A Research About Asymmetry Herding In China’s A-share Market Based On Continuous Belief System

Based on the extended continuous belief system (CBS), this article investigates theexistence of asymmetric herding behavior and the effects on volatility in Chinese A-sharemarket. The data we chosen is the Shanghai and Shenzhen A-share market capitalizationweighted market index from July3,1991to July31,2012, data from CCER database. Andaccording to the extent of regulation and economic background, the entire period is dividedinto three sub-periods: the initial period, the middle period and the crisis period. Wecompare the three sub-periods with each other. After deduced in accordance with theassumptions of the extended CBS can use an asymmetric GARCH model to simulateasymmetric herd behavior. Estimated asymmetric GARCH model suggests: there areasymmetry herding behaviors in all three sub-periods. Interestingly, in the middle periodand the crisis period, the negative return shocks have a greater impact on herding thanpositive return shocks, which is the same as what we usually see in stock marker. However,on the contrary, in the initial period the negative return shocks have a greater impact onherding than positive return shocks, we call it opposite asymmetric herding。Besides, theperiod of crisis is the most asymmetric period. Subsequently, the EGARCH model is usedto analysis whether there is asymmetry in volatility. The estimated outcomes indicate thatasymmetry indeed exists in volatility, and its characteristics are exactly the same asasymmetric characteristics in herd behavior. So there must be a clear link betweenasymmetric herding and asymmetric volatility. This conclusion gives a strong support tothe asymmetric effect hypothesis of herding, and also shows that asymmetric herding canbe used to explain the asymmetric volatility, especially the opposite asymmetry whichcould not be explained with traditional hypothesis (such as leverage, the feedback effect).This conclusion shows that asymmetric volatility in China’s A-share market can be reducedby preventing the occurrence of asymmetric herding.