Random thinning with credit quality vulnerability factor for better risk management of credit portfolio in a top-down framework

Abstract

In the top-down approach of portfolio credit risk modeling, we assess credit risks of sub-portfolios with the so-called random thinning model, which dissects the portfolio risk into sub-portfolio contributions. In this paper, we provide a random thinning model incorporating the sub-portfolio size and the factor called “credit quality vulnerability factor”, in order to take into account credit quality vulnerability of sub-portfolios. With our random thinning model, we estimate credit quality vulnerability of industrial sectors. Numerical examples on assessing the risks of several credit portfolios show that our random thinning model is useful to detect how the proportions of constituent industrial sectors affect portfolio credit risks.