Does the volatility spillover effect matter in oil price volatility predictability? Evidence from high-frequency data

Abstract

This paper examines whether volatility spillovers between oil and equity markets can improve the predictive power of oil's realized volatility. This paper uses intraday data for the WTI futures market and the U.S. and Chinese equity markets from January 4, 2006, to August 1, 2018, and the heterogeneous autoregressive (HAR) model as the benchmark. Our results reveal that the spillover effect has significant predictability and can indeed be applied to the prediction field. First, the results of in-sample estimation show that spillovers transmitted by equity markets have significant predictive information on the oil market. Second, the results of the out-sample provide evidence that net pairwise spillovers transmitted by the Chinese equity market provide significant forecasting gains and those transmitted by the U.S. equity market provide significant forecasting gains only during the precrisis period. Finally, we perform robustness tests, and the results are robust.