Social Preferences

Social preferences are the systematic ways in which human agents care about the payoffs, intentions, and outcomes of others — not merely as instrumental means to their own ends, but as direct components of their utility. They are the formal representation of what lay language calls fairness, reciprocity, altruism, and spite, imported into game theory to explain the robust experimental deviations from purely self-interested behavior.

The canonical models are those of Fehr and Schmidt (1999) and Bolton and Ockenfels (2000). Fehr and Schmidt propose that agents experience disutility from inequity — both advantageous (I have more than you) and disadvantageous (you have more than me) — with the latter typically weighted more heavily. Bolton and Ockenfels propose a simpler model in which agents care only about their own relative share of the total payoff. Both models predict the rejection of low offers in the ultimatum game, positive giving in the dictator game, and conditional cooperation in the public goods game.

But social preferences are not stable individual traits. They vary dramatically across cultures, social contexts, and experimental framing. The same agent who rejects a low offer in the ultimatum game may accept a worse deal in a market context, not because their preferences changed but because the norms activated by the market frame differ from those activated by the fairness frame. This context-dependence suggests that social preferences are better understood as situated norms — culturally transmitted rules that agents apply when a situation is coded as a fairness-relevant domain — rather than as parameters of a fixed utility function.