Value-at-Risk estimation of energy commodities: A long-memory GARCH-EVT approach

- We evaluate Value-at-Risk (VaR) and expected shortfall (ES) for crude oil and gasoline commodities market for both long and short position.
- We adopt three (FI)GARCH models to forecast energy commodities volatility.
- Then we consider Extreme Value Theory (EVT) to model the tail distribution.
- The FIAPARCH-EVT model performs better in predicting VaR for energy commodities for different time horizons (1-day, 5-days and 20-days) and for both short and long trading position.

In this paper, we evaluate Value-at-Risk (VaR) and expected shortfall (ES) for crude oil and gasoline market. We adopt three long-memory-models including, FIGARCH, HYGARCH and FIAPARCH to forecast energy commodity volatility by capturing some volatility stylized fact such as long-range memory, heteroscedasticity, asymmetry and fat-tails. Then we consider extreme value theory which concentrates on the tail distribution rather than the entire distribution. EVT is considered as a potential framework for the separate treatment of tails of distributions which allows for asymmetry. Our results show that the FIAPARCH model with extreme value theory performs better in predicting the one-day-ahead VaR. Using the fitted long-memory GARCH-model and a simulation approach to estimate VaR for horizons over than one day, backtesting results show that our approach still performs for lower estimation frequencies. Overall, our findings confirm that taking into account long-range memory, asymmetry and fat tails in the behavior of energy commodity prices returns combined with filtering process such as EVT are important in improving risk management assessments and hedging strategies in the high volatile energy market.