Beschreibung:
This paper is among the first to investigate high-frequency herding tendencies among exchange traded funds (ETF) traders within the commodities asset class. Operating on 64 popularly traded and liquid ETFs spread across five sub-sectors from 2009 to 2021, we apply four different herding-detection models to capture traditional herding as well the effects of prevailing market volatility state at 15-, 30-, 45-, and 60-minute intervals. This allows us to derive novel insights as to which frequencies attract the most herding attention and which sectors are most prone to it. We further isolate the Covid-19 period to verify if the phenomenon changes in light of market turbulence and economic uncertainty. Our results suggest that agricultural and metal-based ETFs are least prone to herding in general, though the former exhibits opposite tendencies in times of market volatility and Covid-19 pandemic. With respect to frequency, in general market conditions herding mostly occurs beyond a half-hour interval. An exception is agricultural ETFs during the Covid-19 pandemic, whereby herding occurs in all frequencies. This is a striking and crucial finding of this study. Elsewhere, broad basket commodities and energy-based ETFs are generally susceptible to herding across most frequencies. We discuss the implications of our results and recommend avenues for further research by suggesting economic mechanisms which could explain the sectoral idiosyncrasies detected in our study