How much of the geographic clustering of economic activity is attributable to agglomeration spillovers as opposed to natural advantages? I present evidence on this question using data on the long-run effects of large scale hydroelectric dams built in the U.S. over the 20th century, obtained...