Working Paper: CEPR ID: DP4240
Authors: Pierre-Philippe Combes; Gilles Duranton; Laurent Gobillon
Abstract: Spatial wage disparities can result from spatial differences in the skill composition of the workforce, in non-human endowments, and in local interactions. To distinguish between these explanations, we estimate a model of wage determination across local labour markets using a very large panel of French workers. We control for worker characteristics, worker fixed-effects, industry fixed-effects, and the characteristics of the local labour market. Our findings suggest that individual skills account for a large fraction of existing spatial wage disparities with strong evidence of spatial sorting by skills. Interaction effects are mostly driven by the local density of employment. Not controlling for worker heterogeneity biases estimates of agglomeration economies by up to 100%. We also find evidence of various omitted variable biases and reverse causality between agglomeration and high wages. Finally, endowments only appear to play a small role.
Keywords: local labour market; panel data analysis; sorting; spatial wage disparities
JEL Codes: J31; J61; R23
Edges that are evidenced by causal inference methods are in orange, and the rest are in light blue.
| Cause | Effect |
|---|---|
| Not accounting for sorting (C69) | Estimates of agglomeration economies (R12) |
| Individual skills (G53) | Spatial wage disparities (J31) |
| Worker fixed-effects (J29) | Individual wages (J31) |
| Sorting of workers by skills (J24) | True local productivity differences (O49) |
| Skills and interactions (Y80) | Residual spatial wage disparities (J79) |
| Local employment density (J69) | Wages (J31) |