Bounded Rationality, Noise Trading And Excess Volatility

Excess volatility is a famous anomaly in mature markets. For lack of data and for sake of difficulties in empirical implementation, little examinations on such anomaly have yet been done in China. In order to solve this problem, this dissertation uses the Agent-based Computational Finance (ACF) approach to model the phenomenon of excess volatility in Chinese stock markets, and to find the"micro"cause of such anomaly.In this dissertation, an ACF model, in which there are both noise traders and bounded rational investors in the same market, is constructed to investigate the phenomenon of excess volatility on risky assets price. My model implies that the phenomenon of excess volatility exists in such markets as there are both noise traders and bounded rational investors, that the phenomenon of excess volatility is the aggregate characteristics of noise trading and learning behaviors of bounded rational investors, that the learning frequency and the number of noise traders in the market are important factors on the degree of excess volatility. Lower learning frequency and more noise traders will result in more excess volatility.Based on the above model, this dissertation investigates the relationship between excess volatility and the existence of price limit institute. The ACF experiments in this dissertation get the following results. Price limit may have a little effect on excess volatility. While there are plenty of noise traders in stock markets, the existence of price limit institute will decrease the degree of excess volatility. While there are few noise traders, such institute will increase the degree of excess volatility. While there are considerable but not a large number of noise traders, the existence of price limit institute has little influence on the degree of excess volatility.