Table 5

First Stage of Two-Stage Least Squares (Preferred Instruments)

	Share of Immigrants in Age Group		Share of Population aged 18–64 that is Immigrant with
	11–64 (1)	11–17 (2)	Less than 12 Years Education (3)	Exactly 12 Years Education (4)	More than 12 Years Education (5)
Predicted share population 11–64 that is immigrant	0.27^** (0.05)	—	—	—	—
Predicted share population 11–17 that is immigrant	—	0.68^**	−0.00	−0.04	0.05
		(0.26)	(0.07)	(0.02)	(0.06)
Predicted share population 18–64 that is immigrant:
<12 years education	—	0.00	0.72^**	0.04	0.10
		(0.09)	(0.07)	(0.04)	(0.07)
12 years education	—	−1.40^**	−0.82^**	0.24^**	−0.24
		(0.52)	(0.22)	(0.09)	(0.14)
>12 years education	—	0.67^**	0.44^**	0.09	0.43^**
		(0.25)	(0.11)	(0.06)	(0.09)
F-statistic for joint significance	24.9	8.3	65.5	43.4	29.3
Angrist–Pischke F-statistic	—	15.1	103.7	55.0	18.8
Partial R²	0.40	0.39	0.56	0.36	0.42
Observations	294	294	294	294	294

Notes: Estimation is by weighted least squares on ten-year differenced data, with weights 1/(1/h_t + 1/h_t₊₁₀), where h is the denominator of the dependent variable in each regression: population age 11–64 in Column 1, population age 11–17 in Column 2, population age 18–64 with less than 12 years education Column 3, population age 18–64 with exactly 12 years education in Column 4, population age 18–64 with more than 12 years education Column 5. All specifications include year dummies and the (differenced) nonimmigrant covariates of Table 2 Column 7. The excluded instruments (predicted shares) are based on historical patterns of settlement by age and education. The dependent variable and excluded instruments are based on data from 1940–2000 (that is, from t–10), while other covariates are based on data from 1950–2010. Standard errors are clustered by state and reported in parentheses. This table underlies the 2SLS 1 estimates and differenced 2SLS estimates when instrument is not specified.
*** p < 0.01,
↵** p < 0.05,
* p < 0.1.