Herding Behavior in Commodity ETFs

In this study, we use high-frequency microstructure components to explore commodity ETF herding. We employ a new GARCH model incorporating cross-sectional and market volatility at 15-, 30-, 45-, and 60-minute intervals. We document that during market instability and the COVID-19 pandemic, agriculture, and metal-based ETFs herd less, whereas the former does the reverse. Regarding frequency, in normal market circumstances, herding typically occurs beyond 30 minutes. However, broad basket commodities and energy-based ETFs tend to herd at various frequencies. Finally, we document that our results help economists and policymakers mitigate negative outcomes such as asset price bubbles or financial instability.