A Study On Effects Of Investors' Herding Behavior On Stock Market Based On Simulation

In stock market, the decision-making of an investor is very complicated, and investor's decision-making process will be affected by many factors. Different decision-making will bring different effects on the stock market. More and more people realize that subjective factors, such as behavior and emotion of human-being, play an indispensable role in the decision-making process of financial investment. The underreaction and overreaction of investors and so on such anomaly, which can't be explained by modern financial theory correctly, lead to the generation and the development of the behavioral finance study concerning the investor's behavior. Herd behavior of investors could cause stock market forth, reduce the efficiency of stock market and increase the system risk of stock market. Recently, theory related to herd behavior of investors has become one of the most important parts in behavioral finance. In the thesis, theory and simulation methodology for complex system have been used to the financial problems caused by herd behavior of investors.A simulation model of investors'herd behavior in the stock market has been established based on the theory of cellular automata and complex network, which is due to the high operating speed of cellular automata, its simple evolvement rules and the social network's advantage of complex network properly showing the relationship of human-being in real life.It is found that investors'homogeneity become stronger in the typical Moore network structure of cellular automata than in social network structure through simulation, which results in smaller volatility of stock market. Small volatility of stock market can't show the real character of stock market. In addition, in Moore network structure, the distribution of stock market's return called thin tail doesn't conform to fat tail of real stock market.In the paper, we analyzed the effect of mean and variance of investors'herding probability on stock price, return of investors and stock market, and found that suitable herd behavior would bring about maximal stock price, earning rate of invertors and market, and ratio of winner to loser. However, stronger or weaker herd behavior will lead them to minimum. In addition, we analyzed the character of stock market in different network structure. It is found that the larger the mean degree of network is, the smaller the stock price, the earning rate of investors and the ratio of winner to loser will be.