Impact Evaluation of Rural Electrification

The objective of this study is to measure the impact of infrastructure on economic outcomes in Ethiopia, Tanzania, Mozambique, and El Salvador. In Tanzania, the study focuses on the impact of the upgrading of sanitary infrastructures in the poorest neighborhoods of Dar Es Salam. In Ethiopia, the project measures the impact of rural households’ access to electricity as well as optimal levels of subsidies to ensure poorest households’ capacity to connect, and promote the adoption of low-energy lightings. In Mozambique the project is measuring the impact of access to rural electrification in the poorest neighborhoods of Maputo and compares the access of standard electric meters with pre-paid meters. Finally, in El Salvador the project is evaluating the impact of a massive rural electrification program in the Northern Region of El Salvador. The Salvador also provides a very nice opportunity to measure the impacts of different packages of infrastructure given that in addition to electricity, also roads and water and sanitation are being provided.

In Tanzania, the research approach is based on “difference in difference” estimators, which compare the evolution between two periods, of households with access to the program to similar households without access during this period. Since the ex-ante randomization of the intervention across the communities was not feasible, a two-step matching will be implemented: the first one covered communities with similar non-covered ones, then matching participating households to similar households in the counterfactual communities.

In Ethiopia, several experiments are put into place, through lottery-like allocation of discount vouchers for connection to the newly installed grid and the purchase of energy efficient light bulbs. This allows the comparison of service consumption by households in different socio-economic categories, according to different levels of vouchers. Various levels of “treatment intensity” (number of vouchers) at the town level will further allow for the identification of eventual social interactions types of effects.

In Mozambique, regression discontinuity approach will be followed to measure the impact. Finally, in El Salvador a similar approach to the one in Ethiopia is being implemented.