Incorporating severity variations into credit risk

Peter Bürgisser, Alexandre Kurth, and Armin Wagner

Abstract

ABSTRACT

The authors present an approach to modeling credit risk that incorporates the risk of counterparty default and the risk of devaluation of the collateral. The framework is based on a segmentation by industry and collateral type. The systematic risk in both drivers is taken into account by volatilities within and by correlations between the segments. A simple formula is derived for the variance of the loss distribution and an algorithm to compute this distribution is described. Moreover, it is shown that, in the limit of a large portfolio, the loss distribution directly mirrors the assumptions on the economy and depends on the portfolio structure only through the expected loss distribution across the segments.