Measuring Credit Risk And Pricing Credit Derivatives

With the credit derivatives market has grown rapidly both in volume and in the type of instruments it offers, academicians and practitioners focus their effort on the management and pricing of credit risk. Default risk at the level of an individual security has been extensively modeled using both structural and reduced form models. This paper examines default risk at the portfolio level. Default dependencies among issuers in a large portfolio play an important role in the quantification of a portfolio's credit risk exposure. Why are corporate defaults dependent each other? Several explanations have been explored. First, firms may be exposed to common or correlated risk factors whose co-movements cause correlated changes in conditional default probabilities. Second, the event of default by one firm may be"contagious,"in that one such event may directly induce other corporate failures, as with the collapse of Penn Central Railway in 1970. Third, learning from default may generate default correlation. For example, to the extent that the defaults of Enron and WorldCom revealed accounting irregularities that could be present in other firms, they may have had a direct impact on the conditional default probabilities of other firms.The traditional methods and their deficienciesThe standard industrial methodologies like RiskMetrics and CreditMetrics model the dependence structure in the derivatives or credit portfolio by assuming multivariate normality of the underlying risk factors. The advantages of using a normal distribution are its simplicity, its analytical tractability, and the easy estimation of its only parameter, the correlation matrix. However, it has been well recognized that many financial assets exhibit a number of features which contradict the normality assumption—namely asymmetry, skewness and heavy tails. In particular, empirical studies like Junker and May (2002) and Malevergne and Sornette (2004) indicate that especially during highly volatile and bear markets the probability for joint extreme...