Why migrants earn less: in search of the factors producing the ethno-migrant pay gap in a Dutch public organization

Abstract

In many countries, migrant/ethnic minority workers earn less than non-migrant/ethnic majority employees. This pay gap is not only attributable to migrant/ethnic minority employees having acquired less human capital or social capital, to the impact of government policies and to discrimination. Based on both qualitative and quantitative data collected in 2010, this case study of the job segregation component of the wage disadvantages of migrant employees in a Dutch public organization identifies several other factors. Migrant workers’/ethnic minority employees’ lower levels of participation in work-related communication and the application of socio-ideological labour control also widen this earnings gap. Moreover, migrant workers’/ethnic minority employees’ institutional and relational uncertainties, due to their subordinated position in Dutch society, help to explain their lower levels of participation in work-related communication and how socio-ideological labour control works out negatively for them.

Keywords

discrimination ethnic pay gap ethnic penalties immigrants institutional uncertainty job segregation labour market migrant workers wage disadvantages work-related communication

Introduction

Ethno-migrant inequality¹ is a persistent trait of the labour markets of many countries (Heath, 2007; Van Tubergen et al., 2004). It includes an ethno-migrant pay gap – migrants tend to earn less than non-migrants – which is not only due to migrants having attained less human capital, such as level of education or local language proficiency. Migrants may also have acquired less social capital such as involvement in relevant networks, government policies may either curb or widen the ethno-migrant pay gap and migrants may face discrimination.

The case study of a Dutch governmental organization presented in this article contributes to the literature by showing that migrants’ lower levels of participation in work-related communication compared to non-migrant workers and the implementation of socio-ideological labour control (Alvesson and Willmott, 2002) contribute to widening the ethno-migrant pay gap. Furthermore, both institutional uncertainty (Anderson, 2010) and relational uncertainty (Siebers, 2009b), resulting from subordinated positions migrants hold in Dutch society, help to explain these factors. This case study was carried out in 2010 and is based on diaries, diary interviews and questionnaire data (N = 453).

Literature overview²

In many countries, including Canada (Skuterud and Su, 2012), Sweden (Rydgren, 2004), the UK (Blackaby et al., 2002) and the USA (Hall et al., 2010), migrants earn substantially less than non-migrants. Brynin and Güveli (2012) argue that such pay gaps may consist of occupational penalties and/or wage penalties in a strict sense. Occupational penalties refer to job segregation, i.e. to migrants facing more difficulty in getting access to better-paid jobs and having more chance of ending up in lower paid jobs. By contrast, wage penalties in a strict sense point to migrants receiving lower wages than non-migrant workers when doing the same work or having the same job. So, an ethno-migrant pay gap may have two components: across and within occupations. Brynin and Güveli (2012) found that in Britain the relative weight of both components varies from one group to another, but that overall job segregation is mainly responsible for the pay gap they studied.³

The literature provides several explanations for ethno-migrant pay gaps. First, migrants may have acquired less human capital (e.g. level of education and work experience) and therefore end up in lower paid jobs and receive less pay than their non-migrant colleagues. Studies like Portes and Rumbaut (2006) confirm this explanation. More fine-grained research also looks at country-specific human capital, such as local language skills and having received one’s education in the host country (Rafferty, 2012). Duvander (2001) found that although mastering the Swedish language and having obtained diplomas in Swedish institutions did reduce migrants’ occupational penalty in Sweden, the whole pay gap cannot be explained by differences in country-specific human capital credentials.

The inequality in human capital explanation is also fine-tuned by arguing that migrants may have graduated from less prestigious educational institutes or in degree subjects that pay off less. Migrants may also have received lower grade levels. Rafferty (2012) for Britain and Brekke and Mastekaasa (2008) for Norway controlled for those fine-tuned categories, but nevertheless found considerable pay penalties. Their findings suggest that the same human capital does not pay off equally well for migrants as for non-migrants. This suggestion aligns with studies that show that migrants are more likely than non-migrants to be found in jobs below their levels of education. The under-utilization of migrants’ skills and education (some write about over-education) has been demonstrated, for example by Kler (2006) for Australia, by Turner (2010) for Ireland, by Lindley (2009) and Battu and Sloane (2004) for the UK and Britain (but not for all migrant groups) and Chiswick and Miller (2009) for the USA. Apparently, there is more to ethno-migrant pay gaps than just inequality in human capital, which is why the authors set out to explore other factors that produce ethno-migrant pay gaps.

It is not necessary to start from scratch. The second explanation that the literature offers for ethno-migrant pay gaps is that migrants may have less access to social capital. Because of their recent arrival or because of being over-represented in lower classes, they may have fewer contacts with people who could help them to get a better-paid job. This argument builds upon Bourdieu’s (1986) discussion of social capital and on Ibarra’s (1995) work among different racial groups in the USA and is confirmed for example by studies in Sweden (Behtoui and Neergaard, 2010; Rydgren, 2004), the UK (Bloch, 2013) and Germany (Kalter, 2006).

The third explanation for ethno-migrant pay gaps points to the impact of government policies. Those policies can work out either way. They may aim to reduce such pay gaps through affirmative action and equal opportunities legislation, such as in the USA since the 1960s, but may also be part of the problem of ethno-migrant labour market penalties. Building upon Piore’s (1979) work showing the negative impact of migrants’ legal status of temporariness on their labour market position, Anderson (2010) highlights the precariousness of this position caused by government regulations that give migrant workers a temporary or illegal status. Such migration regulations force migrants into conditions of institutional uncertainty including insecure employment conditions and deteriorating social relations (Anderson, 2010). As a result, migrant workers are willing to accept any work available to them, even when below their level of education, and to say yes to lower payment than non-migrant workers would. This willingness on migrants’ part is a major reason for employers to recruit them for low-skilled jobs, as MacKenzie and Forde (2009) found in the UK and Waldinger and Lichter (2003) in the USA. Bloch’s (2013) study of illegal migrants in Britain confirms these conclusions. Restrictions on citizenship status also work out negatively for migrants’ labour market position (Corluy et al., 2011; Heath, 2007).

Such government policies, including ‘guest workers’ programmes (Heath, 2007), widen the ethno-migrant pay gap. Yet, not only specific regulations fuel ethno-migrant pay inequalities; there is also a strong symbolic and discursive impact of migration policies. Siebers (2010) and Siebers and Dennissen (2014) found that the migrant-hostile discourse reflected in Dutch government policies and media triggers ethno-migrant conflicts in work settings that enlarge the pay gap.

A final explanation for the ethno-migrant pay gap refers to discrimination. In most studies discrimination appears as a residual category, i.e. the unexplained part of the pay gap after controlling for ethno-migrant inequality in human and social capital and other factors. Numerous case studies demonstrate discrimination in work relations, including Van Laer and Janssens (2011) in Belgium and Ogbonna and Harris (2006) in the UK, or in recruitment, such as Andriessen et al. (2012) for the Netherlands and Hoque and Noon (1999) for the UK. However, existing knowledge of how discrimination practices actually produce an ethno-migrant pay gap is limited. Moreover, discrimination cannot be understood as just a residual category as it may be very much intertwined with the factors that have been controlled for. For example, discrimination may be one of the reasons why migrants get less access to influential persons who may provide them with better paid jobs, and governmental policies may trigger discrimination at work (Siebers, 2010; Siebers and Dennissen, 2014).

In brief, more research is needed to explore and unravel the complexity of factors and mechanisms that produce ethno-migrant pay gaps (see McGovern, 2007; Reskin, 2003), especially of those parts of these pay gaps that are not attributable to differences in human capital. In line with Behtoui and Neergaard (2010), it may be argued that case studies will be especially fruitful to detect these factors and mechanisms as they allow for in-depth and detailed analysis of workplace processes. The case study presented in this article aims to contribute to this goal.

Methods

In 2010, research was carried out among employees of the Dutch Ministry of Agriculture, Nature and Food Quality – now part of the Ministry of Economic Affairs – more specifically the Dienst Landelijk Gebied (Department for Rural Areas – DLG). DLG coordinates the planning and development of rural areas in the Netherlands. It is a typically Dutch governmental organization as it focuses on the construction of consensus among various societal parties and stakeholders, in the DLG case provincial and local governments, farmers’ organizations and environmental agencies, although these kinds of organizations are increasingly threatened by current neoliberal government policies. At the time, DLG employed about 1250 civil servants in various regional offices.

The Dutch Central Bureau of Statistics (CBS) classifies migrants as those who have at least one parent born abroad. CBS differentiates this ‘abroad’ into so-called Western countries, so-called non-Western countries and Indonesia/former Dutch East Indies. The non-West includes Latin America, the Caribbean, Africa, the Middle East and Asia, except for Japan and Indonesia. As the aim of this study was to explore factors that produce ethno-migrant pay gaps, it made sense to focus on the largest of these pay gaps. CBS data on the labour market as a whole (statline.cbs.nl; see also Behtoui and Neergaard, 2010 for Sweden) show that non-Western, first and second generation migrants face the most problems in labour market participation compared to non-migrants, so these data underscored the expectation that the pay gap between these two categories of DLG employees would be the largest. Therefore, this study focused on this particular pay gap.⁴ The estimated share of non-Western first and second generation migrants was between 2.5 and 3.2 per cent of the total DLG workforce, which is lower than their 10.8 per cent share of the total population (statline.cbs.nl).

Officially, these people were called ‘non-Western allochthones’ or ‘ethnic minorities’, but these terms have taken on board stigmatizing connotations in public discourse. Therefore, members of the DLG diversity network were asked which terms to use. They suggested the distinction between ‘majority employees’ and ‘bi-cultural employees’, terms that subsequently were used in the research in an emic way. In the following part of this article, the term migrants is used to refer to ethnic minority members/migrants with a first or second generation background in the non-Western areas mentioned above.

First, differences in work-related experiences and in the role that their background plays in everyday life at work were explored to detect factors that might produce the ethno-migrant pay gap. In all, 30 employees, 15 with a non-migration background and 15 with a migration background, were sampled. The aim to spread both sub-samples over gender, age and job levels was not totally achieved since the limited numbers of migrant employees were not equally spread over these categories. Instead, the first 15 migrants that offered to participate were taken and each of them was matched with a non-migration employee with the same gender, age category and job level (see Table 1) to allow for comparison. These pairs do not necessarily work with each other. In addition, the five HR coordinators in the regional sub-units were interviewed to get their views on HR policies and practices.

Table 1

Distribution of same-gender pairs of respondents over age and job level categories.

Job level*/ age category	< 35 years	36–50 years	> 51 years
High	–	F/F	–
Medium	M/M, F/F	M/M, F/F, F/F	F/F
Low	M/M, F/F, F/F	F/F, F/F	M/M, M/M, F/F

Notes: M/M means male pair, F/F means female pair.

High: salary scale 11 or higher, medium: salary scale 8–10, low: salary scale 7 or lower.

Practical reasons ruled out doing direct observation, therefore the diary/dairy interview method (Czarniawska, 2007) was applied. This method offers an extra source of data next to the interviews. It allows one to start the interviews by discussing these self-reported practical experiences and thus to avoid general and abstract answers that may reflect social desirability. First, these 30 respondents were asked to keep a diary for a week to report at the end of each day on what they had done, what aspects were difficult, whether there were events when background played a role and what happened in these events. Then, in semi-structured, open-ended interviews these experiences and events were discussed in more detail; it was asked whether they were incidental or represented a wider pattern; and the subjects’ experiences with HRM practices were discussed. After open, selective and axial coding, statements by non-migrant and migrant respondents were compared to single out differences and subsequently phrase propositions that might explain the ethno-migrant pay gap.

These propositions were used to create a questionnaire drawing on existing scales available in the literature or, if those were not available, scales were developed using characteristic statements of the interview respondents as items. After testing the questionnaire in a meeting with five (both migrant and non-migrant) employees, all DLG employees were mailed to fill in the questionnaire on-line. The response rate was 39.6 per cent (N = 493), with 39 respondents having a first (30) or second (9) generation non-Western migration background. The 40 respondents with Indonesian or Western migration backgrounds were left out of the analysis.

Constructs and measures

Salary scales

The first aim of the questionnaire was to measure the ethno-migrant pay gap. DLG applied a hierarchy of salary scales that corresponded with the hierarchy of job levels. Each job level had its own specific salary scale, scales went from scale 5 to scale 16 and each scale consisted of a hierarchy of about 8 increments. The higher one’s scale or increment, the higher one’s actual salary, but the actual amounts of money of the highest increments of one scale overlapped with those of the lowest increments in the scale above. Access to a higher salary scale or increment did not just depend on age or tenure. Getting an increment was conditional upon a positive performance assessment and one could only pass to a superior salary scale after having successfully applied for a vacancy at the corresponding job level or if one’s current job was upgraded in terms of job level. Neither salary scales nor increments were negotiable. One simply got the salary scale that corresponded with one’s job level and started with the first increment, or, in case of promotion, with one increment above the increment that corresponded with the actual payment one had received in the previous job. HR advisors and interviewed employees agreed that the system actually worked out that way.

The questionnaire test disclosed that many employees were unable to answer a question on their actual payment or increment without consulting their administration. They all knew their salary scale, though. Therefore, the questionnaire asked about their salary scale, not their actual payment. Thus, it actually measured the job segregation or occupational (across occupations) component of the ethno-migrant pay gap indicated by one’s salary scale, not ethno-migrant wage penalties within occupations indicated by one’s increment (see Brynin and Güveli, 2012).

However, at DLG, access to a particular job level and corresponding salary scale primarily determined someone’s actual payment; increment rise had a much smaller effect on one’s salary. Actual payment primarily followed from that access to a particular scale, but not in a linear way since payment differences between two successive scales at the higher end of the hierarchy were substantially larger than at the lower end.

Ethno-migrancy

In line with CBS definitions (see above), the questionnaire asked for place of birth of the respondents themselves and their parents. Subsequently, a classification was made between migrants (first and second generation, non-Western origins) = 1, non-migrants = 0.

Human capital

The second aim of the questionnaire was to explain ethno-migrant inequality in access to salary scales. Part of this inequality could be attributable to ethno-migrant differences in human capital. As human capital variables the following were used: level of education, relevant work experience and Dutch language proficiency (see Heath and Cheung, 2007; Tesser and Dronkers, 2007). Level of education was measured by 2 to 0, corresponding with the categories that DLG uses to differentiate between job levels: academic (2), higher vocational training (1) or lower (0). Tenure in number of years was used as a proxy for relevant work experience. Dutch language proficiency was measured by three items (see Table 2) and had a Cronbach’s alpha of .694. The management indicated that filling in the questionnaire should not take more than 20 minutes; that ruled out including more fine-tuned human capital variables.

Table 2

Variables, items and reliability (translated from Dutch original).

Variable	Items	Cronbach’s alpha
Dutch language proficiency	- Over the whole period I have been working for DLG, I sometimes find it difficult to address others in good Dutch*	0.694
	- I never have found it difficult to understand my colleagues well when they speak Dutch
	- I have always found it easy to read a Dutch text
Getting work and tasks below my capacity	- Currently, I do work below my level of education	0.753
Getting work and tasks below my capacity	- I can do more than my current work demands from me
	- I find it easy to do my work regarding its content
	- I get challenging tasks and assignments that give me a chance to develop myself*
	- Usually the not so nice assignments are assigned to me
	- I often get tasks and assignments that help me to become visible in the organization*
Labour control preferences	- It does not matter how I am, as long as I deliver good work
Labour control preferences	- I would consider it inappropriate if my supervisor assessed me on the question whether I can get along well with my colleagues as long as I do my work properly	0.696
	- The quality of your work and not how you are as a person should matter in your career development
	- I want my supervisor to assess me at performance assessment interviews on the results of my work and not on the characteristics of me as a person
Participation in work-related communication	- In the team, I usually keep my opinions to myself*	0.720
Participation in work-related communication	- I do not like to give my opinion in team meetings*
	- I think it is important to give feedback to colleagues
	- If a colleague does not do his work properly, I comment to him on that

Reversed scores.

Gender, age and class

While inequality in access to salary scales produced by differences in human capital may not be considered problematic, such inequality stemming from factors like gender, age and class background is certainly ‘inappropriate’. Gender was measured as male = 0, female = 1. Age was measured in years. Parents’ level of education was taken as proxy for class background and was measured in the same way as respondents’ level of education (see above). Gender, age and class were used as control variables, next to level of education, tenure and Dutch language proficiency.

Beyond human capital differences

To explore other factors that produce ethno-migrant inequality in salary scales, differences in experiences between non-migrants and migrants emerging from the interview data analysis were taken to develop a number of variables. To construct these variables, typical statements from the interviews and items derived from existing scales in the literature were included as questionnaire items. A five-point Likert scale ranging from 1 = ‘totally disagree’ to 5 = ‘totally agree’ was applied. Variable construction was based on a principal component analysis to check for commonalities in the questionnaire items and subsequently an internal consistency or reliability test was done, taking Cronbach’s alpha values of .7 more or less as a minimum, accounting for the mainly inductive way of construct making.

Only those variables that actually turned out to explain ethno-migrant inequality in salary scales in the quantitative data analysis are discussed here (see Model 2 in Figure 2). The variable Getting work and tasks below my capacity is in line with the under-utilization of migrants’ skills and education as discussed above. Interview statements and items from Greenhaus et al. (1990) were included in a scale that has six items (see Table 2), Cronbach’s alpha .753. The variable Participation in work-related communication was constructed using interview respondents’ statements about their participation in communication about work-related issues with colleagues. This variable has six items, Cronbach’s alpha .720. These variables will be discussed in more detail below.

The interview data analysis suggested that migrant respondents preferred to be assessed on technocratic criteria whereas non-migrant respondents had more positive views on what the literature coins socio-ideological labour control. Here the conceptualizations by Alvesson and Willmott (2002) and Alvesson and Kärreman (2004) were followed. In a technocratic way, employees’ performance is evaluated assessing their input (human capital credentials) and output (quantity and quality of their work). In a socio-ideological way, their personality characteristics (norms and values, identity traits, attitudes) and process (the ways in which they reach their results) are also assessed. A one-dimensional scale between technocratic control preferences and socio-ideological control preferences was constructed using interview statements as items and items from previous research (see Siebers, 2009a). A higher value means more preference for technocratic control; a lower value means more preferences for socio-ideological control. The scale Labour control preferences has four items, Cronbach’s alpha .696.

Next to SPSS, AMOS structural equation modelling software version 19 was used for data analysis. Two categories of fit indices (Curşeu et al., 2007: 132) were applied. First, absolute fit indices show the general fit between the theoretical model and the data. Second, relative fit indices compare the tested model with the null model. The null model suggests that the variables in the model are mutually independent and that there is no covariance among the variables (Widaman and Thompson, 2003). The fit indices CFI, TLI and NFI provide a measure of the fit between the tested model and the null model. The indices used here measure the degree to which constructed models fit real relations between variables in the data file and models were constructed that match the standards set for such fit (see Browne and Cudeck, 1993; Widaman and Thomson, 2003).

Findings

Table 3 shows that migrants are on average located 1.8 salary scales lower in the salary scale hierarchy than non-migrants at DLG. Note, this 1.8 does not refer to one scale and eight increments, one scale is the unit of measurement here. This 1.8 scales inequality boils down to a payment gap of between 415 and 1056 euros a month, depending on whether this inequality is situated at the lower or higher end of the salary scales hierarchy.

Table 3.

Regression analysis of ethno-migrancy on salary scales (N = 453).

Variable	B	B	R² Change	F Change
Ethno-migrancy	−1.821*	−.238*	.057	27.044*

p< .001.

Table 3 reflects a model with a single arrow from ethno-migrancy to salary scales. Subsequently, control variables were included in this model to see to what degree the direct impact of ethno-migrancy on salary scales would diminish after this inclusion. Tenure and class background turned out not to be significantly related to salary scales, so level of education, Dutch language proficiency, gender and age remained as control variables.

Model 1 (see Figure 1) meets all the model fit standards and all regressions are significant. It shows that the inclusion of control variables leads to a drop (compared to Table 3) of the direct effect of ethno-migrancy on salary scale to 0.6. Thus, about two thirds of the ethno-migrant inequality in salary scales is due to the control variables. The remaining 0.6 scales inequality represents a payment gap of between 138 and 352 euros a month depending on whether this inequality is situated at the lower or higher end of the salary scale hierarchy.

Figure 1.

Model 1. Path diagram ethno-migrancy, control variables and salary scales (N = 453).

Model 2 (see Figure 2) includes all the additional variables that turned out to have a mediating effect between ethno-migrancy and salary scale. Model 2 meets the model fit criteria and includes only significant regressions. The direct effect of ethno-migrancy on salary scale is now reduced to -.019 and is no longer significant (p = .945). This means that Model 2 explains the whole effect of ethno-migrancy on salary scale. All the factors that produce the occupational component of the ethno-migrant pay gap have been traced. Table 4 shows the hierarchy in the impact these factors have. These factors will be discussed below and related to the qualitative findings.

Figure 2.

Model 2. Path diagram ethno-migrancy, control variables, other intermediary variables and salary scales (N = 453).

Table 4.

Ranking of variables explaining ethno-migrant inequality in salary scales at DLG.

Rank	Variable	After erasing*	Difference to –.019**
1.	Level of education	−.526	.507
2.	Age	−.283	.264
3.	Getting work and tasks below my capacity	−.200	.181
4.	Gender	−.113	.094
5.	Labour control preferences	−.101	.082
6.	Language proficiency	−.098	.079
7.	Participation in work-related communication	−.088	.069

The direct regression of ethno-migrancy on salary scales adopts this new value after deleting the indicated variable from the model (Model 2).

The difference between the new value of the direct regression of ethno-migrancy on salary scales after deleting the variable from Model 2 and the original value of this regression of -.019 before deleting the indicated variable (see Model 2). The larger this difference, the larger the mediating effect this variable has on the regression of ethno-migrancy on salary scales. In other words, the larger this difference, the stronger the role of this variable is in producing ethno-migrant inequality in salary scales.

Discussion

These findings confirm several explanations of ethno-migrant pay gaps in the literature. First, part of the ethno-migrant inequality in salary scales is produced by human capital disparities, including ethno-migrant disparities in level of education (comes on top in Table 4) and in Dutch language proficiency. Second, migrants tend to work more often on job levels below their level of education and do less challenging or interesting tasks compared to their non-migrant colleagues, which corroborates the literature on the under-utilization of migrants’ skills and education. Apparently, DLG selectors hire migrant candidates for jobs below their level of education, but several non-Western interview respondents indicated that they decided themselves to apply for jobs below their levels of qualification. Their explanations echo what Anderson (2010) calls institutional uncertainty. One of them said:

I did not dare to apply for a higher job … I was scared. It is different from your own country. You do not know what they [people in Dutch work situations] expect from you … I first had to prove to myself: ‘Yes, I can do this’. I was unsure.

Having started one’s DLG career at a job level below one’s qualifications, it was hard to catch up afterwards since vertical mobility was limited and vacancies scarce, as HR advisors indicated. Moreover, one had to proactively approach one’s team leader and project leaders to ask for interesting and challenging tasks that may serve one’s future career (see Greenhaus et al., 1990). Four migrant respondents argued that uncertainty kept them from taking such proactive initiatives, so they were easily overlooked when interesting tasks were distributed. Two non-migrant interview respondents had successfully requested an upgrading of their jobs; no migrant respondent had asked for the same.

Migrant respondents’ uncertainty about their relatively weak labour market position intersected with uncertainty about being able to meet what the organization expected from them. Nine out of 15 migrant interview respondents expressed such uncertainty; no non-migrant interview respondent did the same. Siebers (2009b) found a similar institutional uncertainty among migrant employees in another Dutch organization.

Getting less interesting tasks is connected to lower levels of participation in work-related communication. Migrant employees communicated less often with colleagues than non-migrant employees did, especially regarding work-related issues in team meetings and potentially delicate matters like giving feedback or critique. They tended to avoid such matters and to remain silent in team meetings. Non-migrant interview respondents expressed their eagerness to participate in work-related communication, whereas eight out of 15 migrant respondents indicated that they felt hampered here. HR advisors stated that visibility in work-related communication was needed to get a positive performance assessment and to be spotted as a high potential. Their lower levels of participation in work-related communication harmed migrant employees’ chances of getting promoted (see Table 4).

Levels of participation in work-related communication were not just determined by language proficiency, as Model 2 shows. Interview data indicated that uncertainty kept migrant employees from participating in work-related communication. This uncertainty was relational rather than institutional. Six migrant respondents said they felt uncertain about whether their colleagues and supervisors recognized their migration background as something positive, but only two of them indicated they had experienced negative behaviour by colleagues. It seems that relational uncertainty can persist even without experiences of overt discrimination at work, like Siebers (2009b) found. Recognition of migrants in Dutch society is far from self-evident and migrants take this uncertainty to work, reinforced by subtle and implicit forms of exclusion (see Van Laer and Janssens, 2011). One interviewee said, ‘You are an outsider.’

In particular HR advisors attributed their migrant colleagues’ lower levels of participation in work-related communication and less access to interesting tasks to them being less insolent, more modest, more introvert, less assertive, etc. These findings recall what DiFruscia (2012) calls the individualization and psychologization of work relations. She argues that HRM takes workers in work relations as individuals. It understands these relations as depending on these individuals’ psychological traits, thereby ignoring their social and collective embeddedness. Likewise, HRM draws attention away from the social and interactive nature of communication by redefining it as something that is determined by individuals’ ‘communicative skills and competences’. Similarly, non-migrant interview respondents said their migrant colleagues’ difficulties in participating in work-related communication stem from shortcomings in their individual ‘competences’ and personality traits. Thus, by redefining work-related communication as something that depends on an individual’s communication skills, non-migrants are able to blame their migrant colleagues themselves for their lower participation in work-related communication by assuming lower levels of communication skills on their part. In this way, they hide their own role in limiting migrant colleagues’ room for participation.

This psychologization of communication is closely related to labour control preferences. Model 2 shows that, in relative terms, migrant employees wanted their supervisors to assess them applying technocratic standards (quality of the results of their work), whereas non-migrant employees rather want to be assessed on socio-ideological standards (identity, personality characteristics, getting along well with your colleagues). Differences in these preferences produced ethno-migrant inequality in access to salary scales (Table 4), since socio-ideological aspects played an important role in performance assessments at DLG.

The form used to structure performance assessment interviews indicated that employees were not only assessed on the quality of their work (product) but also on how they did their work (process). These process aspects refer to ‘competences’ that employees should have. A very elaborate competences profile document prescribed the various competences needed for each job level and function. Except for the ‘charismatic personality’ prescribed for higher management jobs, these competences were not so much phrased in socio-ideological terms of attitudes or personality characteristics – what you are supposed to be – but in rather classical terms of what employees are supposed to do and know (cf. Grugulis et al., 2004). Nevertheless, for most functions, eight or nine competences were stipulated, each competence specified into four to seven sub-competences that work cumulatively from lower to higher functions. Taken together, each competence profile came close to a detailed and encompassing prescription of a role (process) an employee was supposed to fulfil. The scope of control was wide and detailed.

The language in which these competences were phrased was very personalized and fuzzy. It recalled Urciuoli’s (2008) discussion of the denotational indeterminacy of ‘soft skills’ like ‘communication skills’ that play a central role in HRM discourse but have no fixed or stable referent, i.e. they have no fixed or unambiguous object in the world to which they would refer. She argues that this vagueness, fuzziness (cf. Moss and Tilly, 1996) and ambiguity give free space to an assessor or selector to decide whether the employee or candidate has a particular competence. For example, regarding the competence Passing Judgement, the DLG document said: ‘[The employee] makes the right decisions and draws the right conclusions’. What this ‘right’ meant remained unspecified, so it was up to the assessor or selector to interpret, drawing on his or her subjectivity, preferences and dislikes. So, although the introduction of competence profiles at DLG was probably meant to curtail arbitrariness in assessment and selection procedures, the socio-ideological framing of these profiles created new opportunities for such arbitrariness to re-emerge.

All interviewees said that they were assessed on aspects of socio-ideological control and several respondents pointed to the subjective nature of those aspects. However, only migrant interview respondents expressed their critique of socio-ideological control aspects. Similar to Campbell and Roberts (2007), who showed that migrants find it more difficult than non-migrants to integrate their personal discourse into the organizational normative discourse in application interviews, the non-migrant interview respondents were very well able to discursively relate their own persons to the required competences and to talk about their personal ability to develop such competences, whereas most migrant interview respondents talked about these competences in a general way without making any connection to themselves. Those who did make that connection to themselves did so to express their disapproval of their supervisor’s assessment.

Given their institutional (Anderson, 2010) and relational (Siebers, 2009b) uncertainty, migrant employees had good reasons to dislike any kind of labour control that entailed a wide scope of assessment aspects and that brought their identity and personality under supervisors’ control. Their recognition as a good colleague was uncertain for them (Siebers, 2009b) and therefore any control focusing on their identity and personality aspects represented a risk to them. The assessment of these aspects relied heavily on subjective interpretations by both their supervisors and their colleagues. Assessments applying technocratic control criteria (quality and quantity of output) can be measured more objectively and depend less on supervisors’ subjectivity (Moss and Tilly, 1996). Thus it becomes intelligible that the application of socio-ideological control at DLG fuelled the production of ethno-migrant inequality in access to higher salary scales.

Conclusions

The findings of this study confirm the roles of human capital disparities and of the under-utilization of migrants’ skills and education in producing a considerable across-occupations pay gap, although it is unknown whether this across-occupations pay gap is (partially) offset or reinforced by within-occupations pay differences. This study contributes to an understanding of how ethno-migrant pay gaps are generated by showing that migrants’ lower levels of participation in work-related communication as well as the application of socio-ideological labour control constitute two other factors that fuel unequal payment. Migrant employees’ lower levels of participation in work-related communication harm their career perspectives and the psychologization of communication in terms of competences leads to migrants being blamed themselves for their lower participation. The application of socio-ideological labour control (Alvesson and Willmott, 2002) harms the interests of migrant employees since it increases their risk of non-recognition (Siebers, 2009b). The denotational indeterminacy (Urciuoli, 2008) of socio-ideological criteria like competences and soft skills opens the door for the arbitrary subjective preferences and dislikes of selectors and assessors to become operational.

Furthermore, this study suggests that migrants’ institutional and relational uncertainty is very much intertwined with their lower levels of participation in work-related communication and with the negative consequences socio-ideological control has for them. Anderson (2010) and Piore (1979) argue that institutional uncertainty characterizes migrants with a temporary or illegal status, but this study suggests that migrants with a permanent and legal status also suffer from such uncertainty; institutional uncertainty may thus be something that characterizes the position of migrants as such.

Institutional and relational uncertainty should not be seen as personality characteristics but as social phenomena, stemming from migrants’ position in society. Institutional uncertainty relates to migrants’ subordinated position in the Dutch labour market (Tesser and Dronkers, 2007) as well as to the impact of migration legislation (Anderson, 2010). Their relational uncertainty stems from migrants’ ethnic markers (skin colour, language accents, dress, specific rituals like Ramadan, etc.) that render the recognition by their colleagues uncertain – even when no overt acts of discrimination occur (Siebers, 2009b) – fuelled by migrant-hostile discourses in Dutch politics and media (Siebers, 2010; Siebers and Dennissen, 2014).

The current study’s limitations call for future research. First, the interviews gave no indications of social capital differences so no social capital items were included in the questionnaire and Model 2 was able to explain the total ethno-migrant across-occupations pay gap. This limited role of social capital differences contrasts with the findings of studies like Behtoui and Neergaard (2010) and may be due to the high degree of formalization and structuration of application and promotion procedures at DLG. Further research is needed to see whether this formalization and structuration moderate the impact of social capital differences.

Second, the impact of socio-ideological labour control and the importance attributed to visibility in work-related communication are certainly not limited to Dutch governmental organizations and may be found particularly among higher-level jobs. However, the current study’s number of interview respondents among those jobs was limited, so more fine-grained research on those jobs, also in the private sector, is recommended.

Finally, the factors that produce the ethno-migrant pay gap may result from the migrant status and/or from the ethnicity of first and second generation migrants (cf. note 1), but current knowledge does not allow the relative weights of migrancy and ethnicity to be determined. Some examples: this study found that part of the pay gap between migrants and non-migrants stems from the former having acquired lower levels of human capital than the latter, but the relative weight of migrancy – because they may have come from a country with a lower developed or less relevant education system with spill-over effects to their children – or ethnicity – because they may face ethnic discrimination in the (adult) educational system in the resident country – is unknown. Next, studies like Behtoui and Neergaard (2010) found that migrants tend to have acquired less social capital than non-migrants, but it is not known whether this discrepancy is explained by migrants having had less time to develop such capital due to their recent arrival (migrancy) or by non-migrants’ unwillingness to develop relationships with them because they may identify them as the ‘ethnic other’ (ethnicity). In addition, migrants’ lower levels of participation in work-related communication may be produced by their recent arrival in the country, therefore having few clues as to what is exactly expected from them (migrancy), or by the risk that their ethnic markers may trigger ascribed ethnic identification and discrimination by their colleagues (ethnicity). More research is called for to unravel the complex intertwining of such migrancy and ethnicity effects.

Footnotes

Acknowledgements

Our warm gratitude goes to DLG officials, especially Hetty van Leur and Andries Bouma, for enabling us to carry out our research in their organization, and to the reviewers for their useful comments to previous drafts of this article, as well as to Paul Mutsaers, several student assistants and all others who made our research possible.

Funding

This research received no specific grant from any funding agency in the public, commercial, or not-for-profit sectors.

Notes

Hans Siebers is an associate professor at Tilburg University, the Netherlands. His research focuses on superdiversity, ethnic identity formation and the production of ethnic inequality in work and educational settings as well as on discourses of ethnic exclusion in media and politics. He is the director of the master’s programme: Management of Cultural Diversity.

Jilles van Gastel is a lecturer at the Austral University of Chile. Before completing her PhD at Wageningen University in the Netherlands she held posts in the Dutch Ministry of Foreign Affairs. She is interested in the study of organizations, public policy, meaning-making at work and ethnography.

References

Alvesson

Kärreman

(2004) Interfaces of control: technocratic and socio-ideological control in a global management consultancy firm. Accounting, Organizations and Society 29(3/4): 423–44.

Alvesson

Willmott

(2002) Identity regulation as organizational control: producing the appropriate individual. Journal of Management Studies 39(5): 619–44.

Anderson

(2010) Migration, immigration controls and the refashioning of precarious workers. Work, Employment and Society 24(2): 300–17.

Andriessen

Nievers

Dagevos

(2012) Ethnic discrimination in the Dutch labour market: its relationship with job characteristics and multiple group membership. Work and Occupations 39(3): 237–69.

Battu

Sloane

(2004) Over-education and ethnic minorities in Britain. Manchester School 72(4): 535–59.

Behtoui

Neergaard

(2010) Social capital and wage disadvantages among immigrant workers. Work, Employment and Society 24(4): 761–79.

Blackaby

Leslie

Murphy

O’Leary

(2002) White/ethnic minority earnings and employment differentials in Britain: evidence from the LFS. Oxford Economic Papers 54(2): 270–97.

Bloch

(2013) The labour market experience and strategies of young undocumented migrants. Work, Employment and Society 27(2): 272–87.

Bourdieu

(1986) The forms of capital. In: Richardson

(ed.) Handbook of Theory and Research for the Sociology of Education. New York, NY: Greenwood Press, 241–58.

10.

Brekke

Mastekaasa

(2008) Highly educated immigrants in the Norwegian labour market: permanent disadvantage? Work, Employment and Society 22(3): 507–26.

11.

Browne

Cudeck

(1993) Alternative ways of assessing model fit. In: Bollen

Long

(eds) Testing Structural Equation Models. Newbury Park, CA: SAGE, 136–62.

12.

Brynin

Güveli

(2012) Understanding the ethnic pay gap in Britain. Work, Employment and Society 26(4): 574–87.

13.

Campbell

Roberts

(2007) Migration, ethnicity and competing discourses in the job interview: synthesizing the institutional and personal. Discourse and Society 18(3): 243–71.

14.

Chiswick

Miller

(2009) The international transferability of immigrants’ human capital. Economics of Education Review 28(2): 162–9.

15.

Corluy

Marx

Verbist

(2011) Employment chances and changes of immigrants in Belgium: the impact of citizenship. International Journal of Comparative Sociology 52(4): 350–68.

16.

Curşeu

Stoop

Schalk

(2007) Prejudice towards immigrant workers among Dutch employees: integrated threat theory revisited. European Journal of Social Psychology 37(1): 125–40.

17.

Czarniawska

(2007) Shadowing and Other Techniques for Doing Fieldwork in Modern Societies. Milton Park: Marston Book Services.

18.

DiFruscia

(2012) Work rage: the invention of a human resource management anti-conflictual fable. Anthropology of Work Review 33(2): 89–100.

19.

Duvander

AZE

(2001) Do country-specific skills lead to improved labor market positions? An analysis of unemployment and labor market returns to education among immigrants in Sweden. Work and Occupations 28(2): 210–33.

20.

Greenhaus

Parasuraman

Wormley

(1990) Effects of race on organizational experiences, job performance evaluations, and career outcomes. Academy of Management Journal 33(1): 64–86.

21.

Grugulis

Warhurst

Keep

(2004) What’s happening to skills? In: Warhurst

Keep

(eds) The Skills that Matter. Basingstoke: Palgrave Macmillan, 1–18.

22.

Hall

Greenman

Farkas

(2010) Legal status and wage disparities for Mexican immigrants. Social Forces 89(2): 491–513.

23.

Heath

(2007) Crossnational patterns and processes of ethnic disadvantage. In: Heath

Cheung

(eds) Unequal Chances: Ethnic Minorities in Western Labour Markets. Oxford: Oxford University Press, 639–95.

24.

Heath

Cheung

(2007) The comparative study of ethnic minority disadvantage. In: Heath

Cheung

(eds) Unequal Chances: Ethnic Minorities in Western Labour Markets. Oxford: Oxford University Press, 1–44.

25.

Hoque

Noon

(1999) Racial discrimination in speculative applications: new optimism six years on? Human Resource Management Journal 9(3): 71–82.

26.

Ibarra

(1995) Race, opportunity, and diversity of social circles in management networks. Academy of Management Journal 38(3): 673–703.

27.

Kalter

(2006) In search of an explanation for the specific labor market disadvantages of second generation Turkish migrant children. Zeitschrift für Soziologie 35(2): 144–60.

28.

Kler

(2006) Graduate overeducation and its effects among recently arrived immigrants to Australia: a longitudinal survey. International Migration 44(5): 93–128.

29.

Lindley

(2009) The over-education of UK migrants and minority ethnic groups: evidence from the Labour Force Survey. Economics of Education Review 28(1): 80–89.

30.

McGovern

(2007) Immigration, labour markets and employment relations: problems and prospects. British Journal of Industrial Relations 45(2): 217–35.

31.

MacKenzie

Forde

(2009) The rhetoric of the ‘good worker’ versus the realities of employers’ use and the experiences of migrant workers. Work, Employment and Society 23(1): 142–59.

32.

Moss

Tilly

(1996) ‘Soft’ skills and race: an investigation of black men’s employment problems. Work and Occupations 23(3): 252–76.

33.

Ogbonna

Harris

(2006) The dynamics of employee relationships in an ethnically diverse workforce. Human Relations 59(3): 379–407.

34.

Piore

(1979) Birds of Passage: Migrant Labour and Industrial Societies. Cambridge: Cambridge University Press.

35.

Portes

Rumbaut

(2006) Immigrant America: A Portrait. Berkeley, CA: University of California Press.

36.

Rafferty

(2012) Ethnic penalties in graduate level over-education, unemployment and wages: evidence from Britain. Work, Employment and Society 26(6): 987–1006.

37.

Reskin

(2003) Including mechanisms in our models of ascriptive inequality. American Sociological Review 68(1): 1–21.

38.

Rydgren

(2004) Mechanisms of exclusion: ethnic discrimination in the Swedish labour market. Journal of Ethnic and Migration Studies 30(4): 697–716.

39.

Siebers

(2009a) (Post)bureaucratic organizational practices and the production of racioethnic inequality at work. Journal of Management and Organization 15(1): 62–81.

40.

Siebers

(2009b) Struggles for recognition: the politics of racioethnic identity among Dutch national tax administrators. Scandinavian Journal of Management 25(1): 73–84.

41.

Siebers

(2010) The impact of migrant-hostile discourse in the media and politics on racioethnic closure in career development in the Netherlands. International Sociology 25(4): 475–500.

42.

Siebers

Dennissen

MHJ

(2014) Is it cultural racism? Discursive oppression and exclusion of migrants in the Netherlands. Current Sociology. Epub ahead of print 9 October 2014. DOI: 10.1177/0011392114552504.

43.

Skuterud

(2012) Immigrants and the dynamics of high-waged jobs. Industrial and Labor Relations Review 65(2): 377–97.

44.

Tesser

Dronkers

(2007) Equal opportunities or social closure in the Netherlands. In: Heath

Cheung

(eds) Unequal Chances: Ethnic Minorities in Western Labour Markets. Oxford: Oxford University Press, 359–401.

45.

Turner

(2010) The jobs immigrants do: issues of displacement and marginalization in the Irish labour market. Work, Employment and Society 24(2): 318–36.

46.

Urciuoli

(2008) Skills and selves in the new workplace. American Ethnologist 35(2): 211–28.

47.

Van Laer

Janssens

(2011) Ethnic minority professionals’ experiences with subtle discrimination in the workplace. Human Relations 64(9): 1203–27.

48.

Van Tubergen

Maas

Flap

(2004) The economic incorporation of immigrants in 18 Western countries: origin, destination, and community effects. American Sociological Review 69(5): 704–27.

49.

Waldinger

Lichter

(2003) How the Other Half Works: Immigration and the Social Organization of Labor. Berkeley, CA: University of California Press.

50.

Widaman

Thompson

(2003) On specifying the null model for incremental fit indices in structural equation modeling. Psychological Methods 8(1): 16–37.

Why migrants earn less: in search of the factors producing the ethno-migrant pay gap in a Dutch public organization

Abstract

Keywords

Introduction

Literature overview 2

Methods

Constructs and measures

Salary scales

Ethno-migrancy

Human capital

Gender, age and class

Beyond human capital differences

Findings

Discussion

Conclusions

Footnotes

Acknowledgements

Funding

Notes

References

Literature overview²