This work studies threshold models in credit risk modeling. Measuring credit risk is an important subfield of financial mathematics. At first glance the area turns to be very heterogeneous, due to rapid and decentralized development of many different model approaches. However, this paper develops a way to compare different threshold models. The focus is on the analysis of default probabilities for individual borrowers and credit portfolios. In this context various approaches — Merton’s model, Moody’s KMV, J.P. Morgan’s CreditMetrics and Li’s model — are introduced with respect to individual borrowers and expanded at portfolio level by using the copula theory. Finally, this paper discusses the choice of the Gaussian copula, which the presented multivariate threshold models use explicitly or implicitly, and points to the model risk behind them.
«
This work studies threshold models in credit risk modeling. Measuring credit risk is an important subfield of financial mathematics. At first glance the area turns to be very heterogeneous, due to rapid and decentralized development of many different model approaches. However, this paper develops a way to compare different threshold models. The focus is on the analysis of default probabilities for individual borrowers and credit portfolios. In this context various approaches — Merton’s model, Mo...
»