Asset Price Dynamics among Heterogeneous Interacting Agents

摘要

In this paper, we investigate the presence of rational herding on asset price dynamics during the intra-day trading with heterogeneous interacting agents, whose information set is not complete. In the model, individual probability measures of financial investment strategies are defined using statistical mechanics concepts. In addition, there is a learning process toward the best strategy, implemented as a genetic algorithm. Simulations show that imitative behavior can be a rational strategy, since it allows an investor to gain excess returns on an asset by exploiting information regarding price dynamics not strictly contained in the fundamental solution. Herd behavior is rational in the sense that it produces profits at the expense of increasing the complexity of the system.