Why do cities differ so much in productivity? We document that most of the measured dispersion in productivity across US cities is spurious and reflects granularity bias: idiosyncratic heterogeneity in plant-level productivity and size, combined with finite plant counts. As a result, economies...

This paper exploits a short-lived cooperation program between the U.S.S.R. and China, which led to the construction of 156 "Million-Rouble plants" in the 1950s. We isolate exogenous variation in location decisions due to the relative position of allied and enemy airbases and study the long-run...