Static and regime-dependent herding behavior: An emerging market case study

We contribute to behavioral finance literature by demonstrating the relative superiority of a dynamic regime-sensitive approach in unraveling herding phenomenon. Employing daily data in Bursa Malaysia from 1995 to 2016, we first apply two orthodox techniques: cross-sectional standard deviation of returns (CSSD) model of Christie and Huang (2005) and the cross-sectional absolute dispersion (CSAD) model of Chang et al. (2000). The ensuing results appear inconsistent, with only one model capturing herding. In contrast, the dynamic approach with a two-state Markov Switching model reveals that herding is a heavily regime-dependent and non-linear phenomenon. A deeper dive via sectoral decompositions shows that the financial sector and large- and mid-capitalization segments are more herding-prone.