Welfare impacts of ultra low-cost solar: Evidence from a large-scale randomized trial in rural Rwanda

How to cite

Clarke, Rowan Philip. 2018. "Welfare impacts of ultra low-cost solar: Evidence from a large-scale randomized trial in rural Rwanda." AEA RCT Registry. AEARCTR-0003218. https://doi.org/10.1257/rct.3218-1.0

Download: BibTeX · RIS — for reference managers without a browser translator. Zotero and most others pick up this page's citation metadata directly.

Rowan Philip Clarke*

* Primary investigator

Registry ID
AEARCTR-0003218
DOI
https://doi.org/10.1257/rct.3218-1.0
First registered
Last updated
Status
On going
Trial dates
2016-09-09 – 2019-06-01
Countries
Private
Registry record
http://www.socialscienceregistry.org/trials/3218

Abstract

Over 1.3bn people worldwide lack access to modern energy. In Sub-Saharan Africa more than 600 million live without electricity and this is expected to rise given grid expansion is not predicted to keep pace with population growth. Moreover, almost 6 million children under 5 die each year from preventable diseases, the main cause being respiratory infections, and smoke from traditional lighting sources is a risk factor in correlational and small-scale studies. Further, the US and UK alone, via Power Africa's Beyond the Grid, have recently committed to invest over $1 billion in off-grid and small-scale solar solutions in Sub-Saharan Africa, and this is a fraction of the total amount invested. Yet, there have only been small-scale rigorous studies in Africa which examine welfare impacts. Given the significant sums invested, it is important for governments and impact investors to know the welfare and development effects of low-cost off-grid solar lighting.

This will be the first truly large-scale study to use a randomized control trial to evaluate solar energy in rural Africa. We carry out our research in Rwanda, a country with low levels of rural electrification, which it has in common with the broader region. We evaluate welfare effects of an ultra low-cost program which has distributed solar lighting to over 150,000 households in rural Rwanda. We collect survey data on roughly 16,000 individuals and 3100 households in 283 villages, half of which were randomly assigned solar in 2016. In addition, we use big data on light usage which is captured automatically and remotely transmitted via GSM technology designed for the purposes of this research. We will also use administrative program cost data for a cost-effectiveness analysis.

Recently a randomized evaluation of rural grid electricity rollout, and another on solar lighting, found no impacts, or that where these existed they were not welfare-improving because of high costs. The benefit of solar lamps and phone charging is their very low cost and ease of distribution. However, impacts, if they exist, can also be expected to be small. Thus, a large sample is required to detect these and conclusions about zero impacts should not be made based on studies with insufficient statistical power. Although impacts are likely small, the low cost of solar lights could make them a cost-effective welfare-enhancing intervention, unlike grid electricity which has failed in this respect.

Intervention

The intervention, which has been rolled out to 150,000 rural households in Rwanda so far, includes one free light per household with the opportunity to purchase additional lights.

Experimental design

Our sample consists of 283 rural villages without electricity in Rwanda. These villages were already being visited by enumerators for a different study so this led to cost economies. Half received solar lighting in December 2016, while the remaining serve as the control group. We collect data on roughly 16,000 individuals and approximately 3100 households in 283 villages. We also collect big data on light usage captured automatically and remotely transmitted via GSM technology designed for the purposes of this research. We also use administrative program cost data which will allow us to conduct an accurate cost-effectiveness analysis.

Design details

Randomization method
Computer
Randomization unit
Village
Clusters
283
Observations
16,000 people from 3100 households
Arms
142 treatment villages, 141 control villages
Primary outcomes
We study the general socioeconomic welfare impacts as well as air quality related health including labor supply, income and employment, consumption and expenditure, assets, time-use, air quality related health, study time, safety, and happiness and life satisfaction.

Cited by

No citation data available for this registration yet.

Keywords