the variables LARGE*ASSET_MAT, LARGE*GROWTH, LARGE*SIZE, LARGE*VOL_EBIT, LARGE*FIRM_QUALITY and LARGE*LEV capture the differential effect of the firm-level determinants of debt maturity for large firms. The results shown in Table 5 reveal that the effect of country-level determinants on debt maturity is similar to that obtained in Table 4. RULE_OF_LAW, C_RIGHTS and BANK_CONC have a positive relationship to debt maturity, while the relationship is negative between S_RIGHTS and BANK_CREDIT and debt maturity. The exceptions are the coefficients of the protection of investors rights (S_RIGHTS and C_RIGHTS), which lose their statistical significance in columns (2) and (4) when medium-sized firms are not considered. INSERT TABLE 5 ABOUT HERE As far as firm-level determinants are concerned, the results differ substantially between large and small firms. For large firms, columns (1) and (2) show that the coefficients of LARGE*ASSET_MAT, LARGE*GROWTH, LARGE*SIZE, LARGE*VOL_EBIT, LARGE*FIRM_QUALITY and LARGE*LEV are all statistically significant. The negative coefficient of LARGE*ASSET_MAT is consistent with less evidence in favour of the asset maturity hypothesis in the case of large firms. However, asset maturity maintains a direct relationship with debt maturity for all the firms regardless of their size, the effect being more pronounced for small firms. The coefficient of LARGE*GROWTH is in line with the relevance of agency costs in explaining debt maturity in the case of large firms. Regardless of size, firms with long-term investment opportunities prefer to use long-term debt to hedge against liquidity risk, although this interest is stronger in the case of small firms. Table 5 also provides evidence in line with a positive differential effect of SIZE on debt maturity in the case of large firms. The positive coefficient of LARGE*VOL_EBIT lends support to the irrelevance of the tax explanation in the case of large firms. The negative sign of LARGE*FIRM_QUALITY implies that the signalling hypothesis provides a better explanation of the debt maturity of large firms. The negative coefficient of LARGE*LEV suggests that, although more leveraged firms use long-term debt regardless firm size, 14

15 the relevance of leverage as a substitute of maturity in mitigating under- and overinvestment problems is greater for large firms. The results shown in columns (3) and (4) for small firms confirm the previous conclusions. In summary, we find strong support for the asset maturity, signalling and liquidity risk hypotheses and weak evidence for the agency costs hypothesis, regardless of firm size. In the case of large firms, we observe more evidence in favour of the agency costs and signalling hypotheses, while in the case of small firms our findings are in line with a higher weight of the asset maturity and tax hypotheses in explaining corporate debt maturity. Moreover, the results also reveal that the positive relationship between leverage and debt maturity is stronger in small firms. C. Country-level determinants of debt maturity across firm size We also analyse whether the evidence provided in Table 4 for country-level determinants of debt maturity is homogenous across firm size. To do so, we multiply the variables LARGE and SMALL by the country-level determinants of debt maturity, i.e. RULE_OF_LAW, S_RIGHTS, C_RIGHTS, BANK_CONC and BANK_CREDIT. Columns (1) and (3) show the results when the overall sample is used in the estimations, whereas medium-sized firms are excluded from Columns (2) and (4). In column (1), RULE_OF_LAW, S_RIGHTS, C_RIGHTS, BANK_CONC and BANK_CREDIT show the effect of country-level determinants of debt maturity for small and medium-sized firms, while the variables LARGE*RULE_OF_LAW, LARGE*S_RIGHTS, LARGE*C_RIGHTS, LARGE*BANK_CONC and LARGE*BANK_CREDIT capture the differential effect of country-level determinants of debt maturity for large firms. Table 6 documents the results with respect to the variation in country-level determinants of debt maturity across firm size. It reveals that there are differences between large and small firms depending on the country-level variables. Similar to what we found in Table 4, debt maturity increases with RULE_OF_LAW, C_RIGHTS and BANK_CONC and decreases with S_RIGHTS and BANK_CREDIT in the case of small and medium-sized firms (column 1).The protection of investors rights, bank concentration and the weight of banks in the economy have a differential effect in the case of large firms. There is no difference in the effect of RULE_OF_LAW on debt maturity between large and small firms. INSERT TABLE 6 ABOUT HERE The negative coefficient of S_RIGHTS and the positive coefficient of LARGE*S_RIGHTS in column (2) suggest that improvements in the quality of institutional environment increases debt maturity only 15

16 in the case of large firms, while in the case of small firms a strong protection of property rights extend the credit market to low quality borrowers and hence lenders reduce the term of debt. As for the C_RIGHTS variable, the results show that the influence of the protection of creditors rights varies widely across firm size. The joint effect of C_RIGHTS and LARGE*C_RIGHTS or SMALL*C_RIGHTS, depending on whether we consider columns (1) and (2) or columns (3) and (4), suggests that the influence of the protection of creditors rights is positive for medium-sized firms, in line with more favourable terms of credit when the rights of creditors are strongly protected, and is negative in the case of small firms. Furthermore, the protection of creditor rights has no effect in the case of large firms. The negative coefficient of LARGE*BANK_CONC provides evidence in line with the higher relevance of relationship banking for small firms. In line with the expected result, bank concentration serves to mitigate information asymmetries between creditors and debtors, and, due to the fact that information asymmetries are greater in small firms, these will benefit more from bank concentration. Column (4) reveals that bank concentration has no effect on debt maturity for large firms, seeing as the coefficient of BANK_CONC is not significant. The positive influence of bank concentration on debt maturity is concentrated only in small and medium-sized firms. This result is consistent with the evidence of greater financial constraints for US small firms in less concentrated credit markets provided by Petersen and Rajan (1994, 1995) and Berlin and Mester (1999). It is also in line with Hernández-Cánovas and Koëter-Kant (2008), who show that stronger firm-bank relationships lengthen the maturity of bank loans for a sample of 3,366 SMEs from 19 European countries. The negative coefficient of BANK_CREDIT in column (1) reveals that debt maturity decreases with the weight of banks in the economy, in line with the preference of banks for short-term debt in the case of small and medium-sized firms. The positive coefficient of LARGE*BANK_CREDIT shows that the weight of banks in the economy has even a positive influence on debt maturity in the case of large firms. The coefficient of BANK_CREDIT in column (4) is positive, in line with a direct relationship between the weight of banks in the economy and debt maturity in large firms. These results are consistent with the expected negative relationship between the weight of banks in the economy and debt maturity being particularly important in the case of smaller firms. In summary, we find that most of the country-level determinants of corporate debt maturity are size dependent. In line with the expected relationships, the role of bank concentration and the weight of banks in the economy play a particularly relevant role in the case of smaller firms. 16

17 D. Interactions between the institutional and legal environment and banking structure The following step in this research study was to analyse whether the influences of legal enforcement and banking structure are independent of one another, or whether the effects are complementary or substitutive. We create new variables as an interaction between the legal and institutional environment variables RULE_OF_LAW, S_RIGHTS and C_RIGHTS and the banking structure variables BANK_CONC and BANK_CREDIT. For instance, the RULE_OF_LAW*BANK_CONC variable measures the differential effect of bank concentration when the level of legal enforcement is high. The estimations are presented in Table 7 for the overall sample and for the subsamples of large and small firms. INSERT TABLE 7 ABOUT HERE The results for the overall sample reveal similar relationships for firm-level determinants with debt maturity to those highlighted in Table 4. As for the country-level determinants, we can see that RULE_OF_LAW, S_RIGHTS and BANK_CREDIT maintain the already mentioned effect, regardless of the value of the other country variables. That is, RULE_OF_LAW has a positive influence on debt maturity, while S_RIGHTS and BANK_CREDIT have a negative relationship to debt maturity. However, C_RIGHTS has a negative influence on debt maturity and BANK_CONC has no effect on debt maturity. Both variables only maintain a positive relationship to debt maturity when they co-exist in a country. Thus, C_RIGHTS has a negative influence on debt maturity when the level of bank concentration is low. Similarly, BANK_CONC only has a positive effect on debt maturity if, in addition to a high level of bank concentration, creditors rights are better protected. Summing up, the protection of creditors rights and bank concentration do not have a direct positive influence on debt maturity, as we concluded in Table 4. The positive effect of bank concentration and the protection of creditors rights on debt maturity is only significant when both features concur. The negative coefficient of RULE_OF_LAW*BANK_CONC shows that the higher the bank concentration, the lower the positive effect of the efficiency of the legal system on debt maturity. This result is consistent with bank concentration and legal enforcement as alternatives for extending corporate debt maturity 7. The negative coefficient of BANK_CREDIT suggests that suppliers of debt (banks) have a preference for shorter-term debt. However, the positive sign of the interaction term of this variable with S_RIGHTS reveals that this preference is relaxed in high quality institutional environments. 7 As we highlighted before, bank concentration requires strong protection of creditor rights to have a positive influence on debt maturity. 17

18 The results for small firms present only one difference with respect to those obtained for the overall sample. The preference of suppliers of credit for short-term debt is relaxed not only when there is strong protection of property rights, but also when creditors are better protected. The main difference between the results for small and large firms is that the protection of creditors rights acts by increasing the preference of suppliers of credit for short-term debt in the case of large firms. E. Influence of country-level determinants on the firm-level determinants of debt maturity The results in Tables 4 to 7 show that firm, institutional, regulatory and banking structure variables are important determinants of corporate debt maturity. Moreover, their influence depends on firm size. We now analyse whether traditional explanations of debt maturity vary depending on country environment. To do so, we multiply the proxies of the efficiency of the legal system, protection of investors rights and the banking structure variables by the firm-level determinants of debt maturity and carry out the estimations for the overall sample and for the subsamples of large and small firms. On the one hand, the coefficient of the firm-level determinants of debt maturity shows the influence of these variables on debt maturity in countries with low legal enforcement, protection of investors rights, bank concentration or weight of banks in the economy, respectively. On the other hand, the interaction terms RULE_OF_LAW*ASSET_MAT, RULE_OF_LAW*GROWTH, RULE_OF_LAW*SIZE, RULE_OF_LAW*VOL_EBIT, RULE_OF_LAW*FIRM_QUALITY and RULE_OF_LAW*LEV, for instance, capture the differential effect of the variables ASSET_MAT, GROWTH, SIZE, VOL_EBIT, FIRM_QUALITY and LEV for countries with higher levels of legal enforcement. Thus, RULE_OF_LAW*ASSET_MAT in column (1) indicates the difference in the relevance of the matching hypothesis depending on the efficiency of the legal system. The results can be seen in Table 8 for the overall sample and in Table 9 for the subsamples of large and small firms. The results reveal that country-level determinants establish important differences in the role played by firm-level determinants. In countries with low levels of legal efficiency, the firm-level determinants that explain debt maturity are asset maturity, growth, size, firm quality and leverage. The coefficients of the RULE_OF_LAW*ASSET_MAT, RULE_OF_LAW*GROWTH, RULE_OF_LAW*VOL_EBIT and RULE_OF_LAW*FIRM_QUALITY variables are negative and statistically significant. The negative coefficient of RULE_OF_LAW*GROWTH, RULE_OF_LAW*VOL_EBIT and RULE_OF_LAW*FIRM_QUALITY suggest that the evidence in line with the agency cost, tax and signalling hypotheses is greater in countries with high levels of legal efficiency. The negative sign of the interaction term between RULE_OF_LAW and ASSET_MAT reveals that there is more evidence in line with the asset maturity hypothesis in countries with low levels of legal efficiency. The positive 18