Mark Bedau, Shareen Joshi, Jeffrey Parker
Paper #: 99-03-023
We use game theory and the Santa Fe Artificial Stock Market, an agent-based model of an evolving stock market, to study the properties of strategic Nash equilibria in financial markets. We discover two things: there is a unique strategic equilibrium in the market, and this equilibrium in suboptimal since traders' earnings are not maximized and the market is inefficient. The inevitability of this strategic equilibrium is due to an analogue of the prisoner's dilemma; the optimal global state is unstable because each individual has too much incentive to "defect" and use forecasting rules that pull the market into the suboptimal equilibrium.