Revisiting the Relationship between Geopolitical Risk and Oil Price Realized Volatility: A Markov-Switching Analysis

Résumé

This paper examines the impact of geopolitical risks (GPR) on oil price volatility over the past three decades, using specifications that capture both long-term patterns and short-term disruptions during periods of stability and crisis. In particular, we analyze oil price volatility’s response to geopolitical risks, considering various geopolitical tensions including oil price shocks, political revolutions, wars, trade conflicts, economic sanctions, as well as measures of geopolitical risk acts and threats. We compare a linear benchmark model with proposed nonlinear econometric frameworks that account for structural breaks, regime switching, and asymmetries. The results show that geopolitical risks exert strong and immediate effects on oil volatility, while threats also have significant lagged impacts through expectations and speculative behavior. Furthermore, when accounting for structural breaks and nonlinearities, the proposed nonlinear models – Markov Switching Model with Breaks (MSMB) and Dynamic Factor Markov Switching model with Breaks (DFMSB) – confirm the significant effect of GPR on oil volatility. In sample and out-of-sample forecasting tests indicate that incorporating geopolitical risks within nonlinear models reduces prediction errors by roughly 50% and improves variance tracking measures two to five -fold. Overall, GPR – especially threats – exerts a regime-dependent, asymmetric influence on oil volatility, with the DFMSB (built only on GPR measures and realized oil volatility) delivering the most accurate predictions and highlighting the predictive value of geopolitical information in nonlinear settings.