Forecasting oil futures price volatility with geopolitical risk: A SV-MIDAS model

Abstract

This paper proposes the stochastic volatility model with mixed data sampling and geopolitical risk (SV-MIDAS-GPR) for modeling and forecasting the volatility of INE crude oil futures. The model is capable of capturing both the impact of geopolitical risk on the volatility of INE crude oil futures and the high persistence of volatility. We develop the maximum likelihood method based on continuous particle filters to estimate the model parameters. Our empirical results show that the SV-MIDAS-GPR model outperforms a variety of benchmark models in both in-sample fit and out-of-sample volatility forecasting, including the GARCH model, the stochastic volatility (SV) model, and the SV-MIDAS model, thus highlighting the value of incorporating both the component volatility (MIDAS) structure and geopolitical risk into volatility modeling and forecasting.