Forecasting the realized volatility of the oil futures market: A regime switching approach

Abstract

Considering nonlinear and highly persistent dynamics of realized volatility, we introduce Markov regime switching models to the Heterogeneous Autoregressive model of the Realized Volatility (HAR-RV) models to forecast the realized volatility of the crude oil futures market. In-sample results demonstrate that the high volatility regime is short-lived. Out-of-sample results suggest that HAR-RV models with regime switching increase the forecasting ability significantly than those without regime switching. Moreover, these findings are robust for different actual volatility benchmarks, forecasting windows, and model settings.