Examining the Value-at-risk Performance of Fractionally Integrated GARCH Models: Evidence from Energy Commodities

Abstract

This study examines the out-of-sample value-at-risk forecasting performance of the GARCH, FIGARCH, HYGARCH and FIAPARCH models for West Texas intermediate crude oil, Europe Brent crude oil, heating oil#2, propane and New York Harbour Conventional Gasoline regular under the standard normal, Student's t and skewed Student's t distribution assumptions. Additionally, the expected shortfall is calculated in all cases. The results clearly show that the HYGARCH model under the normal distribution is the most accurate for short trading positions, whereas the FIGARCH model under the Student's t distribution is preferred for long trading positions. This further implies that it is important to consider downside and upside risk separately to obtain   more accurate results.Keywords:  FIGARCH models, value-at-risk, expected shortfall, energy commoditiesJEL Classifications:  C58; C53; G31; Q40