Can Religious Affiliation Explain the Disadvantage of Muslim Women in the British Labour Market?

Introduction

The past 15 years have witnessed particular macro and micro changes in relation to British Muslims in general and Muslim women in particular. The global image of Muslims has negatively shifted due to increased levels of Islamophobia on the one hand (Gottschalk and Greenberg, 2008; Meer and Modood, 2009) and growing resentments among various segments of the population in relation to migration generally on the other hand (Geddes, 2014). Parallel to this, Muslim women now seem to be outperforming Muslim men in relation to their school achievement and are more likely to obtain higher qualifications and degrees (Khattab and Modood, 2017). These changes are likely to affect the opportunity structure and employment prospects of Muslim women in two different and opposing directions. The increased levels of discrimination against Muslims and greater resentments towards migration are likely to restrict these opportunities, whereas an increased proportion of Muslim women obtaining higher qualifications and degrees is likely to improve their labour market opportunities and facilitate better integration. However, the thorough literature review undertaken for this research did not yield even a single systematic study that focuses exclusively on the disadvantages of Muslim women in the UK labour market. Evidence-based knowledge in relation to how well Muslim women are incorporated within the British labour market is therefore limited.

This article draws on the ‘intercategorical complexity’ approach (McCall, 2005), and the intersectionality literature more generally (Choo and Ferree, 2010), to examine how and to what extent the labour market experience of British-Muslim women is multiplied via criteria of ascription such as ethnicity, migration trajectories (first and second generations), race, gender and religion, rather than criteria of achievement, such as education (Treiman, 1970). Moreover, by analysing a range of labour market outcomes, such as economic activity, obtaining professional and managerial jobs (salariat jobs) and earnings from work, the study will help address the question of whether the labour market experience of Muslim women is fixed across the labour market outcomes, or is contingent upon the specific outcome.

Furthermore, this study provides an up-to-date analysis of the position of British-Muslim women in the labour market by using the Labour Force Survey (LFS) data from 2002 to 2013. These together provide a fine-tuned picture of the labour market penalties facing Muslim women in the UK, while advancing the readers’ understanding of how assuming multiple identities can often shape the experience of multiply-marginalized individuals and groups (Choo and Ferree, 2010).

Intersectionality and multiply determined inequality

Muslim women do not enter the labour market solely as women; they face a multitude of challenges invoked by their ethnicity, race, migration status, culture and religion. They also cannot be treated as a homogeneous group as issues of race, ethnicity, migration and class are of great importance in understanding their multiple identities as well as experience in the UK labour market (Brah, 1993; Ruwanpura, 2008). In this context, Muslim women occupy multiple positions and identities that are complex and multi-directional. The visible identity for some, especially when wearing the veil or hijab, as well as other markers such as colour and dress code, play an important part in the construction of ‘otherness’ (Brah and Phoenix, 2013; Brown, 2006) and reflect the intersectionality and shifting nature of race and ethnic boundaries (Archer and Francis, 2006; Arrighi, 2001; Crenshaw, 1991; Lewis, 2004; Reay et al., 2007). This suggests that the experience of an individual within a given racial or ethnic identity (e.g. being white or black) is not shaped or determined only by his or her membership within that particular identity, but is relational to several identities. The effect of each intersection is likely to be amplified or indeed defused depending on how salient other intersections are and vice versa. For example, whiteness can be an asset for one group (e.g. white middle-class women), but it loses its importance, or at least some of it, for other classes or religious groups (i.e. white working-class women).

Following a relational positioning concept, one can argue that the value that is attached to certain categories of class or ‘colour’ can shift and change in how it is perceived by others (e.g. majority group) depending on their other identities; for example, religion or migration (Franks, 2000). Here, the formulation of ‘otherness’ is never static and is always ‘in process’; consequently, racism and racializing practices are not limited to one identity (Fox et al., 2012), do not apply equally to all groups (Khattab, 2009; Khattab and Johnston, 2013) and do not operate at the same extent across the labour market outcomes (Khattab and Johnston, 2015). For example, in the past, both Jews and Irish in Britain faced racism and discrimination (Ghaill, 2000; Knepper, 2007), whereas at present East European migrants and even White-British women who convert to Islam face different forms of racism (Fox et al., 2012; Franks, 2000; Suleiman, 2013). Through this process, some groups might experience multiple layers of racialization ‘otherness’ (e.g. being black, a migrant and a Muslim). Other groups, such as White-British Muslim women, are likely to experience a peculiar and unique position due to subscribing to the same ethnicity as the majority population on the one hand and are identified as others because of their religious background on the other hand, especially if they wear the hijab (Franks, 2000; Moosavi, 2015; Suleiman, 2013). These variations are also likely to be reflected in relation to migration trajectories; for example, between first and second generation migrants within and between groups.

Intersectionality and the effect of criteria of achievement

Similar to other migrants, Muslim women, especially first generation and those newly arrived into Britain, are likely to face labour market penalties due to the lack of language skills, knowledge and familiarity with the system of the new host country (Borjas, 1994; Chiswick, 1978, 1999; Chiswick and Miller, 2002; Reitz, 2001) and other socio-cultural factors such as women’s gender role attitudes and preferences (Koopmans, 2016). In the classical work of Chiswick (1978) on the earnings of foreign-born men in the US, he argues that the economic opportunities of migrants will improve rapidly over time and the initial gaps between them and the host population will be closed as migrants will adapt their skills to the host society. The de-valuation of their human capital and qualifications is likely to be a naturally occurring adjustment period that blurs the linkage between human capital and labour market outcomes. During this period of adjustment, incidents of over-qualification, unemployment and low-paid jobs are to be expected (Nielsen, 2011).

However, obtaining language skills and improving human capital by migrants and minorities can elevate their socio-economic status within the host society only if their multiple identities are neutral. In the case of Muslim women in Britain, none of their multiple identities are neutral. Their ethnicity, religion, migration status and race are all significant, linked and operate simultaneously in the process of shaping their opportunity structure. If migrants and their descendants (the second generation and beyond) are racialized by the majority group and are seen as ‘others’, their human capital can be devalued and they are subsequently marginalized in the labour market, as is shown in relation to Chinese migrant women in Canada (Man, 2004). Under these circumstances, the initial disadvantages of some ethnic groups can become long-term penalties that persist during their working lives (Borjas, 1994) and, in the case of Muslims in some western societies, these penalties are being transmitted to second generations.

As much as the labour market penalties may vary according to whether a group has been racialized and to its level of visibility (physical and cultural), it may also vary according to labour market position or outcome. It is possible that the main barrier or penalty facing minorities in Britain is entering the labour market and finding jobs commensurate with their qualifications (see Cheung and Heath, 2007), especially jobs within the salariat class (Cheung, 2014; Khattab and Johnston, 2015). These studies suggest that when minorities (men or women) eventually find these jobs, they tend to receive the same level of returns; for example, wages, as majority workers. Thus, theoretically, Muslim women in Britain are likely to face substantial penalties in finding employment and obtaining jobs within the salariat class. However, their pay is likely to be similar to the pay of the majority group, especially in the case of second generation. If this hypothesis turns out to be true, then the multi-impact of intersectionality in relation to the labour market experience of Muslim women would be particularly salient at the stage of transition into the labour market and matching jobs with skills and qualifications.

In light of the above discussion, this study will significantly advance the literature on the labour market incorporation of ethnic and religious minority women in general and Muslim women in particular. It helps explain how different social markers and multifarious identities have contingent relationships to multiple determinants and outcomes.

Methods and data

An ‘intercategorical complexity’ methodological approach is adopted to capture the complex intersectionality and co-constitution of gender, race and religion in relation to labour force outcomes (McCall, 2005). Existing analytical categories are utilized through this approach to investigate relationships of inequalities along the multiple, and sometime conflicting dimensions of religion and race. Existing inequalities are thus examined in relation to already constituted social groups with the aim of explicating those relationships. Within this approach, gender, religion and race are treated as ‘anchor’ points, which are not static but are used to identify relative inequalities (Nakano Glenn, 2002).

In this article, data are utilized from 12 April to June quarters of the LFS between 2002 and 2013. The Quarterly Labour Force Survey is the single most important source of labour market data in the UK. It is a large random survey of around 60,000 households sampling over 155,000 individuals. Each quarter is made up of five waves of around 12,000 households interviewed over five successive quarters. Within any one quarter, the survey captures around 70,000 individuals of the working population. However, in some cases, where the sample is broken down by a large number of variables, such as ethnicity, religion, gender, labour market outcomes, qualifications and so on, some estimates might be unreliable due to sample size issues. One way of increasing the statistical power and reliability of survey data is to increase the sample size on which it is based. While the annual sample size is fixed, several years’ worth of data can be pooled to produce estimates for the average of the combined years (see, for example, Almond and Healey, 2003; Gauchat, 2012; Rafferty, 2012).

The availability of information on religion, ethnicity, qualifications and employment patterns on the one hand and the pooling together of a large number of quarters on the other hand make the LFS a valuable source for such analysis. For theoretical purposes, the analysis has been restricted to Christian and Muslim women aged 19–65 years. All women who have self-subscribed to a religious category other than Christian or Muslim (e.g. Jews, Hindus, Sikhs, No Religion and Others) have been excluded from the analysis. The prime interest here is to understand how Muslim women perform in the labour market compared to the majority group of Christian White-British women, which was used as the reference group in the multivariate analysis. The pooled data generated an initial sample of 245,391 women, of which 8444 were Muslim. However, the sample size varies when different analysis and different variables are used, mainly because of missing observations in the original data and due to the relatively lower labour market participation rates among Muslim women in general.

Findings

Modelling economic activity

This labour market outcome has been modelled using a multinomial logit regression model which estimates the odds-ratios of falling in the category of unemployment or economic inactivity compared to the reference category (being in employment). A co-efficient that is less than 1 indicates lower odds of falling within the specific category relative to the reference category (being employed as in this case). A coefficient that is greater than 1 indicates higher odds of falling within the specific category relative to the reference category.

In the first two columns of Table 1, the results of being unemployed or economically inactive are compared to being in employment (the reference category) without introducing the interaction terms. The results show that all of the groups were more likely to be unemployed or economically inactive than being in employment. The only statistically insignificant result is in relation to White-British Muslims in the case of unemployment. This suggests that White-British Muslim and White-British Christian women had the same probability of experiencing unemployment. Furthermore, the coefficients associated with second generation are substantially smaller than the coefficients associated with first generation, suggesting a reduced penalty for second generation Indian-Muslim, Pakistani-Muslim and Bangladeshi-Muslim women and lending support to the assimilation theory (Chiswick, 1983; Chiswick and Miller, 2002; Koopmans, 2016). The results in the case of unemployment show that Black-Muslim women were similar to other first generation groups, whereas Muslim White-Others were similar to second generation groups. This might suggest that Muslim White-Others faced a reduced penalty, which might be attributed to race but also to their knowledge and awareness of the culture.

OPEN IN VIEWER

Table 1.

Multinomial logistic regression (odds-ratios) for economic activity among women aged 19–65 years, UK LFS 2002–2013 (N=225,215), with reference group ‘in employment’.

	No interactions		With interactions
	Unemployment	Inactive	Unemployment	Inactive
Intercept	−1.75	3.31	−1.73	3.29
Age	0.95**	0.80**	0.95**	0.80**
Age squared	1.00**	1.00**	1.00**	1.00**
Marital status, base=Married
Single	2.46**	1.32**	2.46**	1.32**
Separated/divorced	2.44**	1.12**	2.44**	1.12**
Dependent children under 10	1.28**	2.04**	1.28**	2.04**
Region, base=Inner London
Outer London	0.85	0.74**	0.85	0.75**
Rest of UK	0.77**	0.71**	0.78**	0.71**
Qualification, base=No qualification
Low and high tertiary education	0.25**	0.17**	0.23**	0.17**
Low and high secondary education	0.54**	0.32**	0.51**	0.32**
Ethnicity, base=Christian White-British
1st generation Indian-Muslims	3.43**	4.41**	2.15	7.90**
2nd generation Indian-Muslims	2.94**	3.07**	1.78	3.61**
1st generation Pakistani-Muslims	6.88**	10.62**	4.00**	13.45**
2nd generation Pakistani-Muslims	3.59**	3.29**	2.13	5.60**
1st generation Bangladeshi-Muslims	6.45**	9.48**	6.77**	22.52**
2nd generation Bangladeshi-Muslims	2.36**	2.23**	4.50	8.01**
White-British Muslims	1.32	3.03**	3.09*	4.35**
Muslim White-Others	2.76**	6.01**	1.66	6.22**
Black-Muslims	5.23**	4.42**	4.14**	6.72**
Year, base=2002–2004
2005–2007	1.06	0.92**	1.07	0.92**
2008–2010	1.58**	1.20**	1.59**	1.19**
2011–2013	1.96**	1.16**	1.96**	1.16**
Interaction effect
1st generation MI × tertiary education			1.20	0.23**
2nd generation MI × tertiary education			2.67	0.42
1st generation MP × tertiary education			3.06**	0.52**
2nd generation MP × tertiary education			1.73	0.36**
1st generation MB × tertiary education			0.82	0.16**
2nd generation MB × tertiary education			0.30	0.17**
MWB × tertiary education			0.42	0.40*
MWO × tertiary education			1.59	0.66
MBL × tertiary education			1.63	0.35**
1st generation MI × secondary education			2.31	0.63
2nd generation MI × secondary education			1.33	1.07
1st generation MP × secondary education			1.88*	0.79
2nd generation MP × secondary education			1.91	0.68
1st generation MB × secondary education			1.19	0.41**
2nd generation MB × secondary education			0.61	0.29*
MWB × secondary education			0.34	0.76
MWO × secondary education			2.11	1.12
MBL × secondary education			1.38	0.67
Likelihood Ratio Tests	χ2=42,819.07, p<0.001		χ2=43,042.76, p<0.001
Cox and Snell	0.17		0.17

Notes: MI: Indian-Muslims; MP: Pakistani-Muslims; MB: Bangladeshi-Muslims; MWB: White-British Muslims; MWO: Muslim White-Others; MBL: Black-Muslims; *p<0.05, **p<0.01.

Source: Labour Force Survey 2002–2013, authors’ calculation.

Turning to the model with interaction terms, Table 1 shows that in relation to unemployment, qualifications seemed to operate in the same way across all groups. The only exception is the impact of qualifications among first generation Pakistani-Muslim women. For them, higher qualifications (tertiary and secondary) were associated with a greater risk of unemployment. This might relate to various factors including their original migration trajectories and whether they had accompanied their partners and did not seek employment at the time of migration. It also may relate to qualification recognition and their language skills. Furthermore, adding the interaction term to the analysis reduces the differences between most of the Muslim groups and White-British Christian women, suggesting that some of the initial differences observed in the first model might be linked to different impacts of qualification among these groups. Interestingly, including the interaction term between ethnic-generation and qualifications reveals that the main difference in the risk of unemployment between White-British Muslim and White-British Christian women relates only to women with no qualifications. This suggests that White-British Muslim women did not face any penalties in unemployment if they held secondary or higher qualifications when compared to the group of White-British Christian women.

The last column in Table 1 shows that qualifications were very important in reducing economic inactivity among all groups of Muslim women. Muslim women with tertiary education were in fact less likely to be economically inactive than White-British Christian women relative to being in employment. All but two coefficients are statistically insignificant: for second generation Indian-Muslim and White-Other Muslim women. This does not automatically suggest that Muslim women are less likely to be economically active than White-British Christian women due to discrimination, although this is a viable possibility given the strong evidence of discrimination based on names (Bertrand and Mullainathan, 2004; Blommaert et al., 2014). Equally, this outcome could be a result of cultural preferences and the impact of other socio-cultural factors that have not been examined here (Koopmans, 2016).

Modelling salariat jobs

Table 2 summarizes the logistic regression analysis for predicting the salariat jobs in three models. The first model includes the ethnic-generation background, whereas in the second and third models this variable has been replaced by the generation variable so that the interaction with qualifications can be added in the third model. Adding the interaction term using the ethnic-generation background variable destabilized the analysis, yielding unreasonably large exponents due to the small sample size for some of the groups. In order to allow some examination for the differential impact of education across generations, the second best option was to utilize the generation variable instead.

OPEN IN VIEWER

Table 2.

Logistic regression (odds-ratios) for obtaining a salariat class among women aged 19–65 years, UK LFS 2002–2013 (N=219,342), with reference group ‘all other classes’.

	Model 1	Model 2	Model 3
Constant	0.004	0.004	0.004
Age	1.17**	1.17**	1.17**
Age squared	1.00**	1.00**	1.00**
Marital status, base=Married
Single	0.89**	0.89**	0.89**
Separated/divorced	0.91**	0.91**	0.91**
Dependent children under 10	0.85**	0.85**	0.85**
Region, base=Inner London
Outer London	0.97	0.97	0.97
Rest of UK	0.76**	0.76**	0.76**
Qualification, base=No qualification
Low and high tertiary education	32.65**	32.65**	31.93**
Low and high secondary education	4.40**	4.40**	4.33**
Ethnic-generation, base=Christian White-British
1st generation Indian-Muslims	0.43**
2nd generation Indian-Muslims	0.60**
1st generation Pakistani-Muslims	0.19**
2nd generation Pakistani-Muslims	0.50**
1st generation Bangladeshi-Muslims	0.15**
2nd generation Bangladeshi-Muslims	0.51**
White-British Muslims	0.68**
Muslim White-Others	0.22**
Black-Muslims	0.25**
Generation, base=British natives
1st generation Muslim women		0.22**	0.04**
2nd generation Muslim women		0.51**	0.05**
Year, base=2002–2004
2005–2007	1.08**	1.08**	1.08**
2008–2010	1.02	1.03	1.03
2011–2013	0.96**	0.96**	0.96**
Interaction effect
1st generation × tertiary education			6.78**
2nd generation × tertiary education			12.20*
1st generation × secondary education			5.55**
2nd generation × secondary education			9.32*
Likelihood Ratio Tests	χ2=57,612.59, p<0.001	χ2=57,490.25, p<0.001	χ2=57,549.29, p<0.001
Cox and Snell	0.23	0.23	0.23

Note: *p<0.05, **p<0.01.

Source: Labour Force Survey 2002–2013, authors’ calculation.

Model 1 shows that while controlling for other factors, all groups of Muslim women were significantly less likely to obtain jobs within the salariat class when compared to their British-Christian women counterparts. While all Muslim women seemed to face a penalty here, the coefficients (odds-ratios) associated with second generation are somewhat larger than the coefficients associated with first generation, suggesting that they faced a reduced penalty compared to first generation. White-British Muslim women have the largest coefficient (odds-ratio of 0.68), suggesting that they faced the lowest penalty, yet this penalty is still significant. Consistent with previous results, findings related to White-Other Muslim and Black-Muslim women were similar to first generation Pakistani and Bangladeshi women.

Model 2 confirms this generation difference (OR=0.22 and 0.51 for first and second generation, respectively). However, when the interaction term between generation and qualification is included in model 3 (third column in Table 3), this initial difference between first and second generation was diminished (OR=0.04 and 0.05 for first and second generation, respectively). Model 3 also shows that the likelihood for obtaining salariat jobs for both generations of women with no qualifications was significantly reduced further. The interaction terms show that the effect of tertiary education on obtaining salariat jobs for first and second generation is 6.78 and 12.20 times (respectively) the effect of tertiary education for White-British Christian women. The equivalent effect of secondary education for first and second generation is 5.55 and 9.32 times (respectively) the effect for White-British Christian women. This difference in the impact of education is attributed to the very large and significant differences between first and second generation Muslim women and White-British Christian women in the odds of obtaining salariat jobs within the no qualification category. This implies that education is more important for Muslim women, especially second generation, for attaining a salariat job than for White-British Christian women.

OPEN IN VIEWER

Table 3.

Mixed (multilevel) linear model for logged gross hourly pay (N=38,288).

Parameter	Model 1	Model 2	Model 3
Intercept	2.168	1.568	1.565
Ethnic-generation, base=Christian White-British
1st generation Indian-Muslims	0.143	−0.040	−0.370
2nd generation Indian-Muslims	0.141	−0.097	−0.005
1st generation Pakistani-Muslims	−0.092	−0.099*	0.014
2nd generation Pakistani-Muslims	0.005	−0.004	−0.053
1st generation Bangladeshi-Muslims	−0.268**	−0.294**	−0.200
2nd generation Bangladeshi-Muslims	−0.145	−0.140**	−0.189
White-British Muslims	−0.135	−0.060	−0.100
Muslim White-Others	−0.194*	−0.164*	−0.092
Black-Muslims	−0.133	−0.201**	0.117
Age		0.028**	0.028**
Age squared		0.000**	0.000**
Marital status, base=Married
Single		−0.006	−0.006
Separated/divorced		−0.011*	−0.011*
Dependent children under 10		0.017**	0.017**
Region, base=Inner London
Outer London		−0.086**	−0.087**
Rest of UK		−0.285**	−0.286**
Qualification, base=No qualification
Low and high tertiary		0.275**	0.276**
Low and high secondary		0.088**	0.089**
Part-time		−0.062**	−0.062**
Public sector		0.045**	0.045**
Length of employment in months		0.001**	0.001**
Year, base=2002–2004
2005–2007		0.120**	0.120**
2008–2010		0.204**	0.204**
2011–2013		0.246**	0.246**
Interaction effect
1st generation MI × tertiary education			0.454*
Black-Muslim × tertiary education			−0.443**
Black-Muslim × secondary education			−0.312*
Occupational control at level-2	No	Yes	Yes
Residual	0.290	0.14	0.14
Level-2 variance	–	0.07	0.07
Schwarz’s Bayesian Criterion (BIC)	61,203.000	32,650.953	32,816.821

Note: *p<0.05, **p<0.01.

Source: Labour Force Survey 2002–2013, authors’ calculation.

Discussion

This article analysed LFS data between 2002 and 2013 in order to examine how and to what extent the labour market experience of British-Muslim women is multiply determined via criteria of ascription such as ethnicity, race, gender and religion and to what extent this experience is contingent upon the specific labour market outcome. The overarching finding suggests that most Muslim women, regardless of their multiple ascriptive identities, generation and levels of qualifications, still face significant penalties compared with their White-British Christian counterparts. For example, except for White-British Muslim in the case of unemployment, all other groups (Indian-Muslims, Pakistani-Muslims, Bangladeshi-Muslims, Muslim White-Others and Black-Muslims) were more likely to be unemployed or economically inactive and less likely to obtain jobs in the salariat class.

However, the data presented in this article leave little doubt that the labour market experience of Muslim women varies enormously by ethnicity, the type of labour market outcome under study and by generation. It seems that their labour market performance is highly sensitive to their complex intersectionality of gender, race/ethnicity and religion (McCall, 2005). They do not experience the labour market only as women, but also their different dimensions of identity (ethnic, religious and migration status) intertwined in a way that constructs complex social locations and complex inequalities.

In the case of Muslim women, their religious identity might have greater influence than other identities because it is likely to operate extrinsically through unfair practices by employers (Blommaert et al., 2014; Dale, 2002; Ghumman and Ryan, 2013) and intrinsically through women’s choices and cultural preferences (Dale et al., 2002). The latter impact is likely to be associated with a narrower range of jobs that Muslim women are prepared to take (Spierings et al., 2010). The results in relation to the intergenerational difference and ethnic differences found in this study might relate to various factors including resilience and adaptation developed by second generation Muslims, as well as the way second generation Muslim women and some ethnic groups such as Indian-Muslims express their religious identity, both through dress code and practice (Brown, 2006).

Similar to previous research, this study demonstrates that education plays a key factor in determining the probability of Muslim women joining the labour market and obtaining salariat jobs (Dale et al., 2002). In the case of Indian-Muslim women, it was shown that their earning returns on qualifications were significantly greater than the returns among White-British Christian women. This being said, the findings in relation to earnings have demonstrated that all Muslim women, except for Black-Muslim women, do not experience any significant penalties in relation to ethnic-religious identity. Such findings support the hypothesis that the main labour market penalty facing Muslim women in Britain is finding jobs that are relevant to their qualifications, but once they do, they tend to receive the same level of returns (Cheung and Heath, 2007; Khattab and Johnston, 2015).

Black-Muslim women were the only group to experience a significant pay penalty. Together with the penalties they experience in unemployment and obtaining salariat jobs, they appear to be the most disadvantaged female Muslim group in the UK. This, to a certain extent, supports initial anticipations and predictions that Black-Muslim women might face additional penalties due to their visible social markers through skin colour (blackness) more than any other Muslim group (Khattab and Modood, 2015).

This study also highlights some interesting findings in relation to the experience of White-British Muslim women. In spite of belonging to the same ethnicity as their British-Christian counterparts, and having the understanding of the labour market as well as the understanding of the British culture, they seem to experience some significant penalties, especially in attaining salariat jobs. This may suggest that they could be facing a Muslim penalty, especially as many of them wear the hijab (Franks, 2000; Moosavi, 2015), which makes them visible and consequently identified as ‘others’, suggesting a racializing processes that is not based primarily on race but on other differences, in this case religious identity. Compared to other Muslim groups (first and second generation), the penalty White-British Muslim women seem to face is much lighter than the penalty facing other Muslim groups, confirming some expectations related to the multiple disadvantaging factors the latter groups may face. However, it is difficult to ascertain the source of such penalty and whether it is directly associated with their conversion to Islam. Although a recent study suggests it might be linked to increased Islamophobia (Moosavi, 2015), further research is required using data on religiosity, practice and attitudes as well as qualitative data on the actual experience of these women to develop further understanding. Moreover, data on their labour market profile and experience before converting to Islam would be needed to allow further analysis and comparison.

The finding in relation to intergenerational differences does not directly confer with that of a recent study of Cheung (2014) in which she found that second generation shows little advancement in all labour outcomes examined, but it lends more support to the migrant assimilation theory (Chiswick, 1978). These differences are significantly large and cannot be attributed to factors other than a converging effect. Further studies using data covering third generation would validate this hypothesis further.

Conclusions

Using pooled LFS data, this study depicts the complex inequality structure facing Muslim women in the labour market. On the one hand, Muslim women are disadvantaged within the labour market but, on the other hand, the extent of their disadvantage varies by their ethnicity migration status (first or second generation) and other social markers. This demonstrates how different social markers and multiple identities have contingent relationships to multiple determinants and outcomes. This may also encompass the different articulation of Muslim identities adopted by different groups of Muslim women and how they are perceived by a wide range of stakeholders, reflecting the variable mode of incorporations including variable opportunities for different groups (Haller et al., 2011). The variation in the experience of first and second generation may suggest differences in religiosity but also differences in specific dress code, expression of own Muslim identity, as well as developing coping strategies and resilience including theological strategies to counter institutional and individual racialization (Brown, 2006). Within such context, the specific impact of migration trajectories, gender, race and religion in terms of specific labour market penalties are more difficult to establish for Muslim women, due to both the diversity within this group as well as the complexities of their own and perceived identity of otherness.

Moreover, despite the importance of religious identity and the level of religiosity in determining the engagement of Muslim women in the labour market, a direct examination of the effect of religiosity was not possible in this study due to lack of information about this factor in the LFS data. While the observed differences between Muslim women and the majority group might be a result of discriminatory practices, further research that takes into account measures of religiosity and other social-cultural factors is required in order to make a clear inference in relation to the labour market penalties that Muslim women face in Britain.