Modeling the Volatility of World Energy Commodity Prices Using the GARCH-Fractional Cointegration Model

Abstract

Energy commodity prices usually fluctuate, non-linear and non-stationary. These stylish facts pose a big challenge in predicting the volatility of energy commodity prices because they usually contain long memory. In the energy market, energy commodities are empirically cointegrated, and this characteristic is a consideration for combining GARCH with Fractional Cointegration. This study aims to model and compare the GARCH and GARCH-Fractional Cointegration on the price return volatility of each energy commodity. The results show that the GARCH-Fractional Cointegration model is better for long-memory non-stationary data, while the GARCH model is better for long-memory stationary data.