Geopolitical risk uncertainty and oil future volatility: Evidence from MIDAS models

Abstract

Using a textual analysis based geopolitical risk (GPR) index, this paper exploits the effects of geopolitical risk uncertainty on oil futures price volatility within a mixed data sampling (MIDAS) modeling framework. With a variety of MIDAS specifications, our in-sample estimation results suggest that the short-term (e.g. one-day-ahead) oil realized volatility is positively associated with GPR uncertainty, and our out-of-sample forecasting exercise indicates that the GPR index is useful for improving short-term oil futures volatility prediction. In addition, we find that the categorical GPR index: GPR action related index (GPA), contributes more to the long-term oil volatility forecasting, compared with GPR threat related index (GPT).