Selecting Information in Financial Markets : Herding and Opinion Swings in a Heterogeneous Mimetic Rational Agent-Based Model

As expectations are driven by information, its selection is central in explaining common knowledge building and unraveling in financial markets. This paper addresses this information selection problem by proposing imitation as a key mechanism to explain opinion dynamics. Behavioral and cognitive approaches are combined to design mimetic rational agents able to infer and imitate each other's choices and strategies in opinion making process. Model simulations tend to reproduce stylized facts of financial markets such as opinion swings, innovation diffusion, social differentiation and existence of positive feedback loops. The influence of imitation reliability and information precision on opinion dynamics is discussed. The results shed light on two competing aspects of imitation behavior: building collective consensus and favoring innovation diffusion. The role of contrarian and individualistic attitudes in triggering large-scale changes is highlighted. From the results, some policy recommendations to reach better financial markets stability through opinion dynamics management are finally presented