Empirical investigations of herding behavior in fi nancial markets can be divided into two approaches.* e fi rst approach is to employ dynamic time series models to explore whether stock index returns display comovements. For instance, using correlation analysis, Boyer et al. (2006) estimated and compared the degree to which accessible and inaccessible stock index returns comove with the index returns of the country in crisis. eir study suggests that in emerging stock markets, there is greater comovement during high-volatility periods, indicating that crises that spread through the asset holdings of international investors are mainly due to contagion rather than to a change in fundamentals. Chiang et al. (2007) applied a dynamic conditional-correlation model (Engle, 2002) to examine daily stock-return data from six Asian markets from 1990 to 2003. eir empirical evidence suggests that the contagion eff ect takes place during the early stage of the Asian fi nancial crisis and that herding behavior dominates the later stage of the crisis.