European migration : A study of the migration flows in the European Union

J

I

B

S

JÖNKÖPI NG UNIVER SITY

E u ro p e a n m i g r a t i o n

A study of the migration flows in the European Union

Bachelor Thesis within Economics Authors: Jenny Larsson 831107

Albert Söderlind 820721

Tutors: Hyunjoo Kim, Johan Klaesson, Johanna Palmberg

i

Kandidatuppsats inom nationalekonomi

Titel: European migration

Författare: Jenny Larsson, Albert Söderlind

Handledare: Hyunjoo Kim, Johan Klaesson, Johanna Palmberg

Datum: 2008-01-30

Ämnesord: Migration, EU, utbildning,

____________________________________________________________

Sammanfattning

Syftet med denna uppsats är att analysera vilka faktorer som påverkar migrations-strömmarna mellan de nya (EU10) och gamla (EU15) EU-länderna. Uppsatsen under-söker även hur de faktorer som är signifikanta påverkar migrations in- och utflödena positivt eller negativt. En jämförelse har gjorts för åren 1999 och 2004 för att analysera hur situationen för de nya medlemsstaterna har förändrats till följd av EU-medlemsskapet.

De modeller som används är humankapital modellen och Todaros migrations modell. Humankapital modellen grundas på idén att en individ kan tänka sig att migrera ifall den förväntade inkomstskillnaden överstiger de monetära såväl som icke-monetära flytt-kostnaderna. Todaros modell rör sig om flyttströmmarna mellan landsbyggd och urbani-serade områden. Några variabler som används i regressionerna är valda utifrån dessa modeller och förväntas ha påverkan på en persons val att migrera från sitt hemland till ett annat. För att skilja på de gamla och de nya EU-länderna används en dummy varia-bel; 1 för EU10 och 0 för EU15. Resultaten visar att EU10-länderna 1999 hade lägre in- och utflöden än EU15-länderna. Högre utbildning har negativ relation med inflöde och även negativ relation med utflöde. Faktorer enligt regressionerna som inte var signi-fikanta för migrationsströmmarna in- och utflöden är arbetslöshet och ekonomisk till-växt. En korrelationsmatris har också gjorts för att upptäcka samband mellan de obero-ende variablerna.

Slutsatsen av uppsatsen är att fler än de ekonomiska faktorerna påverkar en persons be-slut att migrera eller inte. Exempel på andra faktorer kan vara familjerelaterade orsaker, kulturella skillnader och språkliga barriärer.

ii

Bachelor thesis within economics

Title: European migration

Authors: Albert Söderlind, Jenny Larsson

Tutors: Hyunjoo Kim, Johan Klaesson, Johanna Palmberg

Date: 2008-01-30

Subject terms: Migration, EU, education

_____________________________________________________________

Abstract

The purpose with this thesis is to analyse what factors influence the migration flows be-tween the new (EU10) and old (EU15) EU-members. The thesis also examines how the factors that are significant affect the migration in- and outflow positive or negative. A comparison has been made for the years 1999 and 2004 in order to analyze how the situation for the new member states has changed as a consequence of the EU-membership.

The models that have been used are: the Human Capital Model and the Todaro migra-tion model. Human Capital Model is based on the idea that an individual can consider to migrate if the expected difference in income surpasses the monetary and the non-monetary cost of movement as well. The Todaro model concerns the migration flows between rural areas and urban areas. Some of the variables used in the regressions are chosen from these models and are expected to influence a person’s choice of migration from their home country to another. To separate the old and the new EU-members, a dummy variable is used; 1 stands for EU10 and 0 stands for EU15. The results show that the EU10 members in year 1999 had lower in- and outflow than the EU15-countries. Higher education has a negative relation with inflows and also a negative re-lation with outflows. The factors according to the regressions that were not significant for the migration flow, in- and outflows are unemployment and economic growth. A correlation matrix has been made to discover relationships between the independent variables.

The conclusion of the thesis is that there are more than the economic factors that influ-ence a person’s decision to migrate or not. Examples of other factors can be family re-lated reasons, cultural differences and language barriers.

iii

Contents

1

Introduction... 1

1.1 Problem... 2

1.2 Background... 2

1.2.1 Why do people migrate?... 2

1.2.2 Migration history in Europe ... 3

1.2.3 Mobility and migration rules in the EU... 4

1.2.4 Labour market situation in the European Union ... 5

2

Earlier studies ... 6

3

Theoretical framework... 8

3.1 Todaro migration model ... 8

3.2 Human capital model... 8

3.2.1 The mobility of labour ... 9

3.2.2 Incentives for moving... 10

4

Methodology and results ... 13

4.1 Methods ... 13

4.2 Empirical analysis... 14

4.3 Analysis ... 15

5

Conclusion ... 19

6

Suggestions for further studies ... 20

References ... 21

Internet sources ... 22

Appendices ... 23

Appendix 1... 23

Appendix 2... 24

Figures and tables

Figure 3-1

Unemployment rates for 1999 and 2004

Figure 3-2

Minimum wages 2005

Table 2-1

Largest immigrant country for the EU-countries

Table 3-1

Percentage increase of labour costs in 2004

Table 4-1

Description of the independent variables in the

em-pirical analysis.

Table 4-2

Inflow/population as dependent variable 1999

iv

Table 4-4

Inflow/population as dependent variable 2004

1

1

Introduction

Migration is in general defined as a country’s in- and outflow of people. But the most common definition is more precise and defines migration as the change of permanent residence. The process of migration has been going on for a long time, and is still going on. Migration can be caused by a pull factor, which means that someone migrates be-cause of some kind of attraction from another location, or by a push factor, which im-plies that someone are forced to migrate of some reason.

The movement between countries is sometimes depending on major incidences and can therefore be more intensive during one period to another. A couple of major international occurrences that have affected the migration flows are the post-war reconstruction with its intensive demand for labour, the end of the colonialism which opened up for migra-tion from many African countries to Europe, and the break down of the iron curtain be-tween Eastern and Western Europe (OECD international migration outlook 2007). The migration trends in Europe have been shifting over time. In the beginning of the 20th century there were only a small number of obstacles. Migration almost stopped during the period between World War I and II. After World War II migration in Europe became more liberal. The trend of liberal migration legislations ended during the oil crisis be-cause of decreasing demand for labour (Gustavsson, Oxelheim, and Wahl, 2006). The creation of the European Community, which by the Maastricht treaty in 1992 be-came the European Union, had one of its main goals of creating the so-called “four free-doms” in the union. The four freedoms are: free movement of labour, capital, goods, and services.

The Lisbon-strategy was created in 2000 and is aiming to make Europe the world’s most competitive economy year 2010. The strategy is a method to reach the creation of a common European internal market (Europeiska gemenskapernas kommission). To ac-complish the creation of the internal market, a lot of reforms have to be made, many of them within the area of labour movements. Around 2004 the debate was intensive about consequences for the social situation in the old member states because eight of the ten new member states where former Eastern Europe countries. The largest Swedish labour union LO argued in an newspaper article (Mårtensson, Andersson, 2004) why the labour migration must be limited in order to decrease the risk of social dumping and a down-wards pressure of the wages (Migration och Arbetsmarknad, p 12, 2006). Another con-flict about the free movements of labour is the planned service directive with the purpose to improve the trade of service in the European Union (TT, November 15, 2006).

According to reports made by the Swedish labour unions, there are tendencies of upcom-ing mass-migration over the borders between the ten countries that joined in 2004 (EU10) and the fifteen countries that were members before 2004 (EU15).

2

1.1

Problem

The European Union has by its legislations concerning free movement of labour opened up the borders between its member states. However, even if the legislations should guar-antee free movement of labour, people are not completely free to move between coun-tries when applying for jobs. The expansion of the European Union to 25 member states in 2004 made it possible for inhabitants in the former Eastern Europe to move between the states. There are a couple of differences between EU15 and EU10 which could lead to a situation where a relative large amount of people from EU10 are expected to migrate to EU15. Differences like an average lower unemployment rate and relatively higher wage rates in the EU15 than in EU10.

There are not only economic issues that matters concerning migration. It could be family related, the costs of movements, and the transition rules. The expectation is that people in EU10 are more attracted to migrate to EU15, than the other way around. The migra-tion flows can affect the social situamigra-tion such as brain-drain for the emigrant’s home countries. The purpose of the thesis is to compare the years 1999 and 2004 concerning the migration flows in Europe, and what variables lie behind the migration flows.

1.2

Background

1.2.1 Why do people migrate?

In the population theory there are three components; birth rates, death rates, and migra-tion. Migration is the most difficult component to explain and measure. The migration component differs from the two other components because it must be explained in physical and social terms, and not in numbers (Jones 1990).

In order for migration to take place, mobility must exist. There are two different mobili-ties, spatial- and social mobility. Spatial mobility contains all sorts of territorial move-ments, while social mobility refers to the socio-economic status only. All types of spatial mobility can not be classified as migration. Movements that can be directly excluded are commuters involving no change in residence and movement of seasonal or temporary workers. These kinds of movements should be concerned as circulation instead of migra-tion. The definition of migration is all kinds of permanent or semi-permanent change in residence.

The geographers Kulldorf, Willis, White and Mueser connected the migration rates to the size, shape, and internal population distribution of the area where the migration takes place. The number of migrants will be lower for a large and compact area but will at the same time have more of internal movers (Jones 1990). OECD categorizes migrants in two groups; permanent migration inflow and temporary migration inflow (OECD inter-national migration outlook 2007).

OECD divides the permanent migration into three main groups: Work related migration associates people that migrate because of work-related reasons. Family related migration which means that a person migrates because of having family members in the country where the person plans to migrate to. According to Jones, work-related and family-related migration can be classified as pull-factors, since they are attractive forces which

3

lie behind the decision of migration. An article published in Göteborgsposten explains the situation in a couple of the new member states (Poland and the Baltic states) as alarming. The countries become drained of qualified labour and a large number of young people leave EU10 to seek the luck in EU15. A lot of these people who emigrates are expected to not return (Wolmesjö, 31 december 2007).

The third group of the permanent migration is because of humanitarian reasons. Some examples of humanitarian reasons are listed by the Swedish migration authority (Migra-tionsverket). The most common is when a fugitive in its home country runs the risk of being persecuted because of race, nationality, religious views, political opinions, gender or sexual disposition, or being a part of a specific group in the society. Jones categorizes humanitarian reasons as push-factors since people are forced to migrate.

The difference between migration caused by push and pull factors are not always that clear. Even if migration takes place because of some kind of attractiveness for creating a better standard of living, it can be a migration because of a push factor that has forced the person out of the previous country. It can be a natural disaster, or some kind of armed conflict. One big example of push factor was when a huge quantity of Europeans migrated to the U.S. because of the starving situation in Europe during the 19th century (Jones 1990).

1.2.2 Migration history in Europe

During the industrial revolution in the 1800s the European society faced a demograph-ical change, people moved from the countryside to the cities. At the same time migration between nations took place. The movement between nations included less obstacles than today. A vast number of Europeans immigrated to USA because of poverty and hunger. Between the years 1860 and 1914 approximately 34 million people emigrated from Europe to the U.S.

Examples on countries that can be classified as traditional European emigration coun-tries are Sweden and Germany. In these councoun-tries the citizenship is founded by the prin-ciple of origin. In countries like USA and Australia where the immigration is large, the citizenship is derived more in political terms since it is quite difficult to tie citizenship to origin for a country that is built up by immigrants.

During World War I, the open borders between the countries in Europe suddenly closed. Between the two World Wars, immigration was seen as something bad by the nations. One reason why this sudden change was: the racist ideas about “better” and “worse” people. This point of view became very common in many European countries. (Gustavs-son, Oxelheim, Wahl 2006).

The period after World War II was characterized by economic growth. Both Sweden and Switzerland had a massive recruitment of foreign labour. When the colonies became independent, the former colonists U.K., France, the Netherlands, and Belgium also ex-perienced a remarkable increase in the migration from their former colonies.

According to OECD international migration outlook 2007, traditional immigration coun-tries in Europe after World War II were Germany and U.K. The only barrier for migra-tion was the iron curtain towards the Eastern European countries as a result of the cold

4

war. A lot of the labour migration, especially to Germany and Austria was classified as guest-migration. Guest-migration implied that people came for temporary work, and the plan was to return to their home country after the work was finished. The immigrants brought a foreign culture into the countries where they immigrated. In some European countries the government accepted that the immigrants had their own culture, and in other countries the government tried to integrate the immigrants into the national culture. Even if migration in Europe became relatively free after World War II, it was still regu-lated because of social and political reasons (Gustavsson, Oxelheim, Wahl 2006).

Between 1950s and 1960s there was a huge demand for foreign labour in most of the Western European countries. Until 1970 most of the migrants in Europe were labour- migrants. After 1970 the labour migration decreased drastically because of economic cri-sis like the oil cricri-sis removed most of the demand for foreign labour. The migration after 1970 have been mostly related to human factors (Södersten, 2004).

The majority of the immigrants who immigrated to Europe after the oil crisis were fugi-tives (Gustavsson, Oxelheim, and Wahl). According to the Geneva Convention from 1951 the government is forced to give a sanctuary to all refugees to fulfil the Geneva Convention criterions. This regulation, together with some other national regulations al-lows reuniting with relatives.

1.2.3 Mobility and migration rules in the EU

All citizens of the European Union have the right to stay in all EU-countries for three months assumed that they have a valid passport or identity card. Family members who do not have any citizenship in the European Union also have the same right, but can be required to hold a visa. Passport or identity-card is the only formality that is required of any EU-citizen; according to EC number (539/2001)

If staying longer than three months in a member state, some requirements have to be ful-filled by the citizen:

• The citizen must be able to prove that he/she has an employment or is running his/her own economic activity.

• The citizen must also have a social insurance and cannot be in addiction of any social aid.

• Students from member states must have a health insurance and also proof of economic stability.

• EU-citizens do not need any residence permit, but registration can be required.

• Family members to EU-citizens without citizenship have to apply for a special residence permit for family members which will last at least five years forward from the date it was issued. (these residence permits will not change because of divorces or other family splits.

Article number 8 from directive 2004/38EC describes what is required from each cate-gory of migrants. All EU-citizens will be issued a permanent residence after living in a member state for five years. The same rules are valid for family members of EU-citizens outside the union, but with a time limit that can be refreshed every tenth year. The

per-5

manent residence permit will be removed if the citizen is away for more than two years from the specific member state.

There are some limitations for the free movement: If citizens from an EU-country or a family member from at third nation disturb the general order concerning security or health, he or she can be deported. Decisions of deporting is sensitive since it affects the regulations of free movement and must therefore be founded with respect to the principle of proportionality and the person’s own behaviour. The misconduct from the individual must be associated with a threat against the society. Economic reasons or earlier convic-tions cannot be a reason for deportation.

The borders of the European Union to the rest of the world are regulated by the council’s ordinance number 562/2006 which is working as a common codex for the whole Schen-gen-area. The outer borders are only open at specific places and on specific opening hours. Travelling documents are required, like certification of the purpose with the trip and visa is required in some cases. The person is required to not be in the register for the Schengen information system (SIS), and not be concerned as a threat against the Schen-gen-area’s general order and security. Border controls between member states can be imposed if it is concerned as necessary for the general order in the member state in ques-tion. A report about the border controls between the member states should be handed in to the European parliament by the commission in 2009 and also eventual proposes how the appearing problems by the borders can be solved (eur-lex)1.

1.2.4 Labour market situation in the European Union

An article from Göteborgsposten describes a situation in Eastern Europe where the agri-cultural areas becomes depopulated as a result of young labour who move into urban ar-eas, and also abroad to EU15 countries (Wolmesjö, 31 December 2007).

The unemployment rates differ to a big extent between the countries in EU, especially between countries in EU15 and the 10 members that joined in 2004. The report made by the commission in 2005-2006 named “The social situation in the EU” and the Eurostat Yearbook 2007 reported some interesting topics.

For 1999:

• Unemployment rates differed a lot between EU10 and EU15. The lowest rate was in Luxembourgh with 2.4%, and the highest rate could be found in Slovakia, 16.4%.

• The unemployment rate had increased with 1.3 % compared to the previous year for EU25.

• The average unemployment rate for EU25 for men was 7.8% and 10.8% for women. • Unemployment for persons under 25 was 18.5% and 7.6% for persons over 25 • The long-term unemployment rate was 4.1%.

6

For 2004:

• After having an increased level of unemployment after 2001 related to 9/11, the unem-ployment rate started to decrease in Europe in 2004.

• Poland and Slovakia had in 2005 unemployment rates around 18%, while at the same time Austria and the Netherlands had unemployment rates below 6%.

• The unemployment rate differs between genders. The unemployment in 2004 was about 10.2% for females in the EU25 while at the same time it was 8.1 % for males. Most of these statistics were consistent for most of the member states of EU25.

• 4 % of the unemployed hade been unemployed for more than 12 months in Denmark, Sweden, Austria, Luxemburg. United Kingdom and Cyprus had long-term unemploy-ment rates around 1.5 % while Poland and Slovakia had long term unemployunemploy-ment rates around 10 %. According to the report, females are more affected by long-term unem-ployment than males. The women represent more than half of the long-term unemploy-ment.

• People between the age of 15 and 24 are more likely to become unemployed than older people. In the Netherlands only 8% of the youths where unemployed, while the youth unemployment rate in Poland was 39%.

Source: (European commission)

2

Earlier studies

The largest labour union in Sweden, LO, made a study (Pettersson, Mårtenson, Bard, 2004) about the changes in the Swedish labour market as a result of the expansion of the union; it was released six months after the expansion was made. This report shows that the immigration to Sweden from the new member states, especially Poland, was doubled when the new member states entered the union. The total immigration from Poland to Sweden was around 800 persons in 2003, and after the first six months of 2004 the num-ber increased to somewhere around 1800 people. The purpose with the report from LO was to analyze what happened with the Swedish labour market after the expansion, not for the whole EU15 (Arbetskraft till salu – LO, 2004). Before the expansion took place, many debaters from the labour unions warned for mass immigration which could result in dumping of wage levels and social tourism, which means that people immigrate only to take part of social benefits from the authorities. An investigation from the Swedish in-stitute for political studies of Europe proved that this statement was wrong, the expected mass immigration has so far been absent (Doyle, Hughes, Wadensjö, 2006).

OECD releases every year a publication named “International migration outlook”. The 2007 edition shows a clear tendency that people from countries close to the immigrant country often is the largest immigrant group. People from former colonies are also usu-ally large immigrant groups. Table 2-1 shows from which countries the largest immi-grant groups come from (OECD international migration outlook 2007).

7

Table 2-1 Largest immigrant country for the EU-countries (Source: OECD).

Country Largest immigrant country

Austria Germany

Belgium France

Czech republic Ukraine

Denmark China Finland Russia France Algeria Germany Poland Hungary Romania Italy Romania Luxembourg Portugal

The Netherlands Germany

Poland Ukraine

Portugal Brazil

Slovakia Czech republic

Spain Romania

Sweden Denmark

United Kingdom Australia

OECD’s report from 2007 indicates that inflow of people from EU10 countries has been larger to the EU15 countries without transition rules than to those with.

A report made by the Swedish institute for European political studies (SIEPS), predicts the incentives for people in EU10 to migrate to EU15 countries. The author Jonas Eriks-son predicts the future migration flows from EU10 to EU15 by analyzing how the earlier migration flows from EU15 countries to Sweden had taken place. Eriksson explains the migration both from a microeconomic perspective, what incentives and factors that af-fects the individuals’ choice to migrate, and also from a macroeconomic perspective, how supply and demand for labour is characterized in different areas.

The results from the SIESP´s report illustrate the migration flows from EU10 as depend-ent on the degree of free movemdepend-ent for labour. Differences in incomes and employmdepend-ent have large effects on the migration flows, but there is no tendency for any mass migra-tion across the borders (Nettomigramigra-tion, 2004).

A World Bank report from 2006 named “Migration and Remittances – Eastern Europe and the former Soviet Union” explains that the breakdown of the communism in the early 1990s was a radical change in the migration trend in Eastern Europe. This report analyzes not only Eastern Europe, but also an area called ECA which includes Eastern Europe and Central Asia. The report uses push-and-pull factors like economic condi-tions, demographic pressures and unemployment as determinants of the migration rates. (Mansoor, Quillin 2006).

The World Bank report tries to forecast the future migration flows by looking at the ear-lier migration patterns from Ireland and Southern Europe to the other Western European countries. Ireland and Southern European countries were emigrant countries that have become immigrant countries because of wage equalization (World bank 2006). The re-port focuses on the wage equalization as a result of the migration. The rere-port clarifies

8

that the outflow from EU10 has increased when the quality of life is improved (Mansoor, Quillin, 2006).

3

Theoretical framework

3.1

Todaro migration model

The Todaro migration model describes a theory about rural-urban migration and has been useful to describe the reasons of the emigration from Europe to the United States. Migration was related to the fact that people moved from rural areas where the standard of living was low, to urban areas where people expected an increase in wealth and a bet-ter life (Todaro, Smith, 2006).

People who have been through the rural-migration process lack the access of perfect in-formation, and therefore expected income rather than real income matters when analyz-ing the migration flows. People who live in rural areas have expectations about gettanalyz-ing another job and achieving a higher income by migrating to an urban area. (Todaro, Smith, 2006).

The Todaro migration model is based on a scenario where the economy is close to full employment. In such an economy, an individual would take the highest paying job inde-pendent of its location. Such migration would therefore lead to reduced differences in wage levels (Todaro, Smith 2006).

3.2

Human capital model

Labour differs in productivity because of differences in knowledge and experience. The human capital model is created as a model for analyzing differences in wage levels de-pending on the employee’s education and working experiences. The term “Human capi-tal” is a general definition for all skills that a labour brings to the workplace. A worker is willing to make an investment in human capital like higher education, only if he or she will receive some kind of compensation for it; mostly in form of a higher wage. This theory determines to a large extent over an individual’s decisions to make investments in tertiary education and trainee education on the workplaces. Other investments in human capital are investments in health care and decisions concerning individual’s possibilities to move. In the human capital model labour is a factor of production that can be traded in the same way as physical capital (Björklund, Edin, Holmlund, Wadensjö, 2006).

According to the model of human capital, education is a way for the individual to im-prove its future opportunities. There are many reasons why someone decides to educate him or herself. The most important reason is probably the differences in wages between labour with tertiary education and labour without tertiary education. The difference of wages must be at a level where it at least compensates the loss of income when getting a higher education. If the demand for high educated labour increases, for example as a re-sult of technological development, the wages are expected to increase. In the short run the supply of labour will not be able to adjust to the increased demand for high-skilled

9

labour. The wage level will for a while be above the equilibrium wage, but in the long run this will adjust towards the equilibrium state. The conclusion of the theory is that the larger the difference in wages between high and low skilled labours, the larger is the de-mand for an individual to get an education (Björklund, Edin, Holmlund, Wadensjö, 2006).

University and college are not the only way to attain a higher education. Internal educa-tions at workplaces also result in higher income levels and the theory of these kinds of educations is almost the same as a university education. The internal education increases only the competence for the specific company while a university education increases the general level of human capital. If the worker switches job, the investment in the internal education will be lost for the company. Both higher education and internal education are knowledge which can be useful in other workplaces, also known as spill-over effects. (Björklund, Edin, Holmlund, Wadensjö, 2006).

A short summarize of this model is that people who become educated expect a higher wage than without an education. Therefore, it must be a wage gap between educated and non-educated people in order for an incentive to get a tertiary education. Shortly, ine-qualities generate more educated people. With open borders within European Union, a European Union-citizen can go to other EU-countries to look for higher wages (Björk-lund, Edin, Holm(Björk-lund, Wadensjö, 2006).

3.2.1 The mobility of labour

The human capital model describes a quite simple way how an incentive to migrate oc-curs. If an individual with the yearly income Is gets the chance to move to another place where he or she has the possibility to receive the income It the individual starts to think about moving. In this moving process, the cost of moving C must be taken into account. This includes for example monetary costs like hiring a transport firm, and also non-monetary costs like the loss of moving from one environment to another, and also family related factors. For making it profitable for the individual to move the increase of in-come switching from one job to another must exceed C. The equation 3-1 shows that the change in income minus the cost of movement is higher than the real income at the cur-rent living place, and therefore the individual will migrate. I1 is the income on the actual

workplace, I2 is the income on the eventual new workplace, r is the discount interest rate,

C is the cost of moving including non-monetary costs, and T is the number of years a person will stay at the new location.

C r I r I _/(1+ ₎< _/(1+ ₎Τ − 2 1 (3-1)

The new income will be larger the longer time the employee chooses to stay at the new job according to equation 3-1. Therefore, it can be more attractive for younger people to move to get a new job than for older people who do not have much time left before re-tirement. Low movement costs will make it easier for people to move, and therefore con-tribute to a more flexible labour market.

Family related reasons can also be a factor that decreases the mobility for labour. If a person has a family, other family members will fall into difficulties after a movement if they have a job at the current place or children in school. Younger people are in general more mobile than older people, since most of them have not yet started a family.

10

Sometimes people decide to move even if C is larger than the future income. It can be emigration from war-zones, in such cases the presented formula does not work. Another problem with this model is the impossibility for people to have perfect information about the labour market situation. Therefore, the mobility of the individual will be managed by expectations of future incomes, rather than facts (Björklund, Edin, Holmlund, Wadensjö 2006).

The matching-model for migration is explained by the unemployment rather than the wage-levels. In an economy with more inflexible wages the unemployment rate will be the most important migration factor instead of the wage levels. Migration will take place from the region with relatively high unemployment rate to a region with relatively low unemployment (Björklund, Edin, Holmlund, Wadensjö 2006).

3.2.2 Incentives for moving

The average unemployment rate for EU25 was 9.1% for both 1999 and 2004 according to the Eurostat Yearbook 2006-2007, but there is a large difference between the unem-ployment rates for each individual country. The figure 3-1 shows the unemunem-ployment rates for all member states in 1999 and 2004. Slovakia and Poland had the highest un-employment rates while the lowest rates were found in United Kingdom, Denmark, and Ireland. Differences can also be found between countries in EU15. In figure 3-1 it is easy to see that countries in Eastern Europe in general have a higher unemployment rate than countries in Western Europe.

0 2 4 6 8 10 12 14 16 18 20 Belgi um Czec h re publ ic Denm ark Germ any Eston ia Gree ce Spain_France Irelan d Italy Cypr us Latv ia Lithu ania Luxe mbo urg Hung ary_Malta Neth erlan ds Austr ia Polan d Portu gal Slov enia Slov akia Finla nd Swed en Unite d Ki ngdo m

Unemp. rate % for 1999 Unemp. rate % for 2004

Figure 3-1 Unemployment rates for 1999 and 2004 (Source: Eurostat)

According to the matching-model in human capital model, unemployment is one main reason to move. Migration will take place more easily if the borders are open for free movement.

As mentioned earlier, another reason to migrate can be differences in wages. Figure 3-2 shows the minimum wage levels for the EU countries in 2004. Observe that the countries which are not represented in the diagram are countries which do not have any regulations

11

of minimum wages by law. These countries regulate their wage levels by collective wage agreements between employer associations and labour unions. Sweden is one example of a country with strong collective agreements (Eurofound). According to the human capi-tal model, people living in countries with lower wage levels than other countries will tend to move if the expected change in the wage level is larger than the cost of the movement. Because of higher wage- and lower unemployment rate in EU15 than in EU10, the expectation when taking the human capital model into consideration, is that citizens from EU10 will move to EU15. Figure 3-2 shows a clear picture that the wage level in EU15 is higher than in EU10. Figure 3-1 shows the EU10 members have in gen-eral an unemployment rate higher than in EU15. These are variables that can affect the migration flows between EU10 and EU15.

0 200 400 600 800 1000 1200 1400 1600 Belg ium Cze ch re publ ic Esto nia Gre ece Spai n Fran ce Irela nd Latv ia Lith uani a Luxe mbo urg Hun gary _Malta Net herla nds Pola nd Portu gal Slov enia Slov akia Uni ted King dom

Figure 3-2 Minimum wages 2005 in Euro/month (Source: Eurostat)

Free movement of labour is a requirement for a borderless European labour market to be possible. It is one of the main ideas about the four freedoms in EU. There cannot be any kind of barriers for people who want to migrate. Today transition rules exist between EU10 and most of the EU15 countries which prevents the free movement. Most of the EU15 countries have transition rules towards the EU10 countries which prevents the citizens in those countries to fully take part of the free movement. The transition rules implies that the countries in EU15 are permitted to make national restrictions concerning the four freedoms against the new members during a period of maximum seven years from the time EU10 entered the union. After seven years these rules will expire. All countries in EU15 except Sweden, Ireland, and United Kingdom have imposed the tran-sition rules (EU-upplysningen).

12

Statistics from Eurostat indicates tendencies of convergence of labour costs are taking place. The labour costs increases over the whole EU but increases in a faster pace for member states in EU10 than for the member states of EU15. The Euro-indicator news shows the result for 2004 that is shown in table 3-1.

Table 3-1 Percentage increase of labour costs in 2004 (Source: Eurostat)

Country Minimum Wage Euro/month

Percentage increase of labor costs in 2004

Belgium 1186 3.4 % Czech republic 210 4.1 % Estonia 159 5.4 % Ireland 605 4.7 % Greece 555 6.7 % Spain 1185 4.8 % France 1128 3.6 % Latvia 122 10.5 % Lithuania 135 6.4 % Luxembourg 1403 3.0 % Hungary 199 8.5 % Malta 546 1.6 % The Netherlands 1265 2.2 % Poland 179 3.0 % Portugal 426 2.4 % Slovenia 469 3.9 % Slovakia 150 7.4 % United Kingdom 1115 7.9 %

Table 3-1 represents the percentage change as compared to the same quarter in the pre-vious year. The percentage change differs between the member states, but is on average higher in EU10 than in EU15. There is also a negative relationship between the wage levels in a member state and the level of wage increase. The increase of labour costs in-creases faster in countries with the lower wage levels than countries with higher wage levels, and vice versa.

13

4

Methodology and results

4.1

Methods

The focus of the thesis is to examine what independent variables affect the migration flows. In order to examine this, two models will be used, both with one dependent vari-able and four independent varivari-ables.

The empirical analysis will show how the dependent variables inflow and outflow of people will be affected by the independent variables. The regressions will show if the in-dependent variables are significant or not for the in-dependent variables.

The independent variables or explanatory variables that are chosen are expected to in one way or another influence the dependent variables. The independent variables that are used in the regressions are growth rate, unemployment, tertiary education, and the dummy variable.

The null hypothesis tested for inflow of population states that there is no relationship be-tween migration inflow and the independent variables. This will be tested against the al-ternative hypothesis that claims that at least one of the independent variables can explain the migration inflow.

The null hypothesis tested for outflow of population states that there is no relationship between migration outflow and the independent variables. The alternative hypothesis claims that at least one of the independent variables can explain the migration outflow. A correlation matrix will be used to discover if there are any significant relationships be-tween two independent variables.

To decrease the probability of heteroscedasticity both the dependent variables and the independent variable tertiary education have been divided by the population. Both the dependent variables and all the independent variables except the unemployment rate are logged.

To test if there is a difference between EU15 and EU10, a dummy variable is used to dis-tinguish the two “blocks”. The dummy variable is set as a 1 for EU10 and a 0 for EU15 in all four regression sets. By making the regressions it is possible to see how significant each independent variable is.

14

4.2

Empirical analysis

The Todaro migration model states that people move to areas where the wage level are the highest. Instead of moving from rural areas to urban areas, people are expected to move from EU10 to EU15 because of the wage differences and the unemployment rates that are represented in figure 3-1 and 3-2. These factors are two of the four independent variables that are used in the models.

According to the human capital model there are other variables that influence the migra-tion decision than only unemployment and wage levels. The human capital model is based on the expected outcomes from the migration. Positive growth rates can lead to expectations about improved incomes. Many people in a low-wage country who have a higher education have high expectations about getting a good job if he or she migrates. The multiple regression models is used for making the empirical analysis is shown in equation 4-1 (Gujarati, fourth edition). In the model there is one dependent variable and four independent variables stated below.

ε

β

β

β

β

β

Table 4-1 Description of the independent variables in the empirical anaysis.

Variables Definition of the variable

Unemployment The percentage rate of the total

la-bour force that is unemployed

ln_tertp Number of people with at least three

years of college/university educa-tion divided by the populaeduca-tion (logged).

Ln_growth the percentage change of gdp when

prices are held constant. (logged)

DEU25 Dummy variable, 1 if it is a EU10

member otherwise 0

15

When outflow is the dependent variable, the function tested is:

ε

β

β

β

β

β

ln_outflowp _{1 2}Unemployment ₃ terp ₄ growth ₅DEU

(4-2)

When inflow is the dependent variable, the function tested is:

ε

β

β

β

β

β

ln_ lowp _{1 2}Unemployment ₃ terp ₄ growth ₅DEU

(4-3) Unemployment should have a positive relationship with outflow, the larger the unem-ployment is, the more people would try to find jobs elsewhere. Unemunem-ployment should have a negative relationship with inflow. If a country has a high unemployment rate, fewer people chose to move in.

Tertiary education should have a positive relationship with inflow. A country with a relatively high education has probably a well developed labour market for people with higher education. Higher education usually means relatively higher wages and this could attract people from abroad. Tertiary education is expected to have a negative relationship on outflow.

Growth rate should have a positive relation with inflow, since a high growth rate indi-cates increasing wages. When the growth rate is high, it has a negative relationship with outflow.

DEU25, (dummy variable) if it turns out to be significant in the regression, it shows a

positive or negative effect on the dependent variables if it is a new member. EU10 is marked as a 1 and EU15 is marked as a 0.

4.3

Analysis

A comparison has been made to see if there exist any differences between the years 1999 and 2004. Unfortunately not all data tested for the 25 countries were found which re-duced the number of observations. For the year 1999 the total amount of observations were 22 and for the year 2004, the total amount of observations were 24. Another prob-lem which can affect the regression is the outliers, especially when having such a small test.

Another thing worth to mention is that the regressions did not show any signs of multi-collinearity because the regressions in the thesis did not have a variance inflation factors (VIF) higher than 10 and did not have a lower tolerance value (TOL) than 0.1. If the the-sis would show the opposite results then there would be a problem with multicollinearity (Gujarati, fourth edition).

For 1999 both DEU25 and tertiary education have a negative relationship with both out-flow and inout-flow. The dummy variable DEU25 indicates that the new EU-members have a lower inflow and outflow of migration than the old EU-members. The negative β-value (-0.007) in 1999 of the inflow can imply that the EU10 countries are less attractive to immigrate to than EU15. The negative β-value (-0.003) of the outflow might be

ex-16

plained by that EU10 at that time had not yet joined the European union and therefore the citizens had limited possibility to migrate.

Table 4-2. Inflow/population as dependent variable 1999

Variable Coefficient S. Error P.value tolerance VIF

Constant _{-0.009 0.01 0.415}

DEU25 _{-0.007 0.003 0.021(*) 0.744 1.345}

Log_tertp _{-0.005 0.002 0.018(*) 0.694 1.441}

Log_growht 0.001 0.002 0.627 0.723 1.384

Unempl _{-0.016 0.04 0.69 0.53 1.887}

R2 _{= 0.553 Adjusted R}2 _{= 0.454 No. of}

observa-tions = 22

**.Regression is significant at 0.1 level (2-tailed) *.Regression is significant at 0.05 level (2-tailed)

Table 4-3. Outflow/population as dependent variable 1999

Variable Coefficient S. Error P.value tolerance VIF

Constant -0.004 0.007 0.629

DEU25 -0.003 0.002 0.085(**) _{0.744 1.345}

Log_tertp -0.003 0.001 0.056(**) _{0.694 1.441}

Log_growth 0.000 0.001 0.731 0.723 1.384

Unempl -0.019 0.029 0.518 _{0.53 1.887}

R2 _{= 0.461 Adjusted R}2 _{= 0.341 No. of}

observa-tions = 22

**.Regression is significant at 0.1 level (2-tailed) *.Regression is significant at 0.05 level (2-tailed)

By looking at the regressions from 2004 it can be seen that the dummy variable in both cases of inflow and outflow are no longer significant. This means that there is no signifi-cant difference between the new and the old member countries for in- and outflow of migration. When the EU10 countries entered the union in 2004, the standard of living in these countries has improved a lot since 1999 (Björklund, 3 May 2006). Wages are also increasing in a much faster pace than in EU15 members. Another possible evidence for some kind of catch up for EU10 can be found by looking at the correlation table. A ten-dency of convergence between EU15 and EU10 is taking place and the migration be-cause of unemployment has decreased. The tendencies of equalization could be a reason why the dummy variable did not turn out to be significant in 2004.

17

Tertiary education is a robust variable since it is significant in every regression for both 1999 and 2004. It is difficult to interpret that the relationship between both the depend-ent variables, inflow (β-value -0.005 for 1999 and β-value -0.008 for 2004) and outflow (β-value -0.003 for 1999 and β-value -0.005 for 2004) are all negative. However the ter-tiary education still remains significant for the regression.

Table 4-4. Inflow/population as dependent variable 2004

Variable Coefficient S. Error P.value tolerance VIF

Constant -0.013 0.013 0.325

DEU25 -0.005 0.004 0.167 0.903 1.107

Log_tertp -0.008 0.003 0.010(*) _{0.912 1.097}

Log_growth 0.001 0.002 0.647 0.920 1.086

Unempl -0.031 0.045 0.509 _{0.886 1.129}

R2 = 0.395 Adjusted R2 = 0.274 No. of

observa-tions = 24

**.Regression is significant at 0.1 level (2-tailed) *.Regression is significant at 0.05 level (2-tailed)

Table 4-5. Outflow/population as dependent variable 2004

Variable Coefficient S. Error P.value tolerance VIF

Constant -0.009 0.009 0.332

DEU25 -0.004 0.002 0.144 _{0.903 1.107}

Log_tertp -0.005 0.002 0.029(*) _{0.912 1.097}

Log_growth 8.13E-005 0.001 0.953 0.920 1.086

Unempl -0.012 0.031 0.703 _{0.886 1.129}

R2 _{= 0.333 Adjusted R}2 _{= 0.199 No. of}

observa-tions = 24

**.Regression is significant at 0.1 level (2-tailed) *.Regression is significant at 0.05 level (2-tailed)

The expected relationship between tertiary education and inflow was positive but ac-cording to the regression the relationship is negative between the two variables. The negative relationship could be explained due to the fact that a country with a high num-ber of well educated people does not have a significant demand for foreign tertiary edu-cated labour. Considering the human capital model, where the expected income must ex-ceed the cost of moving, there will be risk for tertiary educated people who choose to migrate to a country with many high educated people will become unemployed. The cost of moving will exceed the cost of staying and therefore the people will decide not to move. However the results especially for tertiary education are difficult to explain since there is not enough information about the migrating people.

18

No significance was found for the independent variable growth. This could be because of the major differences in the economic situations in EU10 and EU15. Most of the EU10 members are transition economies where the economy is changing from a centrally planned economy to a free market. Even if EU10 economies are increasing in a much faster pace than EU15 it has still not reached the same level, which could be a reason why growth rate do not affect the dependent variables.

By comparing DEU25´s coefficients for the dependent variables inflow and outflow in year 1999, it was found that both dependent variables were affected negative by the dummy variable with a β-value of (-0.007) and (-0.003). However the dummy variable has a larger negative effect on the inflow than the outflow. A possible explanation for this could be that EU10 is less attractive to immigrate to than the rest of the member states.

Tertiary education in year 1999 has a negative effect on both inflow and outflow with a β-value of (-0.005) and (-0.003). Tertiary education has a larger negative effect on the inflow- than the outflow of population which is hard to interpret. The negative effect on the inflow is opposite of the expectations. One explanation for the odd negative value might have to do with the strict European immigration policy at that time.

In year 2004 the tertiary education had still a negative effect on both the dependent vari-ables inflow- and outflow of population with a β-value of (-0.008) and (-0.005). Not much difference from the year 1999. The EU10 became members in the European Un-ion in 2004 which could indicate that it is too early to see any differences between the chosen years.

The Pearson correlation tables in the appendix have shown some interesting results worth to mention. In year 1999, the growth rate affected the unemployment negatively with a β-value of (-0.495). This indicates that when there is a high growth rate, there will be a low unemployment. The relationship between tertiary education and unemployment is a positively correlated with a β -value of (0.434) which is difficult to explain (this could be because of too few observations). The inflow of population has a negative rela-tionship with unemployment with a β-value of (-0.512). This could indicate that when there is a low unemployment rate in a country, people tend move in.

The result of the relation between the outflow of population and the unemployment rate was also negative with a β-value of (0.420) which is awkward. However, when compar-ing the inflow and the outflow of population, inflow affects the unemployment more and is also more significant at the 0.1-level compared to outflow which is significant at the 0.05-level. Growth rate has a positive relation with inflow with a β-value of (0.541); this could indicate that when there is a high growth rate, people tend to move in. The Dummy variable EU25 and the inflow of population are positive related with a β-value of (0.508). This is not easy to explain, since the countries that later became the new EU-members where not at that time even EU-members in the union and also the expectations of the countries were that they are more unattractive to immigrate to than then old member states.

In year 2004 there were not as many significant values compare to year 1999. This could be due to a catch-up made by the new member states. There were however one in-teresting significant correlation between the dummy variable and the growth rate with a

19

β-value of (-0.579). This indicates that the new member states EU10 affect the growth rate for the whole European Union negative and shows that the new member states still differs from the old member states.

Since at least one of the independent variables is significant for both the dependent vari-ables inflow and outflow, both of the null hypotheses for inflow and outflow can be re-jected. The rejection of the null hypotheses indicates that the regressions are significant.

5

Conclusion

Tertiary education is the significant variable which has the largest influence on a per-son’s decision of migration according to the regressions. But there is no clear pattern since the coefficient in both the regression and the correlation gave negative results for both the dependent variables inflow and outflow for both the years 1999 and 2004. It is hard to analyse the result since lack of information of the migrant’s education level. The social situation in most of the EU countries has improved the last couple of years, especially in EU10 in the form of increasing wage levels. When the living conditions improves, the incentives to migrate declines. Since the dummy variable in the regression was significant 1999 but not in 2004, this could indicate convergence, the EU10 mem-bers are catching up.

The results from the regressions and the correlations made it clear that the variables do not have the influence on the migration flows to the extent which first was expected. The transition rules that are imposed for the EU10 members have probably slowed down the expected reactions from the expansion of EU in form of massive migration from EU10 to EU15. One of the variables that were expected to have a large influence in the regression was the unemployment rate. But it did not turn out to be a significant variable. The explanation could be the limited possibilities to migrate in 1999 for EU10 citizens. In 2004 the unemployment rate in the regression did not turn out to be significant either. This could be because of the transition rules implied for the new members. For the growth rate, the size of the GDP per capita is still very different between EU10 and EU15. The EU10 has still a long way to go before reaching the same level as EU15. Therefore it could be a reason why growth rate does not have a significant influence on the migration flows.

The major reasons why the independent variables did not affect the dependent variable as much as expected are bigger influence from personal factors like family relations, friendships, cultural barriers, and language rather than economic factors. Being unem-ployed are not a reason enough to migrate across national borders. Migration caused by unemployment is more likely to take place between national regions than between coun-tries. In most EU countries the social security system are well developed and guarantees a standard of living that is not below the subsistence level. The non-monetary costs are more likely to determine the migration than the monetary. People are not like goods.

20

6

Suggestions for further studies

In the end of 2011, the transition rules that were imposed against EU10 will expire. An interesting topic would be to make a similar study of the migration flows a couple of years after the expiration of the transition rules, when the borders between EU10 and EU15 are open.

The United States is almost as large as Europe, but without the differences in languages and other cultural barriers. A suggestion is to make a study of the migration flows over time in the United States and compare it to the migration flows in the European Union. Migration has taken place between the Western and Eastern parts of Germany after the re-union. It has been a migration in an area with basically the same cultural characteris-tics but with economic differences. A study of the domestic German migration flows would be a highly interesting topic that might explain the influence on people’s deci-sions of migration caused by economic factors and the non-economic factors.

21

References

Björklund Anders, Edin Per-Anders, Holmlund Bertil, Wadensjö Eskil (2006) –

Arbets-marknaden, tredje upplagan, SNS förlag, Tallinna Raamatutrükikoda, Estland

Doyle Nicola, Hughes Gerard, Wadensjö Eskil (2006) – Freedom of movement for

workers from Central and Eastern Europe – Experiences in Ireland and Swede, Swedish

institute for European political studies, Stockholm

Europeiska gemenskapernas kommission (2002)- Översyn av strategin för den inre

marknaden , Brussels 11.04.2002 (KOM2002) 171 slutlig

Eurostat (2007) - Eurostat Yearbook 2006-2007),Office for official publications of the European communities, Luxembourgh,.

Gustavsson Sverker, Oxelheim Lars, Wahl Nils (2006) – En gränslös europeisk

arbets-marknad , Europaperspektiv. Santérus förlag, Talinn

Article from ”dagens nyheter” Alla vannn på EU-utvidgning – Marianne Björklund 2006

Article from “Göteborgsposten” Länderna i öst dräneras på arbetskraft – Wolmesjö Lars Gunnar, 31 december 2007

Articel from ”Göteborgsposten” Landsbygd utarmas i jakt på lyckan – Wolmesjö Lars Gunnar, 31 december 2007

Jones Huw (1990) – Population geography, Paul Chapman Publishing LTD, London,

Mansoor Ali, Quillin Bryce (2006) - Migration and remittance – Eastern Europe and the

former Soviet Union, The International Bank for Reconstruction and Development/The

World Bank, Washington

LO (2006) - Migration och arbetsmarknad, En rapport från LO:s förbundsgemensamma projekt, LO

Mårtensson Christina and Andersson Dan (2004) – LO varnar för fri

arbetskraftsinvand-ring, LO

Mårtensson, Pettersson, Bardh (2004) – Arbetskraft till salu – sex månader med öppna

gränser, LO,

OECD (2007) - international migration outlook 2007

Spidla Vladimir, Almunia Joaquin (2006) - EU social situation 2005-2006, The balance

between generations in an ageing Europé. An analysis made by the European

Comis-sion about the social situation in Europe, Luxembourgh, 2006

Södersten Bo (2004) – Globalization and the welfare state, Palgrave Macmillan Ltd, New York, USA

22

Todaro Michael P., Smith Stephen C. (2006) – Economic development, ninth edition. Pearson education, Harlow

Internet sources

eu-upplysningen www.eu-upplysningen.se

Eur-lex – The EU official newspaper Facts about the EU-legislation

http://eur-lex.europa.eu/sv/index.htm

The European union official homepage http://europa.eu

The European Commission http://ec.europa.eu

Eurostat - http://epp.eurostat.ec.europa.eu

IMF World economic outlook database – www.imf.org

Istat – the Italian statistical office www.istat.it/english

Migrationsverket www.migrationsverket.se

Spanish statistical office www.ine.es

Svenska institutet för europapolitiska studier – www.sieps.se

United nations www.un.org

UNESCO www.unesco.org

The U.N. Geneva convention from 1951 about refugees

23

Appendices

Appendix 1

Glossary

EU 15 – The 15 European countries that were members before the expansion in 2004.

The countries are: France, Italy, Germany, the Netherlands, Belgium, Luxemburgh, Por-tugal, United Kingdom, Ireland, Austria, Denmark, Sweden, Finland, and Greece.

EU 10 – The 10 countries that joined the union in the beginning of 2004. The countries are: Poland, Czeck republic, Slovakia, Hungary, Slovenia, Estonia, Latvia, Lithuania, Cyprus, and Malta.

Lissabon strategy – A strategy in the creation of a European common market with the

main goal that the European Union should be the most competitive economy in the world. The creation of a labour market without borders is one part of this strategy.

SIS – Schengen information system. A part of a European police cooperation which has the function as a common European register for well-known criminals. Persons that are concerned as a threat against the general order, being wanted by any national police au-thority, or suspected for terrorism can be registret in SIS. It is not permitted to register individuals in SIS because of their political reasons, private relationships, or religious belongness (EU-upplysningen).

Brain-drain – The tendency for well educated labour to emigrate to other countries with

higher wages which is a loss in human capital for the emigrant country (Todaro, Smith, 2006 p 807).

Pearson correlation – It is a measure of the correlation between two variable. The

cor-relation measures the degree of linear cor-relationship between two variables. It ranges from +1 to -1. When the correlation is +1 then there is a perfect positive linear relaitonship between variables. When the correlation is -1 then there is a perfect negative linear rela-tionship between variables.

24

Appendix 2

Pearson Correlation for 1999 and 2004

unemployment growth rate tertiary education/pop D25 inflow/pop outflow/pop unemployment 1 growth rate -0.495(*) 1 tertiary education/pop 0.434(*) -0.112 1 D25 -0.351 0.309 0.171 1 inflow/pop -0.512(**) 0.541(**) -0.384 0.508(**) 1 outflow/pop -0.420(*) 0.357 -0.385 0.381 0.907(**) 1

**. Correlation is significant at the 0.1 level (2-tailed) *. Correlation is significant at 0,05 level (2-tailed)

Pearson Correlation for 2004

unemployment growth rate tertiary education/pop D25 inflow/pop outflow/pop unemployment 1 growth rate 0.139 1 tertiary education/pop 0.219 0.331 1 D25 -0.25 -0.579(**) -0.135 1 inflow/pop -0.305 -0.307 -0.459(*) 0.304 1 outflow/pop -0.255 -0.323 -0.383 0.333 0.835(**) 1

**. Correlation is significant at the 0.1 level (2-tailed) *. Correlation is significant at 0,05 level (2-tailed)