A multivariate default model with spread and event risk

Market quotes of credit derivatives feature two fundamental properties. On the one hand, daily movements of observable credit spreads contain information about time- varying default probabilities. The standard approach to reproduce these movements is the intensity based ansatz. On the other hand, intensity models, producing only limited dependence, typically fail to adequately capture default clusters, see e.g. Das et al. (2007). Default clusters are better described by models supporting cataclysmic events, such as a number of (static) copula models. This paper presents a new model for a portfolio of defaultable assets that takes into account both features and still allows for the computation of numerous dependence properties and the application of fast, non-simulation based pricing routines. The idea is to combine two existing instances of the so-called conditionally independent and identicallly distributed (CIID) class, namely the intensity based ansatz in the spirit of Duffie and Gârleanu (2001) and the Lévy subordinator concept introduced in Mai and Scherer (2009), in a way that preserves the distinct advantages and excludes most of the shortcomings. The model is calibrated to iTraxx Europe index CDS and CDOs. It is shown how to achieve this numerically challenging task within fractions of seconds by exploiting the special structure of CDOs. The fitting results are very promising.      «

Market quotes of credit derivatives feature two fundamental properties. On the one hand, daily movements of observable credit spreads contain information about time- varying default probabilities. The standard approach to reproduce these movements is the intensity based ansatz. On the other hand, intensity models, producing only limited dependence, typically fail to adequately capture default clusters, see e.g. Das et al. (2007). Default clusters are better described by models supporting catacly...     »