This thesis comprises three studies on the behavioral properties of the stock market by constructing an artificial stock market using the agent-based modeling.
The first study investigates the influences on the price volatility when uninformed traders increase in the market due to the decrease of trading fees. Nowadays, as trading fees decrease, the number of participants of the uninformed traders and also the trade frequency of those participants have increased. This can have influence on the price volatility in the view of an increase in flexibility and heterogeneity. While volatility increases due to the increase in heterogeneity of the total participants, volatility decreases as flexibility increases with the number of participants and trade frequency. So we can see price volatility decrease as uninformed traders increase in the market.
The second study is an inquiry into the relationship between time-horizon of strategy and market, which is conditioned by the trend and volatility of the market price. In markets with the strongly increasing trend of price, long-term investment strategy results in a higher cumulative return than short-term strategy during the same time period. And short-term investment strategy can work better in markets with weakly increasing trend. We can also see that while volatility gives the incentive to use short-term strategy, it is negatively correlated with overall market investment returns.
The final study investigates whether the forecasting behavior of traders and other factors can have influence on the time-series property of the market equilibrium price. In this model, trader forecasts the price of the next period based on the price of the current one. But the time-series of the market equilibrium price does not have an exact autoregressive process of order 1(AR (1)). We also find that the number of traders, variance of expected prices and market growth have influences on the behavioral characteristics of the market price.