Research for REGI Committee - Indicators in Cohesion Policy

Transcription

1

2

3 DIRECTORATE-GENERAL FOR INTERNAL POLICIES Policy Department for Structural and Cohesion Policies REGIONAL DEVELOPMENT Research for REGI Committee - Indicators in Cohesion Policy STUDY

4 This document was requested by the European Parliament's Committee on Regional Development. AUTHORS Christian VANDERMOTTEN Gilles VAN HAMME (IGEAT, Université Libre de Bruxelles) Research manager: Jacques LECARTE Project and publication assistance: Virginija KELMELYTĖ Policy Department for Structural and Cohesion Policies, European Parliament LINGUISTIC VERSIONS Original: EN ABOUT THE PUBLISHER To contact the Policy Department or to subscribe to updates on our work for the XXXX Committee please write to: Manuscript completed in May 2017 European Union, 2017 Print ISBN doi: / QA EN-C PDF ISBN doi: / QA EN-N This document is available on the internet at: Please use the following reference to cite this study: Vandermotten C., Van Hamme G., 2017, Research for REGI Committee Indicators in Cohesion Policy, European Parliament, Policy Department for Structural and Cohesion Policies, Brussels Please use the following reference for in-text citations: Van Hamme and Van Hamme (2017) DISCLAIMER The opinions expressed in this document are the sole responsibility of the author and do not necessarily represent the official position of the European Parliament. Reproduction and translation for non-commercial purposes are authorized, provided the source is acknowledged and the publisher is given prior notice and sent a copy.

5 DIRECTORATE-GENERAL FOR INTERNAL POLICIES Policy Department for Structural and Cohesion Policies REGIONAL DEVELOPMENT Research for REGI Committee - Indicators in Cohesion Policy STUDY Abstract GDP per capita is the sole criterion for identifying specific conditions of eligibility to the benefit of the structural funds. This criterion does not reveal really the well-being of local people. This study examines alternative measures, like final consumption expenses or a more sophisticated synthetic index, and their impact on the eligibility of the regions. The impact of the UK Referendum is examined, either using the present criterion or the alternative ones. IP/B/REGI/IC/2016_080 May 2017 PE EN

6

7 Indicators in Cohesion Policy CONTENTS CONTENTS 3 LIST OF ABBREVIATIONS 5 LIST OF MAPS 7 LIST OF FIGURES 8 EXECUTIVE SUMMARY 9 GENERAL INFORMATION THE GDP CRITERION AND ITS CRITICS The basis of the concept is that any economic production is «positive» for the development, since it produces added value The concept as a tool for international and inter-regional comparison GDP vs. GNI Final Consumption Expenses: a better concept than GDP THE QUESTION OF THE GEOGRAPHICAL SCALE TOWARDS A MORE INCLUSIVE AND GLOBAL INDICATOR? Economic development Physical wellness (health) Life expectancy at birth (E0) Infant mortality rate (IMR) Conclusion Social vulnerability Gini coefficient Population at risk of poverty after social transfers The part of households at risk of material deprivation Other indicators Conclusion Education, technological development and access to information Percentage of year-old population with a high level of graduation (total, male, female) Percentage of population using the Internet Other indicators Conclusion Population dynamics Environment Conclusions about a synthetic indicator GEOGRAPHICAL PATTERNS OF THE MAIN INDICATORS AND GENERAL CONCLUSION 35 3

8 Policy Department for Structural and Cohesion Policies 4.1. The geographical pattern of physical wellness The geographical pattern of social vulnerability The geographical pattern of the educational and technological dimension The geographical pattern of the regional attractiveness A synthetic overview THE IMPACT OF THE UK REFERENDUM CONCLUSION 47 REFERENCES 52 ANNEX 1. ANNEX 2. ANNEX 3. TABLE OF THE POSITION OF THE REGIONS ACCORDING TO THE GDP, THE PRIVATE AND PUBLIC FCE AND THE FOUR OTHER SERIES OF CRITERIA 53 CHANGES IN THE ELIGIBLE AND TRANSITION REGIONS, PASSING FROM THE GDP CRITERION TO THE PUBLIC AND PRIVATE FCE CRITERION 65 CHANGES IN THE ELIGIBLE AND TRANSITION REGIONS, PASSING FROM THE GDP CRITERION TO THE SYNTHETIC INDEX 69 4

9 Indicators in Cohesion Policy LIST OF ABBREVIATIONS CF Cohesion Fund BGS Balance of Goods and Services E 0 Expectancy of Life at Birth EAFRD European Agricultural Fund for Rural Development EMFF European Maritime and Fisheries Fund ERDF European Regional Development Fund ESF European Social Fund ESI European Structural and Investment Funds ESPON European Observation Network for Territorial Development and Cohesion EU European Union EUROSTAT Statistical Office of the European Union FCE Final Consumption Expenses GCF Gross Capital Formation GDP Gross Domestic Product GNDI Gross National Disposable Income GNI Gross National Income Greens/EFA Greens/European Free Alliance GRI Gross Regional Income GRP Gross Regional (Domestic) Product HDI Human Development Index IMR Infant Mortality Rate NPISH Non-Profit Organisations Serving Households NUTS Nomenclature of Territorial Units for Statistics PPS Parity of Purchase Power RD Research and Development REGI Committee on Regional Development SILC European Union Statistics on Income and Living Conditions TFEU Treaty on the Functioning of the European Union YEI Youth Employment Initiative 5

10 Policy Department for Structural and Cohesion Policies 6

11 Indicators in Cohesion Policy LIST OF TABLES TABLE 1 Level of GDP, GNI and Final Consumption Expenses per capita in the EU, at the level of the States (2015) 15 TABLE 2 Main aggregates of the GDP (2015, Current prices) 17 TABLE 3 Percentage of the national population living in eligible areas, following the kind of criterion and the geographical scale (2013 data) 23 TABLE 4 Absolute level of the allocations after the UK referendum following the four scenarios and relative losses by comparison to the present situation. 43 LIST OF MAPS MAP 1 GDP per capita, in PPS. 19 MAP 2 Private and public final consumption expenses (FCE), in PPS 20 MAP 3 Private and public final consumption expenses (FCE), in PPS 20 MAPS 4 AND 5 Geographical pattern of the physical wellness dimension (male expectancy of life at birth) 36 MAP 6 AND 7 Geographical pattern of the social vulnerability dimension 37 MAP 8 AND 9 Geographical pattern of the educational and technological dimension 38 MAP 10 AND 11 Geographical pattern of the global attractiveness (migratory balance) 39 MAP 12 Geographical pattern of the synthetic indicator, by comparison to GDP and FCE per capita (see Maps 1 and 2) 40 7

12 Policy Department for Structural and Cohesion Policies LIST OF FIGURES Figure 1: Level of correlation between health and two economic indicators (GDP and FCE per capita) 27 Figure 2: Level of correlation between social vulnerability indicators 30 Figure 3: Level of correlation between educational and R-D indicators 31 Figure 4: Correlations (R) coefficients between GDP or private and public FCE per capita and each of the four other dimensions 35 Figure 5: Reallocation of the EU28 population between the three categories on the basis of GDP, FCE and the synthetic index 49 Figure 6: Share of the population of each EU28 country in the three categories on the basis of GDP, FCE and the synthetic index 50 8

13 Indicators in Cohesion Policy EXECUTIVE SUMMARY Background Gross (Regional) Domestic Product per capita is presently the only criterion of eligibility for the less developed or intermediate regions in the framework of the European Structural and Investment Funds. However, this indicator considers implicitly that any economic production has a positive contribution to regional development and does not take into account the transfers of income between regions (as Gross Regional Income should do, but is unavailable). It reflects also a purely economic conception of the development and it doesn't measure the standard of living nor the well-being of the population. This study suggests to replace the use of Gross (Regional) Domestic Product by total (i.e. private and public) Final Consumption Expenses (FCE), also computed by EUROSTAT. In addition, a full eligibility statute is proposed for all the regions in countries with a very low level of investment (Greece and Cyprus; Gross Capital Formation less than 15 % of the GDP). Indeed, the very low level of investment in some areas is an obstacle to future development. Another, and probably better solution is to produce a more inclusive indicator taking into account, aside FCE, four other dimensions: physical wellness; social vulnerability; educational and technological development; population dynamics. Using these four dimensions is coherent with the aims of the Europe 2020 strategy, though it may raise issues of data availability. Hence, it is proposed to build a synthetic index which takes into account indicators for each of these dimensions. In most cases, ideal indicators are not available in the EUROSTAT data bank, and it is necessary to rely on proxies, although some of these are not fully satisfactory: for physical wellness, male life expectancy at birth, since more contrasts are observed between regions for males; for social vulnerability, an average between the share of the young adults with a low graduation level, the percentage of young people not in work, education or training, unemployment and long-lasting unemployment; for the educational and technological development, a weighted average of male and female population with a high level of education on one side and three indicators of the technological development on the other side (intra-muros RD expenses in % of the GRP, RD manpower and researchers in % of the active population and patents asked in relation to the active population); for the regional attractiveness, the migratory balance. The study proposes also not to include the environmental dimension at this scale, since the regional scale is not the best one for identifying environmental issues. Of course, environmental issues could evidently be included in actions financed through structural funds. Despite it is not in theory the best geographical division to consider the allocation of structural funds, in practice, NUTS 2 seems to remain the best geographical level for considering the eligibility of the regions, surely if GRP is replaced by FCE or by a synthetic indicator. By comparison to the present eligibility criteria, using FCE instead of GRP leads to reduce the intra-national disparities, in relation with the internal transfers and in particular the disparities between metropolitan cores and surrounding regions. 9

14 Policy Department for Structural and Cohesion Policies By comparison to FCE only, using the synthetic indicator reintroduces more intra-state variations, like North vs. South in France, East vs. West in Germany, and the North-South dichotomy in Italy and to a lesser extent in Spain. The synthetic index also reveals the transition statute of Ireland, in contrast to GDP figures which positions this country at very high levels but doesn't consider the very high amount of financial transfers outside the country. Measuring the impact of the UK referendum on the future of the structural funds is a difficult and very conjectural exercise. A lot of scenarios could be built, including increasing contributions of the remaining EU27 countries. It was not possible to examine all the scenarios in the framework of this study. Only keeping the contributions of the countries at their present level has been considered. In this case, the reduction of the total available budget should be very significant: the UK contributes for 15.6% to the European budget and receives only less than 3.5 % of the allocations for the structural funds. So, if one considers a linear reduction of all the European budgets, the expenses for the structural funds will be reduced by 12.1 %. Another consequence to consider is that the absolute value of the European average GDP per capita will decrease: regions that are now below the 75 % or the 90 % thresholds could exceed the same thresholds computed on the basis of the new average. Combining in a scenario the use of the new averages, the reduction of the total budget and using FCE instead of Gross Regional Product, with the same 90 % and 75 % levels as now, should mean a strong loss of allocations for the big Western European countries where the intra-national transfers of income are important (from the North to the South in Italy, from Paris to the rest of the country in France, from Western to Eastern Germany). Conversely, the losses should be more limited in Central-Eastern European countries. Considering the use of the synthetic index, the losses for the Western European countries would be even higher than for FCE, as well also for the Czech Republic, and lower for the other Central-Eastern European countries and the Mediterranean countries. Spain should become the second recipient of the allocations in volume after Poland, due to the weight of its social vulnerability and its negative migratory balance. Finally, the report also pleads for a special attention paid to Greece and Cyprus, in relation with their very low investment rate. 10

15 Indicators in Cohesion Policy GENERAL INFORMATION KEY FINDINGS The European Structural and Investment Funds to jobs and growth support all regions. However, investments aiming to reduce economic and social disparities focus on EU member states under the threshold of 90% of the average EU GDP per capita, expressed in purchasing power parity. Regarding the ERDF and the ESF, 80% of the funding is allocated to regions with a GDP per capita of less than 90% of the EU27 average and 67% for the regions with less than 75% of this average. GDP per capita is the only criterion to determine into which category a region should be classified within the framework of the Structural Funds. But it is a purely economic criterion, and hence doesn't fit well with the objectives of the cohesion policy, based on a larger approach, which includes a social dimension. Territorial, economic and social cohesion funds (the Cohesion Fund (CF), European Regional Development Fund (ERDF), European Social Fund (ESF, focusing on employment and education measures) and the Youth Employment Initiative (YEI, focusing on education and employment for year-old people)) are the main investment policies of the EU. Their key objective is to promote economic development and convergence across the EU. They account for about one third of the EU budget, i.e. 325 bn EUR (in 2011 prices) for the period of Of this total, 56.2% is assigned to the ERDF, 24.7% to the ESF, 18.1% to the CF and 0.9% to the YEI 1. The most important funds, ERDF and ESF, belong to the European Structural and Investment Funds (ESI), along with the specific European Agricultural Fund for Rural Development (EAFRD) and European Maritime and Fisheries Fund (EMFF). To benefit from these funds, EU members at the national and regional level need to provide some local co-financing. In the context of the objectives of economic, social and territorial cohesion promoted by the Treaty on the Functioning of the European Union (TFEU) amended in 2007 by the Treaty of Lisbon, the investment funds for growth and jobs support all regions. However, the CF, aiming to reduce economic and social disparities, focuses on EU member states under the threshold of 90% of the average EU Gross National Income per capita (NI per capita), expressed in purchasing power parity (PPS). PPS allows to take into account the differences in the price levels between the different countries and, for countries outside the EURO area, possibly the impact of variations in the rates of exchange. Regarding the ERDF and the ESF, resources are allocated in function of the level of GDP per capita: 80% of the funding is allocated to regions with a GDP per capita of less than 90% of the EU27 average (Croatia not included) and 67% for the regions with less than 75% of this average. GDP per capita is thus the only criterion to determine into which category a region will fall within the framework of the Structural Funds. However, this criterion is problematic, since it is purely economic, merely reflecting the level of the global production of a state or a region, while the objectives of the cohesion policy are based on a larger approach, which includes a social dimension, for instance standard of living, and other non-economic dimensions, like education. Therefore, the Committee on Regional Development (REGI) of the European Parliament wished to commission this study on Indicators in Cohesion Policy, developing a critical 1 European Structural and Investment Funds. Total Allocations of Cohesion Policy Breakdown by Spending Categories. Update 6/4/

16 Policy Department for Structural and Cohesion Policies analyse of the use of the GDP as an unique indicator and examining the possibility of using other indicators measuring regional development, their role in eligibility issues, and the feasibility of introducing indicators available at the NUTS 2 level complementing (or even replacing) GDP into the cohesion policy. This study is thus in line with the aims and recommendations expressed in the 2014 Sixth Cohesion Report, entitled Investment for jobs and growth: Promoting development and good governance in EU regions and cities, to explore using additional indicators in cohesion policy. It examines the possibility to develop additional or alternative indicators to GDP for measuring medium and long-term economic and social progress, in particular at the regional level. Previous studies have already approached this question: In March 2011, the study, Shaping EU regional policy: looking beyond GDP 2, was commissioned by the Greens/EFA Group in the European Parliament to examine the impact of the introduction of social criteria on the geography of European structural and investment funds; In February 2016, a briefing by the European Parliamentary Research Service focused on the Beyond GDP approach 3 ; In March 2016, a study from EP Pol Dep A assessed the important role the social economy plays in the EU. Priority policies identified to reach its full potential include, inter alia, improving definitions and developing alternative indicators to GDP to focus policies on bringing added value to the EU. The eligibility conditions to benefit from the structural funds also consider the subsidiary question of the adequate geographical levels to be considered. Indeed, European regulations lay down that the eligibility conditions to ERDF and ESF are defined at the NUTS 2 level. 4 This study will thus also examine the pertinence of using this geographical level. In conclusion, the present study will examine the research question regarding the following subjects: what critiques can be done to the GDP/capita criterion from a general economic perspective and as a measure of regional development in particular; the impact of the geographical scale used and the spatial configuration of the regions on their eligibility for ERDF and ESF, for instance using NUTS 3 in place of NUTS 2 regions; whether it is possible and feasible to build a more inclusive indicator, taking also into consideration social and even environmental criteria; and finally, the study will also revisit these questions regarding the consequences of the UK referendum. 2 Green New Deal Series, 7, Green European Foundation. 3 Beyond GD: Regional development indicators, EPRS, February The NUTS (Nomenclature of Territorial Units for Statistics) system is a hierarchical system used by EUROSTAT for dividing up the territory of the EU for the collection, development and harmonisation of European regional statistics. NUTS 0 is the level of the States, NUTS 1 corresponds to the major socio-economic regions, NUTS 2 to the basic regions for the application of regional policies and NUTS 3 to smaller regions for specific diagnoses. The population of the NUTS 1 units comprises between 3 and 7 million, for the NUTS 2 between 0.8 and 3 million and for NUTS 3 between 150 and 800 thousand. For practical reasons, the study cut out favours to administrative divisions and tried to respect intra-regional global geographical cohesion. For instance, in Germany, the NUTS 3 level corresponds to the Kreise, the NUTS 2 to the Regierungsbezirke (when existing) or to the Länder elsewhere, the NUTS 1 to the Länder. In France, NUTS 3 are the departements, NUTS 2 the regions and NUTS 1 specific groups of regions. In Italy and in Spain, NUTS 3 are the provinces, NUTS 2 the regions and NUTS specific groups of regions; etc. 12

17 Indicators in Cohesion Policy 1. THE GDP CRITERION AND ITS CRITICS KEY FINDINGS GDP is an indicator which supposes that any economic production contributes to the development. The concept doesn't include the transfers of income and is not a valuable measure of the well-being of the population. We state that FCE, used as a proxy of the Gross Regional Income, is a better indicator of the well-being of the population. GDP is an aggregate measure of production, equal to the sum of the gross values added of all resident institutional units engaged in production (plus any taxes and minus any subsidies on products not included in the value of their outputs), measured in purchaser's price (or transformed for comparison between countries in purchasing power parity). Thus, GDP is equal to the sum of the primary incomes distributed by resident producer units. Some major critiques can be done regarding this indicator, first as a concept and second as a tool for measuring regional development The basis of the concept is that any economic production is «positive» for the development, since it produces added value Added Value is the basic concept to calculate Gross Domestic Product, or any related concept (GNI, GNP, etc.). It can be defined as the value added by human work or machine to produce a good or a service sold on the market. This added value generates revenues for workers as well as for capital owners. Hence, any production or service creates added value as soon as it is sold on the market or even is only generating the distribution of wages. In this perspective, not only do morally debatable productions offer added value, but also productions which are only necessary to counter social, environmental or other problems, possibly linked to bad economic choices. For instance, if a factory does not respect environmental regulations, it may perhaps sell a bit cheaper and much more and thus improve economic performances in terms of added value. However, such production will compel the public authorities to develop more utilities for cleaning the environment, which by the way will also generate more added value. For another example, growing traffic jams could produce more added value, as people may consume more fuel, perhaps causing more accidents, implying more car repairs or new purchases, more health care expenditure, etc. Thus any production results in added value and increases GDP. Another important issue is that, by definition, GDP only includes production for the market with money transactions. For example, activities belonging to the domestic sphere are not included because they do not result in money transactions, though they represent probably half of the total working hours in European societies 5. Yet such domestic activities have equivalent activities in the market sphere: eating at home does not produce added value while eating in a restaurant does; taking care of children at home is not included in the GDP while putting them in a nursery is. Therefore, growth in the GDP may only be caused by a transfer from the domestic to the market (private or public) economy for the same service, although it does not necessarily improve social well-being. 5 See the results of a French enquiry for example: Insee (2010), Enquête Emploi du temps

18 Policy Department for Structural and Cohesion Policies 1.2. The concept as a tool for international and inter-regional comparison For comparison between countries, it is evidently necessary to use a common monetary unit, for instance the euro, which supposes the conversion into euro for all countries using other currencies. This conversion is based on the average annual rate of exchange between euro and the other currency. However, even for countries using the euro, a correction is needed to take into consideration the differences in price levels; this correction is the purchasing power parity (PPS) (Table 1). This issue is well solved by EUROSTAT for the international comparisons; because price levels are generally higher in richer countries the use of PPS results in a narrowing of the standard deviation between poor and rich countries. However, such a correction is not done at the inter-regional level inside each country, since no statistics about differences of price level are produced at regional level. Yet, the level of the prices can be very different inside a country, depending on the kind of region, for instance between the capital city and more remote areas; thus, in general, the statistics overestimate the level of development of the central metropolitan regions and underestimate it in more rural regions (if not in ultraperipheral regions, where prices could also be higher). Calculating parity on a regional scale would be an interesting request to present to EUROSTAT. However, this calculation should be a long and costly work since it supposes to assess the price levels for all European regions, which would require large enquiries similar to those carried out at national level GDP vs. GNI Let us recall that the GDP measures the added value produced by the resident economic actors. But some primary incomes produced within a state or a region can be exported and used outside this territory, while others can be imported. This includes the transfer of revenues of capital, social transfers, etc. For example, revenues of capital generated by an enterprise producing (or having registered its manpower) in a country X can be distributed to the shareholders residing in another country, but they are still included in the GDP of the country X. Therefore, Gross National Income (GNI), i.e. GDP plus primary incomes earned by residents from outside, minus primary incomes earned in the domestic economy by non-residents, is a better concept to approach the development and the global richness of the population of a country 6. GNI has replaced GDP as the basis of the budgetary calculation of the European Union and also for the comparisons between countries in the international statistics, for instance by the World Bank. At the national level, the difference between GDP and GNI is in general not very important. But inside the European Union, Ireland is a significant exception, as a big part of the GDP growth is transferred to other countries, which does not result in increasing wealth for the resident population (Table 1). 6 Another concept is Gross National Disposable Income (GNDI), which differs from GNI by adding to it the balance of the current transfers. The difference between GNDI and GNI is generally weak, not more than 1 or 2%. 14

19 Indicators in Cohesion Policy Table 1: Level of GDP, GNI and Final Consumption Expenses per capita in the EU, at the level of the States (2015) 2015 GDP/capita at current prices GDP/capita in PPS GNI/capita in PPS Total final consumption expenses in PPS/capita Private final consumption expenses (net disposable income of the households) in PPS/capita EU (28 countries) Belgium Bulgaria Czech Republic Denmark Germany Estonia Ireland Greece Spain France Croatia Italy Cyprus Latvia Lithuania Luxembourg Hungary Malta Netherlands Austria Poland Portugal Romania Slovenia

20 Policy Department for Structural and Cohesion Policies Slovakia Finland Sweden United Kingdom Source: EUROSTAT and own computations. For Bulgaria, Greece, Spain, Croatia, Italy, Cyprus, Luxembourg and Hungary, GNI data are those of the World Bank, with 0,761 taken as the rate of exchange between international dollar and EURO PPS. However, at the regional level, the differences between GDP and GRI could be very strong. Such a case occurs when a city corresponds to a narrowly delimited NUTS 2 region, excluding an important part of its commuting basin, for instance Berlin, Hamburg, Bremen and Inner and Outer London. So, the Brussels-Capital Region is supposed to be one of the «richest» regions of Europe, considering its GDP/hab. in PPS (211% of the EU average), but if this region produces 18.4% of the Belgian GDP, its inhabitants (10.3% of the Belgian population) benefit of only 6.1% of the Belgian incomes, as 52% of Brussels' manpower lives outside the limits of the NUTS 2 area 7. The question could also be asked regarding regions with a quite weak productive basis, but that are inhabited by many (possibly wealthy) retired people, or people benefiting from any other kind of social transfers. Unfortunately, at the regional level, the equivalent to GNI, let us say the Gross Regional Income (GRI), taking into consideration the primary incomes transfers between a region and the rest of the world, i.e. the other regions of the same country and the other countries, is not computed in the EU, and should presumably be very difficult to estimate. Therefore, GDP remains used for the measure of regional development and for the allocation of the ERDF and ESF Final Consumption Expenses: a better concept than GDP A better proxy than GDP to estimate the regional development and the wealth of regional residing people could be the private and public Final Consumption Expenses (FCE), a concept that excludes from the GNI investments and the balance of saving. To include or not the public expenditures 8 to the consumption as a proxy for wealth has important consequences for comparisons, since the share of public consumption is different from one country to another: for instance, public consumption represents 26.1% of the GDP in Denmark (and is high in the Nordic countries in general as well as in the Netherlands), but only 13.5% in Romania (Table 2). FCE is already computed by EUROSTAT at the NUTS 2 level, but only for households and Non-Profit Organisations Serving Households (NPISH). In this study, the total consumption expenses have been used, making the reasonable hypothesis that the level of the public consumption by inhabitant is the same in each region of a country 9. 7 Let us add that, until the mid-nineties of the last century, the population in the Brussels-Capital Region declined strongly, to the benefit of its suburban area, inducing mechanically a growth of the GDP per capita, as the denominator shortened, in a region confronted at this time to a decline of its population and its relative wealth. In some Central-Eastern European peripheral regions with many aged people and a strong outmigration, the shrinking population could partially explain a relative improvement of the GDP per capita. 8 Public consumption includes all expenditures made by public authorities, for example in the field of education or health. Hence, some services such as health services can be paid through private consumption or public consumption, when the person benefiting from this service does not pay it itself but is paid by the community. 9 Minor adjustments have been done to the NUTS 2 household consumption expenses published by EUROSTAT to adjust regional data to the national values produced in the EUROSTAT main aggregates of the GDP statistics. 16

21 Indicators in Cohesion Policy Table 2: Main aggregates of the GDP (2015, Current prices) Title Gross Domestic Product (GDP) Gross Capital Formation (GCF) Balance of goods and services (BGS) Final Consumption Expenses (FCE) Public expenditure of general government (of which individual consumption) (of which collective consumption) Household and NPISH final consumption expenditure (of which households) EU (28) 100,0 19,8 3,4 76,9 20,5 12,9 7,6 56,4 54,6 Belgium 100,0 23,2 1,7 75,1 23,9 15,4 8,5 51,2 50,0 Bulgaria 100,0 21,2 0,1 78,7 16,1 8,1 8,0 62,5 61,9 Czech Republic 100,0 27,4 6,1 66,5 19,5 10,3 9,2 47,0 46,3 Denmark 100,0 19,6 6,3 74,2 26,1 18,6 7,5 48,0 46,6 Germany 100,0 19,2 7,6 73,2 19,2 12,5 6,8 53,9 52,1 Estonia 100,0 24,7 4,1 72,6 20,3 11,1 9,1 52,4 50,7 Ireland 100,0 21,7 31,7 46,7 12,5 8,5 4,1 34,1 33,3 Greece 100,0 9,8 0,1 90,0 20,1 8,7 11,4 69,9 67,2 Spain 100,0 20,1 2,4 77,5 19,4 11,0 8,4 58,1 57,1 France 100,0 22,4-1,4 79,0 23,9 15,5 8,4 55,1 53,0 Croatia 100,0 18,8 2,8 78,5 19,7 9,9 9,8 58,8 57,7 Italy 100,0 17,0 3,1 79,9 18,9 11,3 7,7 61,0 60,4 Cyprus 100,0 14,5 0,3 85,2 15,7 6,7 9,1 69,5 68,5 Latvia 100,0 22,0-1,1 79,1 18,1 8,5 9,5 61,0 59,9 Lithuania 100,0 19,9-0,7 80,8 17,6 9,9 7,7 63,2 62,9 Luxembourg 100,0 19,6 33,0 47,4 16,7 10,3 6,4 30,7 29,2 Hungary 100,0 21,7 8,9 69,4 20,0 10,1 9,9 49,3 47,7 Malta 100,0 24,9 2,9 72,2 19,3 10,7 8,6 53,0 51,6 Netherlands 100,0 19,3 10,8 69,9 25,3 16,9 8,5 44,6 43,8 Austria 100,0 23,5 4,0 72,6 19,9 12,5 7,4 52,7 50,9 Poland 100,0 20,4 3,1 76,5 18,0 9,9 8,1 58,5 57,7 Portugal 100,0 15,5 0,7 83,8 18,2 9,7 8,5 65,6 63,6 17

22 Policy Department for Structural and Cohesion Policies Romania 100,0 25,6-0,5 74,9 13,5 6,5 7,0 61,4 61,1 Slovenia 100,0 20,1 9,1 70,8 18,7 11,1 7,6 52,1 51,3 Slovakia 100,0 23,2 2,4 74,4 19,5 10,7 8,8 54,9 54,0 Finland 100,0 21,1-0,5 79,7 24,4 16,4 8,0 55,3 52,7 Sweden 100,0 24,2 4,7 71,0 25,9 18,7 7,2 45,1 43,7 United Kingdom 100,0 17,6-2,1 84,4 19,4 12,4 7,0 65,0 61,9 Source: EUROSTAT Evidently, the level of FCE/inhab. can be linked to social transfers from other regions of the country and it could be argued that a part of these social transfers are the consequence of the low level of development of the region, which could so be masked by this indicator. It is surely not always the case; let us consider the wealthy suburban areas around the toonarrowly demarcated NUTS metropolitan regions or the affluent retirement areas. Generally speaking, and even if the GRP and the private and public final consumption expenses (FCE) are quite well correlated (r = 0.78), replacing the GRP per head by the FCE per head may reduce the proportion of the European population living in eligible areas, if we keep the 90% or the 75% threshold, resulting in the higher geographic concentration of funds (Maps 1 and 2). This is because the FCE includes social transfers and hence reduces the differences of level between regions compared to GRP. As a consequence, the number of regions below the 75% or 90% threshold will be reduced. The percentage of the population living in the eligible areas should fall from 44.9% of the European population to 32.0% at the 90% threshold and from 26.9% to 19.4% at the 75% level. However, the possibility may also be examined to raise the thresholds to maintain the same percentage of the European population as now in the eligible areas (Map 3). 18

23 Indicators in Cohesion Policy Map 1: GDP/capita, in PPS. The legend uses the 75% and 90% of the EU average thresholds. Source: EUROSTAT, IGEAT

24 Policy Department for Structural and Cohesion Policies Map 2 (left): Private and public final consumption expenses (FCE), in PPS. As on Map 1, the 75% and 90% of the EU average thresholds are used. Map 3 (right): Private and public final consumption expenses (FCE), in PPS. To keep the same percentage of the EU population than for 75% and 90% of the average GDP per head, the threshold used corresponds to 80.8% and 105.3% of the EU average. Source: EUROSTAT, IGEAT 2016 In conclusion, if wealth is understood as the volume and quality of goods and services a population can benefit from, GDP is not a good proxy of the population wealth at the regional level and a fortiori of its distribution among the different strata of the population and this study recommends to replace it by FCE (as Gross Regional Income data are missing). The question of the distribution of this consumption among the population remains evidently not scrutinized. Moreover, beyond the question of using Gross Regional Income or by default FCE instead of GDP, a fundamental political question from the territorial policy perspective is posed: should any European region have a strong productive basis, or is it enough that people living in each region benefit from good living standards, possibly linked to social transfers? It remains, however, a question in the case of Greece: FCE in «sustained» in this country by a very low level of investment, so the level of eligibility of the Greek regions should be reduced if using FCE at the place of the traditional GDP criterion. At the same time, this very low level of investment reflects the difficulties of the Greek economy. This question will be discussed later on. 20

25 Indicators in Cohesion Policy 2. THE QUESTION OF THE GEOGRAPHICAL SCALE KEY FINDINGS NUTS 2 is the best geographical scale for determining the eligibility of regions. The issue of the pertinent geographical scale for measuring regional development is related to the relevant scale at which the policy objective of territorial cohesion must be achieved. Indeed, territorial cohesion supposes the reduction of inequalities between territories within the EU but the scale at which inequalities must be limited is unclear and still subject to debate (should it be between countries, between NUTS 2 regions, between neighbourhoods in a city,?). Some would defend the idea that the national scale is the most relevant one. They argue that the national scale is the most effective to benefit from the best use of the agglomeration economies, through a concentration of the investments in the most efficient areas, mainly metropolitan areas 11. According to this view, the concentration of activities in a country results in better efficiency (i.e. GDP growth), and other areas will finally benefit from this progress through different mechanisms, such as out-migration from peripheral regions to the most efficient regions, income transfer or diffusion processes. In fact, the use of national level data combined with the propensity to exclusively maximize national GDP growth could lead to an aggravation of the intra-state inequalities, at the expense of the remote regions, with really no true guarantee of higher economic efficiency 12. Conversely, using a lower geographical level (NUTS 3) with the present GDP criterion should lead in many countries to strongly worsen the issue of the dissociation between central urban areas (where value is produced) and their suburban fringes (where a more or less important part of the population producing this value is living). This should, for instance, be the case in a country like Germany, where core cities correspond to NUTS 3 units separate from their semi-rural surroundings (Stadtkreis vs. Landkreis). Replacing NUTS 2 with NUTS 3 as the geographical scale for the determination of the eligibility could also encourage some countries to proceed to a kind of gerrymandering 13 of their basic statistical organisation, cutting or recomposing some statistical areas to reach eligibility thresholds. Conversely, GDP per capita could be recomputed at the level of ad-hoc areas aggregating NUTS 3 units, covering more or less the whole functional urban areas of the main metropolitan regions (core employment area plus employment basin). Recomputing the GDP per capita at the scale of the main urban areas is a feasible task. Indeed, it has already been achieved punctually, for example in the framework of the ESPON programme (see for a full discussion the ESPON Database Portal 14 ) as well as the Urban Audit data produced by EUROSTAT itself. However, this could be a difficult political question, in particular in federal countries where some suburban NUTS 3 units are not pertaining to the same federated entity than their core city. Moreover, this should not solve the question of considering the social deprivation pockets in some parts of 11 This is the conception of the New Economic Geography, which states the «Efficiency vs. Cohesion debate». See for example: World Bank (2009), World Development report 2009, Spatial disparities and Development Policy, Washington D.C.; Martin, R. (2008), National growth versus spatial equality? A cautionary note on the new trade-off thinking in regional policy discourse, Regional policy and Practice, 1, pp See David Q., Peeters D., Van Hamme G., Vandermotten C. (2013), «Is bigger better? Economic performances of European cities, », Cities, 35, December 2013: Gerrymandering means a cutting off of political units, designed to favour some specific political parties, mainly used in countries using electoral systems with uni-nominal constituencies. 14 ESPON database portal, Morphological and Functional Urban Areas. 21

26 Policy Department for Structural and Cohesion Policies these metropolitan wealthy areas, located in their centres or in their suburban areas following these countries. Therefore, although NUTS 2 is not in theory the best possible geographical division to allocate resources, it remains the best available level in terms of size and it does correspond to actual administrative level in most countries. Hence, NUTS 2 has to remain the level at which cohesion policies should be considered: the State level or even the too big NUTS 1 level could lead to an aggravation of the intra-state inequalities, at the expense of the remote regions, and the NUTS 3 level is a too narrow one to consider territorial cohesion policies, often smaller than the main functional metropolitan areas. Moreover, many indicators are not available at this level. The question of the geographical scale at which the eligibility criteria should be defined has completely changed whether GRI or FCE per head is used rather than GDP per capita. Table 3 shows the impact of the change of scale or of criterion on the spatial distribution of the population potentially residing in eligible areas. In general, using FCE per head (as a proxy for income) would result in narrowing territorial inequalities in lower shares of population in regions under the 90% or 75% criteria, with the few exceptions of capital regions in Central- Eastern European countries. In contrast, using GDP per capita at the NUTS3 level rather than NUTS2 would result in higher shares of the population residing in regions where the 75% threshold is not reached, though there may be some exceptions at the national level (in Portugal or Poland, for example). 22

27 Indicators in Cohesion Policy Table 3: Percentage of the national population living in eligible areas, following the kind of criterion and the geographical scale (2013 data) GDP per capita, NUTS 2 Private and public FCE per capita, NUTS 2 GDP per capita, NUTS 3 Threshold 90 % of the EU average Threshold 75 % Between 75 % and 90 % Threshold 90 % of the EU average Threshold 75 % Between 75 % and 90 % Threshold 90 % of the EU average Threshold 75 % Between 75 % and 90 % Belgium Bulgaria Czech Rep Denmark Germany Estonia Ireland Greece Spain France Croatia Italy Cyprus Latvia Lituania Luxembourg Hungary Malta Netherlands Austria Poland Portugal Rumania

28 Policy Department for Structural and Cohesion Policies Slovenia Slovakia Finland Sweden United Kingdom EU Source: author 24

29 Indicators in Cohesion Policy 3. TOWARDS A MORE INCLUSIVE AND GLOBAL INDICATOR? KEY FINDINGS In accordance with the Europe 2020 Strategy, the study pleads for a synthetic indicator considering four dimensions: economic development, physical wellness, social vulnerability, and technological development. Population dynamics will also be included as an additional dimension. Environmental issues are very important, but the relevance of NUTS 2 region to tackle this issue is questionable, e.g. because pollution generated in a specific place may have consequences somewhere else, as it is obvious in the case of greenhouse gas emissions. Beyond the proposition to replace GDP per capita by the criterion of private and public FCE per head, which remains a purely monetary criterion, it could be interesting to examine the possibility to use also other economic, social and environmental indicators easily available at the NUTS 2 level in all EU Member States and which might provide more information on the territorial development, within an economic but also in the perspective of social cohesion. This analysis classifies indicators in five categories: physical wellness (health), social vulnerability, education and technological development, population dynamics and environment. In addition to critically analysing these indicators and their meanings in terms of cohesion, the study will examine, for those which are available at the NUTS 2 scale, their level of correlation for the whole EU (some correlations could be quite different if observed between the regions of one country alone 15 ). Doing so will help improve the understanding of the indicators and help eliminate or keep some of them, either because they are redundant, or, on the contrary, because their statistical independence takes into account different dimensions of the economic and social phenomena. Rather than examining each indicator and the correlations between them individually and critically, the study could have used more sophisticated statistical methods, like the principal component analysis, to reveal the main independent dimensions related to the spatial variance of the whole set of the different indicators. However, doing so could lead to the proposal of a final synthetic indicator that is difficult to understand and to interpret for the political authorities and the public opinion; tools used for defining eligibility criteria have to be very transparent and understandable for anybody One could imagine, by way of a hypothesis, a strong statistical link in the EU between GDP per capita and the part of young people in higher education, as a result of strong variations in GDP per capita between the most and the least developed countries. Inversely, in some EU countries, young people from the poorest regions could be more desirous to graduate because they consider diplomas as springboards to find a job, or even to emigrate to the wealthiest parts of their country or elsewhere in the world, while in richer areas, access to employment is easier, even without a high school diploma. 16 We have also not considered the possibility to use an indicator like HDI (Human Development Index) or its gendered by-indicators, used by the United Nations in their development studies and programs. Moreover the fact that these indicators are not statistically well-funded (arithmetic averages between logarithmic values and values standardized between theoretical maxima and minima), using logarithms crushes the high values and so reduces the efficiency of this indicator in developed countries. 25

30 Policy Department for Structural and Cohesion Policies 3.1. Economic development Following our previous conclusions, the estimation of the global economic development of the NUTS 2 regions will be approached through the private and public Final Consumption Expenses by inhabitant (FCE per capita) 17. However, as GDP per capita (in PPS) remains until now the main indicator to allocate the European funds, it will also be kept in mind into the further analysis Physical wellness (health) It seems legitimate to consider the population s global health among the criteria assessing the level of the social cohesion Life expectancy at birth (E0) E0 (i.e. average number of years a generation should live under the current mortality conditions at each age) is the most global indicator of population health. It reflects the sanitary, environmental, nutritional, etc., living conditions of a population and is available for males, females, and the average of both. This analysis will only retain male life expectancy at birth, because the correlation of this indicator with material well-being (FCE per head) and economic indicator (GDP per capita) is higher for men than for women (respectively 0.73 vs. 0.55, and 0.46 vs. 0.34; it is worth noting that these correlations are significantly higher regarding FCE per capita than for GDP per capita 19 ). Male life expectancy also reflects a higher sensitivity to the improvement of the economic, social or even environmental conditions. Hence, the coefficient of variation (mean distance to the mean, i.e. standard deviation, expressed as a percentage of the mean; unweighted) of female life expectancy between European NUTS regions amounts to 2.5% vs. 3.7% for men. For this reason too, this study is opting to use Male life expectancy at birth as a measure of global health development. 17 Let us recall that we have considered that the public FCE per capita was identical for each region of a country. 18 Let us however recall once more that the significance of the GDP per capita is particularly untrustworthy in some NUTS 2 units because it overestimates the central city where administrative borders are small underestimates surrounding administrative areas. This is the list of the NUTS 2 units the most affected by this issue: in Belgium: Brussels-Capital, compared to Flemish and Walloon Brabant; in the Czech Republic: Prague compared Stredni Cechy; in Germany: Berlin, compared to Brandenburg, Bremen, compared to Lüneburg and Weser-Ems, Hamburg, compared to Lüneburg and Schleswig-Holstein; Gd. Duchy of Luxembourg, compared to the Belgian Luxembourg, Trier and Lorraine; in Austria, Wien, compared to Niederösterreich; in Slovakia, Bratislava compared to Zapadne Slovensko; in Britain, the two Inner London and the three Outer London, compared to Bedfordshire and Hertfordshire, Essex, Berks, Buckinghamshire and Oxfordshire, Surrey, East and West Sussex and Kent. 19 A coefficient of correlation between two sets of variables varies between 0 and 1 (or -1). 0 means that it is not any statistical relation between the two variables, and 1 that they are perfectly linked, that is by knowing the value attributed to an individual for the first variable, its value for the second one can be calculated exactly (-1 if they are perfectly negatively linked, one set of variables being growing when the other one is declining). 26

31 Indicators in Cohesion Policy Figure 1: Level of correlation between health and two economic indicators (GDP and FCE per capita) The graph shows the level of the correlation at the NUTS 2 level between each pair of indicators, here GDP per capita, FCE per capita, expectancy of life (male, female and total) and infant mortality rate. Source: Author Infant mortality rate (IMR) On the world scale, infant mortality rate 20 is a good indicator of sanitary conditions and of the quality of the health system. However, this rate is so low in developed countries that it shows very limited national and regional variations. The variations between countries at a so low level may sometimes only reflect different conditions of stillbirth registration (as well as the access to abortion of foeti recognized as deformed in prenatal tests, or even the quality or efficiency of these tests) 21. A synthetic health indicator taking into account infant mortality rate, female life expectancy and male life expectancy could be calculated, but for reasons already mentioned, it is better to avoid as much as possible the synthetic indicators when they are not indispensable. Moreover, in order not to multiply the number of indicators, and since the correlation between Male life expectancy is better with FCE per head (and with GDP per capita) than for Female 20 Infant mortality rate is the ratio between the number of deaths between 0 and 1 year and the number of people of this same age class. 21 The Gini coefficient is a measure of inequality in the social distribution of incomes. It varies from 0, perfect equality, to 1 or 100 (one individual hasis in possession of all the revenues). 27

32 Policy Department for Structural and Cohesion Policies life expectancy and for IMR, it is considered in this study that male life expectancy rate at birth is a sufficient and satisfying indicator for global health conditions Conclusion Male life expectancy at birth can be considered as the best indicator for global physical wellness conditions Social vulnerability Regional economic indicators, either GDP or FCE, only allow to assess average economic wellbeing. But this average may hide more or less intense differences among the regional population. This is why we also need to measure social vulnerability within each region, by using indicators of social inequalities, of poverty, of unemployment, Some prosperous areas with very high average level of GDP or FCE per capita, may also have high proportion of vulnerable populations, as it can be observed in many European large cities Gini coefficient The criterion of FCE per capita can be considered as an indicator of the average economic well-being, that is the level and quality of goods and services the population can benefit from. But, regarding social cohesion, it indicates nothing about the income distribution between the different social classes in a specific region. For instance, it is true that in large metropolitan areas benefiting from globalization, average income may be high, but a large part of their population is excluded from this prosperity, since growth is generally boosted by highly qualified activities, whereas those areas also include numerous low qualified people, who therefore face high underemployment rates and low incomes. The Gini coefficient 22 is a global indicator of inequality in income distribution. It measures the gap between real and uniform income distribution. Unfortunately, this indicator is not available at the regional level and is not even calculated on an annual coherent basis on national scales. It is often calculated on tax revenues, which do not correspond to the total revenues (notably in cases in which the poor are not required to fill in income tax forms). In addition, the Gini index suffers from the fact that one and the same value of this indicator can represent both an inequality against the most deprived and a more equitable income distribution within the most wealthy classes (less «very rich» among «rich» people). It therefore appears both practically impossible and scientifically non pertinent to use the Gini coefficient as an indicator combined to FCE per head. However, it is very important to note that national and European statistical apparatus have not elaborated relevant and homogeneous methodology and hence do not produce reliable statistics on social inequalities, even at the national level. This certainly presents a necessary improvement to be made Population at risk of poverty after social transfers This indicator is based on a national reference, as it considers the proportion of the population whose adjusted (that is taking into account households' size) disposable income is inferior to 60% of the national median 23. This is thus not an indicator of absolute poverty but of relative poverty compared to the national reference. 23 The national median income is the level which divides the national population in two equal groups. 28

33 Indicators in Cohesion Policy It is an indicator of social inequalities only taking into account the lower revenues, and not considering the distribution of incomes among the wealthiest parts of the population. As such, this indicator can appear as another measure of social inequalities faced by the poorest and a possible alternative to the Gini coefficient. However, it raises the issue of a national reference to measure the distribution of poverty risk on a regional scale. Indeed, the population below the 60% threshold of median national income (which defines the risk of poverty) can be spatially concentrated in some regions. For example, in Belgium, most of the population at risk of poverty is located in Wallonia, especially in old industrial cities. Consequently, this indicator may be a bit confusing: why using national reference to measure poverty across European regions? Why relative poverty should be assessed regarding national reference rather than a regional or a European one? In addition, statistics are based on SILC survey 24 and are not available for all countries at the NUTS 2 level, because of the insufficient sample of population in the inquiries. Hence SILC survey allows providing reliable statistics only for a part of NUTS 2 areas but not for all, especially for NUTS 2 areas belonging to large countries. Therefore, it is proposed that the use of this indicator is not included The part of households at risk of material deprivation This indicator measures the percentage of the population deprived of the possibility to achieve at least 4 out of the 9 following items: ability to face unexpected expenses, ability to pay for a one-week annual holiday away from home, existence of arrears on bills (mortgage or rent payments, utility bills, or hire purchase instalments or other loan payments), capacity to have a meal with meat, chicken or fish every second day, capacity to keep the home adequately warm and ability to afford a washing machine, colour TV, telephone or car. It is, in theory, more valuable than the Gini coefficient and the Population at risk of poverty after social transfers, since it focuses on the measure of social deprivation in a homogeneous and practical way. Indeed, rather than considering the insufficiency of incomes, it concretely measures deprivation of social needs considered as basic in all European societies. However, as for Population at risk of poverty after social transfers, this indicator is not based on exhaustive statistical sources but on the SILC survey and is thus not available at the regional level in several big countries (UK, Germany, France) 25. For the future, it is suggested that this indicator should be available at the NUTS 2 level in Europe, allowing for a reliable and homogeneous measure of social vulnerability. Of course, it supposes to increase the enquiry sample, and to ensure that this sample is representative of each NUTS 2 area Other indicators However, as under the current conditions indicators of social inequalities and material deprivation at the regional level are not available, proxies are here suggested. If they are surely not the best measures of social vulnerability, they are at least already available at the NUTS 2 level. They include the following indicators: 24 SILC survey (the Survey on Income and Living Conditions) is an European household survey covering a broad range of issues in relation to income and living conditions. It is the official source of data on household and individual income and also provides a number of key national poverty indicators, such as, moreover the at risk of poverty rate, the consistent poverty rate and rates of enforced deprivation. 25 SILC enquiries use sample of population representative mostly of the national population and not of the population of each region. For statistical reasons, the size of the sample is similar for large and small countries, making more difficult to use these statistics for NUTS 2 regions of large countries because the sample size is smaller for each NUTS 2 region. 29

34 Policy Department for Structural and Cohesion Policies Share of the population aged with a graduation of the inferior secondary level or less (total, male, female) Young people aged not in work, education or training (total, male, female) Total unemployment rate (total, male, female) Long-term unemployment rate These indicators are relatively poor proxies of social exclusion and precariousness. Indeed, the Share of the population aged with a graduation of the inferior secondary level or less or the share of young not in work or education can mean different things from one country or region to another. For example, in some countries, as it was not so long ago the situation in Germany, lower graduation does not necessarily mean difficulties in getting a job, while in other parts of Europe it may be the case. Unemployment figures also raise a major issue: even if the European statistics of unemployment (total and long term) are based on a homogeneous definition, their levels reflect not only the economic situation or even the local labour market situation, but also the level of social protection in each country and the more or less easiness to benefit from (and to keep) the unemployment grants. Therefore, a lower level of unemployment does not necessarily mean a lower level of precariousness or social exclusion, as indicated by comparisons between for example France and the UK. However, unemployment remains a valuable indicator of social deprivation as well as the situation of young people. Figure 2: Level of correlation between social vulnerability indicators The graph shows the level of the correlation at the NUTS 2 level between each pair of indicators, here people with no job, not in training or not at school (male, female and total), people with a low educational level (male, female and total), rate of unemployment (male, female and total) and rate of long term unemployment (total). A high level of correlation appears between the two first and the two last groups of criteria. Source: Author For each of these indicators, the levels of the correlation between the total, the male and the female values are very high, so it is enough to consider the average rates. Even if the correlation levels between these four indicators are all very significant, two groups appear: on one side, unemployment and long-lasting unemployment, and on the other the level of exclusion of young people and the low level of education among young adults. 30

35 Indicators in Cohesion Policy Conclusion As a consequence, this study proposes to keep two indicators to assess social vulnerability: the average between the index of unemployment and of long-lasting unemployment, and the average between the index of young people excluded from work, education or training and of people aged with a low educational level. The average of both indicators is taken as a synthetic indicator of social vulnerability. However, this synthetic indicator has to be interpreted very cautiously, as a worse index could in some case mean better access to unemployment grants, hence not necessarily reflecting really higher social deprivation Education, technological development and access to information The issue of education of the population and access to technology seems important to assess the situation of European regions in terms of territorial cohesion. They can reflect the quality of human capital for economic development, but also have a social value independent from any economic consideration. The following indicators will be examined: share of manpower working in RD; the RD expenses as a ratio of GDP; the number of patents per head; the proportion of years old with a high level of graduation. The figure below shows the correlations between these four indicators at regional level in the EU. Figure 3: Level of correlation between educational and R-D indicators The graph is built on the same way as Fig. 1 and 2. Source: Author 31

36 Policy Department for Structural and Cohesion Policies Percentage of year-old population with a high level of graduation (total, male, female) 26 This indicator is well correlated with GDP per capita (r=0,51) and to a bit lesser extent with FCE (r=0,46). The male and female population have to be considered separately, because data are more discriminating when one considers men rather than women (the correlation between high levels of female and male education is moreover not so high: r = 0.78). This probably results from the recent and spatially unequal trend of higher shares of women accomplishing higher education than men (42.9% vs. 33.8% in the years-old cohort), and maybe also from a differentiated propensity to migrate between males and females Percentage of population using the Internet With the spread of the Internet among the European population, such an indicator becomes less pertinent than it was some years ago Other indicators Concerning the technological development, three indicators are available: Intra-muros RD expenses, that is Research and Development made in the entreprises, in % of the GDP; RD manpower and researchers, in% of the active population; Patents asked in relation to the active population. The correlations between these three dimensions are pretty high and statistically significant as well as for each of them with GDP or with FCE. The study will keep these three indicators for further analysis Conclusion It is proposed in this study to keep the share of the male and female population with a high level of education and the three indicators related to the technological developments as good indicators to assess the average educational and technological levels reached by the population Population dynamics This dimension has been included in this analysis considering that a negative migratory balance can mean a regional shrinkage: not enough jobs, environmental problems, a bad quality of life, etc. At the same time, a shrinking population may impact negatively the regional economy 27, by the impact it has on the level of consumption as well as the negative impact of the outmigration of young educated people. Conversely, a strong positive migratory balance could indicate regional attractiveness (meaning also sometimes attractiveness for deprived migrants, like in some big metropolitan areas). 26 Let us remark that high educational level is only weakly (negatively) correlated with low educational level (r= ), and also that if the level of correlation between high educational level and economy is not very high, it is yet much lower concerning low educational level ( with GDP and - 0,23 with FCE). If high educational level is partially linked to the economic dimension, low educational level is mainly a social matter. 27 See Grasland et al. (2008), «Régions en déclin: un nouveau paradigme démographique et territorial», 2008, Rapport pour le Parlement Européen, 32

37 Indicators in Cohesion Policy Population dynamics in peripheral areas are a very important issue, and can be a sign of a regional vicious circle: bad economic situation lead to outmigration of young people, resulting in lower birth rate, which may impact again the economic situation of the region 28. The correlation rate between migratory balance and the GDP per capita is 0.43, and the one with FCE reaches The migratory balance will thus be considered, even if this indicator has to be used with caution, as it is computed as the difference between the total evolution of the population and the natural balance, thus including possible statistical adjustments Environment It is better not to take the environmental dimension into account at the regional level. For example, one can wonder if it is preferable to equalize emission levels on a state s territory at the level of the national average, or to have regions where emissions are higher and others where they are lower. The objective should obviously be the reduction of the global volume of emissions, at the national if not at the European level. It is thus difficult to assert the relevance of environmental indicators at the NUTS 2 level, as a basis for the allocation of structural funds. First, pollutants can be produced in one area and have negative impacts on other regions. It thus raises the question of which region should be funded. Second, the emission of pollutants can be related to the regional wealth and population density and, as a consequence, to introduce such indicators may lead to help wealthy regions. Third, indicators about the environmental issue are most of the times not produced at regional level and when they are available, they should be used with caution because they are produced by the generalization of punctual measures. In conclusion, the environmental dimension will unfortunately not be taken into account as a relevant criterion of underdevelopment, though of course this dimension can be included in actions financed through structural funds Conclusions about a synthetic indicator In summary, the study will keep an aggregated indicator for each of the five dimensions (economic development, physical wellness, social vulnerability, educational and technological development and population dynamics) to avoid the overweighting of one or another dimension due to the use of more indicators. We consider thus implicitly in this exercise that the territorial cohesion, i.e. unequal geographical distribution of economic and social wellbeing across European regions, includes simultaneously and equally five dimensions: an economic one, physical wellness, social vulnerability, educational and technological level and the level of attractiveness (or of repulsion) of areas for mobile people. To facilitate understanding, each dimension will be defined by the arithmetical average of the standard deviations to the EU28 average of each indicator, rather than using more sophisticated statistical techniques, like principal component analysis. For each of the five dimensions, we will use the following indicators: For the economic dimension: either GDP, or better private and public FCE per capita. For the physical wellness dimension: male expectancy of life at birth. 28 Grasland et al. (2008), Régions en déclin: un nouveau paradigme démographique et territorial, Rapport pour le Parlement Européen, 33

38 Policy Department for Structural and Cohesion Policies For the social vulnerability dimension: the average between four indicators: unemployment, long-lasting unemployment, share of young people (18-24) excluded from work, education or training, and people aged with a low educational level. Let us recall that the values on this dimension could be influenced by the quality of the protection against unemployment. However, this report pleads for the elaboration of more reliable indicators of social exclusion and material deprivation at the NUTS 2 level, because no available indicator is fully satisfying on this essential dimension. For the educational and technological dimension: the share of year-old male and female population with a high level of graduation; intra-muros R-D expenses, in % of the GDP; R-D manpower and researchers, in % of the active population; patents requested in % of the active population. In the synthetic indicator for this dimension, the same weight has been given to the two first indicators and to the three last ones. For the regional attractiveness: the migratory balance. A synthetic indicator taking simultaneously the 5 dimensions into consideration, i.e. using the distribution of the arithmetical mean of the standard deviations for each dimension, has been mapped. 34

39 Indicators in Cohesion Policy 4. GEOGRAPHICAL PATTERNS OF THE MAIN INDICATORS AND GENERAL CONCLUSION KEY FINDINGS The use of FCE instead of Gross Regional Product smoothes the intra-state regional disparities, because social transfers within each member state tends to reduce regional inequalities in terms of incomes and consumption. The use of a synthetic index which takes into account not only the economic dimension but also physical wellness, social vulnerability, population dynamics and technological development reintroduces more intra-state variations as compared to the sole use of an economic indicator. Moreover, the study proposes to grant the benefit of full eligibility to any region of the countries with a very low investment rate (i.e. Greece and Cyprus), because very low investment rate hampers future development. This study has already emphasized, from a theoretical point of view, that FCE per capita should be preferred to GDP per capita as an indicator for the economic dimension. Figure 54 reinforces this conclusion, considering the territorial and social cohesion from a more global point of view: it shows that, considering the five dimensions listed in section 3.7, the correlations are better towards FCE than towards GDP. Figure 4: Correlations (R) coefficients between GDP or private and public FCE per capita and each of the four other dimensions The graph is built on the same way as Fig. 1 and 2. Source: Author The geography of each dimension and of the global synthetic indicator will be examined below, also by comparison to the GDP or FCE patterns (see Maps 1 to 3). To do so, the study will consider as limits the values that include in each class the same share of the European population as those corresponding respectively to less than 75%, 90%, 100% and 120% of the GDP per capita. Let us recall that the population living in regions where the level of GDP per capita is below 75% of the EU average includes 26.9% of the EU population, and the population living in regions below 90% of the EU average includes 44.9% 35

40 Policy Department for Structural and Cohesion Policies of the EU population 29. The patterns will also be examined through maps using the following thresholds for each dimension: less than 0.5 standard deviation to the EU average, less than 0.25, the EU average and more than 0.5. See also the table in the annex The geographical pattern of physical wellness Maps 4 (left) and 5 (right): Geographical pattern of the physical wellness dimension (male expectancy of life at birth) Map 4 (on the left) uses class limits including in each class the same share of the EU28 population as those corresponding respectively to less than 75%, 90%, 100% and 120% of the GDP per capita. Map 5 (on the right) uses limits corresponding to 0.25 or 0.5 standard deviations to the average of the phenomena (standard deviation is the average of the absolute values of the deviations of the individual values to the average, divided by this average). The legend shows the cumulative percentages of the European population included in each class. Source: EUROSTAT, IGEAT 2016 By comparison to GDP and FCE, this indicator improves the position of the Mediterranean countries, and worsens the situation in Northern France and Northern Germany. The bad position of Central-Eastern Europe is reinforced. These differences in life expectancy between countries of similar economic development may be related to differences between the health system or alimentary diet, well documented in the demographic literature % of the EU population live in regions below the GDP per capita level 100, 77.6% in these below level 120. Using FCE/inhab. and the same thresholds of 75%, 90%, 100% and 120% of the European average, the values should be respectively 19.4%, 32.0%, 38.4% and 80.1%. 36

41 Indicators in Cohesion Policy 4.2. The geographical pattern of social vulnerability Map 6 (left) and 7 (right): Geographical pattern of the social vulnerability dimension Map 6 (on the left) uses class limits including in each class the same share of the EU28 population as those corresponding respectively to less than 75%, 90%, 100% and 120% of the GDP per capita. Map 7 (on the right) uses limits corresponding to 0.25 or 0.5 standard deviations to the average of the phenomena (standard deviation is the average of the absolute values of the deviations of the individual values to the average, divided by this average). The legend shows the cumulative percentages of the European population included in each class. Source: EUROSTAT, IGEAT 2016 By comparison with the GDP and FCE patterns, this indicator emphasizes countries with a high level of unemployment, like Spain and Italy. Conversely, it improves the position of the Central-Eastern European countries where unemployment is weak (but partially perhaps due to a strong temporary or definitive outmigration towards Western Europe and/or weak unemployment grants), like Poland and Czech Republic, and to a lesser extent Slovakia, Hungary, Romania and Bulgaria. Such a geographical pattern illustrates the insufficiencies of the available indicators, since poverty rates are very high in Central-Eastern Europe but do not necessarily result from high unemployment figures; in some cases, lower unemployment may reflect lower development, like in eastern Poland for example. The same is true for North-Western Europe: lower levels observed in Northern early manufacturing England compared to old Wallonia basins and Nord-Pas-de-Calais occur due to differences in labour market regulations rather than better living conditions and a lower exclusion rate. 37

42 Policy Department for Structural and Cohesion Policies 4.3. The geographical pattern of the educational and technological dimension Map 8 (left) and 9 (right): Geographical pattern of the educational and technological dimension Map 8 (on the left) uses class limits including in each class the same share of the EU28 population as those corresponding respectively to less than 75%, 90%, 100% and 120% of the GDP per capita. Map 9 (on the right) uses limits corresponding to 0.25 or 0.5 standard deviations to the average of the phenomena (standard deviation is the average of the absolute values of the deviations of the individual values to the average, divided by this average). The legend shows the cumulative percentages of the European population included in each class. Source: EUROSTAT, IGEAT 2016 The pattern of this indicator reveals worse situations than expected from the level of GDP per capita in some peripheral areas of the richest European countries (North-Eastern Netherlands and Zeeland, Eastern Germany, peripheral parts of the United Kingdom) and in early coal-mining and manufacturing basins (Nord-Pas-de-Calais, Northern England, Hainaut in Belgium, ). In contrast, capital cities or other major metropolitan areas show very high levels for this indicator. A low level of educational level is observed in the Mediterranean countries (including in Northern Italy) also characterized by weak figures of the RD expenses, yet more so in the Central-Eastern European countries. 38

43 Indicators in Cohesion Policy 4.4. The geographical pattern of the regional attractiveness Map 10 (left) and 11 (right): Geographical pattern of the global attractiveness (migratory balance) Map 10 (on the left) uses class limits including in each class the same share of the EU28 population as those corresponding respectively to less than 75%, 90%, 100% and 120% of the GDP per capita. Map 11 (on the right) uses limits corresponding to 0.25 or 0.5 standard deviations to the average of the phenomena (standard deviation is the average of the absolute values of the deviations of the individual values to the average, divided by this average). The legend shows the cumulative percentages of the European population included in each class. Source: EUROSTAT, IGEAT 2016 The pattern of these maps reveals a double centre periphery structure, i.e. at the European scale between central countries vs. peripheral countries and, in each country, between central regions (especially capital cities) vs. peripheral regions. The maps reveal the recent loss of attractiveness of Ireland and of Southern Europe in contrast to North-Western Europe. These old emigration countries had progressively become immigration countries in the 1990s, but the 2008 crisis completely reversed the situation, and the migratory balance has become negative again. In contrast to the previous waves of emigration from South European countries (going back to the 1950 s) situation, the recent wave of emigration from Mediterranean countries (after the 2008 crisis) is primarily due to the outmigration of young educated people. This emigration of educated people may have negative impacts for economic development in the long run. Beyond the core-periphery pattern mentioned earlier, other major territorial contrasts can be observed, between Northern (including Paris at the reverse of most other main European metropolises) and Southern France, between Northern and Southern Italy and between Eastern and Western Germany. 39

44 Policy Department for Structural and Cohesion Policies 4.5. A synthetic overview Map 12: Geographical pattern of the synthetic indicator, by comparison to GDP and FCE per capita (see Maps 1 and 2) Source: EUROSTAT, IGEAT 2016 Compared to the FCE map (see Map 2), using the synthetic index that includes five dimensions reintroduces more intra-state variations, like North vs. South in France, East vs. West and South in Germany, and the North-South dichotomy in Italy and to a lesser extent in Spain. The synthetic index also reveals the transition statute of Ireland in contrast to GDP figures which position this country at very high levels. Moreover, this study proposes to grant the benefit of full eligibility, independent of the level of the synthetic index (either FCE or GDP) in each region to Greece and to Cyprus, as these two countries suffer from a very low level of investment (Gross Capital Formation less than 15% of the GDP), which could impede their further economic and social development. Indeed, in countries where the share of consumption is high, the FCE indicator may overestimate their actual development level and, moreover, their future development perspectives. This is why we believe that some countries where gross capital formation is abnormally low should receive specific attention, given that current low investments may result in weak economic growth in the future. 40

45 Indicators in Cohesion Policy 5. THE IMPACT OF THE UK REFERENDUM KEY FINDINGS The consequences of the UK referendum should reduce the allocations for the structural funds, since the UK is a net contributor to the EU budget. The extent of this loss is of course function of the future relation between the UK and the EU. The elementary scenarios illustrated here are based on the total loss of the UK net contribution, which is a maximum scenario, and on a linear reduction of all the budgets. The absolute value of the EU average GDP per capita will decrease. As a consequence, regions presently under the 75% or 90% threshold of the EU average may exceed this threshold after a Brexit and, all things being equal, not to be anymore in the criteria to benefit from EU structural funds. At the reverse, a Brexit will not change the eligibility of the countries to the Cohesion Funds. According to various scenarios and using a synthetic index, the impact of the UK referendum on the regional eligibilities results in higher losses in the most developed Western European countries and lower losses in some Central-Eastern European countries and in the Mediterranean countries, with Spain becoming possibly the second recipient after Poland. Measuring the impact of the UK referendum on the structural funds is a difficult and very conjectural exercise: United Kingdom could want to remain associated to some European policies (and to pay therefore); the European Union could decide to modify the relative funding of its different policies; or to raise additional financial resources. This exercise is yet more conjectural if one attempts simultaneously to add its impact to a modification of the indicators defining the eligibility. A lot of scenarios could be built, with various thresholds. In the following elementary exercise, only three scenarios (plus a null scenario) will be proposed with a rough estimation of their financial impacts. These three scenarios remain in the present framework, considering the percentage of the European Funds allocated to the structural and investment funds and to the regional policies and the contributions to the European budget. The total loss of the UK net contribution has only be considered, along with a linear reduction of each big category of all the budgets. Considering 2015 data, the UK is producing 17.2% of EU28's GNI and contributing 15.6% to the European budget. Even if it is contributing a bit less to the European budget than its part in EU28's GNI, due to its rebates, the UK appears to be a very important net contributor to the budget. Considering a linear reduction of each big category of budget, the budget available for the regional and cohesion funds (including the funds for Far North and Overseas, the territorial co-operation and YEI) should then decrease from 349,350 million Euros to 294,851 million Euros 30 i.e. a loss of 54,499 million Euros. However, part of this reduction will be compensated by the disappearance of the expenses of the structural funds in favour of the UK, which amount to 10,974 million Euros, plus 1,164 million Euros coming from the territorial co-operation (estimation proportional to the part of the UK in EU28's population). 30 Data for the financial allocations for the European structural and investments funds: Regional Policy InfoRegio, The EU's main investment policy. Total allocations of Cohesion Policy Breakdown by spending categories. Update 6/4/

46 Policy Department for Structural and Cohesion Policies The net reduction of the budget for the structural funds available for the 27 countries should then be 54,499-12,138 = 42,361 million Euros. Another consequence of the Brexit to consider is that the absolute value of the European average GDP per capita will decrease: regions that are now below the 75% threshold could exceed the same threshold computed on the basis of the new average (75% of the present average = 74.09% of the new average), or regions presently under the 90% threshold could be above the new one (90% of the present average = 88.90% of the new average). However, considering the criteria for the Cohesion Fund, where the 90% threshold concerns the GNI at the national level, no country should be concerned by the adjustment of the average. The null scenario considers that the present criterion, on the basis of the GDP per capita in PPS, remains as it is, only with a change of the thresholds: 75% of the European average is replaced by 74.09% of the European average and 90% by 88.9%. Exactly the same regions as now (evidently outside UK) could thus benefit from the less developed and transition eligibility conditions. The reduction of the total available budget is linearly transferred on each post of the budget. Each country will thus see its allocation reduced by 12.6%. In the first scenario, called the GDP scenario, it is considered that GDP remains used as now, and also that the 75% and 90% thresholds are kept, but calculated according to the new EU average. Some regions which are now among the less developed should then move to the transition category, and some pertaining now to the transition category should move to the more developed category. Stredni Cechy, Estonia and Molise are in the first situation, and Burgenland, Bourgogne, Bretagne, Corse and Umbria belong to the second one. The reduction of the allocations for the less developed regions moving to the transition category has been estimated at 50%, proportionally to their part of the population of this category in each country, and another reduction by half has been applied on the same basis to the regions moving from the transition to the more developed category. Then, as the total amount of the recomputed budget remained higher than the available amount (337,208 million Euros), each post of the expenses has been linearly reduced to fit the available amount of 294,851 million Euros. 42

47 Indicators in Cohesion Policy Table 4: Absolute level of the allocations after the UK referendum following the four scenarios and relative losses by comparison to the present situation. Millions Euro Present situation ( ) Scenario 0 (decrease of the thresholds to 88.9 % and 74.1 % of the new European average) Scenario GDP (maintain of the thresholds at the levels of 90 % and 75 % of the (new) EU average) Scenario FCE (thresholds at the levels of 90 % and 75 % of the (new) EU average) Scenario Global synthetic index, considering the same percentage of the EU27 population as now in the less developed and transition categories (b) Loss of allocation with the Scenario 0 (in % of the present situation) Loss of allocation with the Scenario GDP (in % of the present situation) Loss of allocaton with the Scenario FCE (in % of the present situation) Loss of allocation with the global synthetic index, considering the same percentage of the EU27 population as now in the less developed and transition categories (b) EU Belgium ,6 11,8 31,1 32,7 Bulgaria ,6 11,8 7,3 20,8 Czech Republic ,6 16,0 3,7 29,2 Denmark ,6 11,8 16,7 28,9 Germany ,6 11,8 32,1 31,0 Estonia ,6 42,5 7,3 20,9 Ireland ,6 11,8 7,3 1,0 Greece ,6 11,8 26,2 (a) -6,3 Spain ,6 11,8 7,1-48,2 France ,6 14,6 32,5 22,5 Croatia ,6 11,8 7,3 20,8 Italy ,6 12,6 37,8 7,8 Cyprus ,6 11,8 7,2 (a) 20,7 Latvia ,6 11,8 7,3 20,8 Lithuania ,6 11,8 7,3 20,8 43

48 Policy Department for Structural and Cohesion Policies Luxembourg ,6 11,8 7,3 20,8 Hungary ,6 11,8 1,3 20,8 Malta ,6 11,8 7,3 20,8 Netherlands ,6 11,8 7,3 20,8 Austria ,6 15,0 10,7 23,8 Poland ,6 11,8 2,7 19,3 Portugal ,6 11,8 24,5 12,5 Romania ,6 11,8 3,6 17,7 Slovenia ,6 11,8-18,8-1,4 Slovakia ,6 11,8 7,3 20,8 Finland ,6 11,8 7,3 20,8 Sweden ,6 11,8 7,3 20,9 Territorial co-operation (a) The loss for Greece and Cyprus could be reduced if some budget was specifically allocated to the countries with a very low level of gross capital formation. In this case, the loss for the other countries should be a bit higher than computed here. (b) To be perfectly correct the five components of the synthetic index should have been recomputed excluding UK data. It has not been done for this first estimation, but the results should not have been very different. Source: author In this first scenario, all EU27 members except four undergo the same level of reduction ofin this first scenario, all EU27 members except four undergo the same level of reduction oftheir allocations, i.e %. Italy undergoes a stronger reduction (12.6%), and even more do France (14.6%), Austria (15.0%) and the Czech Republic (16.0%). In a second scenario, called the FCE scenario, the thresholds of 90% and 75% of the (new) European average have been applied to the total FCE expenses per capita. With this hypothesis, more regions would change category, either because the use of the new criterion reduces the intra-national disparities or due to a lower level of the new European average. The results should be a much stronger loss of allocations for the big Western European countries where the intra-national transfers of income are important (from the North to the South in Italy, from Paris to the rest of the country in France, from Western to Eastern Germany). Conversely, the losses should be reduced in the Central-Eastern European countries, where the capital cities concentrate a high amount of the production of the economic value, but where the real income falls in fact under the threshold levels after transfers (Czech Republic, Romania, Poland, Hungary). The allocations should even increase in Slovenia by comparison to the present situation. As it was explained in the report, special allocations could be considered in favour of Greece and to a lesser extent Cyprus, countries where the FCE level is combined with a very low level of capital formation, which could handicap further economic development. 44

49 Indicators in Cohesion Policy Finally, the third scenario takes into consideration the global synthetic index. Keeping the same proportions of the EU27 population in the less developed and transition categories leads to consider regions under an average of standard deviation as less developed and under as in transition. For countries with the same percentage of their population in the three categories as now, the loss is then 20.8% of their present allocations. The loss is stronger in the most developed Western European countries (Belgium, Germany, Denmark, Austria, France) and also in the Czech Republic. It is less in some Central-Eastern European countries (Poland, Romania) and mainly in the Mediterranean countries, with even a growth of the allocations in favour of Greece and mainly of Spain (also slightly in Slovenia). Spain should then become the second recipient of the allocations in volume after Poland, the latter remaining largely the first one in each scenario. This impact on the position of Spain according this new indicator is mainly due to the weight of the high-level social vulnerability and of the bad migratory balance in this country and also of a weak educational and technological dimension in its Southern regions. 45

50 Policy Department for Structural and Cohesion Policies 46

51 Indicators in Cohesion Policy 6. CONCLUSION In conclusion, this study pleads for the introduction of a synthetic index that includes all main dimensions of social and economic well-being in EU regions, i.e. economic well-being, physical wellness, social vulnerability, technological development and attractiveness. We believe such reform would be more in line with the evolution of the philosophy of regional development policies in general, which state that economic development, social and territorial cohesion are equally important, and especially the Agenda 2020 pleading for a smart, inclusive and sustainable growth (EC, 2010, 2011a, 2011b) 31. In other words if economic development does not benefit to the whole population - e.g. because revenues are exported or because economic growth benefits to a limited proportion of the regional population is a sufficient reason to benefit from European funds. Also, for example, if the education level is too low in a region, future economic development may be hampered, and it also necessitates actions at the EU level to help these regions. For each dimension of regional development, we propose one indicator or one synthetic indicator: For the economic dimension: either GDP, or better private and public FCE per capita. We propose to replace GDP per capita by total FCE per capita, since it is a better assessment of regional economic well-being. Indeed, GDP does not include most transfers between regions while FCE is a better indicator of the average volume and quality of goods and services the resident regional population can have; For the physical wellness dimension: male expectancy of life at birth; For the social vulnerability dimension: the average between four indicators: unemployment, long-lasting unemployment, share of young people (18-24) excluded from work, education or training, and people aged with a low educational level; For the educational and technological dimension: the share of year-old male and female population with a high level of graduation; intra-muros R-D expenses, in % of the GDP; R-D manpower and researchers, in % of the active population; patents requested in % of the active population. In the global indicator for this dimension, the same weight has been given to the two first indicators and to the three last ones; For the global attractiveness: the migratory balance. As has been explained in more details in Chapter 3, some important indicators are not available or not fully reliable at NUTS 2 level in the current situation. Hence, this list of indicators are not the best possible to measure the different dimensions considered but is based on the best available indicators. This is particularly the case for FCE used because GNI, which should replace GDP to assess regional economic well-being, is not produced at regional level. We must also note the absence of data of purchasing power parity at regional level, since economic indicators in PPS are based on national and not regional levels of prices. Concerning indicators of social vulnerability, the situation is much worse in terms of availability and reliability of regional indicators, which constrains us to opt for unsatisfactory proxies. The calculation of regional Gini coefficient or material deprivation index at NUTS 2 level is a much better option, but would probably require to increase sample size of the SILC survey, at least in some countries. 31 European Commission (2010), EUROPE 2020 A strategy for smart, sustainable and inclusive growth, Communication from the Commission. COM(2010) 2020 FINAL European Commission (2011a), The territorial state and perspectives of the European Union, Background document for the Territorial Agenda of the European Union 2020 European Commission (2011b), Territorial Agenda of the European Union 2020 Towards an Inclusive, Smart and Sustainable Europe of Diverse Regions, Agreed 47

52 Policy Department for Structural and Cohesion Policies In addition, we propose to take into consideration potentialities of future development. Certainly, the indicators for educational and technological dimension are important in this perspective, since many studies have shown the positive impact of education and innovation on economic development 32. But it may be not sufficient in some exceptional cases of very low investment. This is why we propose an additional criterion based on Gross Capital Formation, as a measure of investments. Indeed, low Gross Capital Formation may hamper future development for years and European funds may help to avoid an economic and social deterioration. This is why we propose a full eligibility statute for all NUTS 2 regions included in countries with a level of less than 15% of the GDP, which is the current situation in Greece and Cyprus. We understand that such additional criterion may be difficult to introduce. If such a measure cannot be implemented, specific actions should be considered in order to take into account these exceptional situations of very low investment rates. Finally, we analyze the impact of the proposed changes on the spatial distribution of European funds, giving a precise idea of the regions that may be losing or winning in consequences of these changes. Briefly said, the introduction of FCE per capita instead of GDP per capita would result in narrowing economic differences, hence in fewer regions under 75 or 90% EU average threshold. In contrast, the synthetic index proposed would reintroduce more inequalities between regions. Annex 1 gives the list of NUTS 2 regions which should be eligible or in transition following GDP, FCE or the synthetic index. Annexes 2 and 3 give a list of regions remaining in the same category or changing categories using FCE or the synthetic index instead of GDP. These annexes, and the consequences of the use of alternative indexes on the allocation of the resources of the funds are summarized on Fig. 5 and 6, which are built on the data including the United Kingdom and under the hypothesis to keep the same share of the EU28 population in the developed, intermediate and less developed categories as it is presently. 32 DG REGIO, European commission (2003), A study on the factors of regional competitiveness, eness.pdf. David, Q., Peeters, D., Van Hamme, G., & Vandermotten, C. (2013). Is bigger better? Economic performances of European cities, Cities, 35, doi: /j.cities Capello R. (2013), Territorial dimension of the innovation and knowledge economy, KIT, ESPON, Applied Research 2013/1/13, 48

53 Indicators in Cohesion Policy Figure 5: Reallocation of the EU28 population between the three categories on the basis of GDP, FCE and the synthetic index Source: Author 49

54 Policy Department for Structural and Cohesion Policies Figure 6: Share of the population of each EU28 country in the three categories on the basis of GDP, FCE and the synthetic index Source: Author 50

55 Indicators in Cohesion Policy Figure 5 shows that a large majority of the present developed and less developed regions should remain in the same categories using as criteria FCE or the synthetic index instead of GDP. Some countries remain fully developed, whatever the indicator used: Finland, Luxembourg, the Netherlands, Sweden (or nearly: Austria, Denmark). Others remain fully less developed : Bulgaria, Estonia, Croatia, Lithuania, Latvia, or nearly (Romania). The intermediate category should be the most impacted by using these new criteria, with a very significant shift towards the developed category of the present intermediate regions. This is mainly due to the fact that the GDP criterion doesn t take into account the impact of the direct and indirect redistributions of wealth between regions implemented in many Western European countries: it is clearly the case in the United Kingdom, in France and to a lesser extent in Germany, Denmark, Belgium and Austria. In Italy, the redistributive effects lead to a fading of the North-South divide which results in the extension of the intermediate category using FCE instead of GDP. However, including social criteria would worsen the situation of the South, so that the geographical distribution of the regions is more or less the same using GDP or the synthetic index. At the reverse, in Spain using FCE or the synthetic index clearly would worsen the situation: the number of the developed regions would fall at the expense of the intermediate regions, and present intermediate would become less developed regions. Replacing GDP by FCE or the synthetic index illustrates the strong concentration of the most dynamic sectors of the economy of the Central-Eastern and of some Mediterranean countries in their capital cities, which prevent these cities to benefit from the less developed and intermediate statutes, which could be more on line with the wealth of their population and their social problems: it is the case for Athens, Budapest, Warsaw, Lisbon, Bucharest, Bratislava. 51

56 Policy Department for Structural and Cohesion Policies REFERENCES Capello R. (2013), Territorial dimension of the innovation and knowledge economy, KIT, ESPON, Applied Research 2013/1/13, arch/kit.html David Q., Peeters D., Van Hamme G., Vandermotten C. (2013), Is bigger better? Economic performances of European cities, , Cities, 35, ESPON database portal, Morphological and Functional Urban Areas. European commission (2003), A study on the factors of regional competitiveness, DG Regio, eness.pdf. European Commission (2010), Europe 2020 A strategy for smart, sustainable and inclusive growth, Communication from the Commission, COM(2010) 2020 FINAL. European Commission (2011a), The territorial state and perspectives of the European Union, Background document for the Territorial Agenda of the European Union European Commission (2011b), Territorial Agenda of the European Union 2020 Towards an Inclusive, Smart and Sustainable Europe of Diverse Regions. European Commission, Regional and Urban Policy (2014), Sixth Report on Economic, Social and Territorial, Investment for jobs and growth: Promoting development and good governance in EU regions and cities. European Parliamentary Research Service (2016), Beyond GD: Regional development indicators. European Structural and Investment Funds (2016), Total Allocations of Cohesion Policy Breakdown by Spending Categories, InfoRegio, Update 6/4/2016. EUROSTAT, Système européen des comptes. SEC EUROSTAT, European Union Statistics on Income and Living Conditions (EU-SILC). Grasland C. et al. (2008), «Régions en déclin: un nouveau paradigme démographique et territorial», Rapport pour le Parlement Européen, Haas J. & Rubio E. (2017), Brexit and the EU Budget: Threat or Opportunity?, Policy Paper 183, Jacques Delors Institut Berlin & Bertelsmann Stiftung INSEE (2010), Enquête Emploi du temps Martin R. (2008), National growth versus spatial equality? A cautionary note on the new trade-off thinking in regional policy discourse, Regional policy and Practice, 1, Vandermotten C., Peeters D., Lennert M. (2011), Shaping EU regional policy: looking beyond GDP, Green New Deal Series, 7, Green European Foundation. World Bank (2009), World Development report 2009, Spatial disparities and Development Policy, Washington D.C. Wruuck P. & Rosenberger L. (2016), Think Local. What Brexit would mean for regional and cohesion policies in Europe, EU Monitor, September 30, European Integration, Deutsche Bank Research. 52

57 Indicators in Cohesion Policy ANNEX 1. TABLE OF THE POSITION OF THE REGIONS ACCORDING TO THE GDP, THE PRIVATE AND PUBLIC FCE AND THE FOUR OTHER SERIES OF CRITERIA ** : the worse situation * : intermediate situation The situation of the regions (worse or intermediate) is defined in the first and the third columns on the basis of the inclusion of the same share of the European population as for the thresholds of 75% and 90% of the average GDP per capita, i.e. 26.9% and 44.9% of the EU population. The second column considers as thresholds 75% and 90% of the average public and private FCE/inhab., including then respectively only 19.4% and 32.0% of the European population. In the five next columns, the situation for each kind of indicator is considered as worse when its value falls under -0.5 standard deviation from its European average and as intermediate when it falls under standard deviation. For the global synthetic index (last two columns), we have considered the same share of the EU population as presently in the categories full eligibility and intermediate, i.e. 26.9% and 44.9% of the European populations. By chance, these values correspond to the values -0.5 and 0.0 of the average of the values of the five middle columns. 53

58 Policy Department for Structural and Cohesion Policies GDP/inhab. (thresholds of 75 % and 90 % of the EU average, i.e and 44.9 % of the EU population) Public and private FCE/inhab.(thresholds of 75 % and 90 % of the EU average, i.e and 32.0 % of the EU population) Public and private FCE/inhab.(thresholds of 80.8 and % of the EU average, i.e and 44.9 % of the EU population) Public expenditure of general government (of which individual consumption) (of which collective consumption) Household and NPISH final consumption expenditure (of which households) Global synthetic index (average of the standard deviations ; theresholds -0.5 and -0.25, i.e and 33.2 % of the EU population) (Number of asterisks) AT11 Burgenland (AT) * 0 AT12 Niederösterreich 0 AT13 Wien 0 AT21 Kärnten 0 AT22 Steiermark 0 AT31 Oberösterreich 0 AT32 Salzburg 0 AT33 Tirol 0 AT34 Vorarlberg 0 BE10 Bruxelles-Capitale / Brussels Hoofdstad ** 2 BE21 Prov. Antwerpen 0 BE22 Prov. Limburg (BE) 0 BE23 Prov. Oost-Vlaanderen 0 54

59 Indicators in Cohesion Policy BE24 Prov. Vlaams-Brabant 0 BE25 Prov. West- Vlaanderen 0 BE31 Prov. Brabant Wallon 0 BE32 Prov. Hainaut * ** ** * * 5 BE33 Prov. Liège * * ** 3 BE34 Prov. Luxembourg (BE) * * * 2 BE35 Prov. Namur * * 1 BG31 Severozapaden ** ** ** ** ** ** ** ** ** 10 BG32 Severen tsentralen ** ** ** ** ** ** ** ** ** 10 BG33 Severoiztochen ** ** ** ** ** ** ** ** ** 10 BG34 Yugoiztochen ** ** ** ** ** ** ** ** ** 10 BG41 Yugozapaden ** ** ** ** ** ** 4 BG42 Yuzhen tsentralen ** ** ** ** ** * ** ** ** 9 CY00 Chypre * * * 1 CZ01 Praha * * 1 CZ02 Strední Cechy ** * ** ** ** ** 6 CZ03 Jihozápad ** ** ** ** ** ** * 6 CZ04 Severozápad ** ** ** ** ** ** ** ** 8 CZ05 Severovýchod ** ** ** ** ** ** ** * 8 CZ06 Jihovýchod * ** ** ** ** * * 5 CZ07 Strední Morava ** ** ** ** ** ** ** ** 8 CZ08 Moravskoslezsko ** ** ** ** ** ** ** ** 8 DE11 Stuttgart 0 DE12 Karlsruhe 0 DE13 Freiburg 0 DE14 Tübingen 0 DE21 Oberbayern 0 DE22 Niederbayern * 1 DE23 Oberpfalz 0 55

60 Policy Department for Structural and Cohesion Policies DE24 Oberfranken ** 2 DE25 Mittelfranken 0 DE26 Unterfranken 0 DE27 Schwaben * 1 DE30 Berlin 0 DE40 Brandenburg * ** * 2 DE50 Bremen 0 DE60 Hamburg 0 DE71 Darmstadt 0 DE72 Gießen 0 DE73 Kassel * 1 DE80 Mecklenburg- Vorpommern * * * ** * * 4 DE91 Braunschweig 0 DE92 Hannover 0 DE93 Lüneburg * ** 2 DE94 Weser-Ems ** 2 DEA1 Düsseldorf * 1 DEA2 Köln 0 DEA3 Münster ** 2 DEA4 Detmold * 1 DEA5 Arnsberg ** 2 DEB1 Koblenz * ** 3 DEB2 Trier 0 DEB3 Rheinhessen-Pfalz 0 DEC0 Saarland ** 2 DED1 Dresden 0 DED2 Chemnitz * * ** * 3 DED3 Leipzig * 0 DEE0 Sachsen-Anhalt * * ** ** ** * 6 DEF0 Schleswig-Holstein ** 2 56

61 Indicators in Cohesion Policy DEG0 Thüringen * * ** ** * 4 DK01 Hovedstaden 0 DK02 Sjælland * * 1 DK03 Syddanmark * 1 DK04 Midtjylland 0 DK05 Nordjylland 0 EE00 Estonie ** ** ** ** ** ** ** 6 EL11 Anatoliki Makedonia, Thraki ** * ** ** ** ** ** ** 8 EL12 Kentriki Makedonia ** * * ** ** * ** ** 7 EL13 Dytiki Makedonia ** * * ** ** ** ** ** 8 EL14 Thessalia ** * * ** ** ** ** ** 8 EL21 Ipeiros ** * * ** ** ** * 6 EL22 Ionia Nisia ** * * * ** ** * ** 6 EL23 Dytiki Ellada ** * ** ** ** ** ** ** 8 EL24 Sterea Ellada ** ** ** ** ** ** ** 6 EL25 Peloponnisos ** * ** ** ** ** * 6 EL30 Attiki * ** ** ** 4 EL41 Voreio Aigaio ** * * ** ** ** * * 7 EL42 Notio Aigaio * * * * * ** ** * 6 EL43 Kriti ** * ** ** ** * * ** 6 ES11 Galicia * * * ** ** * * 5 ES12 Principado de Asturias * * * ** * * 4 ES13 Cantabria * * * * ** ** * 5 ES21 País Vasco * ** 3 ES22 Comunidad Foral de Navarra * ** 3 ES23 La Rioja * * ** ** * 5 ES24 Aragón * ** ** * 4 ES30 Comunidad de Madrid ** ** * 4 ES41 Castilla y León * * * ** ** * 5 57

62 Policy Department for Structural and Cohesion Policies ES42 Castilla-la Mancha ** * ** ** ** ** ** ** 8 ES43 Extremadura ** ** ** ** ** * * ** 6 ES51 Cataluña * ** ** * 4 ES52 Comunidad Valenciana * * * ** ** ** ** 6 ES53 Illes Balears * * * ** ** * 5 ES61 Andalucía ** * ** ** ** ** * ** 7 ES62 Región de Murcia * * ** ** ** ** ** ** 8 ES63 ES64 Ciudad Autónoma de Ceuta (ES) Ciudad Autónoma de Melilla (ES) * * ** ** ** ** ** 6 ** ** ** ** ** ** ** 6 ES70 Canarias (ES) * * ** ** ** ** ** 6 FI19 Länsi-Suomi 0 FI1B Helsinki-Uusimaa 0 FI1C Etelä-Suomi 0 FI1D Pohjois- ja Itä-Suomi ** 2 FI20 Åland * 1 FR10 Île de France ** 2 FR21 Champagne-Ardenne ** ** * 4 FR22 Picardie * * ** * 3 FR23 Haute-Normandie ** * 2 FR24 Centre (FR) * 1 FR25 Basse-Normandie * ** ** * 4 FR26 Bourgogne * * * 2 FR30 Nord-Pas-de-Calais * * * * ** * 4 FR41 Lorraine * * * ** * 3 FR42 Alsace ** 2 FR43 Franche-Comté * ** * 2 FR51 Pays de la Loire 0 FR52 Bretagne * 0 FR53 Poitou-Charentes * * 1 58

63 Indicators in Cohesion Policy FR61 Aquitaine 0 FR62 Midi-Pyrénées 0 FR63 Limousin * 1 FR71 Rhône-Alpes 0 FR72 Auvergne * 0 FR81 Languedoc-Roussillon * * 0 FR82 Provence-Alpes-Côte d'azur 1 FR83 Corse * 2 FR91 Guadeloupe ** * ** 6 FR92 Martinique * * ** 6 FR93 Guyane ** ** ** ** 9 FR94 La Réunion ** * ** ** 6 FR95 Mayotte ** ** ** ** 10 HR01 Jadranska Hrvatska ** ** ** ** 8 HR02 Kontinentalna Hrvatska ** ** ** * ** 9 HU10 Közép-Magyarország ** ** * 4 HU21 Közép-Dunántúl ** ** ** ** 8 HU22 Nyugat-Dunántúl ** ** ** ** 6 HU23 Dél-Dunántúl ** ** ** * ** 8 HU31 Észak-Magyarország ** ** ** ** ** 9 HU32 Észak-Alföld ** ** ** ** ** ** ** 9 HU33 Dél-Alföld ** ** ** ** ** ** ** 8 IE01 Border, Midland and Western * * ** * ** ** ** * 2 IE02 Southern and Eastern ** ** ** * 2 ITC1 Piemonte ** ** ** ** ** 3 ITC2 Valle d'aosta/vallée d'aoste ** ** ** ** * 4 ITC3 Liguria ** ** * ** ** 3 ITC4 Lombardia ** ** 2 ITD1 Bolzano/Bozen ** ** ** ** 2 59

64 Policy Department for Structural and Cohesion Policies ITD2 Trento ** ** ** 1 ITD3 Veneto ** ** ** ** 2 ITD4 Friuli-Venezia Giulia ** ** * ** ** 1 ITD5 Emilia-Romagna ** ** * ** ** 1 ITE1 Toscana ** ** ** ** 3 ITE2 Umbria * * ** 2 ITE3 Marche * ** 2 ITE4 Lazio * * ** 2 ITF1 Abruzzo * * ** ** 3 ITF2 Molise ** * * * ** * 5 ITF3 Campania ** * ** ** ** 6 ITF4 Puglia ** * * ** ** 6 ITF5 Basilicata ** * ** * * 7 ITF6 Calabria ** * ** ** ** 6 ITG1 Sicilia ** * * * ** 6 ITG2 Sardegna ** * * * * 5 LT00 Lituanie ** ** ** * ** ** 6 LU00 Luxembourg 0 LV00 Latvija ** ** ** ** ** * ** ** 7 MT00 Malte * * * ** ** ** * 6 NL11 Groningen 0 NL12 Friesland (NL) ** ** 4 NL13 Drenthe ** ** 4 NL21 Overijssel ** 2 NL22 Gelderland 0 NL23 Flevoland * ** 3 NL31 Utrecht 0 NL32 Noord-Holland 0 NL33 Zuid-Holland 0 NL34 Zeeland ** ** 4 60

65 Indicators in Cohesion Policy NL41 Noord-Brabant 0 NL42 Limburg (NL) * 1 PL11 Lódzkie ** ** ** ** ** * ** ** 7 PL12 Mazowieckie * ** ** ** * 4 PL21 Malopolskie ** ** ** ** ** * * 5 PL22 Slaskie ** ** ** ** ** * ** ** 7 PL31 Lubelskie ** ** ** ** ** * ** ** 7 PL32 Podkarpackie ** ** ** ** ** * ** ** 7 PL33 Swietokrzyskie ** ** ** ** ** * ** ** 7 PL34 Podlaskie ** ** ** ** ** * ** ** 7 PL41 Wielkopolskie ** ** ** ** ** * * ** 6 PL42 Zachodniopomorskie ** ** ** ** ** ** * ** 7 PL43 Lubuskie ** ** ** ** ** ** ** ** 8 PL51 Dolnoslaskie * ** ** ** ** * * ** 6 PL52 Opolskie ** ** ** ** ** ** ** ** 8 PL61 Kujawsko-Pomorskie ** ** ** ** ** ** ** ** 8 PL62 Warminsko-Mazurskie ** ** ** ** ** ** ** ** 8 PL63 Pomorskie ** ** ** ** ** * * ** 6 PT11 Norte ** ** ** ** ** ** ** ** 8 PT15 Algarve * * * ** * ** ** ** ** 9 PT16 Centro (PT) ** * ** ** ** ** ** ** 8 PT17 Metropolitana de Lisboa * ** ** * 4 PT18 Alentejo ** * ** ** ** ** ** ** 8 PT20 Açores (PT) ** * * ** ** ** ** ** ** 10 PT30 Madeira (PT) ** * * ** ** ** ** ** ** 10 RO11 Nord-Vest ** ** ** ** ** ** * ** 7 RO12 Centru ** ** ** ** ** ** ** ** ** 10 RO21 Nord-Est ** ** ** ** ** ** ** ** ** 10 RO22 Sud-Est ** ** ** ** ** ** ** ** ** 10 RO31 Sud - Muntenia ** ** ** ** ** ** ** ** ** 10 61

66 Policy Department for Structural and Cohesion Policies RO32 Bucuresti - Ilfov * * ** ** * 4 RO41 Sud-Vest Oltenia ** ** ** ** ** ** ** ** 8 RO42 Vest ** ** ** ** ** ** * ** 7 SE11 Stockholm 0 SE12 Östra Mellansverige 0 SE21 Småland med öarna 0 SE22 Sydsverige 0 SE23 Västsverige 0 SE31 Norra Mellansverige 0 SE32 Mellersta Norrland 0 SE33 Övre Norrland 0 SI01 Vzhodna Slovenija ** ** ** ** * ** * 5 SI02 Zahodna Slovenija * ** 2 SK01 Bratislavský kraj ** 2 SK02 Západné Slovensko ** ** ** ** ** ** * ** 7 SK03 Stredné Slovensko ** ** ** ** ** ** ** ** 8 SK04 Východné Slovensko ** ** ** ** ** * ** ** ** 9 UKC1 UKC2 Tees Valley and Durham Northumberland and Tyne and Wear ** ** * * 1 * * 0 UKD1 Cumbria * * 2 UKD2 Cheshire 0 UKD3 Greater Manchester 0 UKD4 Lancashire * * ** * * 3 UKD5 Merseyside * * 0 UKE1 East Yorkshire and Northern Lincolnshire * * ** * 2 UKE2 North Yorkshire 0 UKE3 South Yorkshire * * 0 UKE4 West Yorkshire * 0 UKF1 Derbyshire and Nottinghamshire * * 0 62

67 Indicators in Cohesion Policy UKF2 Leicestershire, Rutland and Northamptonshire 0 UKF3 Lincolnshire * * ** 2 UKG1 UKG2 Herefordshire, Worcestershire and Warwickshire Shropshire and Staffordshire * * * 1 0 UKG3 West Midlands * * * * 1 UKH1 East Anglia 0 UKH2 Bedfordshire and Hertfordshire 0 UKH3 Essex * * 0 UKI3 Inner London - West 0 UKI4 Inner London - East 0 UKI5 Outer London - East and North East * * 0 UKI6 Outer London - South 0 UKI7 UKJ1 UKJ2 UKJ3 Outer London - West and North West Berkshire, Buckinghamshire and Oxfordshire Surrey, East and West Sussex Hampshire and Isle of Wight UKJ4 Kent * * 1 UKK1 Gloucestershire, Wiltshire and Bristol/Bath 0 UKK2 Dorset and Somerset * 0 UKK3 Cornwall and Isles of Scilly * ** 2 UKK4 Devon * 0 UKL1 West Wales and The Valleys ** * * 1 UKL2 East Wales 0 UKM2 Eastern Scotland 0 63

68 Policy Department for Structural and Cohesion Policies UKM3 UKM5 South Western Scotland North Eastern Scotland * ** 2 0 UKM6 Highlands and Islands 0 UKN0 Northern Ireland (UK) * * ** * 2 UKJ4 Kent * * 1 UKK1 Gloucestershire, Wiltshire and Bristol/Bath 0 UKK2 Dorset and Somerset * 0 UKK3 Cornwall and Isles of Scilly * ** 2 UKK4 Devon * 0 UKL1 West Wales and The Valleys ** * * 1 UKL2 East Wales 0 UKM2 Eastern Scotland 0 UKM3 UKM5 South Western Scotland North Eastern Scotland * ** 2 0 UKM6 Highlands and Islands 0 UKN0 Northern Ireland (UK) * * ** * 2 64

69 Indicators in Cohesion Policy ANNEX 2. CHANGES IN THE ELIGIBLE AND TRANSITION REGIONS, PASSING FROM THE GDP CRITERION TO THE PUBLIC AND PRIVATE FCE CRITERION The same percentage of the EU population is considered in each regime as presently for the GDP per capita (26.9 % for the full eligibility regime, 44.9 % for the eligibility and transition regimes). (a) Moreover, including all the Greek regions and Cyprus in the category of full eligibility should be considered, due to the very low level of investment in these countries. 0 : not eligible T : transition regime E : full eligibility GDP > Public and private FCE Full eligible regions with both criteria Transition regions with both criteria Regions losing partially or fully the access to aids Regions gaining partially or fully the access to aids AT BE Burgenland (T>0) Hainaut (T>0) Liège (T>0) Luxembourg (T>0) Namur (T>0) BG Severozapaden Severen tsentralen Severoiztochen Yugoiztochen Yugozapaden Yuzhen tsentralen CY Cyprus (a) CZ Strední Cechy Jihozápad Praha (0>T) Jihovýchod (T>E) Severozápad Severovýchod Strední Morava Moravskoslezsko 65

70 Policy Department for Structural and Cohesion Policies DE Mecklenburg- Vorpommern Brandenburg (T>0) Leipzig (0>T) Sachsen-Anhalt Lüneburg (T>0) Thüringen Chemnitz (T>0) DK Sjælland (T>0) EE Estonia EL Anatoliki Makedonia, Thraki Notio Aigaio Kentriki Makedonia (E>T) (a) Attiki (0>T) (a) Dytiki Ellada Dytiki Makedonia (E>T) (a) Sterea Ellada Thessalia (A>T) (a) Peloponnisos Ipeiros (E>T) (a) Kriti Ionia Nisia (E>T) (a) Voreio Aigaio (E>T) (a) ES Castilla-la Mancha Galicia La Rioja (0>T) Extremadura Principado de Asturias Aragón (0>T) Andalucía Cantabria Cataluña (0>T) Melilla Castilla y León Illes Balears (0>T) Comunidad Valenciana Región de Murcia (T>E) Ceuta (T>E) Canarias (T>E) FI FR Guyane Mayotte Nord - Pas-de- Calais Lorraine Languedoc- Roussillon Martinique Picardie (T>0) Basse-Normandie (T>0) Bourgogne (T>0) Franche-Comté (T>0) Bretagne (T>0) Poitou-Charentes (T>0) 66

71 Indicators in Cohesion Policy HR Jadranska Hrvatska Kontinentalna Hrvatska Limousin (T>0) Auvergne (T>0) Corse (T>0) Guadeloupe (E>T) La Réunion (E>T) 67

72 Policy Department for Structural and Cohesion Policies 68

73 Indicators in Cohesion Policy ANNEX 3. CHANGES IN THE ELIGIBLE AND TRANSITION REGIONS, PASSING FROM THE GDP CRITERION TO THE SYNTHETIC INDEX The same percentage of the EU population is considered in each regime as presently for he GDP per capita (26.9 % for the full eligibility regime, 44.9 % for the eligibility and transition regimes). (a) Moreover, including all the Greek regions and Cyprus in the category of full eligibility should be considered, due to the very low level of investment in these countries. 0 : not eligible T : transition regime E : full eligibility GDP > Global synthetic index Full eligible regions with both criteria Transition regions with both criteria Regions losing partially or fully the access to aids Regions gaining partially or fully the access to aids AT Burgenland (T>0) BE Hainaut Liège (T>0) Luxembourg (T>0) Namur (T>0) BG Severozapaden Severen tsentralen Severoiztochen Yugoiztochen Yugozapaden Yuzhen tsentralen CY Cyprus (T>0) (a) CZ Severozápad Jihovýchod Strední Cechy (E>0) Strední Morava Moravskoslezsko Jihozápad (E>T) Severovýchod (E>T) DE Brandenburg Lüneburg (T>0) Mecklenburg- Vorpommern Chemnitz Sachsen-Anhalt 69

74 Policy Department for Structural and Cohesion Policies Thüringen DK Sjælland (T>0) EE Estonia EL Anatoliki Makedonia, Thraki Kentriki Makedonia Dytiki Makedonia Notio Aigaio Ipeiros (E>T) (a) Attiki (0>E) Peloponnisos (E>T) (a) Voreio Aigaio (E>T) (a) Thessalia Ionia Nisia Dytiki Ellada Sterea Ellada Kriti ES Castilla-la Mancha Galicia La Rioja (0>T) Extremadura Principado de Asturias Aragón (0>T) Andalucía Cantabria Comunidad de Madrid (0>T) Melilla Castilla y León Cataluña (0>T) Illes Balears (0>T) Comunidad Valenciana (T>E) Región de Murcia (T>E) Ceuta (T>E) Canarias (T>E) FI FR Guadeloupe Picardie Bourgogne (T>0) Champagne-Ardenne (0>T) Guyane Basse-Normandie Bretagne (T>0) Haute-Normandie (0>T) La Réunion Nord - Pas-de- Calais Poitou-Charentes (T>0) Martinique (T>E) Mayotte Lorraine Limousin (T>0) Franche-Comté Auvergne (T>0) 70

75 Indicators in Cohesion Policy HR Jadranska Hrvatska Kontinentalna Hrvatska Languedoc- Roussillon (T>0) Corse (T>0) HU Közép-Dunántúl Közép-Magyarország (0>T) Nyugat-Dunántúl Dél-Dunántúl Észak-Magyarország Észak-Alföld Dél-Alföld IE Border, Midland and Western Southern and Eastern (0>T) IT Campania Umbria (T>0) Valle d'aosta (0>T) Puglia Calabria Sicilia Abruzzo (T>0) Molise (E>T) Basilicata (E>T) Sardegna (E>T) LT Lithuania LU LV Latvia MT Malte NL PL Lódzkie Malopolskie (E>T) Mazowieckie (0>T) Slaskie Dolnoslaskie (T>E) Lubelskie Podkarpackie Swietokrzyskie Podlaskie Wielkopolskie Zachodniopomorskie 71

76 Policy Department for Structural and Cohesion Policies Lubuskie Opolskie Kujawsko-Pomorskie Warminsko- Mazurskie Pomorskie PT Norte Centro (PT) Lisboa (0>T) Algarve (T>E) Alentejo Açores (PT) Madeira (PT) RO Nord-Vest Bucuresti Ilfov (0>T) Centru Nord-Est Sud-Est Sud - Muntenia Sud-Vest Oltenia Vest SE SI Vzhodna Slovenija (E>T) SK Západné Slovensko Stredné Slovensko Východné Slovensko UK Lancashire Northumb.,Tyne & Wear (T>0) West Yorkshire (0>T) East Yorkshire & N. Lincoln. West Midlands Northern Ireland (UK) Merseyside (T>0) South Yorkshire (T>0) Derbyshire & Notting. (T>0) Lincolnshire (T>0) Shrop. & Stafford. (T>0) 72

77 Indicators in Cohesion Policy Essex (T>0) Outer London E & NE (T>0) Kent (T>0) Dorset and Somerset (T>0) Cornwall & Scilly (T>0) Devon (T>0) S.W. Scotland (T>0) Tees Valley & Durham (E>T) W. Wales & The Valleys (E>0) HU Közép-Dunántúl Közép-Magyarország (0>T) Nyugat-Dunántúl Dél-Dunántúl Észak-Magyarország Észak-Alföld Dél-Alföld IE Border, Midland and Western Southern and Eastern (0>T) IT Campania Umbria (T>0) Valle d'aosta (0>T) Puglia Calabria Sicilia Abruzzo (T>0) Molise (E>T) Basilicata (E>T) Sardegna (E>T) LT Lithuania LU LV Latvia MT Malte NL PL Lódzkie Malopolskie (E>T) Mazowieckie (0>T) Slaskie Dolnoslaskie (T>E) 73

78 Policy Department for Structural and Cohesion Policies Lubelskie Podkarpackie Swietokrzyskie Podlaskie Wielkopolskie Zachodniopomorskie Lubuskie Opolskie Kujawsko-Pomorskie Warminsko- Mazurskie Pomorskie PT Norte Centro (PT) Lisboa (0>T) Algarve (T>E) Alentejo Açores (PT) Madeira (PT) RO Nord-Vest Bucuresti Ilfov (0>T) Centru Nord-Est Sud-Est Sud - Muntenia Sud-Vest Oltenia Vest SE SI Vzhodna Slovenija (E>T) SK Západné Slovensko Stredné Slovensko Východné Slovensko UK Lancashire Northumb.,Tyne & Wear (T>0) West Yorkshire (0>T) 74

79 Indicators in Cohesion Policy East Yorkshire & N. Lincoln. West Midlands Northern Ireland (UK) Merseyside (T>0) South Yorkshire (T>0) Derbyshire & Notting. (T>0) Lincolnshire (T>0) Shrop. & Stafford. (T>0) Essex (T>0) Outer London E & NE (T>0) Kent (T>0) Dorset and Somerset (T>0) Cornwall & Scilly (T>0) Devon (T>0) S.W. Scotland (T>0) Tees Valley & Durham (E>T) W. Wales & The Valleys (E>0) 75

80 Policy Department for Structural and Cohesion Policies 76

81

82