Description:
In this paper, we apply two state-space models based on the cross-sectional dispersion of the factor sensitivities (betas) of securities for extracting the time series of herding towards the market, size/growth (SMB), value (HML), operating profitability (RMW), investment (CMA) and momentum (WML) factors. The sample covers all stocks listed in Europe from December 1992 to December 2020. We find that herd behaviour is statistically significant and persistent in all factors, regardless of the market state and macroeconomic conditions. However, when conditioning herding levels across market state factors, herding increases at the lower tail of return distribution and during market crises. Herding increases at the upper tail and decreases at the lower tail of the volatility distribution. These results support a loss aversion bias, suggesting that investors herd towards the market consensus when they face higher risks. The study reveals evidence of mispricing and limitations on portfolio diversification, which casts doubt on the market efficiency and has implications for market participants