Saturday, 7 October 2017: 3:15 PM
This paper aims to investigate the herd behavior of investors in international stock markets, and examine whether the herd behavior across national borders moves together. We discuss the similarities and differences of herding behavior across countries and regions, and also investigate what factors affect the global and regional co-movements of the behavior. In addition, we also explore what factors influence national herding activity’s sensitivities to global or regional factors and country-specific components. Here, the stock market herding is measured by the cross sectional absolute deviation (CSAD) for each country. A dynamic latent factor model with Gibbs sampling is employed to decompose national herd behavior into worldwide, regional, and country-specific components. The data from Thomson Financial’s DATASTREAM database covers 48 countries for the sample period from January 3, 2000 through December 31, 2014. Our results indicate that the high frequency comovement of herd behavior could be captured by the worldwide factor, implying the impact of worldwide factor on the national herd behavior is short-lived. Moreover, the worldwide and regional factors play an important role in explaining variation of herd behavior: they explain about 30 percent of herd variability. In addition, the regional factor plays a more important role in explaining herd behavior movements than the worldwide factor. As a whole, this paper contributes to the literature by providing new insights in studying the interdependence of herding activity across different stock markets. Thus, the empirical findings serve to guide not only investors in making their international portfolio allocations, but also policy makers in monitoring the financial markets.
Keywords: Herd behavior, International comovement, Dynamic latent factor model, Gibbs sampling