Loading field map
Social preferences are the systematic ways in which people care about the outcomes of others, not just their own. The term comes from behavioral economics, where it is used to explain choices that deviate from the standard economic assumption of pure self-interest. When a person gives up money to punish a cheater, shares a windfall with a stranger, or refuses an unfair offer even at a cost to themselves, their behavior is often attributed to social preferences. The field studies these behaviors empirically, builds formal models to capture them, and asks how they shape markets, organizations, and public policy.
The core puzzle of social preferences is why people so often act as if they value the payoffs of others, and how exactly they value them. Standard economic theory treats the individual as a self-interested maximizer of their own material payoff. Social preferences research asks what happens when that assumption fails. It does not ask whether people are "good" or "bad," but rather what specific patterns of other-regarding concern exist, how strong they are, and under what conditions they appear.
A second central question is measurement. How can a researcher tell whether an observed action reflects a preference for fairness, a desire to reciprocate kindness, a taste for punishing norm violators, or simply a miscalculation? The field has developed experimental designs intended to separate these motives, but the task is difficult because the same observable action can be produced by different underlying preferences.
A third question concerns aggregation. If individuals have social preferences, what happens when they interact in markets, firms, or political institutions? Do social preferences get amplified, neutralized, or exploited by competition and institutional design? This question connects the subfield to the broader discipline of economics, because it asks whether the standard predictions of market models survive when actors are not purely self-interested.
The modern study of social preferences emerged in the 1980s and 1990s, but it drew on older traditions. Experimental economics had already developed the tools of controlled laboratory games, and social psychologists had long studied concepts like fairness and reciprocity. What was new was the systematic use of economic experiments to measure other-regarding behavior and the development of formal models to explain the results.
The immediate precursor was the rise of experimental economics itself. Researchers such as Vernon Smith and Charles Plott had established laboratory methods for testing market behavior, and the ultimatum game became a standard tool. In the ultimatum game, one player proposes a division of a fixed sum of money, and the second player can accept or reject. If the second player rejects, both get nothing. Standard theory predicts the proposer offers the smallest positive amount and the responder accepts it, since something is better than nothing. In practice, proposers typically offer substantial shares, and responders frequently reject low offers even though rejection costs them money. This result, replicated across many countries and cultures, became the empirical foundation of the field.
The first formal models of social preferences appeared in the late 1980s and early 1990s. These models added other people's payoffs to the individual's utility function, but they did so in different ways. Some models assumed people care about the distribution of final outcomes, while others assumed people care about the intentions behind actions. The distinction matters because it generates different predictions about when other-regarding behavior will appear.
The field is organized around several distinct modeling traditions, each of which addresses a different aspect of the empirical evidence. These approaches are not mutually exclusive, and many researchers combine elements of several.
The simplest approach assumes that people care about the distribution of material payoffs, not about how those payoffs came about. The most influential version is the inequality-aversion model developed by Ernst Fehr and Klaus Schmidt. In this model, a person's utility increases with their own payoff but decreases when their payoff is below or above the average of others. The person is assumed to dislike being worse off than others (envy) and, to a lesser degree, dislike being better off (guilt). The model has two parameters, one for each type of aversion, and it can explain a wide range of experimental results, including the ultimatum game and public-goods games.
A second outcome-based approach is the theory of "social welfare preferences" developed by Gary Charness and Matthew Rabin. This model assumes people care about the total payoff of the group and about the minimum payoff, in addition to their own. It captures a concern for efficiency and for helping the worst-off, rather than a simple dislike of inequality. The two models make different predictions in some situations, and experiments have been designed to distinguish them.
The strength of outcome-based models is their simplicity and tractability. They can be incorporated into standard economic analysis relatively easily, and they make clear predictions. Their weakness is that they ignore the process by which outcomes are produced. A person who rejects an unfair offer in the ultimatum game might be responding to the outcome, the intention, or both, and outcome-based models cannot distinguish these motives.
A second tradition emphasizes reciprocity and intentions. The key idea is that people respond not just to the final distribution but to whether others acted kindly or unkindly toward them. The most influential model is Rabin's theory of fairness, which formalizes the idea that people are willing to sacrifice their own material payoff to reward kind actions and punish unkind ones. In this model, a person's utility depends on their own payoff, the other person's payoff, and a term that captures whether the other person's action was perceived as fair or unfair.
Intention-based models explain behaviors that outcome-based models cannot. For example, in the "moonlighting game," a player can either help or hurt another player, and the other player can respond. People often reward helpful actions even when the final distribution is unequal, and they punish hurtful actions even when the punishment is costly. This suggests that people care about the intentions behind actions, not just the outcomes.
The weakness of intention-based models is their complexity. They require assumptions about how people form beliefs about others' intentions, and they are harder to apply in market settings where intentions are less clear. They also face the difficulty that the same action can be interpreted as kind or unkind depending on context, which makes prediction harder.
A third approach emphasizes that people are not all the same. Some individuals are strongly altruistic, some are strongly spiteful, and many are conditional cooperators who cooperate if others do but defect if others defect. This heterogeneity is often modeled by assuming a population contains several "types" with different preferences. Researchers estimate the distribution of types from experimental data, and they use these estimates to predict behavior in new situations.
This approach is less a single model than a research strategy. It recognizes that the average behavior in an experiment may not represent any actual person, and it allows for the possibility that different people respond differently to the same incentives. It has become increasingly important as experiments have shown that behavior is highly variable across individuals and that this variability is stable over time.
A fourth approach, more sociological in flavor, emphasizes social norms. Rather than assuming people have stable preferences for fairness or reciprocity, this approach argues that behavior is guided by shared expectations about what is appropriate. People follow norms because they fear social sanctions or because they have internalized the norms as their own.
This approach is harder to formalize than the others, but it has gained traction as cross-cultural experiments have shown large differences in behavior across societies. The ultimatum game, for example, produces very different offers and rejection rates in different cultures, and these differences correlate with the degree of market integration and cooperation in everyday life. This suggests that social preferences are not a fixed human trait but are shaped by the social and economic environment.
The different approaches are not rival paradigms in the sense of mutually exclusive worldviews. They are complementary tools that address different aspects of the same phenomenon. Outcome-based models are useful for predicting behavior in anonymous, one-shot interactions where intentions are hard to infer. Intention-based models are useful for understanding repeated interactions and relationships where history matters. Type-based approaches are useful for explaining individual differences and for making aggregate predictions. Norm-based approaches are useful for understanding cultural variation and institutional design.
The field has moved toward a pragmatic synthesis. Many researchers use a combination of outcome-based and intention-based elements, and they allow for heterogeneity across individuals. The debate is not about whether social preferences exist—the experimental evidence is overwhelming—but about how to model them parsimoniously and how to apply them to real-world settings.
Contemporary research on social preferences is characterized by several developments. One is the increasing use of large-scale experiments, including online platforms that allow researchers to study thousands of participants rather than dozens. These studies have confirmed the basic findings of the field and have also revealed that behavior is more variable and context-dependent than early models assumed.
A second development is the integration of social preferences into macroeconomics and public policy. Researchers now ask how social preferences affect tax compliance, charitable giving, labor contracts, and the provision of public goods. The answers matter for policy design: if people care about fairness, then policies that are perceived as unfair may be less effective or may provoke resistance, even if they are economically efficient.
A third development is the connection to other disciplines. Behavioral economists increasingly collaborate with psychologists, neuroscientists, and anthropologists. Neuroimaging studies have shown that punishing unfair behavior activates brain regions associated with reward, suggesting that the taste for punishment is genuine and not just a strategic display. Cross-cultural studies have shown that social preferences vary systematically with social structure, suggesting that they are partly learned.
A fourth development is the study of the limits of social preferences. Researchers have documented that other-regarding behavior is fragile: it can be crowded out by incentives, undermined by anonymity, and weakened by group size. This has led to a more nuanced view in which social preferences are real but conditional, and in which institutions play a crucial role in either supporting or eroding them.
The field remains active and contested. Some economists argue that social preferences are too context-dependent to be modeled as stable preferences, and that the field should focus on norms and institutions instead. Others argue that the formal models are too simple and that a richer account of human motivation is needed. What is not contested is the empirical fact that people systematically care about the outcomes of others, and that this fact matters for economic analysis. The study of social preferences has permanently expanded the scope of economics, from a discipline that assumed self-interest to one that treats other-regarding behavior as a central object of study.