Intersecting Inequalities in Education and on the Labour Market: Gender and Migration Background in Comparative Perspective

/


Why Study Interactions of Gender and Migration Background in Education and on the Labour
Market?
Gender and migration background are prominent in stratification research as important sources of inequality: being a man or a woman 2 and belonging to a (specific) immigrant group and migrant generation are well known to have implications for individuals' educational and labour market attainment.When it comes to education, most Western industrialised countries have witnessed a considerable shift from a female disadvantage to a prevalent female advantage 1 Fenella Fleischman (f.fleischmann@uva.nl) is professor of sociology at the University of Amsterdam. 2The research reported here and related work mostly relies on a binary definition of gender that differentiates between men and women only, due to data constraints that preclude researchers from capturing more diverse approaches to gender.Despite its obvious limitations, this binary approach is also applied in the writing of this contribution for the sake of readability.
across the educational career (see Buchmann, DiPrete and McDaniel 2008).On the labour market, however, women have not yet reached parity with men in most Western societies and they continue to have lower labour force participation rates. 3Similarly, immigrants and their descendants have lower rates of labour force participation, higher unemployment and less often achieve high occupational status in most Western labour markets (see Heath/Cheung 2007), and they face substantial levels of hiring discrimination when applying for jobs (see Quillian, Heath, Pager, Midtboen, Fleischmann and Hexel 2019).
The relevance of gender and migration background for education and labour market attainment is thus clear and widely acknowledged.However, in the quantitative empirical literature these two sources of inequality are typically studied separately from each other (see Schieckoff/Sprengholz 2021).Much of the research on gender gaps in education and on the labour market focuses on the non-migrant population only (e.g., Blossfeld/Rohwer 1997;Stadelmann-Steffen 2008;Van der Vleuten 2021), while studies of immigrants in Western labour markets -even when they do include immigrant women -do not always include nonmigrant women as reference group (e.g., Stichs 2008), or they present separate analyses for men and women (e.g., Adsera/Chiswick 2007;Kogan 2012;Van Tubergen, Maas and Flap 2004).Few quantitative studies have examined the interactions between gender and ethnic origin with regard to labour market attainment, and most of them have been conducted in countries outside Europe with substantially different histories of immigration, differently composed immigrant populations and different welfare state regulations (e.g., Israel: Raijman/Semyonov 1997; US: Antecol 2001, Cotter, Hermsen andVanneman 1999; but see Salikutluk, Giesecke and Kroh 2020 for a more recent study on Germany).
The joint study of gender and migration background in the explanation of inequality in education and on the labour market allows an empirical investigation of the intersectionality perspective (see e.g., Cole 2009, Crenshaw 1991, Settles/Buchanan 2014).This perspective focuses on the qualitatively different ways in which specific identities come together to create unique experiences and disadvantages or privileges.Concretely, it posits that social identities like gender, race/ethnicity, social class, sexuality, ability etc. are typically not functioning in isolation and may not affect all group members' position and experiences similarly; rather, the way in which they are enacted depends on their combination with other social identities.Specifically for gender and migration background, the implication is that the way in which gender identity benefits or harms persons classified as men and women differs depending on their migration background; and vice versa, that migrant penalties or premiums might differ between male and female members of a specific migrant group.Much of the literature on intersectionality focuses on how individuals experience their specific configurations of social identities.In stratification research, where the object of study is inequality in important socio-economic resources such as educational attainment and labour market positions, the research question is rather whether gender differences replicate across migrant groups or vary by migration background, and relatedly, whether inequalities between migrant groups and non-migrants affect their male and female members to the same extent and in the same way.This can be empirically examined by estimating interactions between gender and (specific) migration backgrounds in the explanation of educational and labour market attainment.In the absence of significant interactions, the results would not lend support to the intersectionality perspective but rather describe the context under study as a situation of additive disadvantages (e.g., for migrant groups and women regarding labour market outcomes) that operate independently.
In this contribution, I bring together empirical findings on the interaction of gender and migration background in the analysis of educational outcomes across the educational career in 9 Western countries, and I zoom in on the case of Germany for three labour market outcomes.Germany is a particularly interesting country to study regarding labour market outcomes because gender gaps differ markedly between its Western and Eastern parts.The core research question of both studies is whether the gender gaps in education and labour market outcomes that are found among the non-migrant population replicate in specific immigrant groups, or whether there is significant variation in gender gaps depending on the migration background.In addition to shedding light on the intersectionality perspective, this question is informed by assimilation theory, which asks to what extent immigrants become increasingly similar to the majority population over time and across generations (see Alba/Nee 2003).Assimilation here explicitly refers to the macro-level, and the research is not focusing on determinants of individual social positions but rather investigating whether within a given society, gender gaps in education and labour market attainment vary by migration background.If gender gaps among (the children of) immigrants resemble those of the non-migrant population, this can be interpreted as a sign of assimilation to the gendered patterns of stratification of migrant-receiving societies.
Against the background of this overarching question, it needs to be acknowledged that the gendered patterns of stratification differ markedly across the life course, and more specifically between the realm of education and the labour market.With regard to education, gender inequality has recently reversed in Western societies: girls are outperforming boys in all stages of the educational career (see Breen, Luyckx, Müller andPollack 2010, Buchmann et al. 2008).Thus, compared to their male peers, young women are more likely to avoid dropping out and complete secondary and tertiary education.It is noteworthy, however, that there are no substantial gender differences in skills, test scores or grades at the beginning of the school career.Rather, domain-specific gender gaps develop over time, with a growing female advantage in reading and grades (see Buchmann et al. 2008, Downey/Vogt Yuan 2005) and a growing male advantage in mathematics (see Marks 2007).This development is accompanied by a growing female advantage in noncognitive skills (see DiPrete/Jennings 2012) and is greater among children from families with lower socioeconomic status (see Entwisle, Alexander and Olson 2007).Compared to the predominant female advantage in education, gender gaps in labour market attainment in European societies are to the advantage of men: across Europe, women are less likely to be active on the labour market, work fewer hours and earn lower wages than men.Moreover, working women are overrepresented in the service sector but underrepresented in managerial positions (see Destatis 2022b).With the transition from education to the labour market, we thus observe a shift of gender gaps, from female advantage to female disadvantage, at least among non-migrant populations.
A similar discrepancy across the life course is noticeable with regard to the role of migration backgrounds, where ethnic disadvantages in education are disappearing over time -at least when the social background is held constant -, but ethnic inequality on the labour market persists within the same cohorts (see, e.g., for the Netherlands: Huijnk/Andriessen 2016).The shifting gender gaps and ethnic penalties between the realm of education and the labour market document that educational attainment, despite its indisputable importance for labour market outcomes, is not sufficient in alleviating labour market inequalities.In other words, greater (gender and/or ethnic) equality in educational outcomes does not necessarily translate into parity (by gender and/or migration background) on the labour market.
The variation of gender gaps and inequalities between migrants and non-migrants across the life course and between the domains of education and the labour market call for an integration of empirical research on the intersection of gender and migration across domains, which is the focus of the present contribution.Before addressing why gender gaps in education and on the labour market might vary by migration background, I will first describe the theoretical mechanisms that may explain gender disparities in these domains.Subsequently, I will discuss how these mechanisms affect (the children of) immigrants from specific origin counties or regions, and formulate two hypotheses that structure the empirical part of this contribution.

Theoretical Accounts of Gender Disparities in Education and on the Labour Market
The most prominent explanation for emerging gender differences in education and later on the labour market, is the notion of gender role socialization.The idea here is that children acquire gender stereotypes and norms prevalent within families and educational institutions in the course of developing a gender identity (see Bussey 2011).Such norms frequently depict girls as more sociable, caring and better able to sit still and concentrate, whereas boys are stereotypically expected to enjoy physical activities and solving technical problems more.As a consequence of gender stereotypes, parents, caretakers and teachers may have gender-biased perceptions of children's abilities and performance, for example, that boys have a better grasp on science and maths (see Schofield 2006).In line with the Pygmalion effect (see Rosenthal 2002), gender-differentiated perceptions and behavioural expectations are likely to translate into gender-biased parenting and teaching styles.Gender role socialization also contributes to gender differences in school-relevant behaviour.On average, girls are found to have more social and noncognitive skills (see DiPrete/Jennings 2012), are more attentive, devote more effort to schoolwork, and are less disruptive in class than boys (see Buchmann et al. 2008, Downey/Vogt Yuan 2005).
Much research addresses how children are socialised with specific gender roles in their families and proximate social contexts.There is also research on the societal origins of specific gender roles or the stereotypical expectations of men and women in a society.Specifically, the social role theory of gender stereotypes posits that the prevalent definitions of a typical woman' and a typical man derive from the assignment of men and women into the roles of breadwinners and homemakers and from specific occupations within the labour market (see Eagly, Wood and Diekman 2000).Accordingly, changes in gender roles, such as the increasing participation of women on the labour market, lead to the adaptation of gender stereotypes.Thus, in countries where female labour force participation has been increasing, women are perceived as more agentic over time (see Diekman /Eagly, 2000), a stereotype content dimension in which men are typically rated higher than women (see Fiske 2017).

Implications for Immigrants and Their Children
Gender role socialization and the social role theory of gender stereotypes jointly suggest that gendered expectations in the field of education and in the labour market may be different among children of immigrants, whose parents grew up in countries with a male advantage in education and low rates of female employment and who were thus socialised with gender roles distinct from those prevalent in their destination countries, and non-immigrant children from countries where the female advantage in education is the rule and female labour force participation tends to be higher.If immigrants from societies with male advantage in education and on the labour market hold more traditional attitudes toward gender roles, they might for instance value education more for their sons than for their daughters, and this might result in smaller female advantages or even a male advantage in educational outcomes, which would then contrast with the non-migrant population.However, one could also derive the opposite expectation from the same scenario: Traditional attitudes toward gendered task distributions may also benefit girls' educational outcomes more than those of boys, particularly if the closer monitoring of daughters compared to sons leads to girls spending more time at home than boys, which might in turn facilitate their investments in education (see, e.g., Varner/Mandara 2013).Furthermore, if children of immigrants resist their parents' traditional gender ideology, girls may pursue a higher education and economic independence as a means to turn against traditional expectations.For boys, in contrast, resisting parental expectations may result in poor performance and defiance in school (see Cammarota 2004).
Gender roles differ strongly from the prevalent pattern in Western destinations in many prominent origin countries of migrants in Europe.In the realm of education, the male advantage particularly prevails in sub-Saharan Africa, North Africa and the Middle East, and South Asia (see Grant/Behrman 2010).Migrants' labour market behaviour, too, may be influenced by the prevalent gender roles in their countries of origin.For instance in the US, the labour market participation of immigrant women is predicted by the rate of female labour market participation in their origin country (see Blau, Kahn and Papps 2010).With regard to the origin countries in our study, we observe considerable variation in the ratio of female-to-male labour market participation, which is low in an important source country like Turkey and much higher in Northern Europe and post-socialist countries (see Fleischmann/Höhne 2013).
Where gendered task divisions are more traditional, children will be socialised with gender stereotypes that differentiate more strongly between men and women.Examining gender gaps in education and on the labour market among a broad range of immigrant groups thus greatly enhances our understanding of the variation in gender roles on a global scope.To the extent that the parents' gender role socialisation drives their gender-specific parenting and educational expectations and labour market aspirations for their children, we would expect the gender gaps that are found in immigrants' origin country to replicate in their destination.This would result in significant interactions between gender and migration background in the prediction of educational and labour market attainment, with smaller female advantages in education and larger female disadvantage on the labour market among (the children of) immigrants from non-Western countries as compared to the non-migrant population.Hypothesis 1 therefore reads: Gender gaps in education and labour market attainment in migrants' origin countries are replicated in their destination countries, resulting in a significant variation of gender gaps by migration background.
There are also reasons to suspect, however, that gender disparities in education and on the labour market may be similar for the migrant and non-migrant population, particularly regarding the second generation, resulting in the second, contrasting expectation.Hypothesis 2 reads: Gender gaps in education and labour market attainment assimilate to the patterns found among the non-migrant population, resulting in non-significant interactions between gender and migration background, particularly from the second generation onwards.
There are three reasons behind this second hypothesis.First of all, migration from a country with male advantage in education to a setting with female advantage greatly increases the potential economic returns to female education.Therefore, the contrast in terms of labour market returns to education between origin and destination countries might be larger for immigrant women and their daughters than for immigrant men and their sons.Human capital theory (see Becker 1985) posits that individuals seek to maximise their returns to education in terms of employment and earnings, yet this individual goal can be overridden by household concerns.Based on the household specialisation approach, being married and having children (particularly infants and pre-schoolers) are expected to lower women's labour force participation and to increase their likelihood of working part-time, while the opposite is expected for men (see Van der Lippe/Van Dijk 2002), all else being equal.Increasing returns to female education in terms of labour market outcomes have been identified as an important reason for the disappearance or reversal of the gender gap in educational attainment in the West (see DiPrete/Buchmann 2006).This development has been accompanied by sociocultural changes in expectations regarding life course trajectories for men and women, with a declining number of individuals in Western countries expressing support for traditional gender roles (see Inglehart/Norris 2003).
Secondly, gender gaps might also vary less depending on the migration background due to selective migration.Immigrants are usually not a representative sample of their origin country's population; in most cases, they are positively selected with regards to educational attainment (see Feliciano 2005), a higher one being a strong predictor of less traditional attitudes towards gender roles (see Inglehart/Norris 2003).In countries with a prevalent male advantage in education and on the labour market, families with more egalitarian attitudes may be more likely to leave their origin country in order to improve the educational opportunities for their daughters.There is little research on selectivity in migration with regard to gender ideology.However, within Western destinations, immigrants from non-Western origins are often found to hold less egalitarian gender role values than the non-migrant population of their destination country (see Röder 2014).
Thirdly, educational and labour market outcomes of (the children of) immigrants are not only derived from individual preferences for gender roles but are also influenced by institutional environments.More specifically, gender equality in educational and labour market outcomes is more likely to generalise also to (the children of) immigrants if gender equality is aspired to and practiced in the schools that the children of immigrants attend, and if the labour market and family policies stimulate gender equality.

Research Design
To test the two hypotheses, the following section presents the outcomes of two empirical studies that examined interactions between gender and migration background.Study 1 is a comparative study across 9 countries that tested these interactions for five educational outcomes, capturing variation by gender and migration background in educational achievement and attainment.Study 2 focuses on Germany as major immigrant destination in Europe and examines variation in gender gaps in labour market participation, part-time work and occupational status by migrant generation and origin country or region, while differentiating between the Western and Eastern parts of the country.Across both studies, the overarching research questions is whether the gendered patterns of educational and labour market attainment that are found among the non-migrant population generalise to all immigrant groups and generations, which would provide support for Hypothesis 2. It is important to note that the absence of gender gap variation by migration background by no means implies parity between genders and migrant groups.Similarity in gender gaps can co-occur with significant female advantage or disadvantage in education and on the labour market, as well as migrant penalties and premiums in these outcomes.The question here is, however, whether gender penalties or premiums generalise across migrant groups and whether migrant penalties and premiums affect male and female members of these groups equally, resulting in additive disadvantages or privileges.The contrasting scenario would rather provide support for the intersectionality perspective, implying that the nature and meaning of gendered (dis-)advantage depends on migration background, and simultaneously, that the nature and the impact of having a migration background for social inequality differs for men and women within the same migrant group.This scenario would lead to significant interactions between gender and migration background in line with Hypothesis 1.

Study 1
To study the intersection of gender and migration background in the field of education, an international team of researchers drew together population, register and nationally representative survey data from 9 Western countries that included sizeable samples of local-born children of immigrants -the so-called second generation -, distinguishable by their country or region of origin.For each country, the maximal number of groups that could be meaningfully distinguished was included.For example, in Belgium, this applies to second-generation Turkish, Moroccans and Italians; in Sweden, 19 origin groups can be distinguished based on their country (e.g., Denmark, Poland, Turkey) or region (e.g., Middle East, South-East Asia) of origin.Analyses were restricted to the second generation, defined as children of immigrants who are born in their parents' destination country or who are foreign-born but arrived before the start of compulsory education (country-specific entry ages were applied).For all countries, the second generation was compared to native-born peers with two native-born parents, described as nonmigrant or majority population.A double comparative approach was applied combining malefemale gaps in educational outcomes with ethnic differentials by estimating interactions between gender and country/region of origin.All analyses were standardized across countries, using harmonized coding of gender, migration background, educational outcomes and control variables.Parental education and occupational status as well as family composition were included as controls.For more details on the data, measures and models, please see Fleischmann and Kristen (2014) and Heath and Brinbaum (2014) on the project as a whole.
Five educational outcomes were analysed to capture a large part of the educational career.We started with OLS regressions of academic achievement at age 15 or 16 years, captured either by scores on standardized tests (e.g., PISA reading, national exams) or grades.We continued with probit regressions of the binary outcomes continuation in fulltime education after the end of compulsory schooling, attendance of academic (as compared to vocational) tracks, completion of upper secondary education, and finally completion of tertiary education.Not all countries could provide estimates for all educational outcomes, depending on data availability as well as features of the national educational system.Thus, our analysis includes 9 countries (Belgium, Finland, France, Germany, the Netherlands, Sweden, Switzerland, England and Wales, and the US) for the first outcome, and smaller sets of countries for the other outcomes (see Table 1).

Study 2
To examine interactions between gender and migration background -including country/region of origin as well as migrant generation -in the explanation of labour market outcomes in Germany, data from the 2009 German microcensus were analysed.The microcensus is an annual survey of 1% of German households, providing information about all of their members (for more information on this data source, see GESIS 2012).The sample was restricted to persons of prime working age (25-54 years) living in private households.Taking into account individuals' immigrant status, migrant generation and country or region of origin allowed us to define 22 distinct groups.Within the non-migrant population, we distinguished between residents of West Germany (the 11 federal states belonging to the Federal Republic of Germany before reunification, including Berlin) and East Germany (the 5 federal states of the former German Democratic Republic) due to different gender gaps in labour market behaviour in the two parts of Germany (see, e.g., Matysiak and Steinmetz 2008).Because immigrants in Germany primarily settle in the Western federal states, West Germans were selected as reference group and the remaining 21 generation-origin combinations were compared to this category.Among the migrant population, we distinguished between the foreign-born first generation and the second generation, which consists of German-born children of immigrants.We created distinct dummy variables for each migrant generation within an origin group, thus contrasting the Turkish first and second generation with the reference group, and similarly for all other groups where we could distinguish two immigrant generations.Three outcome variables were studied with heterogeneous choice models and OLS regressions.The first was labour market participation, which contrasts economically active and inactive persons.Economic activity includes paid work (employed, self-employed or working in family business), including workers who are on parental leave or ill, as well as unemployed who are actively seeking for work.The second analysis was limited to (self-)employed participants and contrasts those working part-time (up to 31 hours per week) with those in full-time jobs (32 weekly hours or more).Finally, we studied occupational status as operationalized with the International Socio-Economic Index (see Ganzeboom et al. 1992).We controlled our analyses for age, marital status, the presence and age of children in the household, participation in education, highest level of education completed and migrants' length of stay.For more details on the data, measures and models of this study, please see Fleischmann and Höhne (2013).

Study 1: The intersection of gender and migration background across the educational career
As a first analytical step to examine whether gender gaps in educational outcomes vary among different migration backgrounds, it was tested whether models including interactions between gender and migration background improved the estimation of the five educational outcomes, compared to more parsimonious models without interactions.Most of these tests yielded nonsignificant results, indicating that including interactions does not improve the model fit.Essentially, this means that the gender gap in educational outcomes replicates among the second generation, and where it does deviate, the difference is not large enough to achieve statistical significance in most cases, thus providing support for the second rather than the first overarching hypothesis.
To further unpack this general finding, Table 1 details the instances where significant interactions between gender and migration background were found, sorted by destination country and educational outcome.The table also specifies the direction of the gender gap among the non-migrant population, which is to the advantage of girls for most outcomes and most countries, attesting to the prevalent female advantage in education in Western countries.We find consistent significant gender gaps to the advantage of girls for educational achievement and the attendance of academic vs. vocational tracks.Regarding continuation and completion of upper secondary and tertiary education, there are a few instances of non-significant gender gaps in the majority population (continuation in Finland, completion of tertiary in England and Wales) and a few cases of male advantage (completion of upper secondary and tertiary education in the Netherlands).In all other cases, non-migrant women outperform their male peers after controlling for social background and family composition.Regarding achievement at age 15-16, the only exception to the pattern of female advantage is found in Switzerland, where boys outperform girls among the Iberian second generation.In all other countries and groups, girls have higher grades or better performance on standardized exams and tests, although the magnitude of the female advantage sometimes deviates from the gender gap among non-migrants.Although not of central interest to this contribution, it is important to note that the prevalent female advantage in academic achievement is combined with a strong ethnic variation, with many migrant groups showing ethnic penalties also after controlling for social background and family composition.
Compared to the distinct female advantage in educational achievement, gender gaps in the continuation of full-time education after compulsory schooling are mostly smaller among the majority but still put girls ahead of boys (see Table 1).This female advantage does not vary across migrant groups in most countries in our sample, except for France and the United States.In contrast to the female advantage among the majority group, male second-generation immigrants from Sub-Saharan Africa and the Caribbean (jointly labelled as Blacks in the data source) in the United States show higher continuation rates than their female counterparts.In contrast, the female advantage in continuation is significantly larger among the North African second generation in France than among the French majority.Apart from these two exceptions, gender gaps in continuation after compulsory schooling do not differ between migration backgrounds.At the same time, there is a large ethnic variation in this educational outcome, and again, the most frequently observed combination consists of a female advantage with ethnic penalties.
The outcome that shows the strongest gender gap variation by migration background is tracking in upper secondary education.Out of the seven countries that could be included in the analysis, four show significant improvements in model fit upon inclusion of interactions between gender and migration background.The general pattern in tracking is one of female advantage, meaning that girls follow academic tracks more often than boys.However, ethnic differentials are more pronounced for tracking than gender gaps, and in contrast to educational achievement and continuation, we find more instances of ethnic premiums than ethnic penalties.This is most likely due to the selectivity of the analysed population, which is limited to individuals who continued in full-time education after the compulsory stage.
Regarding the completion of upper secondary education, we find significant gender gap variation by migration background in three out of six countries.The general gender gap is to the advantage of women, but it is reversed for some groups.The prevalent female advantage in the completion of upper secondary education goes together with a strong variation by migration background.Similar to the findings regarding achievement and in contrast to those regarding tracking, ethnic penalties are more common than ethnic premiums at this point in the school career.
Finally, gender gaps in the completion of tertiary education are only found to vary across migrant groups in Sweden, but not in the four other countries included in this analysis.Thus, the female advantage found among the majority populations in Belgium and the United States is replicated among the second generation, as is the female disadvantage in the Netherlands and the absence of a significant gender gap in England and Wales.In Sweden, the female advantage in the completion of tertiary education among non-migrants is larger among second-generation East Asians, but it is significantly smaller among South-East Asians, Iranians, and the Polish second generation (see Table 1).
Summarising across the five educational outcomes, the findings for education rather support Hypothesis 2, which posits that the second generation assimilates to the pattern of female advantage in education that prevails in the Western destination countries under study, despite widespread and persistent female disadvantage in education in many countries of origin.However, the evidence is not entirely clear-cut: in 11 out of 34 country-outcome combinations, there are significant interactions between gender and migration background, indicating that the magnitude of the gender gap does not replicate exactly in all groups and at all stages of the educational career.Yet, even where gender gaps in the second generation differ significantly from those of the majority population in their magnitude, they rarely reverse.Thus, with the exception of four country-outcome combinations (Iberians in Switzerland regarding academic achievement, Turks in Belgium regarding the choice of academic over vocational tracks, and East Asians and Blacks in the United States regarding the completion of upper secondary education), we saw few cases of female advantage in educational outcomes among the majority population paired with male advantage in the second generation, which should have been the case more often if the gendered patterns of educational outcomes in the origin countries were continuing in the destination countries as predicted by Hypothesis 1.The overall result of the comparative findings across nine destination countries and a large number of second-generation groups is that the female advantage in education found in non-migrant populations extends to second-generation migrants, regardless of their parents' country of origin and despite some significant variation in the magnitude of the gender gap.We found no evidence for a consistent double disadvantage for girls or boys such that they were systematically penalized or systematically benefited by both their migrant background and their gender throughout their educational careers.Rather, the magnitude and the differences in educational outcomes depending on the migration background -both penalties and premiums -highlight that this characteristic is an important source of inequality in education.
Study 2: The intersection of gender and migration background on the German labour market Turning to gender gaps on the labour market and zooming in on the case of Germany, the analyses of the microcensus data reveal that there is an almost universal high labour market participation among men regardless of their migration background, and a much stronger variation among women.Due to the increasing participation of second compared to first-generation women within the same origin groups, we can observe an assimilation to the gender gap prevalent among West German natives.With regard to part-time work, there is much ethnic variation against the background of a strong female disadvantage, meaning that women are less likely to work full-time than men.Gender gaps are much smaller when it comes to occupational status, the third outcome under study.For this outcome, we again observe marked ethnic variation and an increase in occupational status across generations.
More pertinent to the research question of this contribution is whether the gender gap in labour market outcomes found among the reference group of West German non-migrants significantly varies by migration background.Table 2 details the direction of the deviation from the West German gender gap for those groups where the interaction between gender and migration background was found to be significant.The findings show that gender gaps vary considerably depending on the migration background with regard to labour market participation and part-time (in comparison to full-time) work.With regard to occupational status, gender gaps are smaller across the board, yet ethnic and generational variation is substantial.Furthermore, the differences in gender gaps by migration background are not fully accounted for by differential distributions of central explanatory variables of labour market behaviour.Even after controlling for age, level of education, marital status, the presence and age of children in the household, and the differential effects of the latter two variables for men and women, significant interactions between gender and migration background were observed for all labour market outcomes under study.These results confirm previous findings from the US (see Blau et al. 2008) and support the second hypothesis that asserts that gender gaps in the labour market behaviour of migrants' origin countries are replicated in the destination country.At the same time, and similar to the outcomes of Study 1, we find support for the hypothesis that gender gaps among migrants assimilate to those of non-migrants.With the exception of part-time work, where gender gaps were stable over generations, significant differences in gender gaps in the first generation are often no longer significantly different in the second generation.Regarding labour force participation, we observe a clear intergenerational assimilation to West German patterns among Turks, Italians and former Yugoslavians.

Discussion and Conclusion
The large number of significant interactions between gender and migration background when it comes to labour market outcomes stands in contrast to the prevalent absence of such interactions in the realm of education that was found in Study 1.The intersectionality thus finds more empirical support when it comes to labour market outcomes than earlier in the life course.This ties in with findings regarding ethnic penalties in the realm of education that seem to be shrinking over time (at least after controlling for social background), but this increased parity in educational attainment does not go hand in hand with shrinking ethnic inequalities on the labour market (see, e.g., for the Netherlands, Huijnk/Andriessen 2016).The discrepancy in the findings concerning the domain of education and the labour market also underlines the relevance of the life-course perspective taken in the present contribution, whereas focusing on only one domain, as in the underlying empirical contributions, would make it harder to gauge at which point inequalities based on gender and migration background add up and when they interact.At the same time, alternative explanations for the discrepant findings across the two studies should not be discarded too quickly.After all, the first study did not include the foreignborn first generation, whereas the second did and there most significant interactions were found in the first generation.Since the gender gap of the second-generation groups included in Study 2 mostly stopped differing from that of West German non-migrants -and where it did, the difference was still smaller than in the first generation -, our results jointly suggest that interactions between gender and migration background are particularly prevalent in the first generation, which replicates gendered patterns of labour market behaviour with which they have been socialized in their origin countries (and where they completed their education).Future research that goes beyond the descriptive account presented here and examines explanations for gender gaps in education and on the labour market as well as variations thereof depending on the respective migration backgrounds -ideally for multiple immigrant generations, multiple immigrant groups and in multiple countries -will be needed to further assess the different explanations for the discrepant findings of the two studies.Altogether, when it comes to gender gaps in education and on the labour market, the results of the two studies suggest that the substantial differences that the first generation faces are levelled by the second generation.The results presented here did not allow a further examination of the mechanisms that cause variation in gender gaps depending on migration background.Other work finds that attitudes towards or norms about gendered task distributions within households can account for a part of the differences in female labour force participation between different immigrant groups in the Netherlands (see Khoudja/Fleischmann 2015) and in Britain (see Wang 2019).Similar work on the role of gender ideology for gender gaps in educational achievement and attainment is, however, lacking to date.To the extent that such attitudes are driving the larger gender gaps in labour market outcomes in the foreign-born first generation, the findings on assimilation in gender role attitudes across generations (see, e.g., Breidahl/Larsen 2016, Röder 2014, Röder/Mühlau 2016) provide further support for the robustness of our descriptive results and might also explain the absence of variation in gender gaps of different migration backgrounds in educational outcomes of the second generation.
The discrepancy in the results between education and labour market remains a cause for concern.Educational inequality based on a migrant background has been found to be shrinking (see e.g., Huijnk/Andriessen 2016), as are gender gaps (see Buchmann et 2008), but without this growing parity in qualification translating into equal opportunities on the labour market.Where the reversal in gender gaps between the realm of education and the labour market could be explained by different expectations of how men and women should combine work and care tasks across their life course, the shift from ethnic parity to ethnic disadvantage is more difficult to explain based on standard explanations of labour market outcomes rooted in (human) capital endowments.Discrimination is an obvious additional explanation, and there is an abundance of field experimental studies documenting the hurdles that job applicants of migrant origin face across most migrant-receiving societies (see, e.g., Quillian et al. 2019, Zschirnt/Ruedin 2016).It would be interesting to examine, in future research, whether ethnic hiring discrimination is gender specific or vice-versa, whether gender discrimination in hiring affects individuals with a migration background more or less, depending on their specific migration background.A recent study on the intersection of gender and migration background in hiring discrimination found no evidence for effects of signalling warmth or competence (as core dimensions of stereotype contents; see Veit, Arnu, Di Stasio, Yemane and Coenders 2022), but to what extent the content and nature of ethnic stereotypes differs between men and women, and to what extent gender stereotypes differ based on the migrant background in question remains to be studied.If stereotypes of the typical woman and the typical man are informed by the actual roles that men and women are observed to occupy in a society (see Eagly et al. 2000), gender gap variation based on different migrant backgrounds could lead to group-specific gender stereotypes, with group-specific (discriminatory) behaviours as a consequence.
While the findings provide support for assimilation in terms of gender gaps in education and on the labour market, it needs to be emphasized, again, that this does not mean that there is parity regarding gender and migration background.To the contrary, the female advantage in education is still not followed up on the labour market, where women continue to be disadvantaged compared to men, and most migrant groups face significant penalties, particularly on the labour market and less so in education, at least after taking important socio-demographic characteristics into account.These findings jointly call for redoubled efforts to achieve gender and ethnic equality in education and on the labour market; yet they do not call for migrant groupspecific gender policies.Instead, where policy frameworks contribute to equal life chances and outcomes for individuals of all genders, people with migrant backgrounds will benefit as well; and policies that are effective in reducing native-migrant gaps in attainment will likely benefit both male and female members of migrant groups, particularly in the second generation.

Table 1 Interactions between gender and migration background across the educational career, by country
Fleischmann & Kristen (2014)eischmann & Kristen (2014).In the columns Non-migrant female to male gender gap, + indicates a female advantage in the outcome under study, -indicates a female disadvantage, 0 indicates no significant gender gap between male and female peers after controlling for family SES and family composition.In the columns with interaction results, + indicates a stronger female advantage or a weaker female disadvantage, and -indicates a smaller female advantage or a stronger female disadvantage compared to the non-migrant population.n.a.= results for this outcome are not available for this country.Please consult the publication for the full regression tables.

Table 2 .
Fleischmann & Höhne (2013)r and migration background on the German labour market Results are based onFleischmann & Höhne (2013).Data source: German Microcensus 2009.In the row Non-migrant female to male gender gap, + indicates a female advantage in the outcome under study after controlling for age, educational attainment, marital status, presence and age of children in the household, and length of stay in destination country.In the rows with interaction results, + indicates a stronger female advantage and -indicates a smaller female advantage compared to the non-migrant population.Please consult the publication for the full regression tables.