Capital Market Tests of Risk Exposure of Loan Sales Activities of Large U.S. Commercial Banks

Beginning of article

Introduction

This paper deals with the risk exposure of loan sales (LNPART) by large commercial banks in the United States. Commercial loan sales involve the sale of loans originated by banks. Loan sales are structured so that the selling bank continues to service the loan, although the purchaser of the loan is entitled to principal and interest payments. There has been a dramatic increase in loan sales activities by banks in recent years. Call report information indicates that loans sold increased from $50 billion in 1984 to $280 billion in 1988, a 460 percent increase over the five year period.

Since the early 1980s commercial banks, spurred by financial market changes, improved technology, and tighter regulatory capital standards, have increased their sales of loans greatly. More recently advances in legal contracting and new data processing technology have made it possible for a large group of loans to be pooled and sold as a tradeable security, a procedure called securitization. Regulators are concerned, however, about the risk that banks retain by providing loan purchasers at least partial recourse or implicit guarantee if problems develop with a loan. Current capital regulation discourages recourse for credit risk by including such loans' full value in banks' capital requirements (Haubrich, 1989).

Previous research has investigated the determinants and riskiness of loan sales (Pavel, 1988; Pavel and Phillis, 1987; James, 1988). These studies explore the riskiness of loan sales either employing theoretical models or empirically regressing implied asset risk from equity and CD rates against on-balance sheet and off-balance sheet loan sales activities. This paper reports five capital markets tests: systematic risk, equity risk, subordinated debt default risk premia, implied asset risk calculated from the Ronn-Verma (1986) option pricing model, and implied asset risk from the Gorton-Santomero (1990) subordinated debt option pricing model. The purpose of these tests is to examine the riskiness of banks with off-balance sheet loan sales activities.

The first test regresses the systematic risk of banks against on-balance sheet risk measures and bank loan sales activities. The second test purports to find an association between equity risk and on-balance sheet risk measures and loan sales. The purpose of these two tests is to determine whether equity participants price loan sales by banks as risk-increasing or risk-reducing activities.

The third test relates the subordinated debt risk premia to LNPARTs and other accounting measures of bank risk. The underlying premise of this test is that subordinated debt holders are more exposed to the risk of bank failure than are CD holders. The fourth test regresses implied asset risk over LNPARTs along with other bank accounting risk measures, where implied asset risks are calculated from a risk premia option pricing methodology (Gorton and Santomero, 1990). This test contends that subordinated debts are contingent claims whose costs are not linear or monotonic functions of bank risk. Moreover, the underlying risk is contingent upon the regulatory closure rule. Without recognizing these complications, linear risk premia regressions may be inadequate in addressing the market discipline of LNPARTs. The motivation of this test is that the implied asset risk from the risk premia option pricing model is better than risk premia in proxying total risk because it considers both the nonlinear nature of the contingent claims model and the impact of closure rules.

The fifth test associates implied asset risk with LNPARTs in addition to on-balance measures of bank risk, where implied asset risks are calculated from market values of bank equity and fixed deposit insurance premia (Ronn-Verma, 1986). The equity risk used in previous studies ignores deposit insurance and the protection of too-big-to-fail regulation. This test measures risk in a way that incorporates fixed deposit insurance premia and regulatory audit rules. The appeal of this test is that asset risk, instead of equity risk, is a better measure of risk because both bank debt holders and equity holders benefit from the current flat deposit insurance policy.

The standard approach of market studies of bank activities is to regress the equity or debt costs and/or market risk measures over on-balance sheet and income statement measures of risk and return. The basic issue in these studies is whether market measures of bank risk are related significantly to book measures of bank risk. If significant relationships are found, then the market appears to control these banking activities. This paper contributes to the existing market discipline studies of bank off-balance sheet loan sales activities by reporting five capital market tests.

The Association Between Market Measures of Risk and Loan Sales

Commercial loan sales involve the sale of loans originated by banks to third parties. Most commercial loans today are sold either in the form of loan participations or loan strips. Under a loan participation, a bank will sell its loan for a period equal to its time to maturity and the buyer of a participation has no recourse to the selling bank if the loan defaults. Loan strips are short-dated pieces of a longer-term loan in which the selling bank retains the risk of borrower default. The selling bank usually puts up funds to support the loan until it matures. The principal buyers of bank-originated loans are foreign and domestic banks, insurance companies, pension plans, savings banks, and mutual funds.

The loan sales contracts are sold in participation forms so that the selling bank maintains a creditor-debtor relationship with the borrower. The selling bank retains servicing rights on any loans sold, enforces loan covenants, and monitors the financial condition of the borrower. The selling bank receives fees in the form of a spread for performing these services. The spread represents the difference between the rate paid by the borrower to the bank and the return promised to the purchaser of the loan.

The payments that a purchaser receives from a loan sale depend on the recourse provisions attached to the loan. In a loan sale without recourse, the purchaser receives contractual payments and, in loan default, receives whatever cash flows the loan generates. In a loan sale with recourse, the purchaser receives a put option that allows the buyer to sell a troubled loan to the selling bank. Loans are rarely sold with full recourse, however, because federal regulations require a bank selling loans with recourse to hold reserves against these loans and to count them as assets when calculating capital requirements.

The theoretical literature suggests different motivations and reasons for banks to originate and sell loans. The ability of banks to originate loans and then sell them appears to contradict the theoretical assertion that bank …