Working Paper: CEPR ID: DP18101
Authors: David Autor; Christina Patterson; John Van Reenen
Abstract: National industrial concentration in the U.S. has risen sharply since the early 1980s, but thereremains dispute over whether local geographic concentration has followed a similar trend. Usingnear population data from the Economic Censuses, we confirm and extend existing evidenceon national U.S. industrial concentration while providing novel evidence on local concentration.We document that the Herfindhahl index of local employment concentration, measuredat the county-by-NAICS six-digit-industry cell level, fell between 1992 and 2017 even as localsales concentration rose. The divergence between national and local employment concentrationtrends is attributable to the structural transformation of U.S. economic activity: bothsales and employment concentration rose within industry-by-county cells; but reallocation ofsales and employment from relatively concentrated Manufacturing industries (e.g., steel mills)towards relatively un-concentrated Service industries (e.g. hair salons) reduced local concentration.A stronger between-sector shift in employment relative to sales drove the net fall in localemployment concentration. Holding industry employment shares at their 1992 level, averagelocal employment concentration would have risen by about 9% by 2017. Instead, it fell by 5%.Falling local employment concentration may intensify competition for recent market entrants.Simultaneously, rising within industry-by-geography concentration may weaken competition forincumbent workers who have limited sectoral mobility. To facilitate analysis, we have madedata on these trends available at concentration trends.
Keywords: concentration; structural transformation
JEL Codes: L16
Edges that are evidenced by causal inference methods are in orange, and the rest are in light blue.
| Cause | Effect |
|---|---|
| structural transformation of economic activity (O14) | divergence in local versus national concentration trends (R12) |
| shifts in economic activity (R11) | concentration trends (D30) |
| reallocation from concentrated manufacturing sectors to less concentrated service sectors (O14) | decline in local employment concentration (F66) |
| if industry employment shares had remained at 1992 levels (L16) | local employment concentration would have increased (R11) |
| falling local employment concentration (R23) | enhance competition for new market entrants (L13) |
| rising within-industry concentration (L19) | weaken competition for existing workers with limited mobility (J69) |