The Impact of Immigration on Local Labour Markets in the Netherlands

(1)

By Wouter de Krom

An Applied Dissertation Submitted to the faculty of Governance and Global Affairs in Partial Fulfillment of the Requirements for the Degree of Msc Economics and Governance

Thesis advisor Dr. A. Afonso

Leiden University 2018

(2)

Abstract

A globalizing world and the European refugee crisis lead to recent surges in migration rates. The economic effect of migration is a highly salient topic and scholars have yet to settle the debate regarding the benefits or drawbacks of the migration phenomenon. This research builds on the migration discussion by making use of Dutch local labour market data, acquired from the Central Bureau of Statistics, in an area analysis. A multiple ordinary least squares (OLS) regression was run using several year-to-year differences in local migration in order to predict the strength of the overall drop in unemployment rates while taking into account endogeneity issues. The results indicate that a higher inflow of migration reduces the downward volatility of the unemployment rates on the short term. The non-Western immigration group in comparison yielded even stronger significant effects. Additionally, the vulnerability of the young age and low educated groups were used as additional outcome variables. The unemployment among the young age groups seemed to be more influenced by increased migration, especially from the Western migration group. The unemployment among the low educated did not yield significant results. Control variables on labour market characteristics include the skill- and educational attainment-levels, job sector types, the logarithm of the population density, the share of female workers and the share of the workforce compared to the total population.

(3)

Acknowledgement

With this dissertation I hope to combine all that I have learned during my Bachelor and Master studies in an independent research project of my own. Here, I learned how to combine

philosophy, management, creativity, statistics, economics, and many other interesting subjects in order to critically assess pressing societal issues. My interest in the immigration effects on the public opinion was sparked during my Bachelor’s thesis, and I wanted to build on this topic by diving deeper into data and econometrics in the final piece, my dissertation. Being able to possibly find new and unique insights through research is an honour and I thoroughly enjoyed writing it. I wish to thank my thesis advisor, Dr. Alexandre Afonso, for his assistance and insightful advice he offered anytime I needed it during both projects. My gratitude goes to the people that surrounded me in these times, and especially to the unwavering support from my parents and my girlfriend.

(4)

CONTENTS

1. Introduction 5

1.1 Relevance………..………6

1.2 Research question………..……...7

1.3 Dutch immigration policy……….………8

2. Literature review 9

2.1 Unemployment………10

2.2 The substitution hypothesis……….…11

2.3 Short and long-term effects………...…..17

2.4 Vulnerable groups………...18

2.5 Immigration as a complement to the native workforce……….. 20

2.6 Opportunities of immigration………..21 2.7 Hypotheses………..…26 3. The data 30 3.1 Explanatory variables………..32 3.2 Control variables……….34 4. Research design 40 5. Descriptive statistics 47

5.1 Immigration in the Netherlands………..47

5.1 The refugee crisis………50

5.2 Descriptive statistics………51

6. Empirical analysis 60

6.1 Assumptions checks………62

6.2 Reverse stepwise regression results………65

6.3 Variants of the dependent variable………..69

6.4 Discussion of the results………..73

6.5 Limitations………..74

7. Conclusion 76

Bibliography 79

(5)

1 Introduction

Due to increasing globalization and violence persisting in several regions around the world, migration and refugee issues are highly relevant phenomena in the present day. This research will try to determine what the effects of the most recent migration flows on the local labour market of an advanced economy like the Netherlands are because it is still unclear what exactly determines these effects and how the existing findings can be generalized. These grey areas of uncertainty could fuel heated and uninformed political debates as well as lead to ineffective policymaking. It is up to the world of political economic research to provide more answers and facts for politicians to work with. Migration has occurred throughout history, but people’s perceptions of its opportunities and threats on the long term are subject to change. Both senders and receivers of migrants are affected by international migration especially in regions which particularly attract people looking for a more promising future or safe refuge. When socio-economic, cultural and demographic differences between the native population and immigrant groups grow larger, concerns especially among the natives tend to arise and politicians respond to the perceptions of migration threats among the public. (Longhi et al., 2010) A recent strong growth in permanent as well as temporary migration contributed to the attention scientists, policy-makers and the public pay to the effects of migration on the labour market. “The number

of foreign-born residents in countries worldwide almost doubled to 200 million in the period 1985-2005”. (Longhi et al., 2010:356) Because many migrants already reside in advanced

economies and the phenomenon continues to be on the rise, carefully examining the impact of migration on host economies and natives is essential. (Jaumotte et al., 2016) Therefore, this research will do just that and will focus on the unemployment rate in the local labour markets as the outcome variable.

1.1 Relevance

The debate on whether migration could benefit a country is an important one since this topic is currently highly political, especially during increasingly polarising attitudes within Europe and the rise of the popular right-wing parties. Policymakers face difficult challenges due to increasingly polarizing attitudes. Voters consider the migration issue to be one of the most notable challenges their countries face due to continuing flows of refugees towards safe zones and high numbers of labour migration to western European countries. The European Social

(6)

Survey (ESS) reports that Eurobarometer surveys on attitudes towards minority groups in the European Union show increasing numbers of Europeans see minorities as threats to society and welfare.1 The number of immigrants moving to and settling in Europe is increasing. However, according to this survey, the openness of European civilians towards migrants and refugees is increasing too.

Existing research on the socio-economic effects of migration shows differing results. These results generally could support either the substitution thesis or the view of migrants as a compliment to the existing workforce. The fact that this debate still needs to be settled could already be a justification for this research. Many scholars in the past measured data at a country level. However, countries have very different characteristics and the differences in which the migration statistics or economic indicators are measured could be quite large. This heterogeneity could be a weakness of national-level comparisons. This research will use geographical-differences data from within a country which has more homogeneity. The outcomes are hopefully generalizable at least for advanced Northern European economies depending on the assumptions checks. It could also give insight into whether people in advanced economies should or should not actually worry about their fellow workers having a foreign background (of course, altruistic considerations set aside). More studies will be seen in the literature have measured the within-country effects in the Netherlands. One has focused on macro-economic considerations and measured the effects per immigrant for the country as a whole. (Roodenburg et al., 2004) However, local labour markets could pose an additional view on the effects of movements towards certain areas within a country. Another study used a similar research setup as this one, but the data has aged since then. (Galloway and Jozefowicz, 2008) Changing times and a new refugee inflow calls for a renewed investigation. Therefore, this study will measure the effects right after the peak of the 2015 refugee crisis. As will be explained in detail in the research design section of this piece, additional measurements will be included to provide more insight into what groups within labour markets are more or less vulnerable. This research will try to add to the literature by answering this question through area differences in the Netherlands.

1.2 Research Question

1

(7)

The two groups that are affected mostly by migration are the migrants themselves (migration will be defined properly in further sections) as well as the people working in local labour markets. This study will be a source of information especially for the latter group. Table 1 shows the different categories of effects of migration for each subject involved. This study will focus on the economic labour market effects for the host country. The dynamics involved in this interplay will be set out in detail in the literature section. In any case, “the Netherlands is

the country par excellence for such a research framework”. (Van der Waal, 2012) To answer

the research question noted below, this thesis will make use of statistical tools which will be described in the research design section and further elaborated on in the empirical analysis. This will mainly consist of an OLS regression of the year-to-year municipality level data differences using readily available data provided by the Central Bureau of Statistics in the Netherlands. Its research question therefore will be the following:

“To what extent does the change in immigration rates in local labour markets in the Netherlands affect the difference in the local unemployment rate?”

The remainder of this study will be structured as follows. The next paragraph will briefly discuss the Dutch migration policies. Then, the literature will be set out in detail, divided into sections covering those studies that saw negative effects on the one hand and those that nuanced these negative effects or even saw positive effects as well as opportunities for solving societal issues through migration on the other. These could include solving the aging problem or diversifying the labour force. An additional section will briefly describe the dependent variable and its implications. Then, the literature will be summarized into workable hypotheses. The research design section will describe the structure of the analysis along the lines set out in research design literature. Then, the dataset and all the variables included in the model will be defined followed by a chapter that poses a factual overview of migration in the Netherlands and the descriptive statistics of the current dataset which will put the findings of this research into context. The empirical analysis will then present the outcome, followed by a discussion of these outcomes by putting them into the context of the literature as well as the shortcomings of this research setup. The conclusion will summarize the findings and will suggest future research propositions.

(8)

Table 1. Roodenburg et al. (2004:17)

1.3 Dutch immigration policy

As will be seen in the literature, governments in different countries respond differently to migration flows and this affects its implications. In the Netherlands, The Modern Migration Policy Act (MMPA) and the National Visa Act took effect from June 2013 onwards. A new feature was the amending of the Aliens Act 2000 with a new emphasis on sponsorship. It also combined the permit application process for the temporary ninety-day residence permit and the permanent or long-term permit into a single procedure. With the new law, the Dutch Immigration and Naturalisation Service (Immigratie- en Naturalisatiedienst, IND), which assesses the applications for residence permits, issues new obligations for both migrants and sponsors including providing information and keeping records to provide for the government. According to the Dutch government website “immigrants are people who come from other

countries to reside in the Netherlands for employment, family reunification or other purposes.”2 Note that the CBS defines immigrants slightly differently. This will be elaborated on in the further sections of this thesis. To receive a permit, one must fall into certain conditions like having a partner with a sufficient income in the case of a family reunification.3 For nationals of EU countries and Liechtenstein, Norway, Iceland and Switzerland, no residence permit nor a report to the IND is needed and all you do need is your passport or identity document. In case of a longer stay than four months, a registration of details in the Personal Records Database (BRP) is sufficient. For Romanian and Bulgarian nationals, employers no longer need an employment permit since 2014. Croatian nationals still require an employment permit however. For non-EU residents the application procedure is more difficult.

2_{https://www.loc.gov/law/help/immigration/Netherlands.php (visited 09/04/2018)} 3

(9)

The MMPA laid down that migrants who are likely to contribute to the Dutch economy, culture or science have higher chances of being admitted. The Netherlands wishes to participate more effectively in the global economy, therefore certain special treatments are made for entrepreneurs and post-doctoral academic staff to create new jobs or improve the quality in the research institutes. Highly skilled migrants who do not hold a nationality within the EU need a recognised sponsor and must be paid an income above a certain threshold.4 Foreign investors with a capital over 1.25m€ can obtain residence more easily.5 In order to obtain a residence permit an asylum seeker needs to meet a set of conditions regarding the nature of the reason one seeks refuge. These include for instance recognized reasons for fear of persecutions, execution or torture, harm by armed conflict or a close family member receiving an asylum permit. When a residence permit is received, migrants might be obliged to enter the integration process. This will be finalized and tested in a civic integration examination, which must be passed within 3 years. After this period the residence permit might be revoked.6

Now that the relevance of the research topic as well as the policy context has been laid out, the upcoming literature review will go in-depth in the scholarly understanding of the phenomenon of immigration and how it benefits or affects socio-economic indicators. The previous section has shown how exactly a migrant wishing to participate in the Dutch labour market can enter the country and which procedure it needs to follow. The descriptive statistics chapter will further present the numbers on exactly which migrant groups have indeed done so and how this changed over time. What the entering of migrants into the Dutch labour market exactly means, will be the research topic of this dissertation.

4_{https://ind.nl/en/work/Pages/Highly-skilled-migrant.aspx (visited 09/04/2018)} 5_{https://www.government.nl/topics/immigration/options-for-entrepreneurs-and-employees-from-abroad (visited} 13/04/2018) 6 https://ind.nl/en/Pages/Integration-in-the-netherlands.aspx (visited 13/04/2018)

(10)

2

Literature review

This chapter will pose an overview of the scientific research scholars have conducted in the past to clarify what process leads up to the formulation of the research question, as well as the multiple aspects building up to the hypotheses and research design. According to Toshkov (2016), theories make assumptions to model the complexity of the social world around us. Since immigration and its effects are a highly complex matter, it would be of the utmost important to distinguish what exactly is investigated and why. As will be seen in the literature review, most studies start with hypotheses derived from classic economic literature. The outcomes show a large variability of the political and policy outcomes which are all relevant to know for both the research setup and its relevance. This section will also give an insight in the mechanisms that lead up to the causal relationships between the variables used in this research. (Toshkov, 2016) The literature splits into sections briefly describing the literature on the outcome variable, before diving deep into the scholarly immigration debate. Many has been written and different research designs, insights and conclusions emerge from this pool of research. Therefore, these studies are structured into two main sections. The first including the studies that find evidence on the substitution hypothesis. The second includes those scholars that found beneficial effects of migration. This dichotomy will be elaborated on further. Finally, this information will be summarized into hypotheses that can be tested in the empirical analysis chapter.

2.1 Unemployment

This paragraph provides a brief description of several writings on the outcome variable of this research; unemployment, and will provide information on what implications unemployment numbers have and how countries compare in terms of unemployment. Further information on the municipality differences regarding this statistic can be found in the descriptive statistics section. At the time of the writing of this thesis, the Netherlands was doing well in regard to unemployment. Unemployment shares have plummeted from 6.9 percent in 2015 to 4.9 percent in 2017. The only bad news is that the unemployment share in 2018 probably will rise in a slower rate than the two years before, to 4.3 percent, due to a slowdown in employment opportunities growth and the labour competition will probably increase. This could be due to the increases of people’s perceived chances on the labour market (however these numbers are

(11)

still lower than expected since the people that pulled out of the job market during the crisis have been slow to return), the rising pension age, the lifting of taxes on labour by the government and last but not least in relation to this research, the asylum influx. Until now though, the number of job opportunities are still increasing faster than the labour supply increases. This is a result of the economic growth. Employment opportunities in 2017 will rise with two percent (measured in the employment opportunities in hours). This rise can be detected especially in the market sector.7

An OECD publication describes the aspects and effects that come along with unemployment. In the first place, households will encounter strain on their finances because there is a lack of sufficient income. This also accounts for the public finances because it will receive less tax income and on the other hand the government might need to spend money on unemployment benefits. Additionally, on the individual level, unemployment might have serious demoralising effects and it could diminish any future career prospects. Young people that are unemployed might therefore especially be affected by being jobless for too long. Unemployment can be a transient or a persistent phenomenon, depending on the country in which it is measured. Unemployment in OECD countries surged during and after the crisis but has since seen a steady decrease in most countries including substantial falls in Estonia, Latvia, Spain, Portugal and Hungary. The first two even halved their unemployment rate since then and the others have lowered unemployment for around 5 percentage points. In most countries however, the youth unemployment rate is double the adult unemployment rate. (OECD, 2016) As will be seen in the descriptive statistics section of this research, the Netherlands fares relatively well in their unemployment rate. For the main part of this research, the unemployment rate will make up the dependent variable.

That unemployment carries with it negative effects to any society is clear. As will be seen in the literature section however, some societies’ social economic characteristics respond differently to an inflow of migration than others. The next section will provide an overview of the studies that relate migration to (among wages and GDP) unemployment in the receiving labour markets. Any negative effects due to a ‘brain drain’ on the labour markets where migrants tend to leave, will not be taken into consideration. This falls outside of the scope of this research which will solely focus on immigration and the share of foreigners.

2.2 The substitution hypothesis

7

(12)

This literature section will pose an overview of studies that focus on the immigration question. In this chapter these will be sorted according to the nature and focus of these studies, and eventually the hypothesis section will link the existing theory to several predictions which will be tested in the empirical analysis. The dynamics regarding this phenomenon will be elaborated, as well as the threats and opportunities that might come along with it. First, the negative consequences of the inflow of labour workers from abroad will be set out in the upcoming paragraphs. The next section will pose alternative views on the immigration phenomenon. In the overall narrative different research setups will also be discussed to show that the literature includes a diverse range of strategies, although with often recurring methods like the area analysis which will eventually be used in this research. This could also be relevant to discuss since some studies write that the variety of models could explain the different findings, along with country-specific policies. Then, the final section of this chapter will conclude the general ideas and devise testable hypotheses, based on the literature.

The substitution hypothesis is a recurring theme in the literature and suggests that the foreign workers could replace natives. Because many studies on the substitution hypothesis result in different findings and more within-country evidence could be relevant to add to the understanding of these differences, Van der Waal (2012) focuses on the Netherlands, in particular on 22 Dutch metropolitan areas. Both studies that focus on the impact of immigration on wages and those that assess the likelihood of unemployment suggest the substitution hypothesis is dependent on the type of urban economy. The substitution hypothesis in this case refers to the idea that immigrants and their descendants could substitute native workers in the labour market. The study by Jean and Jimenez (2012) for instance suggests that the scattered results in the literature indeed depends on differences in policies between the countries. Also, labour demand, according to various studies, is strong at the lower-educated level in the developed countries which implies that there would be a less strong effect of immigration on the unemployment rates of natives in the lower-educated sectors. According to Van der Waal (2012), this could explain why studies focusing on within-country differences yield varying results. Van der Waal (2012) states that the Netherlands is an ideal country for a between-cities within-countries research framework because the impact of immigration on unemployment numbers are less likely to be the result of differences in the labour market policies between cities. The reason for this is that the Netherlands employs quite a centralistic labour market interference compared to other Western states. Van der Waal employs a multiple-level difference between cities and years. Because of this multilevel structure, he used multilevel

(13)

modelling starting with a null model to show the variation of unemployment for lower educated people at both the city and time levels. He also uses the variables producer services and share of lower educated in relation to the unemployment level which both yield negative significant coefficients, underlining the statement that there is a high demand for low-skill workers. Their settlement is not labour market driven completely, because several cities have higher unemployment level than several others. Van der Waal (2012) uses several control variables in a multilevel regression analysis including the working population, share of lower educated, and ten-year age groups ranging from the age of 15 to 54. One of the conclusions resulting from the models is that urban regions with a high employment share in advanced producer services see a weaker impact of immigration on unemployment, compared to regions with a low employment share in this sector. Overall resulting from this study is the fact that the substitution hypothesis holds up, but that it largely depends on the skill categories of workers and policies. Immigrants compete mostly with lower-educated natives and former waves of immigrants which in turn see their wages decline. Van der Waal (2012) further states that future studies that take note of these aspects of urban economies, like the fact that these economies differ in labour demand while the logic of immigrants is not necessarily based on market logic because they also settle in cities with low employment demand, will find more meaningful results especially in the Netherlands.

The finding that certain subgroups have a hard time competing with immigrant labour is one that returns in more studies. Pocher (2011) examines the effects of immigration on wages and employment prospects of natives through case studies in Germany and the United Kingdom. Her study was a response to the Eastern Enlargement of the European Union in 2003, which led to government worries about migrants replacing domestic workers or put downward pressure on native wages. At the time, German policy implemented freedom of labour-movement constraints to avoid increases in native unemployment rates and as a result higher social benefit spending, for the eight new East European countries (A8). Other EU15 countries came with similar policies, like the UK which limited the inflow of non-EU citizens and only highly skilled non-EU workers could migrate to the country. Pocher (2011) found it interesting to investigate labour market effects in both developed, industrial and wealthy countries with usually a positive Gross Domestic Product (GDP). She finds that foreign workers and the British are imperfect substitutes. Middle educated immigrants and natives had the highest elasticity of substitution. In the UK, positive wage effects were higher than negative employment effects. Natives in West-Germany with low and middle education lost in employment to migrants with similar education but gained in their wages. In East Germany, a

(14)

minimal wage increase was detected alongside a small rise in low educated native employment but no changes in the employment of middle- or high-skilled workers.

These studies were examples of scholars that found evidence to support the substitution thesis. An extensive amount of research has been done on the labour effects of immigration. Many different research techniques have been used to define the natives’ labour market position after labour competition increases. Most of it have been used to assess the recipient national economies. This is mainly due to the EU enlargement, more globalization, and an increasing overall global migrant labour flow over the past fifty years. In the EU, the accessions of the eight Central and Eastern European countries and the refugee crisis have led to a larger controversy concerning the labour market, fiscal and macroeconomic impacts on both the exporting and receiving economies. According to Pouliakas (2014), studies that focus on whether migration harms native employment and wages could see the most significant impacts on the sub-national level, while most studies are conducted on a national level. Migrants would be drawn to high-density population areas where better employment opportunities, higher wages and less traditional lifestyles are more common. (Pouliakas, 2014; Blanchflower et al., 2007; Phimister, 2005)

One common denominator of most studies on this topic is the baseline neoclassical economic model of supply and demand which is referred to the literature sections. The neoclassical model of supply and demand explains the uncertainty that comes with migrant inflows for natives in the labour market, because immigrants raise the price of factors they complement and lower those which they perfectly substitute. This uncertainty often leads to popular fears of the adverse consequences of immigration also regarding unemployment rates may rise due to an excess labour supply. (Pouliakas, 2014) Okkerse (2008) also concludes that most studies find negative consequences of immigration. The higher fractions of migrants in specific labour markets could lead to lower wages and higher unemployment rates. Language barriers and cultural misunderstandings could diminish any positive total welfare advantages or effects of a more diverse workforce due to migration. (Brunow and Benzel, 2012) These language and cultural barriers are thought to be more pronounced for non-Western immigrants, because non-Westerners are logically further removed in this sense than are Westerners when they come to a Western country. This of course does not mean that they cannot integrate and become stellar citizens. It might however lower their chances compared to Western immigrants. This idea will be referred to in the hypothesis section.

Not all studies suggest that immigration causes drawbacks alone. Some claim that migration could have a positive impact on the local labour market (which will be elaborated on

(15)

further in this literature review). But these claims are often based on assumptions that could easily be violated. (Coleman and Rowthorn, 2004) These are that immigration usually consists of relatively young working-age people. Another is that they perform just as well as natives, and this claim is usually false in advanced economies, according to Coleman and Rowthorn (2004). They find that consequences of large-scale immigration are negative or affect certain vulnerable group’s interests, and that economic or fiscal benefits do not outweigh these effects. They find no evidence that in the UK anyone other than the immigrants themselves would benefit from migration and that solely altruistic considerations could make an argument for the more immigration-supporting policy changes in the UK in 1997. These findings contradict with other studies done by Winkelmann and Zimmermann (1993) and De New and Zimmermann (1994a), which will also be elaborated on further in this section.

The fact that the probability of losing your job as a native to foreign labour (displacement risk), and the risk that these workers will not find new employers because the increased relative unemployment leads to increased competition (job-search effectiveness), might rise due to immigration often leads to responses from the public. The perceptions people have of these risks often leads to negative public opinion. (ESS) This might be justified in some cases. Winter-Ebmer and Zweimuller (1999) for instance find that natives could indeed be displaced by immigrant workers in the case of Austria. These public responses to immigration flows were also seen in the country of analysis of this study, the Netherlands. The Central Plan Bureau (CPB), an institution that often makes predictions of policies as well as recommendations to the central government, responded to these developments by publishing an article about the effects of additional migration to the workforce on a macro-economic level. This study will be described in the following section.

Immigrants coming from non-Western countries tend to make more use of welfare state benefits and arrangements in the Netherlands. (Roodenburg et al., 2004) Therefore, Roodenburg et al. (2004) state, the immigration case is not just of academic relevance, it also has a fiscal impact. The authors see the fiscal impact as a risk to the welfare state because it could increase its costs and will make the eventual public fiscal balance worse off. This effect could worsen the problem of the ageing population. However, immigrants tend to be of a relatively young age and therefore have the potential to alleviate this burden. Roodenburg et al. (2004) provided an overview of the impact of migrants on the Dutch economy in the year 2003 using a forward-looking approach to offer advice on policy-shaping regarding migration for the future. They note that these choices should be based on non-economic as well as economic decisions, and that other countries have stakes in this type of policy-making as well. They

(16)

however focus on the impact on the Dutch economy and assess the impact of immigration on the labour market as well as the public sector within the Netherlands. They too address the assumptions needed to improve fiscal impact on the long run. They state that indeed migrants need to be as productive as natives to improve the host country situation. Therefore, the immigration of labour is not exactly effective in posing solutions for the ageing problems in the Netherlands. The fiscal impact of immigrants depends on the age of entry and the socio-economic status. Their study finds that immigrants form a burden to the government budget balance even when the immigrant’s social economic characteristics are equal to those of the non-Western residents, while families result in a larger burden. During times of increased public pressure coming from interest parties like employers and the immigrants themselves, the Dutch authorities need to make hard decisions and, according to Roodenburg et la. (2004), a selective restrictive policy regarding providing access to the Netherlands and its welfare arrangements is necessary in order to achieve a successful policy. They argue that the immigration of labour on the large scale is not an effective tool to counter the effects of ageing on the financial balance as long as there are no improving labour market effects to be measured due to immigration. Until the time of the CPB report, the government had correctly used this argument to defend its stance on migration policy. They back this line of thought with a few findings. First of all, the domestic income of the country increases. On the long run, those natives that possess skillsets similar to the immigrants will win. Capital owners will win too. However, these wins will disappear on the long term. The rise of domestic income of the country benefits mainly the immigrants due to the wages they receive. Those natives that have similar skillsets as the immigrants will lose. Overall, the general net gain in income will probably be small and might even be negative. Roodenburg et al. (2004) find that usually non-Westerners (besides the Surinam migrants) have higher dependency ratios on social transfer programmes including disability and unemployment benefits and welfare. A main factor for this relatively high dependency is a lack of education, which explains one third of the difference in employment rates, as well as differing age and sex compositions. They stress the importance of the language problems since 70 percent of the Turkish and 60 percent of the Moroccan migrants experience issues with the Dutch language while preferring to speak their mother tongue in their private circles. An external factor Roodenburg et al. (2004) touch upon is discriminatory behaviour by potential employers.

So overall, the Netherlands as a whole does not seem to benefit from migration. Even though the conclusion of the CPB publishment is quite straightforward and interesting, this however does not yet settle the debate. It has the characteristics of a footprint study, calculating

(17)

the benefits and losses of an additional migrant. Many studies have taken different approaches. The level of analysis affects the outcomes because it changes assumptions and dynamics between the variables. When these assumptions and dynamics change, so does sometimes the rather pessimistic outcome, as will be seen in the next paragraph. The year in which the CPB rapport was written could also mean that it is slightly outdated. In any case this study will too focus on the Netherlands, like the CPB study, but it uses a completely different approach.

2.3 Short and long-term effects

That conclusions on this topic are most certainly not straightforward and all-defining, is an idea that is established when we compare several specific studies. Some articles, for instance, find differences between measuring the effects of immigration on the short and the long term. We already saw this in the conclusions Roodenburg et al. (2004) made. On the short term, Winter-Ebmer and Zweimuller (1999) find that natives indeed could be displaced by immigrant workers in Austria. However, on the long term these effects are less severe because these natives are able to move between industries or regions. Sufficient mobility therefore could be said to eliminate the threat of structural replacement. The authors looked at differences in two fixed points in time, during the 1988-1991 period when a large influx of foreign workers was taking place. They find that even though the results were modest, certain subgroups like already employed immigrants or seasonal workers, were especially affected by the inflow of migrants (even though these effects were not always significant). The increase of immigration share with a percentage point leads to an increase unemployment duration by about 5% or 5 days and was especially pronounced in the case of young workers. These young workers were the focus group of their research and comprised of workers until the age of 35. (Winter-Ebmer and Zweimuller, 1999)

There are more studies in which the effects are negative but differ over time. An OECD working paper assesses the consequences of immigration on the unemployment risk of natives in OECD countries. Jean and Jiménez (2011) find no long-term impact but see a period of approximately five to ten years in which the natives’ unemployment numbers increase. Several factors contribute to the impact of this higher risk of unemployment, including anticompetitive market regulations, a higher replacement rate of unemployment benefits and more sturdy unemployment protection legislation. So, less market interference leads to lower-educated natives and immigrants to be less exposed to unemployment risks. The fact that these effects are dependent on the institutional design regarding immigration within countries could perhaps

(18)

explain the between-countries effects that results from the literature. However, the large differences in the results regarding the thesis that labour migration is a substitution for the native workforce still need to be explained properly. Winter-Ebmer and Zweimüller (1999) regressed in a pooled cross-section the immigration and the displacement of natives under 35 years old while controlling for unemployment rates, wages, education, seasonal workers and the industry foreign share, and found only a minor impact of immigration on the unemployment risk of natives on the short term. Gross (2002) investigates the effect of immigration on the French labour market. The immigrant flows increase unemployment slightly similar in terms of increased competition to an increase in the native labour force in accordance with economic theory that states that a larger sudden supply of labour will lead to unemployment in the short term under unchanging market circumstances. Also, the characteristics of migration flows matter, and the migration flow composition matters including those coming from the European Union and those originating from countries outside of the EU, as well as the distribution between more and less skilled workers. This needs to be properly balanced. On the long run however, all types of immigrant workers lower the unemployment rate, according to the time-series methods employed in these studies. This includes policy measures or the tendency of migrants to gravitate towards certain areas, as the upcoming sections will show. This dissertation will mainly look at the short-term effects of immigration. This choice will be further explained in the research design section. Gross (2002) further elaborates on policy implications including the screening of immigrants before they enter the country of destination in order to adapt the composition of the labour market. Therefore, not all studies reach conclusions that are pessimistic regarding migration. More studies like this will be described in the next paragraph. As stated before, the effects of migration could depend on the context.

2.4 Vulnerable groups

Several factors contribute to the impact of this higher risk of unemployment, including anti-competitive market regulations, a higher replacement rate of unemployment benefits and more sturdy unemployment protection legislation. So, less market interference leads to lower-educated natives and immigrants to be less exposed to unemployment risks. The fact that these effects are dependent on the institutional design regarding immigration within countries could perhaps explain the between-countries effects that results from the literature. However, the

(19)

large differences in the results regarding the thesis that labour migration is a substitution for the native workforce still need to be explained properly.

Musterd and Muus (1995) assess the spatial settlement differences among five groups of immigrants including labour migrants, asylum seekers, migrants from rich countries, migrants from EU countries and migrants from former colonies. They find that many immigrants arrive in large cities because they look for cheap housing opportunities, but their chances of finding a match within the labour market in these same urban areas are bad. There are only scarce employment opportunities within these areas for newcomers which leads to mismatches. Due to these mismatches many policies like the Act on Equal Employment Opportunities from 1993, which should have improved the participation rates among immigrants, lost their value.

Longhi et al. (2010) find that the employment impact of immigration in Europe is stronger than in the United States. Demarcation across geographical differences and skill matters and according to them, rigid wages enhance the employment impact on natives. They note that most studies on the international migration effects analyse at a national level while between-regions or cities evaluations are relatively rare. Immigrants and emigrants are often spatially clustered. Migrants are also subject to a high level of heterogeneity, which is why there is evidence of differences in socio-economic impact of immigrants. Examples of characteristics that lead to this heterogeneity are education, skills, gender, age, cultural background, motivation and welfare position. The wage and employment impacts of immigration shocks are interdependent through the wage elasticity of demand and supply of labour. A meta-analysis results in a .029% native wage decrease and a .011% native employment decrease because of a 1% increase of the share of immigrants in the local labour market. The impact of other workers on a labour supply increase depends on the extent to which different labour types are substitutes in production and the change in the output composition of firms on the supply side. On the demand side, public expenditure on health and social security payments will depend on the characteristics of the migrants, the location and whether education or health sectors were operating at full capacity.

A study by Pouliakas et al. (2014) modelled the effects of immigration on regional economic indicators using a computable general equilibrium (CGE) technique of three small remote European Union regions. Over two scenarios, the total labour supply is affected and the differential labour skill types in the importance of migratory flows is investigated. They found that remote areas are more vulnerable to the ‘brain drain’ phenomenon, the out-migration of skilled workers. In-migration of unskilled workers led to a widening wage inequality because

(20)

small regional economies were relatively unable to adjust to the inflow of low skilled labour supply. Therefore, the skilled wage premium was widened. These findings correspond with those by Borjas et al. (1997), Ozden and Schiff (2005) and Østbye and Westerlund (2007).

2.5 Immigration as a complement to the native workforce

As seen in the previous paragraph, many studies with very different research designs and setups warn for the negative influence of additional workers from abroad on the current workforce and public finances. This usually fuels sentiments among the public that the politicians should create protective policies to preserve the economic and labour market conditions that people are used to. However, as in many sensitive study topics, one could argue in favour and against certain phenomena, preferably backed by scientific evidence. This paragraph will nuance the findings of the studies presented above. This should eventually set the stage for the current research and show how exactly it contributes to the debate. There have been numerous studies that show that migration might have benefits for a country or at least argue that certain factors mitigate the findings that support the substitution hypothesis. When the public tends to quickly form its opinion on these matters, academic achievements could make for a more well-thought out and balanced policymaking environment. Note that this study focuses on the measurements of economic indicators, and that any discussion of public opinion merely provides context. The negative public opinion towards immigration, like the ESS investigates, is usually made up of perceptions of risk. These natives fear for losing their job (displacement risk) and the risk of not being able to find a new employer when part of the unemployed due to increased competition from foreign workers (job-search effectiveness) is higher.8

In Austria the risk for natives losing their jobs to increased competition was highest for young workers because they made up the largest part of the unemployment entries while at the same time low-wage workers face higher probabilities of joblessness. (Winter-Ebmer and Zweimuller, 1999) However, the job-search effectiveness risk associated with immigration is larger for older workers that become unemployed by chance. Politicians that argue foreign workers will replace natives on a one-to-one ratio, do three assumptions that could easily be violated. One of them is that the new workforce will not affect economic growth, if the job possibilities or the number of jobs remains the same. Furthermore, another assumption is that

(21)

the wages for the incoming foreign workforce are lower than those for native workers. The last assumption is that foreigners and natives are perfect substitutes. Another view could be that immigrants take jobs that native would not do, and instead the natives could see prospects of better job opportunities and therefore complement the existing workforce without any substitution taking place. However, both views would be unrealistic and empirical results probably show different stories. (Winter-Ebmer and Zweimuller, 1999)

2.6 Opportunities of immigration

Even though the native workforce might often see migration as a threat, it could also pose opportunities for countries to boost their GDP, employment rates and lower the public deficits. Additionally, many models produce outcomes that are quite dependent on the assumptions of the models or the research design. Changing its assumptions could lead to different conclusions, like assuming that capital is internationally mobile. These capital flows could lead to a balancing of the capital-labour ratio which leads to no effective changes of immigration compared to the pre-migration period. Intersectoral capital mobility could also affect results. In a multisector model, immigration will decrease the wage rate when factor price insensitivity no longer applies. Finally, the openness of an economy could affect the impact of immigration. (Okkerse, 2008; Borjas, 1999; Rivera-Batiz 1983)

Pischke et al. (1997) also describe how there could be a variety of reasons why the negative effect of immigration on native labour due to the standard competitive model could be diminished. The first reason would be that natives and foreigners would be complementary in production, and second that the labour markets might not be well described by this competitive model. Schmidt et al. (1994) explain how a monopoly could affect low skilled wages. One type of analysis Okkerse (2008) writes about is the area analysis, which is also used in this study. It explains how migrant populations could concentrate in certain geographical areas and that econometric analyses could exploit this diversity to find effects of migration on regional labour markets. These studies use regression models in the following form: Yi= α + Xiβ + γ Pi+ ui. The dependent variable is the measure of the labour market performance in area i, like the unemployment rates, participation rates or average wages. Xi would be the slope of the independent variables including size, density, education level and share of females within the population, and the key independent variable would be y, including the share of migration or immigrants within the region Pi. This area analysis needs to deal with some issues however. One is the endogeneity problems arising from the fact that migrants choose their destinations

(22)

on wage or unemployment levels. Immigrants could choose to settle in destinations with the better labour market conditions, which is the opposite causality which this model is looking for. The correlation between the two variables could be a sign of net effects instead of a causal relationship. Ordinary least squares (OLS) estimates require independent variables to be exogenous, which means that the ratio or numbers of immigrants on a regional level cannot be influenced by the wages/unemployment rates. (Okkerse, 2008) According to Gurak and Kritz (2000) this is quite unlikely since regions with strong economic growth attract higher numbers of migrants. Therefore, to avoid these biases due to endogeneity problems this study will incorporate a simple solution. Okkerse (2008) poses one technique to solve this problem using instrumental variables (IV) estimation. A possible instrument would be the regional share of immigrants at the beginning of the period, relating it to the change in migrant shares. The reasoning is that an originally higher number of immigrants tend to attract more migration due to a larger information availability and the connections to friends and family. However, as can be read in the research design section, this research will incorporate different techniques to avoid endogeneity problems.

A second problem with geographical differences analysis comes from the fact that labour or capital could be moved to other areas by natives, which could lead to nothing being measured locally but there would be effects. (Borjas, 1999) Several studies on these balancing outflows show several results. Inwards migration of natives is lower in areas with higher numbers of immigrants. Higher levels of recent immigration also lead to lower chances of receiving or keeping native workers. Finally, another study finds that higher numbers of immigrants in an area lead to low native in-migration and high native out-migration. (Filer, 1992; White and Liang, 1998; Frey, 1995; Borjas, 2005) Several researchers dealt with the problem of native migration responses by using occupation, industry, experience groups or education as the units of analysis instead of the geographical areas. It would be more difficult to switch sectors or industries than areas in times of increased competition on labour markets. (De New and Zimmermann, 1994; Camarota, 1998; Card, 2001; Orrenius and Zavodny, 2003; Borjas, 2003)

Jaumotte et al. (2016) addressed the issue of long-term socio-economic impact of migration while taking caution for the risk of reverse causality. Different studies have suggested that immigration could raise the total factor productivity. Jaumotte et al. (2016) differentiate between microeconomic studies focusing on certain outcome variables (including productivity, innovation, wages and employment) and macroeconomic studies, which includes their own. They expect several effects of immigration on the GDP per capita of host

(23)

economies. The share of the working-age population may be increased. The labour productivity per worker could also be affected in both positive and negative ways, to the extent that their skill sets are complementary to the native workforce. The study analyses the effects on income levels in advanced economies, the difference in impacts as a result from different skill levels of migrants, and the way the gains are shared across the population. Included in their econometric models are robustness tests to address the risk of reverse causality and effects of control variables. They find that immigration through raising the labour productivity has a positive effect on the GDP per capita of host economies. A one percentage point increase in the migrant share leads, according to them, to a two percent increase in the GDP per capita on the long term. This effect can be seen for both high- and low-skilled migrants because the existing skill set within the society becomes more diverse. They also find that immigration gains are broadly distributed among the population. The positive effects are a result of a balance between on the one hand young and dynamic migrants complementing the labour force and solving the problem of the ageing population and thus helping to sustain public expenditures. On the other hand, if immigrants have a hard time integrating or if they lead to the displacement of natives, this could lead to political backlashes because of the increasing social tensions due to culture and language differences and added pressure on social security. (Jaumotte et al., 2016)

Corluy et al. (2014) studied the performance of immigrants in the receiving country’s labour markets in Belgium, a country unique in the EU15 since the employment rate gap is largest between immigrants and natives. Immigrants’ poverty risks are among the worst in Europe in Belgium. This gap is partly due to the low employment levels of immigrant women (the gap in question for women is twice as high as for men), while males slightly improved their situation. Another interesting finding is that the relatively bad labour position of immigrants is mainly due to the differences in levels of education of these groups compared to natives. Human capital includes country specific knowledge of the local language, labour market etc. which can be a very important indicator for employment opportunities of migrants. Natives and EU immigrants had a significant advantage compared to immigrants from a non-Western background over the course of a few years due to their improving educational level. The return on education differences between immigrants and natives are quite large as well. Corluy et al. (2014) writes that the current employment rate gap in Belgium for EU immigrants can be explained entirely by socio-economic characteristics such as educational levels, but also age, household size and the presence of children show relevant structural differences. Corluy et al. (2014) also write that it would be interesting to investigate the second generation of immigrants, since this reduces the heterogeneous characteristics of more short-term migrants.

(24)

Brunow and Brenzel (2012) take a view on the effects on immigration through increased demographic diversity. Their unit of analysis is the region on an EU level and their outcome variable is the GDP per capita. The increased diversity, according to Brunow and Brenzel (2012) might lead to increases of productivity since they have a positive effect on skill diversity within the population while working together with natives. The sole fact that an inflow of workers leads to an increase in the working population cou ld pose advantages too, especially in times of demographic changes like the ageing of society. They combine Eurostat regional data with the European Labour Force Survey and describe the regional classification through the NUTS levels 1 and 2 of aggregation, which, compared to NUTS level 3 overcomes strong spatial interdependencies. They find that an increase in the cultural diversity within a region improves the per capita GDP through the increase of the proportion of foreigners and the diversity between the foreign-born population alongside a tendency towards dominant groups, which leads to higher incomes due to lower integration costs. Brunow and Brenzel (2012) see immigration as an opportunity to balance out future labour shortages, and that usual measurements of participation and employment rates have too little concern for productivity gains.

Cummer et al. (2015) studied the improvement of general human capital due to migration in Canada resulting from new laws focused on changing the composition of skills. They examine the following immigrant entry earnings. After the initial earnings improvements of immigrant cohorts, further policy changes eliminate these improvements. Immigrants, following the policy changes, are subject to higher chances of periodic unemployment compared to natives. The cultural diversity factors explain these changes, like language barriers or their status of minority. Furthermore, increasing levels of immigration in the Canada workforce are concentrated in certain regional labour markets and skill groups. These results support the idea of policy implementation that matches immigrants with employment opportunities before landing.

Overall, according to a literature review by Okkerse (2008), most area and time-series analyses see no significance in the influence of immigration on employment possibilities. In the short run, according to these findings the chances that migrants increase unemployment is low and in the long run it is even zero. Coulon (2005) did not find any significant effect of increased immigration on the employment of natives. The original ordinary least squares (OLS) estimates yielded a significant negative effect but these were corrected after accounting for potential endogeneity due to an upward bias in the estimate. Winter-Ebmer and Zweimüller (1999) regressed in a pooled cross-section the immigration and the displacement of natives

(25)

under 35 years old while controlling for unemployment rates, wages, education, seasonal workers and the industry foreign share, and found only a minor impact of immigration on the unemployment risk of natives on the short term. Venturini and Villosio (2006) researched the time it took for workers to find work after being unemployed to examine the displacement risk of natives as a result of immigration. They found that in Italy, foreigners and natives are compliments for job effectiveness and displacement risks.

Pischke and Velling (1997) analysed the impact of migration on the employment outcomes of natives using country-level variables in Germany. They adapted the 328 countries to 167 regions in order to utilize more meaningful labour market regions. They use the change in the share of migrants and the net flows of immigrants to an area within one year. They also condition on historical labour market outcomes in order to tackle the problem of immigrant selection of labour markets, which, they write, allows for mean reversion in unemployment rate. Two sets of regressions are included in their research. The first one compares changes within the years 1985 and 1989 and the other looks at variables in year-by-year regressions. They use two dependent variables, namely the unemployment and employment rates. They do not include wages as a dependent variable due to the country situation because at the time unions were successful in implementing similar wages across country regions. After analysing substitution effects between natives and immigrants and accounting for the unemployment rate and employment-to-population ratio dynamics, they find no significant effects of immigration and neither do they find evidence to support the hypothesis concerning the native migration patterns evening out any displacement effects. The unemployment rates show no meaningful difference or effect stemming from migration patterns. Other studies found different conclusions. Winkelmann and Zimmermann (1993) and De New and Zimmerman (1994) for instance do find a negative relation between the share of foreigners, the employment share and wages within industries. Pischke and Velling (1997) explain that the difference stems from the fact that De New and Zimmerman (1994) differentiate for blue- and white-collar workers, but state that this is too rough of an indicator for skill. Low-skilled white-collar service workers are not considered. Another difference with this study is the fact that De New and Zimmerman (1994) look at static levels of the variables instead of changes.

Another study by Blackstone et al. (1998) examines the impact of migration on native unemployment and groups together immigrants according to certain characteristics including age, marital status and sex. They use a time series cross sectional and polynomial distributed lag methodology. They only find an impact on aggregate native unemployment through a lag.

(26)

Native workers both benefited and suffered from the inflow of foreign labour depending on the characteristics of the subgroups.

The choices of settlement within certain urban areas made by immigrants have been subject to several studies as well. Bauer et al. (1991) use individual-level data combining economic and demographic postcode level data to analyse the effects of immigration on employment probabilities and wages in Germany. They employ an instrumental variable method to analyse any negative effects of migration on the labour market and differentiate between low- and high-skilled natives. They investigate the hypothesis that migration negatively affects wages and employment and build on similar studies, and actually find that immigration has no-or an irrelevant-effect on employment and wages of natives. (Borjas, 1999, 2003; Friedberg and Hunt, 1995; Longhi et al., 2005; Traca, 2005; Zimmermann, 2005) Bauer et al. (1991) also write that clustering of migrants with the same backgrounds could have positive as well as negative implications. On the one hand, it could facilitate the integration and settlement process as well as making things easier for people exposed to new surroundings. However, it could have drawbacks for integration incentives like studying languages. This cultural diversity could also be related to the productivity and thus the labour probabilities of natives in either a positive or negative sense, but theoretical models often neglect this fact. Effectivity might be increased due to more diversity, innovation and creativity. Indeed, most studies find positive effects on the productivity of natives. (Ottaviano and Peri, 2006; Südekum et al., 2009) They use local labour markets as their level of analysis, defined by Kropp and Schwengler (2008). In this study this leads to an aggregation to 103 labour markets from 1682 postcode regions. Findings indicate that there is a positive relation between employment and share of foreigners, but no significant one between immigration and wages of natives.

The dispersion of the migrant settlement records in different areas leads to interesting research possibilities. The research design section of this study will go into more detail as to how this will be used in the analysis. The next section will link the literature to the hypotheses.

2.7 Hypotheses

The above-mentioned literature should paint a complete picture of the dynamics that are involved between migration measurements and the independent variables and labour market or economic indicators as the dependent variables. Additionally, it has posed an overview of how these measurements had taken place and how this lead to many differences in the conclusions of studies. This section will pose a quick summary of the main findings from the literature

(27)

section into workable hypotheses. From the literature there are usually central tendencies we can derive in order to come up with proper hypotheses. The first would be the effect of overall immigration on employment rates in the receiving regions. A lot has been written about it and this section will summarize these ideas into the main hypotheses. On the one hand, studies find that immigration has a positive effect on unemployment rates. This could be attributed to several causes.

In general scholars refer to the neoclassical model of supply and demand within the labour market. A higher supply of labour leads to more severe competition, all other things equal. Some studies indicate that an increase in the foreign labour force has a similar impact on unemployment as an increase of the native labour force. Other studies say foreign labour has a stronger negative effect. However, most studies conclude that migration is not beneficial for the labour market situation especially on the short term. Any possible welfare advantage resulting from a labour force increase due to immigration is often hard to achieve in practise due to the difficult cultural integration process. Also, migrants often do not perform as well as natives in terms of productivity. As seen in the literature review, this could have multiple explanations like the settlement in areas with poor chances for these same migrants or the fact that cultural integration and social separation are additional obstacles for entering the labour market. (Van der Waal, 2012; Coleman and Rowthorn, 2004; Roodenburg et al., 2004, Musterd and Muus, 1995; Winter-Ebmer and Zweimuller, 1999) The hypotheses will be tested according to geographical-differences analyses, like many other scholars have done as well. (Pouliakas, 2014; Blanchflower et al., 2007; Phimister, 2005; Van der Waal, 2012; Bauer et al., 1991, Musterd and Muus, 1995; Longhi et al., 2010) What makes this research relevant however is described further in the previous section. Especially on the short term most studies find unemployment numbers to increase. (Jean and Jiménez, 2011; Winter-Ebmer and Zweimuller, 1999) But more evidence is needed on a subnational level. (Pouliakas, 2014) Therefore;

H1: An increase in the year-to-year difference in the share of the people with a foreign

background within the labour force increases unemployment measured on municipality level.

Even when non-Western residents have similar social economic background characteristic, they are still a burden to the government budget, according to Roodenburg et al. (2004). The inclusion of the non-Western measure is a product of the underlying idea that cultural integration, as stated above, is an additional barrier or obstacle in the addition to migrants with

(28)

these backgrounds to the local labour market. Brunow and Benzel (2012) also write that non-Western immigrants face more barriers. Corluy et al. (2014) find that natives and EU immigrants are better off on the labour market due to their improving educational background. Cummer et al. (2015) find that cultural barriers lead to higher chances of periodic unemployment. A recent Dutch Social Cultural Plan Bureau (SCP) for instance notes that the latest large inflow of refugees coming from Syria has many difficulties to overcome. They say themselves that learning the Dutch language is hard. Only 12 percent of the large group of Syrians has a paid job, and most of those are in low-skilled work or a temporary employment agreement.9 This study could be an illustration of the findings that stem from the literature.In any case, this line of thought will lead to the following presumption. The second hypothesis will be;

H2: An increase in the year-to-year difference in the share of the people with a non-Western

foreign background within the labour force increases unemployment on a municipality level more than a rise in the same share but with a Western background.

To expand any possible findings of this research, the share of the non-Western population as a share of the total population within a municipality will be included in this research. The theory that non-Westerners have more trouble integrating into the local labour market and therefore might have a stronger negative effect on the unemployment rate than migrants from Western countries could be shown to be more robust when an additional measurement of the non-Western inflows are included in the model. Therefore, the next hypothesis will be a variation of the previous one;

H3: A year-to-year increase in the non-Western population as a share of the total population

within a municipality increases the local unemployment rate.

Additionally, several studies have suggested that there are certain groups within the labour force that are more vulnerable to increased labour competition due to immigration. Blackstone et al. (1998) and Coleman and Rowthorn (2004) write that workers could benefit and suffer from increased competition depending on their subgroups’ characteristics. Winter-Ebmer and Zweimuller (1999) for instance found that effects could be more severe for the workers within

9

(29)

the young age groups. They explain this through the low amount of built-up firm-specific capital that young workers possess, which makes for an easier consideration for the companies they work at. Van der Waal (2012) saw more pronounced effects in the low educated groups, just like Pocher (2011) saw in the UK. This hypothesis could give additional insight into the dynamics of the integration of additional migration flows into the local labour market. Therefore, the final hypothesis will be:

H4: Any higher unemployment effects as a result of immigration are more pronounced in the

share of the labour force with the age of 15 to 25 and in the share of the labour force with a low education background.

These hypotheses will be tested in the empirical analysis chapter. The conceptualization of these hypotheses and the description of the measurements and definitions will follow in the next chapter.