Abstract

We study asset pricing and risk sharing in experimental financial markets. Our experimental design is motivated by the theory of competitive equilibrium in complete market economies with perfectly rational agents. In our data, theoretical predictions are satisfied at the aggregate but not at the individual level. Deviations from theory are not explained by noncompetitive behavior but by mistakes, taking the form of first-order stochastically dominated actions. We propose and structurally estimate a random-choice model of bounded rationality that fits our data well. We finally quantify and estimate to which extent noise in individual behavior is detrimental to price and allocative efficiency.

Co-auteurs : Bruno Biais, Thomas Mariotti, Sophie Moinas and Sébastien Pouget

Practical information