نظارت بانکی، بهره وری سود و مدل کسب و کار وام تجاری

We build a bank-specific, fixed-effects regression model to develop proxies for a bank's monitoring effort. Our results show that banks that devote more resources to monitoring (based on these proxies) are more profit efficient and the effect is large. A very important theoretical literature in finance suggests that monitoring is value enhancing; we provide empirical evidence consistent with the theory. This research thus establishes an important link between the large literature on bank monitoring and the equally large literature on profit efficiency. Monitoring is a key technology in the commercial lending business model (e.g. Mester, Nakamura, & Renault, 2007). Thus, these results point to considerable strengths in the dominant business model used in the banking industry.

We consider the theoretical and empirical literature on bank monitoring, how monitoring proxies have been constructed, and the relation between loan quality and bank efficiency. 1.1. Bank monitoring In an extensive literature survey, Boot (2000) defined a loan relationship as being characterized by multiple interactions between lender and borrower that generally take place over time. Monitoring can be thought of as the act of obtaining private information about the borrowing firm through these interactions. Because monitoring generally takes place in the context of a developed bank–borrower relationship, we consider the two concepts together.4 The importance of this topic is illustrated by the attention it has recently received, with no less than three surveys of this literature being published between 2000 and 2004 (Boot, 2000, Elyasiani and Goldberg, 2004 and McNulty, 2002). Banking practitioners think of monitoring as attempted early detection of problem loans to prevent further deterioration and avoid severe loss (see e.g., Rose, 2002, Chapter 17). Lender-borrower interactions also provide banks with the opportunity to sell more financial services to the borrowing firm, thus adding value to the relationship (Elyasiani & Goldberg, 2004). The academic literature has a broader view of monitoring. Indeed theoretical studies have suggested many additional reasons why banks that invest more resources in this area should be more profit efficient. For example, relationships facilitate information reusability both over time and across firms. Chan, Greenbaum, and Thakor (1986, p. 249) pointed out, “costly information relating to one borrower is often useful in assessing others”. This interrelation is especially true if the lender engages in sector specialization so that the firms are in the same or similar lines of business (Boot & Thakor, 2000). Another idea implicit in Chan et al. (1986) is that information durability is increased when overlapping generations are introduced into a model.5 Additionally, relationships can be a way for banks to survive competition from other banks and nonbank financial service firms (Boot & Thakor, 2000). Monitoring can also create a demand for bank loans among borrowers with lower creditworthiness (Diamond, 1991). The future financing implicit in the relationship provides the incentives for borrowers to manage the firm in an appropriate manner (Boot, Greenbaum, & Thakor, 1993). Relationships can give a bank market power in its dealings with the borrower, allowing it to obtain some of the gain from a customer's positive net present value project (Houston and James, 1996, Rajan, 1992 and Sharpe, 1990). Relationships also provide valuable renegotiation options not generally found with publicly placed debt (e.g., Berlin and Mester, 1992, Bester, 1994, Gorton and Kahn, 2000 and Rajan, 1992).6 Empirical studies confirm the notion that bank monitoring creates value. James (1987) considered the announcement period effects of bank loans contrasted with private placements and reported a statistically significant average abnormal return of 1.93 percentage points for bank loans, whereas private placements generated negative returns. A number of other studies have reported similar results (e.g., Billett et al., 1995, Preece and Mullineaux, 1994 and Slovin et al., 1993). Because monitoring takes place over time, Lummer and McConnell (1989) asked whether new bank loans create as much value as existing loans. Banks require time to understand the private information they have acquired and to form estimates of firm value. They find that bank loans create value only when there is a revision to an existing loan. Lenders may have different monitoring abilities (e.g., Fama, 1985). Chemmanur and Fulghieri's (1994) model contains banks with different monitoring capabilities. They found that loan renewals from better monitoring banks provide more favorable information to the market than the others because these banks devote more resources to the monitoring effort. Billett et al. (1995) reasoned that bank loans from lenders with reputations for superior monitoring capability should produce greater valuation effects. They found that, as hypothesized, announcement period returns are greater when banks with better reputations make the loans. Thus, monitoring by these banks conveys more information to the market than monitoring by banks that are either not very good at monitoring or do not commit as many resources to the effort as their peer banks. The value of monitoring also varies by firm characteristics. For example, when firm value is more difficult to estimate, returns from bank loan announcements are greater (DeGennaro, Elayan, & Wansley, 1999). Cole, Goldberg, and White (2004) and others found that small banks focus on different types of customers than large firms and evaluate credit in different ways. Carter, McNulty, and Verbrugge (2004) and Carter and McNulty (2005) suggest that monitoring may contribute positively to small bank financial performance because risk-adjusted loan yields and spreads are greater for small banks. They point out that one explanation for the positive relation between monitoring and performance is the ability of small banks to find economically valuable information about a firm's financial condition by monitoring the firm's demand deposit account. Banks gain inside information into a firm's financial condition by observing the flow of funds into and out of the firm's transaction account. This monitoring process is only possible when the borrower has an exclusive deposit relationship with the lender. Thus, this opportunity to reduce asymmetric information problems is not open to larger banks lending to large firms because these firms have multiple bank relationships (perhaps several dozen or more in the case of very large multinational firms). No one bank could obtain economically valuable information about trends in receipts and expenditures for these firms by monitoring accounts at their bank. Previous work by Kane and Malkiel (1965), Black (1975), Nakamura (1993), Billett et al. (1995) and Petersen and Rajan (1995) also called attention to the role of checking accounts in bank monitoring. Mester et al. (2007) provided the first empirical study of how this information is used in the credit evaluation process. They obtained confidential data from an anonymous Canadian bank on 100 small business loan accounts for the period 1988–1992. They showed that the flow of funds into and out of the firm's transactions accounts are closely related to the bank's credit decisions about these accounts. They are the first to present empirical evidence of the process at work on individual loans. Billett et al. (1995, p. 703) noted: “if observing a loan applicant's demand deposit account provides better credit information or an enhanced ability to monitor bank performance, bank loans may add more value than non-bank loans can”. 1.2. Measuring a bank's monitoring effort Because monitoring cannot be observed directly, empirical researchers have constructed proxy measures of the quantity and quality of a bank's monitoring effort. Coleman, Esho, and Sharpe (2006) and Lee and Sharpe (2009) use ex ante proxies based on the amount of labor input into the monitoring process. The first study relates monitoring to loan rates and terms; the second study relates monitoring to abnormal returns for firms from commercial bank loan announcements. To construct the proxies they assume that banks having higher ratios of salaries to total non-interest expense, after controlling for other factors, are devoting more resources to monitoring. We discuss their monitoring proxies in more detail below and we use a similar methodology and construct similar proxies in this study. (Our study differs from theirs in that we are interested in the effect of the monitoring proxies on profit efficiency rather than on loan rates and terms or abnormal returns from loan announcements.) Other measures of a bank's monitoring ability that have been used by researchers are the bank's credit rating (Billett et al., 1995), loan loss provisions (Johnson, 1997) and bank size (Cook et al., 2003 and Johnson, 1997). As Lee and Sharpe (2009, p. 35) point out, the first two are “ex post realizations of the outcome of the lending process.” In our study, relating an ex post measure of loan quality, such as loan loss provisions, to profit efficiency would not produce an advance in our knowledge of banking. Banks with superior loan quality would clearly be expected to be more profitable than others. The second measure, a bank's credit rating, is only available for the larger institutions, while we focus on the entire population of banks that follow the commercial lending business model, a total of 2,295 banks. Researchers use bank size as a monitoring proxy based on the assumption that banks that have screened their borrowers thoroughly will grow to a larger size than other banks. In addition, relationship lending may have high fixed costs and thus be subject to economies of scale. However, in recent years there has been considerable discussion of the point that larger banks may have an incentive to take on poorer quality loans and acquire other banks until they grow to a size at which they are considered too big to fail. In addition, a number of very large banks had to be merged into other banks because of loan quality problems. Banks that survived, such as Citigroup, the largest bank holding company in the US, also had serious asset quality problems in recent years. Hence we consider bank size to be an inappropriate monitoring proxy. Thus, we follow the approach of Coleman et al. (2006) and Lee and Sharpe (2009) by using an ex ante measure of the resources a bank devotes to loan screening and monitoring. This monitoring proxy can be constructed for all banks, not just those that are publicly traded, which is important for our study. 1.3. Loan quality and cost efficiency Berger and DeYoung (1997) explore the intersection of the problem loan literature and the cost efficiency literature. They note that, at first glance, there would appear to be little or no relationship since operations and lending are conducted in different areas of the bank by different personnel. However, the quality of senior management provides one link because banks that are poorly managed may be both cost inefficient and have higher levels of problem loans than other banks. In addition, the costs of administering problem loans may add to inefficiency problems. Using Granger causality tests, they find empirical support for both hypotheses. Their study illustrates the importance of further research on the relation between monitoring and bank efficiency since problem loans are the major cause of bank failures.

Bank monitoring occupies a very important place in the literature in finance (Boyd and Prescott, 1986, Boyd and Runkle, 1993, Diamond, 1984, Petersen and Rajan, 1994, Petersen and Rajan, 1995, Rajan, 1992 and Ramakrishnan and Thakor, 1984). Mester et al. (2007) provide an important empirical application. An implication of this literature is that banks that devote more resources to monitoring should have superior financial performance. We provide empirical evidence consistent with the theory. (We use the term monitoring to include both the initial underwriting decision and the subsequent screening of the borrowing firm.) We develop proxy variables based on labor input into the monitoring process to estimate which banks spend more resources in monitoring their loans than other banks. We estimate fixed-effects salary regressions for 2,295 banks classified as commercial lenders for the period from 1999 to 2005 and use the bank-specific fixed-effect coefficients as monitoring proxies. We focus on banks that are characterized by the FDIC as commercial lenders. This is the dominant business model in the banking industry. There are no mega banks in our sample. The sign of the monitoring proxies is positive and significant in the profit efficiency regression equation, suggesting that monitoring improves profit efficiency. Further, monitoring is one of the most important variables affecting profit efficiency. One reason for average levels of profit efficiency of only 76% (a finding in most profit efficiency studies, including this one) may be inadequate attention to monitoring. Our out-of-sample tests show that the monitoring proxies predict problem loan ratios 4 years in advance even during the period of the financial crisis. The weak incentives in the “originate-to-distribute” business model used in mortgage lending led to poor lending decisions which contributed significantly to the recent financial crisis. What is less apparent in popular discussion of the crisis is the point that the commercial lending business model is the dominant business model used in the industry. This model does not have comparable weaknesses because the loans are generally quite large and are kept on a bank's books, so there are major incentives to invest in monitoring. These are often relationship loans. Because of asymmetric information, relationship loans have more value to the originating bank than to any other bank, and are thus much less likely to be sold. Our results are consistent with the notion that the commercial lending business model creates value when the bank devotes adequate resources to the process. Our results should help banking practitioners and regulators consider the strengths of various business models as the industry emerges from the crisis.