Simulation of futures and spot markets by using an agent-based multi-market model

The agent-based multi-market model we propose simulates futures and spot markets. On the basis of trading strategies in real markets, four kinds of agents-arbitragers, hedgers, speculators, and noise traders are included in our model. Interactions of the two markets are generated through various agent trading behavior. We also statistically analyzed futures and spot prices of the Nikkei 225 index, where we found a large positive correlation between the two prices and a fat-tail distribution of the basis. Simulations results show that, instead of the conventional single-market model, only the two-market model reproduces both statistical properties of futures prices.