The Experimetrics of Public Goods: Inferring Motivations from Contributions

In public goods experiments, stochastic choice, censoring, and motivational heterogeneity allow experimentalists to differ over the extent of unselfishness, and whether it is reciprocal or altruistic. These problems are addressed econometrically by estimating a finite mixture model to isolate types, incorporating the two-limit Tobit model with tremble to accommodate censoring and errors. Most subjects act selfishly, but a substantial proportion are reciprocal with altruism playing only a marginal role. Isolating reciprocators enables a test of Sugden’s model of voluntary contributions, which is rejected because they display a selfserving bias.