Sensitivity of Results to Using Metropolitan Areas and Restricting the Sample to 1980–2010
| All Natives | Native Non-Hispanic Whites | Native Blacks | Native Hispanics | |||||
|---|---|---|---|---|---|---|---|---|
| Fixed Effects (1) | Ten-Year Differences 2SLS (2) | Fixed Effects (3) | Ten-Year Differences 2SLS (4) | Fixed Effects (5) | Ten-Year Differences 2SLS (6) | Fixed Effects (7) | Ten-Year Differences 2SLS (8) | |
| Share population 11–64 that is immigrant; t–10 | 0.09** | 0.26** | 0.03 | 0.10 | 0.18** | 0.41 | 0.34** | 0.12 |
| (0.04) | (0.09) | (0.03) | (0.08) | (0.08) | (0.27) | (0.12) | (0.15) | |
| Share population 11–17 that is immigrant; t–10 | −0.04 | −0.18 | 0.02 | 0.16 | −0.10 | −0.49 | −0.20 | −0.20 |
| (0.05) | (0.32) | (0.04) | (0.28) | (0.18) | (0.95) | (0.16) | (0.80) | |
| Share population 18–64 that is immigrant; t–10 | ||||||||
| <12 years education | 0.10 | 0.70** | 0.06 | 0.39* | 0.19 | 1.39 | 0.50* | 0.15 |
| (0.12) | (0.30) | (0.08) | (0.23) | (0.26) | (0.91) | (0.27) | (0.65) | |
| 12 years education | 0.13 | 0.12 | 0.01 | 0.58 | −0.11 | −0.53 | 0.02 | −0.70 |
| (0.15) | (0.86) | (0.13) | (0.71) | (0.31) | (2.54) | (0.30) | (2.23) | |
| >12 years education | 0.12 | 0.32 | 0.00 | −0.50 | 0.52** | 1.38 | 0.80** | 1.14 |
| (0.11) | (0.88) | (0.10) | (0.77) | (0.25) | (2.62) | (0.24) | (2.26) | |
| Other covariates | Yes | Yes | Yes | Yes | Yes | Yes | Yes | Yes |
| Observations | 390 | 260 | 390 | 260 | 390 | 260 | 390 | 260 |
Notes: The dependent variable is the share of natives age 21–27 of a racial/ethnic group who have completed 12 years of education, adjusted at the individual level for age and sex, as well as race and ethnicity (except Columns 3 and 4). Estimation is by weighted least squares, with weights w the inverse of the squared standard errors on the metro area–year interaction coefficients in the individual regression (odd Columns), or with weights 1/(1/wt + 1/wt+10) (even Columns). All specifications include year dummies and metro area dummies. Other covariates are those listed in Table 2, or their equivalent based where relevant on 1980 rather than 1940. The dependent variable is based on 1990–2010 data, the independent variables on data from 1980–2000. Standard errors are clustered by metro area and reported in parentheses. The instruments are based on the 1940 distribution of immigrants from different regions (see text), and correspond to those in 2SLS 1 in earlier tables. The F-statistic for the excluded instrument in the first stage is 13.9 in the upper panel (even Columns). The Angrist–Pischke F-statistics for the four instruments in the first stage of the lower panel are 1.0, 3.9, 2.1, and 3.5 (even Columns).
*** p < 0.01,
↵** p < 0.05,
↵* p < 0.1.