|
Article Excerpt
A KEY ISSUE of interest in the recent literature on financial intermediation has been the role of relationship lending. Relationship lending is particularly common in the case of small business lending because small businesses typically rely on bank loans for a substantial part of their financing needs but also tend to be informationally opaque. An important issue in this context is the use of collateral, which is a common feature of loan contracts between small firms and banks around the world, and a number of theoretical and empirical studies have examined why it is so widespread and how it relates to the incentives for borrowers and lenders and the borrower-lender relationship. For instance, it has been argued that in the presence of information asymmetries between creditors and borrowers, collateral may mitigate the problem of adverse selection (Bester 1985, 1987) and/or the problem of moral hazard (Bester 1994, Boot, Thakor, and Udell 1991). Collateral also affects the incentives of creditors, who will use it either as a substitute for (Manove, Padilla, and Pagano 2001) or complement to (Rajah and Winton 1995, Boot 2000, Longhofer and Santos 2000) screening and monitoring efforts. Another aspect of collateral that studies have concentrated on is that its presence may depend on the length and intimacy of the relationship between creditors and borrowers (Boot 2000, Boot and Thakor 1994, Sharpe 1990). Existing empirical research has yet to reach decisive conclusions about the nature of these relationships.
This paper seeks to contribute to the existing literature on collateral using a unique firm-level data set of the small and medium sized enterprise (SME) loan market in Japan. Explicitly differentiating physical collateral (such as real estate) and personal guarantees by business representatives, we investigate how the use of collateral and personal guarantees affects the incentives of borrowers, lenders, and the relationship between them. More specifically, we examine the following three issues. First, we examine whether riskier borrowers are more likely to be required to provide collateral or personal guarantees. Second, we investigate how collateral and personal guarantees affect banks' monitoring of borrowers. Third, we examine the correlation between the use of collateral and personal guarantees on the one hand and the closeness of borrower-lender relationships on the other.
The data set we employ is based mainly on the "Survey of the Financial Environment" (SFE) conducted by the Small and Medium Enterprise Agency of Japan in October 2002. In order to focus on firms that mostly depend on bank loans for their financing, we limit the sample to firms satisfying the legal definition of an SME in Japan. We then combine the SFE data for each SME with information on their main bank obtained from the bank's financial statements in order to control for lender characteristics as well. Furthermore, to control for the effect of government credit guarantees on collateral and personal guarantees, in the main analysis of this paper we exclude from the sample all firms that enjoyed any form of government credit guarantee. (1) As a result of this screening process, we end up with a sample of 1,702 firms.
Our main findings can be summarized as follows. We find that firms' riskiness does not have a significant effect on the likelihood that collateral is used. Thus, we cannot find firm evidence that the use of collateral mitigates moral hazard. We find, however, that banks whose claims are collateralized monitor borrowers more intensively, and that borrowers who have a long-term relationship with their main banks are more likely to pledge collateral. These findings suggest that collateral is complementary to relationship lending. In contrast, the complementarity between relationship lending and personal guarantees is weaker.
As far as we know, this is the first empirical study that systematically examines the role of collateral and personal guarantees in Japan's SME loan market. The two main contributions of the paper are as follows. First, given that Japan is generally considered to have a relationship-based financial system in which the relationship lender, the main bank, plays a central role in corporate financing (Rajan and Zingales 2003), the study helps to improve our understanding of the role of collateral in relationship lending and complements existing studies that focus on the United States and Europe. Second, and more importantly, by distinguishing collateral and personal guarantees, the study detects an important role of collateral in relationship lending that has not been remarked on much before. As we argue subsequently, although a typical SME in Japan has a long-term relationship with its main bank, it actually engages in transactions with several banks, which is not common in other countries. A possible corollary of this is that because of the informational free-rider problem it creates, this practice may reduce the main bank's incentive to screen and monitor borrowers. Since collateral defines the order of seniority among creditors, using collateral may mitigate the free-rider problem and enhance the main bank's screening and monitoring. This incentive effect for the main bank becomes tenuous for personal guarantees, because personal guarantees do not define the seniority among creditors. Thus, our work provides empirical evidence on how collateral affects relationship lenders' incentives, and complements previous studies that focus on the problem of borrower incentives (moral hazard and adverse selection).
The remainder of the paper is organized as follows. Section 1 develops our empirical hypotheses, which are based on previous theoretical models and empirical research. Section 2 describes the data and variables that are used in the paper, and explains our empirical model. Section 3 presents the results of our empirical analysis, and Section 4 concludes.
1. EMPIRICAL HYPOTHESES
1.1 Borrower Riskiness
Much of the empirical literature in this field examines theoretical predictions of asymmetric information models on the relationship between risk and collateral. If the bank cannot discern borrowers' riskiness (hidden information), then collateral may serve as a screening device to distinguish between borrowers and to mitigate the adverse selection problem (Bester 1985). This follows from the observation that a lower-risk borrower has a greater incentive to pledge collateral than a risky borrower, because of his lower probability of failure and loss of collateral. Hence, the lower-risk borrower will choose the contract with collateral.
On the other hand, if the lender can observe the ex ante risk, but there are information asymmetries with regard to actions taken by the borrower after the loan is extended, collateral potentially provides an incentive to mitigate moral hazard. Thus, opposite to models focusing on hidden information, those concentrating on hidden action suggest that it is observably riskier borrowers that will pledge collateral because collateral induces more effort by the borrower (Boot, Thakor, and Udell 1991) or reduces the incentives of strategic default (Bester 1994).
Because our database only contains measures of firms' observed riskiness (namely, credit scores), we couch our first empirical hypothesis as follows:
Hypothesis 1 (H1). The use of collateral is higher among observably higher-risk (low credit score) borrowers if the lender requires collateral in order to mitigate the extent of moral hazard. Alternatively, if borrowers pledge collateral as a signal of their unobserved high credit quality, then there is negative or no relationship between the use of collateral and the borrowers' riskiness.
Consistent with the theory of moral hazard, most existing empirical studies, including Berger and Udell (1990, 1995), have found a positive relationship between collateral and borrowers' ex ante risk. Jimenez, Salas, and Saurina (2006) directly test the adverse selection and moral hazard hypotheses by separating ex ante and ex post measures of borrower riskiness, namely, defaults prior to and after the loan origination. Their results suggest that although observed riskiness increases the likelihood that collateral is used, there is also a negative association between collateral and default after the loan has been granted, which is consistent with the adverse selection argument.
It should be noted that theories of collateral as a solution to moral hazard and/or adverse selection problems assume collateral is external to the firm. (2) Unfortunately, our measure of the incidence of collateral does not distinguish between firm (inside) collateral and personal (outside) collateral. Hence, throughout our analysis, we will assume that collateral is mostly inside but allow for the fact that there may also be some outside collateral. As for personal guarantees, they clearly represent outside collateral.
1.2 Screening and Monitoring by the Lender
Recent research on collateral also discusses how collateral affects lenders' incentives with regard to information production, that is, the screening of borrowers' quality and the monitoring of their performance. These theories of the effect of collateral on lenders' incentives apply to both inside and outside collateral. Manove, Padilla, and Pagano (2001), for instance, argue that, from banks' point of view, collateral can be considered as a substitute for the evaluation of the actual risk of a borrower. Thus, banks that are highly protected by collateral may perform less screening of the projects they finance than is socially optimal.
However, several theoretical studies argue that collateral may complement lenders' screening and monitoring activities. In the presence of other claimants, lenders' incentive to monitor borrowers is reduced due to the informational free-rider problem. In order to enhance lenders' incentive to monitor, loan contracts must be structured in a way that makes lenders' payoff sensitive to borrowers' financial health. Rajan and Winton (1995) argue that collateral may serve as a contractual device to increase lenders' monitoring incentive because collateral is likely to be effective only if its value can be monitored. Moreover, the use of collateral as an incentive will be more extensive when the value of such collateral (as in the case of accounts receivable and inventories, for example) depreciates rapidly if business conditions deteriorate, than when the value of collateral is relatively stable (as in the case of, e.g., real estate). (3) Longhofer and Santos (2000) argue that collateral serves as an incentive for information production by the principal lender in the presence of several creditors because taking collateral is effective in making its loan senior to other creditors' claims. Thus, the bank that provides collateralized loan is able to reap the benefits of screening and monitoring activities. Note that this argument does not straightforwardly apply to personal guarantees because, in general, personal guarantees do not define seniority among several creditors....
|
