This article advances the debate over legal origins by using satellite imagery of nighttime lights as a proxy for economic activity and by employing geographic re- gression discontinuity analysis to control for observable and unobservable factors cor- related with location, such as climate and culture. The basic legal structure of most countries was imposed by colonial powers, but Great Britain, France and other Eu- ropean nations did not colonize in a random way. The lack of random assignment means that simple cross-country analysis may lead to erroneous conclusions because of unobservables correlated with legal origin. Regression discontinuity analysis is es- pecially promising for Africa, because many borders were drawn in Europe by diplo- mats and bureaucrats who had only the haziest knowledge of local conditions, except in coastal areas. As a result, borders split ethnic groups, and areas on either side of the border are similar along observable dimensions and presumably on unobservable ones as well. By comparing African border regions with the same ethnic group on both sides, but which have civil law on one side and common law on the other, or where one side was colonized by one European country (e.g. Britain) and the other side by another (e.g. France or Portugal), one can partially disentangle the influence of law and other colonial policies, controlling for geography, climate, and pre-colonial culture. Analysis suggests that countries with common law legal origin do not per- form consistently better than those with civil law.