Working Paper: CEPR ID: DP1714
Authors: Simon Burgess; Julia Lane; David Stevens
Abstract: The ?fractal? nature of the rise in earnings dispersion is one of its key features and remains a puzzle. This paper offers a new perspective on the causes of changes in earnings dispersion, focusing on the role of labour reallocation. Once we drop the assumption that all firms pay a given worker the same, the allocation of workers to firms matters for the dispersion of earnings. This perspective highlights two new factors that can affect the dispersion of earnings: rates of job and worker reallocation, and the nature of the process allocating workers to jobs. We set out a framework capturing this idea and quantify the impact of reallocation on earnings dispersion, using a dataset which comprises almost the universe of workers and the universe of employers in Maryland. We show that these factors have potentially large effects in general on earnings dispersion. In the case of Maryland over the period 1985?94, the changing allocation of workers to jobs played a significant role in explaining movements in the dispersion of earnings.
Keywords: earnings inequality; within-group inequality; allocation of workers; matched worker-firm panels
JEL Codes: D33; J31
Edges that are evidenced by causal inference methods are in orange, and the rest are in light blue.
| Cause | Effect |
|---|---|
| covariance between worker and firm fixed effects (J79) | changes in earnings dispersion (J31) |
| changing allocation of workers to jobs (J29) | changes in variance of earnings (C22) |
| labor reallocation (J69) | earnings dispersion (J31) |
| allocation of workers to firms with different wage markups (J29) | earnings distributions (D33) |
| job reallocation (J62) | earnings dispersion (J31) |
| worker turnover (J63) | earnings dispersion (J31) |