Calculation of Operational Loss Distribution via Bayesian MCMC Algorithm: Evidence from China's Commercial Banks

Abstract

this paper reviews the operational risk data of China's commercial banks from 1994 to 2008, and studies its type of distribution. In order to precisely capture the profile of the operational loss and event distribution of China's commercial banks, we select the operational risk loss distribution type with the Bayesian inference and test the GEV distribution on AIC and BIC standard. As closed-form solutions are not available for the operational risk distributions, we turn to the Bayesian MCMC algorithm for robust test on the selection. The result shows that with the increase of the iterations, the variance of estimated parameters becomes smaller, so we conclude that the operational risk loss distribution for China's commercial banks meets the Generalized Extreme Value (GEV) distribution.