Asymmetric And Contagion Effect Of Herd Behavior:Evidence From Chinese Stock Market

Herd behavior has been one of the hot spots of researches in the investment area and the area of behavioral finance,herd behavior in stock market will lead to asset price deviates from their basic value,it will cause the price bubbles or the bubble burst,and exert an bad influence on the capital asset pricing model and investment trading strategies,the researches of herd behavior can make investors more accurately assess the value of financial assets.The stock market has fluctuated more and more often,and the internationalization of Chinese stock market is gradually deepening,the researches of the contagion effects and asymmetric effects of herd behavior in different markets and in different market conditions is a starting point that can not be ignored.In addition,as high-frequency data and limit order books have been widely used in the field of finance research,we can analyze herd behavior from a more microscopic point of view.The study of herd behavior in the financial market is mainly divided into two kinds of methods,one method is to consider the market as a whole,it tests the herd behavior by comparing degree of deviation of the return of individual stock and the return of market.The other method takes the specific investors in the market as the research object,and analyzes whether the different kinds of investors imitate the investment behaviors in the process of investment.Based on this,this paper will use two research methods respectively to study the asymmetric effects and contagion effects of herd behavior.First of all,based on the CSAD method which is used to study the herd behavior,this paper constructs a nonlinear Markov switching seemingly unrelated regression model(MS-SUR model),this model can automatically classify the state of the stock markets,this model combine the Shanghai stock market and Hong Kong stock market as a whole to be estimated,exploring the asymmetric effects and contagion effects of the herd behavior between the two markets in different market conditions.Using the MS-SUR model in this paper can divide the state of the market automatically,and describe the dynamic characteristics of variable transition between state to state specifically when the relationship with each other changes,it can be an more accurate and clear response to the current state of the market,which can describe the specific characteristics of herd behavior more accurately in different market conditions.Furthermore,this paper uses transaction data,and divides investors into noise traders and information traders according to the size of the order,calculating the trading volume index,and by constructing a VAR model to study the volume index,the VAR model can help us judge the feedback effect of different types of investors' investment behavior,and then analyze the interaction between herding behavior of different types of investors.Finally,it explains the imperfections of this study and puts forward the prospect for future research.