Do refugees integrate into the labour market? The case of Belgium.

Transcription

1 Do refugees integrate into the labour market? The case of Belgium. Dries Lens a *, Ive Marx b and Sunčica Vujić c a Herman Deleeck Centre for Social Policy (CSP), University of Antwerp, Antwerp, Belgium b Herman Deleeck Centre for Social Policy (CSP), University of Antwerp, Antwerp, Belgium c Centre for Development Studies (CDS), University of Bath; Centre for the Analysis of Social Policy (CASP), University of Bath; University of Antwerp, Antwerp, Belgium *Dries Lens (corresponding author). Herman Deleeck Centre for Social Policy, University of Antwerp, Sint-Jacobstraat 2 B-2000 Antwerp, Belgium. Tel.: 0032 (0) dries.lens@uantwerpen.be 1

2 Do refugees integrate into the labour market? The case of Belgium. In this study, we examine the labour market integration of refugees who arrived in Belgium between 1999 and 2009 by looking at their first entry into employment, social assistance and unemployment. Based on a linkage of the Belgian Labour Force Survey with the Data Warehouse Labour Market and Social Protection, we estimate random-effects discrete-time hazard models to compare the labour market transitions of refugees to those of labour and family immigrants of the same arrival cohort. The analysis shows that the labour market integration of refugees leaves much to be desired. Not only do refugees take significantly longer to enter their first employment as compared to other immigrant groups, they also run a greater risk of transitioning into the social assistance scheme (right after arrival) and into the unemployment benefit scheme (after some years of residence). These findings indicate the importance of targeted educational and labour market measures to facilitate the best possible integration of refugees into the Belgian labour market and thereby also into Belgian society. JEL classifications: J61; J64; J68 Keywords: refugees; labour market integration; longitudinal research; discrete event history models Introduction The reception and integration of refugees is one of the most pressing and important issues facing Europe today. The recent influx of asylum seekers also makes it one of the most contentious and hotly debated issues of our time. Yet, despite the availability of a huge body of literature on the labour market integration of immigrants, there is only a rapid body of research looking at how refugees are faring in European labour markets (see Bevelander 2016 for Sweden; Bakker, Dagevos and Engbersen 2016 for the Netherlands; Bratsberg, Raaum and Røed 2017 for Norway; Sarvimäki 2017 for 2

3 Finland; and Schultz-Nielsen 2017 for Denmark). Based on high-quality national register data, these studies provide evidence of a refugee gap, which is the employment and wage gap between refugees and other immigrants, particularly visible shortly after arrival to the host country. This paper adds to the limited literature on refugee integration in Europe by focusing on Belgium, a country that presents itself as a particularly relevant and interesting case. Belgium has had a relatively strong influx of asylum seekers over recent decades, alongside very significant family reunification and less sizeable labour migration inflows. Immigration levels have reached unprecedented levels over the past decade and have accounted for the bulk of population growth in Belgium since the 1990s. Yet, while Belgium is a country of permanent immigration, it has not adapted institutionally to this reality (Van Rie and Marx 2014). This is creating a fundamental tension since nowhere in Europe is the gap in socioeconomic outcomes between immigrants and natives as large as in Belgium (Pina, Corluy and Verbist 2015). The question of how social rights so cherished by Europeans, can be balanced with the goal of integrating refugees into the labour market presents itself with particular saliency in Belgium. In order to diminish this institutional gap and, thus, be able to adjust current integration policies, it is essential to have an outlook on the recent evolution of immigrant experience on the Belgian labour market. Therefore, our main aim is to investigate retrospectively the level of integration of refugees living in Belgium as compared to other types of immigrants, by looking at their first entry into employment, social assistance, and unemployment. We start with a short review of the main theoretical arguments, and the available empirical evidence on refugee labour market integration. We then describe the Belgian setting, by framing asylum migration to Belgium and explaining labour market access 3

4 and social security entitlements of asylum seekers. We continue by describing in more detail our data, methods and variables and then report our main empirical findings. The final section summarizes and discusses the central findings of the study. Existing evidence on refugee labour market integration Upon arrival in a new country, immigrants face certain disadvantages, as they lack country-specific human (work experience and language proficiency and the problematic recognition of their home country credentials) and social (social networks) capital. This tends to result in higher risks of unemployment, and lower occupational status and earnings as compared to the native-born (Borjas 1994; Chiswick 1978; Friedberg 2000; Kalter & Granato 2007). Over time, however, immigrants are expected to catch up with the native-born, as they acquire social networks, language skills and other host country human capital that will help them overcome initial shortfalls (Akresh 2008; Chiswick, Lee, and Miller 2005; Reitz 2007). This process of gradual socioeconomic convergence is not uniform and varies considerably for differently situated immigrants (Kogan and Weißmann 2013). While some immigrants arrive in a new country and quickly make a transition into steady employment, others face stubbornly persistent labour market barriers (Fuller and Martin 2012). While considerable research has focused on immigrant employment and wage gaps with natives, little is known about the so-called refugee gap, which is the employment and wage gap between refugees and other immigrant groups (Connor 2010). An important reason for the limited evidence is data limitations, since detailed statistical information on migration motive is not always easily accessible (Bevelander 2016). A number of studies in traditional settler countries such as Australia, the US and Canada focus specifically on labour market integration of immigrants by migration 4

5 motive. What this research shows is that shortly after arriving in the host country, refugees do not fare well in the labour market. Compared to family and labour migrants, refugees are characterized by lower levels of employment (Aydemir 2011; Cobb-Clark 2001) and earnings (Cortes 2004; DeSilva 1997), and higher levels of social security dependence (Devoretz, Pivnenko and Beiser 2004). This refugee gap has been attributed to different types of factors. First, authors have argued that seeking refuge is often less planned and prepared than other forms of migration and, therefore, refugees arrive in the host country with relatively poorer language skills and weaker attachments or links to the host country (Richmond 1988). Consequently, securing both a higher language proficiency and accreditation of skills are relatively slow processes that can delay labour market access (Rooth 1999). Second, due to the forced nature of their migration and the traumatic experiences frequently associated with it, many refugees suffer from psychological distress. Health conditions like post-migration stress or trauma are shown to have strong links with labour market outcomes of refugees (Hartog and Zorlu 2009). Finally, asylum seeker-specific obstacles are legal restrictions to access the labour market, a long duration of the asylum procedure and a temporary, insecure residence status. These barriers prevent refugees from quickly and fully participating in the labour market (Bakker, Dagevos and Engbersen 2014; Bloch 2008). With increased time in the host country, refugees are expected to catch up and show similar or even higher employment and wage levels compared to family immigrants (Bevelander 2016). The most important reason cited for this recovery process is that the likelihood of returning home is smaller for refugees, as they face a higher risk of harm or persecution and often keep fewer social ties with their country of origin (Phillimore 2011). In view of their prospects for settling permanently, refugees 5

6 thus have a greater incentive to invest in the host country s own human capital (by learning one of the national languages, for instance), which ultimately facilitates their integration (Cortes 2004). As such, higher rates of human capital accumulation can lead to refugees catching up with and perhaps even outperforming other non-economic immigrants over time. However, refugees seldom catch up to the economic integration level of labour immigrants (see Cortes 2004 for an exception). Spurred by the recent interest in refugee integration, a line of research is now emerging on labour market integration of refugees in (mainly Northern) European countries (Bevelander 2016; Bakker, Dagevos and Engbersen 2016; Bratsberg, Raaum and Røed 2017; Schultz-Nielsen 2017). These studies measure the labour market participation of refugees in a longitudinal perspective, contrasting their outcomes with different comparison groups (EU immigrants, labour immigrants, family immigrants and the native-born). Their results seem robust with findings for traditional settler countries (US, Canada and Australia), as they too uncover a significant refugee employment gap right after arrival and, subsequently, a recovery process with refugees catching up to the employment level of family immigrants. However, in addition, they show that, after five to ten years of residence, employment rates of refugees and family immigrants stabilize (or even decrease) while the rates of social insurance dependency increase. Furthermore, they find that this trend applies in particular to refugees. According to the authors, this new evidence suggests that refugees and family immigrants find jobs in the periphery of the labour market, and that the quality of the jobs they get into puts them at high risk of falling victim to cuts and redundancies. 6

7 The Belgian setting Asylum migration to Belgium Because of its size and its less known history of immigration, Belgium is often overlooked as a country of immigration. Yet, over the past five decades Belgium became a permanent country of settlement for many different immigrants groups (Petrovic 2012). In the 1950s and 1960s, Belgium set out to attract primarily unskilled workers from Southern European countries and later from Northern African countries and Turkey to work in mining and heavy industry. Assisted by the lenient rules for family reunification at the time, these guest workers stayed, brought their families and settled permanently into Belgian society. Despite the introduction of a formal cap to limit economic migration in 1974 (reference?), the immigration landscape kept on widening and diversifying as decades passed. Labour migration continued, with work permits being limited to more highly skilled immigrant workers. Family migration increased and took over from labour migration as the most important entry channel to Belgium. It was not until 1990 that asylum seekers became an important phenomenon. Asylum applications peaked particularly in 1993, when 25,000 applications or the equivalent of 42% of all entries into Belgium were registered; in 2000 when 42,000 applications, or 62% of total entries were recorded; and in 2011, when 25,000 asylum seekers, corresponding to 19% of total immigration flows, came to Belgium (Burggraeve and Piton 2016). In 2015, Belgium received 44,800 asylum requests, which makes the inflow much bigger than those seen in 1993 and 2011, but it is similar to that of the year 2000 (see Figure 1). The share of asylum seekers who were granted a residence status is nevertheless higher than in the past. In 2015, the Commissioner General for Refugees and Stateless Persons concluded that the asylum seeker needed protection in 61% of its 7

8 decisions. The 61% positive decision rate in 2015 was up from 47% in 2014 and 29% in 2013 (CGRS, 2016). [FIGURE 1 HERE] Applications for asylum requested in Belgium accounted for 3.4% of the total number of asylum requests registered in the EU in 2015 (reference?). With more than 4 applicants per 1,000 inhabitants, Belgium is eighth on the list of host countries taking in the most asylum seekers (Burggraeve and Piton 2016). In absolute figures, the leading hosts are Germany and Hungary (respectively 36 and 13% of all applicants), while Hungary, Sweden and Austria lead the pack in terms of applicants per capita (respectively 18, 17 and 10 applicants per 1,000 inhabitants). 1 Even though asylum seekers and refugees are still a relatively small immigrant group in Belgium, their number is growing (Burggraeve and Pitton 2016). Their migration motive and the context of reception in the host country distinguishes refugees from other immigrants, thus increasing the need for more knowledge on the integration of this specific immigrant group. Labour market access and social security entitlements of asylum seekers For Belgium, the framework legislation on employment conditions falls under the competency of the federal government. The implementation of this law is to a large extent part of the competency of the regional authorities, which includes among others 1 It should nevertheless be noted that Hungary is regarded as a transit country, unlike the other three countries, which constitute the final destination for potential refugees. 8

9 the granting of work permits to third-country nationals. As a general rule, EU/EEA and Swiss citizens and their family members do not require a work permit in order to be employed in Belgium. However, third-country nationals (non-eu/eea/swiss citizens) and their family members who still do not have a permanent right of residence typically require a Belgian work permit (reference?). Regarding the more specific issue of asylum seekers, the process of acquiring a work permit has undergone numerous changes since the 1970s, at times involving long waiting periods and preventions from working (Rea et al. 2014). Between 2007 and 2010 many asylum seekers did not have the right to work as a result of the 2007 legislative changes, which removed the admissibility phase in the asylum procedure. Therefore, asylum seekers were no longer eligible for a work permit. Since 2010, asylum seekers who fulfil certain criteria are allowed to work with a work permit card (type C). It concerns asylum seekers who have not yet received a first instance decision on their asylum case within six months following the registration of their asylum application. Since September 2015, asylum seekers have been able to get into the labour market four months after they have registered with the Belgian Immigration Office (reference). Following this reform, Belgium is now among the European countries with the shortest period for obtaining a work permit. Only Greece and Sweden have shorter waiting periods, as they allow immediate entry, as well as Austria, Romania and Germany, where workers have to wait three months (Fóti and Fromm 2016). In Belgium, work permits for asylum seekers are not conditional on a test to make sure that no national or European resident is interested in the vacancy (unlike in Austria and Greece). There are also no limits on the sectors of activity where asylum seekers are allowed to work (unlike in Austria, the UK and Sweden), and asylum 9

10 seekers are eligible for self-employment (unlike in Germany and the UK), under the condition that they apply for a professional card. 2 Asylum seekers are entitled to unemployment benefits under the same terms as Belgian nationals (reference?). In practice, however, asylum seekers are unable to avail themselves of these benefits since a work history (or at least a minimum qualifying period) is required, which, of course, they cannot meet (Fóti and Fromm 2016). Job mediation and job placement, as well as vocational training and job-related training are available for both refugees and asylum seekers (after four months), although these require a minimum level of language proficiency in Flanders, this is the level of Dutch of 2.1 or 2.2 (depending on the course). In contrast to unemployment benefits, social assistance non-contributory guaranteed minimum resources does not have any requirements such as waiting periods and compulsory payment of contributions. To be eligible for social assistance, a person must be at least 18 years of age, have exhausted all other rights in the social insurance programs, have legal residence in Belgium and be available for work 3 (Carpentier, 2016). Furthermore, eligibility is regulated by a meanstest for the household, according to the de facto living arrangement. Not all immigrants are eligible for all sub-programs. A lot depends on the immigrants length of stay (Carpentier 2016). 4 Nonetheless, every legally residing individual is entitled to 2 Given the uncertain residence for the asylum applicant, it is not allowed that the self-employed activity requires large investments. 3 Except when exempted for health or other reasons considered valid. 4 Persons with a Belgian or another European nationality and their family members (except during the three first months of residence), as well as non-eu immigrants after five years of residence, recognized asylum seekers and stateless persons are eligible for all subprograms. Other legally residing foreigners, including asylum seekers during the investigation of the motivation for asylum request, rejected asylum seekers with a pending 10

11 assistance if the household satisfies the eligibility criteria. In practice however, non-take up rates are substantial (Bouckaert and Schokkaert 2011). Within the context of entitlements, it is worth mentioning that, in Belgium, if asylum seekers work, they have to contribute to the costs of their accommodation and other material support. This applies to those asylum seekers who reside in a reception centre while working; they continue to be eligible both for material support and housing, but will be obliged to contribute (Fóti and Fromm 2016). Belgium offers a wide range of services and measures to asylum seekers in preparation of labour market access. However, the rules for some of the programmes differ between Flanders and Wallonia: in Flanders, asylum seekers are obliged to attend language and orientation courses as well as civic education, but not as yet in Wallonia, although these schemes are available. In both regions, there are plans to extend the mandatory programmes (Fóti and Fromm 2016). Validation of foreign degrees is free in Belgium for asylum seekers. However, despite some improvement on delays in degree recognition, the process remains burdensome, which discourages many immigrants from even attempting it (De Keyser et al. 2012). By way of example, NARIC 5 received 482 applications in 2015 from asylum seekers, recognized refugees or subsidiary protected, 59% concerning higher degree recognition applications (NARIC-Flanders, annual report 2015). Clearly, this is a very small share of the potential number of applications. And in a small-scale study, Caritas International (2014) questioned 54 refugees and found out that, while 37 of them held a secondary or higher education appeal who are not allowed to work and other foreign-born persons residing in Belgium less than five years, are not. 5 The National Academic (and professional) Recognition and Information Centre is responsible for recognising the equivalence of foreign study certificates in Flanders. 11

12 diploma, only nine of them had applied for equivalence. The most important barriers cited were the cost of the application, the long waiting period before receiving an answer, and not having the original diploma and the inability to request a copy in the country of origin owing to the geographical instability. We conclude that, compared to other European countries, Belgium does show higher nominal flexibility when it comes to labour market access for asylum seekers. However, accessing the labour market (having a work permit) is still quite far from actually getting a job. Moreover, while Belgian social security is in principle an inclusive one, the fact that it is a strongly work-based system means that a steady link to the labour market is required in order to build up social security entitlements. In the coming sections, we will examine whether it is the case that newcomers weak ties to the labour market make them vulnerable to recourse to social assistance. Data and Methods Data In order to understand how refugees progress (or regress) in the host society, this paper builds on a longitudinal integrated data set linking the Belgian Labour Force Survey (LFS) with the Data Warehouse Labour Market and Social Protection. 6 For Belgium, the LFS is a representative sample from the National Register and provides, in addition to demographic characteristics, both general and more detailed data on the employment 6 In order to have a large sample size, we pooled together three annual LFS samples (2010, 2011 and 2012) with added retrospective and prospective quarterly information on the socioeconomic position ( ) from the Data Warehouse. The data matching between the LFS and the Data Warehouse was an exact one, in the sense that the national register numbers have been used to link the individuals information in both datasets. 12

13 situation, such as the quality of employment and characteristics of the workplace. An added value of this sample versus administrative data, and absolutely indispensable for sociologic (and economic) labour market research, is that the highest educational level is available. The sample is also large enough to distinguish between broad categories of immigrants (Corluy and Verbist 2014). On order to have longitudinal panel data, we linked the Labour Force Survey with administrative data from the Data Warehouse Labour Market and Social Protection. The Data Warehouse links data from various administrative databases and the national register for scientific use. In terms of longitudinal analysis, the Data contains the quarterly socioeconomic nomenclature, through which the main source of income can be verified. On the basis of the information from the participating social security institutions, a detailed distribution of the population in the Data Warehouse can be made according to socioeconomic positions. Depending on a position in or out of the labour market, the population is divided into employed, unemployed, in social assistance, professionally inactive or other. 7 The situation taken into account is invariably the situation on the last day of the quarter. Important to note is that the socioeconomic nomenclature is assigned hierarchically, giving priority to work over benefits. Consequently, a person in employment could still have a social assistance benefit top-up payment. Another important remark is that our data do not pick up employment in the informal market, which is known as an important source of employment for immigrants in Belgium (Geets et al. 2007; Rezaei et al. 2013). 8 7 A detailed description of the different socioeconomic statuses can be found in the Appendix. 8 Even if immigrants hold a permanent residence status, this does not necessarily imply access to formal employment, since such access depends on the availability of employment, the networks used to find formal employment, the qualifications and skills that one brings to 13

14 A second upshot of using our linked database is that the additional information of the Data Warehouse allows identifying immigrants by first issued residence permit. Consequently, in this paper we are able to categorize immigrants by their legal means of access for stay or residence in Belgium. We recognize the limitation that the first issued residence permit might not completely reflect the true motivations and aspirations of the individuals when they first migrate, as sometimes those motives are hard to fit into rigid administrative categories. Additionally, an immigrant s motive for migration might change during the stay, while no formal amendment is made in the administrative registers. Nonetheless, we believe that the first issued residence permit still provides an relevant aspect when looking at their outcomes later on. We do not possess information on when immigrants receive their residence permit or foreigners card (or in the case of asylum seekers, got recognized as a refugee). Since we start tracking immigrants socio-economic positions based on their selfreported years of residence in Belgium (available in LFS), it is likely that at the beginning of our observation window, immigrants are still in the procedure of applying for residence rights. As a result, our analysed sample includes both immigrants who have not yet completed the recognition process and immigrants who obtained their residence permit, and we cannot observe when the transition between the two states takes place. However, since the LFS sample is drawn from the National Register, 9 we the labour market, and the hiring practices of employers. As a consequence, legally residing immigrants may be forced to find underground employment. Unfortunately, this channel is not picked up by the formal statistics used in this paper. 9 The National Register includes individuals from the population register (Belgians living in a Belgian municipality and foreigners with a permanent residence permit), from the foreigners register (foreigners with a temporary residence permit, recognized refugees and regularized asylum seekers), and from the register of officials of the European Union. 14

15 are confident that at the time of the survey, immigrants have obtained either a temporary or permanent residence permit. As a result, our findings extrapolate only to the immigrants who eventually do get a temporary or permanent residence permit (or in the case of asylum seekers, got recognized as a refugee), and not to the ones who do not. We selected refugees (N=287), family immigrants (N=3,165) and labour immigrants (N=2,429) who arrived in Belgium between 1999 and 2009 and were between the ages of 18 and 55 upon arrival. 10 This is done in order to diminish the potentially negative bias on employment entry due to mobility in school participation and retirement (Kogan and Weiβmann 2013). Note that for labour immigrants, we had to take into account that a sizeable share of them is unknown to the Belgian social security institutions. The reason behind this is that an important share of labour immigrants is working for European or international organisation that do not fall under Belgian social security rules (e.g. diplomats, European civil servants etc.). Consequently, they get assigned the status other in the quarterly socioeconomic nomenclature while they are registered as employed in the LFS. We chose to delete this group from the analyses, as they biased results quite substantially. 11 All in all, the used dataset will enable us to get a first-of-its-kind view on the labour market trajectories of refugees in the Belgian labour market, with different comparison groups: labour immigrants and family immigrants. Belgians abroad, asylum seekers (waiting register) and persons without legal residency are thus not included in the sample (FPS Economy, Directorate General Statistics and Economic Information). 10 An overview of the sample size per migration motive and entry cohort is provided in the Appendix. 11 More information on the identification and magnitude of this group can be found in the Appendix. 15

16 Methods In our core analysis, we estimate random-effects discrete-time hazard models from a person-quarter file using a complementary log-log link function to study (1) the first transition into employment, (2) the first transition into social assistance, and (3) the first transition into unemployment, after immigration to Belgium. We start with the single transition into employment. The starting time of the episode (t = 0) is defined as the quarter and year of migration to the host country. 12 An event occurs when an individual takes up employment. In this case, the point in time at which employment began defines the ending time of an episode. Hence, duration measured in quarters, equals the period between arriving to Belgium and entering employment. Those immigrants who did not enter employment at the end of the observation period are treated as right censored. In such cases, duration is defined as the period of time between the time of arrival and the end of our observation period (the fourth quarter of 2012). Next, we estimate two random-effect hazard models to study the first entry into social assistance and, subsequently, the first entry into unemployment. Both models correspond to the model for employment entry described above. Cases in which respondents were either employed, in social assistance or unemployed in the years before their declared year of arrival are dropped from the analysis, since they fall outside the scope of the present study (N = 194). To approximate the shape of the hazard functions and to estimate the impact of independent variables, we included dummy variables dividing the time axis into yearly 12 The year of arrival is based on the self-reported years since migration in the LFS data. We randomly assigned a quarter of migration to every individual in the research sample. 16

17 intervals. 13 Hence, our models assume that the hazard is constant not over the whole range of time, but within certain specified intervals of time. Furthermore, we include observed socio-demographic characteristics (see infra). We control for unobserved time-constant characteristics affecting the hazard of the single transitions by nesting person-quarters in individuals and including a random disturbance at the individual level. 14 Not controlling for unobserved heterogeneity (frailty) is likely to overstate negative duration dependence (Lancaster, 1990; Wooldridge, 2002). Our models for the quarterly hazard of transitioning into employment, social assistance and unemployment can be written as: ln[ ln(1 Pr(t T < t T t))] = β 0 + k=1 d ik + γx it + u i, (1) 10 β k Where u i N(0, σ 2 u ) is a normally distributed random disturbance at the individual level. The expression Pr(t T < t T t) represents the discrete-time hazard of transitioning in interval t, i.e. the conditional probability (Pr) that the event occurs (at time T) between time points t and t, given that it has not occurred until t. The d ik represents time intervals k for individual i, X it the observed explanatory variables, and β 0, β k and γ are vectors of coefficients to be estimated. Although all three transitions are continuous processes, we estimate discrete-time models because analysis time is measured in quarters and collapsed into larger intervals for the analysis (Jenkins, 13 The first interval covers the first 4 quarters after arrival to Belgium (year 1), the second the following 4 quarters (year 2), and so on. Due to the relatively small number of cases not covered within the first nine intervals, the tenth covers all periods after the ninth year. 14 We assumed that the random effect is normally distributed with zero mean with a fixed variance estimated from the data and is uncorrelated with the observed regressors. 17

18 2007). Given the complementary log-log function, all models are proportional hazard models and exponentiated parameter estimates, e(b), can be interpreted in terms of hazard ratios (Singer and Willett, 2003). Variables In the models presented here, refugees, i.e. immigrants with a first residence permit issued for asylum or international protection, are compared to labour and family immigrants. Also included in our multivariate hazard models are a set of control variables. Gender is one obvious candidate, as in Belgium immigrant women are less likely to participate in the labour market than immigrant men, especially among immigrants originating from countries outside the EU (Corluy and Verbist 2014). Additionally, we expect that (lack of) human capital is a very important determinant of labour market integration. Hence, we consider age at migration as a linear term, and we also include a quadratic term to account for possible nonlinearities in the effect. Additionally, we include the level of education at the time of the survey as a timeconstant variable. 15 Level of education is divided here into three categories: low-skilled (ISCED 0 through 2), medium-skilled (ISCED 3 and 4) and high-skilled (ISCED 5 and 6). We also include a time-variant host country human capital variable that reflects 15 Unfortunately, the design of the LFS questionnaire does not allow a complete reconstruction of education histories. Combining the questions What is the highest level of education you have successfully completed? and In what year did you reach that level? does not provide sufficient information to evaluate whether an individual was continuously enrolled in education up to the date of the highest educational attainment. We acknowledge that the use of level of education at the time of the survey impedes causal inferences to be made due to reversed causality. 18

19 whether or not one obtained the highest educational degree in Belgium for each quarter of observation. Since the Belgian regions differ considerably in terms of economic situation and thus in employment prospects for individuals, geographical spread of immigrants may provide an additional explanation for differences in labour market participation. Region of residence (Flanders, Brussels, Wallonia) too is included as a time-constant variable measured at the time of the survey. 16 In order to assess the influence of region of origin on the probability of employment, immigrants are sorted into eight regional categories based on country of birth: old EU/EEA Member States, new EU/EEA Member States, Central and Eastern Europe, Turkey, North-Africa, South and Central Africa, America and Asia. 17 Lastly, we control for year of migration (ten dummies) and year of the LFS survey (two dummies). [TABLE 1 HERE] Descriptive statistics Descriptive statistics, reported in Table 1, indicate that refugees (i.e. with asylum 16 We acknowledge that the use of region of residence at the time of the survey impedes causal inferences to be made due to reverse causality. What do you mean by this? 17 Old Member States: Germany, France, Luxembourg, the Netherlands, Denmark, Finland, Sweden, the UK, Ireland, Austria, Iceland, Principality of Liechtenstein, Norway, Spain, Greece, Portugal and Italy. Switzerland is also counted among the Old Member States. New Member States: Hungary, Poland, Latvia, Estonia, Lithuania, Czech Republic, Slovakia, Slovenia, Bulgaria, Romania, Cyprus and Malta. Central and Eastern Europe: Albania, Armenia, Azerbaijan, Bosnia and Herzegovina, Croatia, former USSR, Moldavia, Russia, Federal Republic of Yugoslavia (does not exist anymore), former Yugoslavia (what is this?), Macedonia, Belarus, Georgia. Maghreb: Algeria, Egypt, Libya, Morocco and Tunisia. 19

20 motive) and labour immigrants are predominantly male, while family immigrants are predominantly female. Of the three immigrant groups, refugees arrive at the youngest age with a mean age at migration of 30. For family immigrants this is age 31 and for labour immigrants age 32. There exists considerable variation in terms of education: nearly 56% of refugees is low educated compared to 44% of family immigrants, and 28% of labour immigrants. Refugees are also less frequently higher educated (only 17% is high educated). With regard to geographical spread, we observe that, in comparison to other immigrant groups, refugees more often reside in Flanders and less often reside in Brussels. There are also clear differences in terms of the dominant region of origin. Every refugee in our sample originates from a non-eu country. The most common regions of origin are South and Central Africa (36%), Asia (30%) and Central and Eastern Europe (25%). In contrast, the large majority of labour immigrants come from old Member States (45%) or new Member states (32%). An important share of family immigrants come from North-Africa (30%). The other dominant regions of origin are old Member States (14%), new Member states (13%), Asia (12%) and South and Central Africa (11%). With regard to the year of migration, all three immigrant groups are underrepresented in the first arrival cohort (99-02). This is probably due to the large proportion of immigrants with missing information on the migration motive in the calendar years Lastly, labour and family immigrants are equally represented in the different survey years, while refugees are more likely to be captured by the 2012 survey and less likely to be in the 2010 survey. Refugee labour market integration in a longitudinal perspective Before we turn to our core discrete-time event history analysis, we take a simpler 20

21 perspective to get a first impression of the labour market trajectories of refugees. This can be gained by showing state distribution plots, 18 where the proportion in certain labour market statuses is displayed by duration of stay in Belgium. 19 Figure 3 shows the quarterly state distribution plots for refugees. Refugees tend to follow a particular trajectory, which is roughly as follows. In the first years of residence the large majority of refugees has no connection with Belgian social security whatsoever. After entry, and hardly surprisingly, very few refugees are to be found in the formal labour market. Similarly, very few are dependent on social transfers. But after this initial phase follows a very quick and strong rise in dependency on social assistance, the only social protection programme to which refugees can have relatively rapid access in the largely contributions-based Belgian social protection system. In the third year of residence, the share of social assistance beneficiaries reaches a peak, just below 50%. Entry into employment remains very low during the first years after entry. After what we could label a transition phase, which takes about 4 years, dependency on social assistance starts to drop and formal employment participation surges. But sometime into this second phase we also see an important fraction (20%) transitioning into the unemployment benefit scheme. Yet at the same time, an important fraction remains dependent on social assistance (20%) or stays unknown to Belgian social security institutions (16%). 18 For the definition of the different labour market statuses, we refer to the Appendix. 19 We choose to display the state distributions by duration of stay and not by calendar year because we have different arrival cohorts; in the year 2005, the labour market position of those who arrived in 1999 and had therefore been in Belgium for six years does not mean the same as for those who arrived in 2004 and had therefore only been in Belgium for one year. Showing the results by duration of residence, therefore, is a more accurate way of presenting the results since it is independent of arrival cohort. 21

22 [FIGURE 3 HERE] Like refugees, very few family immigrants are employed in the formal labour market after entry (Figure 4). However, family immigrants do seem to be less dependent on social assistance transfers compared to their refugee counterparts. In the former group, the share of social assistance beneficiaries gradually increases to around 7% in the second year of residence, and remains stable thereafter. Employment levels also seem to increase more rapidly, converging towards a level of 50% after ten years of residence. Like refugees, family immigrants have very low unemployment levels in the first years of residence, which is what we would expect considering that they both lack the stable job history that is required to be entitled to unemployment benefits. However, after some years of residence the unemployment rate starts to increase quickly, peaking at a level of 10% after ten years of residence. However, what clearly sets family immigrants apart is their large share of individuals in the status other, representing persons who are unknown to Belgian social security. One out of three family immigrants remains in this other category after ten years of residence. 20 [FIGURE 4 HERE] Labour immigrants clearly have the fastest labour market attachment, with an employment rate of 47% at the end of the first year, 67% at the end of the second year 20 More information on the population in this heterogeneous other category can be found in the Appendix. 22

23 and 75% from the fifth year onwards (Figure 5). During the first years of residence, labour immigrants also have the highest unemployment attachment of the three immigrant groups, mirroring their faster pace of employment entry. This picture changes somewhat thereafter, as their unemployment rate of 8% stands slightly below the unemployment rate of family immigrants (10%) and way below the unemployment level of refugees (20%) after ten years of residence. Lastly, social assistance benefit dependency is only marginal among labour immigrants (below 3%), compared to the dependency levels of family immigrants and, in particular, refugees. [FIGURE 5 HERE] The results show that, compared to other immigrant groups, it takes refugees significantly longer to get a foothold in the Belgian labour market. Because of their weak ties to the labour market in the first years of residence, refugees are particularly vulnerable to recourse to social assistance. Once they get onto the labour market and build up social security entitlements, they run a greater risk of transitioning into the unemployment benefit scheme. The questions following up on this are whether these differences persist and whether they are statistically significant upon controlling for important covariates. To answer them, in the next subsections, we use discrete-event history modelling to study the impact of migration motive on the risk of entering employment, social assistance and unemployment. Transition into employment Table 2 reports the coefficients and standard errors of a discrete event history model on the risk of entering employment. From Model 1 it is obvious that labour immigrants 23

24 enter employment at a quicker pace than refugees. Family immigrants, on the contrary, display similar trends to refugees upon labour market entry. In light of the considerations expressed in the section on descriptive statistics, Model 2 seeks to determine whether the advantage of labour immigrants over refugees might be due to their different composition with regard to socio-demographics. In addition to the independent variables in Model 1, gender, age at migration, level of education, host country education, region of residence, region of origin, year of migration and year of the survey are included. Upon controlling for these additional covariates, the positive effect for labour immigrants decreases only slightly, remaining significant at the 0.1% level. Above and beyond the effects of migration motive, female immigrants enter employment less smoothly than male immigrants do. The impact of age at migration is curvilinear, with the risk of entering employment increasing in the first years but flattening out later on. The level of education is clearly important for labour market entry, though surprisingly, highly educated immigrants do not hold any advantage over medium educated immigrants in entering the labour market. Having a host country qualification exerts a strong positive influence on quick labour market entry, the coefficient (0.46) being significant at a 0.1% level. Results also show that region of origin is an important factor. In comparison to immigrants from old Member States, Asian, Central and Eastern European and North-African immigrants are significantly slower in entering employment, in that order. Immigrants from new Member States on the other hand, enter the labour market at the quickest pace. So far, the analyses have shown that labour immigrants experience a net advantage in obtaining employment over refugees, while family immigrants do not. However, as outlined above, differences might change over time. To test our 24

25 expectations on the time dependency of relative risk of the three immigrant groups, two time-varying covariates are included in Model 3: elapsed time for labour immigrants (Labour*time linear) and elapsed time for family immigrants (Family*time linear). Controlling for these interaction variables, the main effects observed for labour or family groups can now be interpreted as the group advantage (relative to refugees) immediately after arrival (at time 0). The sign of the respective interaction variable then indicates whether the advantage remains constant ( 0), decreases ( ), or increases (+) over time. The estimated parameters for these variables can be found in Model 3. Both for labour and family immigrants a significant positive main effect can be observed. As a matter of fact, this effect now refers to the net advantage of labour and family immigrants immediately after arrival. As expected, the net advantage is larger for labour immigrants (2.48) than for family immigrants (1.14). Furthermore, the negative significant interaction effects indicate that the initial advantages of both labour and family immigrants on the risk of entering employment decline as the time spent in the host country increases. 21 [TABLE 2 HERE] Transition into social assistance Table 3 reports the coefficients and standard errors of a discrete event history model on the risk of entering social assistance. Model 1 indicates that both labour and family 21 The inclusion of interactions between migration motive and elapsed duration significantly improves the model (Δ-2LL = 34.4, Δdf = 2, p =.000). 25

26 immigrants have a lower risk of entering social assistance compared to their refugee counterparts. Both effects are large and significant, though the difference is more pronounced among labour immigrants (-2.92 compared to for family immigrants). After controlling for various other variables in Model 2 these coefficients decrease considerably especially for labour immigrants but remain large and significant at a 0.1 % level. The results in Model 2 additionally show that male immigrants have a higher risk of entering social assistance compared to females. Age at migration is negatively correlated to entry into social assistance, though at a decreasing rate. The direction of the education effect is as expected: the higher the level of education the lower the risk of entering social assistance. We also find that residing in Brussels or Wallonia enhances social assistance entry. Finally, the pattern of social assistance entry differs largely between different origin groups. Africans, Asians and Central and Eastern Europeans are more likely to enter social assistance than immigrants from old Member States, other things being equal. Importantly, we find that the interaction effect between elapsed duration and migration motive is insignificant, indicating that the difference in the risk of entering social assistance between refugees and other immigrants remains constant over time. 22 [TABLE 3 HERE] 22 A likelihood ratio test for the interaction between the labour and family immigrant dummy s and elapsed duration suggests that the effect does not change significantly over time (Δ- 2LL = 2.0, Δdf = 2, p =.848). 26

27 Transition into unemployment The coefficients and standard errors of a discrete event history model on the risk of entering unemployment are displayed in Table 4. From Model 1, which includes only the time intervals and migration motive dummies, it is evident that both labour and family immigrants are at a lower risk of entering unemployment compared with refugees. Interestingly, differences in the socio-economic characteristics almost completely account for the lower unemployment risk of labour immigrants, but fail to explain the lower risk experienced by family immigrants (the significant negative coefficient even slightly increases) (Model 2). As expected, male immigrants have a significantly higher chance of entering unemployment. Age at migration has a positive and curvilinear effect on the risk of unemployment entry. With regard to educational differences, it is notable that only a higher education protects from the unemployment risk, while a medium education does not. Surprisingly, including an indicator on host country education does not seem to have any significant effect on the individual risk of entering unemployment. In Brussels and Wallonia, immigrants have a higher risk of entering unemployment than in Flanders. Finally, origin groups that are particularly vulnerable with regard to unemployment entry are Turks and North-Africans. Asians and immigrants from new Member States on the other hand, transition the least often. In model 3 the interaction effects of elapsed duration and migration motive are included. 23 Looking at the main effects for both immigrant groups, we see that labour immigrants have a significantly higher risk of entering unemployment right after entry, 23 The inclusion of interactions between migration motive and elapsed duration significantly improves the model (Δ-2LL = 19.7, Δdf = 2, p =.000). 27

28 which is what we also found looking at the state distribution plots (see supra). However, the significant interaction effect shows that the initial gap in the likelihood of entering unemployment between labour immigrants and refugees declines over time. Family immigrants, on the other hand, do not differ significantly from refugees upon arrival. However, the significant negative interaction effect shows that the relative risk of entering unemployment increases over time for refugees. [TABLE 4 HERE] Discussion and conclusion The socio-economic integration of refugees has become a key issue in the wake of the recent influx of asylum seekers to Europe. In this paper, we examined the labour market integration of refugees who arrived in Belgium between 1999 and 2009 by looking at their first entry into employment, social assistance and unemployment. Based on a linkage of the Belgian Labour Force Survey with the Data Warehouse Labour Market and Social Protection, we estimated random-effects discrete-time hazard models to compare the labour market transitions of refugees to those of labour and family immigrants of the same arrival cohort. The analysis showed that the labour market integration of refugees leaves much to be desired. First, refugees take significantly longer to enter their first employment as compared to other immigrant groups. Their weak ties to the labour market make them particularly vulnerable to recourse to social assistance, especially in the first years after arrival. Over time, refugees do catch up to some extent and the employment gap with labour immigrants and family immigrants decreases. However, once refugees build up a limited employment history, they also run a greater risk of transitioning into the 28

29 unemployment benefit scheme. The findings hold after controlling for differences in gender, human capital and origin effects. The extent to which the findings outlined in this paper can be applied to more recent arrivals is uncertain conditions have changed in crucial aspects and the composition of more recent inflows is different from the people under focus in the present analysis. Moreover, the findings of this analysis need further validation by larger, more representative samples. Nonetheless, this study is highly relevant for current policy issues concerning the socioeconomic integration of refugees, as it provides a first-of-its-kind view on the labour market integration of refugees in the Belgian labour market, compared to labour and family immigrants. The results indicate the importance of targeted educational and labour market measures to facilitate the best possible integration of refugees into the Belgian labour market and thereby also into Belgian society. The objective of providing fast labour market access not only to refugees but also to asylum seekers is high on the agenda in Belgium. However, our results provide clear arguments in favour of sustainable labour market integration instead of promoting quick access to the labour market. Successful labour market integration can unlock the potential economic benefits of refugees and help minimize the risk of poverty and social exclusion. They offer a potential we can ill afford to ignore. It appears that successful integration comes at a significant cost. However, if we fail to integrate refugees properly and to their full potential, the cost may even be greater. 29

30 References Andersson Joona, P. & Nekby, L. (2012). Intensive coaching of new immigrants: An evaluation based on random program assignment. Scandinavian Journal of Economics, 114(2), Åslund, O., Forslund, A. & Liljeberg, L. (2017). Labour Market Entry of Non-Labour Migrants Swedish evidence. Nordic Economic Policy Review 2017, Åslund, O., & Johansson, P. (2011). Virtues of SIN: Can intensified public efforts help disadvantaged immigrants?. Evaluation review, 35(4), Aiyar, S., B. Barkbu, N. Batini, H. Berger, E. Detragiache, A. Dizioli, C. Ebeke, H. Lin, L. Kaltani, S. Sosa, A. Spilimbergo & P. Topalova (2016). The Refugee Surge in Europe: Economic Challenges. Staff Discussion Note 16/02. International Monetary Fund, Washington DC. Aydemir, A. (2011). Immigrant selection and short-term labor market outcomes by visa category. Journal of Population Economics, 24(2), Bakker, L., Dagevos, J., & Engbersen, G. (2014). The importance of resources and security in the socio-economic integration of refugees. Journal of International Migration and Integration, 15(3), Bakker, L., Dagevos, J., & Engbersen, G. (2016). Explaining the refugee gap: a longitudinal study on labour market participation of refugees in the Netherlands. Journal of Ethnic and Migration Studies, Bevelander, P. (2011). The employment integration of resettled refugees, asylum claimants, and family reunion immigrants in Sweden. Refugee Survey Quarterly, 30(1), Bevelander, P and Pendakur, R.. (2009). The employment attachment of resettled refugees, refugees and family reunion migrants in Sweden. In Bevelander, P. M. 30

31 Hagstrom and S. Ronnqvist. (eds.) Resettled and Included? The employment integration of resettled refugees in Sweden, Bevelander, P., & Pendakur, R. (2014). The labour market integration of refugee and family reunion immigrants: A comparison of outcomes in Canada and Sweden. Journal of Ethnic and Migration Studies, 40(5), Bilgili, Ö. (2015). Evaluating Impact: Lessons Learned from Robust Evaluations of Labour Market Integration Policies. Working Paper, project Integration Policies: Who benefits. Bloch, A. (2008). Refugees in the UK labour market: The conflict between economic integration and policy-led labour market restriction. Journal of Social Policy, 37(1), Bratsberg, B., Raaum, O. & Røed (2017). Immigrant Labor Market Integration across Admission Classes. Nordic Economic Policy Review 2017, Bouckaert, N. & Schokkaert, E. (2011). A first computation of non-take-up behaviour in the leefloon. Leuven: KULeuven, Centrum voor Economische Studieën. Burggraeve, K., & Piton, C. (2016). The economic consequences of the flow of refugees into Belgium. Economic Review, (i), Butschek, S., & Walter, T. (2014). What active labour market programmes work for immigrants in Europe? A meta-analysis of the evaluation literature. IZA Journal of Migration, 3(1), 48. Caritas International (2014). Réfugie ch. Travail: Résultats d une étude sur la formation et l emploi menée auprès de réfugiés. 31

32 Carpentier, S. (2016). Lost in Transition? Essays on the Socio-Economic Trajectories of Social Assistance Beneficiaries in Belgium (Doctoral dissertation). University of Antwerp (Belgium). CGRS (2017). Asylum Statistics. General Overview January Clausen, J., Heinesen, E., Hummelgaard, H., Husted, L., & Rosholm, M. (2009). The effect of integration policies on the time until regular employment of newly arrived immigrants: Evidence from Denmark. Labour Economics, 16(4), Cobb Clark, D. A. (2000). Do selection criteria make a difference?: visa category and the labour market status of immigrants to Australia. Economic Record, 76(232), Connor, P. (2010). Explaining the refugee gap: Economic outcomes of refugees versus other immigrants. Journal of Refugee Studies, 23(3), Constant, A. F., & Zimmermann, K. F. (2005a). Legal status at entry, economic performance, and self-employment proclivity: a bi-national study of immigrants (No. 1910). IZA Discussion Papers. Constant, A., & Zimmermann, K. F. (2005b). Immigrant performance and selective immigration policy: a European perspective. National Institute Economic Review, 194(1), Corluy, V., Marx, I. & Verbist, G. (2011). Employment chances and changes of immigrants in Belgium: the impact of citizenship. International Journal of Comparative Sociology, 52(4), Corluy, V., Marx, I. & Verbist, G. (2012). De impact van regularisatie op de socioeconomische trajecten van immigranten: de Belgische campagne van Belgisch Tijdschrift voor Sociale Zekerheid, 4,

33 Corluy, V. & Verbist, G. (2014). Can Education Bridge the Gap? Education and the Employment Position of Immigrants in Belgium. ImPRovE Working Papers, No. 14/02. Corluy, V., & Vandenbroucke, F. (2015). Polarisation of employment over households revisited: the Belgian case. CES discussion paper series, (DPS15. 08). Cortes, K. E. (2004). Are refugees different from economic immigrants? Some empirical evidence on the heterogeneity of immigrant groups in the United States. Review of Economics and Statistics, 86(2), De Cuyper, P., & Gonzalez Garibay, M. (2013). Het Vlaamse inburgeringsbeleid in het licht van de huidige beleidshervorming. Welzijnsgids, (86), De Klerck, W., Boey, R., Djait, F., Gonzalez-Garibay, G., Jacobs, M. & Raymenants, J. (2016). De Vlaamse arbeidsmarkt na Steunpunt WSE. De Matos, A. D., & Liebig, T. (2014). The qualifications of immigrants and their value in the labour market. Matching Economic Migration with Labour Market Needs, De Silva, A. (1997). Earnings of immigrant classes in the early 1980s in Canada: A reexamination. Canadian Public Policy/Analyse de Politiques, DeVoretz, D., Pivnenko, S., & Beiser, M. (2004). The Economic Experiences of Refugees in Canada. DeVroome, T., & Van Tubergen, F. (2010). The Employment Experience of Refugees in the Netherlands. International Migration Review, 44(2),

34 Dumont, J. C., Liebig, T., & Peschner, J. (2016). How are refugees faring on the labour market in Europe?: a first evaluation based on the 2014 EU Labour Force Survey ad hoc module. EC (2016). An Economic Take on the Refugee Crisis - A Macroeconomic Assessment for the EU. Institutional Papers 33. July 2016 Eurostat (2015). Labour Market Statistics at the Regional Level. Luxembourg: Eurostat. Feld, S., Nantcho. M. & Perin, S. (2006). Educational factors in the economic integration of the foreign population in Belgium. European Population Conference. Fóti, K., & Fromm, A. (2016). Approaches to the Labour Market Integration of Refugees and Asylum Seekers. Publications Office of the European Union. Geets, J., Pauwels, F., Wets, J., Lamberts, M., & Timmerman, C. (2007). Nieuwe migranten op de arbeidsmarkt. Over. werk, Gonzalez Garibay, M., & De Cuyper, P. (2013). An evaluation framework for the Flemish integration policies. Antwerp. Hartog, J., & Zorlu, A. (2009). How important is homeland education for refugees economic position in The Netherlands?. Journal of Population Economics, 22(1), Heinesen, H., L. Husted, & M. Rosholm (2013). The effects of active labour market policies for immigrants receiving social assistance in Denmark. IZA Journal of Migration 2:15. Lancaster, T (1990). The econometric analysis of transition data. Cambridge: Cambridge University Press. 34

35 Luik, M. A., Emilsson, H., & Bevelander, P. (2016). Explaining the Male Native- Immigrant Employment Gap in Sweden: The Role of Human Capital and Migrant Categories (No. 9943). Institute for the Study of Labor (IZA). Martiniello, M. & Rea, A. (2001). Et Si On Racontait : Une Histoire de l immigration en Belgique. Communauté française de Belgique, Brussels Marx, I., Vanhille, J., & Verbist, G. (2012). Combating in-work poverty in Continental Europe: an investigation using the Belgian case. Journal of Social Policy, 41(01), Mussche, N., Corluy, V., & Marx, I. (2013). Migrant access to social security policy and practice in Belgium. Study for the Belgian national contact point of the European Migration Network (EMN). Brussels: European Commission/European Migration Network. OECD (2005). How persistent are regional disparities in employment? The role of geographic mobility. OECD Employment Outlook OECD Publishing. Paris. OECD (2013). Protecting jobs, enhancing flexibility: A new look at employment protection legislation. OECD Employment Outlook OECD Publishing. Paris. OECD (2015a). Indicators of immigrant integration 2015: settling in. OECD Publishing. Paris. OECD (2015b). How will the refugee surge affect the European economies? Migration Policy Debates 8. OECD (2016). Recent labour market developments and the short-term outlook. OECD Employment Outlook OECD Publishing. Paris. 35

36 Pacolet, J., Perelman, S., Pestieau, P., & Baeyens, K. (2007). Een indicator voor de omvang en evolutie van het zwartwerk in de Belgische economie. Petrovic, M. (2012). Belgium: A Country of Permanent Immigration. Migration Policy Institute. Pina, Á., Corluy, V. & Verbist, G. (2015). Improving the Labour Market Integration of Immigrants in Belgium. OECD Economics Department Working Papers, No OECD Publishing. Rea, A., Wets, J., Herman, B., Schepers, W., & Callier, L. (2014). The Long and Winding Road to Employment An Analysis of the Labour Market Careers of Asylum Seekers and Refugees. Reyneri, E. (2001). Migrants involvement in the underground economy in the Mediterranean countries of the European Union. ILO International Migration Working Paper, (41). Rezaei, S., Goli, M., & Dana, L. P. (2013). An empirical study of the underground economy in the Kingdom of Belgium. International Journal of Business and Globalisation, 11(2), Richmond, A. H. (1988). Sociological theories of international migration: the case of refugees. Current Sociology, 36(2), Rinne, U. (2012). The evaluation of immigration policies (No. 6369). Discussion Paper series, Forschungsinstitut zur Zukunft der Arbeit (IZA). Rooth, D. O. (1999). Refugee Immigrants in Sweden-educational investments and labour market integration. Doctoral dissertation. Lund University. Sarvimäki, M. (2017). Labor Market Integration of Refugees in Finland. Nordic Economic Policy Review 2017,

37 Sarvimäki, M. & Hämäläinen, K. (2016). Integrating Immigrants: The Impact of Restructuring Active Labor Market Programs. Journal of Labor Economics. Schneider, F. (2013). Size and development of the shadow economy of 31 European and 5 other OECD countries from 2003 to 2013: a further decline. Schultz-Nielsen, M.L. (2017). Labour market Integration of Refugees in Denmark. Nordic Economic Policy Review 2017, Svantesson, E., & Aranki, T. (2006). Do introduction programs affect the probability of immigrants getting work. Handelshögskolan, Örebros universitet, Working Paper, 3. Tim Van Rie, Ive Marx. Belgium : when growing background inequalities meet resilient institutions, in Changing inequalities and societal impacts in rich countries : thirty countries' experiences / Nolan, Brian [edit.]; Oxford University Press, 2014-p Wooldridge, J. (2002). Econometric Analysis of Cross Section and Panel Data. London/Cambridge: MIT Press. Yu, S., Ouellet, E., & Warmington, A. (2007). Refugee integration in Canada: A survey of empirical evidence and existing services. Refuge: Canada's Journal on Refugees, 24(2). 37

38 Figure 1. Overview of the number of asylum applicants in Belgium. 50,000 45,000 40,000 35,000 30,000 25,000 20,000 15,000 10,000 5,000 0 Source: CGRS (office of the commissioner general for refugees and stateless persons) Figure 2. Share of immigrant by migration motive and entry cohort Family Labour Asylum Study Other Missing Source: LFS-DWH 38