ترجیحات اجتماعی و تعصبات اخلاقی

Abstract An emerging consensus in economics is that three motives are at work in strategic decisions: distributive preferences, reciprocal preferences and self-interest. An important obstacle, however, has been moral biases: distortions created by self-interest can obscure our measures of social preferences. This paper describes a simple experiment to address this. We compare the decisions of implicated “stakeholders” with those of impartial “spectators.” We find that stakeholders are less inclined to respond to the generosity of others than are spectators. We also clarify a result in previous research [e.g., Offerman, T., 2002. Hurting hurts more than helping helps. European Economic Review 46, 1423–1437] that stakeholders punish unkindness more than they reward kindness. We find that this asymmetry in reciprocity has two sources: an asymmetry in the underlying preference that even impartial spectators display and a moral bias; stakeholders punish more and reward less than spectators. In sum, we find that all three motives have important and significant effects on final allocations.

Introduction The assumption of self-interest has served as a powerful axiom of economics. It has helped to explain and predict large swaths of observed facts and to develop rigorous and elegant theoretical models. With mounting evidence of behavior at variance with material self-interest, however, economists are increasingly enriching traditional models with additional motives. The growing consensus is that integrating social preferences into the analysis can explain important economic phenomena, including involuntary unemployment (e.g., Akerlof and Yellen, 1990), pricing policies (e.g., Kahneman et al., 1986b and Kachelmeier et al., 1991), and bargaining behavior (e.g., Güth et al., 1982). Nevertheless, the confluence of self-interest and social preferences has proven an important hurdle to inferring the underlying preferences that theory needs to incorporate. Attempts to identify social preferences have been hindered by what we call “moral biases,” that is, the distortionary effects of self-interest on expressed social preferences. For example, consider the dictator game, an experiment in which one subject (the dictator) may transfer money to an anonymous recipient. A purely self-interested dictator would send nothing. On average, though, dictators send approximately 30% of their endowment to recipients, and this transfer typically is taken as a measure of social preference. But suppose that an impartial spectator motivated purely by social preferences would implement an equal split between the parties. One might consider this as the measure of true social preferences. The observed 30% transfer is neither self-interest, which implies no transfer, nor social preferences, which prescribe a 50% transfer. The difference between the 30% transfer and the 50% transfer is an example of what we call a moral bias; it represents a distortion in our measures of social preferences due to the impact of self-interest on their expression.2 Although the existence of moral biases has been noted (e.g., Forsythe et al., 1994), almost no studies of social preferences have isolated social preferences from self-interest and its distorting effects. One exception is Charness and Rabin (2002), who report two sets of comparisons in binary choice dictator games focusing on preferences for surplus maximization. The current study, on the other hand, has no surplus creation and instead explores distributive and reciprocal preferences over a wide range of allocation conditions. We experimentally isolate the separate effects of self-interest, distributive and reciprocal preferences using a simple two-stage dictator game. We contrast the behavior of stakeholders, or implicated parties, with that of spectators, or third parties, the latter serving to establish a benchmark of pure social preferences. Our design permits us to identify and separate the different motives. The experiment produces evidence of moral bias that affects both the average willingness to act on social preferences as well as responsiveness to inequity, kindness and unkindness. In addition, both spectators and stakeholders exhibit an asymmetry in reciprocity, favoring punishment over reward. There is also a moral bias in this asymmetry: stakeholders are less inclined to reward and more inclined to punish than spectators. The paper is organized as follows. Section 2 presents background and hypotheses, Section 3 describes the experimental design and procedures, Section 4 presents and analyzes the results of the experiment, and Section 5 contains the concluding remarks.

5. Conclusion and discussion Evidence has been mounting from the laboratory and the field that social preferences are economically relevant and statistically important forces in a variety of settings. They are implicated in involuntary unemployment, strikes and lockouts, product pricing, contract negotiations, and other bargaining behavior. A number of competing motivations have been described and formalized. With this study we hope to help unify this stream of research. We define and examine moral biases, the impact of self-interest that distorts the expression of social preferences. We find that moral biases exist: stakeholder decisions differ significantly from those of spectators. Specifically, stakeholders exhibit a mean moral bias, allocating less, on average, to others than spectators, as well as a response moral bias, reacting less than spectators to differences in the allocations of others. These biases are found both with overall social preferences (Hypothesis 1) and with purely distributive preferences (Hypothesis 2). Both stakeholders and spectators reciprocate but do so asymmetrically: in absolute terms, they punish others for low allocations, but they do not significantly reward them for high ones (Hypotheses 3 and 4). Thus, we conclude that asymmetry in reciprocity is partially driven by an underlying asymmetry in social preferences. However, there is also a moral bias in reciprocity (Hypothesis 5): stakeholders punish low transfers more and reward high transfers less than spectators. These results tend to support Adam Smith's assertion about spectators rewarding kindness but stakeholders not: spectators reward generous X subjects significantly more than stakeholders. This obscuring effect of self-interest has impaired other attempts to infer social preferences (distributive and reciprocal) and to gauge the magnitude of their force relative to one another and to self-interest itself. Previous studies have come to conflicting conclusions about the importance of these preferences, sometimes suggesting that distributive preferences, reciprocity, or both have little or no effect. Our experiment identifies and separates self-interested, distributive and reciprocal preferences. The results reported here allow us to conclude with some confidence that all three forces exert considerable influence on both the level and responsiveness of individual allocation decisions. In addition, this study helps to clarify the relationships between self-interest, distributive preferences and reciprocity and to quantify them unobscured by moral biases. One surprising result is that stakeholders are less reciprocal, on average, than spectators. This runs counter to the reasonable expectation that stakeholders might respond more strongly to kindness and unkindness directed toward them than would an unimplicated third party on their behalf. Nevertheless, it is consistent with Smith's conjecture regarding spectators and stakeholders. A related finding concerns the important asymmetry between positive and negative reciprocity that has been seen in previous studies. We find that spectators, like stakeholders, punish more than they reward. However, spectators punish significantly less and reward significantly more than stakeholders, suggesting that the application of reciprocity observed in previous studies included moral biases. The contrast between spectator and stakeholder reciprocity is potentially helpful in building descriptive models, but it is also a difference that might be important to consider for prescriptive analysis. Social choice models built on the abstract impartial spectator, for example, should consider that the positive reciprocity motive appears to be stronger among spectators than among stakeholders. We believe that investigation into spectator preferences is an important direction for further research. Such investigations have at least three advantages. First, they allow us to understand and estimate the sometimes intricate social preferences that self-interest might otherwise obscure. This holds the potential for informing theoretical work on how social preferences came to be and how they are instantiated. We think this method can help us determine what behavior is deemed fair or right. Second, by isolating motives apart from self-interest and comparing them with the behavior of stakeholders, this approach enables us to identify the effects of self-interest, the central construct in economics, more precisely. Finally, impartial spectatorship has a long tradition in social choice and moral philosophy, and empirical analysis in this vein might inform normative theory and economic policy.