Abstract

We study the design of income-contingent subsidies for residential solar panels. We develop sufficient statistics for evaluating the cost-effectiveness of means-tested subsidies and estimate them using remotely sensed data on solar panel installations across the US and a border-discontinuity design. Our estimates reveal that the responsiveness of installation rates to subsidies is strongly decreasing in income. Using these empirical estimates, we estimate a structural model that embeds a solar adoption decision into a dynamic consumption/savings framework with bor- rowing constraints. Counterfactuals reveal that switching to production-maximizing income-contingent solar subsidies leads to a nearly three-fold increase in public funds received by low-income households and a 2.4% increase in national solar production. Means-tested subsidies are justified on equity and efficiency grounds.