University of Amsterdam
Freedom of movement for workers in
the European Union
The economic consequences for low-skilled
Western European workers
Bachelor Thesis Political Science
Author: Daniël Heijtel
Supervisor: Rosa Sanchez Salgado
Second Reader: Julien Jeandesboz
Table of contents
Introduction 3
Theoretical framework 6
Clarification of concepts 6
Literature review 7
The freedom of movement for workers in the European Union 12
Formulation of hypotheses 14 Data introduction 17 Data analysis 21 Conclusion 24 Discussion 27 Bibliography 29 Appendices 31
Introduction
On January 1st 2014 the last restrictions on the freedom of movement for workers for Romanian and Bulgarian citizens have been lifted. After the accession of Romania and Bulgaria to the European Union in January 2007, Romanian and Bulgarian workers were granted the right to work throughout the European Union on the same terms as workers from other member states. However, the old member states of the European Union had the right to impose a transitional period for the freedom of movement for workers for as long as seven years (European Commission, 2014a). Austria,
Belgium, France, Germany, Luxembourg, Malta, the Netherlands and the UK decided to maintain this transitional period as long as possible, but following the Treaty of Accession 2005 these countries had to remove all restrictions for Romanian and Bulgarian workers on January 2014.
This removal of working restrictions has led to a large political debate in many Western European countries. For instance, David Cameron brought this debate to the fore last November by stating that he ‘would take imminent action to stop EU migrants “taking advantage” of the UK welfare system when Romanian and Bulgarian workers will be allowed to come to Britain from 1 January’ (EurActiv, 2013). Cameron launched his critique after rising fears within his party about waves of Romanian and Bulgarian immigrants claiming benefits when restrictions on these
immigrants are lifted. The research agency YouGov reported that half of the British population wants a referendum on the issue and would vote for the employment restrictions on Bulgarians and Romanians to continue past the 1st of January 2014 (YouGov, 2013).
The lifting of working restrictions is also a hot topic in German politics. The German CSU, one of the German coalition partners, released a statement just a few days before the restrictions would be lifted demanding extra measures to limit poverty-driven migration in the EU. The CSU stated among other things that ‘Europe should not block us if we want to have national regulation of poverty-driven migration. It is fatal behavior for Brussels to shut its eyes to this problem and to curb the possibilities for national measures against poverty-driven migration’ (Financial Times, 2014). However, CSUs coalition partner SPD defended the freedom of movement of workers as an essential part of European integration. The German foreign minister and SPD member Steinmeier stated that ‘European freedoms are the core of our idea of Europe, free movement of workers is an
indispensable part of European integration. Germany has derived immense benefit from that, and certainly much more than others’ (Financial Times, 2014).
In the Netherlands, the minister of Social Affairs Asscher stated that he would have limited the influx of Romanian and Bulgarian workers if the European agreements had not forced him to do otherwise. Asscher claimed that there is a problem with the free movement of workers in Europe, because it leads to a downward spiral in working conditions. A large majority of the Tweede Kamer
supported Asscher in his comments (NRC, 2013). In their criticism the Dutch politicians seem to follow their electorate. In a report released by the Sociaal en Cultureel Planbureau (SCP) about the freedom of movement for Romanian and Bulgarian workers it was concluded that almost two third of the Dutch citizens believe that there will be primarily downsides to this policy. Likewise, 58% of the Dutch think that Eastern Europeans are abusing social benefits and 47% believe that Eastern European workers are taking the jobs of Dutch workers. Especially lower educated people have negative attitudes towards Eastern European workers, the SCP stated (de Volkskrant, 2013).
These discussions are just a few examples of the ongoing debate about the freedom of movement for workers in Europe. The opinions about the freedom of movement for workers in general, and the lifting of work restrictions for Romanian and Bulgarian workers in particular are very mixed. One the one hand, there are many good arguments that endorse the claim that lifting these restrictions will have negative consequences for Western European countries. If Romanian and Bulgarian citizens will move to Western European countries on a large scale, this will increase the labour supply in these countries, which could lead to lower wages and more unemployment. On top of that, since the wages in Romania and Bulgaria are relatively low compared to Western European countries, it is likely that Romanian and Bulgarian workers are prepared to perform a lot of heavy work for a relatively low wage. This would put even more pressure on the wages and unemployment in the receiving country. Then there is also much concern about Eastern European workers being exploited. According to the SCP 60% of the Dutch think that Eastern European workers are being exploited in the Netherlands and that the government should act against this exploitation. These people are not only opposed to this exploitation for moral reasons, but also fear that exploitation in the form of salaries below the minimum wage and very long working days will lead to unfair
competition between Eastern European and Dutch workers (de Volkskrant, 2013).
On the other hand, there are also good reasons that a rise in unemployment among native workers in the receiving country will not take place. One of these arguments could be that foreign workers are mostly performing jobs that native workers do not want to do, or that native workers are just not capable of performing these jobs. Then there is also the possibility that Eastern European workers are only easing a shortage of workers in some sectors, and are thus performing jobs which would otherwise remain vacant. This argument was brought to the fore recently by the European commissioner for Employment, Social Affairs and Inclusion László Andor, who claimed that ‘studies have consistently shown the benefits of free movement of workers for the economies of host countries. Mobile workers complement host country workers by helping to address skills gaps and labour shortages’ (Euronews, 2014). Andor added that there are around two million unfilled job vacancies in the EU, according to the European Commission. Moreover, native workers have some
great advantages compared to Eastern European workers. They speak the same language as the employers and they are familiar with the culture and working conditions.
Given all these contradicting arguments, it is understandable that there is such a lively discussion about the freedom of movement for workers. As indicated by the European commissioner László Andor and the German foreign minister Steinmeier, there seems to be a general agreement in the scientific literature that the opening of the borders for other countries will be beneficial for the economy of the receiving country in general. However, in the arguments mentioned before there are good reasons to believe that despite the benefits for the economy as a whole, the freedom of
movement for workers will have negative consequences for low-skilled workers in Western European countries. This could explain the fierce language of many European politicians about this subject and would give rise to some serious questions about the desirability of the freedom of movement for workers in Europe. Hence, it would be very interesting to look at the effects of the freedom of movement for workers in Europe on low-skilled Western European workers. If these effects can be made clearer, this would not only make it easier to deal with the influx of Romanian and Bulgarian workers, but would also allow policy makers to make a thoughtful decision about a possible future expansion of the European Union. This thesis will therefore attempt to give some clarification about this topic by answering the following research question: what are the economic consequences of the freedom of movement for workers in the European Union for low-skilled Western European
workers?
To estimate the effects of the freedom of movement for workers for low-skilled Western European workers, the subsequent part of this thesis is devoted to the clarification of some important concepts. Then the economic theory about the freedom of movement for workers is explained and some important parts of the scientific literature are discussed. Afterwards, the legal framework of the freedom of movement for workers is described. Thereafter the hypotheses concerning the subject are formulated and the quantitative data to test these hypotheses are introduced. Subsequently this data will be analyzed and interpreted in the light of the theories, and finally the conclusions about the freedom of movement for workers are drawn. In the last part of this thesis these conclusions and their limitations are discussed.
Theoretical Framework
Conceptualization
In order to be able to answer the research question, there are a few aspects of this question that require some clarification. Firstly, the research question indicates that only the economic
consequences of the freedom of movement for workers will be taken into consideration. Of course there are a lot of consequences stemming from the freedom of movement for workers without an economic component. For example, it could be that people feel uncomfortable about the culture and language of Eastern European people and therefore do not want them to live in their
neighbourhood. Another often heard non-economic argument against Eastern European workers is that an influx of these workers will lead to more criminality. Although these issues are an interesting and relevant component of the debate about freedom of movement for workers, capturing all the effects of the freedom of movement in this thesis would be extremely hard given the time constraint. Since the economic consequences of the freedom of movement for workers seem to cause the most debate among scholars, this thesis will focus solely on this aspect of the freedom of movement for workers.
Secondly, in order to answer the research question it should be made clear what exactly is meant by Western European low-skilled workers. In this thesis, workers from Belgium, France, Italy, Luxembourg, the Netherlands, Germany, Denmark, Ireland, the United Kingdom, Greece, Portugal, Spain, Finland, Austria, Sweden, Norway and Iceland are considered Western European workers. These are the member states of the European Union before the 2004 expansion plus Norway and Iceland. The reason for treating these countries as Western European countries lies in the
quantitative analysis later in this thesis. The effects of the freedom of movement of workers will be analyzed based on the 2004 expansion of the European Union. This can only be done by making a distinction between pre- and post-2004 members of the European Union. Norway and Iceland are also included since they adopted the freedom of movement for workers policy in the same way the old EU member states did (European Commission, 2014b).
Thirdly, in this thesis workers performing low-paid jobs that require little schooling and unemployed workers who want to perform these jobs are regarded as low-skilled workers. When analyzing the scientific literature, the definition of the respective author is followed. In the quantitative analysis there are multiple indicators for low-skilled workers. These indicators are discussed more extensively in the data introduction, along with the justification for the selection of the different countries.
Literature review
In the scientific literature there is a lot of information to be found on numerous topics that are relevant for the economic consequences of the freedom of movement for workers in the European Union for low-skilled Western European workers. To make all this literature a little more
comprehensible, this part of the thesis will mainly focus on the economic literature about
international labour mobility and studies on the economic effects of immigration on host countries or regions. Despite this relatively narrow scope, the studies that are discussed in the literature review are still just a small selection of the available literature. These studies have been selected by
snowball sampling and selecting on the frequency of references to a certain study in other relevant studies. In this way it has been tried to give a representative overview of the most important literature. Studies on other effects of international labour mobility and studies on European integration could also be relevant for this topic, but are beyond the scope of this thesis.
The macroeconomic consequences of international labour mobility are described in an adapted version of the specific factors model. The specific factors model is a fundamental economic model, developed by Pal Samuelson and Ronald Jones, designed to illustrate the consequences of trade on income distribution (Krugman et al., 2012: 81). In this adapted version, it is assumed that there are only two countries producing a single good using two production factors, namely labour and land. The model further assumes that whenever international migration is possible, workers will want to move from the low-wage to the high-wage country, and that the marginal productivity of labour decreases when the amount of labour increases. Since there is only one good, there is no reason to trade it. However, there will be a ‘trade’ in labour services. In absence of migration, technology differences or differences in the availability of land relative to labour causes the marginal productivity of labour to differ, which will cause wage differences. For instance, a worker in the country with better technology can produce more goods with the same amount of land, and the productivity of a worker increases if he has more land at his disposal. The wage differences caused by the marginal productivity of labour will invoke workers of the low-wage country to move to the high wage country. This movement will reduce the labour force of the low-wage country and thus raise the real wage in that country, while increasing the labour force and reducing the real wage in the high-wage country. This process will continue until the marginal productivity of labour and thus the real wage rates are equalized (Krugman et al., 2012: pp. 99-100). This theory has some important implications:
- International labour mobility leads to a convergence of real wage rates - It increases the world output as a whole
- Those who originally worked in the high-wage countries will receive lower real wages. The second implication, the increase in the world output as a whole, follows from the fact that the increase in productivity in the former high-wage country is higher than the decrease in the
productivity of the former low-wage country. This follows from the shape of the marginal product curve (Krugman et al., 2012: 100). For simplicity I will not elaborate on the mathematical derivation of this result, this would be beyond the scope of this thesis.
Looking at the implications, it becomes clear that according to this theory international labour mobility, while allowing everyone to be made better off in principle, could leave some groups worse off in practice. In this case, the workers in the former high-wage country and the landowners in the former low-wage country are worse off, although the total gains of all people are higher than the total losses. This main result would not change in a more complex model where countries produce and trade different goods, so long as some factors of production are immobile in the short run (Krugman et al., 2012: 100). However, the authors themselves have some reservations about the assumption that all workers will always want to move from a low-wage to a high-wage country whenever international migration is possible. It is quite likely that some workers do not want to leave their friends or family behind for a higher wage. On the other hand, the authors mention that if the wage differentials between countries are high, for many workers this high wage will outweigh personal tastes for particular countries (Krugman et al., 2012: 99).
The model mentioned above does only focus on the short term effects. However, Krugman et al. state that the results of the previous model do not need to hold in the long run, when all factors are mobile across sectors. To prove this point, the authors are employing the Hecksher-Ohlin model. For simplicity, this model assumes that there is a country in an open world economy producing two goods with two means of production (Krugman et al., 2012: pp. 110-111). In this example, the two goods are cloth and food, and the two means of production are labour and capital. The production of cloth requires relatively much labour, while the production of food requires relatively much capital. The amount of labour and capital used in the production of the goods depends on the marginal productivity of these resources and on their prices. If the relative price of cloth to food is given, which is a realistic assumption in a small country trades a lot, the relative price of labour to capital is known as well (Krugman et al., 2012: 118). This relative price, in turn, determines the ratio of labour to capital employed in both the cloth and the food sectors. Now it is interesting to see what happens if the economy’s labour force suddenly grows. A growth in the labour force implies that the
economy’s aggregate labour to capital ratio increases. At given relative good prices, the ratios of labour to capital employed in both sectors remain constant (Krugman et al., 2012: 118). In that case the additional labour hours are employed by changing the allocation of labour and capital across sectors. The labour-capital ratio in the cloth sector is higher than in the food sector, so the economy
can increase the employment of labour to capital by allocating more labour and capital to the
production of cloth, which is labour-intensive. As labour and capital move from the food sector to the cloth sector, the economy produces more cloth and less food (Krugman et al., 2012: 118). In short, this theory predicts that the allocation of production factors between different sectors would change if the labour force grows. However, the wages should remain constant. This result is based on the assumption that both labour and capital is mobile between different sectors. This assumption would only be true in the long run, as it will take some time for workers to acquire new skills and capital can often not be reallocated immediately.
Considering these economic theories, there are good reasons to believe that international labour mobility could lead to lower wages and higher unemployment among low-skilled workers in a high-wage country the short run. However, in the long run the economy should be able to reallocate resources which will bring the wages and unemployment rate back to its original level. In short, these theories suggest that the international labour mobility could have negative economic consequences for low-skilled workers in a high-wage country in the short run, but not in the long run. Nevertheless, the predicted absence of long-term effects does not affect the relevance of the research on this topic. If, for example, EU policy makers think their electorate places much emphasis on its short-term interests, even temporary negative effects could give them legitimate reasons to be opposed to the freedom of movement for workers.
Having outlined the economic theory on international labour mobility, in the following part of the literature review some empirical studies on this subject will be discussed to see if the predicted effects are observed in practice. In their study ‘The Effects of Immigration on the Labor Market Outcomes of Less-skilled Natives’, Altonji and Card use variation in the fraction of immigrants across 120 major cities to measure the effects on immigration on the labour market outcomes of less-skilled natives. For their analysis the authors used information from 1970 and 1980 censuses on labour market outcomes and correlated changes in the immigrant fractions with changes in native outcomes (Altonji and Card, 1991: 201).
In their theoretical framework, Altonji and Card assume that the labour market consists of skilled and unskilled workers and that immigration adds workers to both sectors, with relative additions depending on the nature of immigrant inflows. Furthermore, the model not only allows for supply-side effects associated with the possible crowding out of native workers, but also for demand-side effects associated with the increase in local population (Altonji and Card, 1991: pp.201-202). In discussing the demand side, the authors conclude that the higher the proportion of unskilled labour among the immigrant force is and the lower the proportion of locally consumed output is, the higher the effects on native low-skilled workers are.
In their empirical analysis, the authors find some interesting relations between immigration and labour market effects. The empirical findings indicate a modest degree of competition between immigrants and less-skilled natives. There is little evidence that inflows of immigrants are associated with large or systematic on the employment or unemployment rates of less-skilled natives. The authors state: ‘(…) we find that a 1 percentage point increase in the fraction of immigrants in an SMSA reduces less-skilled native wages by roughly 1.2% (…)’ (Altonji and Card, 1991: 226).
Another finding of the authors concerns the industry distribution of immigrants. Here the authors conclude that a 1% increase in the share of immigrants generates approximately a 1% increase in the supply of labour to industries in which less-skilled natives are employed. Altonji and Card state that there is no indication that immigrants and less-skilled natives are concentrated in particular industries in a manner that would greatly accentuate the labour market competition between them or, on the other hand, substantially reduce the labour market competition between them (Altonji and Card, 1991: 216). Following this result it is likely that immigration has more or less an equal effect on all low-skilled native workers.
Another article that is worth discussing is ‘The Causes and Effects of International Labor Mobility: Evidence from OECD Countries 1980-2005’ by Ortega and Peri. In this article, the authors use a dataset on migration flows and stocks and on immigration laws for 14 OECD destination countries and 74 sending countries for each year over the period 1980-2005. In this article the authors find that already within one year, employment responds to immigration one for one, and capital adjusts so as to maintain the initial capital intensity. As a result, immigration simply increases the size of the economy, with no negative impact on average wages or labor productivity in the short run (one year) or in the long run (five years) (Ortega and Peri, 2009: 28). Since this paper focuses on the effects on wages and labour productivity in general, this does not imply that these effects also hold for low-skilled workers. However, the authors discuss another finding that is worth mentioning in the light of this thesis. The article concludes ‘that during bad economic times in the receiving country, the effects of immigration are less beneficial. Following an immigration flow equal to one percent of the
population, GDP increases only by 0.6%. Even though the stock of capital still responds vigorously to an immigration shock, the economy is unable to employ all of the newly arrived workers. Or, alternatively, it may do so at the cost of some crowding out of native workers’ (Ortega and Peri, 2009: 28). Following this result, it can be concluded that immigration is more likely to cause negative effects on low-skilled workers during bad economic times than during good economic times.
In his article ‘The Economic Consequences of the International Migration of Labor’, Hanson gives an overview of the recent research on labor-market consequences of international migration. Hanson
mentions a recent study by Borjas (2003) using data on the US labour market. According to this study, education-experience cells in which immigrant labor supply growth has been larger had slower wage growth, even after controlling for education or experience specific wage shocks. This evidence is consistent with immigration having depressed wages for low-skilled US workers. However, Hanson notes that the approach used might confound immigration effects with other labour market shocks, like skill-biased technological change (Hanson, 2008: 20).
Overall, there are many studies indicating that there is little or no impact of immigration on regional wages (Hanson, 2008: 23). However, Hanson also mentions a study by Cortes (2008). This study finds that ‘in the 1980s and 1990s US cities with larger inflows of low-skilled immigrants experienced larger reductions in prices for housekeeping, gardening, child care, dry cleaning and other labor-intensive services’. Immigration could have lowered these prices through its effects on wages (Hanson, 2008: 22).
The freedom of movement for workers in the European Union
The freedom of movement for workers is a fundamental principle of the Treaty of Functioning of the European Union and has existed since the foundation of the European Community in 1957. The freedom of movement for workers is laid down in article 45 of the Treaty of Functioning of the European Union (Eurofound, 2011). In the consolidated version of the treaty on the functioning of the European Union, this article reads as follows (Official Journal of the European Union, 2010: pp. 19-20):
Article 45 (ex Article 39 TEC)
1. Freedom of movement for workers shall be secured within the Union.
2. Such freedom of movement shall entail the abolition of any discrimination based on nationality between workers of the Member States as regards employment, remuneration and other conditions of work and employment.
3. It shall entail the right, subject to limitations justified on grounds of public policy, public security or public health:
(a) to accept offers of employment actually made;
(b) to move freely within the territory of Member States for this purpose;
(c) to stay in a Member State for the purpose of employment in accordance with the
provisions governing the employment of nationals of that State laid down by law, regulation or administrative action;
(d) to remain in the territory of a Member State after having been employed in that State, subject to conditions which shall be embodied in regulations to be drawn up by the Commission.
4. The provisions of this Article shall not apply to employment in the public service.
The principle of free movement of workers is further developed by secondary legislation of the European Union and case law of the European Court of Justice. In short, EU citizens are entitled to look for a job in another EU country, work there without needing a work permit, reside there for that purpose, stay there even after employment has finished and enjoy equal treatment with nationals in access to employment, working conditions and all other social and tax advantages. EU nationals may also have certain types of health & social security coverage transferred to the country in which they go to seek work. The freedom of movement for workers also applies in general terms for Iceland, Liechtenstein and Norway, since these countries are member of the European Economic Area (European Commission, 2014b).
As mentioned in the introduction, individual governments of countries that already were member of the EU are allowed to impose a transitional period on the freedom of movement for workers from new EU member countries after an EU enlargement. For the first two years after an enlargement, the national law and policy of an EU member country determines the access to the labour market of that country. Therefore a work permit may still be necessary for workers from the new member countries in order to be allowed to work in another EU country. If an old member state wants these restrictions to last for three more years, it must inform the European Commission before the end of the first two years. After then, an EU member country can extend these restrictions for another two years if it informs the European Commission of serious disturbances in the labour market. All restrictions must end after seven years (European Commission, 2014a).
Formulation of Hypotheses
Having outlined the recent discussion about the freedom of movement for workers in Europe, the academic literature on international labour mobility and the legal framework of the freedom of movement for workers, it is now possible to formulate some hypotheses about the economic consequences of the freedom of movement for workers in Europe for low-skilled Western European workers.
Following the adapted specific factors model, international labour mobility will reduce the labour force of the low-wage country and thus raise the real wage in that country, while increasing the labour force and reducing the real wage in the high-wage country, so long as some factors of production are immobile in the short run. In Western European economies, there are definitely some vital production factors that are immobile in the short run. Land, factories and infrastructure are just a few examples of such an immobile factor. The assumption that workers will always want to move from a low-wage to a high-wage country whenever international migration is possible will not totally hold in Europe, but given the large wage differentials between Western European EU member countries and new EU members, it is quite likely that for many workers this wage difference will outweigh personal tastes for particular countries. Altogether this theory seems very suitable to predict the short-run macroeconomic effects of lifting work restrictions. This leads to the expectation that Western European countries will face a decrease in wages for low-skilled workers after lifting work restrictions for new EU member states in the short run. In this context it is important to note that in many Western European countries, there is a minimum wage which prevents the wages from falling below a certain level. Besides the minimum wage, wages are quite often not very flexible in the short run, as they may have been fixed at a certain level by collective or personal agreements. If the wages cannot fall due to these reasons, we might observe an increase in unemployment instead of a decrease of the wages. This leads to the expectation that, alternatively or complementary to the expectation about decreasing wages, Western European countries will face an increase in
unemployment for low-skilled workers after lifting work restrictions for new EU member states in the short run.
According to the Hecksher-Ohlin model, a sudden increase in the labour force should lead to a change in the allocation of production factors between different sectors. In the case of Western European countries, this would mean that these countries respond to an influx of low-skilled foreign workers by producing more labour-intensive products and less capital-intensive products. In this way all available labour is employed while the wages are held constant. This result is based on the
assumption that both labour and capital are mobile between different sectors. Since this assumption is only true in the long run, the outcome of this model is also only expected to hold in the long run.
Following this model it is expected that Western European countries will not face a decrease in wages for low-skilled workers after lifting work restrictions for new EU member states in the long run.
In their study, Antolji and Card found a modest degree of competition between immigrants and low-skilled native workers. If the lifting of work restrictions by Western European countries has the same effect as low-skilled immigration in American cities, there would be no reason to expect an increase in unemployment for low-skilled workers, but their wages would be expected to fall a bit. However, it is unclear to which extent the results of a study of American cities are transferable to European countries. In most American cities there exists an extensive network of immigrants
speaking the same language as new immigrants, which could make finding employment much easier. Differences between labour-related laws in the US and in European countries could also lead to different effects of immigration on the labour market. It also is important to note that the results of this study are only expected to occur if most of the new workers are less-skilled. In the case of Western European countries, this seems to be a rightful assumption; a recent report by the
Wetenschappelijke Raad voor het Regeringsbeleid (WRR) on Bulgarian and Romanian immigration to the Netherlands found that most labour migrants in Western European countries are relatively low educated (Kremer and Schrijvers, 2014: 6).
According to Ortega’s and Peri’s study, there are no long-term economic effects of the lifting of work restrictions by Western European countries to be expected. Already within one year, an impact on wages and employment level should not be observed anymore. Only in bad economic times, there should be a small negative impact on low-wage Western European workers. However, since these conclusions are partially based on observations outside Western Europe, it is unclear to what extent these results apply to Western European countries.
The articles mentioned by Hanson make contradictory predictions about the lifting of work restrictions in Europe. Although most articles indicate that immigration will have little or no economic effects on low-skilled domestic workers, there are also studies indicating that the wages for low-skilled workers are likely to decrease. However, these predictions are again mostly based on an analysis of the American labour market, which makes it more difficult to make predictions about the European labour market.
Considering all the studies mentioned above, there are good reasons to believe that the freedom of movement for workers will lead to lower wages and higher unemployment for low-skilled Western European workers in the short term. However, there are no reasons to believe that these effects will also be observed in the long term. Additionally, the predicted short term effects are often not observed in similar studies. In order to be able to see what the economic effects of the freedom of
movement for workers on low-skilled Western European workers are, the following hypotheses will be tested in the data analysis:
1. The lifting of work restrictions for workers from new EU member states leads to lower wages for low-skilled Western European workers in the short term.
2. The lifting of work restrictions for workers from new EU member states leads to higher unemployment among low-skilled Western European workers in the short term.
Data introduction
To test the formulated hypotheses about the lifting of work restrictions, in the following part of this thesis there will be an analysis of historical economic data of countries that had opened their borders for foreign low-wage workers. The hypotheses presume an effect of the lifting of work restrictions on wages and employment. However, these factors are influenced by many factors, and therefore a change in these dependent variables can only partially be attributed to the lifting of work
restrictions. To minimize the effects of external factors, the historical events that are to be analyzed should be as similar as possible to the case of the freedom of movement for workers in Europe. It is also important to include as many cases as possible, since this would increase the reliability of the analysis.
Considering these requirements, the 2004 enlargement of the European Union would be an ideal event to analyze the effects of lifting work restrictions for workers from new EU member states. On 1 may 2004, Cyprus, the Czech Republic, Estonia, Hungary, Latvia, Lithuania, Malta, Poland, Slovakia and Slovenia became member states of the European Union. Following the accession of these countries to the European Union, the workers of the new member states were allowed to work throughout the European Union without any restrictions. However, in the Accession Treaty the old and new member states agreed on transitional arrangements regarding the freedom of movement for workers for workers from all new member states except Cyprus and Malta (European
Commission, 2004). The old member states were allowed to impose restrictions on these countries as described in the section ‘The freedom of movement for workers in the European Union’. While some old member states maintained the restrictions as long as possible, others imposed no restrictions at all. In appendix 1 there is an overview attached of the year of the lifting of work restrictions for each country.
The fact that old member states lifted the restrictions on workers from the Czech Republic, Estonia, Hungary, Latvia, Lithuania, Poland, Slovakia and Slovenia simultaneously, gives a unique opportunity to analyze the effects of a large influx of low-skilled workers in Western European countries. Given the sudden increase of potential new low-skilled workers, if there is ever an effect of the lifting of work restrictions on wages and employment to be observed, it should be after the lifting of work restrictions following the 2004 Accession. Another important matter in observing the effects of the lifting of work restrictions is the time span in which the effects are expected to materialize. In a study to the drivers of mobility within the European Union, Benton and Petrovic found that between 2004 and 2008, the number of citizens of new Member States living in the EU-15 increased by more than 1 million. However, their increase happened gradually during these post enlargement years (Benton and Petrovic, 2013: pp. 5-7). Despite this gradual increase in foreign workers, only the
year of the lifting of work restrictions will be taken into account in the analysis. This has been decided for several reasons. First of all, most studies in the theoretical framework indicate that if there is an effect of the lifting of work restrictions to be found, this effect will be the largest immediately after the lifting of work restrictions. The economy is expected to adjust quickly to the increased supply of low-skill workers, which makes an analysis of the short-term effects of the lifting of work restrictions several years after lifting the work restrictions useless. The second reason for looking only at the year of the removal of work restrictions is that the analysis is based on the percentage change in
employment and economic equality compared to the previous year. An analysis of the percentage change in such a variable due to the lifting of work restrictions will not make sense if the restrictions were already lifted in the previous year. Solving this problem by calculating the percentage change over several years would not be desirable, as this would increase the exposure of the analysis to external factors. Analyzing the year of removal of work restrictions does however give rise to another problem; as indicated the new member states joined the EU on 1 may. Since the work restrictions are lifted immediately after this date or two, five or seven years later, the analysis of the year of lifting work restrictions will include four months in which the work restrictions were not yet lifted.
Nevertheless, based on the foregoing arguments the bias in the result of analyzing the year of lifting the work restrictions is considered to be smaller than the bias in the results of analyzing the following years.
The first hypothesis, about the short-term effects of the lifting of work restrictions for workers from new EU member states on wages for low-skilled Western European workers, will be tested by looking at the change in the income distribution in Western European countries in the year of lifting the work restrictions for workers from the Czech Republic, Estonia, Hungary, Latvia,
Lithuania, Poland, Slovakia and Slovenia. Ideally this analysis would be based on the wage development for low-skilled native workers, but this is not feasible due to a lack of reliable data specifically focused on the development of low-skilled wages. Since the income distribution relies heavily on the development of low-skilled wages relative to high-skilled wages, the development of the income distribution will be used to test this hypothesis. The effects of the lifting of work restrictions on the income distribution will be tested in two ways.
First the development of the S80/S20 ratio in the year of lifting the work restrictions will be tested. The S80/S20 ratio is a measure of the inequality of income distribution, calculated as the ratio of total income received by the 20 % of the population with the highest income to that received by the 20 % of the population with the lowest income (EuroStat, 2014a). For each old member state there has been created an overview of the S80/S20 ratio in each year between 2003 and 2012. This has also been done for Norway and Iceland, since these countries, as indicated in the
member states did (European Commission, 2014b). The raw data needed for this overview has been downloaded from the Eurostat database. Then the percentage change in the S80/S20 ratio in the year of the lifting of work restrictions has been calculated by subtracting the value of the S80/S20 ratio in the year before the lifting of work restrictions from the S80/S20 ratio in the year of the lifting of work restrictions, and then dividing this result by the value of the S80/S20 ratio in the year before the lifting of work restrictions. The same procedure has been followed in calculating the average European change in the S80/S20 ratio in each year between 2003 and 2012. Then the European average change in the S80/S20 ratio in the year of the lifting of work restrictions for each particular country has been subtracted from the percentage change in the S80/S20 ratio for each particular country in the year of the lifting of work restrictions of that country. Then these differences between the change in S80/S20 ratio and the average European S80/S20 change for each country in the year of lifting work restrictions are listed as values of one variable. All the tables described are available in appendix 2. If the lifting of work restrictions has no effect on the S80/S20 ratio of a country, the average value of this variable should not significantly differ from 0.
The second way to test the effects of the lifting of work restrictions on the income
distribution focuses on the change in the Gini-coefficient of Western European countries in the year of the lifting of work restrictions. The Gini-coefficient is defined as the relationship of cumulative shares of the population arranged according to the level of equivalised disposable income, to the cumulative share of the equivalised total disposable income received by them (Eurostat, 2014b). The analysis of the Gini-coefficient follows the same procedure as the S80/S20 ratio, leading to a
comparison between the change in the Gini-coefficient of a particular country and the average European Gini-coefficient in the year of lifting work restrictions. Listing the values of the differences for each country then again leads to a variable which should be significantly higher than 0 if the hypothesis is correct. All the relevant tables are attached in appendix 3.
The second hypothesis, about the short-term effects of the lifting of work restrictions for workers from new EU member on employment among low-skilled Western European workers, is tested in three ways. First the effects of the opening of the borders in a particular country are analyzed by looking at the change in total unemployment among citizens of the respective Western European country in the year of the lifting of work restrictions, and comparing this result with the average change in the unemployment rate in the relevant year throughout Europe. This
unemployment rate excludes non-citizens, which makes this variable useful to estimate the effects on the domestic labour force. The drawback of this method is that it looks at the entire labour force, and not specifically at low-skilled workers. The procedure for testing this variable is identical to the procedure used for the S80/S20 ratio and the Gini-coefficient. All the relevant tables are attached in appendix 4.
The second testing method involves a comparison between the change in unemployment among low-educated workers in the year of lifting work restrictions in a particular country and the average change in unemployment among low-educated workers in Europe in that year. In this method, workers who have had pre-primary, primary and lower secondary education are regarded as low-educated workers (Eurostat, 2014c). The procedure for testing again is identical to the procedure used for the S80/S20 ratio, the Gini-coefficient and the total unemployment rate. All the relevant tables are attached in appendix 5.
The last method differs a bit from the previous methods. In this method the change in unemployment among low-educated workers in the year of lifting work restrictions is compared with the total change in unemployment in the respective country that year. This method will thus not make a comparison between similar group in different countries but between different groups in the same country. Apart from that, the exact same method of calculation is used to list the difference in unemployment rate among low-educated workers and the total unemployment rate for all Western European countries in the year of lifting work restrictions. Again, if the hypothesis is correct, the average value of this variable should be significantly higher than 0. All the relevant tables are attached in appendix 6.
As indicated there are multiple methods used for testing one hypothesis. The decision to use more than one method per hypothesis is twofold. Firstly, the methods no not perfectly reflect the estimated effects. If there are different methods employed to capture these effects, the change of finding significant results is higher. Secondly, by using more than one method the probability of drawing false conclusions resulting from external factors can be reduced. However, the drawback of using multiple methods is that there is a change the analysis will give contradictory conclusions.
Data analysis
Having described the variables that are created to test the effects of the removal of work restrictions, these variables can now be tested to see if the predicted effects as stated in the hypotheses are observed in practice. The variables will be tested in SPSS using a one sample t-test. This test has been chosen because the created variables all consist of two different values, which should be equal if the hypotheses prove to be incorrect. Therefore, the mean of these variables should not differ
significantly from 0. Since the one sample t-test is very effective for testing the mean of a variable, this quantitative method is preferred to other quantitative methods. The One-Sample Statistics of the tested variables are as follows:
One-Sample Statistics
N Mean Std. Deviation
Std. Error Mean Domestic Unemployment Compared to Average European Domestic
Unemployment in Year After Lifting Work Restrictions 17 ,465 17,7220 4,2982 Unemployment Rate for Low-Educated Compared to Average European
Unemployment Rate for Low-Educated in Year After Lifting Work
Restrictions 17 -2,071 16,3404 3,9631
Unemployment Rate for Low-Educated Compared to Total Unemployment
Rate in Year After Lifting Work Restrictions 17 2,694 4,5892 1,1130 Change in S80/S20 Rate Compared to Average European S80/S20 Rate in
Year After Lifting Work Restrictions 15 3,9133 9,75587 2,51895 Change in GINI Compared to Average European Change in GINI in Year
After Lifting Work Restrictions 15 1,9267 4,13357 1,06728
In the table above the mean and the standard deviation of each variable can be read. The S80/S20 variable and the Gini variable both include 15 cases instead of 17. This is caused by missing values in the Eurostat database for these variables. Due to these missing values the percentage change in these variables could not be calculated for two countries. These countries have therefore been excluded from this part of the analysis.
The values mentioned above do not indicate if the difference between the observed mean and the expected mean of 0 is significant. The significance of the differences are described in the One-Sample Test Statistics below. In this test, an alpha of 0.05 is used.
One-Sample Test Test Value = 0 t df Sig. (2-tailed) Mean Difference 95% Confidence Interval of the Difference Lower Upper Domestic Unemployment Compared to Average
European Domestic Unemploment in Year After Lifting
Work Restrictions ,108 16 ,915 ,4647 -8,647 9,577 Unemployment Rate for Low-Educated Compared to
Average European Unemployment Rate for
Low-Educated in Year After Lifting Work Restrictions -,522 16 ,609 -2,0706 -10,472 6,331 Unemployment Rate for Low-Educated Compared to
Total Unemployment Rate in Year After Lifting Work
Restrictions 2,421 16 ,028 2,6941 ,335 5,054 Change in S80/S20 Rate Compared to Average
European S80/S20 Rate in Year After Lifting Work
Restrictions 1,554 14 ,143 3,91333 -1,4893 9,3160 Change in GINI Compared to Average European Change
in GINI in Year After Lifting Work Restrictions 1,805 14 ,093 1,92667 -,3624 4,2158
In the table above the significance level of each variable can be read. It is important to note that SPSS calculates the significance level as if this were a two-tailed test. However, the hypotheses only suggest an decrease in wages and an increase in unemployment . Since there is no reason to assume that low-skilled domestic wages will increase or unemployment among low-skilled domestic workers will decrease after the lifting of work restrictions, a one-tailed test would be more appropriate. If the significance levels are divided by two, the 2-tailed significance levels are converted to 1-tailed significance levels. If these values are lower than 0.05, there is enough statistical evidence to assume that the variable is significantly higher than 0.
The variable that compares the change in domestic unemployment with the change in average European unemployment has a positive T-value, but this value is not significant. The significance level of 0.4575 is well above the alpha of 0.05.
The variable that compares the unemployment rate for low-educated with the average European unemployment rate for low-educated has a negative t-value. This is an remarkable result,
as this would indicate that the unemployment rate for low-educated citizens would improve relative to the average European unemployment rate for the low-educated. However, with an significance level of 0.3045 this value is also well above the significance level of 0.05.
The variable that compares the unemployment rate for low-educated with the total unemployment rate within a country has an positive t-value and a significance level of 0.014. This value is lower than the critical value of 0.05, and therefore the mean of this variable is significantly higher than 0.
The variables that compares the change in the S80/S20 rate with the average European change in that rate has a positive t-value and a significance level of 0.0715. This value is slightly higher than the alpha of 0.05, and therefore the mean of this variable does not significantly differ from 0.
The variable that compares the change in the Gini-coefficient with the change in the average European Gini-coefficient has a positive t-value and an significance level of 0.0465. This value is slightly lower than the alpha of 0.05, and therefore the mean of this variable is significantly higher than 0.
Conclusion
Having analyzed all the created variables, it is now possible to draw conclusions about the
hypotheses and the research question. The first hypothesis stated: ‘The lifting of work restrictions for workers from new EU member states leads to lower wages for low-skilled Western European
workers in the short term. ’ This hypothesis has been tested in two ways. First the change in the domestic S80/S20 rate has been compared to the average European change in the S80/S20 rate in the year after the lifting of work restrictions for each individual country. It was concluded that these changes did not differ significantly, and therefore the lifting of work restrictions does not seem to have a negative impact on the S80/S20 ratio. Secondly the change in the Gini-coefficient has been compared to the change in the average European Gini-coefficient in the year after the lifting of work restrictions for each individual country. It was concluded that the change in the increase in the Gini-coefficient was significantly higher, or alternatively the decrease in the Gini-Gini-coefficient was
significantly lower in countries that lifted the work restrictions in a certain year. Therefore the lifting of work restrictions seems to have a negative impact on the Gini-coefficient.
It seems quite remarkable that two variables that are designed to give an answer to the same hypothesis give an different suggestion about the hypothesis. However, looking closer at the
statistics it becomes clear that both variables give more or less the same suggestion about the hypothesis. The T-values of the variables are 1,554 and 1,805 respectively, and the significance levels are 0.0715 and 0.0465 respectively. These are only small differences, but since the alpha is just in between the significance levels both variables give an different suggestion. Here it should be noted that since there are only 15 cases analyzed in these variables, the mean has to differ a lot from 0 to be able to prove the difference is significant. A solution to this problem would be to increase the alpha to 0.10. However, this would increase the probability of a biased result. Since this probability of a biased result already is quite large due to the large amount of external factors influencing the dependent variables, raising the critical value would not be a good solution. Therefore it has to be concluded that the variables give contradicting suggestion about the effects of the lifting of work restrictions for workers from new EU member states on wages for low-skilled Western European workers in the short term. The hypothesis ‘The lifting of work restrictions for workers from new EU member states leads to lower wages for low-skilled Western European workers in the short term’ can thus be proved nor disproved.
The second hypothesis stated: ‘The lifting of work restrictions for workers from new EU member states leads to higher unemployment among low-skilled Western European workers in the short term.’ This hypothesis has been tested in three ways. First the change in domestic
in the year after the lifting of work restrictions for each country. It was concluded that these changes did not differ significantly. This result is not surprising, as this variable measures total domestic unemployment instead of the unemployment among low-educated workers while there is no reason to assume that the lifting of work restrictions leads to more unemployment among high-skilled and average-skilled workers. Secondly the change in the unemployment rate for low-educated has been compared with the change in the average European unemployment rate for low-educated in the year after the lifting of work restrictions for each country. It was concluded that these changes did not differ significantly. Thirdly the change in unemployment rate for low-educated has been compared with the change in total unemployment rate within the respective country in the year after lifting the work restrictions for each country. Here is was concluded that the increase in unemployment among low-educated workers was significantly higher than the increase in general unemployment within the respective country, or alternatively that the decrease in unemployment among low-educated
workers was significantly lower than the decrease in general unemployment within the respective country.
It is again remarkable that these variables give different suggestions about the hypothesis. Although this result seems contradictory, this result can be explained by looking at the components of the different variables. The last variable compares the unemployment rate among low-educated with the general unemployment rate instead of the unemployment rate among low-skilled
throughout Europe. This method ignores the possibility that, independent of the lifting of work restrictions, the inequality increased throughout Europe during the years that the observations were made. Most Western European countries lifted the work restrictions between 2004 and 2007, in times that the economy was doing well. If these prosperous economic times caused a skill-biased economic growth, the relative change in unemployment among low-skilled workers compared to high-skilled workers could falsely be attributed to the lifting of work restrictions. As indicated in the theoretical framework, this methodological mistake may have been made in previous studies (Hanson, 2008: 20). The variable that compares the change in the unemployment rate for low-educated with the change in the average European unemployment rate for low-low-educated does not face this problem and is thus more reliable. Therefore, in answering the hypothesis, the analysis of this variable is followed. Since it was concluded in the analysis of this variable that the
unemployment rate among the low-educated does not significantly differ from the average European unemployment rate among the low-educated, the conclusion is that the hypothesis ‘The lifting of work restrictions for workers from new EU member states leads to higher unemployment among low-skilled Western European workers in the short term’ has been disproved.
Having dealt with the hypotheses, it is now time to focus on the research question. The research question stated: what are the economic consequences of the freedom of movement for
workers in the European Union for low-skilled Western European workers? According to the scientific literature, there are good reasons to believe that the lifting of work restrictions for new EU member states, which follows from the freedom of movement for workers, could lead to more
unemployment and lower wages among low-skilled Western European workers in the short term. Following the scientific literature there are no reasons to believe that these results will hold in the long term. In the analysis is concluded that the freedom of movement for workers does not seem to cause negative short-term effects on the unemployment rate for low-skilled Western European workers. The effects of the freedom of movement for workers on the wages for low-skilled Western European workers prove to be more ambiguous. Further research on this topic would be required to provide clarity about the short-term effects of the freedom of movement for workers on wages of low-skilled Western European workers.
Discussion
In this thesis it has been tried to shed some light on certain economic consequences of the freedom of movement of workers. However, in investigating such an complex and controversial topic in a relatively short time, it is inevitable to make some assumptions and simplifications which could influence the results of this thesis.
A good example of such an simplification is that from the lack of significant differences in employment levels it is concluded that domestic workers are not hindered by low-skilled foreign workers. It could be that domestic workers cannot perform their preferred work anymore, or that they are forced to relocate. These effects are not seen in the unemployment statistics, but they do cause major negative economic effects on domestic workers.
Also it is assumed that Eastern European workers will only come to Western European countries after the lifting of work restrictions. However, in many cases there already are large numbers of Eastern European workers with a work permit employed in Western European countries when the restrictions are lifted. This fact could have led to an underestimation of the effects of the arrival of Eastern European workers on the Western European labour market.
Another factor that could have biased the results of the analysis is the assumption that immigration flows are exogenous to economic conditions in the destination country. Immigrants base the decision to emigrate at least partially on the economic conditions in the country of destination. However, the amount of immigrants from the same country already settled in the destination country also play a large role (Antolji and Card, 1991: 219). Apart from that, the language and geographical distance are also factors to take into account. The assumption that immigration flows are exogenous to economic conditions in the destination country is therefore only partially correct.
Another doubtful assumption in the analysis is that countries that had lifted work restrictions before a certain year are treated as if this does not affect the macroeconomic indicators of that year at all. In other words, all the effects of lifting of work restrictions are expected to materialize within one year. This assumption does clearly not hold in reality, but has to be made in order to make an analysis based on the percentage change of these indicators. This means that the results of the analysis are biased, but it is not clear how large and problematic this bias actually is.
The last matter that has to be discussed is that most observations used in the analysis were made between 2004 and 2007. In these years the economic conditions were more favorable than in 2014. These different economic conditions could lead to different outcomes regarding the effects of the lifting of work restrictions. Therefore one should be cautious in using the conclusions of this
thesis to make predictions about future expansions of the European Union or the effects of the lifting of work restrictions for Romanian and Bulgarian workers.
Bibliography
Altonji, J. Card, D. (1991) The Effects of Immigration on the Labor Market Outcomes of Less-skilled Natives. Published on pages 201-234 in: Abowd, D. Freeman, R. (1991) Immigration, Trade and the Labor Market. National Bureau of Economic Research, University of Chicago Press. Benton, M. Petrovic, M. (2013) How free is free movement? Dynamics and drivers of mobility within
the European Union. Brussels: Migration Policy Institute Europe.
de Volkskrant (2013) Nederlander bang voor komst Oost-Europeanen. 30 December 2013.
http://www.volkskrant.nl/vk/nl/2686/Binnenland/article/detail/3569303/2013/12/30/Ned erlander-bang-voor-komst-Oost-Europeanen.dhtml
EurActiv (2013) Cameron isolated on push for free movement restrictions.
http://www.euractiv.com/socialeurope/commission-alliance-free-movemen-news-531961, last viewed 4 January 2014.
Eurofound (2011) Free movement of workers.
http://www.eurofound.europa.eu/areas/industrialrelations/dictionary/definitions/freemove mentofworkers.htm, last viewed 12 january 2014.
Euronews (2014) Bulgarian and Romanian exodus fails to materialise as EU work restrictions are lifted. 1 January 2014. http://www.euronews.com/2014/01/01/bulgarian-and-romanian-exodus-fails-to-materialise-as-eu-work-restrictions-are-/
European Commission (2014a) Enlargement - transitional provisions.
http://ec.europa.eu/social/main.jsp?catId=466&langId=en, last viewed 12 January 2014. European Commission (2014b) Free Movement - EU nationals.
http://ec.europa.eu/social/main.jsp?langId=en&catId=457, last viewed 12 January 2014. European Commission (2012) Summary table of Member States' policies - Workers from Bulgaria and
Romania. http://ec.europa.eu/social/BlobServlet?docId=119&langId=en, last viewed 13 January.
European Commission (2004) The Transitional Arrangements for the Free Movement of Workers from the New Member States Following Enlargement of the European Union on 1 May 2004. http://ec.europa.eu/social/BlobServlet?docId=144&langId=en , last viewed 30 January. Eurostat (2014a) Glossary: Income Quintile Share Ratio.
http://epp.eurostat.ec.europa.eu/statistics_explained/index.php/Glossary:S80/S20_ratio, last viewed 30 January.
Eurostat (2014b) Dataset Details. Gini coefficient of equivalised disposable income.
http://epp.eurostat.ec.europa.eu/portal/page/portal/product_details/dataset?p_product_co de=TESSI190, last viewed 30 January.
Eurostat (2014c) Glossary: International Standard Qualification of Education (ISCED).
http://epp.eurostat.ec.europa.eu/statistics_explained/index.php/Glossary:International_stan dard_classification_of_education_(ISCED), last viewed 30 January.
Financial Times (2014) Lifting of EU immigration restrictions divides German coalition. http://www.ft.com/intl/cms/s/0/88371c00-73c6-11e3-a0c0
00144feabdc0.html#axzz2pT5oickB, last viewed 4 January 2014.
Hanson, G. (2008) The Economic Consequences of the International Migration of Labor. NBER Working Paper Series, Working Paper 14490.
Kremer, M. en E. Schrijvers (2014) Roemeense en Bulgaarse arbeidsmigratie in betere banen. WRR-Policy Brief, nr. 1. Januari 2014. Den Haag: WRR.
Krugman, P. Obstfeld, M. Melitz, M. (2012) International Economics. Theory & Policy. Pearson Education, Harlow. Ninth edition.
NRC Handelsblad (2013) Asscher: grens voor Roemenen en Bulgaren te vroeg open.
9 December 2013. http://www.nrc.nl/nieuws/2013/12/09/asscher-grens-voor-roemenen-en-bulgaren-te-vroeg-open/
Official Journal of the European Union (2010) Consolidated version of the treaty on the functioning of the European Union.
http://eur-lex.europa.eu/LexUriServ/LexUriServ.do?uri=OJ:C:2010:083:0047:0200:en:PDF Ortega, F. and Peri, G. (2009) The Causes and Effects of International Labor Mobility: Evidence from
OECD Countries 1980-2005. United Nations Development Programme, Human Development Reports, Research Paper 2009/06.
YouGov (2013) British the most worried about Bulgarians and Romanians. 5 September 2013.
http://yougov.co.uk/news/2013/09/05/british-bulgarian-and-romanian-work-restrictions-m/ Last viewed 5 January 2014.
Appendices
Appendix 6: tables on low-educated unemployment compared to average unemployment within the country
