Estimating Credit Contagion in a Standard Factor Model

State-of-the-art credit risk portfolio models and the New Basel Capital Accord consider only symmetric dependencies between borrowers in a portfolio, such as correlations. Recently, asymmetric dependencies have been introduced by Davis/Lo (2001) among others. However, statistical estimation techniques and empirical evidence on contagion is still rather scarce. The present paper provides a simple credit risk portfolio model extension to credit contagion and shows how its parameters can be easily estimated and tested. We apply our methodology to a dataset provided by Moody’s Investor Services and find significant contagion effects. By stress testing we show how contagion can seriously affect credit losses.