Is It Just Legitimacy of Endowments? An Experimental Analysis of Unilateral Giving

Deviations from standard game theoretical predictions have been repeatedly observed in basic Dictator Games. Different interpretations have been provided to these deviations. On the one hand, empirical (among others, Forsythe et al., 1994) and theoretical works (among others, Bolton and Ockenfels, 2000; Fehr and Schmidt, 1999) have adopted the explanation based on other-regarding concerns. On the other hand, potential weaknesses in standard design of the game have been stressed. Evidence collected shows that when controlling for reputation considerations (Hoffman et al., 1996) and for legitimacy of assets (Cherry et al., 2002) behavior observed in the experiments is very close to that predicted by standard game theory. Results from our experiment suggest that the relevance of these two factors in explaining observed behavior may be overestimated by previous contributions. Relevant deviations from selfish equilibrium are registered in a condition of full-anonymity when assets to be shared are earned by the dictators and, simultaneously, recipients are allowed to work without being rewarded for their effort.