| Summary: | The study’s aim is to analyse the herding behaviour of the largest capital market in Vietnam – Ho Chi Minh Stock Exchange from 03/01/2005 to 29/12/2017. The method proposed by Chang et al (2000) is applied to test for the existence of herding due to its superior effectiveness compared to other methods. Following the method, the indication of herding is the negative and statistically significant relationship between the return dispersion and the square of the market return.
As to establish a thorough knowledge of herding in Vietnam, the matter is analysed under different situations of the market: before and after the global financial crisis; in rising and falling markets; herding on industry-level. For all cases, the existence of herding is strongly supported by the analysis. Specifically, the phenomenon in the pre-crisis period is stronger than that in the post-crisis period. Rising markets exhibit stronger herding than falling ones. On industry-level, 43 out of 45 industries indicate strong statistical evidence of herding. Furthermore, for the improvement purpose, a modified model is proposed by the author to test herding in rising and falling markets. The data indicates the efficiency of this model with a higher coefficient of determination (R^2) and less serial correlation effect.
In addition, the robustness check is run to see whether the findings would alter after the influential factors (portfolio’s size and daily price limit) are controlled. The test shows that our findings still strongly hold.
|
|---|