Nonlinear dynamic correlation between geopolitical risk and oil prices: A study based on high-frequency data

Abstract

This study investigates the nonlinear dynamic correlations between geopolitical risk (GPR) and oil prices using nonlinear Granger causality and DCC-MVGARCH methods based on high-frequency data. The relationship between GPR and oil prices is found to have a complex nonlinear relationship rather than a simple linear one. Further, a bidirectional nonlinear Granger causality is found to consistently exist between GPR and oil volatility across different components of realized volatility. In terms of returns, GPR has relatively weak unidirectional nonlinear Granger causation with oil returns. The dynamic correlation analysis shows that GPR mainly affects oil volatility rather than returns. Moreover, GPR mainly affects oil volatility through the jump component of the oil market after the financial crisis, and there is a strong positive correlation between GPR and volatility jumps. Our findings innovatively suggest that GPR can potentially be utilized to improve models of volatility jumps and provide reference for investors and price analysts in oil markets who want to design sensible risk-management strategies.