Some causes are more equal than others? The effect of similarity on substitution in charitable giving

Donation matching and other directed interventions to encourage prosocial contributions may affect contributions through other channels. In an experimental dictator game where subjects may donate to two different real-world charities, we simulate activity-specific interventions by varying the relative productivity of those charities, and introduce several treatments to test whether (i) subjects substitute across charities, and (ii) whether substitution occurs even across (possibly very) dissimilar alternatives. We find that significant substitution occurs in all cases, but that the effect is weaker the more dissimilar the charity alternatives. In our most dissimilar treatment, substitution is only half as large as when alternatives are very similar.