Why do cities differ so much in productivity? Using a split-sample IV strategy, we document that up to three quarters of the large measured dispersion in productivity across US cities is spurious and reflects the "luck of the draw" of idiosyncratically heterogeneous plants. Due to this...