First-Stage Regressions
| Variables | (1) | (2) | (3) | (4) | (5) |
|---|---|---|---|---|---|
| 1976–2015, LBD sample | |||||
| log(HHI–m) | 1.064*** | 0.748*** | 0.829*** | 0.827*** | 0.466*** |
| (0.0120) | (0.0201) | (0.0174) | (0.0173) | (0.0166) | |
| Observations | 5,450,000 | 5,450,000 | 5,446,000 | 5,446,000 | 5,446,000 |
| R2 | 0.504 | 0.773 | 0.930 | 0.932 | 0.956 |
| F-stat | 7,824 | 1,389 | 2,265 | 2,284 | 791 |
| 2005–2015, LBD sample | |||||
| log(HHI–m) | 1.062*** | -0.328*** | 0.503*** | 0.505*** | 0.192*** |
| (0.0130) | (0.0786) | (0.0303) | (0.0300) | (0.0226) | |
| Observations | 1,531,000 | 1,531,000 | 1,527,000 | 1,527,000 | 1,527,000 |
| R2 | 0.537 | 0.792 | 0.974 | 0.974 | 0.985 |
| F-stat | 6,667 | 17 | 276 | 284 | 73 |
| 2005–2015, W-2 sample | |||||
| log(HHI–m) | 1.053*** | -0.131** | 0.505*** | 0.505*** | 0.187*** |
| (0.0128) | (0.0640) | (0.0280) | (0.0274) | (0.0204) | |
| Observations | 1,522,000 | 1,522,000 | 1,519,000 | 1,519,000 | 1,519,000 |
| R2 | 0.540 | 0.801 | 0.975 | 0.975 | 0.986 |
| F-stat | 6,747 | 4 | 326 | 339 | 84 |
| Year FEs | No | Yes | Yes | No | No |
| CZ FEs | No | Yes | No | No | No |
| Industry FEs | No | Yes | No | No | No |
| Market FEs | No | No | Yes | Yes | Yes |
| CZ by year FEs | No | No | No | Yes | Yes |
| Market trends | No | No | No | No | Yes |
Source: Longitudinal Business Database 1976–2015
Notes: Table reports regression estimates of the relationship between local industrial concentration, as measured by the HHI, and its instrument, the leave-one-out mean of the HHI across other markets in the same industry. Columns represent separate regressions, which include the indicated fixed effects (FE) in addition to the instrument. Regressions are employment-weighted. Coefficients represent elasticities. Sample sizes and statistic values have been rounded for disclosure avoidance.