عدالت توزیعی و متقابل: ما چه می توانیم از ناهمگونی ترجیحات اجتماعی یاد بگیریم؟

Abstract This study evaluates people’s concerns for distributive fairness (equality of outcomes and payoffs to those worse-off) and reciprocal fairness (receiving what one is due based on one’s past actions) using dictator, ultimatum, and trust games. In the dictator games we classify individuals’ preferences as self-interested, inequity averse, efficiency maximizing, or compassionate social surplus maximizing. We find that the different utility functions of the social preference types guide participants’ behavior in the ultimatum and trust games. The self-interested and efficiency maximizers make the lowest offers in the ultimatum game, are least likely to reject a low offer in the ultimatum game, and send back the least amount in the trust game. Consistent with the goal of maximizing the sum of payoffs however, efficiency maximizers and compassionate social surplus maximizers make the highest offers in the trust game, and in attempts to ensure equal outcomes, inequity averters make low offers. Because those classified as self-interested or efficiency maximizing do not exhibit concern for pure distributive fairness, we identify positive offers as first movers in both games as evidence of expectations of reciprocity by second movers and some choices as second movers as clear evidence of their own concerns for reciprocal fairness. Furthermore, we find strong evidence supporting general concern for reciprocal fairness in the trust games where the modal response of second movers in all preference groups is to equalize payments across participants.

1. Introduction Most people value fairness, but there is considerable disagreement on what exactly is meant by “fair”. In this paper, we consider two fairness principles, “distributive fairness”, the concern for how money or goods are distributed among individuals, and “reciprocal fairness”, the belief that one should receive what one is due based on one’s actions (i.e. positive actions should be rewarded and harmful actions punished). We begin by identifying participants’ concern for distributive fairness through a set of dictator allocation decisions. The dictator decisions categorize people’s preferences as self-interested, efficiency maximizing, compassionate social surplus maximizing, or inequity averse. Those who are inequity averse or compassionate social surplus maximizing reveal a concern for distributive fairness because they are willing to give up money to reduce inequality or to raise incomes of those who are worse-off, while the self-interested and efficiency maximizers do not. The subjects then participate as both first movers and second movers in ultimatum and trust games. Using utility functions consistent with the different preference types, we can predict behavior in ultimatum and trust games with and without additional attention to reciprocal fairness. Concerns for reciprocal fairness in both games are exhibited both in first-mover expectations that the responder may reward or punish depending on perceptions of the fairness of the offer and in second-mover responses. The results of the games allow us to achieve two goals. First, we show that people with different types of social preferences behave differently in strategic environments. While there is a body of literature showing that social value orientation affects economic behavior in strategic games, most of this work does not distinguish between different ways of being prosocial, in particular, it does not distinguish between inequity aversion and efficiency/social surplus maximization. We show that inequity averters, efficiency maximizers, and compassionate social surplus maximizers behave very differently and that their decisions are predictable given their social preference type. Second, we utilize our preference classifications to distinguish pure reciprocal fairness from distributive fairness for select groups of individuals depending on their preferences.

Conclusion This study provides empirical evidence on people’s concerns for two types of fairness: “distributive fairness”, which focuses on outcomes in terms of equality and the amount received by those worse-off, and “reciprocal fairness”, which refers to people receiving what they are due in response to their past actions. A set of dictator allocation decisions is used to identify concerns for distributive fairness and to categorize subjects’ social preferences. Compassionate social surplus maximizers and inequity averters (42% of the sample) are willing to give up money in order to improve distributive fairness, and those who are not classified (30%) make at least one decision that improves distributive fairness. Only 28% (self-interested and efficiency maximizers) do not give up money or reduce the surplus, respectively, to increase either equality or the payoff to the worst-off person. We use ultimatum and trust games to compare behavior of the different social preference types and to evaluate concerns for reciprocal fairness. Predictions about how these social preference types will act in reciprocal games receive strong support. The self-interested and efficiency maximizers make lower offers in the ultimatum game than the inequity averters and compassionate social surplus maximizers. In addition, efficiency maximizers and compassionate social surplus maximizers, who care about the size of total payoffs, make the highest offers in the trust game. Similarly, the self-interested and efficiency maximizers are less likely to reject low offers in the ultimatum game and send back less in the trust game than the inequity averters and compassionate social surplus maximizers. This consistency of behavior across games is likely due to the rigor with which we categorize social preferences in the ten question modified dictator game and because we distinguish different types of prosocial preferences. The varying distribution of payoffs in 10 allocation decisions of the dictator games requires the participant to display his or her social preferences repeatedly in a manner that differentiates inequity aversion from two types of social surplus maximization and from self-interest. A key finding of this paper is that the two types of prosocial preferences, inequity aversion and social surplus maximization, cause people to behave very differently. With two-thirds of the participants displaying consistent preferences across the ten allocation exercises, it is not surprising that these same preferences extend to the two games involving reciprocity that promise similar payoffs (and trade-offs) and are played within the same hour. We find strong evidence of concern for reciprocal fairness in that more than 90% of the self-interested and efficiency maximizers (who do not care about distributive justice) make positive offers in the ultimatum game, 77% of the self-interested offer a positive amount in the trust game, 54% of the self-interested and 57% of the efficiency maximizers would reject an offer of $1 in the ultimatum game, and 84% of the self-interested and 94% of the efficiency maximizers send back a positive amount of money in the trust game. Further, there is evidence of effort to promote equality in payoffs in the trust game by those who have shown no interest in pure distributive fairness. In response to trust offers where equality can be achieved, the modal response is to equalize payoffs, with 46% of self-interested and 47% of efficiency maximizers giving back the amount necessary to equalize payoffs or more. While it is difficult to separate out concerns for reciprocal fairness from concerns for distributive fairness for inequity averters and compassionate social surplus maximizers, we do find that as responders in the trust game, 20% of the inequity averters increase partner’s payoff to an amount exceeding own payoff, clearly rewarding generous or trusting behavior on the part of the first mover. Together these results provide strong evidence that all social preference types make decisions motivated by considerations of reciprocal fairness.