Abstract:
Time Series Properties of an Artificial Stock Market

B. LeBaron, W.B. Arthur, and R.G. Palmer

This paper presents results from an experimental computer simulated stock market. In this market artificial intelligence algorithms take on the role of traders. They make predictions about the future, and buy and sell stock as indicated by their expectations of future risk and return. Prices are set endogenously to clear the market. Time series from this market are analyzed from the standpoint of some well known empirical features in real markets. The simulated market is able to replicate several of these phenomenon, including fundamental and technical predictability, volatility persistence, and leptokurtosis. Moreover, agent behavior is shown to be consistent with these features in that they condition on the variables that are found to be significant in the time series tests. Inside this experimental model there exists a well-defined linear homogeneous rational expectations equilibrium. This is used as a benchmark in the experiments to assess the overall ability of the agents in learning. It is found that for certain parameters the results in the market are consistent with this benchmark.

To appear: Journal of Economic Dynamics and Control 23, 1487-1516 (1999)

Last Updated: 15-Nov-99