Price Discovery and Time-Varying Causality Dynamics in Energy Markets: Futures Versus ETFs

This paper investigates price discovery and time-varying causality between energy futures and exchange-traded funds (ETFs) across three major markets: WTI crude oil, natural gas, and gasoline. Drawing on a unique dataset spanning almost two decades, the analysis employs rolling Information Leadership Share (ILS) measures with Monte Carlo simulation bands and recursive time-varying Granger causality (TVGC) tests to trace the evolution of leadership dynamics between ETFs and their underlying assets. The results provide novel empirical evidence that futures generally dominate price discovery under stable market conditions, reflecting their superior liquidity and informational efficiency. In contrast, ETFs assume a more prominent role during episodes of market turbulence, such as the 2014–2016 oil price collapse and the COVID-19 crisis, where they shift from being passive followers to active price setters. Bidirectional, time-varying causality emerges in the WTI–USO and natural gas–UNG pairs, particularly during crisis periods, whereas gasoline markets exhibit a more conventional futures-led process. The study contributes to applied computational economics by illustrating how simulation-based and recursive estimation methods can uncover dynamic leadership patterns in financial markets. It further advances the understanding of financialisation in commodity markets, offering insights into how ETFs reshape information transmission during periods of heightened uncertainty.