Geographies of Intergenerational Immigrant Labour Markets International Population Geographies Conference University of Liverpool 19-21 June 2006 Jamie.

Geographies of Intergenerational Immigrant Labour Markets

International Population Geographies ConferenceUniversity of Liverpool

19-21 June 2006

Jamie Goodwin-WhiteDivision of Social StatisticsSchool of Social SciencesUniversity of Southampton

Theoretical Background

Concern with how immigrants and their children are doing over time and across generations

Spatial dimensions of immigrant economic progress (Waldinger, Clark, Ellis, Zhou 2001)

Local contexts of labor market inequality (McCall 2001, Bound and Freeman 1992)

Questions

Does the geography of immigrants matter for estimating wages and wage gaps?

Questions

Does the geography of immigrants matter for estimating wages and wage gaps?

Is there a selectivity to internal migration that relates to wage outcomes? Is it different for immigrants and natives? How?

Models 1:Geographically Weighted Regression

Does the geography of immigrants matter for estimating wages and wage gaps?

How to weight population to account for different spatial distributions

Geo-weighted Regressions: Data and Methodology

25-64 year-old men from 5% PUMS

(currently employed, and with positive wages and hours data)

Split into immigrant Mexicans and native-born whites

Covariates and weights

Ln(hourly wage)

Age, education (+ fb interactions), married, years of experience (age-5-educ), arrival cohort (for fb)

Model iterations with 1) pweight and 2) mexweight

mexweight: proportion of fb mex (by metro) used to weight regression

Findings

Geoweighted models show:

1) higher wages for both groups

2) wage penalty to immigrants increases slightly

General models predicted logged hourly wage:

pweight mexweight

nbw 2.31 ($10.15) 2.40 ($10.98)

fbm -.18 ($8.51) -.21 ($8.94)

Additional findings: covariates

Significant and dramatic wage penalty to recent arrivals in geoweighted models

Additional findings: covariates

Significant and dramatic wage penalty to recent arrivals in geoweighted models

Significantly reduced wage penalty to those native whites with less than college education

Additional findings: covariates

Significant and dramatic wage penalty to recent arrivals in geoweighted models

Significantly reduced wage penalty to those native whites with less than college education

Large returns to BA increase sizably in geoweighted models - especially for immigrants

Summary:Distinctive Contexts of Immigrant Geographies

Does the geography of immigrants matter for estimating wages and wage gaps?

Yes, areas where immigrant (Mexicans) live are higher waged but more unequal

But much of this penalty is for recent arrivals

Payoffs to higher education, especially for immigrants - and fewer penalties for least-educated when recent arrivals taken into account

Models 2:Heckman Selectivity Models

Is there a selectivity to internal migration that relates to wage outcomes? Is it different for immigrants, their second generation adult children and natives? How?

In other words, are those who move likely to be those who expect to benefit from the move?

Models 2:Heckman Selectivity Models

Two-step models: Probit on instrumental variable migration

(moved at metro level last 5 years) Selectivity parameter (lambda or inverse mills

ratio) comes from taking unmeasured variance from this model and using as covariate in OLS wage model

Heckman Migration-Selectivity Models:Data and Methodology

25-54 year-old men from 2000 PUMS

(currently employed, and with near full-time full-year data)

Split into immigrants, natives, and 1.5 generation

Two-stage covariate selection

Wage models include:

race(white,black, asian,hispanic, other), age, education (BA), lambda

Migration models also include:

family size, new employment growth, and percent immigrant

Findings

Initial OLS: age, education, race generally work as expected. Immigrant Hispanics and native Blacks fare particularly

poorly. 1.5 Generation Hispanics, whites, Blacks fare better then native-born counterparts.

Wages have very slight negative relationship to move

Heckman selectivity-corrected models show strong positive selectivity effect of migration on wages

=0.42 (native-born), ~ =.20 (immigrants and 1.5 generation)

Changes due to migration selectivity controls

BA degree effect stronger still in Heckman models - underestimated without migration selectivity controls (especially for immigrants)

Much of penalty to being Hispanic ( for immigrants, and 1.5 generation only) abated by controlling for migration selectivity

=0.42 (native-born), ~ =.20 (immigrants and 1.5 generation)

Changes due to migration selectivity controls

BA degree effect stronger still in Heckman models - underestimated without migration selectivity controls (especially for immigrants)

Much of penalty to being Hispanic ( for immigrants, and 1.5 generation only) abated by controlling for migration selectivity

=0.42 (native-born), ~ =.20 (immigrants and 1.5 generation)

Immigrant concentration has slight negative effect on wages, but not for native-born

Summary

Is there a selectivity effect on how migration effects wages?

Yes - and a strongly significant one. It is stronger (more efficient for natives than the foreign-stock). Uncorrected wage regression models show negative bias to migration’s effect on wages.

Migration is a way in which immigrant and 1.5 generation Hispanics attenuate severe wage penalties relative to the native-born. Or not.

Strong payoffs to BA, especially for immigrants, indicate that internal migration could also be an intergenerational labor market strategy.

Conclusions

It is critical to understand immigrant and second generation geographies when assessing immigrant economic incorporation.

Internal migration and different group’s selectivity thereof is part and parcel of this.

Proximity to co-ethnics or other immigrants is only part of the equation on immigrants’ residential patterns and labor market outcomes - especially over time and for the second generation.