Evaluation of the Value-at-Risk Estimation Methods with applying a Penalty for Risk Overestimation

In this paper, Value at Risk for Gold prices Is estimated by the Extreme Value theory and parametric method with Normal and t-student distribution for disturbance term in the mean equation together with a range of the conditional variances estimation techniques including, GARCH (1.1), TGARCH, EGARCH, PGARCH, FIGARCH and FIEGARCH Models. The two-stage Back-Testing method is used to evaluate the adequacy and accuracy of the calculation methods. Furthermore, we rank the accuracy of the estimation methods by a loss function. Our findings show that the most accurate method, In terms of the value of the loss function and among the applied econometrics methods, is VaR by t-student distribution for gold return and PGARCH for the long position and acceptable performance for the short position.