Volatility forecasting and risk management for commodity markets in the presence of asymmetry and long memory

Abstract

This paper explores the relevance of asymmetry and long memory in modeling and forecasting the conditional volatility and market risk of four widely traded commodities (crude oil, natural gas, gold, and silver). A broad set of the most popular linear and nonlinear GARCH-type models is used to investigate this relevancy. Our in-sample and out-of-sample results show that volatility of commodity returns can be better described by nonlinear volatility models accommodating the long memory and asymmetry features. In particular, the FIAPARCH model is found to be the best suited for estimating the VaR forecasts for both short and long trading positions. This model also gives for all four commodities the lowest number of violations under the Basel II Accord rule, given a risk exposure at the 99% confidence level. Several implications for commodity market risks, policy regulations and hedging strategies can be drawn from the obtained results.