Timescale Patterns of Causality between Climate and Geopolitical Risks and Commodity Markets

We study the multiscale predictive causal relationship from geopolitical risk, physical and transition climate risk, El Niño conditions, and U.S. consumer confidence to a broad cross-section of 45 commodity prices, using monthly data over the period 2008–2025 (202 observations) and Multivariate Variational Mode Decomposition (MVMD) followed by Pattern Causality. Statistically supported predictive links are concentrated at the long-run horizon and are directional (positive and negative); regime-switching ("dark") causality is not statistically supported once multiple testing is controlled. Geopolitical and physical-climate risks account for the largest number of long-run links. These long-run relationships are, however, level-based: they attenuate under a stationary (returns) transform and out-of-sample, and once common global factors (the U.S. dollar, industrial production, inflation, and interest rates) are controlled, the geopolitical and climate links no longer remain significant, whereas consumer confidence retains an idiosyncratic long-run relationship. The evidence thus points to long-run co-movement largely shared with common macroeconomic factors for geopolitical and climate risks, with consumer confidence the exception.