Families, Markets, and Self-Enforcing Reciprocity Norms

In the last 20 years, economists have become steadily more interested in the logic of norms of cooperation. In part, this increased interest was stimulated by a long series of experiments -- beginning with experiments on the Prisoner's Dilemma and the voluntary provision of public goods, and more recently experiments on the ultimatum and dictator games -- which strongly indicate that individuals are guided by such norms rather than the cold selfishness that economists usually assume. The response of some economists to the experimental results has been to assume that agents have preferences for fairness and reciprocity, which lead them to behave as they do in these experiments. But this solution is too easy almost any behavior can be explained by assuming preferences that make that behavior rational. The challenge facing economists is to derive these preferences from more primitive assumptions.

Two types exist in the models to be discussed. One type, the opportunist, is the classical homo economicus, who maximizes his or her expected material success, without regard for norms of reciprocity or fairness. Thus, in the Prisoner's Dilemma game, the opportunist prefers to defect against a cooperating opponent, rather than be in the joint-cooperate outcome. The second type, the reciprocator, prefers to be in the joint-cooperate outcome, in which the joint profits of the two players are maximized. The models to be presented analyze the conditions under which the reciprocator can be expected to survive in competition with the opportunist, when both types play rationally. That is, unlike the evolutionary tradition stemming from Robert Axelrod and others, agents in these models do not play pre-programmed strategies, but rather maximize their expected payoffs, given their preferences. The role of the evolutionary process is to select the agent type whose preferences lead to greater material success. The very interesting result is that, under fairly mild assumptions, the reciprocator type indeed survives, and there is an evolutionarily stable population mixture of the two types. Empirical evidence bearing on the predictions of the models will also be discussed.