Beliefs and Social Preferences

Why do people sacrifice to help others in some situations, but not in others? Besides a direct taste for helping others, I study three additional psychological motivations that involve beliefs: social-signaling, which holds that a person wants others to think of her as fair-minded; self-signaling, which says she would like to think of herself that way; and beliefs-based altruism, which says that she wants others to feel good from knowing that their payoffs come from wellintentioned actions. After developing a formal model of such motivations and differentiating between them conceptually, I experimentally study a dictator game in which the outcome may be determined by the dictator or by chance, and one in which the recipient is uncertain of the options available to the dictator. To distinguish among the three theories I vary the likelihood that the dictator’s choice will count and the recipient’s information about her choice, the conditions under which it was made, and her identity. I find little evidence of self-signaling or beliefs-based altruism, but stronger evidence for social-signaling. Lowering the expected cost of the decision to give by decreasing the likelihood that it will be implemented has little effect on dictator giving when the beliefs of the recipient are held constant. Increasing the recipient’s ability to discern whether his payoff is due to the dictator’s action—either by providing information about the likelihood that the dictator’s choice was implemented or by reducing anonymity—increases giving, which shows that subjects are concerned with others’ beliefs. Finally, subjects show little willingness to pay to provide information to a dictator game recipient about whether or not a dictator sacrificed to help him or withhold such information, even when doing so would allow the recipient to maintain positive beliefs about how he was treated. This suggests that the act of giving is largely tied to what it says about the giver as opposed to how it makes the recipient feel. ∗UC Berkeley. Email: [email protected]. I would like to thank Brenda Naputi, Lawrence Sweet and the UC Berkeley Xlab for invaluable support while carrying out the experiment. Matthew Rabin, Botond Kőszegi, Shachar Kariv, Ulrike Malmendier, and Stefano DellaVigna provided helpful comments and criticism. Participants at the UC Berkeley Psychology & Economics Seminar and non-lunch provided helpful feedback. This research was supported financially by the UC Berkeley Experimental Social Sciences Laboratory (Xlab), the UC Berkeley Institute for Business and Economic Relations (IBER) and the UC Berkeley Program in Psychology and Economics (PIPE).