The Analysis of Power For Some Chosen VaR Backtesting Procedures: Simulation Approach

Everyone who measures the market risk using the Value at Risk (VaR) approach should test if the assumed model is correct. This procedure is called backtesting. There are many different tests available, but usually risk managers are not concerned about their power. The aim of this paper is to analyze some chosen backtesting methods focusing on the problem of power of the tests and limited data sets. The paper is organized as follows. At the beginning a financial aspect of the analyzed problem is presented very briefly. The second part gives information about some chosen, but (in the author’s opinion) the most popular backtests. The main attention is paid to tests based on the frequency of failures and on multiple VaR levels. Next, the results of the simulations are presented. The last part summarizes obtained results and gives hints for the optimal backtesting.