Modeling Dependent Credit Risks for Application to Off-Site Banking Supervision

During the past five years the Oesterreichische Nationalbank (OeNB), together with the Austrian Financial Market Authority (FMA) and university experts, has developed and implemented several modern tools for the purposes of off-site banking analysis and supervision. One of these tools is the Value-at-Risk (VaR) model, which allows for the standardized quantification of every single bank’s economic capital. Within this portfolio model framework, a total VaR is calculated as an aggregation of credit, market and operational VaR, assuming perfect correlation between the risk categories. The methodology for measuring the credit risk of a bank’s portfolio is currently based on the standard CreditRisk+ model, an actuarial model for aggregating risks in a credit portfolio with a single risk factor. In 2005 the OeNB and the Vienna University of Technology launched a research project with the aim of developing an extended version of the credit risk model that is able to account better for portfolio diversification effects. As the background risk factors in the standard CreditRisk model have to be orthogonal, resemblance to real-world industrial sectors or other macroeconomic factors, which often appear to be strongly correlated, is not possible. This paper gives an overview of our approach to modeling correlations among systematic risk factors. Other extensions of the model, like the ability to calculate a single obligor’s risk contribution and the incorporation of stochastic loss given default, are touched upon.