Incentives and prosocial behavior
Roland Bénabou and Jean Tirole present a theoretical model to explain prosocial behavior that goes over and above what previous models had suggested. Their model suggests that motivation to participate in either discrete prosocial behavior (e.g., giving blood) or continuous (giving more or less money to charity) can be seen as an endogenous and unobservable mix of three types of motivation: intrinsic, extrinsic, and reputational. The model presented is based on a signalling model and discusses the roll of social signalling (e.g., letting every else know how generous you are) and what they call "self-signalling" (e.g., feeling that you are a good person or what other studies might have called "warm glow" altruistic behavior). They describe self-signally as being a concept behind self assessment involving questions like, "what kind of person am I?" Their goal seems to be to theoretically break apart self-esteem from social-esteem and to also allow for a pure "joy of giving."
The authors present their model, discuss how their model would be able to explain the phenomena of crowding out, describe the role that social norms plays, examine the possibility that agents could turn down rewards (and how and why that might happen and what its effect would be) and examines the way that private or public sponsors might set incentives and what the welfare effects of competition may be.
Essentially, the authors model posits these three distinct motivation motivations but states that it can be difficult for an audience to understand whether a person is motivated by one or another. In the absence of a reward, we know people will only be motivated to do a costly action if they are internally motivated. That said, the introduction of reward makes understanding this mechanism more difficult and essentially adds noise to the signal.
They show that by introducing an incentive, you may bring new people "in" who were not motivate enough intrinsically if the incentive is big enough to push them over the edge into action. However, it will also deter people who are receiving less intrinsic value. The basic ideas is that rewards amplify the noise associated with determining why people are doing things. Using this basic approach, the authors show show that crowding out can happen over a very large range and that only very large incentives are superior to none at all.
The authors' model fits the results including Mellstrom and Johannesson's Crowding out in blood donation: Was Titmuss right? that suggests that those that are, in the absence of rewards, the most motivated to contribute may be the ones who are also most susceptible to crowding out. They model also matches the result that suggests that this group will also be the group that tend to turn away the rewards most.
When reputation bears on prosocial orientation, a person will attempt to become as altruistic appearing as possible.