Geopolitical fragmentation, climate risk, and crude oil price dynamics: Evidence from TVP-VAR-SV and causal forest models

Against the backdrop of accelerating geopolitical fragmentation and deepening climate-related uncertainty, this study investigates the dynamic spillover effects between geopolitical risk (GPR), climate risk (CR), and crude oil markets (WTI and Brent) using weekly data from 2000 to 2025. Employing a Bayesian Time-Varying Parameter Vector Autoregression model with Stochastic Volatility (TVP-VAR-SV) and the Diebold-Yılmaz connectedness framework, we quantify spillover direction and magnitude, while static and dynamic networks capture structural shifts around key geopolitical episodes—the Russia-Ukraine and Gaza wars. A Causal Forest model further examines how climate variables mediate GPR's impact on oil markets. The findings reveal that GPR persistently acts as a net transmitter of shocks, with WTI and Brent absorbing risk to varying degrees. Spillovers intensify during major geopolitical and climate-related stress events. While WTI is more sensitive to domestic climate policy and U.S.-related risks, Brent reacts to global shocks more moderately, reflecting its role as the international benchmark. A growing synchronisation between the two oil benchmarks is detected during prolonged crises, particularly after the outbreak of the Gaza war. The Causal Forest results reveal that natural disasters are the most influential climate moderator, with GPR's impact on oil prices being strongest during periods of low disaster activity—suggesting that markets reallocate attention toward geopolitical tensions when environmental stress recedes. These findings carry important implications for energy security strategy, climate-aligned investment, and policy design in an era shaped by the dual pressures of geopolitical fragmentation and climate transition.