Working Paper: CEPR ID: DP8119
Authors: Pierre-Philippe Combes; Gilles Duranton; Laurent Gobillon
Abstract: Measures of urban productivity are typically positively associated with city population. But is this relationship causal? We discuss the main sources of bias in the proper identification of agglomeration effects. We also assess a variety of solutions that have been proposed in the literature to deal with them.
Keywords: Agglomeration economies; Instrumental variables; Natural experiments; Structural estimation
JEL Codes: R12; R23
Edges that are evidenced by causal inference methods are in orange, and the rest are in light blue.
| Cause | Effect |
|---|---|
| urban productivity (R11) | city population (R23) |
| city population (R23) | urban productivity (R11) |
| urban scale (R12) | wages (J31) |
| sorting of productive workers (J68) | urban wage premium (J31) |
| fixed effects (C23) | bias mitigation (C46) |