مشتقات اعتباری در بانکداری: ابزارهای مفید برای مدیریت ریسک؟

We model the effects on banks of the introduction of a market for credit derivatives; in particular, credit-default swaps. A bank can use such swaps to temporarily transfer credit risks of their loans to others, reducing the likelihood that defaulting loans trigger the bank's financial distress. Because credit derivatives are more flexible at transferring risks than are other, more established tools, such as loan sales without recourse, these instruments make it easier for banks to circumvent the “lemons” problem caused by banks’ superior information about the credit quality of their loans. However, we find that the introduction of a credit-derivatives market is not necessarily desirable because it can cause other markets for loan risk-sharing to break down.

Credit derivatives are over-the-counter financial contracts that have payoffs contingent on changes in the credit quality of a specified firm or firms; the specified firm is typically not a party to the contract. The market for credit derivatives was developed during the early 1990s by large money-center commercial banks and investment banks. The market is small but is growing quickly. To date, credit derivatives are used to trade risks that are already traded in existing markets. The underlying instruments on which credit derivatives are written are typically corporate bonds, Brady bonds, large leveraged bank loans, or pools of homogeneous small loans such as credit card receivables. Thus, for now, credit derivatives can be thought of as instruments that repackage traded risks into more convenient forms. The question we address here is whether, from a theoretical perspective, credit derivatives can also be used to trade heretofore nontraded credit risks. In particular, we focus on small and medium-sized bank loans for which asymmetric information concerns outweigh reputation concerns of the lending bank. If credit derivatives could penetrate this market of untraded risks, the effects on banks likely would be large. (Here, we view banks as end-users of credit derivatives, and ignore the potential profits to be made by money-center banks as dealers in the credit-derivatives market.) Bank loan portfolios are typically concentrated within business sectors and geographic regions. An important reason for this concentration is an asymmetric information problem: Banks know more about the value of their loans than do outsiders. Banks with high-quality loans will tend to refrain from selling pieces of their portfolio if outsiders cannot distinguish such loans from low-quality loans. Reputation effects in the loan-sales market can help mitigate problems caused by asymmetric information, but the inherent limitations of such effects are evident in the continued concentration of banks’ portfolios. We argue that credit derivatives’ flexibility in repackaging risks can, in some circumstances, allow banks to trade previously untradeable credit risks. The analysis follows an observation by Duffee (1996) that, depending on the nature of a bank's private information about a loan, the uncertainty in a loan's payoff potentially can be decomposed into a component (or components) for which the bank's informational advantage is relatively small and a component (or components) for which the bank's informational advantage is relatively large. If so, the bank can use a credit-derivative contract to transfer the former risks to outsiders, while retaining the latter risks at the bank. For example, we argue that the bank's informational advantage is unlikely to be constant over the life of the loan. Thus the introduction of credit derivatives that temporarily transfer loan risk to outsiders could promote better risk sharing, thereby reducing the expected deadweight costs associated with bank insolvency. This logic suggests that the use of credit derivatives to fine-tune credit risk management can benefit banks. We formalize these benefits in the context of a simple model. However, we also show that the introduction of a credit-derivatives market can harm banks even as they use it to transfer credit risks to others. Banks can be worse off if the introduction of the credit-derivatives market leads to the breakdown of other risk-transferring mechanisms such as loan sales without recourse that pool the risks of banks that make high-quality and low-quality loans. With the introduction of credit derivatives, banks with high-quality loans may choose to shed part of their risk with credit derivatives and refrain from selling any other part of their risk, destroying the pooling equilibrium in the loan-sale market. The net effect can be an increase in the expected deadweight costs associated with bank insolvency. This seemingly paradoxical conclusion is a standard result in the economics of insurance, and an example of Hart's (1975) seminal point that when markets are incomplete, the opening of a new market can make everyone worse off. The health-insurance market provides a useful analogy. Imagine insurance companies allowed individuals to purchase health insurance that excluded coverage for a particular genetically linked disease. The existence of such lower-cost insurance policies would reduce the impact of both adverse selection and moral hazard; low-risk individuals could purchase more insurance and high-risk individuals would take better care of themselves. But society as a whole might be worse off because the costs of exogenously having a bad gene are not shared as widely. We find that the value of the credit-derivatives market critically depends on whether the asymmetric information associated with bank loans is primarily an adverse-selection problem or a moral-hazard problem. For example, if the quality of a bank's loan portfolio is entirely exogenous (the bank does the best job it can of lending money, but sometimes its pool of potential borrowers is weak), a breakdown in the loan-sales market caused by the introduction of credit derivatives would be, on net, socially costly. At the other extreme, if the portfolio's quality is entirely endogenous (potential borrowers are homogeneous, and the bank can spend money to monitor its loans aggressively), the loss in risk-sharing owing to a breakdown in the loan-sales market would be offset by a reduction in moral-hazard problems, and hence the introduction of a credit-derivatives market would be beneficial. To our knowledge, this paper is the first in the academic literature to consider rigorously the implications of credit derivatives for banks’ risk-sharing. A related literature examines the ability of banks to sell loans about which they have private information. Carlstrom and Samolyk (1995) adopt the standard assumption that there is a deadweight cost to bank insolvency. The cost of bank insolvency gives the bank an incentive to sell some of its loan opportunities instead of directly funding the loans. The quality of loans a bank can make is unobservable by others, which typically gives rise to adverse selection. However, in their model, the deadweight cost of bank insolvency is infinite—thus banks face no real tradeoff between holding their loans or selling them. Therefore Carlstrom and Samolyk circumvent the standard lemons problem in which banks with high-quality loans refrain from selling them at low prices. Gorton and Pennacchi (1995) also model a bank's choice between holding loans and selling them, focusing on moral hazard. If a bank holds a loan, it has a greater incentive to monitor the loan (and thus increase its probability of repayment) than if it sells it. They conclude that if a bank can implicitly commit to holding a certain fraction of a loan (or to provide limited recourse), the moral hazard associated with loan sales is reduced. We note that Gorton and Pennacchi's point is broadly applicable to any mechanism that transfers loan risk outside of the bank, including credit derivatives. The next section describes some of the institutional features of the credit-derivatives market. Section 3 presents a model in which only adverse selection, not moral hazard, limits the ability of banks to sell their loans. Section 4 uses the model to evaluate the value to banks of the credit-derivatives market. Section 5 extends the model to consider moral hazard. This section also addresses some effects that credit derivatives can have on capital allocation. The final section concludes.

We construct a model of a bank that has an opportunity to make loans. The risk of loan default can expose the bank to its own financial distress. The bank can sell any fraction of the loan in order to reduce its expected costs of distress, but because the bank has superior information about loan quality, the loan-sale market is affected by an asymmetric-information problem. We build in a role for credit derivatives in the model by assuming that the magnitude of the asymmetric information varies during the life of the loan. A credit-derivative contract that transfers the loan's risk when the lemons problem is smallest can be used by the bank to reduce its risk of financial distress. If the asymmetric-information problem is sufficiently severe, the loan-sale market will be of only limited use to banks, and thus the opportunity to use credit derivatives will be valuable to the bank. However, when we consider the effects that a credit-derivatives market has on other markets for sharing risks, the introduction of a credit-derivatives market does not necessarily benefit the bank. If, prior to this introduction, the asymmetric-information problem was not severe enough to limit the use of the loan-sale market, the addition of a market in credit derivatives can be harmful. The new market can alter investors’ expectations of the quality of loans sold in the loan-sale market and thereby dramatically change the nature of equilibrium in this market. Thus, although the credit-derivatives market will be useful to the bank, its presence makes the loan-sale market much less useful. We find that if the asymmetric-information problem is one of adverse selection, the net effect is to leave the bank worse off, while if the problem is one of moral hazard, the bank is better off. Therefore the increased risk-sharing flexibility created by credit derivatives is not enough to guarantee that such instruments are beneficial. Note that we are not, in any way, claiming that banks should refrain from entering into credit-derivative contracts. Indeed, we find that credit derivatives may improve capital allocation by reducing investment in poor-quality projects. Instead, the conclusion that should be drawn from our arguments is that theory alone cannot determine whether a market for credit derivatives will help banks better manage their loan credit risks. This issue is ultimately an empirical one. For example, the potential value of this market depends, in part, on the extent to which the loan-sale market is currently used to share the risks of loans about which originating banks have private information. This empirically unresolved issue is examined in Berger and Udell (1993) and Gorton and Pennacchi (1995). If credit derivatives will simply replace loan sales as risk-sharing tools, the consequences for banks are ambiguous.