Chun Xia
(Faculty of Business and Economics, The University of Hong Kong)

 

Abstract: In a multi-period trading and learning model, agents initially do not know their abilities which are related to the qualities of private signals. They assess abilities from communicating and comparing quality of their own signals with that of others after each round of trading. Motivated by recent experimental findings that overand underconfidence in ability are often observed, agents are assumed to credit (blame) themselves strongly for favorable (unfavorable) outcomes. We demonstrate that under reasonable conditions excessive price volatility can be associated with under confidence. The relationship between expected volumes and overconfidence is non-monotone, so are agents’ expected profits. We suggest that information communication among agents can alternatively account for a wide range of empirical findings.

 

Key words: social communication; attribution bias; overconfidence; underconfidence; asset supply uncertainty

JEL codes: D80, G12, G14