A New Agent-based Stock Market Model:Consider The Asset Earning

This paper presents a new agent-based financial market model. Compared with the existing literature, it considers the three properties of assets——the price, the dividend and the earnings. This paper introduces the indicators of price earning ratio (PE ratio), and analyzes the impact on stock return of earnings in the framework through the PE ratio. Return on asset in the model show similar features as real market, such as volatility cluster and fat tail. Leptokurtosis and fat-tail is one of the major features of asset return. The main statistics of data fit well. The mean and standard deviation of monthly return are0.0054and0.0431, while the real values are0.0056and0.0458. And the kurtosis is15.3808, compared with the20.1251, which is significantly larger than the value of normal distribution. The first order autocorrelation is positive as the real market.This paper models these four types of the main investors in real market. The first of these is an adaptive investor, the second a fundamental one who forecasts the future returns with PE ratio, the third a "short AR" investor, and the last one holds the "buy and hold" strategy. It yields some insights into the dynamics of agents’ strategies. And it also show us some interesting results as the real market does.