Out-of-sample prediction of the oil futures market volatility: A comparison of new and traditional combination approaches

Abstract

This paper aims to use both the standard and iterated combination approaches to accurately predict the oil futures market volatility. We further make a comprehensive comparison of the out-of-sample forecasting performance between the two paired combination approaches. According to both the Diebold-Mariano test and model confidence set test, the iterated combination approach generates significantly more accurate volatility forecasts than the standard counterpart. The Direction-of-Change test suggests that the iterated combination approach also has substantially higher directional accuracy. We document that these results are robust to various settings. Furthermore, a mean-variance investor can obtain sizeable economic gains when she uses the iterated combination forecasts instead of the standard ones to allocate her portfolio.