Testing the Analytical Framework of Other-Regarding Preferences

This paper aims to assess the empirical validity of the overall theoretical framework of other-regarding preferences. We focus on those preference axioms that are common to all the prominent theories of outcome-based other-regarding preferences. This common set of preference axioms leads to a testable implication: the strict preference ranking of self over a _nite number of alternatives lying on any straight line in the space of material payoffs to self and other will be single-peaked. We elicit the strict preference rankings of experimental subjects in variants of dictator and trust games using a mechanism that induces truthful revelation under quite weak assumptions. The data allow us to document the extent of single-peakedness and identify who violates single-peakedness. Potential reasons for violations of single-peakedness are delineated and the implications of our findings for theoretical modeling of other-regarding preferences are discussed.