رتبه های مهاجرت و چرخه کسب و کار، با استفاده از تست استرس پرتفولیو اعتبار

The turmoil in the capital markets in 1997 and 1998 has highlighted the need for systematic stress testing of banks' portfolios, including both their trading and lending books. We propose that underlying macroeconomic volatility is a key part of a useful conceptual framework for stress testing credit portfolios, and that credit migration matrices provide the specific linkages between underlying macroeconomic conditions and asset quality. Credit migration matrices, which characterize the expected changes in credit quality of obligors, are cardinal inputs to many applications, including portfolio risk assessment, modeling the term structure of credit risk premia, and pricing of credit derivatives. They are also an integral part of many of the credit portfolio models used by financial institutions. By separating the economy into two states or regimes, expansion and contraction, and conditioning the migration matrix on these states, we show that the loss distribution of credit portfolios can differ greatly, as can the concomitant level of economic capital to be assigned.

The evolution of modern risk management can be traced back to Markowitz and his portfolio theory for investments. His fundamental concept of diversification, of considering the joint distribution of portfolio returns, has gradually migrated to risk management. Market risk measurement techniques were the first to mature, mainly due to the richness of available data. Risk managers now routinely measure portfolio change-in-value distributions and compute statistics such as value-at-risk (VaR), which are used to determine trading limits and assess risk capital. Moreover, the regulatory community has broadly accepted such a model-based approach to assessing market risk capital. On the credit risk side, however, even with the new BIS accords of June 1999 (BIS publication no. 50, 1999), formal credit portfolio models (CPMs) are not permitted for use in the determination of bank credit risk capital. Nevertheless CPMs are becoming more widespread in their use among financial institutions for economic capital attribution, and as the use of new risk transfer instruments such as credit derivatives increases, so by necessity will the use of CPMs.3 Recent turmoil in the capital markets has highlighted the need for systematic stress testing of banks' portfolios, including both their trading and lending books. This is clearly easier said than done. Although we have a wealth of data at our disposal in market risk, even there it is not obvious how best to implement stress testing. A short decision horizon, one on the order of hours or days, forces the thinking towards specific scenarios or the “tweaking” of volatilities and correlations between dominant risk factors. Particularly the latter is important when designing a strategy for stress testing, as the LTCM debacle so poignantly demonstrated (Jorion, 1999). It appears, for example, that correlations increase during times of high volatility (Andersen et al., 2000). Much less, however, is known about stress testing credit portfolios, and both practitioners and regulators are clamoring for guidance. We propose that underlying macroeconomic activity should be a central part of a useful conceptual framework for credit portfolio stress testing, and that credit migration matrices provide the specific linkage between underlying macroeconomic conditions and asset quality. Credit migration matrices, which characterize the expected changes in credit quality of obligors, are cardinal inputs to many applications, including portfolio risk assessment, modeling the term structure of credit risk premia, and pricing credit derivatives.4 They are also an integral part of many of the CPMs used by financial institutions. By separating the economy into two states or regimes, expansion and contraction, and conditioning the migration matrix on these states, we show that the loss distribution of credit portfolios can differ greatly, as can the concomitant level of economic capital to be assigned. We believe, therefore, that our analysis provides a useful framework for stress testing a credit portfolio using any of the CPMs currently available. The paper proceeds as follows. In Section 2 we review the ways in which asset values are tied to credit migration. In Section 3 we describe the ratings data on which our subsequent migration analysis depends and proceed in Section 4 to discuss estimation issues and the properties of the migration matrices. In Section 5 we integrate business cycle considerations into a migration analysis, and in Section 6 we apply our methods to stress testing a credit portfolio using CreditMetricsTM. We conclude and offer suggestions for future research in Section 7.

Utilizing an extensive database of S&P issuer ratings, we have presented a systematic study of rating migration behavior and its linkages to the macroeconomic conditions and asset quality. Our analysis suggests that first-order Markovian ratings dynamics, while not strictly correct, provide a reasonable practical approximation, so long as we allow for different transition matrices in expansions and contractions. We are not the first to note that ratings transition probabilities may vary with the business cycle, and surely we are not the last. In an interesting contribution done independently of this paper, for example, Nickell et al. (2000) use Moody's data from 1970 to 1997 to examine the dependence of ratings transition probabilities on industry, country and stage of the business cycle using an ordered probit approach, and they find that the business cycle dimension is the most important in explaining variation of these transition probabilities. Our work complements, enriches and extends theirs, pointing in particular to the potential usefulness of regime switching models in the context of credit portfolio stress testing. As for future research, at least two promising directions are evident. First, for estimating actual credit risk as opposed to stress testing, one may weight the simulations obtained under assumed takeoffs from expansion and from contraction by p and (1−p), where p is the probability that the economy is currently in expansion, obtained for example using the methods of Hamilton (1989). Second, it is of obvious interest to extend out methods to countries other than the US, whose typically less well developed business cycle chronologies present a challenge for our approach.