Out-of-Time Validation of Default Probabilities within the Basel Accord: A Comparative Study

Posted: 4 May 2017 Last revised: 19 Nov 2019

Date Written: February 27, 2019

Abstract

The probability of a default determines capital requirements set by the Basel Accord. To back-test default probabilities is therefore possibly the single most important task when validating credit rating models. A proper validation procedure must fulfill two criteria. First, it must be able to detect default probabilities which are too low, i.e. it must have satisfactory power. Second, it must not reject adequate default probabilities, i.e. its type I error should be in line with the significance level of the test. In this paper we compare several validation procedures via a simulation and empirical study. Our result is that many validation procedures do not fulfill both criteria.

Keywords: Validation Default Probabilities Credit Rating Basel Accord

JEL Classification: G21, G24, G28, C11, C12

Suggested Citation

Blümke, Oliver, Out-of-Time Validation of Default Probabilities within the Basel Accord: A Comparative Study (February 27, 2019). Available at SSRN: https://ssrn.com/abstract=2945931 or http://dx.doi.org/10.2139/ssrn.2945931

Oliver Blümke (Contact Author)

Raiffeisen Bank International ( email )

Am Stadtpark 9
Vienna, A-1030
Austria

Do you have negative results from your research you’d like to share?

Paper statistics

Abstract Views
1,099
PlumX Metrics