Social preferences: from the experimental lab to economic theory

We present a wide collection of experiments which show how human behavior deviates substantially with respect to the predictions derived from standard homo economicus assumptions. Then we review the theoretical literature that this evidence has stimulated. In particular some models are found to be consistent with evidence from a large set of games. As fundamental differences exist among these proposals, new experiments were devised to contrast their effectiveness in predicting behavior. We argue that inequality aversion models are to be preferred to intention based models because the additional predictive power the latter may have comes at a very high cost of complexity. We also find that equality considerations are more relevant than efficiency motives in most economically relevant settings. Results are not conclusive and this gives scope to further research over these issues.