Development Effects of Electrification: A Meta-Analysis for Income, Labor and Educational Outcomes

Globally, substantial public and private investment has been allocated to electrification, the benefits of which have remained largely, and intuitively, undisputed. However, a close look at the literature suggests that those benefits are far from consistent across different contexts. In fact, estimates of the impacts of electrification are widely heterogeneous and inconclusive. For example, almost 40 percent of the literature’s estimates find no significant effects on income, with similar results for labor and educational outcomes. This paper synthesizes the impact evaluation studies published between the mid-1980s and mid-2019, focusing on the effects of electrification on education, labor and income outcome-categories. Those studies indicate electrification effects (and 95 percent confidence intervals (CI)) of around eight percent in school enrollment (CI=0.04-0.13),17 percent in employment (CI=0.02-0.34), and 24 percent in income (CI= 0.15-0.35).This paper shows, however, that those averages are biased upward, suggesting the presence of selective publication. The meta-regression estimates indicate a mean-corrected, statistically significant effect of 2.6 percent for school enrollment (CI= 0.01-0.04), 5.3 percent for employment (0.04-0.07), and seven percent for income (0.05-0.09). The study design helps explain the estimates’ heterogeneity—i.e., methodologies and data structures requiring less stringent assumptions tend to return lower estimated effects. These findings suggest that the observed heterogeneity reflects actual dispersion in the underlying effects associated with gender differences, as well as, contextual characteristics of the case study, such as country income per-capita and length of the evaluation period.