Giving According to Garp : An Experimental Test of the Consistency of Preferences for Altruism

Experimental subjects often do not appear to behave as selfish money-maximizers, especially when "fair" or "altruistic" motives are inconsistent with money-maximizing Nash equilibria. This paper asks whether this apparently unselfish behavior is consistent with some well-behaved preference ordering other than money-maximization. We do this by checking whether choices of subjects satisfy the axioms of revealed preferences, such as GARP. Further, we estimate utility functions that could have generated the data and use these to explore results from outside our experiment. We find that a rational neoclassical approach to altruism works well and provides a foundation for a preference-based approach to altruism and fairness