Forecasting energy futures volatility based on the unbiased extreme value volatility estimator

Abstract

This paper uses the opening, high, low, and closing prices of five energy futures to estimate and model volatility based on the unbiased extreme value volatility estimator (the Add RS estimator). The statistical and distributional properties of the logarithm of the Add RS estimator support the use of the appropriate order linear long-memory Gaussian model (ARFIMA-Add RS model) to model the Add RS estimator. The out-of-sample evaluation analysis indicates that the volatility forecasts based on the ARFIMA-Add RS model provide superior forecasts of realised volatility of energy futures in relation to the alternative models from the GARCH family. This suggests that the ARFIMA-Add RS model can be a viable candidate for generating more accurate forecasts of realised volatility for energy futures. The study has important implications for governments, policy makers, oil importers, and oil traders.

Keywords

• Energy futures
• Volatility modelling
• Volatility forecasting
• Forecast evaluation
• Bias-corrected extreme value estimator