Forecasting the volatility of energy transition metals

Abstract The energy and digital transitions are increasing demand for metals used in renewable energy, electrification, storage, and digital infrastructure. Many of these metals are traded in thin and concentrated markets with limited substitutability, making volatility a key constraint on investment, hedging, and policy design. This paper analyzes the return and volatility dynamics of 16 Energy Transition Metals. We combine exploratory and dimension-reduction methods with in-sample estimation and out-of-sample point and density forecasts, comparing a baseline GARCH(1,1) model with several stochastic-volatility specifications. The results reveal substantial heterogeneity across metals. Commodities commonly grouped by data providers or linked through geological co-occurrence often exhibit markedly different return and volatility characteristics. Forecasting performance is similarly heterogeneous: neither GARCH nor any single stochastic-volatility specification consistently dominates across commodities. Although some associations emerge between empirical features and model performance, they are not robust across metals or evaluation criteria. Overall, the findings support a metal-specific approach to volatility modelling and risk management and suggest that market maturity, liquidity, price discovery, and hedging opportunities may help explain these differences.