Potential East-West Migration after Enlargement

(1)

Hubert Krieger and Bertrand Maitre*

This study presents new empirical evidence on the potential extent and structure of migration from the ac cession and candidate countries (ACC-13) to the EU based on indicators measuring the intention to mi grate. It is based on individual survey data collected in the 13 accession and candidate countries in 2002. It predicts that the volume o f migration is likely to be from 1.0 to 1.5% in the next five years. The prediction concurs with the latest results of an econometric study by the European Commission predicting a gross mi gration of around 1,100,000 in five years. Migrants are young and better educated or college or university students. The sending countries face the prospect of a major youth drain of 2 to 5% of the youngest co hort combined with a brain drain. The sending countries are at risk of losing significant numbers of college or university graduates. This raises important issues for EU developmental policy on all the ACC-13. What are potential effects on the old EU member states? The overall volume of expected EU migration is much less than predicted by some politicians and in the public debate. Most of the old EU countries are hardly af fected. In addition, two thirds of the migration is likely to be of temporary nature.

Keywords: enlargement, East-West migration

Introduction

The accession of ten new member states to the European Union on 1 May 2004 was ac companied by a heated public debate on post enlargement East-West migration. This article contributes to the policy debate on migration in the EU by presenting new empirical evi dence on the potential extent and structure of migration from the accession and candidate countries to the EU. Recent newspaper articles in Ireland and the UK have predicted a mass influx of Roma people from Central and East ern Europe after 1 May 2004. The E conom ist opened this round of media coverage in the UK with the headline 'The Hoards are Coming' (The E conom ist, 15 January 2004). The head line in the British Sun was even more dra matic, '300,000 Gypsies Heading for Britain'

(UK Sun, 19 January 2004) and the Irish edi tion wrote Thousands of Travelers are on their Way (Irish Sun, 19 January 2004). These senti ments are quite different from those in Ger man newspaper articles dated mid-January 2004 which, based on the latest scientific evi dence, did not predict any significant increase in East-West migration (KSTA, 17-18 January 2004).

Particularly in the vicinity of the ACC-13, e.g. in Germany, Austria and Scandinavia, three questions should be posed: How quickly are citizens from the accession countries to be granted working permits in the EU countries? How much of an inflow will this trigger? And what are the possible economic and social im pacts of increased emigration from the acces sion countries?

The fears are somewhat confirmed by

* Hubert Krieger, research manager at the European Foundation (Dublin), addresses various quality of life issues including migration, housing, poverty and families (e-mail: hubert.krieger@eurofound.eu.int). Bertrand Maitre, researcher atthe Economic and Social Research Institute (Dublin), addresses various aspects related to deprivation, housing, quality of life and social cohesion (e-mail: bertrand.maitre@esri.ie).

(2)

econometric studies, which estimate that from five to fifty million people could move to Western Europe from Eastern Europe in ten to fifteen years (Zimmermann 1996). Other eco nomic heavyweights like Layard et al. (1994) and Baldwin (1994) come to less dramatic but still significant estimates of the number of po tential migrants. Layard et al. estimate a migra tion potential of around 3% over fifteen years, which means three million migrants from the Eastern European accession and candidate countries.1 Baldwin's estimates are higher at five to ten million. Newer European Commis sion studies (Bruecker et al. 2001 and 2003) predict that migration from Eastern Europe to the old member states will increase from one to almost four million migrants between now and 2030.

In a recent policy statement, the European Commission addresses policy concerns on la bour mobility in an enlarged Europe. It pre dicts that despite existing fears, 'the most likely scenario is one in which labour mobility will be moderate to limited and will after a likely short upsurge, just after the accession period (...) start declining again to fall below 100,000 persons per year before the end of the decade' (Commission of the European Communities 2002: 6) .

EU policy

Since one of the basic principles of the EU is free worker movement across the borders of its member states, there is a good chance of a mass increase in legal migration after acces sion. This is why the notion of migration from the new EU members to the old ones has played a significant role in the negotiations preparing for accession. The current accession round was launched at the EU Copenhagen Summit in July 1993, where the Copenhagen Criteria for Accession were adopted. Between 1994 and 1998, thirteen countries expressed an interest in joining the EU. The discussion has been gaining momentum since the mid-1990s. As far as migration from the accession coun tries is concerned, the European Commission has mainly been interested in balancing free dom of movement against the specific labour market concerns of a few old member states situated close to the new ones.

The accession treaties were finally signed at the Copenhagen Summit in December Potential East-West Migration after Enlargement

2002. The new arrangements differ somewhat from those in the earlier accessions of low-income countries. In the cases of Greece, Spain and Portugal, restrictions on migration were agreed upon for a transitional period of seven years. The new accession treaties stipulate re strictions on free movement from the new into the old member states for all the accession countries except Malta and Cyprus. For the re maining eight countries, in principle the tran sitional arrangements are in place for two years plus three years plus two years2

In general, each current member state can decide how liberal its national measures are to be. As Austria and Germany are the key target countries, they both have the right to apply flanking measures to address serious distur bances or the threat of serious disturbances in specific sectors of their labour markets, which could arise in certain regions from the cross- border provision of services. Two other current member states, the UK and Ireland, do not in tend to restrict the entry of workers from the new member states after enlargement, but note that some restrictions are being made to the newcomers' access to social welfare bene fits in their countries for up to two years.

Focus

Against this background, our study describes and analyses the potential migration from the accession and candidate countries to the EU.'’ It is based on individual survey data collected in the thirteen accession and candidate coun tries in early 2002. The data collection was part of the Candidate Country Eurobarometer (2002.1) financed by the Director General of Employment and Social Affairs. The study con tributes to the ongoing policy debate by analys ing the socio-economic conditions and m o tives of potential migrants from the candidate countries. The analysis highlights the hetero geneity of potential migrants with a view to enabling policy makers to develop targeted pol icy measures. It raises the question of whether real social and economic migration-related problems will emerge in the old or new coun tries of the EU. The key concerns pertain to the possible brain and youth drain to and from these countries. In addition, there are other specific concerns in the new member states.

(3)

Potential East-West Migration after Enlargement

Specific concerns of the new member states

In the accession countries, the possible labour market effects of accession are viewed as dou ble-edged. Accession countries with high un employment and low economic growth rates benefit from the migration of unskilled and unqualified workers because it reduces their la bour supply and leaves fewer people without a job. The money migrants send home also has a positive impact on their income, consumption and aggregated internal demand. But if coun tries witness the emigration of higher qualified people, it can erode their long-term competi tive position. It is commonly acknowledged that a brain drain has a detrimental effect on the developmental process (Commission of the European Communities 2002: 15).

Detrimental effects are also predicted by scholars sceptical about the neo-classical eco nomic assumptions on equalising the prices of goods and production factors.4 Movements to wards the centre will accelerate growth and de velopment in the core, and the periphery will increasingly lag behind. 'Migration will then promote the structural change in a positive way in the immigration area and in a rather ne gative way in the emigration area' (H ille/ Straubhaar 2001: 88). According to these scho lars, high levels of migration from the acces sion countries to the EU would have the poten tial of long-term negative employment and economic growth effects in those countries. However, in addition to the European Com mission, other observers are also doubtful about a tidal wave of East-West migration after enlargement.

Doubts about a serious influx of migrants from

the accession countries

Four main arguments are usually developed, all of which question the likelihood of a signif icant inflow of migrants from the accession countries to the EU. Firstly it is argued that a significant percentage of the overall potential of migrants from the accession countries were already in the old EU member states before ac cession. The available figures show that this only led to a small increase in the population of the old member states. In 1998 the fifteen EU member states hosted fewer than 900,000 citizens of Central and Eastern European Countries (Bruecker et al. 2001). Secondly, the demographic situation in the accession coun

tries is similar to the demographic situation in the EU. Fertility in the accession countries is even lower than in the EU, and mortality has dramatically increased, producing a decreasing population trend in the next thirty to forty years. So all things considered, the employ ment chances for younger and better educated people will increase in the candidate coun tries. Based on this trend, Fassmann and Muenz (2002) predict a decrease in the migra tion potential. Thirdly, a transfer of resources for economic development, increased interna tional trade and mass inward investment will accompany the accession of the new member states, accelerating their economic develop ment and giving the new member states a posi tive growth differential in relation to the old ones. In addition, contrary to the economic theory predictions, European employees have barely reacted to the opportunity of free move ment in a common labour market. Straubhaar presents an explanation and notes that 'To an important degree, trade has replaced the eco nomic demand for migration in the EU' (2001a: 170).

Research questions and limitations

However, this study has several important lim itations. Firstly, the available data. This is a sec ondary analysis of an omnibus survey de signed to measure the quality of life in the can didate countries. Regional mobility is regarded as a factor contributing to quality of life. Inten tions to migrate to the EU are only a minor as pect in the whole study. Hence important in formation usually collected in migration stu dies is not available. For example, there is no information on the target countries for migra tion, the length of the intended stay in the tar get countries or specific activities preparing for migration (looking for accommodations, try ing to get a work permit). It also cannot distin guish various forms of migration, ranging from long-term, relatively permanent and unidirec tional migration to semi-permanent, multiple ties migration, with migrants commuting fre quently back and forth.

Secondly, the explanation for the intention to migrate based on socio-economic factors should be mainly limited to push factors based on potential migrant' socio-economic condi tions in their home country, since the survey does not provide any information on the target

(4)

Potential East-West Migration after Enlargement country. This limitation can be partly compen sated by indicators on pull factors based on rea sons for migration, which cover important conditions in the host country (e.g. career and income prospects, welfare system).

More in general, in conceptualising and ana lysing migration to the European Union, one should bear in mind that outward migration to the EU is only one regional mobility option. Usually mobility is classified according to proximity: mobility within the town and vil lage, intra-regional mobility, inter-regional m o bility, mobility to other European countries and mobility outside Europe. A recent OECD analysis (2001: 44 ff.) highlights four trends in the Eastern European EU candidate countries: (1) persisting East-West migration but at a much lower level than observed in the early 1990s, (2) an intensification of migration flows inside Eastern Europe, (3) a very strong in crease of flows between the Eastern European and candidate countries and those in the south (Ex-Yugoslavia) and east (CIS), and (4) a significant but numerically small migration of highly qualified managers and skilled workers from West to East.

Lastly, from a conceptual point of view, mi gration should be regarded as one of several possible reactions to unsatisfactory conditions. Under these circumstances, an individual can lower his expectations, try to change the exist ing circumstances, or opt for mobility. Within the range of mobility options, migration to an other country should be regarded as a serious decision, which is usually accompanied by high information and search costs and higher degrees of uncertainty.

Data

This section is based on Eurobarometer data collected for the European Commission, which has launched a series of surveys in the thirteen accession and candidate countries.5 The sur vey was conducted in the early months of 2002 in the thirteen candidate countries, Bulgaria, Cyprus, the Czech Republic, Estonia, Hungary, Latvia, Lithuania, Malta, Poland, Romania, Slovakia, Slovenia and Turkey. In Cyprus, the survey only pertains to people who live in the southern part of the island. An identical set of questions has been posed to representative samples of the population above the age of fif teen in each country. The regular sample in

Candidate Countries Eurobarometers is 1000 persons in each country except Malta and Cy prus, where it is 500. For this study, a sample of 2,000 was used in Poland and Turkey to achieve better coverage.

The data entail an important additional limitation. Due to the number of interviewees with an intention to migrate relative to the sample size and the small sample size in some countries, it is not possible to present a statisti cally significant country-by-country analysis. This is why the decision was made to create country clusters. Since we lack a theory of so cial welfare regimes in the candidate coun tries, the clustering was done on a pragmatic basis. One cluster includes the two remaining candidate countries Bulgaria and Romania, which have similar economic and social struc tures and where migration trends are moving in the same direction. A second cluster in cludes the three Baltic countries. A third clus ter includes the Czech Republic, Slovakia and Hungary and is based on a shared Habsburg Central European history and a joint post-1989 transition experience. However, there are still important differences. The fourth cluster is even more diverse and includes Slovenia and the two new Southern European member states Cyprus and Malta. These three countries are the new member states with the highest standard of living. Poland and Turkey, the two largest candidate countries, have not been clustered, since the survey provides an over- sampling of 2,000 interviews for each of them and thus a sufficient number of interviewees.

Concepts of migration

This section briefly surveys and classifies the theoretical concepts of migration. It has to be selective in several ways, since the large num ber of theories and their wide range make it impossible to give a complete overview.6 Mi gration theory generally addresses three ques tions: (1) Why move? (2) Why stay? (3) Why have many people never considered moving despite their objective conditions?7 To struc ture the various theoretical contributions to the migration debate, let us give a brief classifi cation. There are six basic classification di mensions.

(5)

1 Theories are formulated in various scientific domains. The main distinction is between economic and non-economic theories of mi gration. Economic theories of migration in clude relevant macro- and micro-theories. These theories have gained importance in the current policy debate on migration by es timating expected migration potentials in general and to the EU from the accession countries in particular. They also provide salient policy advice. Non-economic the ories can be found in multifarious social science subjects. In newer writings, they mainly cover sociological and socio-psycho logical aspects.8

2 The level of analysis distinguishes macro, meso- and micro-perspectives, which are summarised in figure 1 by Faist (2000). Macro-economic neo-classical theories of migration focus mainly on conditions lead ing to an equilibrium between regional la bour markets and on the migrant as a suppli er of labour. Other macro-economic theories question the model of rational choice and ar

gue that international migration is mainly caused by the intrinsic labour demand of modern industrial societies. A good example is the dual labour market theory. The world system theory is another macro-approach, which interprets the creation of a mobile po pulation in peripheral non-capitalistic socie ties as a result of penetrating capitalist eco nomic relations world-wide.9 Various institu tional theories are also important at the macro-level (see Massey et al. 1993: 450 ff.). Once migration at the international level has begun, various institutions are established to satisfy the demand created by the large la bour supply from the sending countries and the limited number of workplaces and visas usually offered by the receiving countries. This leads, for example, to the growth of an institutionalised black market and humani tarian support.

M icro-economic neo-classical theories view migrants as investors in human capital, as consumers, or in the economy of the house hold as producers of household specific com modities. Neo-classical migration theory has

MICRO

Value or desires and expectancies

MESO

Collectives and social networks

MACRO

Macro-level opportunity structures individual values and social ties economics

expectancies -strong family and _{- income and unemployment} - improving and securing household ties differentials

survival, wealth, status, comfort, - weak ties in networks stimulation, autonomy, affiliation of potential movers,

politics

-regulation of spatial mobility through nation and morality brokers, and stayers

symbolic ties states and international

- kin, ethnic, national, regimes

political and religious _{-political repression, ethnic,} organisations, symbolic _{national and religious}

communities _conflicts

contents of ties-transactions _{cultural setting} obligations, reciprocity, and - dominant norms and solidarity; information,

control and access to resources of others

discourses

demography and ecology - population growth -availability of arable land - level of technology

(6)

been supplemented in recent years by the n ew econ om ics of migration, which adds two further dimensions to the debate. It challenges the individualistic assumptions of neo-classical micro - theory and highlights the importance of the wider family context in the decision-making on international mi gration. It also stresses the importance of risk minimisation as supplementary factor in decision-making in addition to the ex pected income maximisation.

In the past thirty years, important m icro-so ciological migration concepts have been de veloped in the remit of an RREEMM mod el,10 which is sometimes viewed as the most appropriate rational choice model of human behaviour (De fong and Fawsett 1981). It is based on a specific ideal type of person, which can be described within a wider uti lity maximisation model. Sub-models of this approach include the theory of place utility (Wolpert 1965), the SEU concept and the Ha bit model.11 Other important socio-econom ic micro-approaches are Lee's push-pull mod el (1972) and the concept of relative depriva tion (Stark and Taylor 1991). A meso-level perspective is presented by network theories, which stress the importance of social net works connecting migrants, former mi grants, potential migrants and non-migrants in the origin and destination areas via social ties based on family, friendship or shared ex perience in the community of origin.

3 An additional classification of theoretical contributions is related to different points in time when the causal analysis of initiate mi gration behaviour and attitude starts. Most theories analyse the conditions that the in ternational movement of labour, e.g. wage differentials in neo-classical theory and la bour demand based pull factors in the dual labour market theory. Another set of con cepts analyses the effect on migration of an ensemble of new conditions arising in the course of migration movements of larger groups of people as an independent cause in itself. This includes the spread of migration networks, institutions supporting or repel ling migrants, and the changing social meaning of work in societies with high per centages of migrants. These aspects are also addressed under cumulative causation. Potential East-West Migration after Enlargement

4 A further dimension focuses on various ex- plananda of migration theory by giving ex planations for different types of interna tional migration.12 The standard orientation is towards one-time long-term migration to another country. Some observers, however, highlight the increasing quantitative impor tance of only semi-permanent migration, with migrants sustaining multiple ties to their home country, sometimes commuting frequently back and forth, and thus creating a transnational social space and a kind of transnational community with a multi-local place to live and work (Faist 2000).

5 An additional dimension is the capacity of a specific theory of migration to integrate other migration concepts. Neo-classical mi cro-economic theory has tried by reducing abstraction (Kalter 1997) to preserve its key nomological basis of an income- maximising h o m o econ om icu s and at the same time to integrate a variety of additional concepts covering search and information behaviour as well as the investment in human capital. The n ew economic theory of migration also tries to encompass traditional as well as the enlarged neo-classical theory and concepts on decision-making in families and risk management. Another effort to integrate sev eral concepts on migration is made in the SEU model, which is strongly related to a so ciology-based general theory of action. Here, the complete set of neo-classical rational choice concepts is integrated with wider con cepts of utility maximization in an indivi dualistic and family concept, which also fa vours integrating process-based migration theories. In the past though, most migration concepts such as Anderson's gravity model (1955) and Lee's push-pull model (1972) had a very specific and limited focus and were thus unable to provide any basis for theoretical in tegration.

6 A final distinction should be drawn between real theoretical contributions that give a cau sal explanation of migration behaviour con taining at least one general law of human ac tivity (Esser 1993: 47), and m ere th eories pro viding either classifications, typologies and tautological explanations of migration, or in dividual plausible hypothesis e.g. in the con

(7)

P o te n tia l E a s t-W e s t M ig r a tio n a fte r E n la rg e m e n t

text of a multivariate statistical model, with out any general integrated framework. In this context it is important to note the point made by Kalter (1997: 34 ff.) that the widely used push-pull-model of migration is less a theoretical approach in its own right than a suggestive way of classifying various influ ence factors. It is based on the idea that m i gration is caused by repelling factors in the place of origin (push) as well as attractive conditions at the destination (pull). A push factor becomes relevant if the disadvantages in the country of origin are greater than at the destination. A pull factor is important if the advantages at the destination are greater than in the country of origin (Deutsche Bank Research 2003: 3). Which factors are included and how they interrelate with the actual migration behaviour often fail to be discussed and should be supplemented, e.g. by micro-economic or sociological concepts of migration. The attraction, however, of the push-pull-paradigm is its openness to a vari ety of very different explanatory variables.

Potential to migrate to the EU

Introduction and conceptual background

This section gives a first overview of the poten tial migration from East to West in an enlarg ing European Union when the ten accession countries (AC-10) join the European Union in 2004, and from all the thirteen accession and candidate countries (ACC-13). The focus is on voluntary rather than forced migration.13 All the predictions regarding potential migration to the EU are based on the assumption of free individual movement, which is currently lim ited via the Accession Treaties for the New Member States and the additional restrictions for the remaining three candidate countries. Within this remit the section describes the construction of the dependant variables, which should be viewed against the background of four important conceptual dimensions: (1) the gap between the intention to migrate and the actual migration, (2) the use of softer and hard er indicators, (3) the conceptualisation of mi gration as a process, and (4) the existence of different forms of migration.

1 An important limitation of survey-based stu dies is the gap between the observed inten tion to migrate and the actual migration. Kal ter (2000: 463 ff.) briefly surveys the theoreti cal and empirical research, which shows that although a strong intention to migrate is the most important factor for predicting actual migration, greater parts of the variance of migration behaviour are explained by exter nal factors.

2 In order to narrow the gap between the pre dicted behaviour based on survey data and the actual migration behaviour, it is useful to reflect on the type of indicators used and how they measure the strength of the pre dicted migration potential. There is an impli cit assumption that it is possible to achieve greater predictability by using harder indica tors. Softer indicators on the migration po tential usually measure a general attitude or willingness to migrate. A typical question in the April 2002 Eurobarometer is: Do you in tend to go to live and work for a few months or for several years in a current EU country in the next five years? The answer categories are yes and no. The indicator covers an extremely wide range of kinds of migration, provides a wide time span and permits var ious motives for migration. Since it is con fined to a yes-no dichotomy, there is no gra dation of strength. Predictions of actual mi gration behaviour based on indicators of this kind allow for a high degree of uncertainty and vagueness. An alternative approach uses hard indicators, which measure a series of concrete activities preparing for possible mi gration. Strong indicators might be: asked for a work permit or sold property in the country of origin. Indicators of medium strength pertain to such activities as learn ing a foreign language and contacting people in the target country (IOM 1999; Fassmann and Hintermann 1997).

3 The use of different indicators also reflects the conceptualisation of migration as a pro cess that goes through various stages, an initial start-up situation characterised by in ertia, a phase of developing an inclination to migrate by overcoming natural inertia and integrating the possibility of migrating into the potentially perceived alternatives of ac

(8)

tion, and a stage of developing a concrete in tention to migrate by comparing the alterna tives, e.g. in a rational choice model, and at tributing the target area with greater utility than the existing place of residence, and lastly the actual migration behaviour.

4 The construction of an indicator measuring the intention to migrate should consider var ious forms of migration. Migration used to be viewed as unidirectional and relatively permanent mobility towards a receiving country. The IOM (1999: 15-16) suggests dis tinguishing three main kinds of migration: traditional permanent migration, long-term temporary migration and short-term labour migration, which includes seasonal work, casual work and even cross-border

commut-The key dependent variables of the study have been constructed on the basis of these consid erations.

Construction of the potential to migrate from

East to West

The study measures attitudes to migration to the EU at three different levels: (1) general incli nation to migrate, (2) basic intention to migrate and (3) firm intention to migrate.

1 The general inclination to migrate reflects a basic attitude to migration to the EU. The questionnaire includes a direct question: 'Do you intend to go to live and work for a few months or for several years in a current Eu ropean Union country in the next five years?' To control the validity of the answers to this question and construct the indicator on the general inclination to migrate, positive an swers to this question are combined with the results of two questions that measure the in tention to exhibit regional mobility in the next five years by moving to a place outside one's city, town or village. In this respect two questions are posed: 'Do you intend to move in the next five years?' and if so, 'In the next five years, do you intend to move to another city, town or village in the same region, to another region in the same country, to an other country in Europe or to live in a coun try outside Europe?' For logical reasons and from a common sense point of view, it

should be assumed that anyone who intends to migrate to the EU also has a basic inten tion to exhibit regional mobility in the next five years.16 Respondents with affirmative answers on the combined indicator can be regarded as having overcome their natural inertia as regards migration and as exhibit ing a general inclination to migrate to the EU.

2 The basic intention to migrate provides a stronger indicator. It combines the intention to migrate to the EU as a direct question with the question of intended regional mobility in the next five years 'to another country in Europe' as part of a set of different forms of regional mobility. The indicator identifies Europe as the specific target, thus narrowing the possible target area. Unfortunately the question does not focus on the intensity of the intention to migrate.

3 The firm intention to migrate to the EU pro vides the strongest measurement in the pre sent study, and should provide the greatest probability in predicting actual migration be haviour by capturing, at least in part, the in tensity of the intention to migrate. It is mea sured using four variables, three of which have already been used to measure the basic intention. In addition, these answers are controlled by a fourth indicator, which mea sures the willingness to live in a country where a foreign language is spoken. The question is 'How willing would you be to live in a European country where a foreign lan guage is spoken?' Explicitly accepting the challenge of migrating to a country where a foreign language is spoken is an indicator of medium level seriousness as regards the in tention to migrate. The response categories are very m uch, to som e extent, n ot m uch and not at all. The indicator of a firm inten tion to migrate only pertains to respondents who answer very m uch to this question. As did other studies, the present study aims to measure the strength of the intention to mi grate by using a more sophisticated scale. With this operationalisation, the study has opted for an inclusive approach covering all the relevant indicators in the questionnaire. In moving from softer to harder measurements, it

(9)

Potential East-West Migration after Enlargement only partially uses harder indicators but elimi

nates logical contradictions between answers to different questions and identifies the re spondents with the highest consistency as the ones most apt to migrate. Unfortunately, the Eurobarometer measurement of the intention to migrate lacks important indicators of the strength of the intention to migrate, which have been used in previous studies.

Potential to migrate to the EU

This section addresses the three key depen dent variables: the general inclination to m i grate, the basic intention, and the firm inten tion to migrate to the EU. As this study dis cusses migration in a broader quality of life context, the intention to migrate is calculated for the whole population in the accession and candidate countries above the age of fifteen. In addition to labour market aspects, we consider a variety of other rationales for migration in cluding personal wishes to be reunited with the family or poor housing conditions.17

The first column in table 1 shows that 4.6% of the citizens in the accession and candidate countries have a general inclination to move. This is only 1.5% as regards the basic intention to migrate and just under 1% (0.9%) as regards the firm intention to move to the EU. That means only 20% of the population with a gen eral inclination to migrate should be regarded as a hard core group that exhibits a higher probability of actual migration. The AC-10 ex hibit a similar trend at a lower level. It is strik ing though that the difference between the two groups of countries is reduced from 1.5% re garding the general inclination to 0.1% regard ing the firm intention to migrate. On the basis of these results, the same propensity to migrate

can be predicted in the accession countries as in the candidate countries.

The basic intention to migrate to the EU is shown in column 2 of Table 1. This indicator measures the respondents' intention to migrate to the EU in the context of all the other rele vant target areas for migration. It forces people to make a clear choice. A total of 1.5% of all the respondents in the accession and candidate countries have a basic intention to migrate to the EU. Bulgaria and Romania are the only countries with a higher figure of 3.2%, and only the three Baltic countries have an inten tion to migrate of 2%. At the lower end we find three country clusters of 0.8%. The only sur prising result is Turkey, which has special con ditions that are explained in the previous sec tion.

The final indicator covers the firm intention to migrate by cross-checking the basic inten tion to migrate with a high degree of willing ness to live in another European country where a foreign language is spoken. In three country clusters, 0.5% of the population has a firm intention to migrate to the EU, i.e. the Southern European accession countries and Slovenia, the Central European Habsburg countries and Turkey. As regards this indicator, Bulgaria and Romania are still the top supply countries for potential migrants in relation to their populations with one in fifty (2%). The highest country result is in Bulgaria with 2.5% and the lowest are in Malta and Turkey with 0.1 to 0.3%. With 1%, the Poles have the second highest firm intention to migrate to the Eu ropean Union in the next five years.

To summarise, in May 2004, two years be fore accession, the actual intention to migrate to Europe was relatively low in the AC-10 and

Table 1 Percentages of people with various migration intentions across countries

G eneral inclina tion Basic in te n tio n Firm in te n tio n

Poland 3 .7 (6 3 ) 1 .6 (2 7 ) 1 .0 (1 6 )

Bulgaria, Rom ania 5 .0 (8 4 ) 3 .2 (5 7 ) 2 .0 (3 1 )

Cyprus, Malta, Slovenia 2.1 (38) 0 .8 (1 4 ) 0 .7 (1 1 )

Turkey 6 .2 (1 3 6 ) 0 .8 (1 8 ) 0 .3 (7 )

Hungary, Czech Republic, Slovakia 2 .4 (5 9 ) 0 .8 (2 6 ) 0 .6 (1 7 ) Estonia, Latvia, Lithuania 3 .5 (1 1 2 ) 2 .0 (5 8 ) 0 .8 (2 5 )

A C -10 3.1 (272) 1 .3 (1 2 5 ) 0 .8 (6 9 )

ACC-13 4 .6 (4 9 2 ) 1 .5 (2 0 0 ) 0 .9 (1 0 7 )

Source: Candidate Countries Eurobarometer 2002.1 April, 2002, numbers in brackets indicate the absolute num ber of respondents.

(10)

Potential East-West Migration after Enlargement ACC-13 countries. Despite the various meas urement errors that usually lead to an overesti mation of the propensity to migrate on the ba sis of individual survey data, the present study predicts an overall moderate extent of migra tion within a corridor of upper and lower lim its.

Comparison with estimates of a recent

European Commission macro-study

The last part of the empirical analysis in this section compares the estimates of the migra tion potential based on the 2002 Eurobaro meter with results of a recent European Com mission econometric study (Bruecker et al. 2001), The Im pact o f Eastern Enlargem ent on E m ploym ent an d L abou r M arkets in th e EU a n d M em ber States. Comparability is no sim ple matter, since the two studies use different methods. The present study is based on indivi dual survey data with all the advantages and disadvantages in predicting actual migration.18 It predicts the possible increase in migration to the old EU member states in the next 5 years or by the end of 2006. Its key dependent variables are various estimated migration rates in the countries of origin as percentages of the popu lation above the age of fifteen.

The European Commission study uses a macro-economic approach to estimate the an nual migration flows from 2002 to 2030.19 Un like our study, it only includes ten candidate countries and leaves out the three Southern European candidate countries Malta, Cyprus and Turkey The coefficients for the estimation of the migration potential are based on time series data of the migration flows to Germany from 1967 to 1998. Its key dependent variable is the annual change in the ratio of the stock of migrants to the population in the receiving country. Based on these assumptions, the study estimates an increase in the stock of mi grants from 1,160.000 in 2002 to 3,900,000 in 2030. The annual flow figures fell from 340,000 in 2002 to fewer than 3,000 in 2030.

Because of the differing methods, the two studies can only be compared on the basis of estimated changes in the stock figures in the ten accession and candidate countries from 2002 to 2006. For this comparison, the follow ing calculations should be made for the Eu ropean Commission study: the predicted num ber of migrants is 1,160,000 in 2002 and

2.250.000 in 2006.20 This means a predicted increase of 1,100,000 migrants from 2002 to 2006 for the ten candidate countries. For our study to be comparable, it should provide re sults for the three dependent variables for the ten candidate countries in the Commission study. The results are as follows:

- general inclination ACC-10: 3.7%; - basic intention ACC-10: 1.8%; - firm intention ACC-10: 1.2%.

In the next step it should calculate the popula tion in the ACC-10 above the age of fifteen. The total ACC-10 population on 1 January 2001 was 102,000,000 (Eurostat 2002). Sub tracting the population under the age of fifteen of around 19,500,000 leaves a total of 82.500.000 people in the ten accession and candidate countries above the age of fifteen. Multiplying this basic figure by the three coef ficients provides the following band of a pre dicted increase in the migration stock up to the end 2006 in absolute figures:

- general inclination: 3,050,000; - basic intention: 1,480,000; - firm intention: 990,000.

The result of the comparison is that the two methods lead to similar results, predicting an increase in the total number of migrants from the ten Eastern European candidate countries of between 1,000,000 and 1,500,000 people for the narrow and medium band in the years up to the end of 2006 in the old member states of the EU.

Conclusions

The measurement and prediction of the vol ume of future migration to the EU from the candidate and accession countries is politi cally and methodologically complicated. Our study concurs with the results of most serious micro-studies that the EU should not expect a tidal wave of emigrants from the candidate countries. There is every indication of a vol ume similar to the one experienced after the Southern enlargement of the EU in the 1980s.

It is striking that the present study also con curs with the econometric study conducted by the European Commission in 2001. With a nar row corridor, both studies predict an increase of around 1,100,000 migrants from the East in the next five years. The present study shows a wide migration corridor to the EU of from 1.8 to 3.7% and a narrow corridor of around 1.2%

(11)

Potential East-West Migration after Enlargement excluding any return migration. These results

are also in line with studies by Layard et al. (1994) and Straubhaar (2001b), which estimate a gross migration rate of the home population over fifteen years of around 3 to 4% and a net potential of \V% to 2% from the Eastern Eu ropean candidate countries (excluding Turkey, Malta and Cyprus).

The differences in the volume of migration from various countries are important from a political point of view. According to our study, the highest migration potential in the ACC-13 is from Bulgaria and Romania. With a migra tion potential of less than 0.5%,. Turkey seems to be much less problematic The figures for all the larger accession countries are around 1% in the next five years.

However, our study has serious limitations. It cannot predict the intended target country or the migrants' intended length of stay. Nor does it provide any indicators of the cross-border commuting and return migration potential.

Socio economic-conditions and migration

Introduction

This section discusses two important socio economic variables, age and education, and how they are interrelated with intended migra tion to the EU.21 Its aim is to present a descrip tive analysis of the main socio-economic char acteristics of the respondents who express a willingness to migrate. This provides some ad ditional insight into the conceptual debate, e.g. on the increasing feminisation of migra tion and the importance of individual unem ployment as a driving force behind migration. A second aim is more political. The heteroge neity of the group of potential migrants should be taken into account in designing relevant la bour market and social policy measures and coping with the increasing regional mobility from the accession and candidate countries to the EU. It is important for policy-makers to know if most of the migrants have a higher or lower educational level, are younger or older, or are unemployed. 'It is therefore important that policies are designed according to the op portunities and constraints specific to differ ent groups (...) as well as according to specific groups' trans-national spheres of action' (Ny- berg-Soerensen et al. 2002: 5).

Age

There is a wide consensus in migration litera ture that age exerts a strong influence on mi gration (Fassmann and Hintermann 1997, Kal- ter 1997 Bauer and Zimmermann 1999, Price- waterhouse Coopers 2002). Younger cohorts are regarded as highly mobile, and in cohorts above the age of forty there is a significant de cline in intended migration. From an econom ic perspective, two main explanations are gi ven. Younger cohorts usually have better labour market prospects in the receiving country, especially if they are prepared to accept poorly paid jobs, mainly in the service sector. They of ten increase their labour market chances in the receiving countries by accepting work below their qualifications, thus giving local employ ers a lower wage rate combined with higher productivity. The second economic argument comes from human capital theory and suggests a better return on investment in migration in younger cohorts. Older workers have a lower economic incentive to migrate, since the amor tisation period for their investment is shorter (Bauer and Zimmermann 1999: 15).

Socio-economic models would suggest that younger people are more dissatisfied with the conditions in their country of origin. Com bined with higher aspiration levels, this results in a strong feeling of relative deprivation and frustration so that younger cohorts are more willing to migrate. The specific position of younger people in the family hierarchy can also explain their greater interest in migration. Older members of the family, in particular the head of the household, will remain in the country of origin to maintain the family, whereas younger members are sent abroad to earn additional money.

Between 85% and 90% of the people with a general inclination to migrate in the AC-10 and ACC-13 are below the age of forty (see table 2). In both groups of countries, the youngest co hort (15-24) has the greatest intention to mi grate. This tendency is however 10% higher in the AC-10 than the ACC-13.

A comparison of the general inclination with the firm intention to migrate shows a structural shift in AC-10 (see table 3). The dominant group is now the 25-to-3 4 -year-olds, who represent nearly half the potential mi grants. This is an increase of 20% compared with the general inclination. As to the ACC-13,

(12)

Table 2 Breakdown of individuals generally inclined to migrate by age group across countries Potential East-West Migration after Enlargement

15-24 25 -3 9 4 0 -5 4 55+ A ll

Poland 64.3 27.9 7.8 0.0 100.0

Bulgaria, Romania 61.5 21.7 16.0 0.7 100.0

Cyprus, Malta, Slovenia - - - -

-Turkey 39.8 42.8 13.0 4.5 100.0

Hungary, Czech Republic, Slovakia 60.7 31.8 5.4 2.1 100.0

Estonia, Latvia, Lithuania 44.8 40.1 14.7 0.4 100.0

AC-10 61.0 30.4 7.9 0.7 100.0

A C C -13 50.9 34.5 12.0 2.5 100.0

Source: Candidate Countries Eurobarometer 2002.1, April, 2002.

the trend is in the opposite direction. The youngest cohort increases by 6%, but the sec ond youngest cohort only increases marginally. This reverse trend is explained by changes in Bulgaria and Romania, where with 75% the role of the youngest cohort with firm inten tions is extremely high and shows an increase of 15% compared with the general inclination to migrate.

A further analysis of the general inclination of the youngest cohort to migrate shows that in Bulgaria and Romania, the two new Southern European member states and Slovenia and the three Central European Habsburg countries, from 60 to 65% of the individuals with a firm intention to migrate are in this age group. The lowest percentage is in Turkey with around 40% . There is no country where - based on the total migration potential - less than 40% are in the youngest age group.22 These results differ from those of Pricewaterhouse Coopers (2002) for three countries, Poland, the Czech Republic and Hungary, where people under the age of 25 only account for a third of the ex pected migrants. In our study, the results for these three countries are significantly higher, ranging from 52 to 65%.

To summarise: the results confirm the com mon belief that migration mainly occurs in the earlier stages of life. The relative percentages vary, however, from one country, period, and study to the next. All the studies agree that

after the age of forty, the intention to migrate is negligible. The receiving countries of the EU can expect newcomers for their labour markets and hope for a small improvement in their de mographic structures. Some sending coun tries, however, face the prospect of a major youth drain which, in countries like Bulgaria and Romania, may lead to an outflow of the youngest cohort of from 5 to 10% in the next five years. This raises serious questions for EU developmental policy towards Bulgaria and Ro mania in the pre-accession phase before 2007.

Education

A high educational level is viewed as a factor that exerts a positive influence on migration. From a human capital perspective, higher le vels of education are assumed to engender in creased income returns on specific segments of the labour market. It is also argued that higher levels of education generate a greater ability to collect and process information, which lowers the risk and increases the ten dency to migrate. Based on an analysis of sev eral international studies of migration, Bauer and Zimmermann (1999) develop the opposite hypothesis. They note an insignificant or even negative coefficient between levels of educa tion and the tendency to migrate. This can be explained by the prevalence of unskilled la bour markets for migrants in the destination countries, making migration less attractive for

Table 3 Breakdown of individuals with the firm intention to migrate by age group across countries

15-24 25 -3 9 40-54 5 5 + AH

A C -10 45.6 49.3 5.0 0.1 100.0

A C C -1 3 56.8 32.4 10.0 0.8 100.0

(13)

Potential East-West Migration after Enlargement Table 4 Breakdown of individuals with a general inclination to migrate by terminal education age across coun tries

Primary school

(up to age of 15)

Secondary school

(16-19 years old)

College or

university

graduate

(20+)

Still

studying

All

Poland 3.2 28.9 15.4 52.6 100.0 Bulgaria, Romania 18.9 36.7 13.6 30.7 100.0

Cyprus, Malta, Slovenia - - - -

-Turkey 37.2 23.1 18.9 20.7 100.0

Hungary, Czech Republic, Slovakia 6.8 32.7 17.1 43.5 100.0

Estonia, Latvia, Lithuania 2.8 31.9 31.4 33.9 100.0

AC-10 4.5 29.3 18.5 47.6 100.0

A C C -1 3 22.0 27.8 17.4 32.8 100.0

well-educated individuals.

The migration of students is sometimes viewed as a form of migration of young and qualified labour. In several receiving countries, it has even been made easier to switch from a student to a worker status than to migrate as a fully qualified employee (Kofman 2003).

The AC-10 and ACC-13 present two comple tely different patterns as to educational level (see table 4). Nearly half the potentially mobile population in the AC-10 are studying, another quarter have a college or university education, and only one in twenty have only completed primary school. The potential migrants from the AC-10 thus represent a highly qualified fu ture labour supply for the EU.

The potential migrants from the ACC-13 are distinctly different. A quarter only have a primary school education and another quarter only a secondary school education. Less than a fifth have a college or university degree and a third are still studying. Migrants from these countries have a 20% lower educational level than the ones from the AC-10, so that the inte gration of those migrants into the EU labour markets may prove more difficult. This differ ence can mainly be explained by the fact that in Turkey, nearly 40% of the mobile population only have a primary school education, and in Bulgaria and Romania around 20% of the po tential migrants only have a primary school education.23

Countries with a high percentage of poten tial migrants with a college or university edu cation are the Baltic countries with over 30% and Cyprus, Malta and Slovenia with over

20%. A total of 95% of the potential migrants from the AC-10 and 78% from the ACC-13 have an educational level above primary school.

More than half the young, qualified stu dents come from several more developed coun tries, Cyprus, Malta and Slovenia, or more transformed ones, Poland and the Czech Re public. Here there is the combination of a high educational level and a youth effect.

A comparison of the overall results for the general inclination to migrate with the overall results for the firm intention to migrate con firms the overall trend (see table 5). As to the AC-10, 99.9% of the migrants with a strong in tention to migrate have more than a primary school education and 30% have a college or university education. Nearly 40% are still studying. Around one in ten of the highly m o bile migrants in the ACC-13 only have a pri mary school education, one in four have a col lege or university education, and one in three are still studying. At both ends of the spec trum, this represents a difference of around

10% between the two groups of countries. The potential loss of part of the student population is a more serious matter for both the country groups. They are in danger of los ing one in eight or one in nine of their stu dents in the next five years. All things consid ered, this means a serious combined brain drain and youth drain for these countries and a resulting risk of reduced economic growth and negative developmental consequences as de scribed by Myrdal (1956),24 Wallerstein (1974) and Olesen (2002).25

(14)

Table 5 Breakdown of individuals with a firm intention to migrate by terminal education age across coun tries

Primary school

(Up to 15)

Secondary school

(16-19)

College or

university (20+)

Still

studying

All

AC-10 0.1 30.2 31.0 38.7 100.0

A CC-13 11.5 30.3 24.7 33.5 100.0

the country level. In Turkey, 15% of all the re spondents with a college or university educa tion have a general inclination to migrate; in Bulgaria and Romania this is true of nearly one out of twenty. There is the greatest danger of a student brain drain in Bulgaria and Romania, where nearly a fifth of the students have a gen eral inclination to migrate to the EU. It seems that the current difficult economic and social conditions combined with the delayed acces sion to the European Union have a negative ef fect. In comparison to Cyprus, Malta and Slo venia (8%) and Hungary, the Czech Republic and Slovakia (9%), these figures are extremely high.

The extent of potential migration is, how ever, dramatically reduced as regards the firm intention to migrate. In both groups of coun tries, around one out of forty students and one out of 100 respondents with a college or uni versity education have a serious intention to migrate to the EU.

These results have also been confirmed by other studies. According to Drinkwater (2002: 19) 'an extra year of education increases the probability that the individual is very willing to move abroad by less than one percentage point'.

From a European Union point of view, these figures should help alleviate fears about the ex pected structure of inward migration. A large majority of potential migrants from the AC-10 are well-educated. This means more opportu nities than risks for the old EU member states. For the candidate and accession countries, these figures mean a significant challenge to their internal economic and social develop ment. The extent of the potential brain drain seems more limited however as regards the firm intention to migrate. Many students have a general feeling that migration may be an op tion, but have not yet invested much in mak ing it happen in the near future. The future

will show whether the envisaged 2007 acces sion of Bulgaria and Romania stipulated at the Thessalonika Summit in June 2003 will change the attitude of better qualified gradu ates and students.

Conclusions

Our study presents the most recent and geo graphically comprehensive analysis of the vol ume and structure of potential migration from East to West prior to the EU enlargement in 2004 and the possible further figures on the additional candidate countries. The results contribute to the current political and scienti fic debate on the new East-West migration puz zle. What are the relevant facts ?

a Volume

The results of our study concur with those of most methodologically acceptable micro and macro-studies that the EU should not expect a tidal wave of emigrants from Eastern European and Mediterranean accession and candidate countries. All the indicators point to a volume similar to what was experienced after the Southern EU enlargement in the 1980s.

- The volume of the narrow migration poten tial is likely to be around 1.0% of the popula tion stock above the age of fifteen of the ACC-13 and AC-10 in the next five years un der the condition of free movement.

- The largest but relatively unrealistic migra tion potential is from 3 to 4.5% in the same time span.

- The predictions of our study for the 10 Cen tral and Eastern candidate countries exclud ing the three Mediterranean countries con cur with the latest results of a European Commission econometric study predicting a gross migration of around 1,100,000 to

1,500,000 in five years.

- In view of the existing legal barriers to free movement and the sequential reduction in

(15)

the next three to five years in the current member states, the predicted gross migration will be significantly under a million up to the end of 2006.

The predicted volume varies significantly how ever among the 13 accession and candidate countries. According to our study, the highest migration potential in the ACC-13 is in Bulgar ia and Romania. Turkish respondents are much less likely to consider moving to Europe with a narrow migration potential of less than 0.5% and a wide migration potential of 6%. The figures for all the larger accession coun tries in the narrow band are around 1 % in the next five years.

b Structure

Migrants from the accession and candidate countries are younger, increasingly female, better educated or college or university stu dents and increasingly single people who live alone. The sending countries are facing the prospect of a major youth drain in a range of 2 to 5% of the youngest cohort, which may lead to an outflow in the youngest age group of nearly 10% in the next five years in countries like Bulgaria and Romania.

- The potential youth drain is combined with a potential brain drain. The sending coun tries are in danger of losing from 3 to 5% of their college and university graduates and more than 10% of their students. The net re sults may be a brain drain of 2 to 3% in both groups in the next five years.

- The study confirms newer hypotheses of an increasing feminisation of migration. Within the narrow migration potential, women ac count for 40 to 45% on the average. In larger countries like Poland and Bulgaria and smal ler ones like Latvia and Cyprus, the percen tage of women is over 50% . There are clear signs of an accelerating trend over time. - Unemployment has an influence on migra

tion, but to a much smaller extent than pre dicted and only in a limited number of coun tries.

- Overall 2 to 3% of the unemployed people want to migrate.

- Married people are less willing to migrate. The main migration potential entails single people.

cType o f migration

The IOM study (1999) makes it possible to dis tinguish between potential permanent migra tion and various types of temporary migration. The overall message is that in the accession and candidate countries, most people want to migrate temporarily to earn more money, and only a minority want to migrate for good. Ac cording to IOM, the percentage of permanent migrants in the overall migration potential - excluding very short stays - ranges from 21 to 45% and averages around a third.26 Transposed to the migration potential in the present study, the volume of permanent migration in the 13 accession and candidate countries would roughly be:

- narrow permanent migration potential of 0.3%;

— widest permanent migration potential of 1 tol.5%.

Following this logic, the absolute increase in permanent migrants from the ten Eastern Eur opean candidate countries27 would be be tween 1,300,000 (widest potential) and 300,000 (narrow potential).

From a policy perspective, three angles should be considered. (1) It is important to ana lyse possible policy implications for different political units, i.e. the old member states, the accession countries and the three remaining candidate countries. The central questions are: do the more serious economic and social con sequences of migration occur in the receiving or the sending countries? Is there a difference between sending countries close to EU mem bership and sending countries whose pros pects of membership are still uncertain? (2) Five important characteristics of the migration process are relevant to the political reflections: volume, structure of migrants, type of migra tion, dynamics over time, and regional distri bution in target areas. (3) The main character istics of the expected migration need to be as sessed within a specific policy focus on labour market, social cohesion, social inclusion and demographic issues.28

Notes

1 The 3% estimate over fifteen years is also con firmed by Straubhaar (2001a).

(16)

Potential East-West Migration after Enlargement 2 A two-year period during which national meas

ures are to be applied by current member states to new member states. Depending on how liber al the measures are, they may result in full la bour market access. After this period, reviews are to be held, i.e. an automatic review before the end of the second year after accession and a further review at the request of new member state. The procedure includes a report by the Commission, but leaves the decision on whether to apply the acquis up to the current member state. In principle, transitional arrangements should be ended after five years, but may be pro longed for a further two years in current member states where there would be serious disturbances of the labour market or a threat of disruptions. 3 This article is based on a publication of the Eu

ropean Foundation (Krieger 2004).

4 Myrdal (1956) and Wallerstein (1974) view migra tion as a phenomenon that increases the devel opmental differences between countries. They highlight the importance of increasing econom ics of scale, resulting in a self-sustaining agglom eration process until economics of scale cease to increase or the costs for agglomeration are too high for people to migrate. According to this ap proach, the development of huge economic core areas is the result of an interplay of increasing economics of scale, people's willingness and ability to cluster in core areas, and the existing costs of transport and transaction.

5 The analysis of the Eurobarometer was con ducted by the European Foundation.

6 Migration has always been an interdisciplinary subject with contributions in sociology, econom ics, demography, political science, ethnology, so cial psychology and other fields (Kalter 1997). 7 The last question has attracted a great deal of at

tention in micro-theoretical contributions on migration since it presents a challenge to the dominant models of rational choice. 'The biggest problem with the application of a cost-benefit model to human migration may not be the cru deness of the actual calculation, but the fact that many people never make any calculation at all. A great number of non-migrants we interviewed appear to have never given any serious consid eration to the thought of moving anywhere' (Speare, 1971: 130).

8 One important objective in this context is to de velop a general theory of migration and integrate it into a general theory of human behaviour. 9 Here, demand factors in developed economies

as well as supply factors in peripheral economies are important in explaining regional mobility. 10 It defines social actors as Resourceful, Re

stricted, Evaluating, Expecting, Maximizing, Man.

11 SEU stands for subjective expected utility. 12 An excellent example is provided in the Interna

tional Organisation for Migration study (IOM 1999).

13 Enforced migration is based on political, social or ethnic discrimination or the effects of wars or natural disasters.

14 A good example of this type of approach can be found in Kalter (1997).

15 Others refer to semi-permanent migration as trans-national m obility or trans-national social spaces (Faist 2000, Kalter 2003). In this case mi grants maintain multiple ties with their country of origin sometimes commuting frequently be tween both contexts, thus organising their life in a bi- or multi-local respectively trans-national way.

16 The dimension w ithin the sam e city, town or village has not been included.

17 This approach differs from other studies, where a narrow labour market perspective has been chosen. Consequently those studies base their estimate of migration potential on a population sample between the ages of fifteen and sixty- five.

18 Criticism of this kind of methodological ap proach can be found in Fassmann and Muenz (2002: 61 ff.).

19 The following calculation was used, based on ta bles 7.10 and 7.11 (Bruecker et al. 2001: 104-105): 1,990,000 (stock 2005) plus 250,000 (flow 2006) adds up to 2,240,000 by the end of 2006.

20 A more detailed analysis that considers gender, employment and marital status is given by Krie ger (2004: 19-33). According to this report, un employment is not the main driving force be hind migration.

21 The second youngest age group exhibits a differ ent pattern. Turkey is the only country where this group represents the dominant age group. High percentages are also shown in Latvia and Lithuania. Only Turkey (5%) and Slovakia (6%) have a significant percentage of potential mi grants in the oldest age group.

22 A particularly low percentage of poorly educated migrants come from Cyprus, Lithuania and Slo vakia. In the AC-10, Hungary is the only excep tion with a relatively high percentage of migrants with only a primary school education (21%). 23 More details on the expected brain drain can be

found in Carrington and Detragiache's report for the IMF (1998).

24 Olesen (2002: 9-12), however, also analyses the potential of return migration, combined with a possible brain gain for the original sending countries.

25 This figure has been calculated as follows: the potential migrants are defined as respondents who want to migrate for several months. Then the percentage of respondents is taken who want to migrate for good in relation to the shorter- term migrants.

26 The range between firm intention and general inclination is from 1.2 to 3.7%.

27 A more in-depth policy analysis in presented in Krieger (2004: 67-70).