Herding Behavior in Commodity ETFs

This paper is a thorough investigation into the herding behavior of commodity ETFs, utilizing microstructure components at high frequencies and drawing from price-sensitive information. Our analysis employs a novel GARCH type model with cross-sectional volatility at various (15, 30, 45, and 60 minutes) intervals, offering a nuanced perspective on the topic. To ensure the reliability of our results, we also consider market volatility in our examination of herding behavior. We also separate our analysis of the Covid-19 period to determine whether the phenomenon has changed in the context of market turbulence and economic uncertainty. Our results indicate that agricultural and metal-based ETFs are generally less prone to herding, although the former exhibits opposite tendencies during periods of market volatility and the Covid-19 pandemic. With regard to frequency, herding typically occurs beyond a 30-minute interval under normal market conditions, with the exception of agricultural ETFs during the Covid-19 pandemic, which incurred herding at all frequencies. This is a notable and important finding of our study. Additionally, broad basket commodities and energy-based ETFs are generally susceptible to herding across a range of frequencies. Drawing from our conclusions, we suggest potential pathways for future researchers to explore the drivers of herding at various scales. We emphasize the importance of this research for economists and policymakers, as it can help to mitigate negative outcomes such as asset price bubbles or financial instability.