Religion and new immigrants' labor market entry in Western Europe*

This paper analyzes the effects of religious participation upon a major socio-economic integration outcome, namely employment, among recent Christian and Muslim newcomers in three Western European destination countries: Germany, the Netherlands, and Great Britain. The paper revisits theoretical arguments about religious participation as an ethnic investment strategy or, alternatively, as a bridge to the societal mainstream. Drawing on the longitudinal dataset produced in the international survey project on ‘Socio-cultural Integration Processes among New Immigrants in Europe’ (SCIP), the paper puts these arguments to a rigorous test by analyzing effects of involvement in religious communities on employment and by scrutinizing channeling effects of the ethnic composition of religious congregations for recent migrants’ entry into mainstream versus ethnic niche economies. The paper finds only limited support for either of the two arguments, suggesting that religious participation is structurally decoupled from socio-economic integration. However, persisting net employment gaps between recent Christian and Muslim immigrants might indicate the existence of religiously marked and socio-economically consequential boundaries in Western Europe.


Introduction
The academic literature on immigrant integration has increasingly paid attention to religion. Adding to previous work by qualitative researchers, quantitative scholars have begun to explore trends in religious beliefs, practices, and identities among migrants and their children, and to study their relation to socio-economic integration across various contexts of reception (for review see Breton, 2012;Voas and Fleischmann, 2012). In Western Europe, this emerging scholarly interest in religion is largely driven by public debates over Muslim migrants, their persistent socioeconomic exclusion, and their increasing public and political visibility -debates that continue to be premised on severely under-theorized and empirically untested assumptions about putative effects of religion upon immigrant integration trajectories.
The most prominent attempt to systematize such effects has been developed by Foner and Alba (2008) in their well-known transatlantic comparison of academic approaches to the topic. They suggest that while religion was typically regarded as ''barrier'' in Western Europe, it was rather perceived as a ''bridge'' in North America. Indeed, various studies from North America have documented that participation in religious community life enhances migrants' well-being, helps building social ties and even provides educational and labor market advantages for migrants and their offspring, thus operating as a bridge to the mainstream (Bankston and Zhou, 1996;Connor, 2011;Portes and Rumbaut, 2006;Reitz et al., 2009). In contrast, studies in the European context have rather focused on ethnoreligious ''penalties'' for the first and the second generation, in terms of educational attainment and labor market performance, suggesting that religious affiliation operates as a barrier to the mainstream (Adida et al., 2010;Cheung, 2014;Connor and Koenig, 2015;Heath and Martin, 2013;Johnston et al., 2010;Khattab, 2009;Lindley, 2002;Luthra, 2013;Model and Lin, 2003;Silberman et al., 2007; for a critique see Koopmans, 2015).
However, the effects of religion upon migrants' socio-economic integration require further scrutiny in Western Europe. Due to the strong focus on the exclusion of categorical groups defined by religious affiliation, hardly any attention has been paid to the possible effects of organized religious practice. Moreover, the highly selective attention given to Muslim immigrants and their children has left largely unexplored the performance of other religious groups, notably Christians who, as a matter of fact, constitute the major religious group among contemporary migrants coming to the European Union (see Pew Research Centre, 2012: 54). Last but not least, there is next to no research on religious dynamics in early stages of immigrant settlement. Compared with North America, where scholars have scrutinized religious as well socio-economic dynamics among newly arriving immigrants (e.g. Akresh, 2011;Connor, 2009Connor, , 2011, the highly dynamic period of newcomers' initial adaptation to the receiving society remains particularly ill-understood in European research. In this paper, we take up some of these unresolved issues by analyzing the effects of religious participation upon labor market entry among recent newcomers in Western European countries. Given that for newcomers finding jobs is of more immediate urgency than occupational status, our analytical focus is on employment. We develop theoretical arguments about group-specific effects of migrants' religious involvement for finding jobs as well as for creating channels into mainstream or ethnic niche segments of the labor market. To test these arguments, we study Christians and Muslims in three destination countries: Germany, the Netherlands, and Great Britain. Drawing on a unique longitudinal dataset produced in the international survey project on ''Socio-cultural Integration Processes among New Immigrants in Europe'' (SCIP), we are able to test whether religious community involvement hinders or helps recently arrived migrants finding employment and whether the ethnic composition of religious communities channels them into different segments of the labor market.

Theoretical background
As shown by a large literature, first-generation immigrants' entry into the labor market is associated with a range of factors (for an overview see Heath and Cheung, 2007). These factors notably include pre-migration human capital, e.g. formal education and work experience (see e.g. Chiswick, 2003;Van Tubergen et al., 2004); socio-cultural factors, such as the weakening of ethnic identities, the development of language capacities and inter-ethnic social ties (see Bisin et al., 2011;Kanas et al., 2011;Koopmans, 2015: Lancee, 2012; and receiving society characteristics such as labor market regulations, welfare regimes and citizenship models (see e.g. Kogan, 2006;Pichler, 2011). In Western Europe, recent studies have also suggested that first generation Muslim immigrants face ethnoreligious penalties and catch up only slowly compared to other immigrants (Connor and Koenig, 2015;Heath and Martin, 2014). However, less attention has been paid to the question whether and how active involvement in religious communities impinges upon employment in various segments of the labor market. In the following, we contribute to theory-building in this area by drawing on recent attempts to reformulate the aforementioned metaphor of ''barriers'' and ''bridges'' (Connor and Koenig, 2013). We start by spelling out two theoretical arguments about potential disadvantages and advantages of migrants' religious participation for finding jobs. Subsequently, we spell out the two arguments' implications concerning the effects of religious participation for channeling migrants into mainstream or ethnic niche segments of the labor market.
The first line of argumentation, which resonates with much Western European scholarship on immigrant integration, is based on conventional assumptions of assimilation theory and treats religious practice mostly as an indicator of ethnic orientations that are perceived as detrimental to integration into mainstream society (see e.g. Esser, 2004). Scholars have long emphasized that first generation immigrants often establish ''ethnic churches'' (Breton, 1964) which provide the institutional infrastructure for maintaining origin-country related identities, practices and social ties (see Gu¨veli, 2015;Mol, 1979). Individuals' participation in religious communities is in this perspective primarily seen as an ethnic investment strategy; as such, it might distract from efforts to acquire language competence or to establish social contacts with majority members, thus preventing newcomers from finding employment. In short, religious participation is here assumed to hinder entry into the labor market among new immigrants. As newer versions of assimilation theory stress (e.g. Alba and Nee, 2003), the extent to which individuals maintain ethnic orientations and the ways in which such orientations relate to socio-economic integration are conditioned on group-specific factors such as the existence of symbolic boundaries. Among religious minorities facing ''bright'' symbolic boundaries, religious practices should be particularly detrimental to finding jobs as they provide strong signals of minority membership and may thus trigger mechanisms of social closure.
Despite their intuitive appeal, conventional theories of assimilation have been subject to severe criticism. Not only do they unnecessarily reduce assimilation or integration to a unilinear process, they also too easily treat religious practices, beliefs and identities simply as an aspect of ethnicity (see Mitchell, 2006). It is, therefore, worthwhile to explore a second line of argumentation which emphasizes the distinctive feature of organized religion more directly. Individuals' participation in religious organizations has been argued to provide access to what Hirschman (2004) famously dubbed as three R's (refuge, respectability, and resources) which, in turn, may foster socio-economic performance. These resources include organizational support in language courses, formal education and actual job searches as well as access to social capital, notably, weak ties with the native population which, unlike ''bonding'' ethnic ties, contain non-redundant information and thus may facilitate entry into the labor market (Lancee, 2012: 27). These mechanisms which indicate that religion can function as a ''bridge'' to the mainstream have been amply documented in the form of case studies for the US (e.g. Bankston and Zhou, 1996), although quantitative evidence remains mixed (see notably Connor, 2011). In the Western Europe, ethnographers have recently assembled initial evidence that multipurpose centers of migrants' religious communities may have bridging functions, too (Baumann, 2014: 118-119). In short, this line of argumentation suggests that religious participation helps newcomers enter the labor market. It is conceivable, however, that the bridging mechanism operates in group-specific ways. Although proponents of segmented assimilation theory have argued that regardless of their majority or minority status religious communities strengthen migrants' sense of respectability and motivation and thus foster socio-economic advancement (e.g. Warner, 2007), there are good reasons to assume that bridging social and cultural capital inheres in religious majority organizations only (see Connor and Koenig, 2013).
When turning to channeling effects of religious participation, both lines of argumentations turn out to be slightly less contradictory than they initially appear. Assimilation theory, to start with, states that while ethnically oriented immigrants may actually find jobs, these jobs -whether in small firms, family businesses or selfemployment -tend often to be part of an ethnic niche economy which may, in the long run, be detrimental to socio-economic mobility (see also Wiley, 1967). Religious participation would on this account channel immigrants into an ethnic niche economy; more specifically such channeling effects should be observable among those immigrants who participate in largely co-ethnic religious communities. The bridging argument, in turn, focuses on ethnically mixed religious communities, or ''multiracial congregations'' which have received increasing attention in the US literature (see Edwards et al., 2013). Emphasizing the integrative functions of ethnically mixed congregations, the bridging argument implies that religiously involved immigrants may more easily find jobs on the mainstream labor market.
In our empirical analyses, we test these lines of argumentation by analyzing how religiously active and inactive newcomers find jobs in mainstream as well as ethnic niche economies segments of the labor market in three different receiving contexts. Going beyond previous studies, we compare Christian and Muslim migrant groups, while also taking into account the ethnic composition of religious communities. Before turning to our data and analyses in greater detail, we provide some elementary background information on the migrant groups and countries under consideration.

Contextual background -country and group characteristics
We study recent immigrants in Germany, Great Britain, and the Netherlands. As major destination countries for immigration in Western Europe, these three countries display a number of crucial similarities and differences. The three countries' migration history has followed a largely similar pattern in both political and socio-economic terms (for historical overview, see e.g. Lucassen, 2005). Post-war immigration from (semi-)peripheral countries such as Turkey and Morocco (for Germany and the Netherlands) or former colonies such as Pakistan (for Great Britain) or Surinam and the Dutch Antilles (for the Netherlands), gave rise to a sizeable, typically low-status immigrant population which since the halt to immigration after the oil crisis in the 1970s was replenished largely through family related migration. Eastern enlargement of the European Union set into motion new dynamics of labor migration, although Germany and the Netherlands protected their domestic labor markets slightly longer than Great Britain. New immigration from Eastern European countries such as Poland or Bulgaria today clearly surpasses immigration from Pakistan or Turkey which continues to be familyrelated while being complemented by student mobility and some high-skill migration. The three countries currently share roughly similar legal rules for immigrant employment. Thus, new migrants from Poland are immediately entitled to work, although it has to be noted that Germany required formal EU work permits from Polish citizens until May 2011. Newcomers from Pakistan or Turkey arriving as family migrants are entitled to work if their family members have a work permit or have already become citizens of the receiving countries. Those arriving on student visa, a particularly large amount of current Pakistani immigrants in Great Britain (see Platt and Luthra, 2016, within this special issue), typically have the right to work part-time and, after ending their study, to be employed for one year.
Of course, there are also noticeable differences between the three countries, not only in terms of immigrants' typical origin countries but also in terms of macrosocietal contexts. In 2010, for instance unemployment rates varied between 7.8% in Great Britain, 7.0% in Germany, and 5.0% in the Netherlands according to Eurostat data. In terms of labor market regulations, Great Britain stands out as the most liberal and flexible regime in Europe, whereas the Netherlands and Germany have more protective employment regulations which are known to reduce the chances of newcomers of finding employment (see Kogan, 2006;Lancee, 2012: 48; see also Luthra, 2013Luthra, : 1105. In terms of migration and integration policies, as captured for instance in the Indicators of Citizenship Rights for Immigrants, Germany stands out as a relatively restrictive context (ICRI score À0.14) compared with Great Britain (ICRI score 0.40) and the Netherlands (ICRI score 0.42) (Koopmans, 2013: 154). Clearly, it is impossible to test how such macro-differences affect integration trajectories with just three cases; following Alba's and Foner's (2014) recent suggestion, we, therefore, use country comparisons not for causal-analytical purposes but rather to sensitize for contextspecific dynamics (see also Gu¨veli and Platt, 2011). We analyze the three countries separately, and discuss any differences in light of relevant macro-characteristics.
Within all countries, we compare Muslim and Christian newcomers. The group of Muslims consists of migrants from Turkey, a major origin country of post-war immigration to Germany and the Netherlands, from Morocco and Surinam (in the Netherlands), and from Pakistan (in Great Britain). The group of Christians consists of migrants from Poland, the major origin country of post-Cold War labor migration and from the Antilleans or Surinam (in the Netherlands). Given their countries of origins, Muslim and Christian newcomers in our dataset differ strongly in terms of migration history and current legal status. Of crucial interest in light of the theoretical arguments, however, are their specifically religious characteristics. In all three countries, the majority population is still nominally Christian although the past decades have seen massive processes of secularization in terms of both religious affiliation and participation (see e.g. Voas, 2009). Muslims and Christians thus clearly differ in terms of their (religious) minority or majority status. For instance, Christian migrants from the Eastern Europe, when establishing ethnic churches, often found support from, or were even incorporated into, official churches in the host society thus providing closer links to the societal mainstream than provided by mosques for Muslim migrants. Crucially, although they provide different levels of Muslim accommodation, all three countries share considerable public hostility against Islam, notably since the 2000s (see Helbling, 2012); this attests to the presence of a bright boundary which Muslim migrants, unlike their Christian counterparts, confront upon arrival.
Against this background, we can formulate group-specific expectations in terms of the above-mentioned theoretical arguments. Following assimilation theory, one would expect that religious participation, notably for Muslims, hinders migrants from finding employment and, notably for those attending co-ethnic congregations, channels them into ethnic niche economies, net of human capital and socio-cultural assimilation. Following the bridging argument, one would expect that religious participation, notably for Christians, helps migrants finding employment and channels those who participate in co-ethnic religious communities into the mainstream labor market.

Data and methods
In our empirical analyses, we use data from the international survey project on ''Socio-cultural Integration Processes among New Immigrants in Europe'' (SCIP) that was funded by the NORFACE Research Programme on Migration (Diehl et al., 2015). The SCIP project is a two-wave panel study of selected migrant groups in which about 8000 recent migrants aged between 18 and 60 were surveyed in four European destination countries -Germany, Netherlands, Great Britain, and Ireland. Migrants were interviewed in 2010/2011 after about 18 months upon their arrival and were re-interviewed about one-and-a-half year later. Included in the SCIP survey were Poles as a recent migrant group (as well as Bulgarians in the Netherlands) and groups representing the classical labor/colonial migration to Western Europe (Pakistanis in Great Britain; Turks in Germany; Turks, Moroccans, Surinamese and Antilleans in the Netherlands). These groups contribute greatly to the share of migrant population in the four countries (for a detailed description of the methodological setup of the project, see Gresser and Schacht, 2015). Due to our interest in group comparisons between Christian and Muslim migrants, we exclude the Irish sample which only contains Polish migrants. While there is religious variation within migrant groups in the Netherlands, religious and ethnic background overlaps in Great Britain and Germany. In our analyses, we exclusively focus on differences across religious groups.
Our major dependent variable to measure socio-economic integration is employment. Respondents were asked about their main activity in both waves (working, unemployed, in education, retired, long-term sick or permanently ill, looking after the home, maternity/paternity leave, other). We recoded those working and those in maternity/paternity leave as being employed. Since our major interest is in those having a chance of entering the labor market, we exclude all respondents who were retired or long-term sick, 1 as well as those who indicated that they had already signed a job contract prior to arrival. In our cross-sectional analyses of wave 1, we exclude those who indicate being in full-time education, while in our longitudinal analysis of the balanced panel, we exclude only those who indicated being in education in both waves 1 and 2.
To directly test our theoretical arguments about channeling effects of religious participation, we rely on additional information about the employment sector. Working respondents who were not self-employed or employed by family members were asked to provide information on the ethnic background of their immediate supervisor. As a proxy for integration in the mainstream labor market, we create a dependent variable which treats only those indicating having a native or third country supervisor as employed in the mainstream economy, contrasted to those in the ethnic niche economy which comprise the employed with a co-ethnic supervisor as well as the self-employed, a typical category among newcomers (Van Tubergen, 2005), and the family employed.
Our main independent variable of interest is religious participation. Based on responses to questions about religious affiliation, we restrict our analysis to a comparison of self-identified Christians (the reference category) and Muslims which constitute more than 90% of the respective migrant groups (except for Bulgarians and Surinamese in the Netherlands which are religiously more heterogeneous); the small number of religiously unaffiliated is excluded from analysis. Religious participation is measured by church or mosque attendance which can be seen as major aspect of organized religious involvement and lies at the center of the theoretical arguments mentioned above. We recode attendance as binary variable distinguishing regular attendees (monthly or more) from non-regular attendees. When testing the theoretical arguments about channeling functions of religious communities, we rely on information about their ethnic composition. People who attend a mosque or church were asked whether all, most, half, some or almost no one among the other worshippers was from their own ethnic background. Those indicating that all or most of the attendees are from their own ethnic background are recoded as co-ethnic (the reference category), the others as inter-ethnic attendees. We use the variable of religious communities' ethnic composition when analyzing the mainstream versus ethnic niche economy placement of newcomers. For further interpretation, we draw on responses to a question on self-reported assistance received by religious organizations which was asked to those attending a church or mosque.
In light of the prevalent literature on immigrants' labor market integration, we control for a number of variables. In order to capture the pre-migration human capital, we include years of education and years of work experience in the country of origin, while also controlling for migration motifs that might initially prevent from actively searching employment (family migration versus other). To capture early processes of socio-cultural assimilation which are known to affect entry into the labor market, we furthermore include language competence measured by an additive index of passive and active reading and writing skills (scale of 1-4) as well as interethnic social ties measured by frequency of contact with natives (scale of 1-6); we furthermore account for length of stay (in months). A final set of socio-demographic control variables, not displayed in the analyses, includes age and age squared (given curvilinear age effects on labor market outcomes); sex 2 ; marital status; and children in household. 3 Missing variables are handled by list-wise deletion.
In our analyses, we start by inspecting summary statistics across the three countries displayed in the Appendix. Moving to multivariate analyses, we proceed in two steps. First, we run cross-sectional logistic regressions which capitalize on the large number of respondents in wave 1 and provide preliminary insights into whether and how religious participation is correlated with labor market entry shortly after entering the host country (Tables 1). We analyze the effects of religious participation on employment (Model I) and on the mainstream versus ethnic niche placement (Model II), respectively, controlling for pre-migration human capital, socio-cultural assimilation, and socio-demographic characteristics. To capture group-differences in religious participation effects, we include an interaction term of religious participation with religious affiliation in Model I; the main effect of religious affiliation sheds light on net differences in levels of employment across religious groups. To capture differences across types of religious communities (coethnic versus inter-ethnic), Model II contains an interaction effect of religious participation and religious communities' ethnic composition. Second, we move to the longitudinal analyses of balanced panel data which permit assessing the causal relations hypothesized in the theoretical arguments more strictly, although panel attrition reduces case numbers considerably (Table 2). Here, our main question is whether newcomers' early involvement in religious communities affects their subsequent chances of finding a job and of being channeled into mainstream or ethnic niche economy. To analyze this question, we reduce the sample to those who had not already found a job in wave 1, now including students, and who thus could be safely assumed to be at risk of becoming employed in wave 2. Capitalizing on the panel structure of the SCIP data, we run logistic regressions with timelagged-independent variables plus change scores for language competence and inter-ethnic contacts. 4 In line with our explorative comparative design, both cross-sectional and longitudinal regressions are conducted separately for each country.

Empirical findings
As summary statistics show (see Appendix), recent newcomers in Germany, the Netherlands and Great Britain succeed in entering the labor market to different degrees. When interpreting these summary statistics, one has to bear in mind that our definition of the risk set of entrees into the labor market excludes those who study in wave 1. Looking at the full wave 1 dataset, it appears that employment rates are highest in Great Britain with 65.6% indicating to have a job, followed by Germany (60.7%) and Netherlands (51.3%); the pattern is different, however, when we focus on employment in the mainstream labor market where, compared with ethnic niche economies, rates are highest in the Netherlands (63.2%), followed by Great Britain (40.4%) and Germany (25.8%). In terms of their socio-demographic characteristics, human capital composition and assimilation patterns, migrant populations look rather similar while in Great Britain they display a distinctive pattern which is due to the research design which   excludes large numbers of Pakistani respondents, as student migrants, from analysis in wave 1. When taking a look at the balanced panel which does include those who study in wave 1 (but excludes those who study in wave 2), summary statistics suggest a substantial change over time, with employment rates increasing in all three countries, especially in Great Britain. By the time of wave 2, almost two thirds, and three quarters in Great Britain have found a job. Interestingly, although it is only in Germany that these increases are partly due to entry in the mainstream employment sector; among the employed, the share of people in that sector increases from 26.3% to 28.1%. In the Netherlands and Great Britain, mainstream employment sector decreases among the employed compared to wave 1, suggesting that most newcomers in these countries move into ethnic niche economies.
In terms of their religious characteristics, immigrants display relatively high rates of regular worship attendance at the time of wave 1, although it is known from previous research with SCIP data that religious participation is, for all groups, considerably lower than prior to migration (see Diehl and Koenig, 2013;Van Tubergen, 2013). In Germany and the Netherlands, both in the full dataset and in the balanced panel, around a third of the respondents indicate attending places of worship at least on a monthly basis in wave 1, and the numbers are even higher in Great Britain, notably among Muslims. All three countries are also similar in that, among the religiously participating immigrants, only less than a third attends inter-ethnic places of worship, while the majority participates in co-ethnic religious communities; this already sheds doubt on the assumption in the ''bridging'' argument that religious participation as such should increase employment chances by means of embedded social capital.
To analyze how religious participation is related to employment for Christian and Muslim immigrants, we turn to cross-sectional logistic regression analyses of the full sample of respondents from wave 1. Before turning to the effects of religiosity, we note the importance of pre-migration human capital and socio-cultural assimilation variables which confirms previous studies about labor market entry. Across all three countries, family migrants are less likely to be employed, while those with work experience seem to have a higher probability of finding jobs (albeit not necessarily in the mainstream economy). Employment, especially in the mainstream economy, tends to be higher among those proficient in the language of the receiving society, and inter-ethnic ties are also positively related to employment. Effects of educational background, however, are more ambiguous (and entirely absent in the Netherlands).
With respect to religious effects, we should first of all stress that net of all these factors Muslims tend to have lower employment levels in Germany, the Netherlands, and Great Britain. Upon closer scrutiny, it turns out, however, that such disadvantages pertain to the mainstream employment sector only in the Netherlands where Muslims are 70% less likely to find jobs than their Christian counterparts [exp{À1.102}¼0.332]. In contrast, the odds for employed Muslims to be in the mainstream economy are more than three times higher in Germany [exp{1.342}¼3.827] and in Great Britain [exp{1.119}¼3.062]; the overall higher employment rates among Christians from Poland thus seem to be mostly driven by their involvement in ethnic niche economies.
But what about active religious participation -does involvement in religious organizations as measured by regular worship attendance hamper or help socioeconomic integration? In their most general versions, neither of these two theoretical arguments finds support in our data as evinced by the absence of attendance effects in Models I in each of our three countries. However, when comparing those among the employed who find jobs in the mainstream labor market as compared with those in ethnic niche economies (Models II), it turns out that, while religious participation has a negative main effect, attending an inter-ethnic place of worship increases the odds of finding employment in Germany by 80% [exp{0.594}¼1.811]. Although the effect reaches only marginal levels of statistical significance, this finding lends at least modest support to arguments about the channeling function of religious community involvement. In fact, when excluding host language competence and inter-ethnic ties from Model II in Germany (not displayed), the interaction effect is slightly stronger (b ¼ 0.692) and becomes statistically significant at the level of p < 0.01 which suggests that the ethnic composition of religious communities might indeed play a role in sociocultural assimilation.
In sum, the cross-sectional analysis of wave 1 provides only limited evidence for the theoretical arguments currently proposed in the literature. It could be, however, that the effects of religious participation only unfold over time. Do those who have not yet found a job by the time of wave 1 (t 1 ), profit from religious participation in terms of their employment status by the time of wave 2 (t 2 )? Do these dynamics differ across religious groups? And does the ethnic composition of religious communities channel attendees into different segments of the labor market?
To analyze these questions, we turn to logistic regressions with time-lagged independent variables among those respondents in the balanced panel who had not yet found employment at the time of wave 1. Again, we start by noting some interesting general dynamics of labor market entry. As summary statistics suggests (Appendix), substantial numbers of new migrants move into the labor market between wave 1 and wave 2 in all three countries. However, Germany is the only country where some of them enter the mainstream labor market; very few do in the Netherlands and no one does in Great Britain, which is why Model II is missing for that country. Pre-migration human capital variables tend to be of less relevance for moving from unemployment into employment or for determining the labor market segment where employment is found. Linguistic skills help migrants find employment in the Netherlands, and inter-ethnic ties as well as their increase across waves are positively linked to employment in Germany and Great Britain.
In group-comparative comparison, it is first of all noticeable that the ''penalties'' for Muslims which we found in wave 1 across all three countries weaken in the longitudinal perspective, although initial employment gaps are not necessarily closed. Muslims who are unemployed in wave 1 do, however, seem to face particular difficulties in entering the mainstream labor market; their odds of being in that employment sector are almost 70% lower in Germany [exp{À1.080}¼0.340] and more than 70% lower in the Netherlands [exp{À2.538}¼0.079] compared with their Christian counterparts.
Religious participation, in contrast, turns out to have no statistically significant effects whatsoever in Germany and the Netherlands. Neither the general hypotheses derived from assimilation theory and from the bridging argument nor the more specific hypotheses about channeling effects of co-ethnic and inter-ethnic religious involvement receive any support in these two countries. The situation is different in Great Britain where religious attendance (Model I) has a rather strong negative main effect on the chances of being in employment in wave 2 (b¼À2.709). However, whereas newer versions of assimilation theory would expect such an effect noticeably among members of a group facing strong boundaries, the positive interaction effects (b ¼ 2.628) indicates that Muslim attendees are actually more likely to be employed than Christian attendees.
The findings presented so far yield only limited support to the theoretical arguments developed above. The absence of bridging effects is even further confirmed by descriptive information on perceived assistance from religious organizations. As reported in the summary statistics (Appendix), hardly any religious attendee, whether Christian or Muslims, reports having received assistance from their church or mosque in all three countries. The religious bridging mechanism even seems to be absent for religious majority members in Western Europe. At the same time, contrary to conventional arguments from assimilation theory, involvement in religious community life does not generally hamper socio-economic integration either.

Discussion and conclusion
In this paper, we shed some light on current controversies over the role of religion in immigrant integration by analyzing the effects of religious participation on employment among recently arrived Christian and Muslim migrants in Great Britain, the Netherlands and Germany. Religiosity is increasingly hypothesized as relevant factor for socio-economic aspects of immigrant integration. But to the best of our knowledge, ours is the first study to systematically compare religious participation effects upon new immigrants, labor market entry across religious groups and to scrutinize the potential channeling effects which the ethnic composition of religious communities exercises upon entry into mainstream and ethnic niche segments of the labor market, net of socio-demographic, human capital, and socio-cultural assimilation. Drawing on SCIP data, our study contributes to current theoretical debates on religion and immigrant integration in several respects. First, arguments derived from assimilation theories which regard religiosity as part of an ethnic investment strategy that would negatively affect entry into the labor market receive very limited support from our empirical findings. For Muslims, in particular, religious participation seems to be entirely decoupled from employment in all three countries, which sheds doubt on alarmist public discourse about the disintegrative potential of Islamic organizations. Only among Christians in Great Britain does religious participation seem to be linked to lower employment rates in our longitudinal analysis -which indicates rather strong ethnic orientations among religiously active Catholic Poles in that country. We also find only little evidence for the idea that religious participation generally channels newcomers into ethnic segments of the labor market; the marginally significant negative effects in Germany in the cross-sectional analysis are not corroborated in longitudinal perspective. Taken together, our cross-sectional and longitudinal analyses suggest that religious participation neither hinders newcomers from findings jobs nor channels them into ethnic niche economies. Second, our findings also show little support for the alternative idea that religious participation in Western Europe operates as a ''bridge'' in the same manner as it does in North America, not even for Christian newcomers whose religious participation might be expected to facilitate contact to the native population. Only in Germany, we find some evidence that attendees of inter-ethnic places of worship initially seem to be more active in the mainstream labor market, but these correlations again cannot be corroborated in the longitudinal analysis. This finding confirms previous studies which have found religious bridging mechanisms to operate only in the North American context where religious fields are constituted by voluntary multipurpose organizations with various embedded resources that might provide advantages for immigrants' socio-economic integration (see Connor and Koenig, 2013). In Western Europe, despite cross-national differences in labor market regulations and integration policies, religious participation seems to be strongly decoupled from socio-economic integration.
These null effects of religious participation may be taken as evidence for those versions of secularization theory which posit that religious practices, whether declining or not, have become structurally differentiated from other domains of society in Western Europe (e.g. Casanova, 1994; see also findings by Fleischmann and Phalet, 2012). At the same time, our findings do support ideas about the persisting role of religion as a marker of difference (Brubaker, 2014). While across all three countries, there is considerable socio-economic mobility among newcomers, employment gaps persist between Christian and Muslims. Evidently, the group-comparative design of the SCIP dataset makes it impossible to strictly distinguish religious from ethnic ''penalties''. But as our multivariate models do take into account many relevant socio-demographic characteristics, pre-migration human capital and socio-cultural assimilation measures, the persisting employment gap may indeed signal the existence of salient boundaries which could, in the long run, translate into the processes of social closure which the aforementioned literature on ''penalties'' claims to document. The disadvantages of Muslim migrants in entering the mainstream labor market over time in Germany and the Netherlands clearly point in this direction.
Of course, our analyses provide only tentative findings given that our groupcomparison covers only few origin countries and focuses on simple measures of employment. Future studies may include additional measures for socio-economic integration, notably occupational attainment which is often said to follow a U-shaped pattern from pre-migration through the first years of settlement within the new country (Chiswick, 2003). Sequence analyses of labor market performance (see e.g. Fuller and Martin, 2012) might also provide more nuanced insights into typical patterns of socio-economic integration which could be related to religious adaptation processes. Finally, future studies may go beyond the focus on participation in organized religion by scrutinizing the effects of religious identities and beliefs upon structural integration. For sure, more research needs to be done to understand how precisely religious differences are linked to social inequalities, a question that will likely continue to be on the agenda of Western European immigrant societies in the foreseeable future.

Declaration of conflicting interests
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Appendix I. Summary statistics (SCIP, respondents at risk of entering labour market)