1. Expand sample to include men who live in the US South (see footnote 16)

Transcription

1 Online Appendix for A Nation of Immigrants: Assimilation and Economic Outcomes in the Age of Mass Migration Ran Abramitzky, Leah Boustan, Katherine Eriksson 1. Expand sample to include men who live in the US South (see footnote 16) 1950 occupation-based earnings in $2010 dollars, Full national sample (1) Cross section (2) Pooled cross section and panel (a) Cross section (b) Panel 0-5 yrs in US (142.44) (175.00) (217.54) 6-10 yrs US (146.52) (162.65) (193.75) yrs US (130.39) (128.13) (140.26) yrs US (140.59) (134.71) (141.81) 30 yrs in US (181.97) (169.05) (170.38) Arrive (102.55) (154.30) Native born (865.58) 295, ,252 Notes: Sample follows notes to Table 1 in paper, with the exception of the southern exclusion. Specification follows Table 3 in the paper.

2 2. Following Hatton (1997), replace the quartic in age with a linear effect of age that is allowed to differ before/after age 35 (see footnote 23) 1950 occupation-based earnings in $2010 dollars, Hatton specification (1) Cross section (2) Pooled cross section and panel (a) Cross section (b) Panel 0-5 yrs in US (142.89) (186.79) (237.29) 6-10 yrs US (147.49) (171.93) (213.46) yrs US (131.23) (133.89) (150.44) yrs US (142.02) (138.70) (150.23) 30 yrs in US (184.26) (178.11) (185.46) Arrive (107.04) (160.61) Native born (176.11) N 205, ,093 Notes: Sample follows notes to Table 1 in paper. Column organization follows Table 3 in the paper. Specification follows equation 1 but replaces the quadratic in age with a linear age term, a dummy variable equal to one for individuals over 35 years of age and an interaction of these two terms.

3 3. Separately estimate equation (1) for the repeated cross section sample and the panel sample (see footnote 25) 1950 occupation-based earnings in $2010 dollars, Separate RCS and panel Cross section Repeated cross Panel section 0-5 yrs in US (143.44) (186.59) (452.31) 6-10 yrs US (147.44) (177.51) (420.68) yrs US (131.27) (145.57) (318.92) yrs US (142.10) (151.18) (291.64) 30 yrs in US (184.65) (184.87) (399.93) Arrive (104.37) (217.67) Native born N 205, ,458 53,635 Notes: Sample follows notes to Table 1 in paper. Columns report from estimates of equation 1 conducted separately for the pooled cross section, the repeated cross section (which includes a dummy for arriving after 1891) and the panel sample.

4 4. Use smaller sample based on a conservative matching procedure that requires individuals be unique by first and last name and place of birth within a five-year age band (see footnote 27) 1950 occupation-based earnings in $2010 dollars, Conservative matched sample (1) Cross section (2) Pooled cross section and panel (a) Cross section (b) Panel 0-5 yrs in US (143.44) (189.38) (407.44) 6-10 yrs US (147.44) (177.00) (343.81) yrs US (131.27) (141.66) (224.29) yrs US (142.10) (146.64) (229.70) 30 yrs in US (184.65) (146.64) (279.92) Arrive (106.55) (257.93) Native-born (221.73) N 205, ,682 Notes: The sample derives from the same sample universe reported in Table 1 (column 2). The matched sample here includes only men who were unique by first and last name and by place of birth within a five year age range. Specification follows Table 3 in the paper.

5 5. Use Goldin and Margo (1991) wage series by occupation category to back cast earnings to the early 1920s (see footnote 30) 1950 occupation-based earnings back cast to early 1920s, in $2010 dollars (1) Cross section (2) Pooled cross section and panel RHS variable (a) Cross section (b) Panel 0-5 yrs in US (103.68) (135.66) (173.19) 6-10 yrs US (112.09) (129.22) (157.57) yrs US (101.82) (102.00) (112.86) yrs US (109.68) (106.25) (113.62) 30 yrs in US (139.44) (133.95) (139.02) Arrive (79.00) (119.30) Native born (126.01) N 205, ,093 Notes: Sample follows notes to Table 1 in paper. Specification follows Table 3 in the paper occupation-based earnings are adjusted using the Goldin-Margo time series of earnings by occupation category (clerical, skilled blue collar, unskilled blue collar). See footnote 30 for details of this adjustment.

6 6. Immigrant earnings penalty/premium by state of residence in 1900 (see footnote 32) 8 Estimated gap in occupation score between immigrants and natives in 1900 Census Nevada Rhode Island Massachusetts South Dakota North Dakota Iowa Nebraska California Connecticut Kansas New Jersey Minnesota Oregon Pennsylvania Colorado Montana New York Wisconsin Wyoming Idaho New Hampshire Maine Michigan Illinois Utah Ohio Vermont Arizona Indiana Missouri New Mexico -6 Notes: Results are based on a sample of immigrants and native born men between the ages of 18 and 65 in the 1900 IPUMS cross section. We regress occupation score (in levels) on a vector of dummy variables for countries of origin, a vector of dummy variables for state of residence, and interactions between state of residence and a dummy for being foreign born. We report the coefficient on this interaction term for each state above. To convert the in $2010, as in the paper, multiply each by $900.

7 7. Predicting cross-country differences in immigrants initial occupation-based earnings relative to natives (see footnote 36) Dependent variable = Initial difference in earnings (immigrants versus natives) from Figure 3 Mean/standard Univariate Multivariate deviation of regression * regression: RHS variable Add economic Multivariate regression: Add cultural variables *** Characteristic of sending country (RHS variable) variable ** Share in agriculture (0.172) ( ) ( ) ( ) Real wage (25.636) (24.67) (23.77) (17.28) Natural increase (3.635) (169.49) (156.14) (105.02) Infant mortality rate (54.934) (10.15) (9.48) (8.19) Linguistic distance (0.344) ( ) ( ) ( ) Cultural distance (0.588) (677.38) (961.15) (920.47) Religious distance , , , (0.045) ( ) (11,222.16) (12,943.87) Notes: Initial earnings difference between permanent immigrants and US natives is measured using the coefficient on 0-5 years in US (relative to natives) in panel sample. N = 16 except for the following RHS variables: infant death (missing for Portugal); cultural distance (missing for Russia) and real wage (missing for Finland, Russia and Switzerland). Economic characteristics are taken from Mitchell (2007) and real wage from Williamson (1995). We accessed the cultural distance measures from Sin (2011). These measures were originally collected by Alesina, et al. (2003); Fearon (2003) and Hofstede (1980). Real wage is indexed to US = 100. Natural increase is crude birth rate minus crude death rate. Infant mortality rate measured per 1000 live births. Linguistic distance theoretically varies between 0 (close) and 1 (far). Religious distance measures the chance that two randomly selected residents, one from the country of origin and one from the US, are of different religious backgrounds. Cultural distance varies between 0.3 (close) and 2.1 (far) in this sample. Standard errors are in parentheses. * Column 3 regresses the immigrant earnings penalty on each of the RHS variables listed in the first column in turn. ** Column 4 adds share in agriculture to each regression (except in the share in agriculture row, in which case the real wage is added). *** Column 5 adds the cultural distance and religious similarity measures to each regression. In cultural distance row, we only add religious similarity and in religious similarity row we only add cultural distance.

8 8. Backing out the implied occupation-based earnings gap between temporary and permanent migrants by country-of-origin (see footnote 39) A. Using return migration rates reported by Gould (1980) Implied earnings gap between temporary and permanent migrants B. Using return migration rates reported by Bandiera, Rasul and Viarengo (2013), Table 4 Implied earnings gap between temporary and permanent migrants Note that the implied earnings gap between temporary and permanent migrants is larger when using Gould s figures because his reported return migration rates are lower than in Bandiera, et al. (mean = 26 percent vs. 62 percent), implying that the earnings gap between temporary and permanent migrants must be larger to justify the difference between the repeated cross section and panel.