Regret Aversion and Informational Cascade in a Sequential Trading Model

抄録

This paper examines the effect of regret aversion on trading strategy and information aggregation in a dealer market with asymmetric information. Although the market maker adjusts asset price according to all available information through Bayesian learning, informational cascades can still occur: informed traders herd to buy when the price is extremely high, while they herd to sell when the price is extremely low. This model overcomes the so-called “price critique” in a simple framework. It also implies that the interaction between investor psychology and rational herding may be an important factor causing bubbles and crashes in security markets.