Truth in giving

It is often difficult for donors to predict the value of charitable giving because they know little about the persons who receive their help. This concern is particularly acute when making contributions to organizations that serve heterogeneous populations. While we have considerable evidence that donors are more generous if they know their assistance benefits a preferred group, we know little about the demand for such information. To start closing this gap, we study transfers of income to real-world poor people in the context of dictator games. Our dictators can purchase signals about why the recipients are poor. We find that a third of the dictators are willing to pay a dollar to learn more about their recipient. Dictators who devote resources to acquiring information are individuals whose giving is particularly responsive to recipient type. They use the information mainly to withhold resources from "undeserving" types, leading to a drastic decline in aggregate transfers. With endogenous information about recipients, we find that all types of poor subjects are worse off. Our results suggest that the effects of truth-in-giving policies are highly responsive to recipient heterogeneity and biased against more generous giving.