None

Applying Portfolio Credit Risk Models to Retail Portfolios

Nisso Bucay and Dan Rosen

We present a simulation-based model to estimate the credit loss distribution of retail loan portfolios and apply the model to a sample credit card portfolio of a North American financial institution. Within the portfolio model, we test three default models that describe the joint behavior of default events. The first model is purely descriptive in nature while the other two models are causal models of portfolio credit risk, where the influence of the economic cycle is captured through the correlations of default rates to various macroeconomic factors. The results obtained using all three default models are very similar when they are calibrated to the same historical data. In addition to measuring expected and unexpected losses, we demonstrate how the model also allows risk to be decomposed into its various sources, provides an understanding of concentrations and can be used to test how various economic factors affect portfolio risk.

In recent years, several methodologies for measuring portfolio credit risk have been introduced that demonstrate the benefits of using internal models to measure credit risk in the banking book. These models measure economic credit capital and are specifically designed to capture portfolio effects and account for obligor default correlations. Several portfolio credit risk models developed in the industry have been made public; e.g., CreditMetrics (Gupton et al. 1997), CreditRisk+ (Credit Suisse Financial Products 1997) and Credit Portfolio View (Wilson 1997a and 1997b). Others remain proprietary, such as KMV’s Portfolio Manager (Kealhofer 1996). Although the models appear quite different on the surface, recent theoretical work has shown an underlying mathematical equivalence among them (Gordy 2000; Koyluoglu and Hickman 1998). However, the models differ in their distributional assumptions, restrictions, calibration and solution. Also, empirical work...