Impacts of Rural Electrification in Rwanda

Publication Details

Journal of Development Effectiveness, December 2011, vol.3, iss.4, pp.567-588. Available From:

Link to Source
Author
Gunther Benscha, Jochen Kluvea, Jörg Peters
Institutional affiliations
None specified
Grant-holding institution
None specified
Country
Rwanda
Region
Sub-Saharan Africa (includes East and West Africa)
Sector
Agriculture and Rural Development, Energy
Subsector
Rural Livelihoods, Rural Electrification
Subsector
Rural Livelihoods, Rural Electrification
Equity Focus
None specified
Evaluation design
Difference-in Difference (DID), Propensity Score Matching (PSM)
Status
Journal Article

Methodology

Electrification is widely assumed to positively affect a number of Millennium Development Goals, through fostering social and economic development and improving overall quality of life. In rural Rwanda, only 1.3 percent of the population has access to electricity, prompting a number of governmental and multilateral efforts to improve rural electrification. However, little empirical data thus far documents the linkages between access to electricity and improved outcomes. This study seeks to fill that gap by examining the impact of access to electricity on lighting usage, children’s study time at home, household energy expenditures, and income. To evaluate the impact of electrification, data are collected in areas that will receive electrification in the future and in carefully selected sites that are similar in terms of location and other characteristics but have had access to electricity for on average 4 years. The first analysis compares households in electrified areas that connect to the grid with those that do not.

The authors note that several additional types of analysis are needed in order to account for unseen differences between the two types of households. Within the areas with access to electricity, the authors develop a probit model to predict the likelihood of household connection to the electricity source using covariates that are uncorrelated with the four outcomes of interest. Using the weights from that model, they then predict the same likelihoods in areas without access to electricity, identifying the households most likely to have connected to electricity were it available (“hypothetically connected households”.) These households serve as the counterfactual to the electrified households. Impacts are then estimated using propensity score matching first with stratification and then with nearest-neighbour matching with non-replacement. To further address potential endogeneity, the authors conduct a difference-in-differences analysis using the non-connected households and hypothetically non-connected households to control for regional differences.

Main findings

Comparing households with access to electricity that do and do not choose to connect to the grid, the authors find that those that do connect (1) use 13.4 additional hours of lighting a day, (2) experience 4.1 additional hours of children studying at home, (3) have 3.1 additional FRw energy expenditures per adult, and (4) have 5,600 FRw in additional income, all significant at the 1-percent level.

In the stratified propensity score matching analysis, which better controls for household characteristics influencing the decision to connect, lighting hours per day increased by 7.5 hours (p = .01) and household income increased by 200,000 FRw (p = .05), but hours spent studying and energy expenditures lost significance. The results for nearest-neighbour matching with propensity score matching are similar, with the exception that children spent 2.2 more hours studying at home, significant at the 5-percent level. Finally, in the difference-in-differences estimation, which controls for regional differences as well as household characteristics, lighting hours per day increased by 8.58 (p = .01), but all other outcomes were insignificant.

The authors note that even though the lack of robust results for hours spent studying at home, energy expenditures, and household income may be disappointing, the net welfare gain from significantly increased hours of lighting in the home should not be underestimated. The increased hours may have a significant impact on quality of life and indirect welfare impacts in areas not addressed in this study or over a longer period of time than seen here. Further research should address the long-term effects of access to lighting and the willingness to pay for electricity by rural people as a proxy for gauging its true value.

Scroll to Top