Measuring the Value at Risk of the Price of OPEC's Oil Basket through Use of Long-Memory GARCH Models

In this paper، we evaluate the performance of parametric models in forecasting the value at risk for the OPEC basket price. We compute VAR [2] for 3 Arch/Garch- type models، including FIGARCH، HYGARCH and FIEGARCH for one day and 10 day periods. These models are estimated for three alternative distributions، namely: normal، student and skewed student. Our results show that long-range memory and fat tails are present in the volatility of price returns of the OPEC basket. Moreover skewed distribution performs better in predicting one and ten day ahead value at risk. The FIEGARCH model outperforms the other models in terms of predicting the value at risk for both time frames.