Civic stratification and social positioning: CEE-labour migrants without a work permit Erik Snel, Marije Faber, Godfried Engbersen

Transcription

1 Civic stratification and social positioning: CEE-labour migrants without a work permit Erik Snel, Marije Faber, Godfried Engbersen Department of Sociology Erasmus University Rotterdam PO Box DR Rotterdam The Netherlands E mail: snel@ fsw.eur.nl Abstract Civic stratification is a relatively new dimension of social inequality in contemporary Western societies. In their intention to attract desired immigrants and to avoid undesired ones, states grant, or deny, diferent juridical statuses and related social and economic rights. EU-nationals, for instance, have ful rights to settle and work in other EU-countries. Most EU-countries made an exception for nationals of new member states like Romania and Bulgaria, who joined the EU in Like before, Romanian and Bulgarian nationals need a work permit to work in other EU-countries. However, many Romanians and Bulgarians came to countries like the Netherlands without having a work permit. This studies examines what this ambivalent legal position of many Romanian and Bulgarian nationals means for their social position. Using data from a survey among 654 CEE-labour migrants, we distinguish four migrant categories: 1) Romanians and Bulgarians with a work permit; 2) Romanians and Bulgarians without a work permit; 3) Romanians and Bulgarians who don t need a work permit; and 4) Polish labour migrants who are in a diferent legal position. We use three measures of the social position of migrants: occupational status, earned hourly incomes, and housing position. 1. Introduction: Migration researchers pointed out new types of social inequality in current immigration societies. Traditionaly, social inequality in Western capitalist societies is perceived in terms of social class, which is closely related to (and often measured by) the occupational status and educational levels of individuals. Class inequality in modern Western societies is complemented by other social cleavages such as gender, ethnicity, health, age, sexual preference and consumption (Bradley 1996, Crompton 2008). Civic stratification is yet another, relatively new dimension of social inequality in contemporary Western societies. In their intention to attract desired immigrants and to avoid undesired ones, states grant, or deny, different juridical statuses and related rights with regard to admission, residence, work, and access to social security, housing, health, education, etcetera to various categories of immigrants. The result is a highly fragmented and complex hierarchy of divergent social and economic rights, granted or denied by the state, and related social opportunities that different categories of migrants have in receiving societies; migration researchers speak of civic stratification (Koffman and Kraler 2006; Morris 2003; 2004; 2010). As Morris argues, this notion goes beyond more conventional distinctions like legal versus ilegal (or documented versus undocumented ) migrants or between citizens, denizens, and aliens (Hammar 1990). The notion of civic stratification refers to a far more fragmented hierarchy of different statuses and related social and economic right. As an example of civic stratification, Morris (2003: 80) refers to the complex regulations with regard to EU citizenship. EU legislation makes a fundamental distinction, next to national or naturalized citizens, between

2 migrant residents from other EU countries and migrants from non-eu-countries ( third country nationals ). As freedom of movement is a basic right in EU-legislation, EU-nationals have ful rights to settle and work in other EU countries. The EU tends to see the movements of EU-nationals from one EU-country to another not so much as migration, but as free internal EU-mobility. Migration, on the other hand, refers to the admission of non-eu-nationals to one of the EU-member states and is highly regulated with complex regulations related to admission, residence, access to the labour market and social security, etcetera. However, to make things even more complex, the old -EU-member states made a temporary exception on this general rule of freedom of movement for citizens of the new-eu-member states of 2004 and i After the EU-enlargement of 2004, most old EU-countries i applied a so-caled transitional period for nationals from the new member states. D uring this transitional period, nationals from the new m em ber states could travel to and settle in the old-eucountries but they had only lim ited access to the labour m arket (H o land et. al. 2011). According to EUlegislation this transitional period could last for seven years at most. The Netherlands ended the transitional period for citizens from the new member states of 2004 three years later, in May Also in 2007, Romania and Bulgaria joined the EU. Again, the Dutch government applied the transitional period giving Romanian and Bulgarian nationals only limited access to the (formal) Dutch labour market. Aside from some exceptions (to be discussed later), Romanian and Bulgarian nationals stil need a work permit (Dutch acronym TWV) to be employed in the Netherlands. It is already announced that the transitional period for Romanians and Bulgarians wil last the fu l seven years that are alowed by the EU ịi As a result of these regulations, Romanian and Bulgarian labour migrants who come to the Netherlands are in a rather ambivalent legal situation. As EU-citizens, they have free access to Dutch territory but they have only limited access to (formal) Dutch labour market. Like non-eu-migrants, Romanians and Bulgarians stil need a work permit (TWV) in order to work in the Netherlands. However, significant numbers Romanian and Bulgarian nationals came to the Netherlands without having the required work permit. As they are unable to find formal employment, many of them end up in informal work arrangement. Ruhs and Anderson (2010) introduced the term semi-compliance to describe the kind of ambivalent situations Romanian and Bulgarian labour migrants that fail to have the required work permit are in. The authors distinguish three levels of compliance : Compliant migrants are legaly resident and working in ful compliance with the employment restrictions attached to their immigration status. Non-compliant migrants are those without the rights to reside in the host country (i.e. those ilegaly resident ). Semi-compliant indicates a situation where a migrant is legaly resident but working in violation of some or al of the employment restrictions attached to his/her immigration status. (Ruhs & Anderson 2010: 201) As examples of semi-compliance they refer to foreign students, au pairs and asylum seekers, who are legaly residents but work in breach of the employment restrictions attached to their immigration status (for instance, too many hours). Romanian and Bulgarian migrants may be in a comparable situation. As EU-citizens, they are legal residents in countries like the Netherlands. But if they work without having the required work permit, they

3 work in breach of the employment restrictions attached to their immigration status. To be sure, this complex hierarchy of differentiated immigration statuses and related social and economic rights is created by the state in order to classify and control immigration. Civic stratification and semi-compliance are in this broad sense produced by the state and thus socialy constructed (Morris 2003; Ruhs & Anderson 2010: 196; cf. Engbersen & Van der Leun 2001; Bosniak 2006; Samers 2004). This article examines the social position of four different categories of labour migrants from various Central and Eastern-European (CEE) countries in the Netherlands. The labour migrants in our survey originate three different countries: Poland, Romania and Bulgaria. We wil focus in particular on Romanian and Bulgarian labour migrants in the Netherlands. We distinguish three categories among them: they either have or fail to have the required work permit, or they formaly do not need a work permit ( not applicable ) (for reasons we wil explain later). We compare these three categories of Romanian and Bulgarian labour migrants with Polish labour migrants who have ful access to the Dutch labour market. We classify Polish labour migrants, Romanians and Bulgarians with a work permit, and Romanians and Bulgarians who don t need a work permit as compliant m igrants. Rom anians and Bulgarians who work in the N etherlands without the required work perm it are sem i- compliant migrants (cf. Ruhs and Anderson s 2010). As EU-nationals they are legaly resident in the Netherlands, but they have limited access to the Dutch formal labour market. There are no non-compliant migrants in our sample. A l respondents as EU-citizens reside legaly in the Netherlands. Folowing the theory of civic stratification, we assume that legal position of migrants has an independent effect (particularly independent from social class) on their achieved social position. We wil use three different indicators to measure the social position of our respondents: occupational status, earned hourly incomes, and housing position. We expect Polish labour migrants to have the highest social position since they have ful and free access to the Dutch labour market and related social rights (Hypothesis 1). Furthermore, we expect Romanian and Bulgarian labour migrants who have a work permit or do not need one to have a higher social position than Romanians and Bulgarians that work in the Netherlands without the required work permit (Hypothesis 2). In the remainder of this paper, we wil first sketch the development of CEE-labour migration to the Netherlands since the EU-enlargement of 2004 (section 2). Then, we wil present the data used in this study that are derived from a survey under 654 Polish, Romanian and Bulgarian labour migrants in the Netherlands (cf. Engbersen et al. 2011) (section 3) before we wil present our findings with regards to achieved social position (occupational status, earned hourly wages and housing position) of Polish, Romanian and Bulgarian labour migrants in the Netherlands (section 4). We wil conclude with a discussion of our findings (section 5) 2. Recent CEE labour migration to W estern Europe and the Netherlands

4 The EU-enlargements of 2004 and 2007 resulted in significant new migration flows within Europe. Particularly Polish nationals settled, temporary or permanently, in large numbers to work in various old EU-countries. Although initialy the UK and Ireland were the favourite destination countries for Central and Eastern European (CEE) labour migrants, the Netherlands also received a relatively large numbers of them (cf. Black et al. 2010; Drinkwater, Eade and Garapich 2010; Engbersen et al. 2010; Dagevos 2011). These migration flows were partly a continuation of migration patterns that already existed before After the fal of communism, several Western European countries concluded bilateral treaties with countries such as Poland, Bulgaria and Romania to enable temporary labour migration in particular for agricultural work. After 2004, however, the total number of CEE-nationals living and working in Western Europe increased rapidly. In the Netherlands, for instance, the total number of CEE-nationals living (at least being registered) by Dutch population registers increase from about 41,000 in 2004 to almost 117,000 in About half of these CEE-nationals in the Netherlands came from Poland. The number of Poles residing in the Netherlands increased from 13,000 in 1996 to over 20,000 in Since 2004, the number of Poles in the Netherlands more than tripled (almost 70,000 in 2011). In comparison, the number of Romanians and Bulgarians in the Netherlands is much lower (12,000 Romanians and 15,000 Bulgarians in 2011) (figure 1).

5 Figure 1 Number of residents* from Poland, Romania and Bulgaria in the Netherlands ( ) Poland Romania Bulgaria Other CEE-countries *individuals born in the country of origin with at least one parent born in the country of origin ( first-generation allochtonous ) and registered in Dutch population registrations (GBA). Source: CBS, Statline (own computations) However, as these data are derived from the formal Dutch population registrations and many CEE-labour migrants appear not to register when arriving in the Netherlands (Engbersen et al. 2011), the actual numbers of them are probably much higher. Recent research estimates that in 2009, there were about 305,000 (between 285,000 and 325,000) CEE-migrants present in the Netherlands, either registered or not. About half of them came from Poland (between 134,000 and 154,000), about 100,000 (between 93,000 and 116,000) came from either Romania or Bulgaria, the remaining 57,000 (between 53,000 and 63,000) came from one of the other CEE-countries (Van der Heyden, Cruyff and Van Gils 2011). Another crucial figure relates to the number of Dutch work permits (TWV) issued to CEE-nationals. Before giving the numbers, we should explain the formal regulations with regards to work permits. According to Dutch migration policies, non-eu-nationals require a work permit in order to be employed in the Netherlands. At the moment, this also applies to Romanian and Bulgarian nationals, the new EU member states of However, three categories are formaly exempted from the TWV-requirement. Firstly, self-employed workers do not need a work permit, provided that they are formaly registered either in the Netherlands or in their home country. Secondly, knowledge migrants are exempted from the TWV-requirement. According to Dutch regulations, knowledge migrants are foreign workers with either a high income or specific occupations (for instance, working at Dutch universities or other research institutions) ịv Thirdly, posted workers are exempted from the TWV requirement. Posted workers are persons employed in one EU member state but sent by their employer on a temporary basis to carry out his work in another member state (European Commission, 2012). Since the rights and working conditions of posted workers are recorded in the Posting of Workers Directive legislated by the European Commission v, the Dutch labour legislation does not apply on this specific labour group. Later, we wil further explore the social position of these (posted workers) posted workers (section 4.2). It is the employer, not the migrant, who applies for a work permit (TWV). Before the employer can apply for a work permit for a non-eu-national, he should check whether or not there is someone available for the job in the Netherlands or elsewhere in the EU; the labour market test. Only when this so-caled priority supply is not available, the employer is alowed to hire a non-eu worker (or someone from Romania or

6 Bulgaria). A work permit is only valid for a specific employer, a specific employee as wel as for those activities for which the permit is issued. This implies that foreign workers are not alowed to do other work (or work for other em ployers) with that work perm it. W hen the transitional period was applied to Polish workers ( ), the Dutch state in practice did not require the labour market test from employers. However, given the current unemployment in the Netherlands and elsewhere in the EU, the (former) Dutch cabinet indicated that it wil only issue work permits for Romanians or Bulgarians in exceptional cases. Figure 2 Number of work permits (TWV) issued in the Netherlands ( *) Totaal Polen Roemenie, Bulgarije Source: CWI; De Boom et al. 2010: 58 (our own computation) (2010 estimated numbers) Figure 2 shows the number of work permits (TWV s) issued by the Dutch government to CEE-nationals between 1996 and The figures relate to issued work permits, not to individuals receiving a work permit. A foreign worker may obtain several work permits per year (for different jobs or different employers). Figure 2 gives some interesting clues. Firstly, it shows that the total number of work permits issued over the years is strongly related to the number of work permits issued to CEE-labour migrants (particularly from Poland). Secondly, the figure again shows that Polish nationals were already working in the N etherlands before the EUenlargement in The number of work permits issued to Polish nationals increased from about 2,500 in the year 2000 to 10,000 in Most of these work permits were issued for work in the Dutch horticulture (De Boom et al. 2010). After the EU-enlargement, the number of work permits issued to Polish nationals mushroomed from 20,000 in 2004 to not less than 54,000 in Since May 2007, Polish nationals do not need a work permit anymore in the Netherlands. This explains the declining numbers of work permits issued to Poles as wel as the drastic decline in the total number of work permits issued in the Netherlands since Thirdly, figure 2 shows that relatively few work permits are issued to Romanian and Bulgarian nationals, also in the years after both countries acceded the EU in In 2008, only about 3,000 work permits were issued to Romanians and ample 1,000 to Bulgarian nationals. In later years, the number of work permits issued to Romanian and Bulgarian nationals did not increase. This shows that there is a wide gap between the total number of Romanians or Bulgarians present in the Netherlands (more than 24,000 according to official registrations or 57,000 according to the estimations of Van der Heyden et al. 2011) on the one hand and the number of work permits issued to Romanians and Bulgarians nationals (about 4,000 in 2010) on the other hand.

7 As it is unlikely, that al these Romanians and Bulgarians residing in the Netherlands are students or otherwise non-working individuals vi these figures make very clear that there are significant numbers of Romanian and Bulgarian nationals present in the Netherlands, who work (or want to work) but fail to have the required work permit. Our survey examines how they cope with this ambivalent situation. 3. Data and methodology The survey used in the paper examines the labour market position and incorporation of labour migrants from Poland, Bulgaria and Romania in the Netherlands. The survey was conducted between October 2009 and February We interviewed 654 CEE labour migrants in nine different Dutch municipalities, including two major cities (Rotterdam, The Hague), two middle-sized cities (Breda, Dordrecht) and several rural towns (mainly in agricultural areas). We used a structured questionnaire with 213 questions about issues such as the migration history of respondents, their labour market and housing position, incorporation in Dutch society, transnational activities, and their future plans. The face-to-face interviews were held in the mother tongue of the respondents, for the duration between one and two hours. Native speaking interviewers recruited respondents at places frequented by CEE labour migrants like Polish shops, churches or on internet forums. Respondents were also approached on the street by the interviewers upon hearing their mother tongue or by identifying their national car plates. We emphasised the diversity of sources from which the respondents are selected. For instance, we were careful not to overuse certain locations. Respondents were also recruited by snowbal sampling. At the end of each interview, respondents were asked whether they knew of felow nationals who could be interviewed. In our view, the survey has led to valid results. Sensitive questions on irregular work ( verbal contract ) and on not having a valid work permit were answered without reservation. We learned, for example, that 41 per cent of the Bulgarians had a verbal contract and that 35 per cent of the Romanians and 65 per cent of the Bulgarians resided in the Netherlands without a work permit. We are dealing with a heterogeneous group that includes both highly skiled and low-skiled respondents (cf. Engbersen et al. 2011) (see table 1). The size of the Bulgarian group is proportionaly large because we gave extra attention to this category for two reasons. Firstly, some of the municipalities participating in the research were specificaly concerned about unemployed Bulgarian nationals as wel as about those working informal (or even involved in crime). Our survey intended to find out whether these worries are justified. Secondly, we wanted to examine the labour market position of Bulgarians given the fact that this group, together with Romanian nationals, stil require work permits in order to work in the Netherlands. Level of compliance The main independent variable in the analyses is someone s position in the civic stratification hierarchy, or the level of compliance (Ruhs and Anderson 2010). This variable requires some explanation. As already explained, CEE-labour migrants as EU-citizens by definition reside legaly in the Netherlands. This means, the category non-compliant is not the issue. Romanian and Bulgarian labour migrants are semi-compliant if they work without having the required work permit. In that case, they work in breach of (one of) the employment restrictions attached to his/her immigration status. Romanian and Bulgarian labour migrants are compliant if they

8 work and have the required work permit. As explained, some specific categories of Romanian or Bulgarian labour migrants do not need a work permit (either registered self-employed workers or posted workers ); in these cases, the level of compliance is coded not applicable. Polish labour migrants who do not need a work permit are incorporated in the analyses as a reference category. As a result, we distinguish four categories among the respondents: (1) Romanians and Bulgarians who work with a work permit; (2) Bulgarians and Romanians who work without a work permit; (3) working Bulgarians and Romanians who don t need a work permit ( not applicable ); and (4) Polish labour migrants. In the survey, Romanian or Bulgarian respondents were asked whether or not they had a work permit. The problem was, however, that some many respondents did not understand the question or gave wrong answers (that contradicted other answers in the survey). If Romanian or Bulgarian respondents said they had a work permit, we accepted their answer as it was. If they gave any other answer, we examined their work situation more closely. Respondents employed by a foreign company ( posted workers ) and self-employed Romanian or Bulgarian respondents, who were formaly registered as such, were categorized as not applicable. A l other working Romanian or Bulgarian respondents were categorized as without a work permit. In practice, this meant that respondents who said they were self-employed but were not registered as such (for instance some street musicians or domestic cleaners) were categorized as without a work permit. Non-working respondents of al three nationalities (N= 39) were excluded from the analyses because it is logicaly impossible to determine whether or not they need a work permit. As long as they do not work, Romanian and Bulgarian migrants do not need a work permit. Also females from CEE-countries with a Dutch life partner were excluded from the analysis because they are exempted from the TWV-requirement. Occupational status The occupational status of respondents is based on the Erikson Goldthorpe and Portocarero (EGP) class scheme (Erikson, Goldthorpe and Portocarero 1979; 1983). Respondents were asked after their occupation in an open question. In the analyses, we divided the occupations in three categories: (0) higher, (1) middle and (2) lower occupations. Higher occupations refer to professionals, higher and lower managers, and routine clerical and sales personnel (the first three categories of the EGP class scheme). Middle occupations are smal employers, independent workers, supervisors and foremen, but also skiled manual work such as construction workers (categories IV, V and VI of the EG P class schem e). Lower occupations refer to unskiled and sem i- skiled manual work and agricultural work (also mostly unskiled work) (category VIIa and VIIb of the EGP class schem e). Finaly, the last category also contains what we cal survival strategies : inform al incom e- generating activities such as being a street musician or seling newspapers on the street (probably homeless newspapers). The official EPG scheme does not take these informal work strategies into account. Earned hourly wages In the survey, respondents were asked about their net earnings per hour, per week or per month. Using this information, we reconstructed their net earned hourly wages. If respondents reported their wage per week, we divided this amount by the number of hours they worked per week. If they reported their wage per month, we

9 divided this amount by the number of weekly working hours times four. In this way we could construct a continuous variable which represents the earned hourly wages of respondents. The minimum wage per hour is 0 euro, the maximum euro. The average earnings per hour are 8.77 euro. Housing position In the construction of the variable housing position we took into account whether or not a respondent shares his/her accommodation or bedroom with others than his/her partner. This variable contains four categories (0) living in a tent/caravan, (1) living in a shared room (shared bedroom), (2) living in a shared accommodation (own bedroom, shared bathroom/living room/kitchen) (3) living in an independent accommodation. Duration of stay The duration of stay is measured by the total time elapsed since respondents first visit to the Netherlands for work (including possible period the respondents was in the home country or elsewhere outside the Netherlands). The unit of measure is in months, because this gives a more exact reproduction of how long people are actualy working in the Netherlands. Moreover, some of the respondents are shorter than one year in the Netherlands, so it would be difficult to express their duration of stay in years. When it was not clear in which month respondents came to the Netherlands, they received the number of years since they have been in the Netherlands, expressed in months. The average time respondents stay in the Netherlands is 27.5 months (amply two years). Background characteristics In the analyses the background characteristics gender, age, educational level and origin are taken into account. Males are coded with (0), females with (1). Educational level is divided in five categories: (0) lower education, (1) secondary education (up to about 15 years), (2) higher education ( high school ) (up to about 18 years), (3) university education and (4) other. The last category contains people who are stil studying and people who reported another educational level than above mentioned. The origin of respondents is based on the country of origin. 4. Research findings 4.1 Romanian and Bulgarian labour migrants with and without a work permit In total, we interviewed 654 CEE-labour migrants: more than half (58%) came from Poland, 17% came from Romania, and 25% from Bulgaria. Table 1 gives an overview of who our respondents are. Somewhat more than half of them were male, somewhat less than half were female. The majority of or respondents (52.8%) was rather young (until 30 years old), although Bulgarian respondents on average were slightly older than their Polish and Romanian counterparts. As table 1 shows, CEE-labour migrants in the Netherlands are generaly rather wel

10 educated. Only a smal minority of them (13.4%) had finished only primary or secondary education (lower general or vocational education up to the age of about 15). However, the share of lower educated respondents is higher among Bulgarians (29.9%). Nevertheless, the large majority of the respondents (more than 80%) had at least finished higher education (general or vocational education up to the age of about 18; high school ). Quite a few of them even had an academic level (or had finished higher vocational education): this goes for 13.5% of the Polish respondents, not less than 35,7% of the Romanians, and 19.5% of the Bulgarians. Table 1 also gives a first idea of the employment situation of CEE-labour migrants in the Netherlands. The majority of them (but even more than three quarter of al Polish respondents) works on a temporary contract, for an employment agency or some other intermediary. Only a smal minority of our respondents works on a permanent contract. Note that Romanians more often have a permanent contact; this is a first indication that relatively more Romanian respondents have better qualified employment in the Netherlands. The most remarkable figure, however, is that the majority of the Bulgarian respondents (53.7%) are active in informal work arrangements. In the interview, we asked the respondents whether they have only a verbal employment contract ; we took this as an indicator for informal work. Smal minorities of each migrant category do not have a job; most of them are looking for work (partly recently arrived migrants who have not found employment yet); others (mainly women) say, they are only active in their own household; and some others, again, are active in other income-generating activities such as street musician or iron-colector ( informal survival strategies ). ***insert Table 1 (overview)*** Finaly, table 1 shows the level of compliance of Romanian and Bulgarian labour migrants (from here on our analyses only refer to working labour migrants). Somewhat more than one third of both the Romanian and the Bulgarian working respondents appear to have a formal work permit; they work according the formal regulations ( compliance ). Also about one third of the Romanian labour migrants plus few Bulgarians are exempted from the work permit obligation: they were either self-employed and formaly registered as such, in either the Netherlands or in Romania or Bulgaria (10 persons) or they were posted workers employed by a Romania- based company (26 persons). In the survey, far more Romanian and Bulgarian respondents said they were self-employed but since they also said they were not registered as such we considered them to be informal workers (like quite a few Bulgarian domestic workers). As a result, somewhat less than one third of al working Romanian respondents and not less than 61% of al working Bulgarian respondents did not have the required work permit. They worked in breach of some of the employment restrictions attached to his/her immigration status ( semi-compliance ). 4.2 Occupational status of labour migrants with and without a work permit We wil measure the achieved social position of labour migrants with and without a work permit in three different ways. The first measure is the occupational status of respondents. A crucial general outcome of our study is that CEE-labour migrants in the Netherlands are a more differentiated social category than we expected.

11 Although many CEE-labour migrants work in low-skiled and often temporary industrial, agricultural and domestic service work we also interviewed highly educated migrants active in professional employment (cf. Engbersen et al. 2011). We divided the occupations of respondents in three categories: higher, middle and lower occupations. Higher occupations range from professionals and managers to routine clerical and sales personnel. Middle occupations include smal employers, supervisors, but also skiled manual workers such as construction workers. Lower occupations refer to unskiled and semi-skiled manual or agricultural work, but also to informal survival strategies. Figure 3 gives a first overview of what kind of occupations the various migrant categories have. Figure 3: Occupational status of various categories of labour migrants 100% 80% 60% 40% 20% 0% With TWV Without TWV Not applicable Polish Low Medium high Already the bivariate comparisons give remarkable outcomes. Given the theory of civic stratification we expected that migrants with the weakest legal position (Romanians and Bulgarians working in the Netherlands without the required work permit) would have the lowest occupational level. However, this seems not to be the case. The majority of both Romanian and Bulgarian respondents without a TWV, but rather unexpected also Polish respondents (who were supposed to have the strongest legal position) work in lower occupations (respectively 69% and 75%) There is, however, a difference in the kind of occupations both categories have. Polish labour migrants for the greater part work in the Dutch horticultural industries. Romanians and Bulgarians without a work permit, almost by definition, work in the informal economy (as domestic workers, in unskiled industrial or service jobs or employing the survival strategies we already mentioned). Respondents who do not need a TWV are in majority (70%) working in middle occupations. For instance, the 26 posted workers in this category are al working as skiled construction workers in the Rotterdam shipyard industries. Finaly, figure 3 shows that almost half (46%) of the Romanian and Bulgarian respondents with TWV are working in higher jobs (but note that this category not only includes professionals, but also routine clerical and sales workers). This is more than in any of the other three categories *** insert table 2 (occupational status) *** Table 2 gives the outcomes of lineair regression model that estimates the chances that respondents belonging to the various migrant categories have a higher occupation. The first model only includes level of compliance into the analysis. In essence, it shows the same outcomes as we saw in figure 3. As expected, the occupational level

12 of Romanian and Bulgarian respondents with TWV is significantly higher than that of Romanians and Bulgarians without TWV. There are no significant differences in the occupational status of the remaining three categories (Romanians and Bulgarians without TWV, Romanians and Bulgarians who do not need a TWV, and Polish labour migrants). The second model includes several background characteristics of our respondents into the analysis. It shows that the difference in occupational status between Romanian and Bulgarian respondents with and without TWV can partly be explained by these background characteristics, in particular differences in educational level and duration of stay. However, also after including these variables into the analysis there is stil a significant difference in occupational status between Romanians and Bulgarians with and without a work permit. Including the background characteristics into the analysis does not change the outcomes of Romanian and Bulgarians who do not need a TWV and those of Polish labour migrants. They stil have a comparable occupational level as Romanians and Bulgarians without a work permit. 4.3 Earned hourly wages of labour migrants with and without a work permit The second measure of the achieved social position of labour migrants is the earned wages per hour. The bivariate comparisons (figure 4) give similar outcomes as we saw before. Again, and against our expectation, the average hourly wages of Polish labour migrants hardly differ (and seem to be even lower) than the wages of Romanians and Bulgarians without TWV or Romanians and Bulgarians who do not need a TWV. Only Romanians and Bulgarians with TWV on average have considerably higher wages per hour. Figure 4: Average earned wages per hour (in euro) of various categories of labour migrants 13 12, ,5 8,7 9 8, With TWV Without TWV Not applicable Polish In table 3 (a linear regression analysis with earned wages per hour as dependent variable) we examine whether these wage differentials result from differences in personal characteristics of respondents. *** insert table 3 (wage per hour)*** Table 3 shows the outcomes of and three different models. Model 1 includes only the central independent variable (level of compliance). It again shows, there are no significant wage differences between three of the four migrant categories (Romanians and Bulgarians without TWV, Romanians and Bulgarians who do not need a TWV, and Polish labour migrants). Only Romanians and Bulgarians with TWV on average have significantly higher wages per hour than the other migrant categories. This pattern changes somewhat after including the relevant background characteristics of the respondents into the analysis (model 2). Particularly gender and the duration of stay in the Netherlands appear to have effect. Male labour migrants and migrants who have worked

13 in the Netherlands for longer periods of time have higher earnings than females and than migrants who have just arrived. Including the background characteristics into the analysis reduces the initialy observed differences in wage levels between migrants with and without TWV with about one fourth. However, also after correcting for the relevant background characteristics in model 2 there is stil a significant difference in the earned hourly wages of migrants with and without TWV. Model 3 includes the occupational status of respondents into the analysis. This, again, reduces the difference in wage levels between migrants with and without a TWV with about one third. When comparing model 1 and 3, we can say that about half of the initialy observed wage differences between Romanians and Bulgarians with and without result from differences in personal characteristics and situations. Romanians and Bulgarians with TWV have higher wages per hour, partly, because they have worked in the Netherlands for longer periods of time and they more often work in higher or intermediate occupations than Romanians and Bulgarians without TWV. On the other hand, also after including al these variables into the analysis, Romanians and Bulgarians with TWV stil have significantly higher wages per hour than Romanians and Bulgarians without TWV (and the other two migrant categories). *** insert table 4 (Average wages for six occupations)*** To disentangle the relationships between the legal status of migrants, their occupational status and the earned wages per hour we also compared the wages of migrants from the four categories within the same occupation. Earlier we saw that the wage differences between Romanians and Bulgarians with and without TWV can partly be explained by the fact that the first category more often works in higher and intermediate occupations. But are there also wages differences between the four migrant categories in typical low-skiled jobs? Table 4 shows the average hourly wages for cleaners, agricultural workers, construction workers, gardeners, plant workers and bartenders are showed. Because the minimum wage differ by age for youngsters under the age of 23, only respondents of 23 years or older are taken into account. In most occupations, Bulgarians and Romanians with TWV indeed have somewhat higher hourly wages compared to their felow countrymen without TWV and also to Polish labour migrants. However, these differences are smal (and do not reach significance, partly due to the low frequencies in each occupation). 4.4 Housing position of labour migrants with and without a work permit A final measure of the achieved social position of the various migrant categories relates to their housing position. Again, we expect the most marginalized category among the labour migrants Romanians and Bulgarians without TWV also has the most inferior housing position. And again, things appear to be somewhat different. Labour migrants in general do not have luxurious housing conditions. Less than one third (30%) of al respondents has an independent accommodation. A l others somehow have to share their accommodation, or even their bedroom, with other migrants (their eventual life partners not included) or live in a tent or caravan. Of the four migrant categories, only Romanians and Bulgarians with TWV in majority have an

14 independent accommodation at their disposal. Only a minority of the respondents of the other three categories live independently. Many CEE-labour migrants live in shared bedrooms with other partners. This is, for instance, the case in the so-caled Polenhotels in the rural city of Westland: in these hotels, two or three migrants share a bedroom. Even worse was the situation in a complex of vacation bungalows in Zundert, an agricultural vilage in the south of the Netherlands. Here, up to eight labour migrants share rather smal bungalows with up to three workers per bedroom. These situations may explain why particularly Polish migrants and Romanians and Bulgarians who do not need a TWV (mostly posted workers ) seem to have relatively unfavourable housing conditions (like shared bedrooms) (figure 5). Figure 5: Housing position of various categories of labour migrants 100% 80% 60% 40% 20% 0% With TWV Without TWV Not applicable Polish Tent/Caravan Shared room Shared accomodation Own accomodation Table 5 presents the outcomes of a multinomial logistic regression analysis that estimates the odds that respondents belonging to the various migrant categories live in a tent or caravan, have to share their bedroom (other than with their own partner) or their accommodation, in comparison to the odds that they have their own accommodation (reference category). We start by comparing the outcomes for respondents who live in tents or caravans with those living in their own accommodation. As not many respondents live in tents or caravans (N= 51), these outcomes may not be very reliable. The table, nevertheless, shows that for Polish respondents the odds of living in a tent or caravan compared to the odds living in their own accommodation is higher than for Romanians and Bulgarians without a work permit (reference category). When the relevant background characteristics are taken into account in model 2 the odds for Polish respondents to live in a tent or caravan become even higher compared to odds for Romanians and Bulgarians without a work permit. Thus, contrary to our expectations, Polish respondents have the highest chance to live in a tent or caravan. Model 2 also shows that Romanians (compared to Bulgarians) and males have a significant higher chance to live in a caravan or tent. In addition, the longer respondents stay in the Netherlands, the smaler their chance to live in a tent or caravan compared to the chance to live in an independent accommodation. *** insert table 5 (housing position) ***

15 The next part of table 5 compares the outcomes for respondents who share their bedroom with other migrants with the outcomes for respondent with an independent accommodation. The table shows that, even when relevant background characteristics are taken into account in model 2, Romanians and Bulgarians who don t need a work permit ( not applicable ) have the highest chance to share their bedroom, compared to the chance for living in an independent accommodation (odds ratio = 6.6). For Polish respondents, the chance to share their bedroom (compared to the chance to live in an independent accommodation) is somewhat smaler, but stil higher than this chance for Romanian and Bulgarians respondents without a work permit (odds ratio = 3,0). Romanians and Bulgarians with a work permit have the lowest chance to share a bedroom (compared to the chance to live in an independent accommodation) (odds ratio 0,7). We expected this category to have a better housing position than Romanians and Bulgarians without a work permit, but that they also have a better housing position than Polish labour migrants in somewhat unexpected. Model 2 also shows that women and respondents who stayed for a longer period in the Netherlands have a smaler chance to have to share their bedroom compared to males and those respondents who are staying just for a short period in the Netherlands. Finaly, we compare the estimates of respondents who live in a shared accommodation (shared bathroom/living room/kitchen) with the estimates of respondents who have an independent accommodation. Model 1 shows us that the chance to live in a shared accommodation (compared to the chance to have an own accommodation) is much smaler for Polish respondents and Romanians and Bulgarians with a work permit in comparison with Romanians and Bulgarians without a work permit. These outcomes remain the same when relevant background characteristics are included in model 2. This model shows that Romanian labour migrants, females, young respondents and respondents who work in the Netherlands for a long time have a lower chance to live in a shared accommodation (compared to the chance to live in an independent accommodation) than Bulgarian labour migrants, males, old respondents and respondents who are only working in the Netherlands for a short time. 5. Discussion This study examined the achieved social position of Romanian and Bulgarian labour migrants working in the Netherlands without having the required work permit (TWV). As EU-citizens, Romanian and Bulgarian nationals by definition are legal residents in the Netherlands. However, in an attempt to protect the Dutch labour market, and Dutch jobseekers, from the feared influx of Romanian and Bulgarian workers the Dutch government applied the so-caled transitional period when Romania and Bulgaria acceded the EU in January During this transitional period (that wil probably last until January 2014 in the Netherlands), Romanian and Bulgarian nationals have restricted access to the Dutch labour market. Like before, Romanian and Bulgarian nationals officialy stil need a work permit in order to work in the Netherlands. However, as our survey shows, this does not stop potential Romanian and Bulgarian labour migrants to come to the Netherlands, also when they don t have the required work permit. Almost one third (29.8%) of the working Romanian labour migrants we interviewed, and almost two third (61.4%) of the working Bulgarians, did not have the required work permit to work in the Netherlands. As a result, many Romanian and Bulgarian labour migrants have to rely on informal work arrangements. This also became clear in our survey. More than half (53.7%) of al Bulgarian respondents

16 (including those who have not yet found employment), and almost one in five (19.4%) of the Romanian respondents said they work informaly (having a verbal contract ). We caled this ambivalent legal position of many Romanian and Bulgarian labour migrants in the Netherlands semi-compliance : Romanians and Bulgarians are legaly residents in the Netherlands, but without the required work permit they work in breach of the employment restrictions attached to his/her immigration status (Ruhs & Anderson 2010). In this article we examined what this means for their social position in the Netherlands. In order to do so, we compared the social position of Romanian and Bulgarian labour migrants without TWV with that of three other migrant categories: Romanian and Bulgarian labour migrants who work with a TWV, those who formaly do not need a TWV (like self-employed or posted workers ), and Polish labour migrants. When Poland joined the EU in 2004, there was also a transitional period for Polish nationals. The transitional period for the new EU-member states of 2004 was ended in May At the time of our survey ( ), Polish nationals had free access to the Dutch labour market. Folowing theory of civic stratification that postulates that states grant or deny various migrant categories different social and economic right, thereby influencing their opportunities and life chances in the destination country we expected Romanians and Bulgarians without a TWV to have the lowest social position. We expected Polish labour migrants who have free access to both Dutch territory and to the Dutch labour market to have the highest social position. Finaly, we expected Romanians and Bulgarians who work in the Netherlands with the required TWV and those who formaly do not need a TTWV to end up somewhere in between. In our analyses, we used three different indicators to measure the social position of CEE-labour migrants: their occupational status, their earned wages per hour, and their housing situation. With regard to the occupational status of labour migrants, we found that Romanians and Bulgarians with TWV indeed have higher occupational positions than Romanians and Bulgarians without TWV. Rather unexpected was that the occupational level of Polish labour migrants on average is similar to that of Romanians and Bulgarians without TWV. Differences in personal characteristics, in particular in particular differences in educational level ( human capital ) and duration of stay. But also after correcting for these individual differences, Romanians and Bulgarians with TWV have significantly a higher occupational level than their felow countrymen without TWV. The second indicator, earned hourly wages, gives a similar pattern. As expected, there are significant wage differences between Romanians and Bulgarians with and without TWV. The regression models show that about half of this wage differences can be explained by differences in personal and labour market characteristics, in particul;ar duration of stay in the Netherlands and the fact that Romanians and Bulgarians with TWV more often work in higher and intermediate occupations than Romanians and Bulgarians without TWV. Here again, we saw the remarkable outcome that the hourly wages of Polish labour migrants are not higher than those of Romanians and Bulgarians without TWV. Also when we compared the hourly wages of the four migrant categories for typical low-skiled work, we found that Romanians and Bulgarians with TWV earn more for the same work than both Romanians and Bulgarians without TWV and Polish migrants. The last measure of the social position of the four migrant categories is their housing position. Surprisingly, and contrary to our expectations, Polish labour migrants in the Netherlands have the greatest chance to live in a caravan or tent, compared to the chance to live in an independent accommodation. Romanians and Bulgarians