The labour market participation of recently-arrived immigrant women in Germany

Objective: This article investigates the role of motivation in female immigrants’ labour force participation. Focusing on recently-arrived immigrants (who have resided in the host country for 18 months or less), we compare the outcomes of two different ethnic groups in Germany: Poles and Turks. Background: The immigrant integration literature tends to focus on the role of resources in immigrant labour market integration. However, when examining particularly the labour force participation of female immigrants, their motivation for joining the labour force is also important. Previous studies of female immigrants in Germany have often neglected this consideration, which includes aspects like culturally-specific gender values and perceived ethnic discrimination. Method: We use data from the SCIP project (Diehl et al., 2015) to conduct logistic regressions on female immigrants’ labour force participation. Our sample includes 829 female immigrants from Poland and Turkey between the ages of 18-60, who were either active in the labour force or were ‘at risk’ of entering. Results: In line with previous studies, our analysis shows that female immigrants’ labour market resources, mainly their prior work experience and German proficiency, greatly reduce the ethnic gap in labour force participation rates. Moreover, motivational factors have a large impact on this outcome for both groups, and greatly enhance the picture that our empirical models present. However, we find no evidence that perceived ethnic discrimination plays an important role. Conclusion: Our analysis indicates that when seeking to understand the labour market participation of female immigrants, their resources and motivation should be seen as key components of a gender-sensitive analysis.


Introduction
Germany has been the destination of numerous immigrants since the end of the Second World War including guest workers, ethnic German migrants 1 , intra-EU movers, and a considerable number of refugees, with the highest gross intake of immigrants in the EU since 2012 (and a place in the top five receiving countries since at least 2006) 2 .As of 2019, the foreign population in Germany exceeded 11 million, with another 21.3 million residents having a migrant background. 3For several decades now, immigrants' labour market integration has attracted considerable academic and societal interest.After all, this is a key component not only for their economic independence, but also for their integration in other societal domains.This being said, the focus of much research has been dedicated to the dynamics of male immigrant labour market integration, at the expense of studying females.In fact, a comprehensive overview of the existing quantitative studies on immigrant labour market integration in Germany reveals that, between 2000 and early 2020, only 6 of 50 publications have focused solely on the outcomes of female migrants, while 12 have focused exclusively on males (see Appendix I for the complete overview).In this article, we show that a gender-sensitive approach is fundamental to understanding the dynamics of female immigrant labour market integration by comparing women from Poland and Turkey and their likelihood to join the labour force after their arrival in Germany.
In the past, females' timely integration into the labour market may have been considered less economically-crucial than males', at least as long as families were considered and the traditional 'male breadwinner' perception held sway.However, there are several reasons to put women in the focus of research on immigrants' labour market integration.The failure to integrate female immigrants into the labour market leads to higher poverty rates, prolonged welfare dependency, and ultimately increases the threat of old-age poverty in this demographic group (Tucci 2012;Frick et al. 2009;Matthai 2004).Moreover, low female employment has been shown to negatively affect the chances of migrant children attending public day care (which in itself has a positive impact on their language acquisition and school success, as reported by Peter & Spieß 2015, for example).Even though German legislation 4 has guaranteed a childcare spot for children between the ages of one and three since August 2013, studies have shown that children of nonworking immigrant mothers are less likely to start day care at an early age (Krapf 2014;Gambaro et al. 2017).Additionally, labour market (in)activity is a behaviour found to be transmitted from mother to daughter (Farré & Vella 2013;Thornton, Alwin & Camburn 1983).Thus, inactive mothers are likely to result in inactive daughters, furthering the negative effects of this phenomenon in the ensuing generations.On top of all this, labour market integration is crucial for female migrants' economic and legal independence, including their chances of ending abusive or exploitative relationships.
Given that female integration outcomes deserve to be studied in their own right, not just as caveats to those of males, analyses should take a gender-sensitive approach to account for the different dynamics at play.After all, the way that factors like marital status and children in the household influence male outcomes is much different than the effect they have for females (see, for example, Donato, Piya & Jacobs 2014).Moreover, the motivation of women to join the labour force and seek paid employment is a different consideration than for men, whose labour force participation is often taken as a given fact.
In this regard, our analysis focuses on Poles and Turks in Germany for several important reasons.While Turks have been the subject of much prior research (mostly in the context of guest workers and their descendants), immigrants from Eastern Europe have received considerably less attention (although some comparative studies on ethnic Germans have differentiated immigrants by their country of origin, such as Constant, Gataullina & Zimmermann 2006;Kogan 2011;Luthra 2013;Seifert 1998).Poles, in particular, had dominated immigration inflows to Germany in the years prior to 2014, before being overtaken by growing numbers of immigrants from Romania and Syria thereafter 5 .However, outside studies on ethnic Germans, Polish immigrants have not been prominently featured in the quantitative literature (several qualitative studies on seasonal migration patterns include Glorius 2008; Kepińska 2013 from White's 2016 literature overview).This contrasts with other EU countries, like the UK and Ireland, where Polish immigrants have received much more academic attention (both quantitative and qualitative, for example Drinkwater, Eade & Garapich 2009;Filimonau & Mika 2019;Krings et al. 2013;Mąkosa 2018).Thus, our analysis will provide some empirical findings on the experiences of a more recent cohort of female immigrants from Poland, contextualised by the experiences of a well-researched group in Germany, Turkish women.
Additionally, the comparison between these two groups will be particularly interesting with regard to the motivation of women to participate in the labour force.Poles and Turks come from two very different cultural backgrounds with different gender values regulating women's actions in the public sphere.Middle Eastern societies based on Islamic values are perceived to severely restrict women's opportunities to engage in activities outside of the household.Meanwhile, in most Western countries women's labour force participation is largely seen as acceptable.Since immigrants may bring culturally-specific gender values with them when they immigrate (Blau 2015), these could have an important effect on the motivation for women to join the labour force after their arrival.By comparing Polish and Turkish women, two groups with different cultural backgrounds, we hope to shed light on the influence that motivational factors have on female immigrant labour force participation in Germany.

5
Retrieved from As a final consideration, we focus specifically on the experiences of recently-arrived immigrants (who have been living in Germany for a maximum of 18 months).Studies on the early experiences of immigrants are rare, rendering it more difficult to disentangle post-migration integration dynamics from immigrant characteristics upon arrival (Schwartz 2005).This problem is particularly relevant when it comes to explaining the ongoing disadvantages of those ethnic groups whose integration appears to lag behind that of others (Kalter 2006).Moreover, it is during an immigrant's first months in the destination country that the decision to enter the labour force is most relevant.
Thus, using innovative data on the early phase of migrants' integration, collected in the project Causes and Consequences of Socio-Cultural Integration Processes of New Immigrants in Europe (SCIP; Diehl et al. 2015) we analyse if and why there are differences in the labour force participation of Polish and Turkish female immigrants.We not only consider the role of human capital endowments, the most important explanatory factor in immigrants' labour market integration, but we also investigate the role of additional resources such as language skills and social ties and more gender-specific motivational factors, such as the migration motive, family composition and cultural gender values.

Theoretical framework and existing findings
In line with findings from many other Western industrialised countries6 , female immigrants in Germany have widely been shown to experience disadvantages in their labour market integration.First-generation immigrant women are found to have lower labour force participation rates than native German women, regardless of their country of origin (Fleischmann & Höhne 2013).Those who do work, however, are more likely than native women to engage in full-time employment (Fleischmann & Höhne 2013), at least during the early years of their stay.Female immigrants also tend to hold jobs with lower occupational status than their native counterparts (Fleischmann & Höhne 2013;Kogan 2011) and earn lower wages (Bauer & Zimmermann 1997;Constant 2009;Reitz et al. 1999).Their wages show some weak assimilation, but do not totally attain the same levels of native women (Basilio, Bauer & Sinning 2009).These disadvantages exist not only between immigrant and native women, but have also been found between immigrant women of different ethnic groups on the German labour market (see, for example, Fleischmann & Höhne 2013).
When it comes to explaining these differences, individual labour market resources play an important role.This is true for both men and women.Most importantly, educational endowments and work experience have a strong impact on labour market integration as spelled out in human capital theory (Becker 1993), and empirical findings in the German context have confirmed this assumption (Granato & Kalter 2001;Seibert & Solga 2005).However, even after comparing individuals with similar levels of human capital, immigrant disadvantages often persist (Fleischmann & Höhne 2013;Kogan 2011;Granato & Kalter 2001).For first-generation immigrants, these disadvantages can be partly attributed to the fact that human capital acquired abroad is only partially transferrable to the new context (Chiswick 1978;Friedberg 2000;Kalter & Granato 2018), which is not fully captured by simply controlling for immigrants' human capital endowments from the country of origin.
This basic framework of human capital theory has been extended to include additional factors relevant to immigrant integration and the process of transferring human capital endowments.Host country-specific resources are important to account for, like proficiency in the majority language (see Chiswick & Miller 2002) and the social capital that comes with the establishment of a social network (Putnam 2000).For Germany, many studies have confirmed that migrants with a high level of German proficiency perform remarkably better on the labour market than those with little knowledge of the language (Lancee & Hartung 2012;Hirsch et al. 2014;Koopmans 2016).Similarly, studies show that social ties to German natives in particular increase not just the probability of employment, but also the occupational status of the job opportunities found through these channels (Koenig, Maliepaard & Güveli 2016;Kalter & Kogan 2014;Lancee & Hartung 2012;Kanas, van Tubergen & van der Lippe 2011).Social contacts turn out to be equally important for male and female immigrants during the job hunt (Kanas, van Tubergen & van der Lippe 2011), although language proficiency may be more important for females than for males in terms of their labour force participation (Koopmans 2016).
While all these components of immigrants' labour market resources are important when studying both male and female migrants, the motivation to work becomes particularly important when focusing on female migrants.Males are traditionally considered the main breadwinner in the household, and they seldom grapple with the question of entering the labour market and searching for employment after migration (see Bürmann, Haan, Kroh & Troutman 2018;Haan, Kroh & Troutman 2017).When analyzing female migrants' labour market integration, on the other hand, their selfexclusion from the labour market may play an important role.This could partly be due to cultural norms about women's participation in economic activities outside the household, but also includes their perception of success on the labour market.
An individual's motivation to work may be indicated by their initial migration motive.This is particularly the case for labour migrants -those who immigrate for work related reasons have a high probability of looking for work after arrival.In some cases, particularly immigrants from outside of the EU, a job offer is required before a visa is issued7 .For other types of immigrants, the channel of migration says less about their labour market intentions.Often it is assumed that women who immigrate for family reasons have a very low motivation to enter the labour market.While several studies have shown that this is indeed the case (Ballarino & Panichella 2018;Basilio, Bauer & Sinning 2009;Krieger 2020), it should also be acknowledged that channels of migration to Germany are heavily restricted for immigrants from certain origins.For example, even for economically motivated migrants from outside the EU, the only option to legally enter Germany may be as a family migrant.These restrictions may divert a wide range of individuals with different motives and characteristics onto a similar path.
When it comes to the impact of cultural gender values on labour force participation, factors like marital status and the presence of children come into play.The distribution of tasks within a household often means that partnered women participate less in the labour force than single women, and the presence of children also negatively affects female migrants' labour force participation rates (Fleischmann & Höhne 2013).Depending on cultural norms and the availability of child care, these factors can have different effects on the labour market behaviour of females from different countries of origin, and thus are important to take into account when studying female immigrants' motivation to enter the labour force.
Several studies for Germany have examined the role of culture and gender values in immigrant women's labour market integration.The most direct way to investigate this relationship is by using survey data that includes this type of indicator.For instance, Koopmans (2016) shows that Muslim migrant women in several European countries (Germany included) with high resource endowments and liberal gender values have similar labour force participation rates to native women.An equivalent effect of sociocultural factors is not observed for males, which reinforces the gendered aspect of this finding.Unfortunately, relatively few surveys include this type of question.Researchers often resort to using proxy variables instead when studying this relationship.
One proxy variable that is frequently used in place of unobserved gender values is religiosity.Research has suggested that there is a connection between strong individual religious beliefs and conservative gender values (Diehl, Koenig & Ruckdeschel 2009).Connor & Koenig (2015) analyse the impact of individual religiosity for a sample of immigrants in several Western European countries.They do not find that self-reported religiosity has an impact on labour market integration, though in this case, their sample includes both men and women and their model is not specified to investigate the possible gender differences that this variable could have.Knize Estrada (2018), by contrast, finds evidence for female immigrants in Germany that an individual's level of religiosity, regardless of her confessional faith, is highly correlated with her labour force activity; i.e., immigrant women who pray less often have higher probabilities of working.
On the topic of religiosity, the specific confessional faith (Muslim, Christian, etc.) has also been proposed as another viable proxy for measuring gender values (Höhne & Koopmans 2010;Knize Estrada 2018).These studies suggest that, while religiosity is correlated with conservative gender values, Muslim religiosity is particularly conservative.Indeed, Muslim religiosity was found to have a strong negative impact on female immigrants' employment and their hours worked in Germany, compared to non-religious female immigrants (Köbrich Leon 2013; Knize Estrada 2018).It was also found to increase their unemployment duration relative to individuals with a different confessional faith (Höhne & Koopmans 2010).
While the abovementioned findings on gender values provide important insights into the role of culture in labour market integration, and demonstrate the validity of the specific intensity and/or type of religiosity as a proxy for these values, there are important drawbacks in the ways that these studies have investigated this mechanism.Importantly, for both Köbrich Leon (2013) and Knize Estrada (2018), the outcome variable is operationalised as employment and/or hours worked, which differentiate between individuals who were successful at finding employment and those who were not, and not between individuals who were and were not motivated to enter the labour force.But it is precisely this individual decision to join the labour force in search of paid employment that should be most influenced by gender attitudes, not their success in finding a job and establishing some sort of work schedule.Thus, based on the current state of the literature, it would appear that the relationship between gender values and labour force participation requires more precise research.
In addition to gender values, female migrants' motivation to join the labour force may be influenced by their perceived chances of success in finding suitable work.It is in this context that perceptions of ethnic discrimination need to be taken into account, a variable which is included in many surveys tailored to immigrants.While several articles have recently pointed out that these perceptions should not be taken as an accurate indicator of actual experiences of discrimination (Diehl & Liebau 2017;Small & Pager 2020), individuals may nonetheless be discouraged from starting the job search if they perceive that their group's access to the labour market is hampered by discrimination.
In sum, these indicators of the motivation to participate in the labour force, including gender values and perceived ethnic discrimination, are important to consider when studying female labour market integration.The findings above show that these indicators supplement the factors referenced in human capital theory and can help to explaining the remaining nativity or ethnic gaps for female immigrants.

Poles, Turks, and the German labour market
Poles and Turks represent a large share of the immigration flow into Germany, making up 14 percent and 2 percent of the total inflow respectively in 2014 (Bundesamt für Migration und Flüchtlinge 2016).Turks were the largest immigrant group (in terms of arrivals) from 1971 to 1980, while Poles were the largest group from 1983 to 1990, and again from 1994 to 2013 8 .In the post-war period, Turkey was an important source of economic migrants, who were recruited to work in low-skilled industrial occupations in Germany.They arrived as 'guest workers' with the assumption that their stay would be temporary, though return migration to Turkey after recruitment policies ended turned out to be very limited.After the end of Germany's foreign worker recruitment policy in the early 1970s, immigration patterns shifted focus to family reunification.To this day, this remains an important channel for Turkish immigration to Germany, accounting for 27 percent of Turkish arrivals in 2017 (Bundesamt für Migration und Flüchtlinge 2019).The overall number of Turks coming as students has also steadily risen over the past two decades, from just 747 in 1999 to 3517 in 2017 (Bundesamt für Migration und Flüchtlinge 2019).
Though they might not be as 'negatively selected' in terms of their human capital as the early guest workers, the recent arrivals from Turkey continue to bring with them low human capital endowments.About 56 percent of the Turkish male immigrant population in Germany has completed some sort of occupational training or tertiary education, while this number is only 30 percent for Turkish female immigrants (Schührer 2018).This being said, the German language skills of those arriving after 2000 have been rather high on average, especially compared to the language skills of those who arrived during the 1960s and 1970s (Schührer 2018).
Since the Turkish community in Germany has been built up from the original guest workers, it has strong ties to the lower skilled sectors in the German economy (Kogan 2004) and has created an ethnic niche market addressing the needs of its own clientele.Turks in Germany have comparatively few inter-ethnic friendships, even in the second generation, and even compared with Turks in other European countries (Crul & Schneider 2012).This point is important, because studies show that Turks' co-ethnic contacts have little impact on their unemployment durations, but that contact with native Germans greatly increases their chances of finding employment (Lancee & Hartung 2012).
Along with human capital, immigrants bring cultural capital with them from their country of origin.First-generation Turkish migrants were socialised in Turkey, which has a potential impact on their labour market integration in Germany.Patriarchal norms still hold sway in Turkish society, meaning there is little public support for female-headed households and a high acceptance of the view that women have to have children to feel fulfilled (Dildar 2015).The labour force participation rate of married women in Turkey (as well as those with children) is lower because of both societal pressure to stay in the home and a lack of affordable child care (World Bank 2009).From a low of 25 percent in the early 2000s, the female labour force participation rate has since been on the rise, reaching 30 percent in 2010 and 38 percent by 2018.The low levels of female employment outside the household may also be related to the fact that Turkey is a highly religious Muslim society -the importance of religion has been increasing in Turkish society since the 1990s (Dildar 2015), and about 84 percent of Turks consider themselves religious9 .
Turning to Poles, the current community in Germany mainly arrived in the postcommunist period, either as ethnic Germans or students (Bundesamt für Migration und Flüchtlinge 2010).Circular immigration from Poland has also been an important source of labour for low-skilled sectors in Germany, such as agriculture, with many migrants working in Germany temporarily and returning to Poland after a certain financial target has been reached (Kepińska 2013).Since Poland's accession to the European Union in 2004, Poles have received new rights to freedom of movement within the wider European Union, although Germany continued to restrict labour migration from Poland until May 2011.
Unlike the cyclical labourers, data from a representative investigation of current immigrant groups in Germany shows that Poles who arrived between 2010 and 2015 had fairly high levels of education: 59 percent reported they had completed some sort of vocational training (Berufliche Ausbildung) and an additional 25 percent had completed tertiary education (Babka von Gostomski 2016).Due to European schemes for educational standardisation, the educational qualifications they bring with them from Poland are generally recognised in the German labour market.This being said, their level of German proficiency seems to be slightly lower than the Poles who arrived before 2010 (Babka von Gostomski 2016).
The Polish community is not as prevalent in German society as the Turkish community, since earlier cohorts who arrived as ethnic Germans were met with favourable policies and quickly integrated into the majority population (Kreyenfeld & Konietzka 2002).More recent immigrants have benefited from freedom of movement within the European Union, which has facilitated circular immigration for groups who live geographically close to Germany.Thus, Poles have fewer opportunities in ethnic job markets, but are also less likely to experience the segregation that comes with this.
Like Turks, Poles come from a political and social context where religion continues to play an important role.Catholicism has been an important part of Polish national identity since the period of communist rule, when it was seen as a source of unity and resistance against the state regime (Requena & Stanek 2014).Today, 67 percent of Poles declare they are practising Catholics (Requena & Stanek 2014).The country also has a tradition of female labour force participation, which was supported by the communist regime and viewed as essential to the labour supply for the country's rapid industrialisation (Kepińska 2013).This norm continues to this day: the female labour force participation rate in Poland was 59 percent in 2010, and had reached 63 percent by 2018 10 .After the collapse of this regime, however, there was some tendency to revert to conservative family structures.Data from the early 2000s shows that 67 percent of women and 77 percent of men favoured traditional family gender roles (Titkow & Duch 2004, from Kepińska 2013).Thus, it appears that there are multiple societal expectations for women to both be active in the labour force and to undertake the traditional tasks in the household.Especially in the older age groups, women have been found to allocate much more time to household tasks than men (Siemieńska 2008).These characteristics have also been observed in the group of Poles who choose to immigrate to Germany, at least in terms of their high religiosity (Diehl & Koenig 2013) and their participation in gender-typical sectors of the German labor market (Kalwa 2007).
The immigrant groups in our study accessed the German labour market in 2009/2010, during a period characterised by low unemployment, and even labour shortages in many areas.This typically renders ethnic discrimination more costly for employers than in contexts where unemployment is high, and should be beneficial for newcomers' labour market integration.Nevertheless, research has shown that entering the higher strata of the German labour market in particular is difficult for outsider groups  (Luthra 2013(Luthra : 1105) ) due to high wage floors and strong employment protections (Kogan 2006: 699).Thus, newcomers on the labour market may face challenges that are related to structural aspects of the German labour market (Kogan 2006;Luthra 2013) rather than to individual discrimination.

Expected results
So far, we have discussed the most important factors shaping first-generation immigrants' labour market integration, and have contextualised the immigrant groups we include in our analysis.We have shown that an analysis that seeks to explain the group differences in female immigrants' labour market participation needs to take into account both the labour market resources that immigrants possess, and their motivation to enter the labour market.
In our subsequent analyses, labour market resources should exert a strong influence on the decision to participate in the labour market.Given the lower endowments of Turks with respect to their level of education, and prior work experience, German language proficiency, and ties to natives, we expect to find an initial gap between the labour market outcomes of Polish and Turkish migrant women, which declines after controlling for these resources.
Regarding motivational factors, we expect the ethnic gap in labour force participation to further decrease once indicators for their motivation to find paid employment outside the household are accounted for.Since administrative reports show that for Polish women, work is an important motive for their immigration to Germany, including an individual's migration motive and family status in our models should address many of the compositional differences between the two ethnic groups.Similarly, perceived ethnic discrimination should reduce the motivation to join the labour force for both groups (even though more Turkish than Polish women may expect to experience discrimination on the labour market).
In a similar vein, the strength of individual religiosity should have an overall negative impact on female immigrants' labour force participation.Since previous findings have suggested that Muslim religious affiliation in particular is connected to conservative gender norms, we expect that religious Turkish women are even less likely to be active in the labour force than religious Polish women.Accounting for the differential impact of confessional faith should reduce the ethnic gap in labour force participation even further.

Data, variables, and analytic strategy
The data for this analysis comes from the NORFACE-funded Causes and Consequences of Socio-Cultural Integration Processes Among New Immigrants in Europe (SCIP) project, which was conducted between 2010 and 2013 in four European countries (Germany, the Netherlands, the UK, and Ireland).This survey targeted newly-arrived immigrants (living in the destination country for 18 months or less) from different ethnic backgrounds.In the case of Germany, this dataset provides a rich overview of demographic and sociocultural variables for Polish and Turkish immigrant groups living in five large German cities (Berlin, Bremen, Cologne, Hamburg, and Munich).Face-to-face surveys were conducted for all respondents during the first wave, which had a response rate of 60.5 percent (Gresser & Schacht 2015).Using data from the first wave of the German survey, our analysis explores the labour market situation of female immigrants who had been living in Germany for a maximum of 18 months.Our sample includes female respondents aged 18-60 who were either active in the labour force or 'at risk' of entering (thus excluding respondents who were students, disabled, or already retired at the time of the survey, for whom other factors restrict their labour force participation).Overall, our sample comprises 829 women (490 Polish women and 339 Turkish women).Since we are particularly interested in exploring the role of motivational factors in female labour market integration, the outcome variable in our analysis is labour force participation (at the time of interview).We group together those who were employed11 and unemployed as 'active' in the labour force and designate homemakers as 'inactive'.It could be argued that in some instances, it may be problematic to assume that the unemployed are more motivated to work than homemakers.After all, it may be attractive to register as unemployed even for those individuals who are not actually searching for work if this comes along with access to unemployment benefits.However, we believe that for our specific sample, these incentives are less relevant because recently-arrived immigrants in Germany are not eligible for unemployment benefits.
To account for individual human capital endowments, we include the immigrant's pre-migration work experience and their level of education.Work experience from the immigrant's country of origin is condensed into a dichotomous variable indicating whether the respondent had any prior work experience or not.The educational endowment from the country of origin is aggregated into an ordinal variable with three levels, where individuals who failed to complete more than lower secondary education are placed in the 'low' group, those who completed upper secondary education are assigned to the 'middle' group, and those who completed tertiary education are placed in the 'high' group.
To take into account the labour market capital specific to Germany, we include the individual's self-assessed German language proficiency and the presence of close German friends.Language proficiency is based on a mean index of a respondent's level of reading, speaking, writing, and comprehension skills, in which immigrants are classified as having no German proficiency, some proficiency, or good proficiency.The presence of close German friends is reduced to a dichotomous variable, with those who reported one or more close friends born in Germany being assigned in one category, and those who specify zero close German-born friends being assigned to the reference group.
To look at the impact of motivational factors, we include the variables migration motive, family status, religiosity, and perceived ethnic group discrimination.Migration motive distinguishes between those who came for work reasons and those who came for other reasons.Family status reflects a migrant's marital status and the presence of children in the household, coded either as 'single', 'partnered, no children' or 'partnered with children'.Additionally, the respondent's level of religiosity is included as an approximation of her gender values, based on previous findings that high levels of religiosity are positively correlated with conservative gender values (Diehl, Koenig & Ruckdeschel 2009).Responses are coded into 'rather non-religious' for those who were not religious or had low religiosity, and 'rather religious' for those with moderate to high religiosity.Perceived discrimination is an ordinal variable that differentiates between respondents who were unaware of or perceived little ethnic discrimination, those who perceived a medium frequency of discrimination, and those who perceived high frequency of discrimination.
Table 1 presents descriptive statistics for the dependent variable and the covariates included in the analysis.This table shows that the group of Polish women in our sample is slightly older than the Turks, with an average age of 34 (compared to 30 for the Turks).In terms of human capital, a much higher percentage of the Polish women arrived with work experience from their country of origin than the Turkish women, and on average they had a higher level of education.This is in line with the findings on contemporary migrants from Poland and Turkey mentioned in the preceding sections.Looking at host country capital, we see that neither group had a high level of German proficiency, and few women from either group had a close German friend.Turning to motivational factors, we see that many of the Polish women came for work reasons (58 percent), while the vast majority of the Turkish women (93 percent) came for other reasons (mostly joining family).As was previously mentioned, the legal pathways for Turks to immigrate to Germany are rather restricted, so their official migration motive does not always reflect their working intentions.Over half of Polish women were single, while those who were married are split rather evenly between households with and without children.For Turkish women, very few were single, and many were found in childless households.Looking at levels of perceived group discrimination, we see that the Poles actually perceived more ethnic discrimination than the Turks, with 59 percent indicating 'medium' or 'high' levels of discrimination (compared to 47 percent of the Turks).Moreover, the Poles tended to be more religious than the Turkish women in the sample 12 .
Given that our dependent variable is dichotomous, we use logistic regressions for our analysis.We start with a baseline model controlling only for age and ethnicity, and by adding our covariates -relating to resources and motivational factors -in a stepwise fashion.

Descriptive analysis
If cultural gender values do indeed impact females' motivation to participate in the labour force, then significant differences may already be visible in female immigrants' main economic activity before their arrival in Germany.Figure 1 presents the economic status of female migrants in Germany at the time of the survey, based on their economic status in their country of origin13 .While our analytical sample is restricted to those who were employed, unemployed, or homemakers at the time they were interviewed, these respondents may have occupied a different place in the labour market before their arrival.Those women who were previously students or retired in their country of origin, but in Germany find themselves employed, unemployed, or homemakers, are identified in the 'other' column.

Figure 1:
Activity status at the time of the interview by activity status in the country of origin The height of the bars in Figure 1 shows the share of the group in each economic activity before migration.A large percentage of the Polish women were economically active before migrating to Germany (60 percent were employed and another 16 percent were looking for work).Conversely, a rather small proportion were homemakers in Poland (only seven percent) and an additional 17 percent were occupied with some other activity (mostly students).Looking at the Turks, we see that women who were previously economically active again represent the largest share of the group (46 percent were employed and 25 percent looking for work), although compared to the Poles, the Turkish women were somewhat more spread out across all the different economic statuses.About 19 percent of them had previously been homemakers in Turkey (more than double the share of Polish homemakers) and nine percent had been occupied with some other activity.
The colour of the bars in the Figure 1 reveals the activity of these women after their arrival in Germany.The yellow areas indicate that a large majority of the Polish women were employed, regardless of their status pre-migration.About 66 percent of the Polish women were employed in Germany, and another 14 percent were unemployed (designated by the purple areas).The Turkish women, on the other hand, were particularly likely to be unemployed (50 percent) or homemakers (30 percent, as shown by the green areas) in Germany, regardless their prior economic status.
This figure reveals that the Polish and Turkish women in our sample did not differ substantially in their labour force participation -or at least, not as much as we might expect given the data on their counterparts in their respective countries of origin.The data indicates that 80 percent of the Polish women in our sample were active (employed or unemployed) in the labour market before migrating to Germany, as were 71 percent of Turkish women.Thus, both groups had much higher rates of participation prior to coming to Germany than their respective national levels (59 percent in Poland and 30 percent in Turkey in 2010, mentioned above).Even when taking into account that we excluded students, the retired, and the disabled (at the time of the survey) from our analysis sample, this points towards a process of positive immigrant selection.In the multiple regressions below, we examine the determinants behind these group differences in labor force participation.

Regression analyses
The results of logistic regressions of female immigrants' labour force participation are presented in Figure 2 as average marginal effects.Controls for age, age squared, and the region of residence are included in all models, but since these variables are not the focus of our analyses, we do not present the coefficients here (full tables of results can be found in Appendix II).The baseline model confirms that even after controlling for age and place of residence, the Turkish women were less likely to participate in the German labour force than the Polish women.The resource model shows that relevant labour market capital − like previous work experience, a good knowledge of German, and social contacts with members of the majority population − all increased the likelihood of participating in the labour force.Taking these factors into account decreases the ethnic gap (by almost seven percentage points) and eliminates its statistical significance.

Figure 2:
Predictors of recent female migrants' labour force participation in Germany (logistic regressions -AMEs) Note: Point estimates are presented with 95 percent confidence intervals; all models control for age, age 2 , and region of residence The final models look at the impact of motivational factors on the probability of participating in the labour force.Motivation model A shows that motivational factors do hold considerable explanatory power for labour force participation.In fact, the estimate for ethnicity, which had indicated a disadvantage for Turks in the baseline model, now indicates a slight advantage (although this is not statistically significant at the 95 percent level).As expected, women who came for the purpose of working have a much higher probability of participating in the labour force.Women who were married and had children in the household had a much lower probability of being active in the labour market, another finding consistent with our expectations.Perceived group discrimination, on the other hand, did not seem to keep women from joining the labor force, while religiosity did indeed have a negative (and statistically significant) impact on female labour force participation.
Motivation model B investigates the interaction between ethnicity and religiosity to determine whether the religious Muslims differ from the religious Christians in their labour market participation.Here, we combined the two independent variables into an "ethno-religious" categorical variable to better compare the different groups.We see that, compared to the non-religious Turks, the religious women had a lower likelihood of participating in the labour force (regardless of their ethnicity), whereas the non-religious Poles did not statistically differ from non-religious Turks.

Discussion
In the analysis above, we have looked at the early stage of female immigrants' labour market integration process; i.e., their decision to participate in the labour market after immigrating to Germany.We have argued that an immigrant woman's decision to join the labour force is more deliberate than an immigrant man's (as men have very high employment rates from early on, see Sprengholz et al. this issue).Therefore, when studying female immigrants' labour market entry after their immigration, it is important to take motivational factors into account along with their resources, since both of these dimensions likely contribute to their decision.
Our analysis reveals a slight disadvantage for the Turkish women compared to the labour force participation rates of the Polish women, although the difference was not as big as might be expected based on the characteristics of earlier cohorts.In line with our expectations, we find that differences in both groups' labour market resources (particularly work experience and German language skills) play an important role in explaining the gap between the two groups.Once these are taken into account, the groups no longer differ in terms of their labour force participation.Also in line with our expectations, we find that motivational factors like family composition and religiosity are both significant determinants of the labour force participation decision, as is the migration motive.After accounting for motivational factors, the former disadvantage in labour force participation rates actually turns into a slight (albeit statistically insignificant) advantage for Turkish women, confirming that these factors are important for clarifying the broader picture.This could be an indication that selection processes are at work related to the different channels of migration for Poles and Turks.The descriptive table presented earlier in the article showed that very few of the Turkish women arrived as labour migrants, although those who did arrive through this channel presumably had a job lined up before their arrival (as a condition for receiving their work visa).This rigorous selection process created a more specialised pool compared to the Polish women who also came for work reasons, and who may not have had the same attachment to the German labour market before their arrival.Meanwhile, when we take a closer at the other side of the migration motivation variable, we see that a large number of women who declared that they migrated for family reasons were nevertheless active in the German labour force.In fact, cross-tabulations not shown here reveal that 68 percent of the Turkish family immigrants were active in the labour force after their migration, compared to 59 percent of the Polish family immigrants.
As was previously mentioned, the working intentions of non-labour migrants are difficult to estimate from their channel of entry because the labour migration channel is so restrictive.In prior research, groups like family immigrants have been considered less motivated to find employment outside of the household after migration (Ballarino & Panichella 2018).However, based on the groups we consider, this assumption does not hold.We observed earlier that the Turkish women in our sample, though mainly family migrants, had higher rates of labour force participation at home in Turkey than their nonmigrant counterparts.Thus, the assumption that these women hold conservative gender values similar to those of the general Turkish population may not be valid.Since the decision to immigrate is (largely) a self-driven process, the individuals who end up immigrating likely differ from the rest of the population on both observed and unobserved characteristics -something our results suggest.By using data on recent immigrants, we could take these dynamics at least partly into account.
The findings from our final model refute the idea that Muslim religiosity has a more negative impact on labour force participation than Christian religiosity.Although the intensity of religiosity seems to correlate with lower labour force participation, it does not appear that the religious Muslims in our sample are any less likely to be economically active than the religious Christians.This finding contradicts the oft-cited argument that Muslims (and especially Muslim women) are uninterested in integrating into European society in general, and in the labour market in particular.Both Knize Estrada (2018) and Köbrich Leon (2013) reported a negative impact of Muslim religiosity on female economic activity in Germany, but because their dependent variables were conceptualised around employment, this may partially capture discrimination by the employer (and not just selfexclusion).
Turning to discrimination, we acknowledge that numerous audit studies have exposed the labour market discrimination against Turks in Germany ( Quillian et al. 2019;Weichselbaumer 2019;Kaas & Manger 2011) and the greater social distances that Turks face compared to Western European immigrants (Diehl, Blohm, & Degen forthcoming).While discrimination may play a role in their employment rates 14 , our results do not indicate that this is an important factor in explaining immigrant women's labour force participation.Our analysis show that the ethnic gaps in this outcome reflect mostly resource endowments, family status, and migration channel; and that these gaps disappear or -in our case -even reverse, once these variables are taken into account.Even with respect to a more subjective indicator, namely perceived ethnic discrimination, there is no evidence that it hampers female migrants' motivation to join the labour force.Needless to say, we only studied female migrants' decision to join or not to join the labour force, and not their motivation to apply for certain jobs.We cannot exclude the possibility that perceived discrimination reduces the chances that Muslim women would apply for certain jobs; e.g., those that involve contacts with customers.
Overall, our empirical results supplement the picture shown in the descriptive analysis.They underline the fact that recent cohorts of Turkish women are quite similar to Polish women in terms of their labour force participation.The presence of children in the household appears to be one of the remaining barriers to the labour force participation of all immigrant women (something which has also been noted by Samper Mejia & Kreyenfeld, this issue), which highlights the need to communicate the availability of childcare to recently-arrived immigrants.To be sure, this study was limited to the early stages of a female immigrant's labour market integration, and these findings should be supplemented by research on how this process unfolds over the years after arrival; as well as by considering additional labour market outcomes like employment (for those who are active on the labour market), occupational status and part-time work.As far as women are concerned in this aspect of labour market integration, our analyses suggest that for the more recent immigrants from Turkey, the legacy of disadvantage from earlier cohorts and the often described 'exceptional' challenges that this group faces on the labour market, may slowly be fading away.

Table 1 :
Descriptive statistics of recently-arrived female migrants (column percent)

Table 1 :
Descriptive statistics of recently-arrived female migrants (column percent) (continued) Percentages may not add up to 100 because of missing values.These cases are filtered out of the ensuing quantitative analyses. Note: