Do investors herd in a volatile market? Evidence of dynamic herding in Taiwan, China, and US stock markets

This article investigates whether investors exhibit herd behavior in a high market volatility state. A modified herding model with the Kalman filter and GARCH methodology is used to estimate the time-variation herding corresponding to each influential event. Our proposed model provides comprehensive results for the relationship between investor herding and the market state, which has been argued in the previous literature. We find that investors indeed herd in volatile markets, including the 2001 dot.com bubble and the 2009 global financial crisis. However, in recent years, anti-herding is prevalent and herding is slight even in turbulent markets, such as the 2020 Covid-19 pandemic.