Climate Risks and Predictability of the Trading Volume of Gold: Evidence from an INGARCH Model

We investigate whether text-based physical or transition climate risks forecast the daily volume of gold trade contracts. Given the count-valued nature of gold volume data, we employ a log-linear Poisson integer-valued generalized autoregressive conditional heteroskedasticity (IN-GARCH) model with a climate-related covariate. We detect that physical risks have a significant predictive power for gold volume at 5- and 22-day-ahead horizons. Additionally, from a full-sample analysis, it emerges that physical risks positively relate with gold volume. Combining these findings, we conclude that gold hedges physical risks at 1-week and 1-month horizons. Similar results hold for platinum and palladium, but not for silver.