Chao Li

In view of the contradicting results in existing research, this article proposes a new weighted cross-sectional variance (WCSV) model to re-examine the level of herding behaviour in the Chinese A-share market. Motivated by the original WCSV model, we utilize a Fama-French augmented seven-factor model as the underlying Arbitrage Pricing Theory model, which introduces the trading volume and turnover rate factors to the Fama-French five-factor model. The regression results show the superiority of our new model compared to the WCSV model based on Fama-French three- and five-factor models, which implies that the empirical findings of herding with our WCSV model are more reliable in relative terms. In the empirical aspect, in addition to testing the herding level yearly and integrally, to provide further insight on the relationship between market stress and herding, we apply our model to the Chinese A-share herding behaviour within each of three well-known crisis periods. In addition, we also split the sub-samples into pre-crisis and post-crisis periods to detect the existence of asymmetric herding behaviour for different market directions. Our findings suggest that Chinese A-share herding behaviour is more prevalent during large market turmoil, especially under condition of down market.