Forecasting energy commodity returns: Can weak factors and nonlinearity help?

Abstract

This study investigates whether incorporating nonlinear structures and weak factors can improve the predictive accuracy of energy commodity returns. Existing literature emphasizes the utility of technical indicators and dimensionality reduction techniques, but it often overlooks nonlinear dynamics and the role of weak factors. To address these gaps, we apply the scaled sufficient forecasting (sSUFF) method, a novel dimension reduction approach, to enhance return predictions. Empirical results show that sSUFF outperforms traditional methods both in-sample and out-of-sample. It remains robust across varying economic conditions and performs particularly well during periods of heightened market volatility, such as the COVID-19 pandemic and the Russia–Ukraine conflict. sSUFF’s advantage arises from its ability to capture nonlinear patterns and effectively distinguish between strong and weak predictors. Economically, sSUFF-based forecasts yield higher investor returns, highlighting their practical value in financial forecasting and their relevance to investment strategies, risk management, and policy decisions.