Targeting Incentive Contracts in Heterogeneous Populations: Evidence from Walking Payments

• Development Workshop

Date and Location

Tuesday, March 10, 2020, 4:10 PM - 5:30 PM
ARE Library Conference Room, 4101 Social Sciences and Humanities

Abstract

Providing financial incentives to encourage behavioral change is increasingly common, and has been effective in health, environmental, and educational contexts. In designing such contracts, policymakers face a trade-off between designing contracts for high-behavior-cost individuals (who require large incentives or easy behavior targets) and low-behavior-cost individuals (for whom large incentives are unnecessary and easy behavior targets will induce minimal change). A promising strategy to improve effectiveness is targeting contracts based on individual behavior costs, or “types”. Using a randomized controlled trial offering incentives for walking in India, we explore two strategies for doing this. The first is analogous to third-degree price discrimination: we target contracts based on observable proxies of their effectiveness. The second is analogous to second-degree price discrimination: we allow individuals to choose between a menu of contracts where the favorite choice is correlated with its expected effectiveness. We show that both forms of targeting are highly effective in our context, increasing the treatment effect by about 50%. We then explore whether three behavioral mechanisms (time inconsistency, misinformation, and autonomy) influence the effectiveness of each strategy.