BACK-TESTING APPROACHES FOR VALIDATING VAR MODELS

Abstract

Value at risk (VaR) is one of the important market risk measures. It measures the possible potential loss on given investment in terms of value, with certain probability for certain time horizon. In this paper, our aim is to discuss different back-testing approaches to validate VaR models, and also test it the real market data. We back tested VaR of Nifty 50 index obtained by Variance Co-variance method, Historical simulation method, Monte-Carlo simulation, and cubic polynomial regression method. We have used Total exceptions by binary back-testing over entire population. we have also used Basel Traffic Light Zone Test, Kupiec POF-test, Kupiec TUFF-test, and Haas’ Mixed-Kupiec test and analyzed the above methods.