Testing opvar accuracyan empirical back-testing on the loss distribution approach

Abstract

The application of the Value at Risk (VaR) concept to the Loss Distribution Approach (LDA) is encouraged by the Basel Committee for measuring the operational risk. Moreover, complementary analysis such as the back testing exercise plays an important role in assessing the excedancees beyond Operational Value at Risk (OpVaR) forecasts and providing with valuable feedback on the soundness of such advanced measurement approach (AMA). In this paper, we conduct an empirical back-testing analysis on the LDA by using an Internal Operational Losses Database (IOLD) provided by a medium sized Spanish Savings Bank. We apply different techniques for carrying out the back-testing exercise: the Basic Analysis and Extremal Index, and more complex statistical methods such as Kupiec and Christoffersen�s Tests. Our empirical results bring into light that the application of the LDA model for the Savings bank analyzed would be rejected according to the regulatory framework.