This paper uses a regression discontinuity approach to investigate whether a set of colonial policies adopted in the Diamond District of colonial Brazil have long-run impacts on development. Results regarding household income are still inconclusive. On the other hand, the estimated effects on adult literacy and light density from satellite images are positive. I also try to explore potential channels through which this historical event might influence the present. Using a geospatial road location database, I find that observations inside the District's historical boundaries have denser road networks. Additionally I use microdata from the 1830s to show that slavery was more intense in untreated villages, which has been related in the literature to underdevelopment.

ACCESS PAPER