The Research Of Herd Behavior In Chinese Stock Market Based On High Frequency Data

Since Chinese stock market was established, its volatility shows stronger than those developed capital market. Only in 2015 had Chinese stock market happened two big crash. Those market fluctuation made us think how can we make Chinese capital market run rationally. Affected by frequent fluctuation of the capital market, investor affected easily by policies and other investors have great difficulty in making rational investment. At the same time, due to the frequent policy intervention and serious asymmetry information, investors tend to follow others, which lead to market volatility and have a bad effect on the stability of stock market. Therefore, herd behavior has a great influence on the stability of stock market and the efficiency of resource allocation in the financial system. It is a key factor in price fluctuations of the capital market. So studying herd behavior has great significance to the stock market.To measure herd behavior, most of researches are based on constructing linear regression model to analyze low frequency trading data which include day-data, week-data and month-data, etc. But herd behavior is a fleeting and complex process, whose short duration means that it is difficult to capture the characteristics of its changes by studying low-frequent data. So the traditional researches methods can not make a fundamental description to the time of herd behavior and time of duration. In order to better understand the internal characteristics of herd behavior, we must use intraday high frequency data to depict herd behavior accurately in the stock market.This paper respectively use high frequency and low frequency trading data to test herd behavior in Shanghai stock market. The difference between high and low frequency data is, low frequency data can only detect whether there is herd behavior in a period of time, but high frequency data can not only identify the specific days when herd behavior happened, but also can detect the time of herd behavior in a trade day, duration time, and the influence degree of different size companies to herd behavior. etcThe researches results of high frequency data show that the herd behavior of Shanghai stock market in 2015 mainly happened in two stock market crashes. Market behavior is complicated, which is affected by various factors. So the time of herd behavior is irregular which may happens in any time in trade day(it could appears in opening quotation, intraday and closing quotation) and only sustains a few dozen minutes. In the absence of policy intervention, the contribution of small companies to herd behavior is higher than big companies, which means herd behavior is more likely to happen in small companies. When herd behavior occurs, there is a significant "small company effect" in the market.