Powering education

More than 1.3 billion people worldwide have no access to electricity and this has first-order effects on several development dimensions. In this paper we focus on the link between access to light and education. We randomly distribute solar lamps to 7th grade pupils in rural Kenya and monitor their educational outcomes throughout the year at quarterly frequency. We find that access to lights through solar lamps is a relevant and effective input to education. Our identification strategy accounts for spillovers by exploiting the variation in treatment at the pupil level and in treatment intensity across classes. We find a positive and significant intention-to-treat effect as well as a positive and significant spillover effect on control students. In a class with the average treatment intensity of our sample (43%), treated students experience an increase in math grades of 0.88 standard deviations. Moreover, we find a positive marginal effect of treatment intensity on control students: raising the share of treated students in a class by 10% increases grades of control students by 0.22 standard deviations. We exploit household geolocation to disentangle within-class and geographical spillovers. We show that geographical spillovers do not have a significant impact and within-school interaction is the main source of spillovers. Finally, we provide suggestive evidence that the mechanism through which lamps affect students is by increasing co-studying at school especially after sunset.