Keywords: copula, credit risk models, global sensitivity, latent factor models, one-at-a-time analysis, sensitivity analysis, variance-based methods, credit portfolio, risk assessment, financial services
Global sensitivity analysis for latent factor credit risk models
This paper proposes the use of global sensitivity analysis to evaluate the risk associated with a credit portfolio model. Although successfully applied in many disciplines such as chemistry, biology and environmental science, the use of global sensitivity analysis is still rare in financial contexts. One-at-a-time sensitivity analysis is usually applied despite its shortcomings. The scope of this paper is to demonstrate the advantages of a more comprehensive analysis, that is a global sensitivity analysis. As an example, following the works by Frey et al. (2001) and Kiesel and Kleinow (2002), we analyse the static and time-varying uncertainties of three inputs in a latent factor credit risk model: the multivariate distribution of the latent variables, the correlation, and the default probabilities of the obligors. Results show that the relative importance of the inputs strongly depends on the average default probability of the portfolio and the analysed quantiles of the default distribution.