The case for herding is stronger than you think

In this paper, we challenge the often implemented herding measure by Chang et al. (2000). They regress the cross-sectional absolute deviation of returns on the absolute and squared excess market return. A coefficient on the squared excess market return significantly smaller than zero is interpreted as evidence for herding. However, we show that the true coefficient is positive under the null hypothesis of no herding. Hence, their test is biased against finding evidence in favour of herding. Empirical examinations for the S&P 500 and the EuroStoxx 50 confirm the misleading implications of Chang, Cheng and Khorana's measure, while our modified test provides clear-cut evidence for herding behaviour.