Table 7 Fixed-effects and first-differenced models

From: Immigration and wages: new evidence from the African American Great Migration

		OLS				IV
		Black		White		Black		White
Fixed effects
Annual wages	Prop. Southern	−0.276	0.212	1.354***	1.413***	−6.675*	−3.420	−1.373	−0.862
		(1.948)	(1.345)	(0.405)	(0.299)	(3.731)	(5.360)	(2.993)	(4.749)
	Observations	3150	3150	148,303	141,642	3110	3110	143,248	137,752
	Clusters	29	29	100	92	28	28	80	78
	Covariates	N	Y	N	Y	N	Y	N	Y
Weekly wages	Prop. Southern	−1.193	−0.488	0.847**	0.908***	−4.339	−6.411	−0.217	0.958
		(0.839)	(1.053)	(0.366)	(0.309)	(2.817)	(4.691)	(1.680)	(4.202)
	Observations	3135	3135	147,855	141,232	3095	3095	142,825	(0.839)
	Clusters	29	29	100	92	28	28	80	78
	Covariates	N	Y	N	Y	N	Y	N	Y
First differences
Annual wages	Prop. Southern	−0.188	0.445	0.254	0.297	−9.796***	−7.385**	17.00	−6.143
		(2.154)	(2.150)	(0.253)	(0.372)	(2.584)	(2.905)	(135.7)	(8.574)
	Observations	21	21	93	67	20	20	59	59
	Covariates	Y	N	Y	N	Y	N	Y	N
Weekly wages	Prop. Southern	−0.144	0.154	0.0978	0.217	−5.490**	−5.104**	15.35	−6.362
		(1.261)	(1.294)	(0.190)	(0.275)	(2.215)	(2.262)	(122.0)	(8.788)
	Observations	21	21	93	67	20	20	59	59
	Covariates	Y	N	Y	N	Y	N	Y	N

Notes: Covariates include white and black metro-level percent employed in manufacturing, percent farming and average years of education and well as indicators for age and educational attainment. The dependent variable for the first-differenced models is the mean residual from a regression of wages on indicators for age and educational attainment. Standard errors for the fixed-effects estimates are clustered at the metro level, for the first-differenced estimates they are heteroskedasticity robust
“***”, “**”, and “*” denote significance at the 1, 5, and 10 % levels, respectively

Back to article page