Behavioral Spillovers across Prosocial Alternatives

Abstract: This thesis contributes to the economic literature on prosocial behavior. It includes three papers, all of which relate to the issue of policy-driven spillovers across prosocial alternatives. For example, the common fundraising practice of donation matching could affect contributions through other channels; similarly, recycling policy could affect household efforts on other environmental activities. Given that the performance of one activity has been argued tocrowd out ("moral licensing") as well as in ("moral consistency") other activities, such spillovers could plausibly have either sign. The first paper develops a model where agents may contribute to a single public good through several different activities. For a large set of plausible cases, we predict that policy facilitating one activity reduces effort on other activities, though overall public-good production still increases. Furthermore, the multi-activity framework admits new interpretations that are sometimes at odds with prominent results from single-activity variants of our model. In the second paper, we run an experimental dictator game where subjects may donate to two different real-world charities. To simulate activity-specific interventions, we vary the relative productivity of those charities, and introduce several treatments to test whether spillovers occur even across (possibly very) dissimilar alternatives. We find that negative spillovers occur significantly in all cases, but that the effect is weaker, the more dissimilar are the charity alternatives. The third paper estimates policy spillovers within the context of a natural experiment on food waste in Sweden. We use a difference-in-difference design to measure the causal impact of introducing food-waste collection on the sorting of packaging waste. Results suggest a positive spillover effect corresponding to 5-10% of the population average. Point estimates are relativelysmall, and sometimes insignificant, when we control for shifts in the waste-related incentives facing households. Although unable to control for all such incentive shifts, we argue that the remaining bias may be negligible.