Individual behavior in isolation is known to deviate systematically from predictions of the Bayes' Rule when they are intuitively (i.e., without formal computation). Economic settings, rich in incentives, information and learning, may induce Bayesian behavior. Further, it is possible that the aggregate behavior in markets can be predicted accurately by simply assuming that individuals act as Bayesians. Data from laboratory markets suggest that, compared to specific alternative assumptions about individual information processing, Bayes' Rule is a better predictor of market price but not of allocation and efficiency. Improvement in its performance with trader experience suggests that, as a descriptive theoory of aggregate market behavior, Bayes' Rule may have only asymptotic validity.
Experiments, Bayesian Behaviour, Oral Double Auction
Date of this Version