Cursed Equilibria and Knightian Uncertainty in a Trading Game

TL;DR

This paper introduces a new equilibrium concept combining Knightian uncertainty with cursed equilibrium, explaining why trade occurs despite traditional models predicting no trade, aligning with experimental observations.

Contribution

It develops a novel equilibrium framework that integrates Knightian uncertainty into cursed equilibrium, providing insights into trade behavior under uncertainty.

Findings

Trade occurs in the new equilibrium despite Bayesian predictions

Uncertainty increases trade among cursed and uncertainty-averse players

Results align with experimental observations of trade behavior

Abstract

We introduce a novel equilibrium concept that incorporates Knightian uncertainty into the cursed equilibrium (Eyster and Rabin, 2005). This concept is then applied to a two-player game in which agents can engage in trade or refuse to do so. While the Bayesian Nash equilibrium predicts that trade never happens, players do trade in a cursed equilibrium. The inclusion of uncertainty enhances this effect for cursed and uncertainty averse players. This contrasts general findings that uncertainty reduces trade but is consistent with behavior that has been observed in experiments.