DIRECT TESTS OF INDIVIDUAL PREFERENCES FOR EFFICIENCY AND EQUITY

Abstract

I. INTRODUCTION Economics has a long history of using models of preferences to explain choices. Until recently, the preferences most commonly used have been self-regarding (or economic man) preferences in which an agent cares about his own material payoffs but is indifferent about the material payoffs of others. There is now a large literature that supports the conclusion that self-regarding preferences models are mostly inconsistent with observed behavior in experiments in which fairness is a salient characteristic of the decision tasks. In response, various models of social preferences have been developed and applied to data from experiments. Two prominent types of models of social preferences are inequality (or inequity) aversion models (Bolton and Ockenfels 2000; Fehr and Schmidt 1999) and the quasi-maximin model (Charness and Rabin 2002). These models have been widely applied to data in the past and continue to be applied in current literature (Chen and Li, 2009; Fehr, Klein, and Schmidt, 2007). In the present paper, we focus on testing the distinguishing characteristics of the models rather than fitting their parameters to data. The distinguishing characteristic of inequality aversion models is that utility is decreasing with the absolute (value of the) difference between one's own and others' material payoffs as well as increasing with one's own payoff. The distinguishing characteristic of the quasi-maximin model is that utility is increasing with the lowest of all agents' payoffs (the maximin property) and the total of all agents' payoffs (the efficiency property) as well as increasing with one's own payoff. Inequality aversion, efficiency, and maximin have been described as for behavior, and controversy has developed about the relative importance of these motives for explaining behavior (Bolton and Ockenfels 2006; Engelmann and Strobel 2004, 2006; Fehr, Naef, and Schmidt 2006). We here apply a somewhat different approach with an experiment that includes three treatments. Each treatment implements an experiment design that identifies and tests an observable consequence of a single one of these possible properties of distributional preferences; we refer to this type of test as a test. Treatment 1 implements a direct test for inequality aversion: a subject whose preferences include an aversion to inequality in payoffs favoring another person will make one specific choice in this treatment while other feasible choices are inconsistent with inequality aversion. Treatment 2 implements a direct test in which one feasible choice is consistent with preferences that are monotonically increasing in the total of all agents' payoffs while other feasible choices are inconsistent with such preference for efficiency. Treatment 3 implements a similar direct test for preferences that include the maximin property. Choices made by large majorities of subjects in the three direct-test experiment treatments reported herein are inconsistent with preferences characterized by inequality aversion, efficiency, or maximin. This might seem surprising, given the many applications of the inequality aversion and quasi-maximin models that seem to show that the models fit data from various experiments pretty well. But in many experiments the implications of inequality aversion, efficiency, and maximin are confounded with the implications of the conventional convexity property of preferences. We will explain that most data from our three experiment treatments that are inconsistent with the inequality aversion and quasi-maximin models are, by contrast, consistent with preference models that include conventional properties such as convexity and positive monotonicity for all agents' payoffs. One such model is the egocentric altruism model (Cox and Sadiraj 2007). (1) II. TREATMENT 1: A DIRECT TEST FOR INEQUALITY AVERSION Utility functions for inequality aversion models are increasing with an agent's own material payoff but decreasing with the absolute (value of the) difference between her/his own payoff and others' material payoffs. …