SUBSIDIES FOR DOMESTIC TECHNOLOGY ADOPTION UNDER HETEROGENEOUS TREATMENT EFFECTS

Abstract

Governments and NGOs in developing countries offer subsidies to encourage the adoption of beneficial domestic technologies to generate positive externalities and improve access for poorer households. However, these subsidies may be mistargeted if the benefits come from the continued use of the technology, which is not guaranteed by its initial take-up. This dissertation examines the impact of a subsidy to promote the adoption of a rainwater harvesting (RWH) technology on the water restrictions residents of poor neighborhoods in Mexico City face. I explore this topic theoretically and empirically in three main chapters.

In the first chapter, I outline a simple economic model of technology adoption and treatment effects. The model shows how exogenous changes to the subsidy can identify the treatment effects for different types of households, characterized by their willingness to pay (WTP) for the technology. To overcome the challenge of rare exogenous variation in subsidy rates and unobservable WTP, I propose the use of contingent valuation (CV) methods. These methods can exogenously generate variation in hypothetical subsidies and provide insights into the distribution of WTP in the relevant sample. The model is then completed by incorporating the CV information for empirical analysis. This approach may be valuable when randomized interventions are unfeasible due to institutional or budget constraints.

In the second chapter, I empirically estimate the effects of the RWH Program in Mexico City on the time households spend obtaining water and the likelihood of postponing daily activities due to the lack of water. I employ the framework developed in the first chapter and local instrumental variable methods for the estimation. The data for this analysis was collected among all program participants in 2021 in partnership with the implementing agency. I find that the usage and causal effects of the RWH technology improve with the households' WTP. High-WTP households save 5 hours per week in water procurement time and reduce postponement of daily activities due to water scarcity by 25 percentage points. Conversely, low-WTP households are less likely to use the technology, yielding negligible benefits.

The empirical analysis has significant policy implications. In the third chapter, I simulate counterfactual policies and show that adjusting the subsidy structure could enhance the average benefits of the RWH Program. Specifically, introducing enrollment fees that are a fraction of the total cost of the technology could consistently improve the average impact on recipients. These fees do not seem to disproportionately affect poorer households or those facing more stringent water restrictions, suggesting a potential avenue for policy refinement.