Ranking the Predictive Performances of Value-at-Risk Estimation Methods

We propose a novel ranking model and a complementary predictive ability test statistic to investigate the forecasting performances of different Value at Risk (VaR) methods. The ranking model develops a unified framework which penalizes excessive capital allocation, autocorrelation of violations and violation magnitudes through an anisometric structure. In addition, as an alternative to existing predictive ability tests which compare the forecasting methods two at a time, our complementary test statistic considers all methods at the same time to confirm whether the chosen method outperforms the competing methods. The results show that asymmetric methods such as CAViaR Asymmetric and EGARCH generate the best performing forecasts. This suggests that the performance of VaR methods does not depend entirely on whether they are parametric, non-parametric, semi-parametric or hybrid; but rather if they can effectively model the asymmetry of the underlying data or not