The purpose of this study is to consider the theoretical basis of commercial loan pricing. Is commercial loan pricing most representative of pricing to reflect risk in the Markowitz sense or do banks ration their loanable funds based on credit risk or expected long-term customer value? Alternatively, does each theory contribute to the explanation of loan pricing?
Some of the pricing theories noted in this study have been tested at the aggregate banking level, however, few studies have been performed at the loan level. Moreover, the author is not aware of any study that tests which theory noted here best describes actual pricing practices for bank loans. In fact, DeVany (1984) and Goldfeld (1984) have noted that models of bank behavior have undergone little direct testing. Goldfeld acknowledges that the sparse empirical work in banking exists because much of the theoretical analysis is at the level of the individual bank where appropriate data are not available. This study overcomes that problem by using the loan portfolio of one of the top 50 bank holding companies in the nation as a case study.
Portfolio theory, credit rationing, and customer relationships provide the basis for this investigation of how banks price commercial loans. Portfolio theory indicates that the risk of a particular loan as well as its contribution toward the riskiness of the entire loan portfolio provides the most information about loan pricing. Credit rationing, however, indicates that the contract interest rate an applicant is willing to accept acts as a signal of loan quality and predicts the bank's expected return on the loan. Finally, theories about customer relationships indicate that customer traits such as variability of deposits and length of the relationship play a role in the way banks price loans.
The data used in this study are at the loan level and were obtained from one of the top 50 bank holding companies in the nation. Loan pricing procedures are examined by performing a series of cross sectional generalized least square regressions where the expected return on the loan is the dependent variable in each regression. The non-nested J-test and Cox-test help determine whether any of the model specifications tested in this study provide significantly greater explanatory power in commercial loan pricing than the competing model specifications.
The empirical findings of this study should be considered exploratory in nature because of its reliance on data from one bank. Moreover, these results assume that each of the models have been properly specified. With these caveats in mind, the results are consistent with credit rationing and customer relationship theories (Hodgman and Kane and Malkiel). Moreover, the non-nested Cox-test indicates that the credit rationing specification used in this study provides more explanatory power with regard to loan pricing than the customer relationship specification.
The regression of the portfolio theory specification provided statistically significant results, but with coefficients of the wrong sign. Contrary to theory, the results suggest that the expected return on loans increases as the variance decreases. In addition, the regression results do not provide strong support that loans are priced relative to the risk they contribute to the total portfolio.
In a matter related to loan pricing, this study also found that collateralized loans are associated with a smaller expected return than non collateralized loans. This finding is consistent with Boot, Thakor, and Udell (1991) who suggest that firms use collateral to obtain more favorable loan terms.
The conclusions and implications of this study revolve around the illiquid nature of commercial loans which creates an inefficient market characterized by asymmetric information.
In light of the scarcity of information related to potential commercial loans, it is not surprising that customer relationship theories provide some explanation of current pricing practices. Certain aspects of a customer relationship, such as deposits and length of the relationship can provide banks with valuable information about the riskiness of loans. Moreover, relationships that cover several bank services may enable a bank to supplement thin loan margins.
Finally, the support, albeit weak, of credit rationing can also be explained with asymmetric information. Because of adverse selection and moral hazard, there is a point at which further increases in the contract interest rate on a loan will lead to declines in the expected return to the bank. Beyond this point, the profit maximizing bank should ration rather than loan its funds.
|01 January 1993
|VCU Scholars Compass
|Virginia Commonwealth University
|Theses and Dissertations
|© The Author
Page generated in 0.0027 seconds