Research

With Kate Beach, Thabbie Chilongo, Charles Jumbe, Andrea Mahieu, and Pam Jagger

Abstract: Over 760 million people, primarily in sub-Saharan Africa, lack electricity access. Off-grid solar offers a scalable alternative. Using two waves of survey data from 1,030 rural Malawian households, where solar uptake trends reflect rural Africa more broadly, we assess the use and impacts of solar technologies.

We employ difference-in-difference estimations to assess how solar adoption impacts vary with device capacity and usage. Small solar devices (<5 W) are predominantly used for phone charging, reducing charging-related travel by 1-hour/week but yielding no other impacts. Households adopting 5-50 W devices gain 20 extra hours of lighting/week. They primarily use solar for security while sleeping or for lighting for domestic tasks, which then increases time spent on productive activities. The largest benefits accrue to the few households adopting >50 W, increasing lighting by 35-50 hours/week and time for productive activities by 3.5 hours.

Relying on population-wide solar adoption metrics obscures energy access realities. Low-capacity solar alone is likely insufficient to achieve Sustainable Development Goal 7. Our findings reflect patterns across sub-Saharan Africa where low electricity access and household solar usage is common. By considering device capacity and household priorities, policymakers can design electrification strategies that more effectively improve well-being while advancing electrification goals.

With Meera Mahadevan, Robyn Meeks, and Manisha Pradhananga

Abstract: Expanding access to basic services like electricity and water is a cornerstone of development policy in many low- and middle-income countries, with its effects extensively studied. However, the consequences of upgrading existing infrastructure, though often equally significant and costly, remain underexplored. This paper investigates the impact of a major infrastructure upgrade—the installation of smart meters in Kathmandu, Nepal—on household behavior and electricity consumption, utility revenue, and service quality. While the theoretical effects of such an upgrade are uncertain, potentially boosting utility revenues while imposing higher costs on consumers, our findings suggest a different outcome. The introduction of smart meters resulted in a Pareto improvement, benefiting both the utility and consumers: utility revenues increased without a corresponding rise in consumer payments. In addition, there was a notable increase in the adoption of energy-efficient lightbulbs and households reported perceiving fewer electricity outages. The mechanism through which we observe these impacts is a reduction in late payments by consumers and an increase in online payments. These results highlight the potential for infrastructure improvements to simultaneously make the consumers better off while improving the utility's cost recovery.

With Samuel Miles, Layla Kwong, Jackson Mughuma, and Dan Kammen

Abstract: Health systems across sub-Saharan Africa face major challenges in delivering essential services due to shortages of staff, equipment, and reliable infrastructure. Electricity is among the most critical inputs. Without sufficient capacity power to run all equipment and services and power that is reliable, lighting, refrigeration, diagnostic equipment, and basic infection control can all be compromised. Despite this, more than one billion people globally still rely on facilities with no electricity or a highly unreliable supply. Using survey data, objectively measured sensor data, and monthly health system records from 303 hospitals and health centers across three provinces of the Democratic Republic of the Congo, we provide one of the first causal estimates of the relationship between having sufficient and reliable power on in-facility outcomes, specifically on mortality rates, patient utilization rates, and staffing. Using longitudinal targeted maximum likelihood estimation and two-way fixed effects models, we find that facilities with sufficient and reliable power (>22 hours of power/day) experience a 1.8 percentage point reduction in mortality (a 41% decline from control facilities, where average mortality rates are 4.5%). For a subset of facilities where we measure reliability over time, we find that during months where the uptime >91.7%, mortality rates are 1.6 percentage points lower, suggesting that having good reliability on average is important, but so too is sustaining that reliability. We also find a 17.3% increase in the number of outpatients and a 23% increase in the number of days nurses are scheduled to work each month. Our results underscore the critical role of sufficient and reliable electricity as a foundational health system input with large impacts on patient survival and service quality and demonstrate the potential of low-cost survey tools for capturing reliability in data-scarce settings.