1 Introduction... 2
2 Review of existing migration literature ... 4
2.1 International mobility of people in the CEECs... 4
2.2 Migration flows after previous EU- enlargements... 6
2.3 Conclusion ... 8
3 The migration flow from the CEECs to the Netherlands... 8
3.1 Migration flows after enlargement in the 1980s ... 9
3.2 Comparison of enlargement situations... 10
3.2.1 Income... 11
3.2.2 Unemployment... 12
3.3 Model ... 15
3.4 The decision which country to target by other than economic factors ... 18
3.4.1 Network effects on migration towards the Netherlands ... 18
3.4.2 Knowledge of languages... 19
3.5 Conclusion ... 20
4 Review of existing literature on consequences of immigration... 21
4.1 Consequences of immigration... 21
4.1.1 Labour market consequences... 21
4.1.1.1 Theoretical model ... 21
4.1.1.2 Applicability to the Netherlands ... 22
4.1.2 Public sector effects ... 24
5 The consequences for the Dutch economy ... 26
5.1 Labour market effects ... 27
5.2 Public sector effects ... 27
5.3 Conclusion ... 28
1 Introduction
The expansion of the European Union, almost three years ago, with countries in Eastern and Central Europe worried some of the older members of the EU (the EU-15). These concerns still exist and are mainly about the expectation of a big migration flow of workers from the new EU members to the EU-15. Headlines like: “The Pole’s are coming” are common in newspapers and underline the concerns of people in the richer countries of losing their jobs to the new workers who are willing to work for lower wages. These concerns are reflected by decisions of most “old”-member governments to restrict the access to their labour market. Table 1 shows that almost all EU-15 countries have restricted their labour market. The few countries that did open their labour market restricted, at least partially, the access to welfare by migrants (Boeri and Brückner, 2005).
Table 1: Transitional regulation in the EU-15
Country Access to labour market
Austria Access to labour markets restricted at least for 2 years, quotas for work permits.
Belgium Access to labour markets restricted at least for 2 years.
Denmark General access to labour market, but obligations for work and residence permits. Work permits issued only for 1 year (EU-nationals: 5 years).
Finland Access to labour markets restricted at least for 2 years.
France Access to labour markets restricted at least for 2 years.
Germany Access to labour markets restricted at least for 2 years, prolongation for further 3 years under discussion.
Greece Access to labour markets restricted at least for 2 years.
Ireland General access to labour market, but obligation to register for work and residence permits. Work permits issued first for limited time. Safeguard clause applies.
Italy Access to labour markets restricted at least for 2 years, quotas for work permits.
Luxemburg Access to labour markets restricted at least for 2 years.
Netherlands Access to labour markets restricted at least for 2 years, quotas for work permits.
Portugal Access to labour markets restricted at least for 2 years, quotas for work permits.
Spain Access to labour markets restricted at least for 2 years, bilateral agreement with Poland which permits limited number of Polish nationals to work.
Sweden Community rule for free labour mobility applies.
UK General access to labour market, but obligation to register for work and residence permits. Work permits issued first for limited time. Safeguard clause applies.
The question raises whether these concerns are appropriate and whether a large migration flow can be expected because of the enlargement. In an attempt to answer this question we can learn from previous EU-enlargements, especially the Portugal and Spain enlargement in 1986 and the Greece enlargement in 1981.
This results in the following main question of this paper:
Will the opening of the doors to workers from the east (new EU-countries) lead to a big increase in migration of workers to the Netherlands?
To answer this question with the aid of information of previous enlargements, it is necessary to consider the circumstances under which previous enlargements took place.
The paper is split up in two parts. The first part predicts the migration flow to the Netherlands following the EU enlargement and the second part consists of the consequences of this immigration inflow for the Dutch economy. This second part considers the following question:
What are the consequences of a immigration flow from the Central and Eastern European Countries (CEECs) to the Netherlands for the Dutch society?
We start in the next chapter with a review of the existing literature on the causal factors of migration, and by consequence should be taken into account. In chapter 3 data of the migration flows observed at previous enlargements are described together with the analysis of the economic situations at the time of these enlargements. The output of this analysis is used as input of a model that predicts the migration flow to the Netherlands following the enlargement to the East.
The next subject is whether labour migration, as a consequence of the EU-enlargement,
has a positive or negative influence on the economic situation in the Netherlands. The
existing literature on this subject is described in chapter 4, which allows us to analyze the
effects of the output of the model of chapter 3. This analysis is done in chapter 5, which
answer the research questions.
2 Review of existing migration literature
The focus of the existing literature on economic migration is mostly on economic factors, such as unemployment differentials, wage differentials, economic disparities and differences in GDP per capita. By contrast, migration theories often do not consider EU accession as the major factor that induces migration. In this chapter such theories together with general migration theories are summarized and analyzed.
Section 2.1 deals with the international mobility of residents in the CEECs, which can be used to give an indication of the size of migration flows. Several attempts have been undertaken to predict migration flows following an EU-enlargement in literature. One way is to use earlier enlargements of the EU and to compare the labour market characteristics and the economic situation of a previous enlargement with the situation of an upcoming enlargement. Theories using this approach are described in section 2.2.
2.1 International mobility of people in the CEECs
The mobility of people differs between countries and plays an important role in migration theories. Drinkwater (2003) analyzed the labour mobility of most of the CEECs and concluded that the willingness of residents to move in CEECs is lower than that of residents in the EU-15. The people in the CEECs who are able to understand the German and/or English language are the most willing to move. Furthermore, Drinkwater shows that better educated individuals are the most willing to move. This is an interesting finding because in the public debate in the Netherlands it is often argued that the country is in need of more highly educated employees.
Fidrmuc and Huber (2004) draw more or less to the same conclusion and find by using an econometric analysis that the willingness to migrate in the CEECs is least for family house owners, the less educated, middle income earners, the elderly, as well as persons residing in regions with above-average unemployment rates. These findings give rise to the question what kind of labour immigration the EU-15 countries (and especially the Netherlands) need in terms of educational levels. This question is partly answered in chapter 4.
A Eurobarometer survey on geographic and labour market mobility determined the
unemployed (see figure 1). According to this study, the most important difficulties people expect to face when moving to another EU-country are the lack of language skills, difficulties in finding a job and in finding suitable housing. The most important factors that are encouraging people to move to another EU-country are higher household income, better working conditions and the discovery of a new environment. These factors are again more or less the same as those found by Fidrmuc and Huber (2004) and Drinkwater (2003).
Figure 1: Percentage of people who would be ready to move to another EU country to find a job if they were unemployed, by country.
0% 10% 20% 30% 40% 50% 60%
PL LU SE LT NL LV FI MT SI UK IT FR BE EU-25 ES SK EE DE DK EL PT CY CZ HU IE AT
Source: Eurobarometer 64.1 on geographical and labour market mobility- September 2005
2.2 Migration flows after previous EU- enlargements
Strielkowski and O’Donoghue (2006) consider EU enlargements as a key factor causing migration flows and compare the enlargement situation of 2004 with the situation of Ireland, which was a typical emigration country in the 1970s and 1980s. They show that regions with favorable economic conditions tend to experience high immigration as well as emigration, whereas depressed regions generally display low labour mobility. This implies that a more “mobile” population, measured in terms of regional emigration and immigration responsiveness at home, will have a bigger response in migration to an EU- enlargement. In their study they use a model in which migration is explained by the wage gap and the unemployment gap between the sending and receiving countries. According to this study, enlargement itself may not necessarily play a decisive role in triggering off the migrants from the “new” Member States. The emigrants should also strongly react to the economic changes and be highly mobile within their country. Given the low migration potential (low “mobility”) in the CEECs, Strielkowski and O’Donoghue argue that it is unlikely that the recent trend of very modest migration to the “old” EU will change (Strielkowski and O’Donoghue, 2006).
Bauer and Zimmermann (1999) use the previous enlargements with Greece, Spain and
Portugal as a point of reference to predict migration flows from the CEECs due to the
enlargement of 2004. They use data of migration, population, unemployment rates and
real GDP per capita in the sending and receiving countries of the previous enlargements
to analyze the determinants of migration. These determinants are used to simulate the
potential emigration rates from the CEECs to the EU. The results of these simulations are
reported in table 2.
Table 2:Simulated emigration rates from CEECs to the EU (Bauer and Zimmerman, 1999).
Total
As % of population in sending country.
Restricted Mobility
As % of population in sending country.
Free Mobility
As % of population in sending country.
Poland 1.83 1.29 6.11
Czech Republic 0.46 0.74 0.33
Slovakia 0.41 0.36 0.95
Hungary 1.05 0.94 2.20
Slovenia 0.15 0.22 0.13
Romania* 6.54 4.06 27.73
Bulgaria* 3.16 1.80 15.72
* These countries are not part of the EU-enlargement in 2004. Bulgaria and Romania became members of the EU on the first of January 2007.
The study of Bauer and Zimmerman (1999) shows that the biggest emigration rates are expected from Poland, Bulgaria and Romania which is mainly a result of their relatively high income disadvantage. These estimations should be read with care because several important determinants such as network migration
1and distance are ignored; the study also does not make a distinction between temporary and permanent emigration. In addition the observed income gap between the acceding countries and the EU-countries is now much greater than it was when Greece, Spain, and Portugal joined the EU. The results do show that it is reasonable to expect long-run emigration rates from the CEECs to the EU of between 2-3% of their total population.
As we focus on migration from the CEECs to the Netherlands, the above table raises the question which part of the migration to the “old” EU is expected to go to the Netherlands.
This question is further analyzed in chapter 3.
The theories described above all use labour market properties and income as factors that influence migration flows. In addition, social scientists have been using a modified version of Isaac Newton's Law of Gravitation to predict movement of people between
1 According to the network approach, migration may become a self-perpuating process, because the costs and risk of migration are lowered by social and information networks. Potential migrants who already have relations in the target country face lower monetary and psychological costs.
cities and even continents. This model uses distance as a factor that influences migration flows. Greater distances deter migration because greater distances imply larger migration costs (Borjas, 2000), so the model presumes a negative relation between distance and migration. This could be a good extension of the model used by Bauer and Zimmerman (1999).
2.3 Conclusion
In this chapter an overview of the theories about migration has been given. Several approaches of predicting migration flows have been described. Theories based on earlier enlargement situations come to the conclusion that migration flows to the “old” EU members will be marginally. However, these conclusions must be taken with care because there are several important factors that are not taken into account. Furthermore, the existing literature on EU-enlargements and migration have made clear that data of GDP per capita, unemployment, population, distance and previous migration flows can be used to predict future migration flows as a result of upcoming enlargements. These factors will be analyzed in the next chapter and then used in a model to predict the migration flow to the Netherlands.
3 The migration flow from the CEECs to the Netherlands
This chapter gives an indication of the expected migration flow from the CEECs to the Netherlands following the 2004 EU-enlargement. This “prediction” is based on data of previous enlargements in the 1980s. At first the factors that influence the size of migration flows are described and analyzed. In section 3.1 the migration inflows to the Netherlands from Spain, Greece and Portugal at the time they entered the EU are analyzed. In section 3.2 the circumstances of the accession of countries in 2004 is compared with those of enlargements in the 1980s and in section 3.3 an econometric model is developed to simulate the immigration from the CEECs to the Netherlands.
Finally, section 3.4 discusses some factors that are omitted from the econometric model,
but may be of influence on the migration flow.
3.1 Migration flows after enlargement in the 1980s
Greece entered the EU in 1981, while Spain and Portugal entered in 1986. The EU labour market opened for Greeks in 1988 and for residents of Portugal and Spain in 1992. In 1994 about 1.2% of the Spanish, 4.1% of the Greek and 8.5% of the Portuguese population lived in EU-countries other than their home country. Most of these migrants lived in Germany and France. The foreign population in the “old-EU” since the opening of the labour markets shows some changes. 135.200 Greek (1.3% of the population) moved to the old-EU, and the number of Portuguese and Spanish in EU-countries other than their home country reduced by 108.200 and 99.800 (1.1% and 0.3%) respectively (Alecke et al., 2000). So at first the Spanish and Portuguese population in the EU increased and after a while decreased even more than it had increased, possibly due to the economic growth in their home countries.
If we look at the data of immigrants from these countries to the Netherlands in Figure 2, we see an increase of immigration flows in the year that the labour market opened (1992 for Spain and Portugal and 1986 for Greece). Directly after the opening the number of immigrants from Spain increased with 27%, from Portugal with 32% and from Greece with 24%. For Spain and Portugal this was a temporary peak, but for Greece the number of immigrants increased further after the year of the opening of the labour market. The number of immigrants at the peak is still much lower than the number of immigrants from these countries in the 1970s and is just a small fraction of the whole Dutch population.
In sum, we can see a temporary peak of the immigration flow following the opening of
the EU-labour market. Still this temporary peak is not having a big impact on the Dutch
economy because of its relatively small size. The temporary peak is followed by an
outflow of migrants several years after the enlargement.
3.2 Comparison of enlargement situations
The circumstances at which the 2004 enlargement took place are of course different from the circumstances at the time of previous enlargements. In order to get a good prediction of the migration flows after the 2004 enlargement, it is necessary to analyze the situation of earlier enlargements and to compare those with the 2004 enlargement.
Figure 2: Immigrants into the Netherlands around the years of the opening of the labour market (1992 for Spain and Portugal and 1986 for Greece).
Immigrants into the Netherlands
0 200 400 600 800 1000 1200 1400 1600 1800
1975 1977
1979 1981
1983 1985
1987 1989
1991 1993
1995 1997
Spanish Portuguese
Immigrants into the Netherlands
0 100 200 300 400 500 600 700 800
1975 1977
1979 1981
1983 1985
1987 1989
1991 1993
1995 1997
Greek
Source: CBS
The 2004 enlargement is much larger in terms of population size. At the time of the
“Greek, Spanish and Portuguese enlargement” the EU-population increased with 22%.
The enlargement of 2004 means an increase of the EU-population with 29%. This larger population increase may have a bigger impact on the economies of the existing EU. Other factors that may influence these economies are described and analyzed below.
3.2.1 Income
In Figure 3 GDP per capita for the Netherlands and the acceding countries is shown, at the time that Greece, Spain and Portugal entered the EU as well as at the time of the 2004 EU-enlargement. As comes forward from this figure, the difference in GDP per capita between the acceding countries and the Netherlands in 2004 is (with some exceptions) evidently larger than at the time of the enlargements in the 1980s. This larger difference implies a stronger encouragement for people to migrate and therefore a greater migration flow. The fact that at a particular point in time GDP per capita is significantly lower does not always lead to a migration flow. The growth of this number should also be taken into account. However, GDP per capita grows much faster in most of the CEECs than in the Netherlands in the last few years, as is shown in Table 3.
Table 3: GDP growth comparison between acceding countries and the Netherlands.
Acceding country:
GDP growth acceding country:
GDP-growth of the Netherlands in the year of accession of Greece, Spain, Portugal and in 2004.
Greece (1988) 4.3% 1.9%
Spain (1992) 0.7% 1.3%
Portugal (1992) 1.1% 1.3%
Average CEECs (2004) 5.0% 1.1%
Source: "The Conference Board and Groningen Growth and Development Centre, Total Economy Database, May 2006, http://www.ggdc.net" , own calculations and OECD Factbook 2006:
Economic, Environmental and Social Statistics.
The expected growth of GDP per capita in the coming years may have consequences for
expected migration. If convergence between the CEECs and the Netherlands occurs more
rapidly than expected, those who did not migrate will profit, since they managed to avoid
direct migration costs. On the other hand, if convergence is slower than expected, the individual is still free to migrate later. This induces a certain hesitance to migrate, which can also be motivated by a preference for short term gains (Alecke et al., 2000).
In sum, the bigger income gap indicates a bigger migration flow to the Netherlands from the CEECs than at the time that Greece, Spain and Portugal entered the EU. By contrast the higher growth rate of GDP per capita for the CEECs in 2004, compared with that of Greece, Spain and Portugal when they entered the EU, (partly) compensates the positive migration effect of the income gap.
3.2.2 Unemployment
As became clear from the last section the income gap between the Netherlands and the acceding countries at the 2004 enlargement is on average larger than those at the time of the accession of Greece, Spain and Portugal. This section deals with another factor that could influence the immigration flow, which is the unemployment gap between the sending countries and the Netherlands. In Figure 2 these differences are shown graphically. Portugal had a lower unemployment rate than the Netherlands at the time of their access to the EU-labour market
2. This situation is not comparable with the situation of any acceding countries at the 2004 enlargement. The unemployment gap between Spain and the Netherlands in 1992 is comparable with the unemployment gap between the Netherlands and Poland and the Slovak Republic. The gap between Greece and the Netherlands at time of the accession of Greece is comparable with the gap between Hungary and the Netherlands at the time of the 2004 enlargement.
The on average larger unemployment gap in 2004 than at the time of the accession of Greece, Spain and Portugal intuitively indicates a larger migration flow as a result of the 2004 enlargement. On the other hand, Fidrmuc and Huber (2004) found, as mentioned in section 2.1, that people living in regions with above average unemployment rates are the least willing to move to another region.
2
Figure 3: Differences in GDP per capita between the Netherlands and other countries when the EU labour market for these countries opened.
G D P p e r c a p i t a i n U S $ 2 0 0 5 a t t h e t i m e o f t h e o p e n e i n g o f t h e E U l a b o u r m a r k e t f o r G r e e k s ( 1 9 8 8 )
0 5 0 0 0 1 0 0 0 0 1 5 0 0 0 2 0 0 0 0 2 5 0 0 0
N e t h e rla n d s G re e c e
G D P p e r c a p i t a i n U S $ 2 0 0 5 a t t h e t i m e o f t h e o p e n e i n g o f t h e E U l a b o u r m a r k e t f o r S p a n i s h a n d P o r t u g u e s e l a b o u r ( 1 9 9 2 )
0 5 0 0 0 1 0 0 0 0 1 5 0 0 0 2 0 0 0 0 2 5 0 0 0 3 0 0 0 0
N e t h e rl a n d s P o r t u g a l S p a i n
G D P p e r c a p it a a t t im e o f r e c e n t E U e n la r g e m e n t ( 2 0 0 4 )
0 5 0 0 0 1 0 0 0 0 1 5 0 0 0 2 0 0 0 0 2 5 0 0 0 3 0 0 0 0 3 5 0 0 0 N e th e r la n d s
B u lg a r i a * * C z e c h R e p u b li c E s to n i a H u n g a r y L a tv i a L i th u a n i a P o la n d R o m a n i a * * S lo v a k R e p u b li c S lo v e n i a A v e r a g e C E E C s *
Source: "The Conference Board and Groningen Growth and Development Centre, Total Economy Database, May 2006, http://www.ggdc.net" , own calculations.
*The CEECs consist of Slovenia, the Slovak Republic, Poland, Lithuania, Latvia, Hungary, Estonia and the Czech Republic which are the CEEC countries that entered the EU in May 2004. The value is calculated with the population of the countries taken into account.
** Will join the EU the 1st of January 2007.
Figure 2: Difference in unemployment gaps between the Netherlands and accession countries at the time of their accession to the European Union.
Unemployment rates at the time of the opening of the EU labour market for Greek labour (1988).
0,00 5,00 10,00 15,00 20,00 25,00
Netherlands Greece
Unemployment rates at the time of the opening of the EU labour market for Spanish and Portuguese labour (1992).
0,00 5,00 10,00 15,00 20,00 25,00
Netherlands Portugal Spain
Unemployme nt rates 2004
0,00 5,00 10,00 15,00 20,00 25,00
Netherlands Czech Republic Hungary Poland Slovak Republic
Source: www.imf.org and www.oecd.org
3.3 Model
In the last sections several statistics, that may indicate a larger or smaller immigration flow from the CEECs, have been discussed. In this section these factors are used in order to get a concrete number of the expected migration flow.
By using an econometric analysis of the determinants of earlier enlargements, it is possible to evaluate the potential migration flows from Eastern Europe to the Netherlands after the 2004 EU enlargement. This econometric analysis uses the, in migration studies, widely used gravity model. The gravity model is essentially derived from the Newtonian Gravitation theory (Tat Man, 1981). Sir Isaac Newton’s Law of Universal Gravitation (1687) states that two bodies attract each other in proportion to the product of their masses and inversely to the square of their distance. This is formulated in the formula:
( )
2 2 , 1
2 1
D M G M F
=
where F is the force of gravity each mass pulls each other, G is a universal constant known as the pull of gravity, M
1and M
2refer to the masses and D
1,2refers to the in- between distance. When this model is applied to migration studies F is identified as the movement of population between regions (M
i,j), G has no direct analogy and isrepresented by a regional variable multiplier K, M
1and M
2are identified as destination attracting attributes like population size (P
iand P
j) and distance (D1,2) as the distancebetween regions (D
i,j). This results in the formula:
β
α λ
j i
j i j
i D
P K P M
,
,
=
with the use of the explanatory exponents λ, α and β. In log linear form, it gives:
j i j
i
ij K P P D
M
ln ln ln ln
,ln = + λ + α − β
The log linear formula is suitable for regression analysis. This technique allows the simultaneous input of a great number of variables.
In this paper we use a model, which is based on the gravity model and is brought forward
by Alecke et al. (2000). To this model we add distance as a factor that influences
migration:
(
rt)
st rt st t s,tt GDP GDP U U DIS
Population n Immigratio
ε β
β β
β β
β + + + + + +
=
0 1log
, 2log(
,)
3(
,)
4(
,)
5log( )
where s refers to the sending country (i.e. Greece, Spain, and Portugal), r to the respective receiving country (i.e. the Netherlands), and t refers to the year. U refers to the unemployment rate, GDP refers to the EKS
3Gross Domestic Product per capita and DIS refers to the distance from sending country to the Netherlands.
The equation is estimated using data of migration from Spain, Greece and Portugal to the Netherlands in the period 1985-1997. In Table 4 the results of this estimation are shown.
Table 4: Statistical properties of the model
Variable Coefficient Std. Error t-value P-value
β
0 0.0955 0.491 0.194 0.847β
1 0.0735 0.0561 1.31 0.200β
2 -0.0648 0.0412 -1.57 0.126β
3 -0.528 0.365 -1.44 0.159β
4 -0.0274 0.0754 -3.64 0.0010β
5 -0.0129 0.0237 -0.544 0.590R-squared 0.797 F-statistic 23.62
Log likelihood 111.3 Prob(F-statistic) 0.000000 Durbin-Watson stat 2.19
If we take a look at the estimated β-values, we see that the signs of β
1and β
2corroborate to the theory that a higher GDP per capita gap results in more migration. β
3corroborates to the theory that a lower unemployment rate in the Netherlands attracts more immigrants. β
4, on the other hand, shows a counter intuitive sign, because a smaller unemployment in the sending country generates a higher migration flow. On the other hand, the sign does correspond to Fidrmuc and Huber (2004), who find residents living in regions with lower unemployment rates show a higher willingness to move than residents living in regions with higher unemployment. β
5is in line with the theory that a greater distance results in less migration.
The t-values of β
1,β
2,β
3and β
5show that the coefficients of GDP
r,t, GDP
s,t, U
r,tand
GDPs,tare not significant. For β
5this is probably due to the fact that only three countries
3 Total GDP, in millions of 2005 US$ (converted to 2005 price level with updated 2002 EKS PPPs). The EKS PPP for a pair of countries implicitly assigns weights by giving most weight to products that are
are used in the sample and that the distances between the Netherlands and Spain, Portugal and Greece do not change over time and are more or less the same.
The estimated equation is used to calculate the expected yearly migration in the first 3 years after the EU-enlargement to the Netherlands with unemployment, GDP per capita gaps and distance as input. In Table 5 the expected migration flows from several CEECs are shown.
The table shows that the Netherlands can expect the biggest yearly migration flow in percentages from Hungary and the smallest migration flow from Poland. In absolute numbers the biggest migration flow does come from Poland, which is due to its population size.
If we use the average of the above estimated percentages and apply this to all countries that acceded the EU in 2004, the Netherlands can expect a yearly inflow of approximately 42.000 immigrants. The reader should note that re-migration of people from the Netherlands back to their “mother country” has not been taken into account.
This remigration is signaled after previous enlargements and can be even higher than the immigration flows such as the case of Spain a few years after the accession. This indicates that the net inflow of migrants will be significantly smaller.
The results that are presented in Table 5 should be read with care because several important factors are not taken into account in the model. For instance, every EU country is ‘competing’ with another EU country for immigrants with their unemployment rate and their GDP per capita. For instance: the Netherlands had a higher (7.4%) unemployment rate than (West-) Germany (6.3%) in 1988, while in 2002 this was the other way around (3.7% for the Netherlands and 9.1% for Germany). Other factors that are not taken into
Table 5: Simulated yearly immigration rates from CEECs to the Netherlands based on Econometric results.
Percentage of population
# Persons Czech Republic 0.066% 6,711
Hungary 0.117% 7,223
Poland 0.045% 17,205
Slovak Republic 0.044% 2,387
Total 33,526
account are network migration and cultural similarities between sending and receiving countries. These factors are discussed in the following sections.
3.4 The decision which country to target by other than economic factors
Whether a potential migrant decides to move to country A or B depends on several factors. It depends on factors like GDP per capita, unemployment rate and physical distance, as used in the model. Besides these factors the ‘social distance’ from a home- country could influence the migration decision. It may be possible that a Polish migrant prefers to move to Germany instead of the Netherlands because of a better knowledge of the German language. Also a Polish immigrant may prefer to move to Germany because he/she already has relations in Germany (the so called network migration). The more people with the same cultural background and the same language in a receiving country, the more attractive this region is for a potential migrant. A large social network not only makes a migrant more comfortable in a foreign country, it also facilitates the migrants’
integration into the labour market and makes it easier to access government institutions (Thum, 2000). These network effects for the Netherlands are discussed in section 3.4.1.
The migration influence of language-knowledge of potential migrants is discussed in section 3.4.2.
3.4.1 Network effects on migration towards the Netherlands
In 1997 officially 950.000 persons from the CEECs lived in the European Union. This
amounted to a share of 0,2 % of the total population. Migration, however, was unevenly
spread across European Union Countries: 527.000 CEE citizens lived in Germany and
103.000 in Austria. In population shares, Austria was the most strongly affected country
and Germany the second (Brückner, 2002). In Figure 4 the share of foreign population
from CEECs in Western-European are shown. The figure shows that there is a relatively
low share of migrants from CEECs in the Netherlands compared to other EU-countries
like Germany and Austria. The theory of network effects suggests that the more people
from a country have settled in a certain region, the more attractive this region is for
potential new migrants from the same country. This path-dependent character of
migration together with the low share of foreign population from CEECs in the Netherlands indicate that only a small part of the emigrants from the CEECs will move to the Netherlands.
3.4.2 Knowledge of languages
If a potential migrant already has some knowledge of the language of a possible target country he might prefer this country above another country of which he/she has no knowledge. This means that the knowledge of languages of the inhabitants in a sending country can give an indication of the countries which a migrant from this sending country targets. In Poland, with 38 million inhabitants by far the biggest acceding country, English is with 55% the most spoken foreign language, followed by Russian (17%),
Figure 4: Share of foreign population from Central and Eastern European countries in Western European countries, 1993 (in % of total population). (Bauer and Zimmerman, 1999)
German (15%), French 10% and Spanish 5%
4. The “Western” languages are more spoken by the younger population since they learn these languages at school. The older population who went to school during the communistic period learned mostly Russian as a foreign language.
If we solely look at these facts this would mean that most of the migrants would choose to move to English speaking EU members and Germany and not to the Netherlands. On the other hand, the knowledge of English can be sufficient for working in several sectors of the Dutch labour market, a fact that may be taken into account by potential migrants.
Altogether the “knowledge of languages”-factor shows that the Netherlands are at most as attractive, but most probably less attractive than other EU-members. This indicates that it is a “braking” factor for migration from CEECs to the Netherlands. Also the knowledge of the Western language, which is more developed for the younger people indicates that more young people than older people will migrate.
3.5 Conclusion
In this chapter the income and unemployment gaps between the Netherlands and the accession countries in the 1980s and in 2004 have been analyzed. The results show that the gaps in 2004 are larger as a result of which we can expect a larger migration flow from the East than the immigration from the South of Europe in the 1980s and 1990s.
Our econometric analysis indicates that approximately 42.000 people will migrate from the CEECs to the Netherlands per year, which represents less than 0.1% of the CEEC- population. The re-migration of the immigrants has not been taken into account.
Apart from the above mentioned economic factors, social factors can have an important impact on the migration decision of the CEEC population. If these factors are taken into account the Netherlands is less attractive than other EU-countries because of historical migration flows and the language. The consequences of the predicted inflow is analyzed in the next chapters.
4 Review of existing literature on consequences of immigration
In the last chapter the migration inflow from the CEECs to the Netherlands has been predicted and in this chapter the consequences of this inflow are analyzed. The chapter starts with an overview of the existing literature on the consequences of migration. Next, the generated migration inflow of chapter 3 is used as input to illustrate the consequences of the immigration for the Netherlands.
4.1 Consequences of immigration
In this section theories which can throw more light on the consequences of an inflow of immigrants from the CEECs to the Netherlands are described. These consequences can be separated in “labour market consequences”, such as employment, wages et cetera, and in “public sector consequences”. In section 4.1.1 the labour market consequences are described and in section 4.1.2 the public sector consequences.
4.1.1 Labour market consequences
The labour market effects considered in this paper are those with respect to the wage rate, employment effects and the immigration surplus. These impacts are analyzed using a theoretical model of Borjas (1999) which is first discussed and then applied to the situation of the Netherlands and the CEECs.
4.1.1.1 Theoretical model
In theory, immigration can be beneficial to a country because the national product goes
up. On the other hand, important distribution effects may occur that can affect the
situation of certain groups in a negative way. This is illustrated by a model of Borjas
(1999). He assumes that the labour force L consists of N native employees and M
migrants. Further, he assumes that migrants and native employees are perfect substitutes,
thatnatives are the only owners of capital, that production has constant returns to scale
and that the labour market is flexible.
The effect of immigration is shown in Figure 5, where S is the (inelastic) labour supply curve and f
Lthe labour demand curve. At the point of departure (without immigration), the wage rate is equal to w
0and employment to N. The flow of migrants causes a new equilibrium with a lower wage rate w
1and a higher level of employment (M+N). Native employees receive a lower wage as a result of immigration.
The immigration also affects national income. At the point of departure national income is equal to the surface under the demand curve (ABN0). In the new situation the national income increases to ACL0. Part of the national income increase is represented by the income of the immigrants (DCLN). The other part, triangle BCD, is the so called immigration surplus and is distributed to the native population. In other words, as a consequence of immigration the native population benefits in the sense of a higher average income. On the other hand, important redistribution effects arise: native employees lose income as big as rectangle w
0BDw
1, which goes to (native) capital owners, who pay lower wages.
4.1.1.2 Applicability to the Netherlands
The model as described above assumes a flexible labour market, which is not completely
Figure 5: The immigration surplus in a model with homogeneous labour and fixed capital (Bosman, 2003)
exist which cause a sticky wage rate (Bosman, 2003). In case of sticky wages immigration will lead to unemployment of size M, the people who become unemployed can consist of only the immigrants, but it is also possible that some native employees will lose there job. Also in this situation there is no creation of an immigration surplus.
If there exists a shortage of employees, for example, as a consequence of sticky wages (and thus no equilibrium on the labour market in the initial situation) things turn out differently. Suppose that w
1is the wage rate and N is the supply of native employees.
Without immigration there is a shortage on the labour market of M persons. When these M persons immigrate, it will lead to an immigration surplus of size BDC, with no change of the wage rate and no transfer of income from the native employees to the capital owners. The surplus will benefit the capital owners which is a Pareto efficient improvement.
A situation like the above can exist in certain sectors of the Dutch economy. For example there is a shortage of employees in the education sector in certain parts of the Netherlands and in this sector the wages are relatively sticky.
The above descriptions are of course simple reproductions of reality. If one makes the model more realistic other results may occur. A more advanced edition of the model of Borjas (1999) shows that if the supply of capital is elastic and a distinction is made between low-skilled and high-skilled employees, the immigration surplus increases as immigrants are more different from natives. For a country with a relatively high-skilled working population this would mean that the immigration surplus is maximized if relatively low-skilled immigrants are permitted to enter the country (Bosman, 2003).
Nevertheless, if the supply of capital is inelastic, the model of Borjas (1999) shows that it is more efficient to permit more high-skilled immigrants to enter the country. This opens the question which is most likely situation in the Netherlands.
Roodenburg et al. (2003, p.56) adapted the above model for the Netherlands to predict
the economic effects of immigration. The stylized model is empirically estimated and
they conclude that immigration will lead to an increase of GDP due to the wages of the
immigrants. They also show that the overall gain of residents will be small or even
negative. The amount of redistribution between residents will be substantial and the more
the skills of the immigrants differ from those of the residents the larger the amount of
redistribution will be. The residents with skills comparable to those of the immigrants will lose and residents with complementary skills to those of the immigrants will win in the long run. The capital owners will win in the short run, but in the long run their gains will disappear. Finally the study of Roodenburg et al. concludes that due to labour market imperfections, part of the income effect will lead to employment effects (unemployment instead of wage decrease).
In sum, the social and economic characteristics of the immigrants play an important role in the influence of the immigrants on the labour market. The type of skills of the immigrants determines the redistribution of income and whether a native resident with certain skills will benefit or lose. Generally, Roodenburg et al. (2003) conclude that large scale labour immigration is not an effective instrument to cope with the financial effects of an ageing population in the Netherlands. On the other hand, small scale labour immigration may be beneficial to the Dutch labour market, especially if immigrants are high-skilled, have good prospects for a job, and if they fill vacancies which are otherwise difficult to fill. Note that Borjas’ model would come to the same conclusion under the condition that the supply of capital in the Netherlands is inelastic.
4.1.2 Public sector effects
In the migration discussion the argument that immigration could be a solution to the problems as a result of the ageing population and the costs involved for the government on public pensions and health care is frequently made. Immigration could increase the tax base and therefore partly close the financial gap.
Roodenburg et al. (2003) use two methods in order to analyze the fiscal impact of
immigration and discuss both the cost side and the benefit side. The first method
calculates the present value of lifetime net contribution of an individual immigrant to
public finances as shown in Figure 6.
The second method of Roodenburg et al. computes the aggregate impact of a persistent additional yearly inflow of immigrants on future public budgets based on the above mentioned age profile. The immigrant’s age profile is combined with the size and age composition of the fictitious population
5increase that is generated by the additional inflow of immigrants.
Both methods show that the answer to the question, whether immigration has a positive or negative impact, depends on the age of the individual immigrant. Immigrants aged 25 at the time of entry have the most favorable outcome. These immigrants do not increase the cost of education and yet have the favorable ‘middle ages’ in front of them. The expensive stage of elderness is still far away and thus weighs less heavy in present value terms.
The first conclusion on the public sector effects found in the paper by Roodenburg et al.
(2003, pp. 80-81) is that the fiscal impact of an immigrant depends very much on his or
5 The population is increased by the inflow of migrants.
Figure 6: Age profile of lifetime net contributions, 2001
Source: Roodenburg et al. (2003).
her age at entry and his or her social and economic characteristics (labour market performances). The outcomes are most favorable for the immigrants who are 25 years of age at entry and perform well at the labour market. The second conclusion of this study is that for all entry ages, immigrants turn out to be a burden to the public budget if their social and economic characteristics correspond to those of the present average non- Western resident. Accordingly, budget balances are affected negatively. The third conclusion is that the average negative contribution of immigrants is not fully the result of a lagging performance. Partly this is the reflection of the generous system of Dutch collective arrangements. The fourth conclusion is that immigrants who perform better on the labour market than average Dutch residents alleviate public finances over a wide range of entry ages. Accordingly, an inflow of such immigrants would positively affect the budget balance. The final and overall conclusion of Roodenburg (2003) is that the results indicate that immigration cannot offer a major contribution to alleviate public finances and thus to become a compensating factor for the rising costs for government due to the ageing of the population.
Leibfritz et al. (2003) underline the conclusion that age is a decisive factor whether an immigrant affects the public sector positively or negatively. They add that the fertility rate of the immigrants may differ from the existing population, which changes the age structure and thus changes the ageing problem. If an average immigrant has more children than the average native this changes the net contribution of the immigrant, and thus affects the overall public sector.
5 The consequences for the Dutch economy
In chapter 3 the yearly immigration flow from the acceding countries to the Netherlands was predicted to be approximately 0.1% of the CEEC population
6and will probably be smaller due to network effects and language effects. In this chapter we assume that this number is correct and analyze the consequences of this immigration flow. In persons this number represents an inflow of approximately 42.000 immigrants per year, which is comparable to a city with the size of Doetinchem and represents 0.2% of the Dutch population.
6
5.1 Labour market effects
The labour market effects, are in this paper, represented by wage effects, employment effects and the immigration surplus. Following the theory of Borjas (1999) as mentioned in chapter 4, the immigration would lead to a downward pressure on the wages.
Assuming sticky wages in most of the sectors of the Dutch labour market and no other economic effects because the immigration, this would lead to a rise of the unemployment rate from 5.0% to 5.4%
7in the first year after the enlargement. In the second year the unemployment rate rises from 5.4% to 5.7%. Previous enlargements showed remigration after the first two year, so we only calculate the consequences of the migration in the first and the second year.
If we use the wage curve of Blanchflower and Oswald (1994) and assume wages can decrease, we calculate a wage decrease following the immigration of 1.20 % in the first year and with 0.6% in the second year
8. Of course these calculations are done with some strong assumptions and the remigration has not been taken into account, but it shows the effects are not negligible.
In the sectors where a shortage of employees exists, an immigration surplus is generated, which benefits the capital owners (assuming that the immigrants are qualified to fill in the vacancies). A recommendation for further research on this subject is to determine the qualifications of the (potential) immigrants in order to give a more detailed overview of the labour market consequences.
5.2 Public sector effects
As mentioned in chapter 4 the public sector effects depend mostly on the social and economic characteristics of the immigrants. The age of an immigrant determines whether the immigrant has a positive or negative effect on public finances. The knowledge of
“western” languages is on average better for the younger residents of the CEECs. This
7 The immigration leads to an increase of the labour force of 30.000 (under the assumption that 70% of the immigrants will be part of the labour force) which rises the number of unemployed from 820.000 to 850.000 (because of the sticky wages the number of unemployed rises with the same amount as the labour force increase). This raises the unemployment rate for 5.0% to 5.4%.
8 We use a simple version of the wage curve (ln W =-0.17 * ln U) and the coefficient -0.17 is based on the research on the Dutch labour market by Blanchflower and Oswald (1994). W0=-0.17 ln (0.05)=1.66 and W1=-0.17 ln (0.054) =1.64, which represents a decrease of 1.2%.
