Herding Behavior Existence in MSCI Far East Ex Japan Index: A Markov Switching Approach

Abstract

This paper examines the herding behavior in the MSCI Far East Ex Japan stock exchange (MSCI-FE) using the Markov switching approach. An analysis of daily data from December 26, 2012, to June 17, 2019, is considered to find evidence of herding behavior in MSCI-FE stock market under different market regimes. The market is divided into two regimes (upturn and downturn) to determine whether the herding exists in an upturn or downturn market. In this study we consider four models consisting of linear regression, linear regression with time-varying coefficient, Markov Switching regression (MS), and Markov Switching regression with a time-varying coefficient (MS-TV). To detect the herding behavior, we employ the Cross-Sectional Absolute Deviation (CSAD) method of Chang et al. (J Bank Finance 24(10):1651–1679, 2000) and Chiang and Zheng (J Bank Finance 34(8):1911–1921, 2010). The results show that herding behavior is present in MSCI-FE. From the results of MS and MS-TV models, the herding exists only in the market upturn regime, indicating that investors will neglect their information and choose to mimic larger investors during the market upturn. On the other hand, during the market downturn, we do not find any evidence of the herding behavior.