The Value of Electricity Reliability: Evidence from a Natural Experiment in Senegal

Abstract

Governments and donors invest heavily in expanding access to electricity in low-income countries. However, recent studies document small impacts of access to grid electricity in several low-income countries. This paper provides new evidence on the impact of grid reliability on the willingness to pay (WTP) for grid electricity based on large scale reliability improvement projects in Senegal’s electricity supply network. I use high-frequency geocoded administrative data on outages, customer-level bills covering the universe of electricity users, and rich geocoded data on the economic activities of households and firms to conduct two analyses. First, I estimate the causal effect of improved reliability on consumers through a difference-in-differences approach that leverages the variation in the timing of reliability projects. I find that these projects reduced outage duration by 40%, increased electricity consumption by 8%, and led to greater appliance ownership, a reallocation of time away from household chores to wage labor, and lower disconnection rates. Second, I estimate a quantitative demand model and a random-coefficient discrete choice model of energy sources. I find that reliability projects increase the WTP for an additional hour of grid electricity by 10%, from an average of $0.20. This increase varies over peak and off-peak hours and differs significantly across customer types (firms versus households) and by establishment size (small versus large). This finding corresponds to an internal rate of return for reliability projects exceeding 30%. My findings suggest that reliability improvements can significantly increase the value of grid electricity, even in contexts where baseline reliability and the WTP for grid electricity are low.