German Institute for Economic Research. and EPRC EUROPEAN POLICIES RESEARCH CENTRE. The Impact of EU Enlargement on Cohesion

Transcription

1 German Institute for Economic Research and EPRC EUROPEAN POLICIES RESEARCH CENTRE The Impact of EU Enlargement on Cohesion European Commission Tender No. PO/00-1/RegioA4 The authors gratefully acknowledge financial support from the European Commission, in the context of the Second Cohesion Report and Second Cohesion Forum. Nevertheless the opinions remain those of the authors, and do not necessarily represent Commission views Christian Weise, John Bachtler, Ruth Downes, Irene McMaster, Kathleen Toepel FINAL REPORT Berlin and Glasgow, March 2001

2 DIW 2 EPRC

3 Structure List of Tables... 6 List of Figures... 7 List of Annexes... 7 Glossary of Terms... 7 EXECUTIVE SUMMARY Enlargement and Cohesion: Concepts and Definitions Introduction EU Enlargement and Cohesion Policy Objectives Indicators of Cohesion Aims and Approach of the Study Economic Characteristics of Enlargement Enlargement: Just a Political Project or Deepening of Economic Integration? Possible Impact of Enlargement Alternatives to Enlargement? Economic and Social Cohesion in the Candidate Countries and European Union: Analysis of Disparities Socio-Economic Context at the National Level Macro-economic Developments Labour Market Social Inequality Environment Patterns of Regional Disparity in CEE and EU Countries General Overview Population Density GDP per Head DIW 3 EPRC

4 3.2.4 Unemployment Employment Structure Types of Regional Problems in the CEECs Capital Cities and Major Urban Agglomerations CEE: EU Border Regions Peripheral Regions: Eastern Border and Rural Old Industrial Regions Types of Regions in Applicant Countries and the EU: Cluster Analysis Island Economies Malta Cyprus Effects of Enlargement: Critical Issues Trade Trade Volume Trade Structure Regional and Social Impact Outlook and Conclusion Investment Volume of FDI Allocation of FDI Role of FDI Impact of Increased FDI to CEECs Conclusion Migration Migration Flows Regional and Social Impact Types of Migration DIW 4 EPRC

5 4.3.4 The Cases of Germany and Austria Border Regions Projected Patterns of Disparity Objectives Approach Determinants and Alternative Development Paths of National GDP Patterns for Regional Convergence in the Applicant Countries Results Summary and Perspectives Introduction Economic Characteristics of Enlargement Economic and Social Cohesion in the Candidate Countries and European Union: Analysis of Disparities Effects of Enlargement: Critical Issues Projected Patterns of Disparity Prospects and policy issues References DIW 5 EPRC

6 List of Tables Table 3.1.1: GDP growth in Applicants and selected EU Members (annual %) Table 3.1.2: GDP in Applicants and selected EU Members, 1989 = 100 Table 3.1.3: Gross domestic investment in Applicants and selected EU Members (% of GDP) Table 3.1.4: Inflation in Applicants and selected EU Members, consumer prices (annual %) Table 3.1.5: Labour force in Applicants and selected EU Members, total, 1990 = 100 Table 3.1.6: Unemployment in Applicants and selected EU Members, total (% of total labour force) Table 3.1.7: Female Unemployment in Applicants and sel. EU Members (% of female labour force) Table 3.1.8: Male Unemployment in Applicants and sel. EU Members (% of male labour force) Table 3.2.1: Inner-National Disparities in Applicant Countries (GDP) Table 3.2.2: Inner-National Disparities in Applicant Countries (Unemployment) Table 3.3.1: Regional Disparities at the EU - CEEC - Border Table 4.1.1: EU Member States Exports, Imports and Net-Exports with AC 12 Table 4.1.2: EU 15 Exports, Imports and Net-Exports with individual Applicants Table 4.1.3: Trade data of the EU 15, 1998 Table 4.1.4: EU exports and imports with selected Applicants by selected product group Table 4.1.5: EU 15 Imports from selected Applicants, sum of product group No. 84=100 Table 4.1.6: Regional Trade Data for Poland Table 4.2.1: FDI Stocks of EU members in CEECs, 1997, Assets Table 4.2.2: FDI Stocks of EU members in CEECs, 1997, Assets, EU 15 = 100 Table 4.2.3: FDI Flows from EU members to CEECs, 1998, EU 15=100 Table 4.2.4: FDI Flows from CEECs to EU, 1998, in Mio Euro Table 4.2.5: FDI Stocks of EU members in CEECs, 1997, Assets, World = 100 Table 4.2.6: FDI Flows from EU to CEECs, 1998 Table 4.3.1: Distribution of Immigrants from Central and East European Countries among the Member States of the EU Table 4.3.2: Projections for the Growth of the Population of Citizens of the Central and East European Candidate Countries Resident in the EU Table 4.3.3: Projections for the Stock Population of Citizens of the Central and East European Candidate Countries Resident in the EU Table 4.4.1: Shares of Exports of the New German Bundesländer as a percentage of German Exports Table 5.1: Scenarios of Regional Convergence: Summary of Main Results for 2030 DIW 6 EPRC

7 List of Figures Figure 4.1.1: Germany: Regional Structure of Trade with the CEECs 1997 Figure 4.3.1: Germany: Regional Distribution of Employees from the CEECs List of Annexes Annex 1: Maps on Regional Disparities in the EU and the Applicant Countries Annex 2: Summary Table of Regional Indicators in the EU and the Applicant Countries (NUTS II level) Annex 3: Clusters of Regions of the EU and the Applicants Annex 4: Scenarios of long-run National GDP Growth in the Applicant Countries Annex 5: Scenarios of long-run Regional GDP Growth in the Applicant Countries Glossary of Terms AC CEE CMEA FTA IIT NUTS PPP SME SMP Applicant Countries Central Eastern Europe Council for Mutual Economic Assistance Free Trade Area Intra-Industry Trade Nomenclature des Unités Territoriales Statistiques Purchasing Power Parities Small and Medium Sized Enterprise Single Market Programme DIW 7 EPRC

8 DIW 8 EPRC

9 The Impact of EU Enlargement on Cohesion Preparation of the Second Report on Economic and Social Cohesion, Study Area 11 commissioned by the European Commission, DG Regio FINAL REPORT DIW, German Institute for Economic Research, and EPRC, European Policies Research Centre EXECUTIVE SUMMARY Berlin and Glasgow, March 2001 It is a common view that enlargement poses a severe challenge for EU structural and cohesion policies. Far less clear and uncontroversial, however, is the empirical and analytical basis for that statement. Three broad questions need to be addressed: (1) What is the current state of economic and social cohesion in the applicant countries and how will, as a consequence, the situation in a future EU 27 differ from that in the current EU 15? (2) How will enlargement itself affect cohesion via the expected intensification of economic integration? (3) How long will EU structural policy have to deal with the challenges of enlargement? A detailed analysis of these questions is the purpose of the present study. Cohesion can be interpreted in various ways: a level of stability or a process of convergence; specifically in terms of income or unemployment levels, or elastically to encompass employment opportunities and living standards. EU structural and cohesion policies give primacy to two measures: GDP per capita and unemployment rates. This report interprets national and regional cohesion as meaning the degree of disparity in GDP per capita in PPP. Social cohesion refers to the exclu- DIW 9 EPRC

10 sion/inclusion of sections of the population from/in the labour market as well as to poverty. Enlargement could result in differing scenarios for economic and social cohesion. A decisive factor seems to be the preparedness for structural change in all members of the enlarged EU. A delay of enlargement due to pessimistic expectations would probably contribute decisively to the realisation of these expectations and would harm rather than protect less competitive sectors of EU economies. Analysis of existing disparities Despite recent growth rates above the EU 15 average in the CEECs, economic convergence remains limited. Poland, Slovenia, Hungary and the Czech and Slovak Republics overall display the most positive macro-economic indicators. Considerable labour market changes have occurred associated with the processes of economic restructuring, privatisation and liberalisation. These include a sharp fall in industrial employment and a considerable rise in service sector employment, but there remain substantial differences with the employment structure of the EU Member States. Unemployment has risen in all CEE countries to varying extents. Income levels and standards of living have declined and poverty has spread considerably (with variation between countries and a disproportional effect on certain social groups). Rapid industrialisation, inefficient raw material extraction, obsolete technology and a lack of environmental controls have left a legacy of environmental degradation with significant spatial differentiation. While reduction in pollution levels is evident, the costs of clean-up are still extremely high. The spread of sub-national disparities (in GDP and unemployment) in the CEECs is smaller than in some EU Member States. Disparity patterns (at NUTS II level) include the following: GDP per capita in CEE regions is considerably less than the EU average only Prague and Bratislava lie above this level; regional unemployment is relatively low in CEE in comparison to the EU, with considerable sub-national variation (but again less than in EU Member States); CEE regions are in general more sparsely populated than in the EU; and agriculture dominates regional employment structures in some CEECs (eg. Romania and Poland) to a much greater extent than in the EU. DIW 10 EPRC

11 The types of regional problems in CEE reflect both the unique process of transition, as well as structural changes already undertaken in Western countries but delayed in CEE by geo-political factors. Overall groupings include: Capital cities/major urban agglomerations which demonstrate the most favourable economic indicators, benefiting from high investment, skilled labour and training facilities, better infrastructure, business services and access to decision-makers; Western border regions which benefited from proximity to the EU, encouraging investment, trade, tourism and educational/technological initiatives; Peripheral eastern and rural regions which are among the most economically disadvantaged in CEE. Geographical location, poor infrastructure, low investment, declining agriculture and rural out-migration are all contributory factors. These regions have particularly high rates of unemployment; Old industrial regions, the drivers of economic activity under socialism, which have been particularly negatively affected by privatisation, enterprise restructuring/closures, subsidy loss and market re-orientation. Problems include unemployment, lack of entrepreneurship and environmental decline. A cluster analysis was conducted to classify all ca. 260 EU and AC regions simultaneously in types of regions according to their employment structure and population density. This produced six clusters: Agglomerations; Service dominated; Service biased; Industry; Agriculture biased; Agriculture dominated. The distribution of the regions among the clusters shows the very poor development of the service sector and the importance of agriculture in the transition countries compared to the EU 15. Industry plays a dominant role for employment only in a small number of the AC regions. This economic structure of the AC regions is noteworthy because, in general, regions with above-average GDP per head are more likely to be found in the agglomeration or service clusters than in the industry cluster. An agricultural bias is clearly associated with a low per capita GDP. Labour market problems, however, are not obviously associated with specific clusters. Effects of Enlargement During the 1990 s, the CEECs have managed to re-direct their exports away from the former COMECON members towards the European Union. The trade volume has increased significantly, and the EU has become the most important trading partner of DIW 11 EPRC

12 the CEECs. From the point of view of the EU, the AC are much less important partners. Geographical proximity seems to play a key role in determining bilateral trade flows. Regional trade data available indicate that this pattern also applies at the regional level. CEECs have been able to change the commodity structure of their exports from inter-industry to intra-industry trade but still export mainly products with comparatively low unit values. There is no indication that the CEECs constitute a severe competition for the EU cohesion countries or other EU members. As in the case of trade, recent years have seen a marked increase in FDI flows from the EU to the AC, dominated by the main trading countries but also by France and the Netherlands. Contrary to the situation with regard to trade, there are practically no FDI flows from the AC to the EU. Migration is often cited as the most important post-enlargement effect with automatically associated negative consequences for EU members. But more diligent analyses do not expect a massive influx of migrants after enlargement and see only minor - and by no means necessarily negative - effects on wage and employment in the EU. Migration flows will be directed mostly into Germany and Austria. Inside these countries, they will be directed to centres of economic activity, not necessarily to the border regions. However, the inflow will not be as excessively high as sometimes expected and it will slow down over time. Border regions are potentially most affected by enlargement, accentuating internal disparities inside these regions with both, positive and negative, possible effects. Competitive enterprises, sectors and areas will gain from the proximity of new markets and the supply of a wider selection of inputs. Less competitive ones will suffer from increased competition. Along the EU:CEEC border the impact will most likely be concentrated on the eastern Austrian regions. The impact is not necessarily negative on balance but the adjustment pressure will be highest here. Challenges for support policies The analysis of the challenges to support policies has to start with an assessment of the long-run perspectives for convergence of GDP per head in an EU 27. First, a plausible range for convergence of national GDP per head was derived from the experience of post-war developments in 21 European market economies. A convergence rate of 2 %, i.e. reducing the income gap by 50 % in 35 years, seems to be DIW 12 EPRC

13 most plausible. The ten applicants from CEE would then achieve a per capita GDP of approx. 65 % of the EU 15 average in 2037 (starting from 38 % in 1997). This analysis was complemented by the calculation of eight different scenarios for convergence of regional GDP per head not to identify the most plausible one but rather to cover a wide range of reasonably possible developments. Even in the most optimistic scenario, only 14 of the 51 regions of the applicants would achieve a per capita GDP of more than 75 % of the EU 15 average in The empirical work in this study has shown that enlargement will not pose a severe challenge for today's EU regions. Development problems of the applicants, however, are severe and will persist. Given the severity of the regional development problems in an enlarged EU and the duration of very low standards of living in most of the applicants (which are to be expected under reasonable assumptions), the options for a reform of EU structural policies have to be discussed. DIW 13 EPRC

14 DIW 14 EPRC

15 1 Enlargement and Cohesion: Concepts and Definitions 1.1 Introduction The invitation of the EU to the Central and Eastern Europe countries (CEECs) to join the European Union upon compliance with certain conditions has been part of EU policy since the European Council of Copenhagen in June At the time of the publication of the First Cohesion Report the transition countries had formally applied for membership but substantial negotiations had not begun yet. Since then, significant progress has taken place. The EU has offered the formal status of a candidate country to all applicant countries from Central and Eastern Europe (CEE) as well as to Cyprus, Malta and Turkey in two steps in Luxembourg (December 1997) and in Helsinki (December 1999). With the notable exception of the case of Turkey, formal accession negotiations with all applicant countries (AC) have now begun first with the so-called Luxembourg Group (Cyprus, Czech Republic, Estonia, Hungary, Poland and Slovenia), then with the Helsinki Group (Bulgaria, Latvia, Lithuania, Malta, Romania and Slovak Republic). While negotiations with the individual countries have made varied degrees of progress, some of the applicants had reached an advanced stage by Summer Parallel to the enlargement negotiations, the EU has had to begin preparing its own policies for the inclusion of new Member States almost all of which have considerably lower figures for per capita GDP (relative to the EU average) than any other new or old EU Member State. Hitherto, the main response of the EU to this challenge has included the reform of the EU structural and agricultural policies in March 1999 (Agenda 2000) and the development of specific support programmes for the applicant countries (ISPA, SAPARD etc.). However, the more real the prospect of enlargement becomes, the more important and concrete becomes the question of the impact of enlargement on cohesion in an enlarged EU. Work on this report was carried out by Christian Weise and Kathleen Toepel (DIW Berlin, German Institute for Economic Research) and John Bachtler, Ruth Downes and Irene McMaster (EPRC, European Policies Research Centre, Strathclyde/Glasgow); Annex 4 was provided by Herbert Brücker (DIW Berlin). The authors are grateful to the Commission officials at DG Regio, Eurostat and other EU offices for their support during the preparation of this study. In addition, thanks are due to DIW 15 EPRC

16 research assistants and secretarial staff at DIW and EPRC. This report was funded by the European Commission in the context of the Second Cohesion Report. However, the opinions expressed are those of the authors and do not necessarily represent the Commission's views. 1.2 EU Enlargement and Cohesion At the heart of the debate over the challenges of EU enlargement is economic and social cohesion in a wider Union. Cohesion is an important pillar of the European social market economy, it underpins EU action in the field of regional development and it will take on greater political, economic and social significance in an enlarged EU given the relative underdevelopment of the CEE accession countries. The importance accorded to cohesion derives from the belief that solidarity and mutual support are an equally important basis for progress [as market forces], not only for social reasons but also for optimising overall economic benefits since there is ample evidence of detrimental effects of inequality of growth (CEC, 1996). This commitment to territorial and social justice provides the rationale for the EU Structural Funds and the Cohesion Fund as well as the pre-accession instruments, ISPA and SAPARD. While there is a clear political commitment to economic and social cohesion at EU and national levels, the architecture of future policies is not clear. Several issues need to be taken into account. First, the impact of enlargement on cohesion is still speculative, in particular because reliable sub-national data for the CEECs are only now starting to become available. Second, the size and diversity of an enlarged EU requires a fundamental reappraisal of the rationale and objectives of policies to address economic and social cohesion. Third, the scope for EU intervention will be influenced by the willingness of the EU 15 to commit financial resources (the size of the structural policy budget) and their preparedness to forego the aid provided to current recipient regions (the criteria for allocating the budget). Fourth, the relationship between EU and national policies in the field of regional development is changing, affecting the scope for current and future Member States to implement their own regional policies. Greater coherence is driven partly by regulation (Structural Fund reform, EC regional aid guidelines) and partly by a convergence in thinking about strategies for economic and social cohesion, but the relationship is still uneasy. Lastly, it is becoming increasingly recognised that effective delivery of both EU and DIW 16 EPRC

17 national policy intervention in regional development requires significant investment in institutional capacity at national, regional and local levels. Cohesion is not a straightforward concept and can be interpreted in various ways. For some, it implies a level of stability in territorial and social relations; for others, it involves a process of convergence in disparities between regions and social groups. In some cases, it is defined specifically in terms of income levels or unemployment rates; it is also used more elastically to encompass access to employment opportunities and desirable living standards. Further, the use of the term is associated with very different policy choices. In some territories, cohesion is addressed through policy objectives of equalising regional and social differences through an explicit redistribution of growth, employment etc.; in others, policy is oriented towards maximising the contribution of regions and social groups to national economic efficiency. 1.3 Policy Objectives The rationale for regional policy is the key determinant of the choice and use of indicators, and the timescales and spatial scales used for analysing cohesion. The objectives of cohesion differ greatly between countries and between Member States and the EU. As is well known, regional policies have been introduced for a mix of economic motives (e.g. utilisation of production factors, congestion costs), social factors (commitment to full employment, welfare considerations), environmental arguments (e.g. over-crowding, pollution in congested areas) and political reasons (consequences of disparities for voting patterns) (Vanhove, 1999). The objectives of regional policy are often discussed as a trade-off between: aggregate national efficiency, involving a more efficient allocation of regional resources to maximise net national benefit; and inter-regional equity, involving a more equal distribution of income, employment or infrastructure over space (Bachtler, 2000). Over the long term, EU countries tended to introduce regional policies for reasons of equity but have progressively given greater priority to efficiency since the mid-1970s. Regional policy goals are increasingly concerned with optimising the contribution of regional resources to the creation of economic growth by promoting competitiveness and reducing unemployment (Prud homme, 1994). This is true of smaller EU countries where regional differences are comparatively small (Austria, Denmark, Netherlands, Switzerland), as well as larger Member states (UK, France) which have suffered relatively high, nation-wide unemployment for much of the past 20 years and DIW 17 EPRC

18 which have extensive areas experiencing deep-seated industrial decline and social problems. However, the equity goals of regional policy still exist: the historic concern with spatial equality as an objective of regional development remains to some degree in the Nordic countries, France (in part) and in Germany, where the aims of regional policy continue to advocate the reduction of inter-regional disparities in relation to income generation and employment opportunities (Yuill/Bachtler/Wishlade, 1999). The efficiency-equity trade-off is particularly difficult for the Cohesion countries which have to avoid jeopardising national efficiency in the interests of economic growth while at the same time channelling resources to less-favoured regions. From the perspective of CEECs, these issues are highly pertinent. While policies to address economic and social cohesion were not a priority for CEECs in the early years of transition, the challenge of rising territorial and social inequality is forcing regional policy onto the agenda of CEE governments, encouraged by the preaccession aid requirements of the EU. However, the CEECs face difficult questions: when the performance of a country is relatively poor, should resources be focused on particular regions? What is the balance between the need for national economic growth and the social exclusion of sections of the population? At what level should cohesion be addressed? What is the balance between economic and political considerations? In short, CEE policymakers have to strike a balance between internal cohesion within national boundaries and external cohesion from a European perspective, i.e. how to reduce the economic gap with the EU while addressing internal disparities and social inequality. 1.4 Indicators of Cohesion Across the EU, there are different approaches to identifying areas for cohesion policies. This reflects the perceptions of cohesion, approaches to selecting areas for policy intervention, methodological preferences and political perspectives. At Member State level, the types of indicator for measuring cohesion fall into three main groups (Wishlade/Yuill, 1997). First, there are physical indicators (those associated with geographical or natural conditions) which are particularly relevant in the Nordic Member States where peripherality and sparsely populated areas are a central preoccupation of regional policy. In addition, the Cohesion countries, and to some extent Italy, have problems associated with peripherality. Regional policies in France, Germany, the UK and Austria also give some recognition to physical factors. Second, DIW 18 EPRC

19 concern with economic disparities is most evident in the four Cohesion countries, as well as in Germany and Italy. Belgium, Finland and the Netherlands all utilise GDP per head as a measure of regional prosperity. Most of these countries, plus Germany, France, Sweden and the UK are also concerned with regional differences in employment structure. Germany and Portugal take account of the level of amenity or infrastructure. Many countries consider the potential impact of demographic trends. Third, there are measures of cohesion relating to social inequalities, notably differences in living conditions and income. In Germany and Spain, there is an explicit constitutional commitment to equalising standards of living. Elsewhere, social disparities are primarily concerned with (un)employment, especially in the UK. Apart from Greece, Portugal and Sweden, all Member States use unemployment rates for measuring cohesion to varying extents. Other measures include future employment trends, the economically active population, levels of personal income and quality of the labour force. At the EU level, the choices of cohesion indicators are more restricted. Methodological difficulties are compounded by the technical problems of comparative analysis across countries and regions. For EU policy purposes, economic development (measured in terms of per capita GDP) largely determines access to the Structural Funds (most notably through Objective 1) and is a key consideration in EC competition policy reviews of regional aid maps. Also, EU regional policy and competition policy both use unemployment rates as an important measure of cohesion as well as employment trends (for the Structural Funds). The features of EU regional socio-economic patterns are well-known: regional disparities across the EU are wide by international standards, and there is a significant core-periphery disparity in regional GDP per head and (partly) in regional unemployment. According to EU data, the gaps in GDP per head between the cohesion countries and the rest of the EU tended to narrow over the period, with an improvement in the GDP per head of the four countries from 65 % of the EU average in 1986 to almost 77 % in 1996 and forecast to rise to 78 % in 1999 (CEC, 1996). Such conclusions about cohesion are, however, sensitive to the time period and spatial scale used for analysis. Over the long term, several studies have found a narrowing of regional disparities at different levels up to 1973 but thereafter a freezing (or even widening) of disparities from the 1970s to early 1990s (Dunford/Smith, 1998; DIW 19 EPRC

20 Pettenati/Canullo, 1994). Other research on trends in disparities (Armstrong/Vickerman, 1995; Dignan, 1995) suggests some convergence in EU regional differences from the mid-1980s to the early 1990s, but with a convergence trend that was weak, halting and mainly due to convergence between some peripheral countries and the rest of the EU. Analyses of trends in regional unemployment also find limited evidence of any widespread convergence of regional unemployment rates over the period from the early 1980s to early 1990s. Indeed, it has been argued that European regional unemployment disparities are in fact characterised by an equilibrium state of persistent high disparities in regional unemployment (Baddely/Martin/Tyler, 1998). EC research shows that the 25 regions with the lowest unemployment rates were much the same in the late 1990s as ten years previously, while rates in the most affected regions increased from 20 % to almost 24 % (CEC, 1996). At a different spatial scale, regional differences within countries appear to have been diverging, although there is great variation between Member States and little consensus among the research undertaken. Research on EU disparities is, of course, affected by spatial scale. For example, the choice and size of spatial units alters the measurement of GDP. At high levels of spatial disaggregation, disparities in levels of GDP per head increase, while conversely, high levels of aggregation lead to differences between areas being averaged out. The spatial unit used is also relevant, since the use of non-functional units can lead to a separation of centres of economic activity from their prosperous commuter belts. The NUTS classification used by Eurostat is associated with all of these problems, since it is based on administrative boundaries and involves areas of greatly differing size, population and population density, influencing the genuine comparability of statistical indicators across the EU (Wishlade/Yuill, 1997). There is still scope for improvement in comparability, a fact recognised by the European Commission which is preparing a regulation on NUTS to set standards. With respect to the indicators themselves, measuring regional GDP per capita is associated with several important methodological challenges, including the difficulties in assigning output where activities cross regional boundaries or where the income accrues from off-shore natural resources, and the effects of government transfers and the size of the black economy on the composition of regional GDP. Using purchasing power parities (PPP) to take account of the cost of living in the measurement of regional GDP per head can make a substantial difference to the position of regions DIW 20 EPRC

21 across the EU, making some regions appear poorer and others richer, but overall reducing the disparities between regions. The use of population as the denominator may also give a partial view of economic development compared, for example, to the size of the regional labour force. Methodological difficulties are also associated with the use of regional unemployment rates, notably the limitations on comparability of regional unemployment between countries due to the differences between national labour markets and the varied determinants of unemployment rates (Wishlade/Yuill, 1997). Bearing in mind these caveats, this report interprets national and regional cohesion as meaning the degree of disparity in GDP per capita (measured in purchasing power parities). Social cohesion refers to the exclusion/inclusion of sections of the population in the labour market, captured principally by various unemployment indicators (official rate of unemployment, long-term and youth unemployment etc.). Social cohesion, as proposed, mainly depends on the overall growth and the development of national and regional cohesion. National cohesion refers to the 15 Member States of the current EU, the ten applicants from Central and Eastern Europe, and the island economies of Malta and Cyprus. Regional cohesion refers to the NUTS II level wherever applicable. 1.5 Aims and Approach of the Study The aim of this study is to provide a detailed examination of the possible consequences of enlargement for cohesion in the current EU Member States and in the candidate countries to support the preparation of the Second Report on Economic and Social Cohesion in the EU. The study will clarify the main channels through which enlargement will affect national, regional and social cohesion in Europe, identify the main similarities and contrasts between (relevant groupings of) regions and regional developments in Western and Eastern Europe, and produce a picture of the possible consequences of enlargement for regional disparities and social cohesion, e.g. employment opportunities and unemployment risks. The study is structured as follows. The first step has been to clarify the terms and concepts used. As regards cohesion, this was done in Chapter 1. Discussions on the possible impact of enlargement in East and West are often confused and overburdened if each and any development in the transition countries or in East-West DIW 21 EPRC

22 economic and political relations is attributed to the envisaged inclusion of these countries in the EU. Chapter 2, therefore, discusses the specific aspects that distinguish enlargement from transition and economic integration of the CEECs in the international division of labour that might be expected without the prospect of becoming an EU Member State. In addition, different scenarios of the impact enlargement might have on the economic and social prospects of the (enlarged) EU are developed. The topic impact of enlargement on cohesion can be interpreted in at least three different ways. First, it has to be shown how different intra-eu disparities will be in an enlarged EU as compared to the present EU 15. Chapter 3 provides a detailed analysis of the disparities in the applicant countries and of the range of new development problems associated with the inclusion of these countries in the EU. Second, in addition to this static approach, it has to be asked whether enlargement will ameliorate or worsen inner-eu disparities in a dynamic sense. Chapter 4 tackles this question by an analysis of trade, investment flows and migration between the EU and the applicants and develops a judgement of the possible regional implications of an intensified economic integration. Third, the challenges of enlargement for cohesion policy deserve attention. Chapter 5 develops as a background to this question a plausible upper and lower boundary for the development of national convergence of the CEECs (to the EU average) and presents some plausible scenarios for regional convergence of AC regions. This demonstrates how much time is needed to achieve substantial cohesion. Chapter 6 gives a summary and concludes the study. Empirical work for this study has mostly had to rely on data available from the European Commission. Regional data were provided mainly by DG Regio and Eurostat; data on trade and FDI came from the COMEXT database and Eurostat's EU Direct Investment Yearbook, respectively. The World Development Indicators of the World Bank have been used for the macroeconomic analysis. In addition, selected data came from the UN-ECE Economic Survey of Europe, the World Development Report, the UNDP Human Development Report, the Transition Report of the EBRD and Eurostat's Statistical Yearbook on Candidate and South-East European Countries. Migration figures are from the OECD and national sources. A systematic review of the home-pages of the applicant countries statistical offices brought mixed results. While some countries provide interesting data, others do not. As the empirical work needed to use comparable and reliable data, regional data DIW 22 EPRC

23 from national sources was only used for anecdotal support of the analysis of Eurostat data. However, special emphasis was placed on obtaining regional trade figures for German Federal States and Polish Provinces (see section 4.1). Data analysis had to work with data at NUTS II level. On the one hand, this is due to availability of data and to practical reasons of work efficiency. On the other hand, EU structural policy is directed (in its relevant elements) at the NUTS II level. However, as there are only approximately 50 NUTS II regions in the CEECs (and six applicant countries consist of just a single NUTS II region), these data necessarily sometimes lack detail. Therefore, the analysis on regional disparities in the CEECs (chapter 3) draws on published and unpublished studies that are more selective (and, necessarily, less comparable) but make use of more detailed data from various sources. An extensive literature search was undertaken in support of the data analysis on economic and social cohesion in the CEECs and the effects of enlargement. This involved a review of official publications, academic studies and the grey literature derived from a range of sources including the EC, European Parliament, OECD, EBRD, IMF, World Bank, UNECE, German Institute for Economic Research (DIW), Vienna Institute for International Economic Studies (WIIW), NEI, EastWest Institute, Centre for Economic Policy Research (CEPR), Bank of Finland and Rhine-Westphalia Institute for Economic Research (RWI). The literature review covered the following issues in particular European enlargement, trade and FDI, labour market and structural change, EU/CEE migration and regional development. The references cited are drawn mainly from the last 3-4 years given that the situation is highly dynamic and many texts are outdated relatively quickly. The literature review aims to provide a review of the main points raised in relation to economic and social cohesion in the CEECs and an extensive bibliographic reference list. DIW 23 EPRC

24 2 Economic Characteristics of Enlargement The last decade has seen dramatic changes in the political and economic systems of the applicant countries from CEE. The process of transforming socialist to market economies was associated by a liberalisation of the integration in the international division of labour, in general, and by developing closer economic and political relations with the EU, specifically. This process has resulted in more political freedom and much better opportunities for pursuing individual concepts of life for the people of the transition countries as well as in enhanced economic welfare for, at least, a substantial part of the population. For many, in the EU as well as in the AC, enlargement is a logical and almost inevitable next step to ensure smooth progress of the economic and political development in the AC and economic gains for all European economies. Not surprisingly, however, transition has not been without negative consequences for selected social groups and economic sectors in the AC. Accordingly, fears that EU membership might be associated with even more liberalisation and reform pressures leading to a worsening of cohesion in the new member states are a political reality. At the same time, the argument is raised in the EU that economic benefits of integration with the CEECs are already fully exploited and enlargement would only lead to limited political gains but immense economic burdens for the current members of the EU. To discuss these fundamental positions of enlargement supporters and sceptics with more clarity it seems to be decisive (1) to develop a clear understanding of how - from an economic point of view - membership of the Union differs from membership in a Free Trade Association as it is, by and large, already achieved between the EU and the applicants; (2) to evaluate how enlargement could either benefit or worsen cohesion in Europe and who might be affected; and (3) to check whether there are viable alternatives to enlargement of the EU. The aim of this section is to develop the main lines of arguments without trying to give a prognosis of the most likely outcome or presenting detailed empirical evidence. 2.1 Enlargement: Just a Political Project or Deepening of Economic Integration? According to integration theory, different levels of economic integration have to be distinguished. Creating a Free Trade Area (FTA) while maintaining individual tariffs is DIW 24 EPRC

25 the least intensive form of integration. The next step is the establishment of a Tariff Union that applies the same tariffs to all third countries and no restrictions to internal trade. Current EU members have achieved a much higher degree of integration with the Single Market and, for most of them, the introduction of the Euro. The Europe Agreements between the EU and those transition countries that apply for EU membership basically ensure a Free Trade Area (with some significant exceptions). EU membership, however, is in economic terms equivalent with inclusion in the Single Market and other Common Policies. This obviously has severe economic implications, the most important of them being, most probably: liberalisation of agricultural trade, a ban on any threat to use anti-dumping measures, a potential for more intensive trade with current EU members (today an EU membership dummy contributes significantly to the explanation of trade flows between industrial countries), a change in conditions for investments both from domestic or foreign sources (cancellation of a still existing risk premium due to uncertainty about future developments), free movement of labour, inclusion in transfer programmes of the EU, most notably the Structural Funds, enhanced political and economic stability and insurance against consequences of possible negative shocks in the future, applicability of the acquis communautaire, most notably in competition policy, social policy and environmental matters, and possible participation in the European Monetary Union (although the new member states do not have to take part necessarily). It has to be noted, however, that all these implications of membership (as opposed to 'mere' FTA) will take effect at different points of time that are not always easy to indicate. As in the case of Southern enlargement or SMP '92 it is highly likely that enterprises will take enlargement for granted, at the latest, when the negotiations are de facto concluded. First round effects will therefore begin to materialise prior to the official date of enlargement, i.e. they have, most probably, already begun to materialise. DIW 25 EPRC

26 2.2 Possible Impact of Enlargement The actual consequences of enlargement cannot be determined ex ante with certainty. The empirical evidence available today and the probability of some aspects of possible scenarios are analysed in later chapters of this study. Here, the objective is to outline how the various economic aspects of enlargement can either contribute to a favourable development or result in a more problematic situation regarding cohesion in the enlarged EU. Scenario 1: The yellow brick road to an integrated Europe An optimistic scenario would, most importantly, have to rely on the willingness and ability of all EU member states to accept the need for structural change. With regard to the four freedoms in the Single Market (free flows of goods, services, capital and persons) this would imply, among other things, a further liberalisation of the Common Agricultural Policy (rising share of income support payments according to social policy criteria -, rising share of co-financing by national governments, gradual reduction of support) accompanied by support for the re-structuring of rural areas (whether from national or Community sources). In general, trade volumes would rise significantly confirming past experience that trade between EU members is higher than can be explained by factors like GDP per head or distance alone. Enterprises from the AC would prove able to withstand competition from EU firms and develop market niches in EU core markets. The commodity structure of the applicants exports would become more and more similar to that of the EU members indicating much smaller and less concentrated import competition pressure in the EU. EU membership would signal a sustainable legal, political and economic framework for investors (domestic and foreign). Rising investment shares in GDP would contribute to higher growth rates and create new employment opportunities. This favourable economic situation would ameliorate internal cohesion in the applicants: On the one hand, economic growth opportunities would trickle down also to regions that are not immediately affected from enhanced investment and, on the other hand, intranational migration would become easier because of ameliorated labour market and housing conditions. International migration would not be a major topic as potential migrants see positive prospects in their home countries. DIW 26 EPRC

27 The applicability of the acquis communautaire would improve the framework conditions for competitive enterprises (competition policy) as well as for employees and citizens in general (social and environmental policy). Where the implementation of the acquis would be too costly for either the enterprises or the government, there is either Community support available or a transition rule applies. This scenario would require the allocation of sufficient structural funds transfers to the new members and their efficient implementation, i.e. the funds and national development measures are focused on increasing competitiveness of enterprises, human capital and attractiveness of regions for investment. In this scenario, the over-all implications for the current EU would be positive. The CEECs would constitute stable and growing neighbours that provide an interesting market for EU products and services as well as promising investment opportunities. This might also lead to a significant inflow of foreign capital into EU financial markets and rising indices at EU stock exchanges. The CEECs would become more and more competitive but because of the high diversification and the modernisation of their economies this competition pressure would not focus on specific sectors or regions of the EU. Net migration flows from East to West would be quite low. Scenario 2: Why (and how) things may go wrong Obviously, the consequences of enlargement might also be less positive. On the AC side, the political will to pursue reforms and structural change may slow down as soon as EU membership is achieved. Under the pressure of highly competitive enterprises from 'old' EU members, combined with an excessive burden implied by the EU acquis, slowly developing modern sectors in the CEE economies would not survive. CEECs would have to rely on current comparative advantages in resource- and labour-intensive production. These, too, would be hampered in their development by social and environmental standards of the EU. The trade volume would not rise significantly: The CEECs would not be a promising, growing market for EU 15 exports nor would CEECs exports be able to compete on EU 15 markets. Convergence of the CEE economies to the EU average as well as internal cohesion would slow down. Public funds would be used almost exclusively for subsidies and social cushioning. Structural funds and (limited) national resources would not be used efficiently either due to misguiding incentives or to a lack of monitoring and DIW 27 EPRC

28 evaluation. Governments would only pay lip service to the implementation of the acquis (especially in competition policy) and to the development of an efficient administration. Foreign investment would concentrate on selected regions and on the exploitation of labour costs advantages; the contribution to development would be meagre. There would not be any efforts to achieve a balanced regional development and, consequently, no trickle down to poorer regions. Due to unsatisfactory prospects for young (and especially well-educated) persons migration to the EU 15 would be substantial. On the EU side, the potential strength of the CEECs in agricultural products would be answered by de facto-discriminating special support for EU 15 farmers (which, among other things, leads to welfare losses for the EU 15 due to delayed WTO negotiations). There would be no growth impulses from CEE. Adjustment pressures would be concentrated on low-qualified employees or job seekers. They would suffer from import competition and from competition on the labour market resulting from a massive inflow of cheap labour from the CEE. In general, protectionist tendencies would gain ground at the expense of support for European integration. The EU would not only be weakened economically resulting, among other things, in an outflow of capital to, say, the USA and South East Asia but also politically. In this scenario, the repercussions on the EU would be negative. The more advanced EU regions would miss the opportunity to trade with a growing market in the East while the lagging regions of the EU and some rural areas would possibly come under competition pressure from CEEC suppliers. In addition, fierce battles on the distribution of structural funds would evolve. 2.3 Alternatives to Enlargement? As the possible outcomes of enlargement seem to cover a wide range, including some unpleasant developments, it might be seen as sensible to adopt a cautious 'wait and see'-approach and to monitor possible early indicators of enlargement success such as political stability and administrative capacity, trade volumes and commodity structure, investment motives etc. over a longer time horizon. Alternatives to a firm enlargement policy include a strategy of low EU commitment to enlargement with an intentional delay of accession negotiations or an open termination of negotiations DIW 28 EPRC

29 that might be re-orientated towards a membership in the European Economic Area. The political plausibility and potential costs of such strategies cannot be evaluated here. However, it is interesting to look at the economic impact of such an anti-monde on cohesion in the EU and the AC. The main immediate impact would most probably be a severe setback to economic growth in the applicants. Several studies have shown the potential gains from the implementation of the Single Market (see, e.g., Baldwin, 1989) but also that the periphery might be disadvantaged, at least temporarily, during the process of integration (see, e.g., Krugman/Venables, 1990). Excluding the AC from further integration or offer them integration without any support to develop their competitiveness would diminish their development prospects (under the assumption of an efficient use of EU transfers). Perhaps even more important seems to be the loss of credibility of the reform process in the CEE (see Piazolo, 1999). Analyses of the state of reform in the CEE suggest that, although much has already been achieved, there still is a distinction between the political systems of the transition countries and that of developed democracies and market economies (see, e.g., the ratings of the EBRD). Kaufmann et al. (1999a and b) have developed a method to aggregate various governance indicators in order to rank a sample of 160 countries according to their state of "voice and accountability", "political instability and violence", "government effectiveness", "regulatory burden", "rule of law" and "graft". The method is, according to the authors, still rather imprecise and should, therefore, only be used with caution when analysing the relative position of a specific country. However, it is instructive to note that almost all CEECs lay below the EU members (except Greece) in practically all categories. Only Hungary and Slovenia are close to Italy and Belgium, the worst-performing EU members in this analysis; they are followed by Poland and the Czech Republic. There are differing views on whether FDI in the CEECs will grow after enlargement or whether it might already have achieved its long-term level. However, it is quite clear that delayed or even cancelled enlargement would hamper investment conditions. It would result in potential FDI not being realised and/or FDI stocks being scaled down and/or FDI being directed to less intensive forms of co-operation like outwardprocessing (governed by low wages) instead of developing own R & D-capacities abroad. DIW 29 EPRC

30 As structural change will, in the anti-monde, be more difficult in the AC, import competition in the EU would concentrate on goods that have a comparatively low content of R & D and human-capital. Protectionist measures against these imports would not be an option because of the Europe Agreements (and because of the WTO membership of all transition countries in CEE). In addition, it has to be noticed that imports from the CEE in traditionally problematic sectors like textiles are often much smaller than those from newly industrial countries e.g. from South East Asia. Infrastructure as well as social policy and environmental issues would become worse in the AC because the acquis would not apply and there would be less resources available for necessary investments (see Garvey/Hager, 2000). Legal migration would be restricted. However, as push-factors for emigration would be much more pressing there would be a rising inflow of illegal migrants to the EU. DIW 30 EPRC

31 3 Economic and Social Cohesion in the Candidate Countries and European Union: Analysis of Disparities The emergence of spatial and social disparities occurs within a broader socioeconomic context and is influenced by macro-economic and wider social and environmental factors. The economic and social framework for both the individual citizen and the development prospects of specific regions are often dominated by decisions and developments at the national level. The CEE region, despite a common socialist heritage, is not economically or socially homogenous and, at national as well as regional levels, differences have become increasingly apparent during transformation as individual countries have made reform choices against unique historical, political and institutional backgrounds. While the focus of this section (and report) is on disparity and cohesion, it is important to highlight the context within which these spatial and social aspects are emerging. This chapter starts with an overview of key aspects of the socio-economic framework which forms the background to the spatial disparities subsequently dealt with in more detail. These sections are brief, as many of the issues are dealt with in more detail in other parts of the Second Cohesion Report and they are the context for the main focus of this study. The data analysis (unless otherwise stated) was carried out by the project team for the report and is presented in detail in the tables. The chapter then goes on to analyse in more detail the patterns of regional development in CEE and the EU 15 to provide a picture of cohesion within an enlarged Europe. Types of regional problems currently facing CEECs are then discussed further before clusters of similar regions from the EU and the AC are identified. The chapter concludes with a review of the two island economies of Malta and Cyprus which are in a different situation to the ten candidate countries of CEE. 3.1 Socio-Economic Context at the National Level Macro-economic Developments Growth rates in most of the transition countries have exceeded those of the majority of the EU 15 in the second half of the 1990s, but substantial convergence is still limited (see table 3.1.1). Poland and Slovenia are the only two candidate countries which, in 1998, had exceeded their pre-transition GDP level (see table 3.1.2). The DIW 31 EPRC

32 GDP of the Czech Republic, Hungary and the Slovak Republic had almost reached the level of 1989, while the other CEE candidate countries remain between % of the 1989 level. Latvia had achieved only 57 % of the 1989 level while in Bulgaria and Romania, sustainable recovery is yet to begin. Table GDP growth in Applicants and selected EU Members (annual %) Country Name Bulgaria Czech Republic Estonia Hungary Latvia Lithuania Poland Romania Slovak Republic Slovenia Cyprus Malta France Germany Italy United Kingdom Greece Ireland Portugal Spain Source: World Bank, World Development Indicators DIW 32 EPRC

33 Table GDP in Applicants and selected EU Members, 1989 = 100 Country Name Bulgaria Czech Republic Estonia Hungary Latvia Lithuania Poland Romania Slovak Republic Slovenia Cyprus Malta France Germany Italy United Kingdom Greece Ireland Portugal Spain Source: World Bank, World Development Indicators All major economic theories concur that the level and development of investment patterns is decisive for persistent economic growth. The CEECs should have investment rates clearly above those of the EU. Among the five more successful candidate countries, Poland and Slovenia had (domestic) investment rates comparable to those of the four major EU economies in the 1990s while Hungary and, in particular, the Czech and Slovak Republics, had notably high rates (see Table 3.1.3). Bulgaria had the lowest rates of the CEECs. However, high rates of investment are not sufficient in themselves: Romania had quite high rates until 1996 although Romanian growth is meagre. Even more important appears to be whether or not domestic investment is rising. On this basis, the rates in Bulgaria, Latvia and Romania are currently much lower than in the early 1990s, while the Czech and Slovak Republics, Hungary and Slovenia invest a higher share of their GDP. Foreign direct investment is most important in Hungary and Estonia FDI levels of around five percent of GDP or above DIW 33 EPRC

34 were reached in only a few applicant countries in selected years. More analysis and discussion on the specific issues of investment and trade are provided in Chapter 4. Table Gross domestic investment in Applicants and selected EU Members (% of GDP) Country Name Bulgaria Czech Republic Estonia Hungary Latvia Lithuania Poland Romania Slovak Republic Slovenia Cyprus Malta France Germany Italy United Kingdom Greece Ireland Portugal Spain Source: World Bank, World Development Indicators In the early 1990s, all the transition countries experienced very high inflation rates (see table 3.1.4) and, while many were able to achieve one-digit rates in 1997 and 1998, inflation remained relatively high in Poland (16 and 12 % respectively) and Hungary (18 and 14 %). Romania and Bulgaria were again the most problematic cases with extremely high inflation throughout the 1990s. Inflation peaked in 1997 (1,087 % in Bulgaria and 155 % in Romania) and, although falling dramatically in 1998, still stood at 22 and 59 % respectively. DIW 34 EPRC

35 Table Inflation in Applicants and selected EU Members, consumer prices (annual %) Country Name Bulgaria Czech Republic Estonia Hungary Latvia Lithuania Poland Romania Slovak Republic Slovenia Cyprus Malta France Germany Italy United Kingdom Greece Ireland Portugal Spain Source: World Bank, World Development Indicators Labour Market Labour market changes are a key element in the transformation process of the CEECs. With the fall of socialism, new labour market models have evolved, characterised by liberalisation, increased private sector employment (including selfemployment), an extension of deregulatory functions of the state in the use of manpower and the introduction of new systems of labour relations and negotiation (Bilsen/Konings, 1998; Franz, 1995). The labour market changes in individual CEE countries vary depending on a range of factors including the nature of reform and the existing sectoral and employment structure. In some cases, the growth of the private sector has been sufficient to compensate, to a considerable degree, for the decline and restructuring of the state sector. In other cases, the decline has been more dramatic, and, combined with a weak private sector development, has fuelled employment decline and unemployment. Pol- DIW 35 EPRC

36 icy choices in areas such as privatisation affected the labour market situation, resulting in both employment decline but also private sector employment growth. The dismantling of factors which previously maintained excess employment, such as soft budget constraints and inefficient management, also had, and continue to have, impacts on the CEE labour markets. Wider changes in the economic structure of the CEE economies have also impacted on the labour market. First, significant sectoral change has occurred since the start of the reform process. The pre-reform sectoral structure of the CEE countries was characterised principally by a large industrial sector complemented by an equally significant agricultural sector (in some cases, four times higher than the EU average) and a relatively small share of services. The main post-reform sectoral changes include the marked rise in the share of the service sector and a corresponding fall in the role of manufacturing industry and agriculture (EBRD, 1998; OECD, 1997). Industrial employment (defined in this case as mining, manufacturing and utilities) is estimated to have fallen by between % in different CEECs (UNECE, 2000). However, the decline in most transition countries reflects more previous high levels of over-employment and a deterioration of the industrial and manufacturing base than a positive process of restructuring. A comparison with the employment structure of the EU (see, e.g., Annex 3) shows that the share of the service sector in the CEECs is still significantly below that of the EU. Employment is, compared to the EU 15, still concentrated on industry and agriculture (see section 3.4). Intensified restructuring in 1999 accelerated the rate of employment decline in mining, manufacturing and the utilities in countries including Poland and the Czech Republic, while Hungary was the only CEEC where employment in these sectors has been increasing since 1997 (UNECE, 2000). Industrial production (defined in this case as non-agricultural or service related production) as a share of GDP still exceeds the EU average in a number of transition economies (Martzanis and Petrakos, 1998). Specifically, manufacturing employment has declined to a lesser extent than in the larger category of industry in the majority of the CEECs and has been a motor of economic recovery in some countries. In Latvia and Lithuania, however, manufacturing has fallen much more than industry (more than 40 %), reflecting the breakdown of former close production ties and the loss of key markets in the former Soviet Union. DIW 36 EPRC

37 In the services sector, there has been an overall increase in the share of total employment in all CEE countries, although the countries had different starting points. The Czech Republic and Slovakia, for example, had service sectors which accounted for over 40 % of employment in 1990 while the corresponding figure in Romania was only 27 % (Employment Observatory, various years) and according to UNECE only in Poland and in the Czech Republic, the service sector was significantly growing in terms of absolute employment figures. Certain sub-sectors such as finance, hotels and trade have displayed particularly notable increases. The development of the private sector (see below), as well as a proportionally lower fall in employment in public services such as health, education and other social services, are important contributors to service sector growth in most candidate countries but there is still a long way to go. Employment trends in agriculture are mixed. While employment in this sector has fallen in most acceding countries, there are important exceptions. In Poland and Romania, agriculture plays a key role and, at the start of the transition period, accounted for over 25 % of the total working population. In the Polish case, this is largely related to the fact that full collectivisation was never achieved. Agricultural employment actually increased in Romania (as well as in Bulgaria, Lithuania and Latvia) during the early 1990s, acting partly as a reservoir for job losses in other sectors. Agricultural smallholdings also became a necessary source of additional income and insurance against unemployment. Romania remains the extreme case and, in 1999, agriculture still accounted for 42 % of total employment. The second trend with ramifications for the labour market is the shift in ownership structures and the growth of the private sector. This has emerged partly as a result of the increase in private entrepreneurial enterprise (Johnson/Loveman, 1995; Welter, 1997, 1995; Zemplinerova, 1997; Bilsen, 1997) and partly through the process of privatisation (Pohl, 1997; Iankova, 1998; Ostojic/Scott, 1996; Smith/Cin/Vodopivec, 1997). Privatisation has involved various approaches in different countries, starting in some states with the restitution of property, and then generally including a small privatisation programme bringing small-scale enterprises in the retail, construction and commercial service sectors into the private sector. This was then followed by a more extensive transfer of state-owned industrial assets or the privatisation of large stateowned enterprises. Business start-ups have increased, encouraged by packages of measures liberalising the business environment. This trend has been particularly no- DIW 37 EPRC

38 table in the service sector. The growth of new firm formation and share ownership suggests that considerable entrepreneurship exists in the CEE countries (Jackson, 1996) although a high turnover of businesses is associated with difficult overall economic conditions. Data calculations undertaken for this report show the current labour markets of the candidate countries to be relatively homogenous (see table 3.1.5). After initial falls in overall employment in all the CEECs, the size of the labour force in 1998 in most of the transition countries was between 94 and 108 % of the 1990 level Latvia was the exception with a figure of 90 %. Poland, the Czech and Slovak Republics were the most successful while Hungary, Slovenia, Lithuania and Romania managed to stabilise employment figures. Table Labour force in Applicants and selected EU Members, total, 1990 = 100 Country Name Bulgaria Czech Republic Estonia Hungary Latvia Lithuania Poland Romania Slovak Republic Slovenia Cyprus Malta France Germany Italy United Kingdom Greece Ireland Portugal Spain Source: World Bank, World Development Indicators Unemployment is one of the most obvious phenomena to have emerged as a result of economic reform, both from a political and social perspective. There is much less DIW 38 EPRC

39 homogeneity in this area of the CEE labour markets. Table indicates that, in 1997, Bulgaria and Latvia had the highest rates of unemployment (14 %) with Poland (12 %), Slovak Republic (11 %) and Hungary and Estonia (10 %) all also in double figures. The lowest rates were recorded in the Czech Republic (5 %) and Romania (6 %), with Lithuania and Slovenia also low (7 %). The under-registration of hidden unemployment and ineligibility for benefits among those out of work for prolonged periods are thought to raise the real levels of unemployment above these rates in a number of countries. It is notable, however, that the CEE unemployment levels are comparable to, and in many cases lower than, those in EU countries. However, it must also be considered that the low rates in some CEECs (and Romania especially) are due to delayed economic restructuring. The spatial and social dimensions of unemployment are discussed further in following sections. Table Unemployment in Applicants and selected EU Members, total (% of total labour force) Country Name Bulgaria Czech Republic Estonia Hungary Latvia Lithuania Poland Romania Slovak Republic Slovenia Cyprus Malta France Germany Italy United Kingdom Greece Ireland Portugal Spain Source: World Bank, World Development Indicators DIW 39 EPRC

40 3.1.3 Social Inequality The official policy of communist regimes in CEE was to eradicate inequalities between different social groups and, to some extent, policies were implemented to limit the extent of income disparities. Nevertheless, inequalities in incomes and standards of living persisted during the communist period and, in the post-communist period, have generally deepened while new socio-economic divisions have also developed. Economic transformation in particular has been associated with emerging social problems and widening inequalities within CEE societies (Cox/Mason, 1999; Mareš/Možný, 1995). The emergence of opportunities for the ownership of capital, a growth of employment opportunities in the private sector and a general liberalisation of CEE economies have all contributed to a growth of inequalities in the distribution of incomes and wealth. A commonly used method of measuring income distribution is the Gini coefficient which is scaled between 0 and 100. The coefficient equals 0 in the hypothetical case where all units (usually households) in a society have the same income, and it equals 100 when one unit receives all of the society s income. According to Wyzan s (1996) calculations, measured inequality generally rose in CEE, especially in Bulgaria, Estonia and Lithuania. In comparison with some countries in the West, income disparities in Visegrad countries and Slovenia were not large: all had Gini Coefficients of around 30 in 1993 which is roughly on a par with the more egalitarian EU members states (Wyzan, 1996). Nevertheless, when compared with the situation under communist rule, the present levels of inequality in CEE appear to represent a significant widening of differentials - although problems inherent in using communist statistics as a guide to inequality need to be taken into account. New, privileged, wealthy groups have generally benefited from processes of transformation which have created conditions enabling private ownership and entrepreneurship and the conversion of the managers of state-owned enterprises into managers and owners of new independent privatised firms (Cox/Mason, 1999; Eyal et al. 1997). At the same time, income levels and standards of living have declined and poverty has spread considerably. According to Deacon (1993), living standards dropped by up to one third in the early 1990s, increasing the number of people living on or below the official poverty line. The low levels of economic development in general, and the specific problem of poverty, are undoubtedly factors influencing the DIW 40 EPRC

41 wide discrepancies between East and West Europe in key social indicators such as life expectancy. In 1993, male life-expectancy in Hungary was behind the EU average by eight years, and the figures were six years for Romania, 4.6 for Poland and 4.3 in Bulgaria (Amato/Batt, 1999). Poverty is also inexorably linked to social problems such as increased crime rates and discrimination against minority groups. Analysing the true extent of poverty in CEE can be difficult, as definitions and indices of poverty vary. There is consensus in the literature that levels of poverty in CEE are considerable, although significant variation appears to exist between countries. According to the World Bank, which compares actual incomes against a common poverty line set for all CEECs at $120 per person per month, relatively small numbers of people were below the poverty line in the Czech Republic and Hungary in the early 1990s (Cox/Mason, 1999). In Estonia roughly one third of the population lives below the nationally defined poverty level (European Parliament, 1999a). According to Amato/Batt (1999), the number of people living below the subsistence minimum in Hungary reached three million in 1995, over one third of the population, while in Bulgaria, over 70 % were living below the poverty line in The incidence of poverty also varies across social groups. Evidence suggests that particular sections of CEE society are more adversely affected than others by dramatic socio-economic reform, e.g. the elderly, specific ethnic groups, single-parent families, unemployed, low-paid employees and women. There are a number of reasons for the increase in poverty generally and for the particular vulnerability of these social groups. Loss of income through unemployment is a major contributory factor. Unskilled, poorly educated workers in vulnerable sectors of the economy have been particularly affected by unemployment and youth unemployment has also reached high levels (see Map). New entrants into the job market have found gaining an initial foothold in labour markets particularly difficult when employment was generally falling (Cox/Mason, 1999). Women have tended to be concentrated in low-skilled and poorly paid jobs, and have been particularly vulnerable to redundancies brought about by economic liberalisation and enterprise restructuring. In 1997, female unemployment (as a percentage of the female labour force) was higher than the male equivalent in the Czech Republic, Poland, Latvia and the Slovak Republic (see tables and 3.1.8). This is similar to the situation in the majority of EU countries. Hungary and DIW 41 EPRC

42 Estonia were the only countries where these figures were reversed, unemployment evels among men and women being equal in all other cases. According to Bretherton (1999), women once unemployed were significantly less likely than men to be reemployed. In 1997, the share of long-term unemployed women (as a percentage of all unemployed women) was higher than total long-term unemployment in all CEECs except the Czech Republic and Slovenia (no data for Hungary). Table Female Unemployment in Applicants and selected EU Members (% of female labour force) Country Name Bulgaria Czech Republic Estonia Hungary Latvia Lithuania Poland Romania Slovak Republic Slovenia Cyprus Malta France Germany Italy United Kingdom Greece Ireland Portugal Spain Source: World Bank, World Development Indicators DIW 42 EPRC

43 Table Male Unemployment in Applicants and selected EU Members (% of male labour force) Country Name Bulgaria Czech Republic Estonia Hungary Latvia Lithuania Poland Romania Slovak Republic Slovenia Cyprus Malta France Germany Italy United Kingdom Greece Ireland Portugal Spain Source: World Bank, World Development Indicators Bretherton (1999), Subhan (1996) and Watson (1993) have also noted the emergence of some discriminatory practices in recruitment. Minority groups, and in particular the Roma population of CEE, experience particularly high unemployment levels and discrimination. In 1995, the unemployment rate for the Roma population of Hungary was 45 % compared to 10.6 % for the rest of the population (Szamuely, 1996). The Bulgarian Roma minority has similarly high unemployment rates (European Parliament, 1999a). In many CEECs, specific measures to assist minority groups, women, youth and long-term unemployed back into employment are limited (European Parliament, 1999a), although some exceptions exist. In Slovakia, for example, programmes are in place to target the problems of specific groups, e.g. youth, elderly and the long-term unemployed. Further, a large system of social transfers in the country produces a reduction in measured poverty rates at least in relation to a very basic living standard and there are apparently few people well below the poverty line (OECD, 1996). Another factor affecting poverty levels has been the decline in real wages, and in particular minimum wage levels, following the introduction of price liberalisation and DIW 43 EPRC

44 strict income policies by CEE governments. As Standing/Vaughan-Whitehead (1995) have noted, while the minimum wage helped to provide a barrier to severe poverty in the past, in the new market-oriented economy it very often became the means of intensifying poverty. Moreover, the decline in minimum wage levels has also had implications not only for the lowest paid workers, but in many countries for a wider range of workers whose wage rates are set with reference to the minimum wage. Only 44 % of respondents in a study by Rose/Haerpfer (1998) claimed their regular job paid enough to cover their needs and in Bulgaria the figure was as low as 17 %. According to this and other studies (e.g. Amato/Batt, 1999) individuals often have to turn to the second economy (unregistered economic activity) in order to supplement their incomes. However, the poorest and least skilled groups tend to be excluded from the second economy through lack of skills, resources and personal connections. The social situation is critical for those who have insufficient incomes, especially in countries such as Romania and Bulgaria. Economic difficulties, and the delay by many governments in implementing major social reform, has particularly impacted on certain groups e.g. the unemployed, pensioners and farmers, and made them vulnerable to poverty (European Parliament, 1999a). Vulnerable social groups have also been affected by cuts in service provision and governments difficulties in establishing new institutional structures to deal with social problems (Elster/Offe/Press, 1998). In the Czech Republic, for example, poverty has recently increased because spending on the social security system was the first area to be cut as part of a response to the recent economic crisis and increase in unemployment in the country (European Parliament, 1999a). Formerly, social protection systems were broad and universal, but guaranteed only a relatively low standard of living. In the post-communist period, there has been a general decline in the level of social services provided by the state-owned enterprises as enterprises striving to meet new profit criteria close créches, clinics, subsidised canteens and other fringe benefits. Groups such as single parent families and the elderly have been particularly adversely affected. The greatly reduced availability, and increased cost, of child-care provision has been a particular impediment to single-parent families and low-income groups (Ferge, 1997). Both groups have also been affected by the failure of some benefits, e.g. child support and pensions, to keep up with the price increases of the early 1990s. The European Briefing paper The Social Aspects of Enlargement of the European Union (1999) highlights some of the problems. In Bulgaria, for example, DIW 44 EPRC

45 the level of social benefits offered provide inadequate protection. In Romania, pension systems are facing major difficulties: due to an ageing population, pre-pension schemes and redundancies in state owned enterprises, the number of pensioners in 1998 was bigger than the number of contributors. Further, the sharp decrease in the purchasing power of pensions has exposed the retired population to an increased risk of poverty Environment Heavy industrialisation in the pre-transition period caused extensive deterioration in environmental quality, affecting parts of every CEE country (Manser, 1993). Rapid industrial development combined with the inefficient extraction of raw materials, obsolete technology and a lack of environmental regulatory controls created a legacy of environmental degradation. The risks of old technology in areas such as nuclear power generation are particularly serious, with important nuclear installations in countries such as Lithuania, Bulgaria and Slovakia. Substantial improvements have, however, been made in the last decade in the reduction of environmental pollution. This has been influenced by a range of factors including policy changes, the introduction of newer technologies, the fall in industrial production and the elimination of subsidies, and structural changes in energy supply. Considerable spatial differentiation in environmental problems is evident in CEE, largely as a consequence of concentrated environmentally damaging production close to the centres of raw materials and energy resources. Karl et al. (2000) highlight the spatial concentration of pollution in Poland where Upper Silesia produces around 25 % of all industrial dusts, 30 % of all gases, 25 % of all non-cleaned sewage and around 60 % of all industrial solid waste. Other heavily polluted areas are located in the south and south-west border regions the Legnica-Katowice copper basin and the Glogow-Ryback coal basin as well as Gdansk and Szczecin on the Baltic coast. Similarly in the Czech Republic, polluted areas are concentrated in the capital city region and Northern Bohemia - the area of the Ore mountains, together with the adjacent border districts in Germany (Saxony), and Poland (Silesia), is sometimes referred to as the Black Triangle. The spatial dimension can change the overall national picture in Hungary, for example, while overall levels of environmental degradation are relatively lower than other CEECs, the problems in the so- DIW 45 EPRC

46 called industrial crescent, stretching from Ajka, Gyor and Tatabanya in the west to Miskolc and Ozd in the east, are more comparable (Karl et al., 2000). Overall, the CEE countries face the same environmental challenges as other industrialised European countries but from a base of more accentuated difficulties. The cost of environmental clean-up and the conflict with other policy priorities negatively affected the progress made in this area. Shifting the costs onto private industry has perceived risks of slowing the process of industrial restructuring, while concern has been voiced that the strict application of environmental regulations could dissuade foreign investors from taking on privatised enterprises or making greenfield investment. On the other hand, foreign companies have been the source of new environmental practices in many cases, and legislation against pollution has been accelerated in part by the need to conform to the EU acquis. From a governmental perspective, the problems have been compounded by the lack of reliable data (on various pollution variables) and limited monitoring and enforcement mechanisms available to policy-makers (OECD, 1996). This has led to political difficulties in determining where the priorities in environmental expenditure should be, or indeed what the most appropriate policy instruments are to encourage compliance with regulations (Ambler and Marrow, 1998; OECD, 1994b). The Sixth Periodic Report notes that the legislation being introduced in the CEE still generally imposes requirements which are less than those in the EU and that there are further problems with implementation. The costs both of reducing pollution and dealing with the legacy of the past are unlikely to be able to be borne by the CEECs alone (CEC, 1999). 3.2 Patterns of Regional Disparity in CEE and EU Countries Within the socio-economic context outlined above, regional and social disparities in the CEECs are becoming increasingly evident. The following section, based on data analysis carried out by the project team and using principally 1997/98 figures, provides more detail on the patterns of disparity in key economic areas both within CEECs and in comparison to the EU. Annex 1 includes a list of all AC regions (at NUTS II level) as well as an overview map to locate them and maps for each of the indicators discussed below for AC regions and for EU and AC regions combined. Full Tables are provided in Annex 2. DIW 46 EPRC

47 3.2.1 General Overview A summary of data on sub-national disparities in the applicant countries is provided in tables and and show a differentiated picture for individual countries. A country-specific analysis is hampered at NUTS II level due to the small overall size of many applicant countries. Six CEECs comprise a single NUTS II region and Bulgaria is divided into only three regions which gives little space for distinctive regional profiles. The Tables include, in their first column, all 6 applicants with more than one region with their national average in per capita GDP and employment. As a point of reference, values are provided for Germany and Italy, the two EU members with the most severe internal regional disparities. In each of the six candidate countries with more than one NUTS II region, the capital region is the wealthiest in the country. The general pattern throughout CEE is a significant concentration of economic activity in the political centre of the country (see Summary Table in Annex 2 as well as section 3.3.1). However, the very high relative values of some of the capitals are influenced by commuter flows. Nyugat-Dunantul in Hungary, Slaskie and Wielkopolskie in Poland and Centru in Romania are the only regions, apart from the capital city regions, which are clearly above their respective national averages. The poorest regions in Poland, the Slovak Republic and Hungary are located at the eastern border, and the very poor region of Nord-Est in Romania also fits this pattern. The regions of Bulgaria and the Czech Republic are relatively homogenous. The national spread, or difference between the richest and poorest regions, is often distorted by artificially high GDP values in the wealthiest regions resulting from commuter flows. A better measure for sub-national disparity levels is the standard deviation, calculated from the weighted distance between the regional per capita GDP and the national average. Compared to the most extreme case of the EU 15, sub-national disparities in the candidate countries are not high. Only Hungary and Poland pose a possible problem in this respect while the situation in the Czech Republic is dominated by the exceptional case of Prague. DIW 47 EPRC

48 Table 3.2.1: Inner-National Disparities in Applicant Countries (GDP) Applicant Country (in brackets: GDP p.c. EU15 = 100) Wealthiest Region Poorest region GDP p. c. 1997, National Average =100 GDP p. c. 1997, EU15=100 National Spread (1) Standard Deviation, National Average = 100 BULGARIA (22.6) SOFIA STOLITSA SEVERNA BALGARIJA CZECH REP. (63.2) PRAHA STREDOCESKY HUNGARY (47.5) KOZEP-MAGYARORSZAG ESZAK-MAGYARORSZAG POLAND (35.4) MAZOWIECKIE SWIETOKRZYSKIE ROMANIA (30.4) BUCURESTI NORD-EST SLOVAKIA (45.4) BRATISLAVSKÝ KRAJ VÝCHODNÉ SLOVENSKO Memo Items: GERMANY (108.4) HAMBURG DESSAU ITALY (101.7) EMILIA-ROMAGNA CALABRIA (1) Difference between wealthiest and poorest region (in percentage points of national average). Source: Eurostat. DIW 48 EPRC

49 Table 3.2.2: Inner-National Disparities in Applicant Countries (Unemployment) Applicant Country (in brackets: total unemployment rate 1998) Region with: Lowest Unemployment Highest Unemployment Unemployment Rate 1998 National Spread (1) Standard Deviation BULGARIA (16.0) SOFIA STOLITSA SEVERNA BALGARIJA 19.5 CZECH REP. (5.9) PRAHA SEVEROZAPAD 9.6 HUNGARY (8.9) NYUGAT-DUNANTUL ESZAK-MAGYARORSZAG 13.0 POLAND (9.9) LUBELSKIE WARMINSKO-MAZURSKIE 15.6 ROMANIA (5.6) SUD-VEST NORD-EST 7.2 SLOVAKIA (14.4) BRATISLAVSKÝ KRAJ VÝCHODNÉ SLOVENSKO 21.6 Memo Items: GERMANY (9.8) OBERBAYERN DESSAU 22.3 ITALY (12.3) TRENTINO-ALTO ADIGE CALABRIA 27.0 (1) Difference between regions with highest and lowest unemployment in percentage points of national unemployment rate Source: Eurostat. DIW 49 EPRC

50 In Bulgaria and the Slovak Republic, the wealthiest and poorest regions are also those with the lowest and highest rates of unemployment. Sub-national disparities with regard to unemployment are less pronounced in the CEECs than the EU 15. With the exception of Bucharest, unemployment in the capital regions is sometimes clearly below the national average. There are only few regions with especially high unemployment (compared to the respective national average) and these include Severna Balgarija in Bulgaria, Ostravsky and Severozapad in the Czech Republic, Eszak-Magyarorszag in Hungary, Warminsko-Mazurskie and Zachodnio-Pomorskie in Poland and Vychodne Slovensko in the Slovak Republic Population Density Most of the CEEC regions are sparsely populated (see Map). This holds true in particular for Estonia and Latvia, where population density is only inhabitants per km 2. Overall, most of the CEE regions have an average population density below 100 inhabitants per km 2, fulfilling the characteristics of a rural region. In addition to capital cities, large agglomerations and their hinterlands, industrial areas are more densely populated e.g. the industrial belt in the south of Poland (Slaskie, Malopolskie) and the north-east of the Czech Republic (Ostravsky). The EU is generally more densely populated (see Map) with an average population density above 100 inhabitants per km 2. The main exceptions are the Nordic countries of Finland and Sweden which are very sparsely populated and have considerably fewer inhabitants per km 2 than any CEE region. The more densely populated regions in CEE can be compared with EU regions such as Muenster, Lombardia and West Vlaanderen. The CEE capital cities, apart from Prague, are in general less densely populated than the large EU equivalents GDP per Head The general picture of disparities between the CEECs and the EU 15 is well known. In general, the GDP per head of the CEE regions is considerably less than the EU average (see Map). Only Prague (120 %) and Bratislava (102 %) are above the average GDP per head in the EU while Budapest (Kozep Magyarorszag in the Tables), Slovenia and Cyprus approach two-thirds of the EU average figure. The next group of regions has GDP per head levels of around 50 % of the EU average and includes the whole of the Czech Republic, the region around Warsaw in Poland and a Hungar- DIW 50 EPRC

51 ian region close to the Austrian border. The remainder of the CEE regions are all well below 50 % of the EU average figure, with the most underdeveloped regions located at the eastern borders with Russia, Belarus, Ukraine, Moldova and to the south bordering the former Yugoslavia and Greece. The poorest regions are found in Bulgaria with only % of the EU average. Below NUTS II level, even greater differentiation is found with the poorest CEE region identified in the Sixth Periodic Report as Latgale in Latvia with a GDP per capita of 16 % of the EU average (CEC, 1999a). On the basis of the current Structural Fund regulations, the majority of the CEE regions would qualify for Objective 1 support. Only Prague and Bratislava are above the 75 % threshold. Even the poorest EU region (Ipeiros in Greece) reaches 43 % of the EU average and therefore has a comparable GDP per head with the Czech Republic, one of the most developed CEECs. This indicator provides one of the sharpest disparity levels between CEE and the EU Unemployment Disparities in regional unemployment figures must always be interpreted with caution as official figures depend strongly on the labour market rules and methods of measurement within individual countries. Levels of officially registered unemployment can be seriously affected by the nature of the social benefit system, including ineligibility for those unemployed for long periods, which can hide the true level of unemployment. In contrast to the situation with GDP per head, official unemployment figures for many CEE countries and regions are relatively low in comparison with the EU. In 1998, the Czech Republic (5.9 %), Romania (5.6 %) and Slovenia (7.4 %), as well as Cyprus (3.3 %) and Malta (5.1 %), all had (official) unemployment rates considerably below the EU average (10.1 %). Indeed, the official statistics appear to suggest almost full employment in four regions North West and South West Romania and Central and South West Bohemia in the Czech Republic. Hungary (8.9 %), Estonia (9.6 %) and Poland (9.9 %) also had national unemployment rates below the EU average. Bulgaria, the Slovak Republic, Latvia and Lithuania had unemployment rates exceeding the EU average, the highest in 1998 recorded in Bulgaria at 16 %. As in the EU, the variation at sub-national level is considerably higher. Even in countries with relatively low national rates, such as the Czech Republic and Romania, re- DIW 51 EPRC

52 gional differences are considerable. In the Czech Republic, for example, the lowest rate in 1998 was recorded in Prague (3.1 %) while the industrialised region of Severozapad in the north-west had an unemployment rate of 9.6 %. In Hungary, the regions close to the eastern border of the country have higher rates of unemployment (Eszag-Magyarorszag, 13 %, and Eszak Alfold, 11.6 %) although still not markedly over the EU average. Rural regions in Bulgaria, Slovakia and northern Poland have among the highest unemployment rates in CEE, with the eastern region of Slovakia (near the Ukrainian border) recording the highest level of 21.6 %. Unemployment levels in some EU regions are considerably higher than in CEE counterparts and this is particularly true in the south of Italy and Spain where unemployment rates of up to 29 % are recorded (see Map). The highest unemployment rates in the CEE regions correspond better to rates in regions such as Dessau, Halle, Magdeburg in Eastern Germany and Sardegna in Italy. Only some Dutch, Austrian and southern British regions have low unemployment figures comparable with those in Czech and Romanian regions. While the overall employment rate in the applicant countries is slightly lower than in the EU 15, the female employment rate is higher. Female unemployment follows by and large the same regional pattern as total unemployment. In Latvia and Bulgaria as well as Lubuskie/PL, Central Hungary, Central and South West Romania female unemployment reaches the same level as the overall unemployment. In the other Baltic states, the Eastern part of Romania and the Western part of Hungary, female unemployment is lower than the overall unemployment rate. Especially the Czech Republic, Cyprus and the western part of Poland suffer from highest female unemployment compared to the overall unemployment rate. In these regions female unemployment is between 20 to 50 per cent higher than general unemployment. It seems that the regions with a greater share of agriculture in employment are suffering less from female unemployment. However, in absolute terms, the highest female unemployment can be found in Eastern Slovakia (22.8 %), Zachodnio-Pomorskie/PL (19.2 %) and Northern Bulgaria (19.6 %). In Slovakia having the highest unemployment rates in the CEEC men and women are suffering almost equally from unemployment. Compared to the EU15, female unemployment of the CEEC regions in absolute terms lies close to the DIW 52 EPRC

53 EU average, but it is considerably lower than female unemployment in Spain, the south of Italy, Eastern Germany and France. Regarding the overall labour market problems and the future strategies, it is necessary to note that youth unemployment is in all CEEC regions remarkably higher than female unemployment. The regions with high overall unemployment (by national standards) are consistently those regions with high youth unemployment. In addition, Sud-Est and Bucuresti in Romania and Podkarpackie in Poland show high youth unemployment, although overall unemployment is less significant. The problem of youth unemployment is most severe in Bulgaria, the Slovak Republic and the labour market problem regions of Poland Employment Structure An analysis of regional level employment structure reveals a clear dominance of agriculture in many CEE regions. This is especially true in Romania where, outside Bucharest, the lowest share of agriculture in total employment is 32 % rising to a maximum of 57 % in Nord-Est. These findings are confirmed in other literature sources. Croft et al. (1998), for example, state that, in the mid-1990s, CEE countries employed proportionally four times as many people in agriculture as the EU (22.5 % of the total workforce in comparison with 5.3 %). The capital cities are, unsurprisingly, dominated by services, with figures ranging from 64 % in Sofia to over 70 % in Bratislava and Prague. In the case of Budapest and Sofia, both cities also have a considerable share of industry (40 and 33 % respectively). Regions dominated by industry include the northern border regions of the Czech republic, the industrial belt in the south of Poland and the north-east of Hungary all with shares in excess of 40 %. The Czech, Slovakian and Slovenian regions also have considerable shares of industrial employment, as well as agricultural labour. The EU regions tend to be much more diversified in terms of employment structure. It is rare to find a dependency on agriculture in EU regions which is similar to that found in CEE, and shares in agricultural employment are much lower even in regions where agricultural employment is dominant e.g. in Spain and southern Italy. Only two Greek regions (Anatoliki Makedonia and Peloponnisius) come close to the Romanian levels. DIW 53 EPRC

54 3.3 Types of Regional Problems in the CEECs The type of regional development problem to emerge in the CEECs is a product of a wide range of factors. To some degree, the transformation of the 1990s in the CEECs resembles the start of structural change which took place in more developed economies a few decades previously but which was prevented in the CEE case by the geo-political changes following the Second World War. Regional economic development problems and disparity patterns associated with similar structural change in Western Europe can, therefore, be anticipated in CEE. While this may be the case, the process of economic and political transition in CEE is also quite distinct and, in spatial terms, faces the particular legacy both of internal socialist development patterns and a re-orientation away from the formerly imposed economic integration links with eastern, rather than western or global, markets. The result is a complex mix of regional development disparities and problems which have common elements both between the individual CEECs and with Western equivalents but also influenced by deep-seated historical and cultural factors, ethnicity, and specific national characteristics. The ability of individual regions to adapt to the fundamental changes in economic environment rests on a range of issues including their socio-economic structure, level of initial development and proximity to capital and innovation as well as the way in which they are affected by national policy decisions. An early conceptual framework classifies regions on the basis of their relative position at the start of transition and their likely response of the process of economic change (Gorzelak, 1994) and identified certain groups of relatively advantaged or disadvantaged regions in the CEECs. The regional out-working of this framework has not altered significantly, and a similar group of four inter-related types of regional disparity is currently identifiable: the contrast between urban and rural areas; a core/periphery disparity, especially in countries with a mono-centric urban structure; a west/east difference, particularly evident in border areas; and concentrations of restructuring problems in old-industrial areas (Bachtler/Downes, 1999). Clearly, the detailed patterns of regional disparity are more complex but the overall groupings hold broadly true and are useful in providing an overview of the type of regional difficulties which have emerged in the CEECs. The analysis in this section has been drawn principally from research by the project team and the wider review of available literature. DIW 54 EPRC

55 3.3.1 Capital Cities and Major Urban Agglomerations Throughout the available literature, major CEE agglomerations and urban areas are consistently identified as leaders in the transformation process (Bachtler/Downes/Gorzelak, 2000; Boeri/Brücker et al., 2000; Aru, 1998; Vaidere, 1998; Gorzelak, 1994). Most prominent is the dominant role of core and capital city regions; indeed in countries such as Hungary, the Czech Republic, Estonia and Latvia there is no centre which rivals the capital city. The impact of mono-centric settlement structures is formidable: the Tallinn area (Estonia) has % of foreign investment and tourism and 40 % of all registered enterprises; Riga (Latvia) has 30 % of the national population and almost half of the FDI stock; in Prague (Czech Republic), GDP per capita is approaching a level twice the national average, and wage levels are almost one third higher than the Czech average; and Budapest (Hungary) accounts for 40 % of the total urban population, 35 % of service sector employment, and nearly twothirds of all FDI flowing into Hungary, contributing to a GDP per capita level three times that of the worst-placed county in Hungary (Szabolcs-Szatmár-Bereg) (Bachtler/Downes/Gorzelak, 2000). The Bratislava region (Slovakia) was found to be similarly dominant in studies by Buček (1999) and Smith/Fernečíková (1998): in 1997, the region accounted for 62 % of inflows of FDI, 92 % of banking and insurance sector employees and 41 % of R&D and business service sector employees. Boeri/Brücker et al. (2000) found that, in comparison to the situation in EU countries, CEE capitals were relatively small in terms of population share (except Hungary) but had significantly higher shares of overall GDP. The absence of major secondary centres in many of these CEECs means that, outside the capital cities, spatial disparities in growth are more limited. In the CEE countries with more multi-polar urban structures, while the capitals still demonstrated more favourable economic indicators (especially for foreign investment), other important centres for economic development can be identified. These include Poznań, Kraków, Wrocław, Gdańsk, Katowice, Szczecin and Łódź in Poland, Kaunas and Klaipeda in Lithuania and Varna and Plovdiv in Bulgaria (Kozak, 1998; Bachtler/Downes/Gorzelak, 2000). In Slovenia, the topography of the country means that 75 % of the population live on one-fifth of the territory. Central Slovenia, which includes the capital Ljubljana, together with the north-western regions, were found to display generally better economic indicators (for 1997) than the rest of the country, DIW 55 EPRC

56 but the disparities are not as marked as in other CEECs (Bachtler/Downes, 1999; Piry, 1998). As established centres of economic development, major cities benefit from a high level of investment, a skilled labour force, more developed infrastructure, business services, access to key decision-makers, a higher standard of accommodation and retail facilities. These conditions are recognised as supporting accelerated progress in key areas of economic transformation areas such as privatisation and restructuring. More specifically, urban agglomerations demonstrated advanced rates of service sector growth, FDI inflows, and new business development. On the whole, capital city regions have lower unemployment levels than surrounding regions. According to the European Commission s Sixth Periodic Report (CEC, 1999a), regional unemployment in Hungary is lowest in Central Hungary (containing Budapest) and urban centres in Poland, such as Warsaw, Poznań, and Katowice, have the lowest unemployment levels in the country. Extremely low rates of unemployment in Prague (Czech Republic) have even given rise to concerns about worker shortages. As rapid development in the capital cities has advanced, specific urban-related problems have started to emerge including, for example, pollution, congestion and housing shortages. Fassman (1998) and Duke/Grime (1997) have also identified the emergence of intra-urban social inequality as the larger urban areas become increasingly stratified by social group and polarised along Western European lines. Recent research on housing privatisation (Grime et al., 2000) suggests that social inequalities are developing rapidly, with unemployed residents particularly disadvantaged. The study concluded that societal distribution patterns of post-communist cities are moving towards Western models of polarisation CEE: EU Border Regions The western border regions of CEEC during the socialist era were in an unfavourable, peripheral position because of the geo-political orientation of the CEECs. Post- 1989, these regions started to benefit from their position within a new European political and economic geography. Proximity to the EU, relatively developed infrastructure and low labour costs combined with labour force skills all contributed to stimulate markets and encourage investment into western border regions (Boeri/Brücker et al., 2000; Gorzelak/Zanycki, 1995; Bachtler/Downes/Gorzelak, 2000). Other advantages, such as relatively developed infrastructure highlighted by Boeri/Brücker et al. (2000), DIW 56 EPRC

57 also compound their advantage. Western Polish border regions, for example, have a more diversified industrial structure and a higher number of SMEs. In Hungary, western regions have witnessed falling unemployment in recent years and a positive inflow of investment, building on existing industrial structures, proximity to western markets and cultural factors which make them attractive for investment from neighbouring countries (Lorentzen, 2000; Bozóki, 1997). In addition to this type of economic development, other benefits experienced by western border regions include increased tourism (Turnock, 1997), cross-border retail and educational/technological initiatives such as the international university in Frankfurt/Oder-Słubice and the ACCESS cross-border technology and business park on the Austrian-Czech border. Western European centres that lie relatively close to CEE borders Vienna, for example, is only 65 km from Bratislava and Berlin lies only 110 km from the Polish border have been found to exert a strong influence on neighbouring border regions (Gorzelak/Zanycki, 1995). Boeri/Brücker et al. (2000) note the particular growth of regions along the main east-west transport axes and the western border areas in Hungary, Slovakia and Poland. A study of Euro-Regions in Poland (Bojar, 1996) found that local, economically motivated cross-border co-operation could promote population stability, stimulate prosperity and revive company co-operation in border regions. The Győr-Moson-Sopron region in north-western Hungary has benefited from Austria and Hungary s interregional plans to develop road and train links and upgrade the region s basic infrastructure. Local firms in Hungary s western Transdanubian region are also profiting from cross-border business with Austria. Bozóki (1997) suggests Austria s entry into the EU further encouraged cross-border co-operation and the INTERREG II PHARE CBC programme offers Győr-Moson-Sopron, Vas and Zala counties new, co-ordinated developmental opportunities. However, in some documented cases, differing economic, social and political conditions in EU and CEE border regions have caused difficulties for cross-border developments, e.g. in the case of German-Polish Euro-Region of Viadrina (Bertram, 1998). It is often thought that enlargement will have the greatest impact on regions on the border between the current EU and the applicant countries. The above section highlights the impact of economic transition on CEE western border regions and section 4.4 will analyse the potential effect of enlargement in more detail. It is, however, DIW 57 EPRC

58 instructive at this point to review selected data for neighbouring regions at the EU:CEE border in a little more detail to provide an indication of the starting position along the current external border of the EU 15. Table shows regional indicators for every pair of regions along the current EU external border, as well as the relative value for the CEEC region as a percentage of the value of its western neighbour. Per capita GDP and productivity (excluding commuters) is lower in the eastern border region than in its western counterpart in all the border pairs. The only exception is Bratislava in comparison to its neighbouring Austrian regions in Niederösterreich and Burgenland. This is most probably related to the inability to take commuters into account, as well as the fact that Bratislava is the capital region in the Slovak Republic. It has to be noted that per capita GDP and productivity levels are well below the EU 27 average not only for all the CEEC border regions but also for the new German Länder, Burgenland (the Austrian Objective 1 region) and the Greek border regions. These EU border regions also have a very low population density, indicating a relatively low levels of economic activity on the western side of the border. In the German border regions, total unemployment is higher than in the neighbouring Polish and Czech regions, while youth unemployment is almost always clearly higher in the CEECs. In the Austrian, Italian and Greek border regions, both selected unemployment indicators show better labour market conditions on the EU side of the border than in the neighbouring CEE regions. DIW 58 EPRC

59 Table 3.3.1: Regional Disparities at the EU - CEEC - Border Neighbouring Regions Density pop/km² GDP p.c (EU27 = 100) Productivity (excluding commuters) 1997 (EU 27 = 100) Unemployment Rate 1998, total Unemployment Rate 1998, youth EUR EU MECKLENBURG-VORPOMMERN (D) ZACHODNIOPOMORSKIE (PL) CEEC Region in % of EU Region BRANDENBURG (D) ZACHODNIOPOMORSKIE (PL) CEEC Region in % of EU Region BRANDENBURG (D) LUBUSKIE (PL) CEEC Region in % of EU Region DRESDEN (D) DOLNOSLASKIE (PL) CEEC Region in % of EU Region DRESDEN (D) SEVEROVYCHOD (CZ) CEEC Region in % of EU Region CHEMNITZ (D) SEVEROZAPAD (CZ) CEEC Region in % of EU Region OBERFRANKEN (D) SEVEROZAPAD (CZ) CEEC Region in % of EU Region OBERPFALZ (D) JIHOZAPAD (CZ) CEEC Region in % of EU Region NIEDERBAYERN (D) JIHOZAPAD (CZ) CEEC Region in % of EU Region OBERÖSTERREICH (AT) JIHOZAPAD (CZ) CEEC Region in % of EU Region DIW 59 EPRC

60 Table continued: Regional Disparities at the EU - CEEC - Border Neighbouring Regions Density pop/km² GDP p.c (EU27 = 100) Productivity (excluding commuters) 1997 (EU 27 = 100) Unemployment Rate 1998, total Unemployment Rate 1998, youth NIEDERÖSTERREICH (AT) JIHOZAPAD (CZ) CEEC Region in % of EU Region NIEDERÖSTERREICH (AT) JIHOVYCHOD (CZ) CEEC Region in % of EU Region NIEDERÖSTERREICH (AT) ZÁPADNÉ SLOVENSKO (SK) CEEC Region in % of EU Region NIEDERÖSTERREICH (AT) BRATISLAVSKÝ KRAJ (SK) CEEC Region in % of EU Region BURGENLAND (AT) BRATISLAVSKÝ KRAJ (SK) CEEC Region in % of EU Region BURGENLAND (AT) NYUGAT-DUNANTUL (HU) CEEC Region in % of EU Region STEIERMARK (AT) SLOVENIJA (SI) CEEC Region in % of EU Region KÄRNTEN (AT) SLOVENIJA (SI) CEEC Region in % of EU Region FRIULI-VENEZIA GIULIA (IT) SLOVENIJA (SI) CEEC Region in % of EU Region KENTRIKI MAKEDONIA (GR) YUZHNA BALGARIJA (BG) CEEC Region in % of EU Region ANATOLIKI MAKEDONIA, THRAKI (GR) YUZHNA BALGARIJA (BG) CEEC Region in % of EU Region Source: Eurostat. DIW 60 EPRC

61 Cross-border economic relations between neighbouring regions are likely to intensify after enlargement where there are concentrations of economic activity on both sides of the border and if labour market conditions and income levels are better on the current EU side. On the basis of this measure, EU border regions that might be most affected by enlargement could include those in Bavaria and Austria, as well as Friaul. However, well-founded conclusions require a much more detailed analysis than that possible on the basis of NUTS II data (see also section 4.4) Peripheral Regions: Eastern Border and Rural The majority of the poorest or most economically disadvantaged regions in Western Europe are located on the periphery of the EU 15 and a similar pattern can be identified in the CEE with implications for the eastern periphery of an enlarged EU. Two important contributory factors can be identified for the weak position of CEE peripheral areas. The first is geographical location. While western CEE border regions are generally winners in the transformation process, the situation is consistently found to be reversed in the case of the eastern border regions, e.g. regions bordering Russia, the Ukraine and Belarus. Boeri/Brücker et al. (2000) and Gorzelak (1996) identified a belt of backward depressed regions which passes along the former Soviet border from north-eastern Poland to the Moldavian districts of Romania. Sometimes termed the eastern wall, the eastern border regions tend to have relatively poor infrastructure, little investment and unfavourable economic structures characterised by a predominance of agrarian activity, high unemployment, low educational attainment and qualifications in the labour force, limited labour mobility, unsatisfactory environmental conditions and unfavourable demographic structures (Gorzelak, 1996; Buček, 1999). The infrastructure endowment of the eastern peripheral regions in CEE is much poorer than that of peripheral regions in the EU 15 (CEC, 2000). There are a number of sources of potential disadvantage for eastern border regions. Communication and transportation routes are generally oriented towards former markets in the East and the re-orientation towards western markets has had a correspondingly negative impact. The current poor economic condition of the post-soviet republics limits opportunities for trans-border co-operation and joint economic initiatives, although some links are slowly being encouraged (Bachtler/Downes, 1999). Border regions between less developed CEECs e.g. Bulgaria and Romania, have also been found to experience similar difficulties to regions on the external eastern DIW 61 EPRC

62 border. Lukas (1999), for example, identified Slovak regions close to the Hungarian, Polish and Ukrainian borders as lagging behind central and western regions in terms of new employment opportunities while Buček s (1999) analysis of Slovak regional disparities found that the east-central region of Presov had an average unemployment 3.5 times higher than that of Bratislava (4.95 and 18.4 % respectively in 1997) and a per capita GDP of only a sixth of that of the capital. Eastern Hungarian regions, similarly, have been worse affected by rising unemployment than their western counterparts (CEC, 1999a). The second factor affecting the economic position of peripheral areas is their rural structure. In many cases, these two factors come together as seen in the strongly agrarian structure of many eastern border regions. In addition, rural regions sometimes known as the internal periphery suffer similar economic development problems. These CEE regions inherited a poor economic structure and position from the socialist period, which has generally worsened through the process of economic transformation and reform. This is characterised particularly by increasing unemployment and declining employment opportunities, both within and out of agriculture. The privatisation of agriculture has, in some CEE countries, left farmers in a very uncertain position. There has been a partial return to subsistence farming on small, uneconomic plots, lacking fertilisers and other costly resource inputs, and not viable for mechanised farming. In some cases, such small-scale farmers have to compete against state-owned farms still receiving preferential government support or rely on agricultural supply or marketing enterprises still under state control. Overall, agriculture in the CEECs is in need of urgent structural reform to meet the current and future competitive environment. Although agrarian price levels are low, product quality (also in hygiene terms) is often barely competitive on Western markets. Overall, many analysts concur that predominantly rural regions on eastern peripheries in particular, e.g. eastern regions of Poland, Latvia and Estonia continue to witness little improvement in their socio-economic position (Boeri/Brücker et al., 2000; Bachtler/Downes/Gorzelak, 2000; Gorzelak, 1996, 1999; Fassmann, 1998; Hajdú/Horváth, 1994; CEC, 1996). Such employment and unemployment problems, largely due to the declining agricultural sector and lack of alternative opportunities, are evident in regions including South-Transdanubia and the Northern Great Plain of Hungary, rural, mountain regions of Bulgaria (see Spiridonova/Grigorov, 2000), eastern Estonia and northern DIW 62 EPRC

63 Poland. Notable exceptions are the southern and eastern regions of Poland where high proportions of workers were still employed in agricultural sectors and unemployment is low (CEC, 1999a). As noted earlier, the share of primary sector employment also remains high in Lithuania and Latvia (ca. 21 % to total employment in 1999) and particularly in Romania (over 40 %), with even higher regional specific rates (see section 3.2.5). Lower unemployment rates are partly attributable to high out-migration, of some rural and underdeveloped agricultural regions (Bachtler/Downes, 1999). As in Western Europe, migrants are often younger or better qualified and there are low rates of return to the origin region. This exacerbates already imbalanced age-sex structures and impacts on the ability of rural regions to undertake successful, long-term economic regeneration. The problem of an ageing population in a number of countries (Slovenia, Czech Republic, Hungary) is concentrated in rural and peripheral areas on the eastern borders of CEE Old Industrial Regions Old-industrial regions can be counted among regions which have been most adversely affected by the process of economic transition. During the socialist period, the heavy industrial regions were the focus of planned development and were the drivers of economic activity (Michalski/Saraceno, 2000). During the transformation period, old industrial regions were severely affected by privatisation, enterprise restructuring and closures, the reorientation of trade from formerly secure markets and the loss of subsidies. The decline in socialist-style heavy industry in particular has played a significant part in widening regional disparities in CEECs (Bachtler/Downes/Gorzelak, 2000). Many old industrial regions have high rates of unemployment with a large proportion of groups which are more difficult to re-integrate into the labour market e.g. low qualified and long-term unemployed. The Slovak Republic inherited an industrial legacy from the former Czechoslovakia, and the subsequent reduction in military production in particular (in some cases by almost 90 %) has created very high unemployment and persistent socio-economic problems around the industrial towns of Dubnica, Martin, Povazska Bystrica, and Detva (Buček, 1999). North-eastern Estonia, which is heavily dependent upon power production, heavy industry and the former Soviet military-industrial complex, faces similar problems. Piry (1998) highlights the industrial-related problems of Slovenian regions including Gorenjska, Savinjska, Spodnej Posavska and Koroška which are DIW 63 EPRC

64 burdened with heavy industry including metallurgy in the lead and steel branches. In Poland, regions in the industrial north and west of the country have also experienced large-scale decline. Gorzelak/Jalowiecki (1997) identified Upper Silesia as a region with one of the most difficult restructuring challenges of any CEE industrial region, with an estimated 800,000 people, of a total of four million, employed in endangered sectors. Old-industrial regions commonly face severe degeneration of the environment, in terms of air, water and ground pollution, often associated with the scale and poor technology of former industrial production. Economic restructuring, as well as the requirement to meet environmental components of the acquis, have, in some cases, led to the modernisation or closure of polluting industrial sites (CEC, 2000). Major environmental problems still persist, however, which will add to the cost of further restructuring and may act as a disincentive to progress and investment (CEC, 2000) (see section 3.1.4). While many old-industrial regions have already experienced sharp economic and social transformation costs, others have yet to undergo a full process of restructuring. According to the EBRD s 1999 Transition Report, the restructuring of existing enterprises remains one of the greatest tasks of economic transformation in CEE. Old, mono-structural areas face infrastructural, technological, skill-based and environmental challenges, exacerbated by the increased requirement for the adoption of new technology, managerial and entrepreneurial skills required to remain competitive. Economic reform has not universally led to quality investment and the development of new products and production methods. Large numbers of enterprises have not fully embarked upon restructuring programmes and inefficient enterprises have been allowed to remain open, e.g. in Upper Silesia (Poland) and the Ostrave Karviná region (Czech Republic). Lack of restructuring has largely been due to weak political commitment, often associated with fears about the political and social consequences of massive job losses. Eser (1998) identifies, for example, the case of the coal and steel-producing region of Ostrave Karviná. Inefficient and outdated plants in vulnerable industrial sectors were able to postpone restructuring by competing on the international market, undercutting prices of CEE competitors (e.g. from Poland and Bulgaria) on the basis of the devaluation of the Czech koruna. DIW 64 EPRC

65 3.4 Types of Regions in Applicant Countries and the EU: Cluster Analysis The preceding sections showed the severe challenges for cohesion in the EU that result from the specific regional development problems of the applicant countries. In this section, the starting point of the analysis is not the specific situation in the CEECs (and therefore the difference between the EU and the CEEC regions) but the search for similarities of the regions in Eastern and Western Europe. The tasks were the development of a typology of regions and the classification of all approx. 260 European regions into this various types of regions. To achieve these tasks it was, first of all, necessary to select adequate indicators. It has to be distinguished between indicators that measure basic socio-economic framework conditions for the development of a region ("structural indicators") and performance indicators. Performance indicators are not useful for the typology of regions because they either may change rapidly (e.g. GDP growth) or they might separate Western and Eastern regions too strictly in different groups (e.g. GDP level). Most importantly, they are better used in analysing the situation inside the various groups of regions instead of defining these groups (e.g. GDP levels or unemployment indicators). A structure indicator often used for identifying rural areas and/or agglomerations is population density. Other promising (and available) structural indicators were the shares of agriculture, industry and services in total employment. These four indicators were included in the attempt to classify the regions of the EU and the AC into well-defined groups of regions. There was no viable way to postulate, in a first step, specific values for all the indicators by discretion that would, then, have to serve as thresholds in order to distinguish, say, an industry-dominated rural area from a service-driven agglomeration. First, this would have included too high a degree of arbitrariness. Second, there are not always easy and unambiguous solutions if there is more than one indicator involved. An established methodology to solve the tasks is the use of a cluster analysis. The main objective of this method is to build clusters of in this case regions that are as homogenous as possible with regard to the elements inside each cluster and as heterogeneous as possible as regards the profiles of the individual clusters. Cluster analyses allow the inclusion of an indefinite number of indicators without having to define critical values for these indicators by discretion. However, some discretion is DIW 65 EPRC

66 also always necessary in a cluster analysis (see Box for technical aspects; see also Backhaus et al. 2000). Box: Technical Aspects of the Cluster Analysis There is no uniform way to conduct a specific cluster analysis. At the various steps of the analysis several questions need to be answered (choice of indicators, choice of proximity measure, choice of clustering algorithm, depth of clustering, etc.). The different approaches lead to different solutions and not all of them deliver useful results. In particular, wrong or inconsistent specifications of individual regions are unavoidable to a certain extent. Therefore, some discretion is necessarily also associated with each cluster analysis. This box summarises some technical aspects of the cluster analysis conducted to classify the NUTS II regions of the EU and the AC into different types of regions. We checked various combinations of the selected variables. Because the three employment shares are not independent of each other we only included two of them in each clustering. (The three shares should add up to 100; however, due to data problems this is not always the case in practice.) As the indicators are measured in different units they had to be standardised. A so-called "z-transformation" was used ensuring an average of zero and a variance of 1 for each variable. There are also different ways to measure the proximity or similarity of elements in the analysis. We compared results for all so-called Minkowski-Measures (City-Block, Euclidean Distance, Squared Euclidean Distance). As regards the method used to combine the various elements to clusters, the Ward-Procedure has been proven to produce the most consistent results; this procedure requires the prior elimination of exceptional cases using the Single Linkage Procedure. However, we also checked other approaches (Complete Linkage). We conducted separate cluster analyses for the 210 EU regions and for 51 AC regions as well as an analysis that combined all regions. Malta could not be included because comparable Eurostat data on employment shares of economic sectors were lacking. The most consistent simultaneous classification of all NUTS II regions of the EU and the AC was produced by using the Ward-Procedure and the Squared Euclidean Distance. (One exceptional case had to be eliminated but could easily be assigned to a cluster afterwards.) In this approach, we used the shares of agriculture and services in total employment as well as the population density. The computer analysis (using SPSS 9.0) lead to eight different clusters that were re-assigned to six clusters. Only 13 regions, i.e. 5 %, had to be assigned to new clusters by discretion. Detailed results are reported in the text and in Annex 3. DIW 66 EPRC

67 The cluster analysis that produced the most consistent results (see Tables in Annex 3) led to eight clusters that were re-assigned by discretion to the following six: 1. Cluster 1 "Agglomerations": Membership in this cluster was clearly determined by exceptional high values for population density. The cluster analysis produced two groups of agglomerations with different levels of population density. However, to simplify the classification it seemed sensible to combine both of them to a single cluster. This new cluster also includes now the exceptional case of the analysis which was "Inner London" with its extremely high population density. 2. Cluster 2 "Service dominated": All regions in this cluster show remarkably high shares in service employment (69 % and more compared to the EU 27 average of 62 %). Industrial and agricultural employment is below the average (8 % and 30 % respectively). 3. Cluster 3 "Service biased": This cluster combines regions with an aboveaverage share in services that are not dominated by services, i.e. they have also a significant share (close to average) of employment in either agriculture or industry. 4. Cluster 4 "Industry": These regions all have above average shares of industrial employment and no specific strength in either agriculture or services. 5. Cluster 5 "Agriculture biased": These regions show above average shares of agricultural employment. However, the values are less pronounced than in Cluster 6 and all regions in this cluster also have comparatively high shares in either industrial or service employment. 6. Cluster 6 "Agriculture dominated": This cluster combines two clusters that were produced by the computer analysis. Both were characterised by very high shares of agricultural employment. As one of these only consisted of three Romanian regions with extremely high values for agricultural employment (50 % and more) it seemed sensible to combine all agriculture dominated regions. First of all, it is interesting to note how Eastern and Western regions are distributed among the different clusters; this is shown in a map in Annex 1. Prague is the only Eastern region in the "Agglomeration"-Cluster that includes eleven Western regions (among them London, Wien, Berlin, Brussels and, admittedly, Ceuta y Mellila). There is also only one Eastern region, Bratislava, where employment is DIW 67 EPRC

68 dominated by services. Western regions in the service dominated cluster include tourist destinations like Islas Baleares and modern economic centres like Luxembourg; 16 of the 48 EU regions in this cluster come from the UK. (According to national Maltese data, Malta would probably be included here.) The service biased cluster shows also only very few AC regions, i.e. Sofia and Kozep-Magyarorszag (incl. Budapest). Among the 55 EU regions in this cluster are approx. half of the French and the Dutch as well as a quarter of the German regions, Denmark and almost all of the remaining UK regions. With eleven out of 66 regions the applicants have roughly a proportionate share in industrial regions. These comprise most of the remaining Czech as well as three Hungarian regions, Lubuskie and Slaskie in Poland and Bucharest. More than half of the industrial EU regions are from Germany. In addition, there are mainly some Italian and French regions in this cluster. The important role of agriculture in the AC is well-known and becomes evident also in the results of the cluster analysis. 19 of 50 agriculture biased regions and 17 of 27 agriculture dominated regions in the sample come from the applicants. Agriculture biased EU regions, i.e. with an above average share of agricultural employment but also some notable signs of industry and/or services, are mostly Spanish and Portuguese regions as well as Ireland and some Greek, Austrian and Finnish regions. From the AC come seven of the 16 Polish regions, all of Slovakia (apart from Bratislava), Cyprus, Estonia, Lithuania and Slovenia. Agriculture dominated is defined here as a share of agriculture in total employment of at least 20 % and a clearly underdeveloped service sector; some of the regions do show, however, an above average share of industrial employment (e.g. Centru/ROM and Dytiki Makedonia/GR). In this cluster are all Romanian and Bulgarian regions (apart from Bucharest and Sofia) as well as Latvia and seven Polish regions. EU regions in this cluster include Galicia in Spain, Centro in Portugal and eight Greek regions. This overall picture already hints at one severe development problem associated with the economic structure of the AC economies. Services are often a key to economic growth and a favourable employment situation. The service sector, however, is clearly underdeveloped in the applicant countries while it is quite important for the EU regions. Consequently, one half of all EU regions have an economic structure that is similar to only four of the 51 AC regions. In this part of the spectrum, the cluster DIW 68 EPRC

69 analysis accentuates differences between East and West rather than identifying similarities of specific regions. Services are, of course, not the only option for obtaining a satisfactory economic situation as the industrial orientation of the German economy demonstrates. Unfortunately, though, a clear sign of an "industrial tradition" that might serve as a basis for future economic development is only evident in a fifth of all AC regions. Two thirds of the AC regions even show an above average share of agricultural employment which is the case for only one fifth of the EU regions. Annex 3 comprises all regional data available for the six clusters constructed in this analysis; the table "national averages" serves for comparison purposes. The following text cannot aim to give a complete overview of all the information included in this annex but presents cluster by cluster some instructive results derived from a comparison of the AC region(s) included with its Western counterparts in the same cluster. With a view to a comparative analysis of AC and EU regions the first three clusters only have to be discussed briefly. Six of the 106 million inhabitants of the AC live in agglomerations (Prague) or in regions with a notable share of services in employment (Bratislava, Sofia, Budapest-Region). Mostly, the regions in these three clusters have a per capita GDP of at least the EU 15 average. In the case of the agglomerations, the sometimes quite high values are obviously distorted by commuter flows. The service dominated cluster comprises also some of the capital regions of the EU members and other economic centres with distinct above average GDP per capita. While there are some quite wealthy regions in the service biased cluster (Antwerpen, Oberbayern, Darmstadt, Groningen) the GDP per head tends to be somewhat lower than in the service dominated cluster and six EU regions in the cluster would qualify as Objective 1 region in Prague and Bratislava were the only AC regions with a GDP per capita above the EU 15 average in 1997 and were by no means the poorest members of their cluster, either then or in Both regions significantly improved their position relative to the EU 15 average from 1994 to 1997 (by 13.4 and 15.3 percentage points). This distinguishes them clearly from the EU regions in their clusters; only very few of them managed to improve their GDP per capita relative to the EU average significantly and only two, Noord-Holland and Uusimaa/SF, gained more than ten percentage points. The situation for Sofia and Budapest-Region was different. Both were poor by the DIW 69 EPRC

70 standards of their cluster in 1997 Sofia exceptionally so and Budapest-Region at a level with some Greek and Eastern German regions and Sardegna. Worse still, Sofia had a significantly lower GDP per head (relative to the EU 15 average) in 1997 than in 1994; Budapest's improvement was only meagre. The employment rate was comparatively high in Prague (and, consequently, productivity relatively low) as well as in Bratislava while GDP per head and productivity were roughly in line in Sofia and Budapest-Region. The labour market situation in Prague was notably better than in the other agglomerations by all available measures (total, female, youth and long-term unemployment). In Bratislava, the labour market situation was also quite favourable in 1998 but most of the Dutch and the UK regions in the service dominated cluster had a lower total unemployment rate than the Slovak capital (6.4 % in 1998). This holds also true for female and youth unemployment. However, long-term unemployment was less pressing in Bratislava than in almost all other service-dominated regions. The various unemployment measures for Sofia and the central Hungarian region are worse than for Prague and Bratislava but still below the EU 15 average (with the exception of youth unemployment in Sofia). They fit well with the rates of other regions in the "Service Biased"-cluster. The overall picture in this cluster is heterogeneous (and seems to be determined by country specific factors) but total unemployment rarely exceeds the EU 15 average. In the industry cluster, only some German and Italian regions as well as North Eastern Scotland have a GDP per head clearly above average. The values for the AC regions are without exception below those of their EU counterparts in this cluster with the richest AC region (Ostravsky/CZ, 59.3 %) just matching the poorest EU region (Dessau, 60.1 %). The AC regions managed to improve their relative GDP position slightly from 1994 to 1997 (with the exception of Bucharest and the Czech Severozapad) while quite a few of the EU regions in the cluster experienced a severe setback. Brandenburg, Staffordshire and Karlsruhe were most successful in this. Judged by their GDP per head there are clear intra-ceec differences in this cluster between, on the one hand, the Czech regions and Nyugat-Dunantul/HU with 50 % or more of the EU 15 average and, on the other extreme, Eszak-Magyarorszak/HU and Lubuskie in Poland with less than a third of the average. DIW 70 EPRC

71 In general, differences in productivity were less pronounced among the AC industrial regions than differences in GDP per head of population, i.e. the employment rate in the Czech regions was comparably high. However, the EU regions tend to show a higher employment rate than the AC regions in this cluster. The industrial cluster comprises some of the well-known labour market problem regions in Spain, Eastern Germany and other regions. Apart from these, unemployment was below average for the EU regions. Unemployment in the Eastern industrial regions was notable in the two very poor regions but also in the relative rich Czech regions Ostravsky and Severozapad (ranging from almost 10 to 13 %). Labour market problems among the Eastern regions concentrated in these regions (plus high youth unemployment in Bucharest) but were not as pronounced than, for example, youth unemployment in some of the Spanish, French and Italian regions in the cluster. The "agriculture biased" cluster consists of regions that have an above average share of agriculture in total employment (average: 8 %; in the cluster: between 8 and 20 %) but do also show strength in either industry or services. In the case of the AC regions in this cluster, this means almost always a significant share of industry. Just Cyprus had a distinctly above average share of services, few of the other regions showed a balanced structure (Mazowieckie/PL, Lithuania). The regions of the EU members in this cluster were more heterogeneous in this respect. The Austrian regions combined their agriculture bias with high values for industrial employment while the Greek and the Italian regions tend to show a higher share of employment in services; the employment structure of the regions from other EU members was mixed. Only three of the regions in this cluster, Ireland, Oberösterreich and Navarra, had at least an average GDP per head (compared to EU 15). The rest of the cluster includes, among others, some of the poorest regions of the EU and the AC. With Cyprus, Slovenia, Jihozapad/CZ, Jihovychod/CZ and Mazowieckie/PL the cluster includes some AC regions that are comparatively well-off. From 1994 to 1997 most of the regions in this cluster improved their relative GDP per head (most of all Ireland, Alentejo and Mazowieckie). This applied for all AC regions in the cluster (apart from the three Hungarian regions). In the comparatively well-off AC regions the employment rates were above average. This was also the case for the Austrian and Portuguese regions while they were quite low in some of the Spanish and Italian regions. With only few exceptions (mostly in DIW 71 EPRC

72 Austria and Portugal) the agriculture biased regions had dissatisfying unemployment rates. Female and youth unemployment was also high, especially in Spain and Italy. Among the AC regions in this cluster, labour market problems were concentrated in a few regions, mostly Warminsko-Mazurskie and Zachodniopomorskie in Poland as well as Zapadne and Vychodne Slovensko in the Slovak Republic. The composition of the agriculture dominated cluster differs from that of the other clusters because only a few countries have regions in the cluster and these mostly form large parts of the affected countries. The cluster mainly consists of Romania and Bulgaria (complete except the capital regions), Latvia, eight of 13 Greek and seven of 16 Polish regions. All these regions had a very low per capita GDP, the distinctly richest region being Crete with 70 % of the EU 15 average. However, there still is a clear East-West-difference in this cluster. The poorest EU region (Ipeiros, 42 %) still was well above the richest AC region in the cluster (Centru/ROM, 34 %). None of the regions in the cluster managed to grow above average between 1994 and 1997 (with the exception of Centro/PT and Galicia/ESP). Only very few regions in the cluster had to cope with distinctly above average total unemployment (Bulgaria and Galicia) but the unemployment rates of the EU regions were markedly higher in 1998 than they have been in Female and youth unemployment was quite high, however, in most of the EU and some of the AC regions. 3.5 Island Economies The island economies of Cyprus and Malta, together with Slovenia, are the most prosperous of the 12 accession countries. Cyprus and Malta are differentiated from the CEE candidate countries in two critical areas. First, they are both island economies with very small populations in 1997, Cyprus had 741,000 and Malta 374,000 inhabitants. Second, neither country is undergoing economic, political, social or environmental upheaval on the scale of the CEE countries. Both islands gained independence from the United Kingdom in the early 1960s and are strongly dependent on foreign trade and FDI as economic growth drivers in light of their small domestic markets. The following sections outline the main characteristics of the two island economies in more detail and, where they are evident, highlight relevant issues relating to regional disparities and socio-economic cohesion. DIW 72 EPRC

73 3.5.1 Malta The Maltese economy is characterized by a protected domestic-oriented sector and an open export-oriented sector (UNECE, 2000). The former sector comprises both state and non-state owned monopolies, particularly in utilities, as well as small manufacturing firms producing goods and services for the domestic market, most of which have been broadly protected from external competition. The export-oriented sector is built up of manufacturing firms, established with FDI, and a tourism sector, which is an important source of growth and is dominated by Maltese capital. Overall, the protection of the former sector has diverted resources from the latter and limited economic growth. The application for EU membership has highlighted the need for restructuring in the protected manufacturing industries and, from October 1999, Malta has begun to remove remaining trade barriers with a target completion date of Malta s level of per capita GDP has increased by approximately 20 per cent in real terms since 1993, but slowed towards the end of the 1990s and remains relatively low in comparison to the EU average. While Maltese statistics do not allow for direct international comparison in purchasing power parities, figures from the UNECE (2000) show that Malta had a comparative per capita GDP level in current prices of 42.7 (assuming EU=100) in Private consumption was the main component of growth in 1999 and the rate of fixed investment was weak. In terms of economic structure, the share of agriculture and fisheries in GDP is small. Agriculture accounted for 2.8 % of GDP in 1998 and 1.6 % of total employment in The fishing industry accounts for three percent of GDP and provides direct employment for about 2,500 people. 95 % of the production is exported to EU Member States (mainly Italy). Industrial production accounts for less than one third of GDP (27.4 % in 1998) (Government of Malta, 1999). Shipbuilding and repair continue to be the largest industrial sub-sector although, despite restructuring attempts, it makes a negative contribution to GDP. Indeed, shipbuilding and repair have had a high profile in EC reports on Maltese membership, particularly in light of the high level of state aid received by this sector which conflicts with EU competition rules. Special derogations are likely to be requested by the Maltese government for a transitional period after membership (UNECE, 2000). Electrical machinery production is an important contributor in terms of exports, investment and employment and its DIW 73 EPRC

74 share of total manufacturing value added increased from 15.2 % in 1993 to 27.9 % in The role of services has strengthened over time (accounting in 1998 for nearly 70 % of GDP) and private market services now account for around 33 % of total employment (UNECE, 2000). This reflects growth both in traditional service industries (e.g. tourism and market services) and expansion in new sectors such as financial services, transhipment activities from the Malta Free Port and ship registration. Tourism remains the most important services sub-sector - more than one million tourists visit Malta each year, four-fifths of whom are from the EU (Government of Malta, 1999). Tourism income accounted in 1998 for 21.3 % of the value of exports of all goods and services and 6.3 % of total employment. It also has an important multiplier effect on the remainder of the economy. Malta's economy overall is dominated by SMEs, mainly involved in the manufacturing industry and services. The most important SME sectors in Malta are the construction industry, transport and motor vehicle repair, clothing, food and beverages and tourism. SMEs in all sectors share a number of structural weaknesses associated with their size including distance from export markets, machinery operating below capacity, a tight labour market and lack of expertise in management and sales promotion, and limited access to finance (CEC, 1999b). The EU continues to be Malta's principal export market, absorbing just over half of Malta's exports in 1999 and supplying nearly 70 % of its imports (UNECE, 2000). The level of exports to the EU has declined in the 1990s, from a starting level in 1990 of nearly 85 %, due in part to a sharp drop in exports to Italy (falling from 37 % in 1990 to five percent in 1999). Clothing was the dominant Maltese export to the EU until the start of the 1980s after which the share of electronic goods increased to become the leading export. Change in export composition has also contributed to an amended trade direction with North American and Asian markets increasing in importance in the 1990s exports to America have increased from 3.9 to 22.2 % over the period while the Asian export share rose from 5.2 to 20.5 % over the same time period (UNECE, 2000). A recent Maltese government report (1999) highlights the susceptibility of the island s trade performance to sudden shifts in international demand as a result of the high degree of concentration of merchandise exports in a single sector product (one semi- DIW 74 EPRC

75 conductor firm accounts for approximately 50 % of Malta's total merchandise exports). The ability of the Maltese economy to cope with the competitive pressures within the Union remains a challenge. This is particularly true for small enterprises in sectors such as agro-industry, services, handicraft and furniture. Unemployment has historically been low but has increased recently from an average rate of % to 5.1 % in June 1999, reflecting the slowdown in the economy (CEC, 1999b). The increase in the unemployment rate primarily reflects a fall in total employment, although the increased size of the labour force is also a factor (European Parliament, 1998). One challenge for the Maltese labour market is a reduction in public sector employment. Further, the processes of restructuring and privatisation, likely to be accelerated by the prospect of EU membership, will have negative implications for the labour market in the short- to medium-term in the absence of compensatory economic growth Cyprus Cyprus is a slightly larger island economy which, according to UNECE figures, had a comparative per capita GDP level in 1998 (in current purchasing power parities) of 78.5, assuming that EU=100 (Discussion in this section generally does not include data for the Turkish occupied part of the island). The Cypriot economy has grown steadily over the past three decades, driven in the 1970s and 1980s mainly by manufacturing and, in the last two decades, by tourism and other services, including financial and business services. Inflation is very low (1.7 % in 1999) and unemployment, although rising slightly in recent years, has remained below four percent since the mid-1990s and stood at 3.7 % in 1999 (UNECE, 2000). Agricultural production accounts for around five per cent of GDP. Manufacturing industry accounted for about 12 % of GDP (a decline from the position at the end of the 1980s) and 16 % of employment in 1998 (Government of Cyprus Press and Information Office). The most important sectors in terms of value added are food and beverages, apparel and footwear, and metal products. The development of manufacturing was initially geared to supplying the local market but, as opportunities for importsubstitution became exhausted, the emphasis switched to exports. More recently, manufacturing industry has experienced periods of decline in production, exports and employment. This is related to an erosion of competitiveness, both abroad and in the local market, due to rising production costs and insufficient productivity gains, at a DIW 75 EPRC

76 time of intensified international competition (Government of Cyprus Press and Information Office, 1999). According to official estimates, the per capita GDP of people employed in manufacturing is just % of the level of Spain (CEC, 1999a). Overall, Cypriot industry faces structural weaknesses such as labour shortages, lack of domestic raw materials and a small domestic market, leading to competitiveness problems in the face of low cost competitors or high wage-producers competing on quality (Bachtler/Downes/Helinska-Hughes/Macquarrie, 1999). Accession is likely to increase these pressures, and the Cypriot government has recently introduced measures to promote the development of high-tech industries, attract FDI and improve export competitiveness of domestic firms. In terms of services, Cyprus has become an important off-shore centre for many (mainly European) companies providing financial, legal, accounting, maritime and other services which have been attracted by favourable tax treatment (UNECE, 2000). A number of the tax implications of EU accession and the adoption of the acquis could have important negative ramifications for the island. As in Malta, tourism has also grown considerably and, including indirect impact on other sectors such as construction and retail/wholesale trade, accounted in 1998 for 20 % of GDP and 15 % of total employment. The EU accounts for over 60 per cent of income from tourism (CEC, 1999a). An upgrading of tourist services is underway as part of an attempt to compete within the Mediterranean market. The service sector is an important growth area, though there are some concerns about over dependence upon tourism. In 1995, jobs in services accounted for over 63 % of the total following a significant shift away from agriculture and, to a lesser extent industry (CEC, 1999a). Tourist arrivals in Cyprus reached the level of 1.95 million in 1996, 2.1 million in 1997 and 2.2 million in 1998 (Government of Cyprus, Press and Information Office, 1999). Revenues form tourism also increased. During 1999, an acceleration of foreign demand for services was mainly due to favourable development in the tourist sector (Central Bank of Cyprus, 1999). A more recent particular growth area has been in the field of international business and offshore banking (Bachtler/Downes/Helinska-Hughes/Mcquarie 1999). Registered unemployment has risen only slightly from 3.1 % in 1996 to 3.4 per cent in Unemployment remains low even among young people and women, the only group for whom the rate is high are those over 50 (CEC, 1999a). Labour shortages have materialised, espe- DIW 76 EPRC

77 cially in some activities demanding high skills and may constrain growth in the next few years (CEC, 1999a). Cyprus has a relatively open economy and is strongly dependent upon foreign trade (Ayres, 1999). Over the period , the value of foreign trade increased by a moderate annual average growth rate of seven per cent (Government of Cyprus Press and Information Office, 1999). The composition of trade generally reflects the structure of the economy. The main exports include clothing, footwear, potatoes and citrus fruits (CEC, 1999a). EU countries accounted for 54.7 % of total Cypriot imports and 40 % of exports in The United Kingdom and Greece are the main trading partners of Cyprus in terms of both exports and imports. The transition countries of CEE, and Russia in particular, have become increasingly important trading partners for Cypriot exports, with their share rising from 8.5 % in 1992 to 25.9 % in 1998 (UN- ECE, 2000). DIW 77 EPRC

78 4 Effects of Enlargement: Critical Issues The process of economic change and liberalisation in the CEECs has had important spatial ramifications, with patterns of increasing regional disparity emerging (see chapter 3). The enlargement of the European Union will also not be a spatially homogenous process, but will have differing impacts both on the regions of the current EU 15 and those of the CEECs. Much of the literature written on the effects of enlargement has focussed on the national level, with the term regional often denoting either the CEE region as a whole or national level breakdowns within it. Much less empirical research has been done on the effects of enlargement on sub-national units within the CEECs or the EU 15. The uncertainties of the enlargement process, including a shifting timetable, differing patterns and speeds of economic reform and development in the CEECs and economic change in the EU 15, and continually evolving links between CEECs and EU 15 countries in terms of trade, investment and labour flow make the assessment of effects of enlargement difficult even at national level. A recent analysis of various research studies looking at the costs and benefits of enlargement to include the Czech Republic, Hungary, Poland and Slovenia over the period concludes that while several questions have been clarified, there still is a high degree of uncertainty in important aspects of eastern enlargement. This exacerbates the constraints of applying model-based methodologies and makes robust assessments of accession-related costs and benefits extremely difficult to produce (Mortensen/Richter, 2000). The challenge of assessing the effects of enlargement at regional level is even greater - although it is likely that some of the greatest effects of enlargement will be felt at this level. In this context, the following section analyses the critical issues of enlargement, focusing on trade, investment and migration. The following discussion will tackle these issues separately. However, they are also interconnected issues. For instance, capital movements to the CEECs can increase trade and, thus be neutral or beneficial for wages and employment in the affiliated sector. The spatial implications of these issues are often, by necessity, given at national level although, wherever possible, the regional impact is drawn out. The chapter also includes a short analysis of the particular issue of border regions in the eastern enlargement of the EU. DIW 78 EPRC

79 4.1 Trade Trade patterns have been fundamentally realigned since the initiation of economic and political reform at the start of the 1990s. The major shift has been the reorientation of economic links from the former CMEA countries to the EU 15. Measures to liberalise domestic trading conditions and support international trade were made early on in the transition process and included, for example, the removal of national trade restrictions and barriers of the benefit of domestic companies and the introduction of currency convertibility at international rates (Brada, 1998; Desai, 1998; Drabek and Brada, 1998). The initial Interim or Free Trade Agreements, and the subsequent Europe Agreements also encouraged trade with the EU 15 through the dismantling of tariff barriers for many products. While these Agreements were asymmetrical, leaving the CEECs more time for adjustment, restrictions were maintained on EU sensitive products which continued to protect important sections of the EU market from CEE exports. The EU is now clearly the main trading partner of the CEECs, built on factors such as a freer trade regime, geographical proximity, the rapid exploitation of new market opportunities by some EU countries, government policy and the advantages also to the CEECs of lower transport and information costs associated with market entry. As early as 1993, EU countries accounted for around half of Polish, Czech and Hungarian imports and, by the late 1990s, the EU market accounted for percent of all CEE exports approximately the importance of the EU market for EU nations themselves. Despite high growth rates of trade, the CEE market is much less important to EU exports and imports, taking only around 4-5 percent of the EU 15 exports in 1997 although significant country variation is evident, with the CEECs being more significant for the main border countries of Germany, Greece, Finland, Italy and Austria (Baldwin et al., 1997; Boeri/Brücker et al., 2000; Grabbe/Hughes, 1997; Richter, 1998) Trade Volume The beginning of the 1990s was marked by a significant rise in the volume of trade between the CEECs and the EU. In 1998, EU exports to the CEECs were on average 7 times higher than in 1990 and imports were 5 times higher (see table 4.1.1). DIW 79 EPRC

80 Table 4.1.1: EU Member States' Exports, Imports and Net-Exports with AC-12 Exports in Mio Euro Imports in Mio Euro /'90 98/' /'90 98/'95 FRANCE BEL/LUX NETHERLANDS GERMANY ITALY UNITED KINGDOM IRELAND DENMARK GREECE PORTUGAL SPAIN SWEDEN FINLAND AUSTRIA EUR (of reported data) EU Members Net-Exports in Mio Euro FRANCE BEL/LUX NETHERLANDS GERMANY ITALY UNITED KINGDOM IRELAND DENMARK GREECE PORTUGAL SPAIN SWEDEN FINLAND AUSTRIA EUR (of reported data) Exports, EU 15=100 Imports, EU 15= FRANCE BEL/LUX NETHERLANDS GERMANY ITALY UNITED KINGDOM IRELAND DENMARK GREECE PORTUGAL SPAIN SWEDEN FINLAND AUSTRIA EUR (of reported data) Source: Eurostat Comext. DIW 80 EPRC

81 The trade surplus of the EU in 1998 over the CEECs was 25 billion Euro. The biggest net-exporters to the CEECs among the EU 15 were Germany (7.2 billion Euro), Italy (4.3 billion Euro), France, Finland and Austria (see table 4.1.1). For 1999, preliminary figures show a decreased surplus due to sharply increased imports from the CEECs (Statistics in focus external trade, 10/2000). Among the CEECs, Poland has the highest trade deficit with the EU followed by the Czech Republic and Hungary (see table 4.1.2). Germany is by far the most important trading partner of the CEECs. In 1998, 42 % of all EU exports to the CEECs and 47 % of all EU imports from the CEECs were conducted with Germany. Italy followed with 13 % of all EU exports to the CEECs and 11.5 % of all imports (see table 4.1.3). Austria has also gained from rapidly expanding economic relations with CEECs, according to Richter (1998). Palme (1999) estimates that approximately 200,000 production workers in Austria have worked on additional exports to key CEE markets between 1989 and In the same period, Austria s manufacturing jobs are estimated to have grown by an additional 0.2 to 0.5 % points per annum owing to additional foreign trade with CEECs. On the basis of this type of evidence, Mayhew (1998) and Egger (1999) anticipate that enlargement in combination with continued economic growth in CEE countries, is expected to further increase the importance of CEE markets to EU exporters. Malta has strong trading relations with the EU. In 1998, 70 % of all imports came from the EU and 50 % of the Maltese exports went into the EU with France and Italy being the most important trading partners (Statistics in focus external trade, 7/2000). Cyprus on the other hand has stronger trade relations to countries outside the EU than most other candidate countries. In 1998, only 38 % of their exports went into EU member states and 55 % of their imports came from the EU with the UK and Greece being the most important trading partner among the EU member states. Malta and Cyprus both accounted for 2 % of all exports from the candidate countries to the EU and for 1 % of all their imports from the EU in 1998 (see table 4.1.2). Geographical proximity plays a key role in trade between the EU and the CEECs. The border countries Germany, Greece, Finland, Austria as well as Italy account for two-thirds of the trade with the CEECs. Their share of EU exports to and imports from the CEECs is considerably higher than in overall EU trade, whereas the share of France and the UK is rather small (see table 4.1.3). DIW 81 EPRC

82 Table 4.1.2: EU 15 Exports, Imports and Net-Exports with individual Applicants Exports in Mio Euro Imports in Mio Euro /'90 98/' /'90 98/'93 MALTA ESTONIA LATVIA LITHUANIA POLAND CZECH REP SLOVAKIA HUNGARY ROMANIA BULGARIA SLOVENIA CYPRUS CEEC (of reported data) EU-15 Net-Exports in Mio Euro MALTA ESTONIA LATVIA LITHUANIA POLAND CZECH REP SLOVAKIA HUNGARY ROMANIA BULGARIA SLOVENIA CYPRUS CEEC (of reported data) Exports, CEEC=100 Imports, CEEC= MALTA ESTONIA LATVIA LITHUANIA POLAND CZECH REP SLOVAKIA HUNGARY ROMANIA BULGARIA SLOVENIA CYPRUS CEEC (of reported data) Source: Eurostat Comext. DIW 82 EPRC

83 Table 4.1.3: Trade data of the EU 15, 1998 Trade of EU 15, in Mio Euro Trade with AC-12 Total foreign trade Exports Imports Exports Imports FRANCE BELGIUM/LUXEMBOURG NETHERLANDS GERMANY ITALY UNITED KINGDOM IRELAND DENMARK GREECE PORTUGAL SPAIN SWEDEN FINLAND AUSTRIA EUR Trade of EU 15, EU 15 = 100 Trade with AC-12 Total foreign trade Exports Imports Exports Imports FRANCE BELGIUM/LUXEMBOURG NETHERLANDS GERMANY ITALY UNITED KINGDOM IRELAND DENMARK GREECE PORTUGAL SPAIN SWEDEN FINLAND AUSTRIA EUR Share of trade with AC-12 in total foreign trade Exports Imports FRANCE 3 2 BELGIUM/LUXEMBOURG 2 2 NETHERLANDS 3 2 GERMANY 8 8 ITALY 6 4 UNITED KINGDOM 2 1 IRELAND 1 1 DENMARK 4 3 GREECE 13 4 PORTUGAL 1 1 SPAIN 2 1 SWEDEN 4 3 FINLAND 8 4 AUSTRIA EUR 5 4 Source: Eurostat Comext. DIW 83 EPRC

84 DIW 84 EPRC

85 This country variation is also true for the CEECs. In 1998, Poland, the Czech Republic and Hungary accounted for two-thirds of CEECs exports and imports with the EU (see table 4.1.2). These three countries now belong to the ten most important trading partners of the EU. According to calculations by Schumacher/Trübswetter (2000), EU trade with the CEECs reached in 1997 on average only about 50 % of exports and 60 % of imports to be expected from normal trade relations estimated on the basis of distance and GDP. This indicates that overall trade between the EU and the CEECs will most probably rise significantly during the next years. However, this high growth potential does not apply to all EU member states. Estimates from Boeri/Brücker et al. (2000) show that Austria, Germany and other countries bordering the CEECs have reached or even surpassed the normal volume of trade with the CEECs in Nevertheless, there are no indications that the growth of trade between the neighbouring countries of the EU 15 and the CEECs is slowing down Trade Structure According to international trade theory such as the Heckscher-Ohlin-theory, comparative advantage arises from differences in national factor endowments. Consequently, the CEECs should have a comparative advantage in the production of labour-intensive and resource-intensive products, whereas the EU member states should have a comparative advantage in the production of (human-) capital-intensive and R&D-intensive products. By and large, this expectation was confirmed by data on the commodity structure of trade for the early 1990s between the EU and the CEECs with CEE exports to the EU leaning towards resource-based and labourintensive products and EU exports to the CEECs leaning more towards capitalintensive goods and higher labour and R&D skill products (Raines/Bachtler, 1997). During the 1990s, the development of trade between the EU and the CEECs was characterised by a decline in inter-industry trade and a strong increase in intraindustry trade (IIT). Recent studies emphasise the future role of intra-industry trade and vertical differentiation of products in trade between the EU and the CEECs (Boeri/Brücker et al., 2000; Aturupane/Djankov/Hoekman, 1997; Landesmann/Burgstaller, 1997). The increase of intra-industry trade might seem to imply that CEECs managed to increase their exports of highly specialised products. However, an DIW 85 EPRC

86 analysis of the unit values does not confirm this view. Intra-industry trade can be distinguished between horizontal (two-way trade in products of similar quality, but different characteristics or attributes) and vertical IIT (trade of similar products of varying qualities and therefore varying unit values). Theory suggests that the more similar countries are in terms of their factor endowments, their technological development and other production conditions, the greater is the share of horizontal IIT (Aturupane et al., 1999). The trade analysis in Boeri/Brücker et al. (2000) shows that unit values of goods produced in the EU and goods produced in the CEECs differ strongly within the same commodity group. The EU members specialise in the higher price and quality segments of markets with a high content of human capital and R&D, whereas the CEECs specialise in the lower price and quality segments of markets with a lower content of human capital and R&D. The absolute levels of unit values in 1996 in EU imports from the CEECs were more or less comparable to those of developing countries (Boeri/Brücker et al., 2000). Burgstaller/Landesmann (1999) identify Hungary as the only exception as it is moving towards more technology- and skill-intensive engineering industries. Other CEECs such as Poland, Bulgaria and Romania tend to concentrate more on labourand less skill-intensive (variations of) products. The overwhelming share of intraindustry trade, in fact around 80 to 90 %, between the EU and the CEECs is trade in vertically differentiated goods, which indicates great differences in factor endowments (Aturupane et al., 1999). In 1995, horizontal IIT accounted for much less than 10 % of total trade with the EU. It was comparably high for the Czech Republic and Slovenia with around 8 % and comparably low for Bulgaria with about 2 %. These figures also imply that due to the large differences in unit values CEECs do not yet compete with Southern EU 15 member states in the same market segments. Thus the fear of these EU members that the Eastern Enlargement would bring a displacement of their exports cannot be verified. It was not possible within the framework of this report to calculate and to analyse bilateral trade flows between the EU and the CEECs in an extensive way. Nevertheless, some results can be drawn by regarding trade data from selected bilateral trade flows between the EU and three CEECs (Poland, the Czech Republic and Hungary). Data for EU imports from Poland, the Czech Republic and Hungary and exports to these countries on CN2 basis (2-digits of the Combined Nomenclature) seem to DIW 86 EPRC

87 show a rather similar commodity structure. EU exports and imports are mostly concentrated on three product groups: vehicles, machinery and mechanical appliances as well as electrical machinery and equipment (see table 4.1.4). Thus it seems that the commodity compositions of these CEECs are increasingly reaching proportions characteristic of intra-eu trade (Korhonen/Randveer, 2000; Richter, 1998). Still, significant differentiations can be distinguished. The share of Electrical machinery and equipment and vehicles and parts thereof in Polish and Czech exports to the EU was in 1998 with 20 % and 30 % relatively high but still significantly lower than the export share of these product groups from the EU to these two countries. The Hungarian export structure is the most similar to the EU export structure. In 1998, the three product groups vehicles and parts thereof, machinery and mechanical appliances and electrical machinery and equipment accounted for 58 % of all EU imports from Hungary (see table 4.1.4) and for 56 % of all EU exports to Hungary. This corresponds to other findings, which claim that Hungary s exports to the EU, are based on more human capital intensive, medium-to high technology products (Kaminski, 2000) while those of other CEECs such as the Czech Republic and Poland are still relatively more labour intensive (Lukas, 1999). A closer look on CN4 digits, i.e. a more disaggregated commodity structure, for the product group 84 machinery and mechanical appliances and parts thereof, shows even further difference between import and export goods of Poland and the Czech Republic and the EU. Polish and Czech exports among the CN2 digit 84 are concentrated on parts for engines and machinery (8409, 8431, 8466) and appliances for pipes and pumps (8481, 8413), which are not distinguished from finished products on CN2 basis and which do not require a very high level of R&D or human capital (see table 4.1.5). Again, the export structure of Hungary looks the most similar to the structure of EU exports with the product groups of automatic data-processing machines and units thereof (8471) and parts for machines (8473) accounting for 29 % of all Hungarian exports into the EU among the CN2 group 84. These two product groups also accounted for 25 % of EU exports to Hungary in this CN2 group. Thus these results support the findings of Boeri/Brücker et al. (2000) stated earlier that the trade between the CEECs and the EU is still dominated by vertical IIT. DIW 87 EPRC

88 Table 4.1.4: EU exports and imports with selected Applicants by selected product group Poland EU-Exports = 100 EU-Imports = 100 NACE categories Mineral fuels, oils Pharmaceutical products Plastics and articles thereof Wood and articles of wood Paper & paperboard Art of apparel & clothing Iron and steel Articles of iron or steel machinery/mechanical applian Electrical mchy equip/ parts Vehicles, parts thereof Optical, photo, cine, meas Furniture; bedding Sum of chosen products Sum of all exports Czech Republic EU-Exports = 100 EU-Imports = 100 NACE categories Mineral fuels, oils Pharmaceutical products Plastics and articles thereof Wood and articles of wood Paper & paperboard Art of apparel & clothing Iron and steel Articles of iron or steel machinery/mechanical applian Electrical mchy equip/ parts Vehicles, parts thereof Optical, photo, cine, meas Furniture; bedding Sum of chosen products Sum of all exports Hungary EU-Exports = 100 EU-Imports = 100 NACE categories Mineral fuels, oils Pharmaceutical products Plastics and articles thereof Wood and articles of wood Paper & paperboard Art of apparel & clothing Iron and steel Articles of iron or steel machinery/mechanical applian Electrical mchy equip/ parts Vehicles, parts thereof Optical, photo, cine, meas Furniture; bedding Sum of chosen products Sum of all exports Source: Eurostat Comext. DIW 88 EPRC

89 Table 4.1.5: EU 15 Imports from selected Applicants, sum of product group No. 84=100 Poland Czech Republic NACE categories Parts for engines (mostly aircraft and marine engines) Pumps for Liquids Refrigerators. Freezers Parts for machinery (mostly lifting or loading machinery) Harvesting machinery Parts for machine-tools Automatic data-processing machines Parts for offices machines Aplliances for pipes sum of chosen products sum of product group No Hungary NACE categories Parts for engines (mostly aircraft and marine engines) Pumps for Liquids Refrigerators. Freezers Parts for machinery (mostly lifting or loading machinery) Harvesting machinery Parts for machine-tools Automatic data-processing machines Parts for offices machines Aplliances for pipes sum of chosen products sum of product group No Source: Eurostat Comext. In general the trade composition of CEECs is in line with expectations based on their expected comparative advantages. The CEECs trade composition is currently more labour- and resource-intensive. It has to be noted, though, that trade patterns among the CEECs are very diverse. Romania, Poland and the Slovak Republic have a distinct comparative advantage in labour intensive goods, the Baltic countries and Bulgaria have a distinct comparative advantage in resource intensive goods (Freudenberg/Lemoine, 1999). Slovenia, Hungary and the Czech Republic, which have a significantly higher share of intra-industry trade with the EU (Unguru, 1999) reveal a much lower comparative disadvantage in human capital intensive goods vis-à-vis the EU. The EU member states in general concentrate on (human-)capital-intensive and technology-intensive products. Nevertheless, the comparative advantage of the EU DIW 89 EPRC

90 vis-à-vis the CEECs in human capital and technology intensive goods can be seen to decline over time (Quaisser, 1999). Goldin (1998) argues that advanced CEECs have a relatively highly educated workforce and their long run comparative advantage is anticipated to lie in more high skilled, high-tech sectors than at present. Mayhew (1998) further claims that as restructuring progresses and as domestic CEE producers and foreign investors improve the quality of locally produced consumer goods, the potential for CEE exports to the current EU members will increase and the amount of horizontal IIT will expand. Trade in agricultural products is the area where most potential post-accession impact may occur. However, even in this area, existing barriers to agricultural trade will start to be dismantled now to help prevent subsequent enlargement overly disrupting markets on either side. This is expected to bring significant pre-accession increases in trade. There are considerable questions over the competitiveness of this sector in the CEECs, particularly in light of its current protection from strong competition and the structure and productivity of the industry. In the short-term, adjustment strains from exposure of the CEE countries to competition could be considerable, not least in the shedding of surplus labour (Avery and Cameron, 1998). The dismantling of remaining trade barriers in this area, together with parallel developments such as the integration of new CEE Member States into the Common Agricultural Policy (in whatever form) and the adoption of western production techniques, is expected to improve the competitiveness of the CEE agricultural sector in the medium- to long-term. In sectoral terms (Huber and Pichelmann, 1998; Illés, 1997), the product structure in agriculture in the CEE countries was found to correspond extensively to that of the rural regions of Western Europe, with production in Mediterranean countries viewed as complementary to CEE. If the CEECs succeeded in increasing their currently low rates of productivity in comparison to EU levels, the peripheral rural regions of the western EU Member States could eventually come under particular adaptation pressure. Within the CEE countries, this would require continued privatisation as well as land reform, capital intensification and infrastructure renewal Regional and Social Impact As was stated earlier there are immense differences in trade volume and structure between the individual CEECs, but there are also significant differences between regions within a country. So after the overview over national trade data, it is important DIW 90 EPRC

91 to have a closer look at the regional impact as well as at the social impact of an eastern enlargement. Unfortunately, data on the regional level is less reliable because of different accounting standards and also less available and accessible. For these reasons, this report had to concentrate its regional analysis on Germany and Poland, both the most important participants in East-West trade among the EU and the CEECs respectively. For Germany, the regional data is more or less complete. The search for regional trade data of Poland was more difficult. All 16 statistical offices of the Polish Provinces were contacted but only four of them provided useable trade data. The key role of geographical proximity in the national distribution of trade between the EU and the CEECs suggests that EU regions bordering on the CEECs should trade at more than proportional levels (Boeri/Brücker et al., 2000). The following data from Germany and Poland supports this view: The shares of the CEECs in the trade of the German Länder neighbouring the candidate countries are twice as high as in the trade of the Länder in the west of Germany. In fact, in Eastern Germany not only high levels but also strong dynamics in trade with candidate countries can be observed. On average the share of CEECs in total trade turnover in the Länder in Eastern Germany in 1997 was over 15 %, whereas this share in the Länder in Western Germany was only around 6 % (see figure 4.1.1). The case for Poland seems to be similar, although the data basis is much smaller here. Data was available for the Provinces Opole, Zielona Gora, Jelenia Gora and Mazowieckie. In 1998 there was a restructuring of all Provinces in Poland and the number of 49 Provinces earlier to 1998 was reduced to 16 Provinces. The data for the first three Provinces Opole, Zielona Gora, Jelenia Gora is from the old structure of Provinces, whereas the data for the fourth Province is from the new Province Mazowieckie. DIW 91 EPRC

92 Figure Germany: Regional Structure of Trade with the CEECs 1997 DENMARK SCHLESWIG- HOLSTEIN 5.2 MECKLENBURG- WESTERN POMERANIA 16.6 NETHERLANDS NORTHRHINE- WESTPHALIA 6.5 BRE- MEN (STATE) 3.2 HAM- BURG (STATE) 3.5 LOWER SAXONY 8.4 SAXONY- ANHALT 14.7 BRANDENBURG 14.2 BERLIN (STATE) 9.4 PO LAND BELG IUM LUXEMB. RHINELAND- PALATINATE 6.7 HESSEN 5.5 THURINGIA 11.0 SAXONY 20.2 CZECH REPUBLIC SAAR- LAND 4.6 BAVARIA 8.9 FRANCE BADEN- WURTTEMBERG 6.5 AUSTRIA SWITZERLAND LIEC HTENSTEIN Share of CEEC-10 in total trade turnover in %: < 5 5 to 8 > 8 to 12 > 12 Source: Unpublished data of the German Federal Statistical Office, Boeri/Brücker et al. (2000). DIW 92 EPRC

93 Even though the economic activity in Western Poland is much lower compared to the economic centre around Warsaw, the two Provinces Zielona Gora and Jelenia Gora in Western Poland as well as the Province Opole trade with the EU at more than average level (see table 4.1.6). Between 80 and 82 % of all exports from these three Provinces go into the EU and between 73 and 81 % of all imports come from the EU. In contrast the national average level is with 64 % for both exports and imports much lower. Nevertheless, the economic impact of the three Provinces Zielona Gora, Jelenia Gora and Opole is rather limited. These three Provinces together account only for 5 % of Poland s exports, whereas the Province Mazowieckie alone accounts for around 20 % of Poland s exports. Table 4.1.6: Regional Trade Data for Poland Vojevoidship Opole (1997) Vojevoidship Zielona Gora (1998) Vojevoidship Jelena Gora (1997) Vojevoidship Mazowieckie (1999) Poland (1997) in Mio $ in % in Mio $ in % in Mio $ in % in Mio $ in % in Mio $ in % Exports Total Share of EU Imports Total Share of EU Exports in % of reg. GDP ) Imports in % of reg. GDP ) Regional GDP p.c. in % of nat. GDP p.c Source: National and Provincial Statistical Offices. A European Parliament Briefing Paper (1999b) reports that accession will generate sectoral and regional winners and losers, notwithstanding the overall positive balance for the European Union. Large expansions of industrial production are anticipated for the Länder of Vorarlberg, the Tyrol, Vienna, Lower Austria, Upper Austria and Styria (Palme, 1999). Palme views industries around major cities and in the highly centralised areas of the Länder as well positioned to exploit increasing returns to scale brought about by eastwards enlargement of the EU. DIW 93 EPRC

94 Conversely, EU producers of labour intensive goods (e.g. textiles, footwear and leather products) and of capital and scale intensive goods with a low degree of technological sophistication (printing, chemicals, plastics and rubber product) will continue to lose out through increasing competition (European Parliament, 1999b). Bachtler, Downes, Helinska-Hughes, Macquarie (1999) predict that enlargement would create competitive pressures particularly for those EU countries (and regions) which depend on: agriculture, with a typical production range for central Europe (grain, vegetables, fruit, cattle and pig farming); industrial operations with labour intensive but not particularly human resource intensive production; primary chemicals plants or shipyards with relatively low level product ranges; and assembly plants. It is likely that there will be localised instances where increased competition from CEE producers will have a detrimental effect. For instance, labour and energy intensive production sectors, e.g. in eastern Austria, will come under increased competition from low cost producers in CEECs. Southern and Eastern border regions of Austria (e.g. Styria and Burgenland) have larger concentrations of these industries and consequently these regions would be likely to suffer. Conversely, energy-intensive branches of industry in CEECs, e.g. in Slovakia, remain highly competitive, as long as energy prices in the country remain subsidised and low. Abraham and Koning (1999) concur with the predicted employment losses for unskilled workers in EU countries, especially eastern border regions Outlook and Conclusion The immediate trade effects of enlargement are not likely to be very significant, given the already existing levels of trade liberalisation with the EU. However, Baldwin et al. (1997), Avery/Cameron (1998) and Boeri/Brücker et al. (2000) predict that accession will generate gains for EU 15 countries. They also find that these gains are likely to be unevenly distributed. According to Breuss/Schebeck (1999), the extent to which Eastern enlargement will affect EU countries will not least depend on their existing trade relations and interdependencies with CEE countries (see also RWI/EPRC, DIW 94 EPRC

95 2000). Gains from EU enlargement are also likely to benefit particular sectors over others. For instance, Mayhew (1998) anticipates that EU exports in investment goods are expected to remain high as CEE industry undergo restructuring processes and businesses consequently re-tool. EU business and financial services are also expected to retain a clear competitive advantage (RWI/EPRC, 2000). High productivity industries are also likely to benefit. Over the medium- to longer-term, the adoption of the acquis, including standardisation, certification and product liability rules, will reduce transaction costs for trade in goods and services. As has already been discussed, it is also possible that the commodity composition of CEECs may shift. In the short-term, however, accession is expected to increase production costs in the CEECs at the same time as firms are facing increased competition from a fully integrated market. Polish and Romanian manufacturing firms, surveyed as part of a study of enterprise readiness for EU accession (Carlin/Estrin/Schaffer, 1999), expressed particular concerns about the likely costs of compliance with EU regulations, competition from the EU and the loss of skilled labour to the EU. Costs and benefits of increased levels of trade will vary between CEE countries. Costs and benefits are also likely to vary between sectors. Based on current conditions, Avery/Cameron (1998) anticipate accession to increase pressure on large sectors of CEE industry, including small and medium sized enterprises (SMEs) and services. Rollo (1998) also anticipated that CEE producers of high value added goods, of traded services, and in the finance industry could suffer from increased competition. Further within CEECs, sectors with a high R&D intensity which substituted for imports from the world market have already been most affected by the removal of trade barriers according to Boeri/Brücker et al. (2000). However, CEE regions are also in a position to become the lowest cost producers inside the EU in textiles steel and bulk chemicals (Rollo,1998). Bröcker/Jäger-Roschko (1996) suggest that manufacturing industries in advanced CEECs are likely to diversify away from any areas of competition between CEE and EU peripheries. Urban (1998) speculates that as ongoing specialisation of the CEECs takes place in the direction of the more sophisticated engineering branches rather than towards low-skill, labour intensive industries, some advanced CEECs may be taking an intermediate position between the industrially advanced EU-North DIW 95 EPRC

96 and the less advanced EU-South countries. Taking everything together, concerns that peripheral EU regions will be replaced by the CEECs as trading partners of other EU countries are not necessarily well founded (Avery and Cameron, 1998). Some countries, e.g. Greece and Italy, may even gain from proximity. However, the precise nature of the economic costs and benefits of enlargement will depend upon the conditions in which the single market is enlarged, which in turn depends upon the progress which the associated countries are able to make prior to accession in aligning their laws and practices with those of the EU (see also RWI/EPRC, 2000). Economic gains will also depend upon the adequacy of transport, telecommunication and energy infrastructures and networks in acceding countries, which are necessary to support increased trade and economic activity. 4.2 Investment Foreign capital flows into CEECs largely take the form of foreign direct investments. Cumulative net inflows of FDI into the CEE region amounted to USD 50 billion between 1991 and 1997 (Boeri/Brücker et al., 2000). In terms of FDI, a similar realignment towards the EU has taken place as was the case for trade flows. Foreign investment in the CEECs has been encouraged since the start of the political and economic reform process by expanding domestic markets, low cost labour and a range of other incentives and by the privatisation process (Hughes/Helinska-Hughes, 1998; OECD, 1997; Zemplinerova, 1997; Stankovsky, 1995). FDI data for the CEECs has been variable and often displays inconsistencies in the amount of investment recorded both for individual countries and the region as a whole. Undoubtedly, FDI levels have increased considerably over the 1990s, with Hungary, Poland and the Czech Republic taking the majority of the total (Estrin et al., 1997, Mayhew, 1998; Grabbe/Hughes 1997; Estrin/Hughes/Todd 1997) Volume of FDI Boeri/Brücker et al. (2000) observe that from the perspective of EU members, capital flows to CEE countries are negligible: an annual net capital flow of around USD 15 billion corresponds to a share of 0.15 per cent of GDP and 0.8 per cent of gross fixed investment in the EU However, from the stand point of candidate countries, those annual capital inflows amount to 5 per cent of GDP and more than 20 per cent of gross fixed investment and thus have contributed significantly to capital formation, DIW 96 EPRC

97 relieving domestic capital markets in CEE countries and having a substantial impact on growth, interest rates and wages (Boeri/Brücker et al., 2000). It is widely assumed that accession will lead to increased FDI flow into CEE countries. Experience from past enlargements demonstrate that accession to the EU can considerably increase capital inflows, at least for a transitional period. Boeri/Brücker et al. (2000) predict that capital flows to the CEE countries may double in the wake of accession and the inflow of portfolio capital will pick-up as the harmonisation of the regulation of financial markets gains momentum. The main countries receiving FDI among the CEECs are Hungary (30 % of total EU- FDI stocks in CEECs in 1997), the Czech Republic (28 %) and Poland (26 %). Almost 90 % of total FDI stocks in the CEECs were located in the candidate countries of the Luxembourg-group in 1997 (see table 4.2.1). Countries with sales strategies in privatisation (e.g. Hungary) reached significantly higher shares of FDI than countries with voucher strategies (e.g. Slovakia). The sale of public utilities and telecommunications had an important impact on the size of FDI inflows (Boeri/Brücker et al., 2000). Table 4.2.1: FDI Stocks of EU members in CEECs, 1997, Assets Applicant Countries FDI Stocks from EU Members in CEECs in % of reported CEEC in Mio data Poland Estonia Lithuania Latvia Czech Republic Slovakia Hungary Romania Bulgaria Slovenia Sum of reported CEECs data Source: Eurostat, European Union direct investment, Yearbook DIW 97 EPRC

98 Analogous to trade and migration, geographical proximity seems to play an important role in determining bilateral FDI flows. The main investing countries of the EU are Germany, France, Austria and the Netherlands (see table 4.2.2). Germany alone accounts for about 38 % of all reported FDI stocks of the EU 1997 in the CEECs. All these main investing countries account for approximately two thirds of all EU FDI stocks in the CEECs. In 1998, EU-FDI flows to Slovakia were dominated by Germany and Austria, EU-FDI flows to Poland mainly came from Germany and almost all FDI flows to the three Baltic states originated in Sweden or Finland (see table 4.2.3). This evidence suggests that bordering regions in CEECs might be affected by FDI much more strongly than the average. Unfortunately, FDI figures are not reported at the regional level in a concise and overall picture so far. However, selected data for Poland and Hungary show that the capitals and its surrounding regions as well as the industrialised regions bordering to the EU profited most from FDI inflows (see also section 4.2.4). Table 4.2.2: FDI Stocks of EU members in CEECs, 1997, Assets, EU 15 = 100 Applicant Countries EU Germany France UK Austria Finland Netherlands Portugal Sweden Poland Estonia Lithuania Latvia Czech Republic Slovakia Hungary Romania Bulgaria Slovenia Sum of reported CEECs data EU FDI-Stocks in the World Source: Eurostat, European Union direct investment, Yearbook DIW 98 EPRC

99 Table 4.2.3: FDI Flows from EU members to CEECs, 1998, EU 15=100 Applicant Countries EU Germany France UK Italy Austria Finland Bel/Lux Poland Estonia Lithuania Latvia Czech Republic Slovakia Hungary Romania Bulgaria Slovenia Sum of reported CEECs data Sum of FDI flows from EU members to the World Applicant Countries Denmark Spain Netherlands Portugal Sweden Poland Estonia Lithuania Latvia Czech Republic Slovakia Hungary Romania Bulgaria Slovenia Sum of reported CEECs data Sum of FDI flows from EU members to the World Source: Eurostat, European Union direct investment, Yearbook DIW 99 EPRC

100 Table 4.2.4: FDI Flows from CEECs to EU, 1998, in Mio Euro Applicant Countries EU Germany France UK Italy Austria Finland Bel/Lux Equity+ Equity+ Equity+ Total Equity+ Total Total Equity+ other other other capital other capital capital other Poland Estonia Lithuania Latvia Czech Republic Slovakia Hungary Romania Bulgaria Slovenia Sum of reported CEECs data Sum of FDI flows from the world to EU members Applicant Countries Denmark Spain Netherlands Portugal Sweden Equity+ Equity+ Equity+ Equity+ Equity+ other other other other other Poland Estonia Lithuania Latvia Czech Republic Slovakia Hungary Romania Bulgaria Slovenia Sum of reported CEECs data Sum of FDI flows from the world to EU members Source: Eurostat, European Union direct investment, Yearbook DIW 100 EPRC

101 So far there are almost no capital movements from the CEECs to the EU. In 1998, Hungary, which accounted FDI flows in the EU of 261 Mio Euro, was the main investing country among the CEECs. The Czech Republic followed at a distance with 30 Mio Euro (see table 4.2.4). If capital movements to the CEECs can make themselves felt at all in the present EU, then few sectors in few countries and regions may be affected by FDI. Overall, the share of the CEECs in world-wide FDI stocks 1997 of the EU members was only 2 %. Only in the case of Austria this share exceeds the EU average significantly: 28 % of all Austrian FDI goes to the CEECs (see table 4.2.5). Table 4.2.5: FDI Stocks of EU members in CEECs, 1997, Assets, World = 100 Applicant Countries EU Germany France UK Austria Finland Netherlands Portugal Sweden Poland Estonia Lithuania Latvia Czech Republic Slovakia Hungary Romania Bulgaria Slovenia Sum of reported CEECs data EU FDI-Stocks in the World Source: Eurostat, European Union direct investment, Yearbook Allocation of FDI CEECs already differ according to the importance of FDI in their economies. Publications, based on IMF and World Bank figures, show the accumulation of FDI over the period to be US$ 12.9 billion in Hungary, US$ 12.8 billion in Poland, and US$ 5.5 billion the Czech Republic, then falling to US$ 1.2 billion for the next highest recipient, Romania. At the bottom of the scale, Lithuania received US$ 286 million and Bulgaria US$ 467 million (Pavlínek/Smith, 1998). Hungary has been the most successful economy in terms of attracting FDI (Kaminski, 2000). Foreign affiliates accounted for two thirds of the equity capital in manufacturing and produced more DIW 101 EPRC

102 than 60 % of the Hungary s manufacturing output in 1996 (Hunya, 1998). The restructuring of Hungarian industry seems to have been largely FDI led. Short-term problems have emerged due to the rapid restructuring, capacity destruction and layoffs (Kaminski, 2000). Resultant enterprise closures and unemployment have exacerbated social and regional disparities in the country. More generally however, the high level of FDI and the presence of multi-nationals in the economy contributed towards the increased international competitiveness of a growing number of Hungarian enterprises (Agh, 1999). Lorentzen (1998) identifies similar experiences in Poland and Hungary: manufacturing firms with foreign ownership outperformed domestic firms in terms of export growth, the technology intensity of exports and labour productivity. Huge differences in the allocation of FDI between the candidate countries indicate that the credibility of institutional and economic reforms plays a key role in the location of foreign capital flows. Not surprisingly, countries with a stable institutional and economic environment benefit more significantly from capital inflows than other countries. However, while the favourable and stable economic and institutional environment seems to play a key role for the selection of a particular destination country it has less influence on the choice of the particular location inside a country as the basic institutional framework is often determined at the national level. The concentration of around 85 % of EU-net capital flows into the CEECs in the Luxembourg Group candidates in 1998 may be interpreted as an indication that the prospects for accession contribute considerably to the credibility of the CEECs (see table 4.2.6). Causality works here in both directions: stability may have contributed to the inclusion of countries in the Luxembourg Group, while this inclusion may have contributed to the credibility of the new institutions and economic policies and hence triggered FDI (Boeri/Brücker et al., 2000). The distribution of FDI is expected to continue to be uneven between CEECs and between regions in countries. FDI is likely to be heavily concentrated in the more advanced CEE countries. Further, accession may even widen disparities according to some studies. For instance, CEECs which remain outside the EU lose out as FDI diverts to the new member states. Zuleeg (2000) estimates the costs of delayed accession to be high for the CEECs furthest from EU membership (Romania and Bulgaria). DIW 102 EPRC

103 Table 4.2.6: FDI Flows from EU to CEECs, 1998 Applicant Countries FDI Flows from EU Members to CEECs Equity+ other, in % of reported in Mio CEEC data Poland Estonia Lithuania Latvia 46 1 Czech Republic Slovakia Hungary Romania Bulgaria Slovenia Sum of reported CEEC data Source: Eurostat, European Union direct investment, Yearbook Role of FDI Two main arguments exist in the literature regarding the longer-term role of FDI in the national and regional economies of the CEECs. The first proposes that the CEECs, in order to restructure their legacy of inefficient industries, use FDI as a means to position themselves within flows of global capital, often resulting in restructuring based on low-cost competitiveness and an erosion of wage relations and worker flexibility (Dunning, 1993; Grabher, 1997; Pavlínek, 1998, Smith/Swain, 1998). The second argument supports the view that FDI can, under certain conditions, lead to a progressive upgrading of a national/regional economy through the creation, for example, of good local supplier networks, upgrading of technology, increased efficiency and increased competition (Dicken et al., 1994; Lorentzen, 1998; Malmberg et al., 1996; Amin/Thrift, 1994). According to the latest survey of the Institut der deutschen Wirtschaft (Institute of the German Economy) in which 480 German companies were questioned about their motives for FDI, the most important reasons named were market access, production for home market and exploitation of country specific factor endowments such as lower labour costs and lesser taxation and regulations. According to this survey market access was of less than average importance for European companies to invest in DIW 103 EPRC

104 the CEECs, whereas production for home markets was the most important aspect named (Beyfuß/Eggert, 2000). Most studies of actual FDI flows in CEE arrived at different conclusions, though. The results of the analysis of Boeri/Brücker et al. (2000) highlight the fact that foreign direct investments to the CEECs are not motivated exclusively by low labour costs. The branch structure of investment indicate that market access was the primary motive for investment. Only around one fifth of foreign investment was found to be located where low labour costs play a significant role and the share of unskilled labour is relatively high. Boeri/Brücker et al. argue that quite often an important investment motive is to supply the markets of the FDI-host country and the exploitation of firstmover advantages in markets with no or limited competition. Nearly half of FDI to CEECs are made in non-tradable sectors such as public utilities (electricity, transport and communication), construction and service sectors. According to Boeri/Brücker et al. (2000), market access already seems to be an important investment motive even for investments in tradable sectors and is likely to gain in relative importance with growing incomes in the CEECs. These investments have no or only negligible effects on wages and employment in the EU. Other recent in depth studies of motives for investment in CEE (Altzinger/Winklhofer, 1998; Lankes/Venables, 1996), also found that factor-cost considerations are less important. CEE based production places EU firms within growth markets, which implies that products will be sold in the local markets and consequently would not imported back (Hardy 1998; Pavlínek/Smith 1998). Furthermore, in the long run, revenues from FDI may outstrip initial capital exports and stabilise parent companies in the EU. In the short run, employment in the parent companies may increase by the so-called home-office effects, i.e. by the establishment of necessary back offices in the headquarters. Finally, the investments in public utilities contribute to the improvement of infrastructure for further business activities in the CEECs. There is still a considerable demand for investment in infrastructure in the CEECs at the moment but it may be estimated that investment in the nontradable sectors will decline in importance in the long run. The share of primary sectors in FDI is negligible. Within manufacturing, the investments are concentrated in the production of motor vehicles, food stuffs, petroleum, chemicals, rubber and plastics. DIW 104 EPRC

105 Insofar as foreign investment in the tradable sectors is motivated by market access FDI in the CEECs complement rather than substitute trade, i.e. it raises the value added of parent companies in home countries relative to a case without foreign investment. However, the high labour intensity of FDI in the CEECs is an indication that the exploitation of the wage differential may be an important investment motive for at least some tradable sectors (e.g. textiles, clothing, electrical machinery, rubber and plastics, motor vehicles). If a negative impact of FDI on employment and wages can be felt at all in the present EU, then on unskilled labour in the above mentioned sectors in countries and regions bordering the CEECs. The picture regarding labour intensity and skill and wage levels is ambiguous. For instance, the shares of low tech, low wage and unskilled industries in German FDI in the CEECs is higher than in German FDI in the rest of the world, but this is not the case for Austria. Capital intensive industries play the crucial role in this case (Boeri/Brücker et al., 2000). Thus, Boeri/Brücker et al. (2000) concluded overall that only a minor part of FDI is driven by low-wage costs in CEE and replaces home production. The major part of FDI is aimed at market access. Accordingly, many of the companies which have invested in CEE and those which will do so in the future are not necessarily investing in CEE at the cost of investment in EU member states. In this case, investment is created rather than diverted from elsewhere in the EU Impact of Increased FDI to CEECs Within CEECs, major urban agglomerations have traditionally attracted the highest levels of FDI and will continue to do so. Western border regions can also expect to be foci for FDI investment, though they have not proved to be as uniformly attractive as urban agglomerations. Pavlínek (1998) found western border regions of the Czech Republic to be particular foci for cross border export-oriented investments. These types of investment, according to Pavlínek, merely exploit low labour costs in the service of external markets and are not likely to contribute to durable regional economies. Conversely, western border regions of Poland were found not to be attractive for German investors. High wage levels relative to the rest of the country, out migration of skilled labour, mainly attracted by high wages in the Berlin urban agglomeration and the adjustment shock faced by East German enterprise have limited cross-border investments, according to Bertram (1998). DIW 105 EPRC

106 Empirical research on investment patterns also indicates that foreign investments are motivated by market proximity (Döhrn 1996; Markusen, 1995; Boeri/Brücker et al., 2000). Thus, border countries and regions are particularly well placed to exploit the new investment opportunities offered as enlargement opens up new markets and investment opportunities. Austria already ranks among the top investing countries (Richter, 1998). Austrian firms have more than 11,000 affiliates and joint ventures in the CEECs and are among the most significant foreign investors in the region (Richter 1998). German enterprises have also expanded into CEE regions (Mayhew, 1998). Penetrating eastern markets and outsourcing part of the production to lower cost locations may strengthen the international competitive position of EU firms. This is especially the case for small and medium sized companies in German and Austria, for which proximity for the new markets and production locations matter more than for globally oriented corporations. Anticipated gains for CEE countries and regions have sparked concerns about possible job and investment losses in current EU countries and regions. There are particular concerns in some member states concerning the delocalisation of employment to CEE (i.e. EU companies may close factories at home and move operations to lower cost CEE sites) and the impact of competition from some CEE producers. Concerns appear to be particularly pronounced in border regions of countries neighbouring CEECs. As EU enlargement becomes more of a certainty and as unit costs remain high in some EU countries, it might be expected that enterprises, especially in Austria and Germany, would be attracted to new production locations in nearby CEE regions, particularly western border regions. FDI investment in CEECs in order to cut costs and exploit cheaper labour costs are also likely to be concentrated within particular industrial sectors. Djankov/Hoekman (1998) and Boeri/Brücker et al. (2000) predict that job losses due to delocalisation will be concentrated in construction, food and beverages, leather and footwear, textiles and clothing, electrical machinery, precision instruments and furniture. Pavlínek (1998) and Smith et al. (2000) suggest that even within these sectors job losses may be confined to outsourcing of more labourintensive tasks. Moreover, nearly half of FDI from the EU to CEE is directed at nontradable sectors (i.e. public utilities and communication, financial inter-mediation and other services) (Boeri/Brücker et al., 2000). Additionally, some FDI may tend to foster the specialisation of production in human capital intensive processes in the EU and labour intensive production in CEE. This outcome could conceivably hurt wages and DIW 106 EPRC

107 could conceivably hurt wages and employment of unskilled workers in specific sectors and branches. However in these branches, large trade surpluses vis-à-vis the CEE countries can be observed, so that undesirable effects of FDI on wages and employment of unskilled workers could be compensated by increasing exports (Boeri/Brücker et al., 2000). Thus, fears of large-scale job and investment losses in the EU appear to be less well founded. FDI is highly correlated to increasing shares of intra industry trade, which is an indication that production processes might become increasingly segmented in human capital intensive activities on the side of the EU and labour intensive activities in the CEE countries (Boeri/Brücker et al., 2000). Investment attracted to CEECs by potential reductions in production costs implies that enlargement could encourage restructuring based on low-cost competitiveness (Dunning, 1993; Grabher, 1997; Smith/Swain, 1998). However, as economic restructuring continues, it may also be possible that FDI could focus on technology and skill intensive production, at least in the more advanced CEECs. These types of investment are more likely to capitalise on skilled labour and improve local supplier networks, upgrade of technology, increase efficiency and increase competition in CEECs (Dicken et al., 1994; Lorentzen, 1998; Malmberg et al., 1996; Amin/Thrift, 1994) Conclusion FDI notably increases the capital endowment of the CEECs and, furthermore, enforces growth through the transfer of technologies, knowledge and human capital. In the short and medium term it is not expected that all regions will benefit evenly from FDI inflows. In the best positions are the regions immediately bordering the present EU member states and the capital regions. This trend will succeed in future because these regions offer the best conditions regarding the main influence factors for FDI: (1) Proximity to the donor country reduces the transaction costs; (2) the infrastructure of almost all CEECs is directed towards the capital cities - although CEEC-wide concise figures about the regional differentiation of infrastructure are not available, selected non-standard data show that the quality and availability of infrastructure is better in the capital cities as well as along the main transport corridors; (3) even more important, large shares of the population in the CEECs are concentrated in the capital regions (e.g. 28 % of the whole Hungarian population live in Budapest and its sur- DIW 107 EPRC

108 roundings), and as market access is a decisive investment motive, the population share to be reached will play a crucial role for location decisions in the future. Without accession and related transfers from the EU, the CEECs economic catching-up process will be slower (wage gaps will take longer to close), and the incentive to utilise the wage gap will persist much longer. Further, accession will be conditional to the introduction of all EU regulation in environmental protection, technical and social standards in the new member countries. This will necessarily imply considerable increases in the cost of production in CEECs, thus the low-cost advantages of CEE based production are gradually eroded. 4.3 Migration The anticipated post-eastern enlargement migration flows, and their subsequent impact on EU labour markets in particular, are one of the most sensitive issues in the EU-internal debate on the accession of CEECs to the European Union. The wage differential between the CEECs and the EU Member States (much larger than in any previous enlargement round), the relative economic underdevelopment of the candidate countries, and the more highly integrated EU market have all led to fears of massive post-enlargement migration flows into the EU. A key component of current, as well as future post-enlargement, migration is its geographical concentration, primarily in the neighbouring countries of Germany and Austria. Of current CEE residents in the European Union, 73 % of the working age population and 80 % of the employees are in these two countries. The share of current CEE nationals in total employment was 1.1 % in Austria and 0.5 % in Germany in 1995 in Greece and Sweden the figure was 0.2 % while in the remainder of the EU countries, the total was negligible (Boeri/Brücker et al., 2000, see table 4.3.1) Migration Flows While some estimates, reported in the media, came to possible inflows of million migrants from the CEECs, recent empirical studies have been much more conservative in their estimates of numbers of potential migrants. They range, for example, from a net immigration of ca. 335,000 residents with an assumed removal of barriers to migration in 2002 (falling below 150,000 within a decade) (see table 4.3.2, Boeri/Brücker et al., 2000) to at least three percent of the eastern population within 15 years i.e. an immigration flow of ca. 3 million people or 0.81 % of the 1995 EU DIW 108 EPRC

109 population (Bauer/Zimmerman, 1999). Individual country estimates also range widely - estimates in the Austrian case, for example, range from immigration of ca. 150,000 people to an annual inflow of ca. 18,000 and 23,800 commuting into border regions and large urban areas (Freihsl, 1998). In dealing with any figure for migration flows, care must be taken to ensure that inclusions and exclusions are clarified, as well as the underlying assumptions regarding time scales and the nature of the future enlargement process. For example as Mayhew (1998) notes, levels of migration will be influenced by how CEE and EU economies perform in the future and whether capital moves to the sources of labour or the reverse. Table 4.3.1: Distribution of Immigrants from Central and East European Countries among the Member States of the EU Cumulative Net Migration from the CEEC-10 (1) Stock of Citizens of CEEC- 10 residing in Host Countries (1) Stock of Citizens of CEEC-6 (2) working in Host Countries (2) January 1, % of the % of the Persons Persons Persons Population Population in % of employees Austria n.a. n.a (6) Belgium Denmark Finland France n.a. n.a (6) Germany Greece n.a. n.a Ireland n.a. n.a. 200 (6) n.a. n.a. n.a. Italy n.a. n.a (4) Luxembourg (6) Netherlands (5) Portugal n.a. n.a (7) 0.0 Spain n.a. n.a Sweden (7) 0.2 United Kingdom (4) total 3) EU-15 (estimate) ) Bulgaria, Czech Republic, Estonia, Hungary, Latvia, Lithuania, Poland, Romania, Slovak Republic, Slovenia.- 2) Bulgaria, Czech Republic, Hungary, Poland, Romania, Slovak Republic.- 3) Non-estimated countries only.- 4) ) ) Figure is estimated on basis of national records of labour participation of foreigners.- 7) Figures estimated on the basis of national statistics on foreign residents. Sources: EUROSTAT, Hönekopp (2000), Brücker/Trübswetter/Weise (2000). DIW 109 EPRC

110 Table 4.3.2: Projections for the Growth of the Population of Citizens of the Central and East European Candidate Countries Resident in the EU Residents from the CEEC-10, absolute Share of Country in % 1998 Increase in the Number of Residents, absolute Austria (1) Belgium Denmark Finland France (1) Germany Greece Ireland (1) Italy Luxembourg (1) Netherlands Portugal Spain Sweden UK Total EU The extrapolation of the projection for Germany on the EU-15 countries relies on the assumption that country shares in the number of foreign residents from the CEEC-10 remains constant. 1) The number of foreign residents is estimated for a number of countries on basis of employment figures. Sources: EUROSTAT, Brücker/Trübswetter/Weise (2000). A recent OECD study (OECD, 1998) concludes that, for a number of reasons, largescale migration flows from east to west are not likely to occur and should not be overemphasised in the enlargement agenda. First, it is anticipated that within the process of membership, there will be delays in the introduction of the free movement of people beyond the accession date as was the case when Portugal, Greece and Spain joined the EU. Second, such temporary restrictions on the movement of people are likely to be accompanied by bilateral agreements between current EU Member States and the candidate countries. Austria and Germany, the two countries most affected by migration flows, have already signed such agreements with a number of CEECs. Third, the membership requirements may, in themselves, reduce the incentives to migrate if a move towards convergence in economic and social indicators and the harmonisation of labour standards reduces existing differences in living and DIW 110 EPRC

111 working conditions (Amato/Batt, 1999). The study also points to the possibility that, following enlargement, migration flows may be directed towards the relatively more developed CEE Member States, rather than any of the current EU Regional and Social Impact Many existing recent migration studies suggest that the effects on wages and labour markets in the EU as a whole will be negligible. Bauer/Zimmerman (1999), for example, working on the assumption of immigration levels of ca. 200,000 a year over the next 15 years, estimate that the fall in the wages of workers in the current EU Member States would, at most, be 0.8 % of current wage levels. Further, they conclude that while the empirical evidence on the employment effects of immigration is more contradictory than that on wages, the majority of the evidence further points to relatively small effects. An important factor influencing any migration impact is the skill level of the migrants. The inflow of low-qualified workers is likely to result in an increase in the unemployment of domestic low qualified workers and a decrease in the unemployment of high-qualified workers and vice versa in the case of an inflow of high qualified workers. The time element also affects the impact of migration, with initial high levels following the introduction of the freedom of movement, likely to fall over time and overall (potentially significant) increases in the foreign population in the most affected countries (Germany and Austria) occurring over a longer time period (see tables and 4.3.3). The annual increase of the foreign population from the ten CEE candidate countries in Germany, for example, is expected to be negligible 30 years after the introduction of freedom of movement (Boeri/Brücker et al., 2000). While migration may benefit EU 15 states, the economies of the applicant countries may suffer from the negative impacts of brain drain and loss of a valuable part of young mobile workforce. European labour market mobility can be hindered by the importance of formal educational degrees for specific jobs that differ between countries, by different languages and cultural traditions, by rigid housing and labour markets (Abraham/Konings, 1999). The same factors may also apply to countries in Central and Eastern Europe. Past enlargement experiences have already demonstrated that only a relatively small fraction of the workforce may be expected to migrate, even given large differences in incomes and wages (Boeri/Brücker et al., 2000). Further, recent evidence from Hungary and Bulgaria shows a strong regional divergence in DIW 111 EPRC

112 unemployment, suggesting that unemployed Hungarian and Bulgarian workers do not even move within their own country to find a job (Burda, 1998). Table 4.3.3: Projections for the Stock Population of Citizens of the Central and East European Candidate Countries Resident in the EU Residents from the CEEC-10 in Persons 1998 Share of Country in % Resident Population from the CEEC Austria Belgium Denmark Finland France Germany Greece Ireland Italy Luxembourg Netherlands Portugal Spain Sweden UK Total EU-15 (1) memo items:... the Resident Population of the Country of Origin... the population in the EU-15 Resident population of the CEEC-10 as a % of The extrapolation of the projection for Germany on the EU-15 countries relies on the assumption that country shares in the number of foreign residents from the CEEC-10 remains constant. 1) The number of foreign residents is estimated for a number of countries on basis of employment figures. Sources: EUROSTAT, Brücker/Trübswetter/Weise (2000). Much attention has been paid to East-West migration but, since 1989, CEECs have rapidly become hosts to migrants from the former Soviet Union and Romania. Hungary has received large influx of Romanian citizens. Immigration to Poland, from Russia, the Ukraine and Belarus also increased. Polish Authorities estimate that be- DIW 112 EPRC

113 tween 100,000 and 200,000 illegal migrant workers were in the country (Koslowski, 1998). The prospect of EU membership is identified as a factor which contributes to increasing migration to some advanced CEE countries. The criteria these state must fulfil in order to achieve membership include major policy changes in the areas of immigration and law enforcement. However, enlargement would eliminated the buffer zone that key CEE countries have formed between EU frontier and the likes of Russia, Croatia, and Romania - potentially very large migrant and refugee sending countries (Koslowski, 1998). One aspect to arise from recent studies on the impact of migration relates to the ageing of the west European labour force (Bauer/Zimmermann, 1999; Freihsl, 1998; Mayhew, 1998; Amato/Batt, 1999). The anticipated ageing of the west European population and corresponding decline in working age population will contribute to labour market tensions which can attract migrant workers. The predicted ageing process is lowest in the UK, with the size of the over-65 group expected to increase from 15.7 % to 19.4 % between , and highest in Greece, with a rise from 13.8 to 22.2 % in the same group over the same period. With the exception of Ireland, the working age (15-64) population share is expected to decline in all EU countries by 2-5 percentage points over the same time period. In the CEECs, Bulgaria and Hungary display similar demographic developments but all the other countries are characterised by relatively smaller age groups beyond 65 and larger ones for the 0-14 age range (Bauer/Zimmermann, 1999). This difference suggests a migration potential for young people in the East, due to labour shortages in the West, particularly in occupational areas usually taken by younger age groups. An inflow of working migrants could not only combat labour market shortages but also make an important contribution to the alleviation of falling tax income and the difficulty of adequately financing pension and social security systems Types of Migration Migration between the CEECs and the EU Member States is characterised by a relatively high gap in per capita incomes over a short geographical distance. This changes the potential for different types of migration and significantly increases the options for short-term, temporary migration as well as cross-border commuting - which almost exclusively affects Germany and Austria. Temporary migration is already very significant, and has risen at the expense of permanent migration, and full DIW 113 EPRC

114 post-enlargement integration is likely to increase the potential for this type of activity even further. The regional impact of different forms of migration varies. The choice of destination for daily commuters from the CEECs is strongly influenced by spatial proximity and accessibility factors, overtly favouring border areas and accessible larger urban centres. Non-daily commuters, even if still short-term migrants, are likely to include job possibilities, wage levels and living conditions to a much greater extent in their decision and operate within a much wider spatial area The Cases of Germany and Austria The patterns of sub-national distribution and impact of migration in Germany and Austria also differ. In the German case, the share of migrants in the east German Länder is well below average even in the direct border regions. The share of migrant employees in total employment along the German-Polish border is negligible whereas, in the rest of Germany, migration from the CEECs is concentrated at the border with the Land of Bavaria and the Czech Republic (see figure 4.3.1, Boeri/Brücker et al., 2000). The migration patterns into Germany, therefore, do not just follow immediate geographical proximity but also agglomeration of prosperous industries. The number of border commuters and guest workers in Germany is currently very small, with the much more significant group of temporary migrants largely confined to the agriculture and construction sectors. In 1996, for example, 80 % of the temporary workers were seasonal (maximum of three months employment per year) and 90 % of these were employed in agriculture. The employment of agricultural seasonal workers has had broadly complementary effects on the incomes of native farm workers whereas the wages and employment of native workers in construction was negatively affected in some cases through the subcontracting of CEE firms. The Austrian situation differs from the German case, principally because of its common borders with four CEECs and the close proximity of centres of population and economic activity to these borders). The Austrian capital, Vienna, two Land capitals, Graz and Linz, and a number of other important regional centres such as Villach, Klagenfurt, Wolfsburg, Wiener Neustadt and Baden are all within a 90 minute drive from the border regions of the neighbouring CEECs. The case of Vienna is also particular given the close proximity of the Slovakian capital of Bratislava, only 65 kilometres away. An influx of migrants from eastern Europe and other parts of the region DIW 114 EPRC

115 into Austria has stoked up fears of job losses among blue-collar workers who, however flimsy the evidence, perceived their jobs to be threatened (Caplen, 2000). DIW 115 EPRC

116 Figure Germany: Regional Distribution of Employees from the CEECs Source: Federal employment services, Boeri/Brücker et al. (2000). DIW DIW 116 EPRC