وثیقه های کسب و کار و تعهدات شخصی در وام دهی شرکتهای کوچک و متوسط

Using a database of SME credit approvals from a large Belgian bank, this paper extends the empirical evidence on the determinants of collateral by examining the determinants of business collateral simultaneously with the determinants of personal collateral/commitments. Our results suggest that firm and relationship characteristics seem to be more important determinants of collateral/commitment protection than loan and lender characteristics. Family firms are more likely to offer a higher degree of collateral/commitment protection while introducing competition between banks decreases this likelihood. The collateral requirement decreases in the length of the bank-borrower relationship. Furthermore, trade credit seems to have a signalling effect. The ‘lazy banks hypothesis’ was not supported. Our results suggest that beside risk arguments, also commercial arguments help explain the pledging of collateral. Using a continuation-ratio logit model, we discover several differences in the determinants of the collateralisation decision and the determinants of the type of collateral/commitments.

The use of collateral is a common feature of credit contracts between firms and lenders. The question why the use of collateral is so widespread has been the subject of several theoretical contributions. In general, the contractual relationship between borrowers and lenders may be hampered by the presence of asymmetric information, adverse selection and moral hazard, usually leading to credit rationing. As such, the risk of lending may be reduced by collateral. Collateral may play the role of a signalling device for borrower quality (e.g. Chan and Kanatas, 1985, Bester, 1985, Besanko and Thakor, 1987a, Besanko and Thakor, 1987b and Boot et al., 1991), may lower the agency costs of debt by preventing the problem of asset substitution (Jensen and Meckling, 1976) and mitigate Myers’ (1977) underinvestment problem (Stulz and Johnson, 1985). In general, when moral hazard risk shows up in the lending relationship, collateral may play a disciplinary role in the behaviour of the borrower. Consequently, stronger creditor protection from collateral would lead to better credit terms or even the approval of credit that otherwise would not be granted. According to Mann, 1997a and Mann, 1997b, secured credit limits the firms’ ability to obtain future loans from other lenders or reduces the risk of excessive future borrowing. Recently, Manove et al. (2001) criticized the unrestricted reliance on collateral and argued that this might have a negative impact on credit-market efficiency. They argue that banks are in a good position to evaluate the future prospects of new investment projects. Collateral will weaken the bank’s incentives to do so. Especially for small firms, banks would do little screening and rely excessively on collateral. From the point of view of banks, collateral and screening can be considered as substitutes. Despite the considerable amount of effort that has been devoted in the theoretical literature to the role of collateral in business lending, only few theoretical studies (e.g. Chan and Kanatas, 1985 and Mann, 1997b) make the explicit distinction between personal and business collateral. Chan and Kanatas (1985) postulate that business and personal collateral are very similar. Nevertheless, Mann (1997b) argues that personal collateral is more effective in limiting the borrower’s risk preference incentives by enhancing the likelihood that the principal will feel any losses personally. The empirical literature (Leeth and Scott, 1989, Ang et al., 1995, Avery et al., 1998, Harhoff and Körting, 1998, Degryse and Van Cayseele, 2000 and Hanley, 2002) concerning the determinants of collateral is rather scant, partly due to data limitations. While it is well documented that small and medium-sized firms rely primarily on financial intermediaries as lenders, especially commercial banks (Cole et al., 1996), only partial clues exist as to the role of personal wealth or business wealth in the contractual details of lending arrangements. To date, as far as we are aware off, only two empirical studies (Ang et al., 1995 and Avery et al., 1998) examine the topic. Both studies found that personal commitments are an important component of SME lending. However, no efforts have been made to refine such results by distinguishing the factors related to both personal commitments1 and business collateral usage. Moreover, no European empirical evidence is available to date. This paper contributes to the literature in two ways. First, using a recent database of a large Belgian bank, new empirical evidence on the determinants of both types of collateral protection is presented. Prior research only concentrated on the determinants of business collateral or the determinants of personal commitments separately. We extend this literature by examining the determinants of business collateral simultaneously with the determinants of personal collateral/commitments. Therefore, we differentiate between three degrees of collateral/commitment protection demanded by the bank. These three categories are considered as ordinal or ordered categories ranging from ‘no protection’ to ‘highest protection’: (1) unsecured debt, (2) secured debt with only business collateral and (3) secured debt with personal commitments (with or without business collateral). The theoretical rationale for considering these three categories as ordinal can be found in the information asymmetry of the borrower-lender relationship. The ‘implicit value’ of personal commitments as a disciplining device that limits the borrower’s risk preference incentives is higher than for business collateral (Mann, 1997b), especially under corporate organisational forms such as is the case for the firms in our database. The lender receives explicit claims on personal assets and/or future wealth of the borrower (Ang et al., 1995), which he cannot rely on in the case of business collateral. Moreover, the likelihood that the borrower will feel any losses personally is higher when granting personal collateral. Based on these three categories of collateral/commitment protection, we estimated two kinds of econometric models. In order to test the determinants of collateral/commitment protection as a continuum, we estimate an ordered probit model. Furthermore, we investigated the differences between the determinants of business collateral and personal commitments by a continuation-ratio logit model with two binary choice models: (1) the choice with or without collateral/commitments and (2) the choice between business collateral or personal commitments, given that the bank has already decided to ask for some kind of collateral. Second, the database contains specific data that allowed us to examine determinants that were not included in previous empirical studies such as the signalling role of trade credit (Biais and Gollier, 1997), the family factor and the ‘lazy’ banks hypothesis (Manove et al., 2001). Our results suggest that firm and relationship characteristics seem to be more important determinants than loan and lender characteristics. Furthermore, we find that beside risk arguments, also commercial arguments help explain the pledging of collateral. Using a continuation-ratio logit model, we discover several differences in determinants of the collateralisation decision and the determinants of the type of collateral/commitments. The organization of the paper is as follows. Section 2 discusses the determinants of business collateral and personal commitments and develops the hypotheses. Section 3 explains the data set, the variables and the method. The empirical results are analysed in Section 4. Section 5 concludes the paper.

The determinants of collateral have intrigued scholars for decades. Although in reality many SMEs have to pledge personal collateral/commitments, there is barely empirical evidence available about the determinants of these personal collateral/commitments. As far as our knowledge, we are the first that empirically investigate the determinants of personal collateral/commitments simultaneously with the determinants of business collateral for SMEs in a bank-oriented economy. Moreover, we provide empirical evidence about the differences in determinants of the collateralisation choice and the type of collateral decision. Our results suggest that firm and relationship characteristics seem to be more important determinants of collateral/commitment protection than loan and lender characteristics. The results indicate that family firms are more likely to offer a higher degree of collateral/commitment protection while introducing competition between banks decreases this likelihood. Trade credit is negatively related to the degree of collateral/commitment protection, indicative of a possible signalling effect of trade credit. In line with previous studies, we find indications that collateral requirements decrease in the length of the bank-borrower relationship. Furthermore, when a firm introduces more competition between banks for a credit request, it decreases the likelihood of having to offer collateral or in case of collateral pledging, to offer personal commitments. Our results also suggest that beside risk arguments, also commercial arguments determine the pledging of collateral. If a financial institution operates as ‘main bank’ for a firm, this firm is also more likely to offer collateral/commitments. A theoretical explanation for this behaviour is provided by Mann, 1997a and Mann, 1997b. Banks often take collateral because secured credit limits the firms ability to obtain future loans from other lenders and reduces the risk of excessive future borrowing. We add to this a more pragmatic explanation. Formalized interviews with the account managers revealed that banks in Belgium usually act in this way because of commercial reasons. Collateral creates a barrier-to-entry for other competing banks if these banks try to capture the client-firm from the main bank. Lines of credit are more likely to be granted with a higher degree of collateral/commitment protection than other loans. The provocative ‘lazy banks’ proposition of Manove et al. (2001) is not supported by our results: the pledging of collateral does not seem to be a substitute for the screening efforts of the bank as measured by the time needed to approve the credit request. Using a continuation-ratio logit model, we discovered several differences in determinants of the collateralisation decision and determinants of the type of collateral/commitments. The signalling effect of trade credit and the amount of the loan are only significant in the type of collateral choice model while the main banker, lines of credit and the maturity of the loan show only significant effects in the collateralisation decision. Generally, our models seem to explain better the collateralisation decision than the type of collateral decision. One may question if a banking policy of taking more collateral than necessary from a point of view of risk such as suggested by our results, may lead to the phenomenon of ‘discouraged borrowers’ which has been discussed recently in the academic literature (Levenson and Willard, 2000 and Kon and Storey, 2003). Previous research (Binks and Ennew, 1995) indicates that the lenders’ demand for personal commitments augments the perceived credit constraints in SMEs. Future research should scrutinize the correlation between banking policy and the phenomenon of discouraged borrowers and demand side-credit rationing. A possible positive relation would shed a different light on the credit rationing debate. Finally, several issues remain unexplored. We measure collateral protection as an ordinal variable and with dummy variables. Although the majority of current and past empirical studies measure the dependent variable in a similar way, an important dimension of the problem remains unexplored being the amount of collateral/commitments a bank requires. Future research should implement the amount of business collateral and personal commitments as a dependent variable in order to test the robustness of current findings.