Banking Environment and Syndicate Structure: A Cross-Country Analysis
This version: September 2008
Christophe J. GODLEWSKI University of Strasbourg, EM Strasbourg Business School, LaRGE
Pôle Européen de Gestion et d’Economie 61 avenue de la Forêt Noire 67000 Strasbourg, France Mail :
[email protected]. Tel.: +33390242121 / fax: +33390242064.
1
Banking Environment and Syndicate Structure: A Cross-Country Analysis
Abstract We investigate how syndicate structure is influenced by the nature of the banking environment, including the structure of the banking market, financial development, banking regulation and supervision, and legal risk. The results of a cross-country analysis performed on a sample of 15,586 syndicated loan facilities for borrowers from 24 countries over a period of 15 years confirm that syndicate structure is influenced by banking environment consistent with minimizing agency costs and efficient recontracting objectives. Key words: Banking environment, Agency costs, Loan syndication, Syndicate structure. JEL Classification: C31, F30, G21, G32.
2
1. Introduction In 2006, the syndicated loans1 market reached 2.8 trillion USD (Thomson Financial, 2006) and represented more than one third of the funds raised on the worldwide financial markets (Altunbas, Gadanecz and Kara, 2006). This tremendous growth can be attributed to the advantages inherent in syndicated lending for both borrowers and lenders. Lenders can diversify loan portfolios and sources of income, and exploit the comparative advantages of syndicate members through financing and information sharing. Syndication leads to more competitive pricing and more flexible funding structures than bonds or a series of bilateral loans (Allen, 1990; Altunbas and Gadanecz, 2004), thus benefiting borrowers. However, such advantages come at a cost because of the agency problems linked to differences in information between syndicate members. Moral hazard problems may appear if some of the monitoring of the borrower is delegated to an arranger whose efforts are unobservable. Furthermore, if the private information collected by the arranger through due diligence or previous lending relationships cannot be credibly communicated to the participants, then an adverse selection problem arises. These specific problems expose the lenders and the borrower to potential agency costs that might outweigh the benefits of syndication. In organizational terms, such costs can be mitigated through the structure of the syndicate (Pichler and Wilhelm, 2001). The arranger can consequently adapt the size and the composition of the syndicate, which involves explicit and implicit costs and revenue tradeoffs. For instance, the arranger decides on the institutions to invite, chooses the
1
A syndicated loan is a loan which is provided to the borrower by two or more banks that is governed by a single loan agreement. We present the process of bank loan syndication in section 2.
3
initial menu of designated amounts for participation, and the dollar size and associated fees for each bracket. Furthermore, the arranger can adjust his or her portion of the loan to align with monitoring incentives, establish a signal of borrower’s quality, and form a smaller syndicate to fund opaque and risky borrowers. Existing empirical research on syndicate structure by Lee and Mullineaux (2004), Jones, Lang and Nigro (2005), Bosch and Steffen (2006), and Sufi (2007) has demonstrated that syndicates on the US and the UK markets are structured to enhance monitoring efforts and to facilitate renegotiation. However, these studies focus solely on single markets and do not account for the influence of the banking environment on syndicate structure. A large body of research into law and finance pioneered by La Porta, Lopez-De-Silanes, Shleifer and Vishny (1997, 1998, 2000) and recently completed by Djankov, McLiesh and Shleifer (2007) and Qian and Strahan (2007) has shown that the banking environment has a significant impact on corporate ownership, financing policies, capital allocation, design of loan contracts, banking performance and economic and financial development. The nature of the banking environment plays an important role in shaping financial contracts. Lenders must understand their rights as creditors and the efficiency of local enforcement before lending in a given country. Esty and Megginson (2003) found that syndicates funding project finance are larger and more widespread in countries with poorly defined creditor rights and weak legal systems. However, their results focus on a specific loan purpose and so do not provide in-depth insight into how other aspects of the banking environment influence syndicate structure. The banking environment is much more than just legal risk, including aspects like market structure, financial development
4
and regulation, which should all be taken into account by syndicate members when making a decision about whether to provide a loan (Fight, 2004). As the banking environment impacts the agency costs related to lending, syndicate structure should also be adapted to specific features of the banking environment. Therefore, empirical investigation of the relationship between this environment and syndicate structure should provide important normative and policy implications for market structure, financial development, and the regulatory and legal aspects that drive syndicate structure. This could, in turn, lead to more efficient syndicated lending and syndicated loan market development. The aim of this article is to add to existing research by empirically investigating how several features of the banking environment influence the structure of syndicates on a cross-country basis. After carrying out control testing for various loan agreement terms and the financial characteristics of borrowers, we focused on several characteristics of this environment, including banking sector structure, financial development, bank prudential regulations, and legal risk. Using the power of cross-country analysis, we performed our study on a sample of more than 15,000 syndicated loan facilities covering 24 markets over a period of 15 years. We then move on to discuss the theoretical and empirical background of syndication (Section 2) and present the empirical design of our work and discuss our results (Section 3). Finally, in Section 4 we set out our conclusions. 2. Loan syndication and syndicate structure 2.1 Loan syndication and agency problems
5
The process of bank loan syndication can be summarized as follows2: after having solicited competitive offers to arrange and manage the syndication, the borrower chooses one or more arrangers, which is/are entrusted with forming a syndicate, and consequently negotiates a preliminary loan agreement3. The syndication process involves drafting a preliminary loan agreement and preparing a documentation package for potential syndicate members, called an information memorandum. The memorandum contains information about a borrower’s creditworthiness and the loan terms. A roadshow is then organized to present and discuss the content of the memorandum, to present fees, establish a timetable for commitments and closing, formally invite potential participants and determine allocations. Finally, after completion, the loan becomes operational, binding the borrower and the syndicate members to the debt contract. Loan syndication results in several agency problems. First, private information about the borrower can create adverse selection issues, as the arranger may be inclined to syndicate loans for unreliable borrowers. However, such opportunistic behavior can damage the arranger’s reputation, having a negative impact on the success of future syndications (Pichler and Wilhelm, 2001). Secondly, participating banks may delegate monitoring to the arranger, but the banks are not in the loop as to what the arranger is doing, thus potentially resulting in situations of moral hazard. In addition, the arranger has less incentive to monitor the borrower than if it were to lend the full amount of the loan (Pennacchi, 1988). Thirdly, the borrower's financial distress is an important factor in syndication. It is more complicated to reorganize and reformulate the agreement for the
2
See Esty (2001) for a detailed description of the syndication process. The arranger acts as the syndicate’s agent, which includes tasks like administering funds, calculating interest and ensuring covenants are enforced.
3
6
borrower as a collective decision needs to be taken by the lenders (Bolton and Scharfstein, 1996). The structure of the syndicate should be such that it can tackle such agency problems. For example, to provide a credible signal about the quality of the borrower and to align monitoring incentives, the arranger should adjust the portion of the loan it has provided (Jones, Lang and Nigro, 2005; Sufi, 2007). Having a larger number of arrangers generally leads to better handling of agency problems in terms of monitoring the borrower, as several delegated monitors are present, reducing adverse selection problems linked to private information since such information is now spread among several arrangers. Syndicates are usually smaller and more cohesive when there is limited information about the borrower available, when credit risk is relatively high and when a loan is secured (Lee and Mullineaux, 2004). Furthermore, having multiple specialized coagents can mitigate adverse selection problems resulting from arranger monitoring. In addition, it is likely that some of the arrangers will act as specialized agents during the syndication, thus resulting in the process being handled better, with increased cost efficiency and reduced informational asymmetry (François and Missionier-Piera, 2007). Overall, syndicates appear to be structured in a consistent manner that makes it possible to handle agency costs. Syndicate structure is adapted in order to enhance monitoring and facilitate renegotiation. The business, regulatory and competitive environment where the syndicate operates can also influence a syndicate’s structure. Indeed, a country’s legal framework can have an impact on agency problems for syndicated loans and the efficiency of debt restructuring as legal institutions and legal risk affect the way banks carry out
7
governance, especially monitoring and re-contracting (Esty and Megginson, 2003). For example, as shown by La Porta, Lopez-De-Silanes, Shleifer and Vishny (1997, 1998, 2000), the legal framework has a significant influence on corporate ownership, financing policies and capital allocation. Recently, Qian and Strahan (2007) showed that the price and non-price terms of bank loan contracts are linked to the legal and financial environment. For bank loan syndication, Esty and Megginson (2003) have shown that the size and concentration of bank syndicates funding project finance is linked, to some degree, to the legal risk involved in the loan process. Professionals also point out the important role that the banking environment plays in driving the syndicated lending market (Fight, 2004). Given this, we also looked at the impact of the legal environment, especially since this influences both agency and re-contracting risks, and thus can influence syndicate structure. 2.2 Determinants of syndicate structure Now that we have looked, both theoretically and empirically, at the ties between syndicate structure, agency problems and the banking environment, we can turn to examine various country-level determinants that can reasonably be expected to influence syndicate size. Syndicate size is defined by two dependent variables: Number of Lenders and Number of Arrangers, similar to Lee and Mullineaux (2004) and Sufi (2007). We make this distinction because senior members of the syndicate have different concerns and motivations than other participants4. Banking structure is a two-variable proxy. We considered Overheads, defined as the ratio of banking overhead costs to total assets, as a proxy of overall cost inefficiency
4
To avoid biased results, we do not distinguish the number of participants, as the same financial institution can have several roles in a syndicate, being simultaneously an arranger and a participant.
8
in a banking industry5. Cost-efficient banks should be more effective at risk management, screening and monitoring and consequently less exposed to hidden information problems within the syndicate. Indeed, Berger et DeYoung (1997) have shown that more costefficient banks are better at risk management as they have fewer non-performing loans in their portfolios. In addition, since syndicated loans imply the sharing of origination costs, cost inefficiency is expected to encourage the formation of larger syndicates. Consequently, we could expect this variable to have a positive coefficient. Concentration, defined as the assets of the three largest banks as a share of all bank assets, is a proxy for market structure in the industry. Several arguments suggest this variable should have a negative influence on syndicate size. First, greater concentration means a lower number of potential syndicate members. Secondly, banks with greater market share already benefit from diversified loan portfolios and have little incentive to diversify further. Indeed, Demsetz and Strahan (1997) found that large US bank holding companies were better diversified than small ones, while Acharya, Hasan and Saunders (2006) provided empirical evidence in favor of a negative correlation between size and assets as well as geographical concentration for Italian banks. Finally, the incentive provided by increased revenue from syndicated loans should exert less pull on more profitable banks, largely due to greater market power. We also included two variables linked to the development of financial markets. Stock Markets is defined as the value of listed shares to GDP and measures the development of stock markets. Allen and Gottesman (2006) have shown that stock markets and syndicated loan markets are highly integrated, facilitating information flow
5
A more common proxy would be the cost to income ratio but such information is unavailable in our dataset.
9
among markets. The development of stock markets contributes to information disclosure. The availability of information mitigates adverse selection problems that result from the private information that the lead bank has about the borrower. Consequently, the coefficient for this variable should be positive. However, some companies use stock markets as an alternative source of financing for large loans, meaning that as a stock market becomes more developed, loan syndication becomes a less attractive option. Subsequently, one sees an increase in the share of non-syndicated bank loans. Such influence should be even more prominent for the development of bond markets, measured by the sum of private and public domestic debt securities to GDP. When it comes to major financing for companies, bonds compete directly with syndicated loans. However, this negative side might be offset by the positive effects of bond markets, since they contribute to increased information for participant banks in loan syndicates and thus limit adverse selection problems. Banking regulation is our third category of variables for the banking environment. We started by constructing the Mincar x Credit Risk variable. This is the product of the minimum capital requirement value and a dummy variable equal to one if the minimum regulatory capital ratio varies with bank credit risk. On the one hand, this variable should have a positive coefficient. The existence of capital requirements should favor syndication because of the need to respect lending limits. This also takes into account the notion that stronger requirements increase the motivation to meet this requirement. On the other hand, it might also have a negative coefficient. The capital requirement reduces the number of potential syndication participants that have adequate capitalization and can benefit from such funding advantages. Regulation on lending abroad should have a
10
positive impact on syndicate size, because such regulation reduces the opportunities that domestic banks have to diversify. Therefore, Abroad Loan Prohibited should have a positive coefficient. If there is a ban on loan funding abroad, it is a dummy variable equal to one since such prohibitions make syndication an attractive way to diversify a loan portfolio. Three variables are used for supervisory mechanisms: NPL Definition (a dummy variable equal to one if a formal definition of non-performing loans exists), Public Risk Disclosure (a dummy variable equal to one if regulations are imposed on a bank’s public disclosure of their risk management procedures), and Examination Frequency (a variable equal to one, two or three if the frequency of onsite inspections occurs once, twice or three times a year). If binding, these regulatory features should have a positive influence on syndicate size as they enhance the transparency of participant banks loan portfolios through supervisory discipline6. Indeed, enhanced transparency within the syndicate reduces agency costs and thus allows larger syndicates. Our fourth and final category of banking environment variables considers the legal environment, which is taken as the framework within which contracts are enacted and enforced, that is, the legal framework of where the borrower is located. In contrast to financing contracts that are typically governed by New York or UK law, operating contracts and the enforcement of security provisions depend on the legal system in the country where the borrower is located (Esty and Megginson, 2003). As a result, lenders must understand their rights as creditors and the efficiency of local enforcement before lending in a given country. 6
The other variables considered for regulatory discipline and disclosure were the obligation to publicly disclose off-balance sheet items and the presence of a public or private credit registry, but we did not include them in our estimations as they account for more than 95% of the loans in the sample.
11
We included two indicators in our analysis for legal institutions. Protection of creditor rights and law enforcement are measured with the Creditor Rights and Rule of Law indices. The Rule of Law ranges from zero to ten with a higher score indicating better law enforcement while Creditor Rights are scored on a scale from zero to four with a higher score indicating better creditor protection. The expected sign of the coefficient for both variables is ambiguous. Esty and Megginson (2003) found that syndicates funding project finance are larger and more diffuse in countries with poorly defined creditor rights and weak legal systems. Thus, lenders structure syndicates to facilitate recontracting in countries where creditors have strong and enforceable rights. Additionally, better legal protection for banks mitigates the moral hazard problem induced by syndicated loans. Indeed, better creditor protection decreases the need to monitor the borrower. This reduces agency problems resulting from monitoring efforts by those involved in the syndicate. Furthermore, in countries where legal risk is high, the efficient reorganization of a borrower in trouble might be difficult. Hence, a larger syndicate structure is better suited to deal with these issues. However, on a more global basis, agency problems resulting from lending decisions should also be mitigated, which may favor the choice of a standard loan rather than a syndicated loan for the lead bank. Indeed, the desire to share risk should play less of a role in well-protected legal environments. Hence, monitoring should be more important where legal risk is high (few legal rights and low contract enforcement), through a smaller syndicate and/or a larger number of arrangers. Finally, we also took into account the origin of the legal structure in the borrower country through a dummy variable that is equal to one if it is French.
12
Following previous studies of syndicate structure (Lee and Mullineaux, 2004; Sufi, 2007), we decided to use control variables for several different loan agreements and borrower characteristics. We included the following loan characteristics: size and maturity, the availability of public information (dummy variable equal to one if a Standard and Poor’s senior debt rating is available) and borrower reputation (equal to the occurrence of a particular borrower in the sample7), and four dummies that take lender protection into account (Guarantors, Sponsors, Covenants, Senior Debt). We also took into account the number of syndicated loan facilities per country as a control for syndicated lending market development and type (Term Loan), purpose (General Corporate, Debt Repayment, Project Finance, Working Capital) and the loan benchmark rate (Libor and Euribor). Once again, this is done through the inclusion of dummy variables8. The regressions also have dummy variables for year, region and industry. Regarding borrower’s characteristics, following Sufi (2007) and Bosch and Steffen (2007), we focused on its creditworthiness, through the inclusion of four control variables: size (logarithm of Total Assets), profitability (EBIT / Total Assets), interest coverage (EBITDA / Interest Expenses), and leverage (Total Debt / Total Capital). 3. Empirical design and results 3.1 Data and methodology The syndicated loans sample was obtained from the Dealscan database, provided by the Loan Pricing Corporation (LPC, Reuters). Data about financial structure and regulatory and supervisory characteristics were gathered from Beck, Demirgüç-Kunt and 7
Bharath, Dahiya, Saunders and Srinivasan (2007) and Champagne and Kryzanowski (2007) construct similar indicators to investigate the benefits of lending relationships for banks and the determinants of alliances in syndicates respectively. 8 We do not provide variables for other types and purposes in our regressions, since they represent less than 5% of the loans in the sample.
13
Levine (2000) and Barth, Caprio and Levine (2005). Legal environment indicators came from La Porta, Lopez-De-Silanes, Shleifer and Vishny (1998) and Djankov, McLiesh and Shleifer (2007). Borrower’s characteristics were extracted from the Compustat database9. The sample size was determined by the availability of data for the variables used in regression analyses. Following Lee and Mullineaux (2004), we only used completed and fully confirmed deals, excluding private placements. We ultimately generated a full sample of 15,586 loan facilities for borrowers from 24 countries for the period between 1990 and 2005. The number of loan facilities, the average number of lenders and arrangers, and lending amount by country are displayed in Table 1, while Table 2 provides descriptive statistics for the variables. Definitions of the variables are provided in Table A.1 in the appendix. Japan appears as the largest syndicated loan market with more than 4,000 loan facilities operating over the period in question, followed by Australia, Germany, Spain, South Korea and Taiwan, each with more than 1,000 facilities. On average, the number of lenders ranges between 5 and 14, with between 2 and 7 arrangers. Finally, the average loan size is between 100 and 1,000 million USD, with the largest average deals approved in Germany. The results from the full sample descriptive statistics show that the average syndicate has almost nine lenders and three arrangers, with an average loan size of 333 million USD for an average maturity of five years. In comparison, from 1987 to 1995 in the USA, Lee and Mullineaux (2004) report an average number of nine lenders, with an average loan size equal to 221 million USD and an average maturity of four years. Using
9
We match borrowers by their country, name and industry sector. This procedure reduces the size of the full sample. Borrower’s variables have a one-year lag compared to the year of the syndicated loan.
14
a more recent time span (1992-2003), Sufi (2007) observed an average number of eight lenders and two arrangers. The average loan amount was 364 million USD with an average maturity of three years. On average, bank markets are concentrated and relatively cost efficient, and stock and bond market capitalization are important. Regulatory mechanisms such as nonperforming loans are quite common, unlike abroad lending prohibitions and public risk disclosure. On-site inspection of banks occurs more than once a year. Finally, the average legal environment is satisfactory, with creditor rights indices above seven and the rule of law index in the mid-range of the scale. Following Sufi (2007) and Bosch and Steffen (2007) and due to the fact that the dependent variables have large numbers of values (from 2 to 80 for Number of Lenders and from 1 to 27 for Number of Arrangers), we estimated the following set of individual equations using OLS regressions with robust standard errors and clusters at the borrower level: Number of Lenders = f(Banking Environment, Loan Characteristics, Borrower Characteristics) (1) Number of Arrangers = g(Banking Environment, Loan Characteristics, Borrower Characteristics) (2)
3.2 Results and discussion We performed three series of regressions for equations (1) and (2) with different sets of explanatory variables. The results are provided in Tables 3 and 4. In Table 3, the first set of regressions (1.1 and 1.2) is a benchmark displaying only loan agreement and borrower characteristics. We dropped borrower characteristics in the second set (2.1 and 2.2), including banking environment characteristics instead. Finally, the third set (1.3 and
15
2.3) in Table 4 includes all characteristics and also serves to evaluate the robustness of our results. First, we observed that all regression analyses have satisfactory explanatory power with R² equal to at least 30%. We also noted that most of the loan agreement characteristics were significant and robust across all the regression analyses. Loan Size came out with a positive coefficient and was significant for all regressions, suggesting that larger syndicates form around larger loans to diversify loan portfolios and in line with regulatory-driven issues (as in Lee and Mullineaux, 2004; Sufi, 2007). The Maturity coefficient demonstrates notably positive results for most estimates. This can be explained by a negative coefficient between maturity and credit risk (Sharpe et al., 2000) and is congruent with prior research on the US syndicated market. We also observed that borrower transparency positively influenced syndicate size, in line with previous results from Lee and Mullineaux (2004) and Sufi (2007). These results confirm that increased transparency reduces both adverse selection and moral hazard problems within the syndicate. This result was confirmed by a significant and positive coefficient for Borrower Presence, as increased reputation leads to lower information asymmetry and thus fewer syndicate agency problems. Variables assessing lender protection mechanisms reveal that Senior Debt always results in a negative coefficient and is significant in equation (2) and positive in equation (1), while the presence of financial covenants positively affects both the number of lenders and of arrangers, the latter also being influenced by the presence of guarantors. Hence, debt seniority works as an effective protection device for all lenders, reducing agency problems within the syndicate and allowing larger syndicates with fewer
16
arrangers. The presence of a guarantor suggests increased loan risk (Berger and Udell, 1990; Jimenez and Saurina, 2004) and, consequently, a loan plagued by greater agency problems, for which a larger syndicate “core” composed of numerous arrangers aids more effective monitoring. Finally, restricting the discretionary power of the borrower through covenants effectively reduces the risk of loan default (Rajan and Winton, 1995), and enhances the ability to monitor the borrower, thereby reducing monitoring costs and leading to larger syndicates10. Overall, the results for the loan characteristics support prior findings by Lee and Mullineaux (2004), Bosch and Steffen (2006) and Sufi (2007). As regards borrower characteristics, larger firms are associated with larger syndicates, in a similar manner as loan size. In addition, more profitable borrowers reveal smaller syndicate structures, while leverage has no statistically significant effect on syndicate size. In the analysis of banking environment variables in specifications (1.2) and (2.2), the results show significance for most variables. Our results demonstrate that all banking environment characteristics are important for the syndicate structure since we noted significant coefficients for legal environment, financial development, and banking structure and regulation. As expected, the cost level of the banking industry exerts a positive influence on the number of arrangers. Inefficient banks might have less incentive to monitor the borrower and therefore require a greater number of arrangers to perform this task. Moreover, the sharing of origination costs by arrangers helps to explain the findings. Our results show that the correlation between banking industry concentration and the number
10
Most of the dummies for loan type, purpose and benchmark rate appear as significant in the regressions, suggesting that all loan agreement characteristics influence syndicate structure.
17
of lenders is negative. Greater concentration lowers the number of participants that might join and form a syndicate, since banks with greater market share already benefit from diverse loan portfolios and have little incentive to diversify further. In addition, the development of bond markets has a positive correlation to the number of lenders and a negative one to the number of arrangers, suggesting that improved financial development helps information flow and transparency, thus mitigating agency problems in the syndicate and allowing the syndicate to grow with fewer arrangers. The Mincar x Credit Risk coefficient is negative (2.2) suggesting that binding capital requirements can reduce the number of syndicate members that have adequate capitalization and can benefit from such funding advantages. As expected, regulation on lending abroad had a positive influence on syndicate size. Such regulation reduces the number of opportunities for domestic banks to diversify and increases their “appetite” to fund a share of a syndicated loan as part of diversifying their portfolio. Among the proxies for regulatory discipline and transparency, we saw that Examination Frequency and Public Risk Disclosure have negative coefficients. Public Risk Disclosure only influences the number of lenders, suggesting that such supervisory devices tend to inform the arrangers that a large portion of potential local lenders have weak or inefficient risk management procedures and thus do not qualify for the syndicate. Less frequent on-site inspections has a negative relation to the number of arrangers, suggesting this form of supervision takes the place of monitoring by the arrangers. The NPL Definition coefficient is significant and positive in both specifications. These results suggest that greater transparency regarding problem loan classification could reduce friction linked to information within the syndicate and allow the establishment of larger groups of lenders.
18
Finally, we noted that enhanced protection of creditor rights has a negative and significant influence on the number of lenders but no impact on the number of arrangers. Better protection of creditors might reduce the incentives for lenders to monitor borrowers. However, this may exacerbate free-riding problems, which can be tackled through an adapted small-size syndicate structure. Furthermore, creditors could benefit from such protection and monitoring of the borrower by the syndicate to avoid inefficient re-contracting should problems arise. Therefore, smaller syndicates with larger cores are more suitable for such tasks. As the quality of institutions increases (i.e. legal risk decreases), the number of arrangers diminishes as monitoring is more effective in such a legal framework. These results are congruent with the findings of Esty and Megginson (2003). Our last set of regressions (1.3 and 2.3) in Table 4 was performed as a control for all individual characteristics (both loan agreement and borrower) and the banking environment, and served as a robustness check. We noted that most borrower risk proxies are significant and the more risky firms are associated with smaller syndicates, which serve to mitigate agency problems. Furthermore, we also saw, from the controls we carried out to test for borrower risk, that several banking environment characteristics influenced both lenders and arrangers in areas such as bank concentration and examination frequency, while most of the remaining coefficients gained in magnitude. Thus, the controls for borrower characteristics showed once again the influence of the banking environment on syndicate structure, suggesting that this environment is an important driver of loan syndication. 3.3 Robustness checks
19
We performed several robustness checks for different explanatory variable proxies and other econometric techniques. First, we used Poisson and negative binomial regressions, achieving virtually the same results. We also performed several reduced form regressions omitting potentially internal variables such as Maturity, Guarantors or Log (Loan size), as a lack of data prevented us from using an instrumental variables approach. These regressions led to similar coefficients - and thus conclusions - regarding banking environment variables. Second, using other measures of borrower transparency, such as Moody’s Senior Debt rating or a dummy variable equal to one if a Moody’s or a Standard and Poor’s rating is available, led to similar results. Including alternative proxies for borrower risk, such as the Quick Ratio or the ratio of Net Income to Total Assets, also produces similar results. The latter are also robust when alternative proxies for financial structure, such as the ratio of private credit of financial institutions to GDP, are used. Qualitatively, our results were unaffected by the use of the Creditor Rights index components from La Porta, Lopez-De-Silanes, Shleifer and Vishny (1998)11. Replacing Rule of Law with alternative proxies from La Porta, Lopez-De-Silanes, Shleifer and Vishny (1997, 1998) such as Risk of Expropriation, Repudiation of Contracts, Corruption or Judicial System Efficiency did not alter our results12. Finally, we obtained similar results when we replaced the legal origin variable with the English Law dummy.
11
The components are: Secured Creditor Paid First, Restriction on Reorganization, or Management Stays. These dummies equal one if secured creditors are ranked first in the distribution of proceeds that result from the disposition of assets of a bankrupt firm, if the reorganization procedure imposes restrictions, such as creditors’ consent, to file for reorganization, or if the debtor keeps the administration of its property pending the resolution of the reorganization process respectively. 12 These variables are defined as indexes, scaled from 0 to 10 with lower scores for higher risks, corruption or inefficiency, assessing the risk of “outright confiscation” or “forced nationalization” and of the ‘‘risk of a modification in a contract”, the level of corruption in government or assessing the “efficiency and integrity of the legal environment as it affects business” respectively.
20
We also performed several regressions that took the banking environment of the lender’s country into account. Descriptive statistics for these variables are displayed in Table A.213. Indeed, syndicated loans are usually funded by lenders from different countries, and taking borrower country banking environment variables into account could be considered as only controlling for borrower country characteristics. In our sample, on average, 26.81% of syndicate members are from the same country as the borrower, while they have the same legal origin in 27.28% of the cases. The banking environment variables for the lead arranger country are linked to the lead arranger’s hierarchy in the syndicate, which we inferred from the lender’s title. When the lead arranger’s title did not enable us to determine its rank, we used banking environment variables for the lender with the most prestigious title and/or holding the largest portion of the loan. In line with the arguments developed in sub-section 2.2 related to factors driving the operating contracts, we did not include legal risk measures for the lender country. These results are displayed in Table 5. We performed two series of regressions: (1.4) and (2.4) that only had lender banking environment and loan characteristics, and the other series that added in borrower country legal risk measures, (1.5) and (2.5)14. We compare these results to (1.2) and (2.2) in Table 3. First, we observed that most of the loan characteristics coefficients remained unchanged across the regressions in Table 5. Borrower legal risk characteristics in regressions (1.5) and (2.5) exhibit similar coefficient signs as in regressions (1.2) and 13
Lender banking environment variables are defined like the borrower banking environment variables but are computed for the lender country. Due to the correlation structure, we cannot include all of the borrower and lender banking environment variables in the same regressions. Ideally, we would perform the regressions on sub-samples where the percentage of lenders from the same country or legal origin as the borrower is the lowest (respectively the highest) to provide better insight into the effect of banking environment on syndicate structure. Unfortunately, as this information is scarce in our data, we cannot proceed that way. 14 The available data made it impossible to include borrower characteristics in the regressions.
21
(2.2). Regarding lender country market structure, we observed that less costly and more concentrated markets lead to larger syndicates. More efficient leaders are more likely to enter a syndicate because they are valuable partners for better management and monitoring during the syndication. Also, more concentrated leaders enter syndicates to diversify. We found that the development of the stock market has a negative influence on syndicate size. Bank regulation characteristics, such as regulatory capital requirements, the definition of non-performing loans and a lending abroad prohibition had a similar influence to that in regressions (1.2) and (2.2) and thus led to similar interpretations. 4. Conclusion Our study addressed the influence of the banking environment on syndicate structure in 24 countries over a period of 15 years. Our results demonstrated that most banking environment characteristics, such as financial development, banking regulation, and legal environment, have a significant influence on syndicate structure. Overall, the structures of syndicates are adapted to enhance monitoring of the borrower and to increase the efficiency of the re-contracting process in case of borrower troubles. The primary reasons for syndication include diversifying loan portfolios, regulatory pressure and reducing management costs. The influence that the tested variables had suggests that certain reasons are more prominent when it comes to forming syndicates with adapted structures. Syndicates are structured to minimize agency problems related to loan and borrower characteristics and a country’s financial, regulatory and institutional environments. In more costly banking industries, there are larger syndicates, while more concentrated banking industries reduce the number of lenders. In certain bond market, financial development positively
22
influenced the size of the syndicate. Bank capital, banking regulation and loan portfolio transparency all have a positive influence on syndicate size. Finally, syndicates are structured consistently to mitigate legal risk. Our results are congruent with previous research, which shows that syndicate structure is adapted to the specific agency problems related to syndication and recontracting issues (Esty and Megginson, 2003; Lee and Mullineaux, 2004; Jones, Lang and Nigro, 2005; Sufi, 2007). Our results were obtained using a cross-country sample of more than 15,000 syndicated loan facilities from 24 countries and over a period of 15 years, further strengthening previous findings. Furthermore, we also showed that apart from legal risk, syndicate structure is also influenced by other components of the banking environment, such as banking market structure, financial development, banking regulation, and prudential supervision. Indeed, syndicate lenders must understand market structure, financial development, regulation and legal risk before lending in a given country. Therefore, we can infer that financial regulators, for normative policy, should take all of the banking environment components into account in order to promote efficient syndicated lending markets with adequately designed syndicates. This would allow for more efficient syndicated lending and syndicated loan market development. References Acharya, V.V., I. Hasan and A. Saunders, 2006. Should banks be diversified ? Evidence from individual bank loan portfolios, Journal of Business 79, 1355-1412. Allen, T., 1990. Developments in the international syndicated loan market in the 1980s, Bank of England Quarterly Bulletin, February, 71-77.
23
Allen, L. and A. A. Gottesman, 2006. The informational efficiency of the equity market as compared to the syndicated bank loan market, Journal of Financial Services Research 30, 5-42. Altunbas, Y. and B. Gadanecz, 2004. Developing country economic structure and the pricing of syndicated credits, Journal of Development Studies 40, 143-173. Altunbas, Y., B. Gadanecz and A. Kara, 2006. Syndicated Loans A Hybrid of Relationship Lending and Publicly Traded Debt. (Palgrave MacMillan Studies in Banking and Financial Institutions). Barth, J. R., G. Caprio and R. Levine, 2005. Rethinking Bank Regulation: Till Angels Govern. (Cambridge University Press). Beck, T., A. Demirgüç-Kunt and R. Levine, 2000. A new database on financial development and structure, World Bank Economic Review 14, 597-605. Berger, A. N. and G. F. Udell, 1990. Collateral, loan quality, and bank risk, Journal of Monetary Economics 25, 21-42. Berger, A. N. and R. DeYoung, 1997. Problem loans and cost efficiency in commercial banks, Journal of Banking and Finance 21, 849-870. Bharath, S. T., S. Dahiya, A. Saunders and A. Srinivasan, 2007. So what do I get ? The bank’s view of lending relationships, Journal of Financial Economics 85, 368-419. Bolton, P. and D. Scharfstein, 1996. Optimal debt structure and the number of creditors, Journal of Political Economy 104, 1–25. Bosch, O. and S. Steffen, 2006. Informed lending and the structure of loan syndicates – Evidence from the European syndicated loan market, Working Paper, Goethe University Frankfurt.
24
Demsetz, R.S. and P.E. Strahan, 1997. Diversification, size, and risk at bank holding companies, Journal of Money, Credit, and Banking 29, 300-313. Dennis, S., N. Debarshi and I. G. Sharpe, 2000. Determinants of contract terms in bank revolving credit agreements, Journal of Financial and Quantitative Analysis 35, 87109. Djankov, S., C. McLiesh and A. Shleifer, 2007. Private credit in 129 countries, Journal of Financial Economics 84, 299-329. Esty, B.C., 2001. Structuring loan syndicates: A case study of the Hong Kong Disneyland project loan, Journal of Applied Corporate Finance 14, 80-95. Esty, B. C. and W. L. Megginson, 2003. Creditor rights, enforcement, and debt ownership structure: Evidence from the global syndicated loan market, Journal of Financial and Quantitative Analysis 38, 37-59. Fight, A., 2004. Syndicated Lending. Elsevier. François, P. and F. Missionier-Piera, 2007. The agency structure of loan syndicates, The Financial Review 42, 227-245. Jimenez, G. and J. Saurina, 2004. Collateral, type of lender and relationship banking as determinants of credit risk, Journal of Banking and Finance 28, 2191-2212. Jones, J. D., W. W. Lang and P. J. Nigro, 2005. Agent behavior in bank loan syndications, Journal of Financial Research 28, 385-402. Champagne C. and L Kryzanowski, 2007. Are current syndicated loan alliances related to past alliances?, Journal of Banking and Finance 31, 3145-3161. La Porta, R, F Lopez-De-Silanes, A Shleifer and R W. Vishny, 1997. Legal determinants of external finance, Journal of Finance 52, 1131-1150.
25
La Porta, R, F Lopez-De-Silanes, A Shleifer and R W. Vishny, 1998. Law and finance, Journal of Political Economy 106, 1113-1155. La Porta, R, F Lopez-De-Silanes, A Shleifer and R W. Vishny, 2000. Investor protection and corporate governance, Journal of Financial Economics 58, 3-27. Lee, S. W. and D. J. Mullineaux, 2004. Monitoring, financial distress, and the structure of commercial lending syndicates, Financial Management 33, 107-130. Pennacchi, G. G., 1998. Loan sales and the cost of bank capital, Journal of Finance 43, 375-396. Pichler, P and W Wilhelm, 2001. A theory of the syndicate: Form follows function, Journal of Finance 56, 2237-2264. Qian, J and P E. Strahan, 2007. How laws and institutions shape financial contracts: The case of bank loans, Journal of Finance 62, 2803-2834. Rajan, R. and A. Winton, 1995. Covenants and collateral as incentives to monitor, Journal of Finance 50, 1113-1146. Sufi, A., 2007. Information asymmetry and financing arrangements: Evidence from syndicated loans, Journal of Finance 62, 629-668. Thomson Financial, 2006. Syndicated Loans Review (Thomson).
26
Table 1 Syndicated loan facilities and average number of lenders, arrangers and loan size by borrower country The table provides frequencies of loan facilities, the average number of lenders and of arrangers, and the mean loan size (in million USD) by country for the full sample. Borrower Syndicated Number of Number of Loan Country loans Lenders Arrangers Size Argentina 246 8.59 2.92 195 Australia 1,146 7.02 2.59 407 Austria 43 11.33 5.26 370 Belgium 170 14.18 4.40 847 Brazil 330 9.11 2.85 218 Chile 226 9.98 3.65 231 Denmark 115 10.35 3.84 666 Finland 116 10.78 7.09 585 Germany 1,006 11.65 5.66 1,000 India 429 8.61 2.63 119 Indonesia 686 10.49 2.44 128 Ireland 165 11.24 3.91 464 Italy 688 11.59 3.52 758 Japan 3,954 5.91 1.25 162 Korea (South) 1,636 8.61 3.37 201 Malaysia 377 7.52 2.36 204 Mexico 463 10.73 3.72 302 Netherlands 722 9.86 3.34 604 Peru 39 5.46 2.67 123 Philippines 146 9.31 2.75 162 South Africa 124 14.23 5.63 359 Spain 1,042 11.64 4.07 526 Taiwan 1,163 10.11 2.03 169 Thailand 554 9.58 2.68 116 15,586 8.86 2.53 333
27
Table 2 Descriptive statistics The table provides descriptive statistics computed on our dataset of loan facilities. Definition of variables appears in table A.1 in the appendix. Std. dev.: standard deviation, Min.: minimum, Max.: maximum. Variable Number of Mean Std. dev. Min. Max. observations Dependent variables Number of Lenders 15,586 8.8611 7.4639 2.0000 80.0000 Number of Arrangers 11,550 2.5324 2.7058 1.0000 27.0000 Borrower banking environment variables Overheads 15,586 0.0275 0.0160 0.0023 0.1448 Concentration 15,586 0.5008 0.1551 0.2701 1.0000 Stock Markets 15,586 0.7012 0.3701 0.0773 2.8243 Bond Markets 15,586 0.8706 0.5859 0.0168 1.8780 NPL Definition 15,586 0.7005 0.4581 0.0000 1.0000 Mincar 15,586 8.1578 0.6387 8.0000 11.500 Credit Risk 15,586 0.1600 0.3666 0.0000 1.0000 Abroad Loan Prohibited 15,586 0.2236 0.4167 0.0000 1.0000 Examination Frequency 15,586 1.6148 0.6882 1.0000 3.0000 Public Risk Disclosure 15,586 0.3947 0.4888 0.0000 1.0000 Rule of Law 15,586 7.7484 1.9307 2.5000 10.0000 Creditor Rights 15,586 2.2029 0.9272 0.0000 4.0000 Loan agreement control variables Loan Size 15,586 333 1020 22230 81100 Maturity 15,586 56.5119 42.8570 1.0000 480.0000 Guarantors 15,586 0.0932 0.2908 0.0000 1.0000 Sponsors 15,586 0.0956 0.2940 0.0000 1.0000 Covenants 15,586 0.1759 0.3807 0.0000 1.0000 Senior Debt 15,586 0.7350 0.4414 0.0000 1.0000 S & P Rating 15,586 0.0694 0.2542 0.0000 1.0000 Borrower Presence 15,586 8.0387 12.5907 1.0000 314.0000 Borrower control variables Log (Total Assets) 12,473 6.9727 1.6869 2.9929 12.9182 EBITDA / Interest Expenses 11,311 19.0546 264.4381 -6.6916 18668.0000 EBIT / Total Assets 8,655 0.1443 0.0990 -0.1898 2.5506 Total Debt / Capital 12,249 81.4441 581.7350 3.7590 58404.5500
28
Table 3 OLS regressions explaining the structure of syndicates The table below shows the results of OLS regressions with robust standard errors (s.e.) and clusters at the borrower level. The dependent variables are Number of Lenders and Number of Arrangers (equations 1 and 2 respectively), equal to the number of lenders and arrangers forming the syndicate. Definitions of variables appear in table A.1 in the appendix. Dummy variables for loan type (Term Loan), loan purpose (General Corporate, Debt Repayment, Working Capital, Project Finance), benchmark rate (Libor, Euribor), year, region, industry sector, and Syndicated Loans and French Law variables are included in the regressions but are not reported. ***, **, and * indicate statistically significant coefficients at the 0.01, 0.05, and 0.1 level respectively. Specifications (1.1) (2.1) (1.2) (2.2) Variables Borrower banking environment Overheads Concentration Stock Markets Bond Markets Mincar x Credit Risk Abroad Loan Prohibited NPL Definition Public Risk Disclosure Examination Frequency Creditor Rights Rule of Law Loan characteristics Log(Loan Size) Maturity Guarantors Sponsors Covenants Senior Debt S & P Rating Borrower Presence Borrower characteristics Log (Total Assets) EBITDA / Interest Expenses EBIT / Total Assets Total Debt / Capital Intercept N R²
coef.
s.e.
coef.
s.e.
2.18175*** 0.0087*** -0.0005 0.5164* 1.7881*** 2.6624*** 1.7679*** 0.0371*
0.09 0.00 0.21 0.30 0.23 0.44 0.26 0.02
0.9740*** -0.0027 0.3903** 0.9436*** 0.4564** -2.7349*** 0.3292** -0.0176
0.08 0.00 0.15 0.23 0.18 0.30 0.16 0.01
0.3189*** -0.0002
0.07 0.00
0.0356 -0.0001**
0.05 0.00
-4.1963*** 0.75 -1.4216*** 0.45 -0.0001 0.00 -0.0001 0.00 -33.9686*** 5.88 -12.7928*** 1.21 8,516 4,997 0.3544 0.4026
coef.
s.e.
coef.
s.e.
-12.6337 -2.9464*** 0.4243 0.4746** 0.0205 3.2922*** 0.8296*** -2.7688*** -0.0132 -0.8529*** -0.1007
6.57 0.80 0.25 0.23 0.03 0.23 0.30 0.30 0.14 0.14 0.07
23.1500*** 0.9108 0.0052 -0.6840*** -0.0336** 0.1132 0.3141* 0.0399 -0.2506*** 0.0643 -0.1960***
3.31 0.47 0.10 0.12 0.02 0.12 0.19 0.17 0.09 0.06 0.03
2.5906*** -0.0024 0.126 -0.2587 0.8582*** 1.1604*** 0.1101 0.0424***
0.06 0.00 0.17 0.18 0.16 0.23 0.20 0.01
0.5617*** 0.0026*** 0.0186 -0.1116 0.0872 -0.6835*** 0.1499* 0.0099***
0.02 0.00 0.07 0.07 0.08 0.11 0.09 0.00
-36.5257*** 2.33 15,586 0.3157
-8.0426*** 1.08 11,550 0.3052
29
Table 4 OLS regressions explaining the structure of syndicates – robustness check with borrower characteristics The table below shows the results of OLS regressions with robust standard errors (s.e.) and clusters at the borrower level. The dependent variables are Number of Lenders and Number of Arrangers (equations 1 and 2 respectively), equal to the number of lenders and arrangers forming the syndicate. Definitions of variables appear in table A.1 in the appendix. Dummy variables for loan type (Term Loan), loan purpose (General Corporate, Debt Repayment, Working Capital, Project Finance), benchmark rate (Libor, Euribor), year, region, industry sector, and Syndicated Loans and French Law variables are included in the regressions but are not reported. ***, **, and * indicate statistically significant coefficients at the 0.01, 0.05, and 0.1 level respectively. Specifications (1.3) (2.3) Variables Borrower banking environment Overheads Concentration Stock Markets Bond Markets Mincar x Credit Risk Abroad Loan Prohibited NPL Definition Public Risk Disclosure Examination Frequency Creditor Rights Rule of Law Loan characteristics Log(Loan Size) Maturity Guarantors Sponsors Covenants Senior Debt S & P Rating Borrower Presence Borrower characteristics Log (Total Assets) EBITDA / Interest Expenses EBIT / Total Assets Total Debt / Capital Intercept N R²
coef.
s.e.
coef.
s.e.
55.8104*** 12.31 33.9359*** -3.2173* 1.66 -3.9612*** -1.7274*** 0.58 -0.0381 1.8175*** 0.39 -2.1948*** -0.2137*** 0.07 0.0752 3.1469*** 0.58 0.1716 -0.2205 0.74 1.5261*** -2.3506*** 0.60 -0.1615 -0.9509*** 0.32 -1.1076*** -0.5447*** 0.15 -0.068 -1.2156*** 0.38 0.4281
10.80 1.40 0.36 0.22 0.06 0.43 0.52 0.47 0.22 0.12 0.25
2.2915*** 0.0146*** -0.1163 -0.4464* 1.8098*** 1.5146*** 0.9337*** 0.0830***
0.12 0.00 0.27 0.27 0.29 0.54 0.29 0.03
0.3962*** 0.0036* 0.8367*** 0.0025 0.0556 -0.9982*** 0.5268*** 0.0125
0.05 0.00 0.16 0.11 0.12 0.34 0.15 0.01
-0.2383*** -0.0007***
0.09 0.00
0.055 -0.0001***
0.03 0.00
-1.9728** 0.80 -0.0001*** 0.00 -29.6724*** 3.83 4,721 0.3889
0.4389 0.35 -0.0001 0.00 -4.0459* 2.43 3,699 0.5512
30
Table 5 OLS regressions explaining the structure of syndicates – robustness check with lender country banking environment The table below shows the results of OLS regressions with robust standard errors (s.e.) and clusters at the borrower level. The dependent variables are Number of Lenders and Number of Arrangers (equations 1 and 2 respectively), equal to the number of lenders and arrangers forming the syndicate. Definitions of variables appear in table A.1 in the appendix. Dummy variables for loan type (Term Loan), loan purpose (General Corporate, Debt Repayment, Working Capital, Project Finance), benchmark rate (Libor, Euribor), year, region, industry sector, and Syndicated Loans and French Law variables are included in the regressions but are not reported. ***, **, and * indicate statistically significant coefficients at the 0.01, 0.05, and 0.1 level respectively. Specifications (1.4) (2.4) (1.5) (2.5) Variables coef. s.e. coef. s.e. coef. s.e. coef. s.e. Lender banking environment Overheads (Lender) -38.2461*** 11.73 -11.2453 7.35 -15.1949 12.16 -1.2223 6.85 Concentration (Lender) 5.2451*** 1.47 -0.0153 0.73 3.3560** 1.61 0.2577 0.85 Stock Markets (Lender) -2.2161*** 0.60 -0.2840 0.32 -0.6721 0.64 0.3181 0.35 Bond Markets (Lender) -0.2576 0.48 0.1750 0.25 -0.1149 0.49 0.4643* 0.26 Mincar x Credit Risk (Lender) -0.1726*** 0.06 -0.0716** 0.03 -0.0927 0.07 -0.1124*** 0.03 NPL Definition (Lender) 1.9487*** 0.40 -0.0635 0.24 2.0136*** 0.47 -0.2053 0.30 Abroad Loan Prohibited 4.3909*** 1.20 0.9752 0.67 5.2142*** 1.54 0.7596 0.61 (Lender) Public Risk Disclosure 0.4774 0.36 0.2117 0.20 0.2978 0.45 -0.0627 0.23 (Lender) Examination Frequency 0.0303 0.25 -0.1398 0.14 -0.0605 0.27 -0.0722 0.17 (Lender) Borrower banking environment Creditor Rights -0.7632*** 0.18 -0.6315*** 0.14 Rule of Law -0.3270** 0.15 -0.2413* 0.12 Loan characteristics Log(Loan Size) 3.2451*** 0.16 1.3166*** 0.11 3.4303*** 0.20 1.1875*** 0.13 Maturity -0.0070* 0.00 -0.0020 0.00 -0.0072 0.00 -0.0110*** 0.00 S & P Rating -1.0214* 0.52 -0.2105 0.30 -0.8770* 0.53 0.1508 0.35 Guarantors -0.5738 0.37 0.2193 0.31 -0.7472* 0.43 -0.2711 0.35 Sponsors 0.4135 0.50 0.0842 0.28 0.6488 0.61 0.2142 0.32 Covenants 0.8284* 0.42 -0.4992* 0.27 0.2483 0.49 -0.5618* 0.33 Senior Debt 2.7745*** 0.69 0.4853* 0.29 2.4373*** 0.83 0.5037 0.31 Borrower Presence 0.0224*** 0.00 0.0007 0.00 0.0213*** 0.00 -0.0004 0.00 Intercept -55.8339*** 4.16 -23.2168*** 2.29 -54.8771*** 5.07 -19.2643*** 2.77 N 3,502 1,922 2,543 1,436 R² 0.367 0.3251 0.414 0.4124
31
Table A.1 Brief description of all variables and their sources Variable Description Dependent variables Number of Lenders Number of lenders in the syndicate. Number of Arrangers Number of arrangers in the syndicate. Borrower banking environment variables* Overheads Ratio of banking overhead costs to total banking assets Concentration Assets of the three largest banks as a share of total banking assets Stock Markets Value of listed shares to GDP Bond Markets Public and private domestic debt securities to GDP Mincar Minimum capital requirement value Credit Risk =1 if the minimum regulatory capital ratio varies with bank credit risk Abroad Loan Prohibited =1 if banks are prohibited from granting loans abroad NPL Definition =1 if a formal definition of non-performing loans exists Public Risk Disclosure =1 if regulations impose on banks public disclosure of their risk management procedures Examination Frequency =1, 2 or 3 if the frequency of onsite inspections is 1, 2 or 3 times a year Creditor rights An index aggregating four aspects of creditor rights. The index ranges from zero (weak creditor rights) to four (strong creditor rights) Rule of Law An index indicating law enforcement. The index ranges from zero (weak enforcement) to ten (strong enforcement) French Law =1 if the legal system is based on French law Loan Size Maturity Guarantors Covenants Senior Debt S&P Rating Borrower Presence Syndicated Loans Term Loan Corporate Purposes Debt Repayment Working Capital Project Finance Euribor Libor Log(Total Assets) EBITDA / Interest Expenses EBIT / Total Assets Total Debt / Capital
Loan agreement control variables Size of the loan in million USD Maturity of the loan in months =1 if there is at least one guarantor =1 if the loan agreement includes covenants =1 if debt is senior =1 if the borrower has a senior debt rating by Standard & Poor’s Number of times a particular borrower is present in the full sample Number of syndicated loan facilities by country =1 if the loan is a term loan =1 if the loan purpose is general corporate purposes funding =1 if the loan purpose is debt repayment funding =1 if the loan purpose is working capital funding =1 if the loan purpose is project finance funding =1 if the benchmark rate is Euribor =1 if the benchmark rate is Libor Borrower control variables Logarithm of total assets Earnings before interest, tax, depreciation and amortization / Interest expenses Earnings before interest and tax / total assets Total debt / total capital
Source Dealscan Dealscan Beck et al. (2000) Beck et al. (2000) Beck et al. (2000) Beck et al. (2000) Barth et al. (2005) Barth et al. (2005) Barth et al. (2005) Barth et al. (2005) Barth et al. (2005) Barth et al. (2005) La Porta et al. (1998) / Djankov et al. (2007) La Porta et al. (1998) / Djankov et al. (2007) La Porta et al. (1998) / Djankov et al. (2007) Dealscan Dealscan Dealscan Dealscan Dealscan Dealscan Dealscan Dealscan Dealscan Dealscan Dealscan Dealscan Dealscan Dealscan Compustat Compustat Compustat Compustat
32
Table A.2 Descriptive statistics for lender banking environment variables The table provides descriptive statistics computed on our dataset of loan facilities. Definition of variables appears in table A.1 in the appendix. Std. dev.: standard deviation, Min.: minimum, Max.: maximum. Variable Number of Mean Std. dev. Min. Max. observations Lender banking environment variables Overheads (Lender) 11,541 0.0338 0.0110 0.0020 0.1417 Concentration (Lender) 11,544 0.5627 0.1694 0.2267 1.0000 Stock Markets (Lender) 11,493 0.9399 0.5133 0.0178 4.9921 Bond Markets (Lender) 11,014 0.9276 0.3626 0.0215 2.1817 NPL Definition (Lender) 10,768 0.4257 0.4944 0.0000 1.0000 Mincar (Lender) 9,060 8.1156 0.6383 8.0000 12.0000 Credit Risk (Lender) 10,764 0.2625 0.4400 0.0000 1.0000 Abroad Loan Prohibited (Lender) 10,769 0.0056 0.0750 0.0000 1.0000 Examination Frequency (Lender) 6,045 1.4138 0.7146 0.0000 3.0000 Public Risk Disclosure (Lender) 8,806 0.3399 0.4888 0.0000 1.0000
33
Economics and Management of Financial Institutions - Working Paper Series (Emfi Wps) is published by ADEIMF, the Italian Association of Scholars in the Economics and Management of Financial Intermediation and Markets (www.adeimf.it) Emfi Wps 1 - 2008
Edito nel mese di Novembre 2008 a cura di ADEIMF presso l’Università di Parma via Kennedy, 4 – 6 I - 43100 Parma ISBN 978-88-6351-000-3 Stampato a cura di Lulu Enterprises, Inc. 860 Aviation Parkway, Suite 300 Morrisville, NC 27560 www.lulu.com ID: 4718813
34
