Abstract
This paper examines the impact of gender empowerment on labor market performance of U.S. married immigrants across different family types. Three types of families are defined based on the nationality of immigrants’ spouses; ‘Home-country Marriage’ where both spouses come from the same home-country, ‘Native Marriage’ where one spouse is an immigrant and the other is a Native, and ‘Foreign Marriage’ where both spouses are immigrants, but each comes from a different country of origin from the other. Recognizing the role of gender ideology across different family structures, this study provides empirical evidence of the impact culturally based gender status has on immigrants across family types. Labor market performance is measured by wages and the labor force participation decision of immigrants, using the Ordinary Least Squares and Logit models. The gender empowerment measure is utilized to reflect different cultural and institutional conditions, which shape gender status in the immigrants’ home countries. Results indicate the positive effect of gender equality on women’s earnings and labor force participation, as well as that home-country gender status strongly influence women’s labor force participation and earnings when they share the same cultural norms as their husbands. Interestingly, women are more influenced by their spouses’ gender norms than men when immigrants come from a different cultural gender norm than their spouse.
This is a preview of subscription content, log in via an institution to check access.
Similar content being viewed by others
Notes
Empirical results on the family investment model are somewhat ambiguous regarding the labour supply of married female immigrants, with different evidence across researchers. For example, Worswick (1996) and Baker and Benjamin (1997) find results supportive of the family investment hypothesis, while Blau et al. (2003) and Basilio et al. (2009) provide counterexamples to the family investment model, finding both wives and husbands have a similar pattern of wage increases relative to native workers in the U.S. and Germany. Meanwhile, Cobb-Clark and Crossley (2004) find mixed results for the family investment hypothesis depending on the gender of the primary worker in the household. They find supportive results on the family investment hypothesis for the traditional household but non-supportive results for the non-traditional household.
Most empirical studies on tied movers deal with internal migrant couples rather than international migrants due to difficulties related to data collection and identifying tied movers or the strength of tie between husband and wife for dual earning couples. Most of these studies show that married women benefit less from migration in the job market in terms of earnings, employment, job market conditions, etc. (Geist and McManus 2012; Blackburn 2010; Quinn and Rubb 2011; Cooke 2003a, b; Büchel and Battu 2003; Bielby and Bielby 1992; Sandell 1977).
Epidemiological technique is an approach to analyse the influence of culture separate from the influence of markets and institutions using immigrants or immigrants' descendants' data (Fernández 2008).
The three categories assigned to foreign couples based on the degree of cultural differences from the U.S. are as follows: Foreign-born couples with a different culture from the U.S., foreign-born couples with a somewhat similar culture to the U.S., and foreign-born couples with a very similar culture to the U.S. This culture of origin difference is by ancestry not by country of birth. In addition, they do not consider family type such as intercultural marriage.
Natives only include people who are born in the U.S and exclude people who are born in U.S outlying areas such as Puerto Rico, Guam or other U.S. Possessions.
Basilio et al. (2009) report 37% of immigrant husbands arrive ahead of their wives, utilising a detailed data set that includes the time of marriage for immigrant couples in Germany. This implies that married immigrant couples from the same home-country may include many couples where spouses arrive in succession. For robustness, possible tied mover couples who move simultaneously, a subset of all ‘Home-country Marriage’ couples, are tested.
For details on the construction of the GEM, please see supplemental online Appendix I in Huh (2017).
Although the component for economic resources represented by GDP per capita may be more reflective of a country’s overall development status than women’s empowerment, it is beneficial to include in this study since the level of skill transference between host and home-country should be controlled.
These correlations are even higher between 1999 and 2007 when more consistent methods of measurement were implemented for GEM. The correlation among GEM ranks and values for the time period between 1999 and 2007 are 0.877 and 0.910, respectively.
The youngest individuals in the sample are 25 years old in 2006 and if they migrated at the age of 18 (youngest migration age) then they migrated in 1999.
The subsample of ‘Home-country Marriage’, where both spouses moved simultaneously, is 21,908 and about 29% of ‘Home-country Marriage’ women migrated with their husband in the same year. The simultaneously-moved ‘Home-country Marriage’ immigrants have similar characteristics with the total sample of ‘Home-country Marriage’ immigrants, showing 47% of females and 15% of males do not work. Among the 15,007 labour market participants, 61% are male and 39% are female.
Linking spousal birthplace information within the sample results in the same number of male and female immigrants in ‘Home-country Marriage’ couples. ‘Foreign Marriage’ couples are composed of 49% male and 51% female while ‘Native Marriage’ couples are composed of 40% male and 60% female.
Sources of income-wage include wages, salaries, commissions, cash bonuses, tips, and other money income received from an employer; payments-in-kind or reimbursements for business expenses are excluded.
About 35% of women and 38% of men in ‘Home-country Marriage’ couples have a bachelor’s degree or higher education and about 49% of women and men have more than a high school education. For labour market participants, 41% of women and 40% of men in ‘Home-country Marriage’ couples have a bachelor’s degree or higher education.
Through the remainder of the paper, italicised terms represent the variables’ names as presented in the table.
Immigrants responding to the fluency question on the ACS survey who state they “speak English only”, “speak English very well”, and “speak English well” are classified as “speaking English Fluently” while those who respond to the same question by selecting “does not speak English” or “not well” are denoted as “speaking English Non-fluently”.
Numerous ACS categories for region and race were combined in this analysis into larger divisions: The ACS regional category of New England Division and Middle Atlantic Division to the “East”; East North Central Division and West North Central Division were combined and classified as the “Middle West”; Mountain Division and Pacific Division were coupled to form the “West”; and South Atlantic, East South Central, and West South Central Division are incorporated in the “South”.
For the race category, information for “Hispanic” and “Single race identification” are combined because a single race category in the 2006 ACS survey does not provide Hispanic information and the portion of Hispanic among immigrants is large: 42% of female immigrants and 55% of male immigrants classify themselves as Hispanic. People who did identify themselves as Hispanic in the sample are classified as Hispanic no matter how they respond to the “race” question.
Wage is the earned wage and salary income of each individual.
Considering the influence of assortative marriage on earnings, spousal income and the number of children under 5 are tested in the earning equation, both together and separately. Although spousal income shows almost zero influence on both labour market participation as well as earnings, the number of children shows a positive significant influence on earnings across all immigrant men. However, the main results of the paper are not changed much.
An alternative dependent variable, imputed hourly wages, instead of annual salary-wages, is also tested across all regressions. All regression results using imputed wages are the same as the main results. In addition, considering the possibility of endogeneity of wages with respect to hours worked, wage equations without working hours are tested as a robustness check. All regression results without working hours are similar to the main results quantitatively and qualitatively.
The GEM shows strong significant effects on the selection equation and wage equation for all women in the Heckman two-stage model, implying OLS results might be biased due to selection bias on labour market participation for women, but not for men.
The standard deviation of the GEM is 0.121067.
A standard deviation factor change is calculated as follows: Exponent (Log Odds Ratio × S.D. of GEM).
Thus, the likelihood of being in the labour force increases 1.34 times [Exp (2.41217 × 0.121067) = 1.34] for ‘Home-country Marriage’ women and 1.14 times [Exp (1.10499 × 0.121067) = 1.14] for ‘Native Marriage’ women at the time of arrival.
This is similar to findings on work behaviour of second generation American women in Fernández and Fogli (2009) in that the impact of husbands’ culture is more significant than that of wives’.
The Gender Development Index (GDI) measures the conditions in a country necessary to provide individuals with the opportunity to develop their ‘basic capabilities’, to pursue their goals and achieve well-being in terms of standard of living. It also includes three components to measure standard of living in health, education, and economic resources. It incorporates the longevity of women and men, the adult literacy rate and the combined primary, secondary and tertiary enrolment ratios of women and men, and estimated earned income for women and men based on GDP in a nation. The regressions with the GDI and the interaction between years in the U.S. and GDI, as well as regressions including both GDI and GEM and their interactions with years in the U.S. are tested. Replacing GEM, GDI shows qualitatively similar results but with much more pronounced impact on men’s outcome and no significant impact on women’s labour force participation, demonstrating the GEM is a better measurement for gender inequality. When both indices are controlled together, the significance of GDI decreases across all models, reflecting the fact that part of GEM subsumes GDI’s impact. The significance of GEM on both earnings and labour force participation on women did not change.
Utilising the Barro-Lee World Educational Attainment Dataset (2000), a binary dummy variable which indicates whether immigrants are more or less educated than the average school years of their corresponding home-country population is created.
Models including the interaction between GEM and educational level are tested in an attempt to transform GEM into a within-country effect.
Huh (2017) demonstrates that all immigrant workers from 42 countries are more educated than their home-country population counterparts and low GEM leads to more positive selection of immigrant women. Positive (Negative) selection occurs when immigrants have attributes which help (or hinder) labour market success more than those who remained in their home-country. Findings in Huh (2017) reinforce the results of this study since those results would make it more likely to underestimate the positive GEM effects on earnings of immigrant women.
Simultaneously-moved couples from the same home country is tested in all models to examine possible bias from tied movers.
Women’s low labour market participation suggests a possible sample selection issue for women in the labour force not being representative of the whole population. If women who participated in the labour market are not a random sample and are a more positively (or negatively) self-selected group, the estimates in the earning equations could be biased upward (or downward). In this case, GEM effects on earnings could be overestimated (or underestimated). Complications regarding labour market participation and earnings can also cause selection bias. For example, if women from high GEM countries are more likely to participate in the labour market, then relatively higher numbers of low wage women from these countries may appear in the sample, whereas only high wage women from countries with low GEM would appear in the earnings sample. In this case, GEM effects would be biased downward. This will reinforce the main results, a positive GEM effect on earnings.
This is also supported by the regression results using the Heckman two-stage model, which considered selection bias regarding labour market participation. Once selection into the labour market is controlled, ‘Home-country Marriage’ immigrants show the strongest influence from GEM in the outcome equation.
This is to consider the fact that Heckman’s two-stage model sometimes underestimates standard errors in the second stage (outcome) equation. But, Heckman (1979)’s two-stage procedure, the LIML (Limited Information Maximum Likelihood) estimator, is also more robust than FIML in that FIML more heavily relies on its normality assumption.
Accounting for cohort effects in some specifications must be done with caution since with a single cross section, years -in the -U.S.S and year of arrival are jointly determined. Estimation of both in the current case will turn on functional form.
A discrete variable which contains 9 levels of education (No school, primary education (1–6 grade), middle school (7–9 grade), high school (10–12 grade) drop out, High school graduate, Some college education, Associate degree, Bachelor's degree, Master's degree, and Doctoral/Postgraduate degree) is created.
Huh (2017) demonstrates that all immigrant workers from 42 countries are more educated than their home-country population counterparts and low GEM leads to more positive selection of immigrant women. Positive (Negative) selection occurs when immigrants have attributes which help (or hinder) labor market success more than those who remained in their home-country. Findings in Huh (2017) reinforce the results of this study since those results would make it more likely to underestimate the positive GEM effects on earnings of immigrant women.
Regional dummy variables in 6 categories (Asia, Europe, Middle East, Latin America and Caribbean, and Others) are created. All results are similar to the main findings in both quantitative and qualitative ways, except for the regressions on earnings of married men when regions are controlled.
References
Antecol, H. (2001). Why Is there interethnic variation in the gender wage gap?: the role of cultural factors. The Journal of Human Resources, 36(1), 119–143.
Antecol, H. (2003). New evidence on culture and the gender wage gap: a comparison across ethnic origin groups. Research in Labour Economics, 22, 447–464.
Atkinson, A. B. (1970). On the measurement of inequality. Journal of Economic Theory, 2, 244–263.
Baker, M., & Benjamin, D. (1997). The Role of the family in immigrants’ labour market activity: a evaluation of alternative explanations. The American Economic Review, 87(4), 705–727.
Barro-Lee (2000) World educational attainment dataset. Education attainment for population aged 25 and over. Harvard University Department of Economics, http://barrolee.com/ (Accessed January 6, 2010).
Basilio, L., Bauer, T., & Sinning, M. (2009). Analysing the labour market activity of immigrant families in Germany. Labour Economics, 16(5), 510–520.
Bielby, W. T., & Bielby, D. D. (1992). I will follow him: Family ties, gender role beliefs and reluctance for relocate for a better job. American Journal of Sociology, 97(5), 1241–1267.
Blackburn, M. L. (2010). Internal migration and the earnings of married couples in the United States. Journal of Economic Geography, 10(1), 87–111.
Blau, F. D., & Kahn, L. M. (2011) Substitution between individual and cultural capital: Pre-migration labour supply, culture and us labour market outcomes among immigrant women. NBER Working paper No. 17275.
Blau, F. D., Kahn, L. M., Moriarty, Y., & Souza, A. P. (2003). The role of family in immigrant labour market activity: An evaluation of alternative explanation: comment. American Economic Review, 93(1), 429–447.
Blau, F. D., Kahn, L. M., & Papps, K. L. (2011). Gender, source country characteristics and labour market assimilation among immigrants. The Review of Economic and Statistics, 93(1), 43–58.
Büchel, F., & Battu, H. (2003). The theory of differential overqualification: Does it work? Scottish Journal of Political Economy, 50, 1–16.
Cobb-Clark, D. A., & Crossley, T. F. (2004). Revisiting family investment hypothesis. Journal of Labour Economics, 11(3), 373–393.
Cooke, T. J. (2003a). Family migration and relative earnings of husbands and wives. Annals of the Association of American Geographers, 93(2), 338–349.
Cooke, T. J. (2003b). All tied up: tied staying and tied migration within the United States, 1997 to 2007. Demographic Research, 29(30), 817–836.
Duleep, H. O., & Sanders, S. (1993). The decision to work by married women. Industrial and Labour Relations Review, 46(4), 677–690.
Fernández, R. (2007). Alfred Marshall lecture women, work and culture. Journal of the European Economic Association, 2(2–3), 305–332.
Fernández, R. (2008) ‘Women, work, and culture.’ NBER Working Paper No. 12888.
Fernández, R., & Fogli, A. (2006). Fertility: The role of culture and family experience. Journal of the European Economic Association, 4(2–3), 552–561.
Fernández, R., & Fogli, A. (2009). Culture: An empirical investigation of beliefs, work and fertility. American Economic Journal: Macroeconomics, 1(1), 147–177.
Fu, V., Winship, C., & Mare, R. D. (2009). Sample selection bias models. In M. Hardy & A. Bryman (Eds.), The handbook of data analysis (pp. 409–430). London: SAGE Publications Ltd.
Geist, C., & McManus, P. (2012). Different reasons, different results: Implication of migration by gender and family status. Demography, 49(1), 197–217.
Heckman, J. (1979). Sample selection bias as a specification error. Econometrica, 47(1), 153–161.
Huh, Y. (2011). The effect of home-country gender status on the labour market success of immigrants. Journal of Human Development and Capabilities, 12(3), 367–392.
Huh, Y. (2017). Gender empowerment and educational attainments of U.S. immigrants and their home-country counterparts. Feminist Economics, 23(2), 120–145.
Long, J. E. (1980). The effect of americanization on earnings: Some evidence for women. The Journal of Political Economy, 88(3), 620–629.
MacPherson, D. A., & Stewart, J. B. (1989). The labour force participation and earnings profiles of married female immigrants. Quarterly Review of Economics and Business, 29(3), 57–72.
Ong, P. M. (1987). Immigrant wives labour force participation. Industrial Relations, 26(3), 296–303.
Oreffice, S. (2014). Culture and household decision making. Balance of power and labour supply choices of US-born and foreign-born couples. Journal of Labour Research, 35(2), 162–184.
Quinn, M. A., & Rubb, S. (2011). Spouse overproduction and family migration: Evidence from the U.S. Journal of Family Economics, 32(1), 36–45.
Ruggles, S., Alexander, J.T., Genadek, K., Goeken, R., Schroeder, M.B., & Sobek, M. (2010) ‘Integrated Public Use Microdata Series: Version 5.0.’ [Machine-readable database]. Minneapolis: University of Minnesota. https://usa.ipums.org/usa/cite.shtml (Accessed January 21, 2010).
Sandell, S. H. (1977). Women and economics of family migration. The Review of Economics and Statistics, 59(4), 406–414.
UNDP (2007) ‘Human Development Report 1990–2007.’ United Nations Development Programme (1990–2007), New York: Oxford University Press.
Worswick, C. (1996). Immigrant families in the canadian labour market. Canadian Public Policy, 22(4), 378–504.
Author information
Authors and Affiliations
Corresponding author
Ethics declarations
Conflict of interest
The author declares that she has no conflict of interest.
Appendix: Robustness checks and alternative specifications
Appendix: Robustness checks and alternative specifications
A number of robustness checks and different model specifications for Eqs. (1) and (2) are conducted. First, a selection corrected model for earnings is tested, which considers women’s potential selection bias into the labor market. Equations (1) and (2) are used for the output and selection equations in the Heckman two-stage selection model, respectively (see “Appendix Table 9”). As expected, all married women show selection bias regarding labour market participation, but no selection bias for married men is found in the selection corrected model. The results from the selection corrected model demonstrate a strong positive GEM effect on both labour market participation and earnings for ‘Home-country Marriage’ and ‘Native Marriage’ women. Additionally, the Heckman FIML (Full Information Mffaximum Likelihood) estimation, which consists of finding the parameter values that maximize the likelihood of the data, is also tested but no differences were found.Footnote 34 Considering the paper includes a large sample (1000 + observations), the robustness of the Heckman estimator is supported by existing Monte Carlo evidence and theoretical knowledge from OLS estimates (Fu et al. 2009) (Appendix Tables 10, 11 and 12).
Second, all regressions for Eqs. (1) and (2) are repeated with additional dummy variables which consider the selection bias of immigrants relative to non-migrants who remained in their home countries, as well as cohort differences by time of arrival for immigrants. Using 10- year intervals (besides individuals who migrated before the 1970s which is a single interval) all regressions include 5 categories, pre1970, C7079, C8089, C9099, and C2000 s which correspond to immigrants who migrated before 1970, during 1970–1979, 1980–1989, 1990–1999, and 2000–2006 respectively.Footnote 35 Examining selection bias between immigrants and their home-country populations, a binary dummy variable (More_Edu) is employed, which indicates whether immigrants are more or less educated than the average school years of their corresponding home-country population who did not migrate by utiliszing the Barro-Lee World Educational Attainment Dataset (2000). In addition, the interaction between GEM and educational levelFootnote 36 is also tested in an attempt to transform GEM into a within-country effect which examines heterogeneity of gender-based socioeconomic status within a country (“Appendix Table 13”). All results under these specifications do not change the main results of Tables 5 and 7.Footnote 37
Additionally, additional regional dummy variables which indicate the broader region in the world from which an immigrant came from are also tested to control for country-specific effects, since it is not possible to consider the full set of country fixed effects (“Appendix Table 14”).Footnote 38 The results under this specification reinforce the strong and consistent GEM effects of the main results for all women, but OLS results for married men show some difference, namely insignificant GEM effects across all family types.
Finally, subsamples of the data are also tested as a robustness check. The subsample of ‘Home-country Marriage’, composed of simultaneously-moved couples from the same home-country, is tested for all models to examine possible bias from tied movers (“Appendix Tables 15, 16 and 17”). As expected, the results for simultaneously-moved couples from the same home -country were very similar to the main findings of ‘Home-country Marriage’ couples. In addition, all regressions are repeated with datasets excluding either Mexican or Filipino immigrants or both, which compose the largest immigrant groups in the data. The results from all of these different subsets of data were consistent with the main results (Appendix Tables 18 and 19).
All regression results in the paper are reported with robust standard errors. Testing for multicollinearity in the regression, a Variance Inflation Factor test was conducted and no issues were found for anyall models. The results demonstrate consistently statistically significant effects for GEM on the earnings of immigrant women (besides ‘Foreign Marriage’ women) across different model specifications and different segments of the data set.
Rights and permissions
About this article
Cite this article
Huh, Y. Family typology and gender empowerment: the labour market performance of married immigrants. J Pop Research 35, 237–288 (2018). https://doi.org/10.1007/s12546-018-9208-9
Published:
Issue Date:
DOI: https://doi.org/10.1007/s12546-018-9208-9