Conditional value at risk (CVaR) is widely used in risk measure that takes into account losses exceeding the value at risk level. The aim of this research is to compare the performance of the EVT-GJR-vine copula method and EVT-GARCH-vine copula method in estimating CVaR of the portfolio using backtesting. Based on the backtesting results, it was found that the EVT-GJR-vine copula method have better performance when compared to the EVT-GARCH-vine copula method in estimating the CVaR value of the portfolio. This can be seen from the statistical values ??, and of EVT-GJR-vine copula method which is generally smaller than the statistical values , and of the EVT-GARCH-vine copula method.
Value-at-Risk Forecasting Ability of Filtered Historical Simulation for Non-Normal GARCH Returns