Geopolitical Uncertainty and Gold Market Volatility

This paper investigates whether geopolitical risk (GPR) affects gold returns and volatility over time in an increasingly multipolar world with growing geopolitical tensions and increasing de - dollarization. Given recent interest in gold as a safe haven asset and tool for portfolio diversification, this study seeks to confirm whether GPR influences these aspects of gold returns and volatility. Thus, we use a generalized econometric framework that incorporates both traditional and augmented volatility models with daily gold prices and the GPR Index. Our approach follows that of GARCH (1, 1) (with GPR integrated into the modelling specification) and GARCH - X (where GPR becomes a covariate). We also employ the Akaike Information Criterion and the Bayesian Information Criterion, in addition to using several tests on the residuals, to select a sufficient model. Lastly, we augment our methodology to improve its predictability by using Gaussian Process Regression (GPR), an ML technique useful for forecasting volatility with potentially nonlinear patterns and measuring forecasting out - of-sample. We show that GPR is negligibly significant for predicting mean gold returns, but is effective at capturing variations in conditional volatility and does improve prediction relative to our benchmark when there′s significant uncertainty over politics. We infer that diversification against geopolitical events afforded by gold is beneficial for portfolio managers given the shift in monetary power and geopolitical strategic reserve changes around the world.