Immigration, labor market mobility, and the earnings of native-born workers: an occupational segmentation approach

American Journal of Economics and Sociology, The,  April, 2006  by Roberto Pedace

• E-mail
• Print
• Link

The samples for the migration calculation of Equation (3) include males and females aged 26-44 who were living in MSAs identified on both the 1980 and 1990 Public Use Samples. (13) The age restrictions are such because those aged 16--25 in the 1990 Census were not in the labor force in 1980 and, thus, no migration calculation of this type is possible for those individuals. Those aged 16-25 reporting themselves in the labor force in 1990 are simply counted as net in-migrants in their corresponding labor market segment. In addition, the regression samples were restricted to native-born individuals aged 16-64 who reported all the necessary personal and employment information and were civilian, nonstudent, wage and salary workers.

VI

Measuring the Impact of Immigration by Labor Market Segment

BOSTON'S (1990) CLASSIFICATION SCHEME IS USED AS A MODEL for the occupational segments created for this analysis. Although Boston (1990) uses the 1983 Current Population Survey to cluster occupations into a primary and a secondary sector, the occupations are matched with compatible IPUMS codes. Tables 4 and 5 provide detailed descriptions of the primary and secondary sector segments.

Once labor market segments are created and estimates of native labor market migration are obtained, an augmented human capital equation is estimated separately for the primary and secondary sector:

[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII] (4)

where the i, j, and s subscripts represent individual, MSA, and labor market segment, respectively; log w is the natural logarithm of the weekly wage (annual earnings/number of weeks worked); EDUC is a vector of categorical variables representing education groups (e.g., high school graduate, some college, etc.); AGE represents the respondent's age; [AGE.sup.2] is AGE squared; REG is a vector of categorical variables representing region of residence; MARRIED is a marital status dummy; HEALTH is a health status dummy; and IND is a vector of categorical variables representing industry (e.g., manufacturing, construction, etc.).

The variables of interest are I/N and M/N, which are the percent of recent immigrants and the percent of net native and earlier immigrant (pre-1980 arrivals) labor migration, respectively. Both of these are by labor market segment and MSA, raising the issue of group effects, which has been largely overlooked in this literature. Since the immigration and labor mobility variables are at a higher level of aggregation than the dependent variable, the residuals will no longer be independent across all individual observations. Instead, the error term will contain a component that is common to all individuals belonging to the same group (i.e., MSA). The consequence is that standard errors will tend to be understated and the likelihood of finding statistically significant coefficients will increase (Moulton 1986). The wage equation, therefore, is estimated with unadjusted and group-effects-adjusted standard errors.