International Database on Employment & Adaptable Labour (IDEAL)

Amsterdam, February 2007 Mind the gap International Database on Employment & Adaptable Labour (IDEAL) Ernest Berkhout Christian Dustmann Piet Emmer

SEO Economic Research carries out independent applied economic research on behalf of the government and the private sector. The research of SEO contributes importantly to the decision-making processes of its clients. SEO Economic Research is connected with the Universiteit van Amsterdam, which provides the organization with invaluable insight into the newest scientific methods. Operating on a not-for-profit basis, SEO continually invests in the intellectual capital of its staff by encouraging active career planning, publication of scientific work, and participation in scientific networks and in international conferences. SEO-report nr. 968 ISBN 978 90 6733 380 1 Copyright 2007 SEO Economic Research, Amsterdam. All rights reserved. Reproduction of material from this publication is permitted without prior permission, providing that the source is acknowledged clearly and accurately.

Preface It is a great pleasure to introduce the second edition of the International Database on Employment and Adaptable Labour (IDEAL). This publication, published jointly by Randstad Holding and SEO, once again provides a comprehensive overview of international employment trends and more specifically the market for flexible, and notably agency, work. This edition also contains a chapter which focuses entirely on the role of labour migration. Besides an analysis of future labour market shortages & the current role of migration, you will find contributions of two professors, specifically assigned by the University of Amsterdam to this topic. Prof. Piet Emmer presents an analysis of labour migration in the historical context, prof. Christian Dustmann analyses the role of migration on wages and employment in an economic context. The publication is unique in that it is the only publication of its kind which, as well as analyzing the existing data on employment trends and flexible work, seeks to integrate these data with international data on agency work. Bearing in mind that agency work is far and away the fastest growing form of flexible employment, the lack of comparable international data on the subject constitutes a serious flaw in the existing academic analysis of international labour markets. Even today, now that the contribution of agency work to well functioning labour markets is universally recognized most notably by the International Labour Organization, policy mistakes are still being made because data fails to distinguish between temporary (fixed term contracts) and part time work and agency work. Randstad s core values since its foundation in 1960 have been: the simultaneous promotion of all interests; to know, serve and trust; striving for perfection; In order to implement these core values, Randstad needs a thorough knowledge of the markets in which it operates, the labour markets. As a complement to its existing knowledge of these markets at a local level, Randstad therefore strongly welcomes this publication which results from our sponsorship of scientific research at the University of Amsterdam. SEO Economic Research is delighted that, as a result of this sponsorship, it is being given the opportunity to undertake research into an area where so little international analysis so far exists. Together, we hope that this publication will help society better understand the labour market of tomorrow. Ben Noteboom, CEO Randstad Holding Jules Theeuwes, General Director SEO Economic Research

Contents Executive Summary... I Introduction... 1 1 The Lisbon targets: participation and unemployment... 3 1.1 Participation...3 1.2 Participation and gender...6 1.3 Participation and age...8 1.4 Participation projections EU, 2005-2050...11 1.5 Unemployment...15 2 Flexible labour relations & agency work... 23 2.1 Part-time work...23 2.2 Temporary work...31 2.3 Agency work...40 3 Immigration, labour market & the enlarged European Union... 47 3.1 Demand & supply in the future labour market...47 3.2 Migration as a solution to labour market frictions?...57 3.3 History of immigration...72 3.4 Migration, wages and employment...81 Index of figures & tables... 89 Literature... 91 Appendix A - Employment gap calculations... 93 Appendix B - Employment gaps by sector & occupation without migration... 113 Appendix C - Net migration since 1960... 114

I Executive Summary The aim of IDEAL is twofold: To present a useful overview of the current international labour market situation, with a specific focus on the Lisbon targets on participation and unemployment (Chapter 1) and on several types of flexible labour relations, including agency work (Chapter 2). To address special issues that influence international labour markets and labour relations. This publication deals with the issues of ageing, the resulting future employment gap and labour migration (Chapter 3), and includes two special contributions from Prof. Emmer and Prof. Dustmann. 1. Lisbon targets The participation targets for 2010 might be met by the EU-25 on average, but not by every country individually. The United States, Canada and Japan are all above the EU average. To reach an overall participation rate of 70%, economic development will have to be positive and participation will have to grow faster than it has during the past five years, especially in the larger countries Poland, Italy, France and Germany. The unemployment target of 4% is unrealistic and will not be met. In most EU countries, female employment has become a part of labour market culture, but not in Eastern Europe. In Italy, female employment is even lower than it is in Eastern Europe. Still, although some countries are clearly lagging behind, meeting the Lisbon target of 60% female employment seems realistic on average. Participation of the elderly (age 55-64) did not rise during the 1990s, but between 2000 and 2005 it increased by 7 percentage points. Therefore, it is not unrealistic to suggest that the Lisbon target of 50% employment for this age group might be reached in 2010 in most countries, with the exception of Poland, Italy, Austria, Luxembourg, Belgium, Hungary and probably France. In 2006, unemployment is 8.2% on average in the EU-25 (7.4% in the EU-15) and thereby at its lowest in more than two decades; it will not reach 4% in the coming years. Japan, the United States and Canada have much lower unemployment than Europe does, although it is still more than 4%. Eurostat population projections for 2050 show that, in the baseline scenario, the EU-25 workingage population (age 15-64) will decline from 305 million persons to 255 million persons. On the other hand, the number of elderly (age 65+) will increase from 80 million to 140 million; this will cause the future supply of labour to be much lower while the demand for publicly financed goods like health care and pensions will be much higher. This stresses the importance of high participation rates, but reduces the suitability of the Lisbon targets: will a 70% employment rate be enough in the long run?

II EXECUTIVE SUMMARY 2. Flexible labour relations Part-time work is becoming more popular, mostly among youths (including those who are still in education). Nearly half of all Dutch employees work fewer than 35 hours a week, as do 30-40% of employees in the rest of Europe except the Southern and Eastern countries. Part-time jobs are typically a female phenomenon (except in Denmark and the Netherlands) and are more popular among the low-skilled. Most part-time workers do not want a full-time job. Differences in national part-time rates are therefore employee-driven (which makes it difficult to stimulate these people to work full-time hours). Temporary work is more popular among young workers, the low-skilled and immigrants; the worker s sex does not come into the equation. Spain, Poland and Portugal have the largest number of temporary workers. Temporary agency work is most prominent in the UK (4.3% of total employment), followed by France, the Benelux countries and the US (around 2%). It is a leading indicator for GDP growth, is subject to special regulation in most countries, is becoming more popular and often functions as a stepping stone towards a permanent job. 3.1 Future frictions: the European employment gap Although age distributions differ between countries, the ageing challenge is apparent everywhere. In the next forty years, the average age will increase all over Europe. What will be the impact of this on the future labour markets? Which countries will suffer most? These are the questions discussed in Chapter 3.1 using the Eurostat population projection mentioned above. Ageing decreases the labour supply: as the effect of ageing alone in 2050 the EU-25 employment level will have decreased by 31 million persons if the employment rate would remain at its current 63.3%. This is called the employment gap caused by ageing. It is calculated by applying the age structure of 2050 to the total population size of 2005 and comparing that with the actual 2005 population. To keep the employment level constant at its current 195 million persons, the employment rate will have to rise above the Lisbon target, to 75.5%. 1 But even if the EU succeeds in closing the employment gap, the balance between the working-age population and the elderly population will be different in 2050; the same number of employed persons will have to finance a larger number of elderly people. Increasing demand (for health care, leisure goods and services, etc.) might also add to the employment gap but its exact impact is much harder to assess for long-term horizons. Knowing that demand will rise, our calculations can therefore be seen as minimum employment gaps. The implicit assumption of constant labour productivity is discussed below. Besides by ageing, the employment gap will widen even further because of declining (and often negative) population growth. This negative growth is projected to be small on average (-1% for the EU-25) but quite large in Germany and Eastern & Southern Europe (even as large as -13% in the Czech Republic). The separate impacts of both the ageing effect and the population decline effect are presented in the graph below. The black circle indicates the present employment rate, the red dot the employment rate needed in 2050 to offset the ageing effect. For example, in the Czech Republic the ageing of the population causes the employment rate to rise from 65% to 1 In some countries the effect is much worse, in others it is less. Employment gap calculations for each country separately are included in Appendix A.

III 81%. Though different in magnitude, the ageing effect is clearly present in all countries. But on top of that, total population will decline (in most countries), causing a further widening of the employment gap. To bridge the population decline effect as well, employment rates will have to reach the level indicated by the end of the bars. For the Czech Republic, this means a further increase of the employment rate to an unrealistic 94%. On the other hand, in Luxemburg the population effect is positive because of the estimated population growth of 43% (mainly due to large net migration). Therefore, employment levels can be held constant without the need to increase the employment rate. How to keep employment levels constant: ageing effect and population growth effect in 2050 Poland Hungary Italy Belgium Spain France EU-25 Luxembourg Czech Republic Germany Ireland Portugal Austria United Kingdom Sweden Netherlands Denmark 40 50 60 70 80 90 Employment rate 2005 Pure ageing effect incl. baseline growth Source: SEO calculations based on Eurostat (EUROPOP 2004, lfsi_act_a, lfsq_ergan). There are several options if employment gaps subsist: Further increase employment participation, in terms of persons or of hours worked per person (the former seems more plausible than the latter). Decrease output (GDP). Probably the most unfavourable option, though the necessary outcome if everything else fails. Productivity increase. An efficient and realistic option in the manufacturing and the agricultural sector. However, such is harder to achieve in labour-intensive service sectors; here, constant productivity seems a realistic assumption. Sectoral shifts of labour. As labour is scarce it might move from the lower-paying sectors to the higher-paying sectors. This of course will not bridge the overall employment gap, but merely shift the problem from one sector to another. Extra labour migration. If arranged so as to be beneficial to both the host country and the immigrant, it is an economically sensible option. However, it can be useful only if immigrants supply the skills the labour markets need. Temporary migration should also be considered.

IV EXECUTIVE SUMMARY 3.2 Immigration as a solution? As discussed above, some countries might mitigate the ageing and population effects somewhat through net migration, or even offset the ageing effect completely (Luxembourg, Ireland & Sweden). To get an idea of the impact of immigration: if no net migration is allowed, the overall EU-25 employment gap in 2050 will be 55 million persons rather than 32 million persons. To keep employment at the current level, the employment rate will have to reach over 88% on average. In some countries the importance of migration is even higher (see the no migration scenarios in Appendix A). Contrary to many beliefs, immigrants are over-represented among both the low-skilled and the high-skilled; two different types of immigrant appear to exist. On average they have higher unemployment and temporary work rates. Currently, net migration is generally less than 1% of total population per year, but with large differences between individual EU countries. In Austria, Italy, Ireland & Spain migration has doubled recently, while in the Netherlands net migration has decreased dramatically and has been negative since 2004. Historical overviews of net migration by country are included in Appendix C. Summarizing, we conclude that disallowing migration will definitely not help; immigrant labour has to be made beneficial to both the immigrant and the host country. Another conclusion is that migration has many different faces. It is not only low-skilled labour that immigrates and highskilled labour that emigrates, as some people want us to believe. In most countries at least two different migrant groups can be distinguished, that is, one that is relatively high skilled and one that is relatively low skilled. Even when looking only at immigrants from non-western countries, they turn out to be on average high skilled in the one country and on average low skilled in the other; a lot seems to depend on institutional differences between countries, such as current and historical immigration policies, cultural and colonial links, and the structure and wage structure of the economy. 3.3 History of labour migration Recent history shows us that if immigrant skills do not match the local labour market needs, the immigration debate might become characterised by negative emotions and stereotypes. From a historical perspective, Prof. Emmer shows in a special contribution (see page 72) that this is a much older phenomenon than most of us think, and that we can learn from history to prevent future failures. The central question in this contribution is: why do Europeans have such a low opinion of migrants in spite of the fact that in the past they themselves have been the prime beneficiaries of migration? Why do Europeans frown upon Third World migrants who leave their home in order to improve their living conditions as well as their earning capacity? During the time of mass emigration from Europe, migrants were seen as dynamic and entrepreneurial. Now, politicians and public opinion alike seem to deplore the relatively modest yearly immigration surplus, which in the past would have been a reason for pride. Even Italy, Spain and Portugal in spite of their recent emigration history, seemed to have forgotten the benefits of migration now that labour immigrants from the South in turn face them.

V The explanation for this dramatic change in the European attitude towards immigration can be found in the substantial payments of government subsidies and social security benefits to the immigrants. The ready access to these benefits has blurred the difference between talented and untalented immigrants. In the past, only talented immigrants (like talented entrepreneurs) had reason to stay as they had succeeded in finding a niche in the labour market of the host country. In the 19 th century for example about one fifth of the European migrants to the New World came back. During the past fifty years there certainly were large numbers of talented migrants coming to Europe, but there also arrived a disproportionate number with no such talents, mainly from Turkey, Morocco, and some of the former colonies. Their failure on the European labour markets did not force them to return as they received much higher social security benefits than they would have earned with a full day s work at home. The European social security systems were not designed to cope with such immigrants. These systems were established in the 1950s and 1960s when unemployment was low, most of the population consisted of two-headed families, the average number of children per woman was at least two or more, and most people worked near the place where they lived and were willing to work long hours. It is obvious that all of these conditions have changed and that these changes would have taken place even without immigrants. However, immigrants tend to be kept under close scrutiny and have their actions exaggerated, which is why they are blamed for many of the present shortcomings of the various European welfare systems. The traumatic immigration experience of the past decades has made it virtually impossible to discuss the resumption of labour migration to Europe. Yet, it seems unlikely that Europe can build a viable future without immigration. The declining population will create tensions on the labour market, and the ageing population will need more labour-intensive services (such as health care and nursing) that are difficult to outsource or to obtain elsewhere. In order to put labour immigration on a new footing, Europe should distinguish clearly between its unselfish obligation to help refugees and asylum seekers, and its selfish need to import labour. In addition to reshaping our policies towards asylum seekers, Europe is also in need of a new policy towards economic immigrants. During the 19 th century, the transatlantic movements were fuelled by wage differences of a factor 1.5 to 2. Today, European and North American wages are more than 20 times that of the wage rate in some of the Third World countries. Most of the EU member states need legal economic migrants and several of these countries are in the process of adopting new immigration laws. They have to decide how to meet the challenge of combining immigration without jeopardizing Europe s highly developed welfare state. 3.4 Migration, wages & employment From an economic perspective, Prof. Dustmann explains in his special contribution (see page 81) the possible ways in which immigration might affect local wages and employment. Looking theoretically at only the short-term effects, the wage and possibly employment effects of migration will occur only if the composition of immigrants in terms of their skill mix is different from that of the native workforce. There will be losers and winners: migration will have distributional effects, with an overall positive surplus for the native workforce. But in the long run, even if immigrant skills differ, the economy can react by adjusting the output mix of goods it

VI EXECUTIVE SUMMARY produces; there may well be short-run effects on wages and employment, but in the longer run they may return to their pre-migration level. Prof. Dustmann s contribution emphasizes the multiple ways an economy can adjust to immigration. Public debate about the possible consequences of immigration for employment and wages often seems to be led by the perception that a fixed number of jobs exist in the economy and that immigration will lead to more competition for these jobs. This is untrue: even in our simplest theoretical model, nothing prevents the economy from increasing output, thus absorbing new workers by creating new jobs. Regarding the empirical evidence, economists have made tremendous progress in addressing serious problems of empirical analysis when assessing the effects of immigration on wages and employment, but still have no conclusive picture of the precise effects, or of the exact mechanisms by which immigration impacts wages and employment. Unfortunately, the public debate is often fed by politically motivated statements from individuals or pressure groups about the employment and wage effects of immigration; they often lack serious argumentation or sources, and are sometimes based on single studies or conclusions of such studies that are taken out of context. Such statements are not helpful and hinder a well-informed debate. Any serious economist working in this area ought to admit that the research agenda on assessing the wage and employment effects of immigration is still very much ongoing.

1 Introduction In most developed countries the use of flexible forms of labour has increased during the last one or two decades. But at the same time very large differences exist between countries in the scale and forms of flexible labour. The EU enlargement with Eastern & Central-European countries increases heterogeneity even more. Differences in regulations & restrictions, the workforce and the economic situation are considered to be the main causes for these differences. The Netherlands are a special case when looking at flexible labour. Not only is flexible labour frequently used in the Netherlands (especially part-time work), but also the role of temporary work agencies is much larger than in most other countries. For Randstad Holding, a major player in the Dutch and European market for temporary work, it is important to learn more about the use of flexible forms of labour, the driving forces behind it and differences between countries in labour market institutions and the relationship with flexible labour. Although much statistical information exists by amongst others OECD, Eurostat, CIETT, national Statistical Offices detailed internationally comparable statistics (both time series and cross section data) on flex-work are scarce. A further problem with these statistics is that definitions differ strongly between countries and that they are adjusted frequently. A third problem is that the distinguished countries, the frequency and the topics covered vary between sources. For Randstad Holding these were reasons to start a project in September 2000 with the aim of collecting labour market data in general and data on flexible forms of labour in particular. The project resulted in the International Database of Employment and Adaptable Labour (IDEAL). This database is created by SEO Economic Research in co-operation with and commissioned by Randstad Holding. The aim of IDEAL is to bring together a large number of comparable international statistics on employment, flexible labour and agency work. In May 2004 this resulted in the first publication of the Randstad Jobs Report, in which an international outlook was presented based mainly on data recent to the year 2002. In this report we present an update of this database with new countries and more recent figures. The last chapter focuses on the special theme of labour migration. Where available statistics are presented for the following countries: Austria (A) Poland (POL) Belgium (B) Portugal (PT) Czech Republic (CZ) Spain (ES) Denmark (DK) Sweden (SE) France (F) Switzerland (CH) Germany (G) United Kingdom (UK) Hungary (H) Canada (Can) Ireland (IRL) Japan (J) Italy (I) Turkey (T) Luxembourg (L) United States (US) Netherlands (NL)

2 INTRODUCTION On some topics we will also be able to compare figures of the above countries with averages for the European Union (EU-15 or EU-25). Reliable figures on China and India appear to be virtually non-existent yet, but where available we will add them as well. The main focus of IDEAL is on international comparability between statistics. Outline of the report In Chapter 1 we will elaborate on the topics of employment participation of the population and unemployment. For both these items the Lisbon 2000 European Council has set target values to be reached by the member states before the year 2010. Present values and target values will be compared. In Chapter 2 we describe other important developments and trends on the European labour markets. In subsequent paragraphs we focus on: part-time work (paragraph 2.1), temporary work (2.2) and agency work (2.3). In Chapter 3 we have a special focus on the theme of labour migration. As this topic is too often discussed in an emotional debate rather than in a debate based on facts, our aim is to lift the immigration debate to a scientific level. We first look at the facts & figures: what are the effects of ageing and EU enlargement and how is what is the expected impact on future frictions in the European labour markets? What would be the role of migration as a possible solution? We look at labour migration in both a historic and an economic perspective. Professor Piet Emmer contributes on the history of migration (3.3), professor Christian Dustmann will analyze the effect of immigration on the receiving countries labour market (3.4). Our concluding remarks can be input for the special seminar on Labour Migration, to be organized in the spring of 2007.

3 1 The Lisbon targets: participation and unemployment In Lisbon in the spring of 2000, the Heads of State and Government decided to endorse a new strategic social goal to be attained by 2010: to become the most competitive and dynamic knowledge based economy in the world capable of sustainable economic growth with more and better jobs and greater social cohesion. Regaining full employment is at the core of the new strategy. The European Council set out -as the key indicator of this success- that by 2010, the employment rate should rise to 70% for the European Union as a whole and to 60% for women. Among the elderly (aged 55-64) participation should be at least 50%. At the same time the Union average unemployment rate should be reduced to 4% or less. In the first Randstad Jobs Report (Berkhout & van Leeuwen, 2004) we showed that the participation targets might be met by the EU-15 on average in 2010, but that such was not likely to happen for the unemployment targets. In this report we take a second look, taking into account three more years of data and some new member states as well. 1.1 Participation In nearly all publications on employment and participation only percentages are used, because they are easily comparable between countries. Of course, absolute numbers of employed persons are not very useful when we want to compare the degree of participation between countries, but they still are useful to give an impression of the size of the labour market in the country. In IDEAL, we have recorded the absolute values of the total population and the population with age 15-64 (the so called potential labour force ), both for men and women. These quantities are presented in Table 1.1. With the EU expansion in 2004 the total population increased with over 70 million people, half of whom lives in Poland. The working age population increased by 51 million people, to the recent number of 305 million (2005 average). From other sources we obtained estimates of the population sizes of China and India. Apart from each of these two countries being far more populated than the EU & US together, they also grow slightly faster than most European countries.

4 CHAPTER 1 Table 1.1 Total population, by gender, 2005 POPULATION POPULATION 15-64 TOTAL MALE FEMALE TOTAL MALE FEMALE Austria 8,109 3,939 4,170 5,516 2,745 2,770 Belgium 10,477 5,127 5,350 6,876 3,459 3,417 Czech Republic 10,229 4,987 5,242 7,270 3,646 3,624 Denmark 5,396 2,671 2,725 3,566 1,799 1,767 France 59,224 28,748 30,476 38,683 19,100 19,584 Germany 81,529 39,938 41,590 54,764 27,559 27,206 Hungary 9,932 4,698 5,234 6,815 3,328 3,486 Ireland 4,149 2,067 2,081 2,831 1,425 1,406 Italy 58,077 28,192 29,885 38,588 19,248 19,340 Luxembourg 450 223 227 304 153 151 Netherlands 16,107 7,992 8,116 10,943 5,519 5,424 Poland 37,527 18,104 19,422 26,211 12,986 13,225 Portugal 10,563 5,115 5,448 7,115 3,516 3,599 Spain 43,141 21,268 21,873 29,755 15,019 14,736 Sweden 9,039 4,479 4,559 5,896 2,993 2,903 Switzerland* 7,364 3,602 3,763 5,035 2,523 2,512 United Kingdom 58,421 28,476 29,945 38,529 18,983 19,546 Canada* 31,946 15,817 16,130 21,882 10,947 10,935 Japan* 127,687 62,295 65,392 84,610 42,460 42,150 Turkey 71,606 35,743 35,863 46,610 23,296 23,314 United States 296,410 146,000 150,411 191,014 94,207 96,807 EU-15 380,563 186,010 194,553 253,973 126,816 127,157 EU-25 453,831 221,215 232,615 305,076 152,081 152,996 China(2006 est.) 1,313,974 676,463 637,510 928,399 482,439 445,960 India(2006 est.) 1,095,352 564,613 530,739 704,058 363,876 340,182 * Total population figures are from 2004 Source: Eurostat (lfsi_act_a, 2006) OECD LFS (2006); US Census Bureau (2006); CIA Factbook (2006). Given the important role of the central party committee on present Chinese development politics, labour market statistics from the National Bureau of Statistics are mainly collected through the Ministry of Labor and Social Security and therefore often doubtful. They mostly illustrate that the measures for macroeconomic regulation adopted by the central government were prompt, correct and effective. Implementation of these measures helped to prevent local problems from expanding into overall problems, to avoid serious inflation and big ups and downs of the economy, and to ensure the stable and rapid development of the economy. 2 2 CEQ Bulletin, 2005, vol 1, page 7. Official text issued by the National Bureau of Statistics.

IDEAL 2006 5 Table 1.2 Employment-population ratio s for persons aged 15-64 1990Q2 2000Q2 2001Q2 2002Q2 2003Q2 2004Q2 2005Q2 2005Q4 Austria 68.5 68.4 68.8 69.1 67.7 68.4 68.8 Belgium 54.4 60.9 59.7 59.7 59.3 60.5 61.0 61.2 Czech Republic 64.9 65.0 65.5 64.9 64.1 64.7 65.2 Denmark 75.4 76.4 75.9 76.4 75.1 76.0 75.5 76.8 EU-15 61.6 63.2 63.9 64.2 64.3 64.5 65.0 EU-25 62.2 62.7 62.8 62.9 63.0 63.6 France 60.8 61.7 62.7 62.9 63.3 63.2 63.4 62.8 Germany 66.4 65.3 65.7 65.4 64.9 64.3 65.3 65.9 Hungary 55.9 56.1 56.2 57.0 56.6 56.8 57.1 Ireland 51.9 64.5 65.2 65.1 65.1 65.5 67.1 67.8 Italy 53.9 53.4 54.5 55.4 56.1 57.7 57.8 57.8 Luxembourg 59.2 62.7 63.0 63.6 62.2 62.5 63.6 Netherlands 61.1 72.9 74.1 74.5 73.8 73.1 73.2 73.6 Poland 55.1 53.7 51.7 51.4 51.4 52.2 53.7 Portugal 65.5 68.2 68.9 69.2 68.2 68.0 67.6 67.4 Spain 49.1 56.1 57.7 58.6 59.7 60.9 63.2 64.0 Sweden 71.1 74.4 74.0 73.6 72.4 72.6 71.5 Switzerland 78.3 79.1 78.9 77.9 77.4 77.2 United Kingdom 71.1 71.0 71.3 71.2 71.3 71.5 71.5 71.6 Canada 70.3 70.9 70.8 71.4 72.2 72.6 72.5 Japan 68.6 68.9 68.8 68.2 68.4 68.7 69.3 Turkey 54.5 48.8 47.8 46.9 45.8 46.1 45.9 United States 72.2 71.0 70.8 71.2 72.0 72.5 71.5 Total participation rates (including unemployed): China 84.2 83.8 83.4 82.9 82.4 82.3 82.0 India 63.2 61.5 61.4 61.4 61.3 61.0 60.9 Source: Eurostat LFS (lfsq_ergan,2006); OECD LFS (2006); ILO (Laborsta-EAPEP) Although the potential workforce in China and India is very large (resp. 930 million and 700 million) the actual employment is less than in most European countries, due to higher unemployment. Unfortunately Table 1.2 shows only the total participation rates for these two countries, which include the unemployed population as well. We have no reliable data on unemployment in the past years, but estimates for 2006 are 8.9% in India and somewhere between 9-20% in China. 3 That would translate into employment rates of 56% for India and 66-75% for China. The EU-15 average is 65 percent, with the EU-25 showing slightly lower employment at 63.6 percent. This is obviously below the Lisbon target of 70 percent, but since 2000 a steady rise can be seen. Figure 1.1 reflects this graphically; the dark arrow shows the development of the employment rate during the ten years of the last decade of the 20 th century, the light arrow shows the development in the first five years of the current century. Some countries met the Lisbon target already in 1990 (but these include US, Canada & Switzerland), the Netherlands made a huge catching-up during the nineties, and some other countries are getting close now. Ireland & Spain 3 CIA Factbook, 2005 estimates.

6 CHAPTER 1 showed high employment growth since 1990 until now, in Belgium the economic trough seemed to have had more impact because the employment rate stayed around 60 percent during the last five years. Italy is still behind but finally growing; Poland and Hungary are different cases. They joined the EU only in 2004, and since then employment shows a positive development. To reach the 70 percent target in 2010 economic development has to be positive and participation has to grow faster than in the last five years, mainly in the larger countries Poland, Italy, France and Germany. Figure 1.1 Development of the overall employment rate, 1990-2000-2004 40 50 60 70 80 1990->2000 2000->2005 Turkey Poland Hungary Italy Belgium Spain France Luxembourg Czech Republic EU-15 Germany Ireland Portugal Austria Japan United Kingdom United States Canada Sweden Netherlands Denmark Switzerland Source: Eurostat LFS (lfsq_ergan,2006); OECD LFS (2006). 1.2 Participation and gender A second target for 2010 is to reach an employment level of 60% for females, on average. In 2002 this figure was 55,5 percent for the EU-15. In the next three years this has risen to 57,4 percent (56,2 if EU-25 is considered). In other words: getting close, but not there yet. Canada, US and Switzerland all have female employment rates well above 60 percent, even Japan is above the EU-average. Within Europe there are structural differences in the female labour markets. In most countries female employment is 10-15 percentage points less than male employment, but in Turkey, Italy, Spain, Ireland & Japan this so-called gender gap is much larger. This can be seen in Figure 1.2 where employment rates are plotted for both men and women, and countries are sorted by the male employment rate. Female employment is still highest in Denmark at 70,8 % (and rising), which is higher than male employment in Italy, France, Turkey, Belgium, Hungary & Poland. If the Lisbon target will be met or not mainly depends on the countries mentioned above, where the gender gap leaves enough opportunity for new people to enter the labour market. Especially if the business cycle conditions are favourable (as is the case now in 2006) efforts for increasing

IDEAL 2006 7 participation will be most effective. But there might be cultural differences to be overthrown, maybe more than economic differences. What is most important might be illustrated by Figure 1.3 which shows female employment developments since 1990. Figure 1.2 The gender gap in employment rates, 2005 Poland Hungary Belgium Turkey France Italy Germany EU-15 Czech Republic Luxembourg Portugal Sweden Spain Austria Ireland Canada United Kingdom United States Netherlands Denmark Japan Switzerland 24 46 45 58 51 63 54 68 68 58 69 70 59 71 57 73 56 73 54 73 62 73 71 75 51 75 62 75 58 76 68 77 66 77 66 78 66 80 71 80 58 80 70 84 0 20 40 60 80 100 Women Men Source: Eurostat LFS (lfsq_ergan,2006); OECD LFS (2006). Figure 1.3 Development of the female employment rate, 1990-2000-2004 Turkey Italy Poland Hungary Spain Luxembourg Belgium Czech Republic EU-15 France Ireland Japan Germany Austria Portugal United States United Kingdom Netherlands Canada Switzerland Sweden Denmark 20 30 40 50 60 70 1990->2000 2000->2005 Source: Eurostat LFS (lfsq_ergan,2006); OECD LFS (2006).

8 CHAPTER 1 Clearly, the Netherlands and Ireland showed a shift in the cultural pattern of women staying at home to women going to work: employment rates rose by 25 percent. The same is true, to a lesser extent, for countries like Spain, Luxembourg, Belgium and Portugal. So in most EUcountries female employment became part of labour market culture, or was part of it already. But in Eastern Europe developments are not so positive in the last years, either they are small or they are negative. It is not clear yet if this pattern will be changed now they are EU-members, latest quarterly figures suggest that such might be happening. In Italy female employment is even lower than in Eastern Europe; it is a growing phenomenon but recent growth is not spectacular, given the fact that not even half of the Italian women are in some form of employment. Still, although there are some countries clearly lagging behind, meeting the Lisbon target of 60 percent female employment seems realistic. 1.3 Participation and age The third target to be reached in 2010 is an employment rate of 50% for the elderly. In this paragraph we take a closer look at the different employment situations of younger and older people. Even though there are no specific targets specified, in some countries low youth employment (because of high unemployment) is more of a problem than low employment among the elderly (because of early retirement). Figure 1.4 Employment rates of the elderly (age 55-64), 2005 Poland Turkey Italy Austria Luxembourg Belgium Hungary France Spain EU-15 Czech Republic Germany Netherlands Portugal Ireland Canada United Kingdom Denmark United States Japan Switzerland Sweden 27 31 31 32 32 32 33 39 43 44 44 45 46 51 52 55 57 60 61 64 65 69 0 20 40 60 80 100 Source: Eurostat LFS (lfsq_ergan,2006); OECD LFS (2006). Figure 1.4 shows that the average employment rate for people aged 55-64 in 2005 is 44 percent. Spain, Germany & the Netherlands are close to this average; Portugal, the Anglo-Saxon and Scandinavian countries are well above. In Eastern & Central Europe and Italy older people work

IDEAL 2006 9 clearly less than 50 percent, for these individual countries such a level to be reached in five years seems unrealistic. Figure 1.5 Development of the employment rate of the elderly (age 55-64), 1990-2000-2005 20 30 40 50 60 70 1990->2000 2000->2005 Poland Turkey Italy Austria Luxembourg Belgium Hungary France Spain EU-15 Czech Republic Germany Netherlands Portugal Ireland Canada United Kingdom Denmark United States Japan Switzerland Sweden Source: Eurostat LFS (lfsq_ergan,2006); OECD LFS (2006). However, from Figure 1.5 it becomes clear that elderly participation is developing in the right direction in most countries (except Poland and Turkey), at least since 2000. In the EU-15 participation of 55-64 did not rise during the last decade of the past century, but from 2000-2005 it increased by 7 percentage points. Therefore, it is not unrealistic to suggest that the Lisbon target of 50 percent employment for this age group might be reached in 2010. A different story has to be told when looking at youth employment and unemployment. Figure 1.6 shows clearly that in all countries except Spain, Austria and Sweden youth employment has declined since 2000, and that in most countries (except Ireland, Denmark & the Netherlands) this process already started a decade before. But among the age group 15-24 employment participation does not tell the whole story. A lot of people are still in education, and if ceteris paribus participation in education rises then employment might fall for a positive reason. On the other hand, youth employment in the Netherlands has risen particularly sharp because the number of people who are both in education and in employment has risen. Students started to work in relatively small jobs to finance their education as education reimbursements were cut. To get a complete picture we also have to look at the youth unemployment rates which are shown in Figure 1.7 (the Lisbon target on overall unemployment will be discussed later, in paragraph 1.3).

10 CHAPTER 1 Figure 1.6 Development of the youth employment rate (age 15-24), 1990-2000-2005 20 30 40 50 60 70 1990->2000 2000->2005 Hungary Poland Luxembourg Italy Belgium Czech Republic France Turkey Portugal Spain EU-15 Sweden Japan Germany Ireland Austria United Kingdom United States Canada Switzerland Denmark Netherlands Source: Eurostat LFS (lfsq_ergan,2006); OECD LFS (2006). Figure 1.7 Development of the youth unemployment rate (age 15-24), 1990-2000-2005 0 10 20 30 40 1990->2000 2000->2005 Netherlands Denmark Ireland Japan Austria United States Canada United Kingdom Luxembourg Germany Portugal EU-15 Czech Republic Turkey Hungary Spain Belgium France Sweden Italy Poland Source: Eurostat LFS (une_rt_a, 2006); OECD LFS (2006). Above we can see that in some countries youth unemployment changed quite a lot during the last 15 years while in other countries the development was less dramatic. A close linear relation between employment rate development & unemployment rate development is not clear;

IDEAL 2006 11 structural changes in the combination of education & work are also significant. For instance, in the Netherlands in the nineteen nineties a huge increase in employment did bring down unemployment only marginally. The increase in employment here came from people who were not in the labour force before. On the other hand, in Ireland during the period 1990-2000 we see a huge increase in employment accompanied by a huge drop in unemployment. Apparently some negative correlation between youth employment rates and youth unemployment rates is clearly present, but other factors (combining education & work) are important as well. 1.4 Participation projections EU, 2005-2050 The previous paragraphs showed the current participation situation in several Western countries. Some countries will easily meet the Lisbon targets of 2010, others will probably not. But after 2010 population developments do not stop; in fact, large labour market effects of an ageing EUpopulation are expected. Eurostat has recently published an outlook (Eurostat, 2006) in which they present seven possible population growth scenarios for the period 2005-2050. 4 In this paragraph we will look at the three basic projections & the no-migration variant, and focus on the possible consequences of ageing based on these scenarios. The four scenarios are: Baseline projection Low growth scenario High growth scenario No migration scenario All scenarios are calculated based on the 2004 population in each of the 25 EU-countries, plus for every country separately assumptions regarding the fertility rate, life expectancy and net migration. In the baseline scenario s these assumptions are based on current values and expected to remain constant over the projection period. In the low growth scenario all assumptions are below the present values and in the high growth scenario all factors are assumed to be higher than presently. The no migration scenario is a special variant of the baseline scenario, useful to show the impact of (net) migration. Table 1.3 summarizes these assumptions. Figure 1.8 & Figure 1.9 show the resulting projections for the EU-25, differentiated by age class. Table 1.3 Projection assumptions Total fertility rate Life expectancy Net migration Baseline base base base Low growth low low low High growth high high high No migration base base zero Source: Eurostat (Statistics in focus 3/2006). 4 The projection database is officially called EUROPOP2004 : EUROstat POpulation Projections 2004-based and originally consisted of only four scenarios; the three basic projections baseline, low growth & high growth plus the no-migration variant. In this publication we mainly use the baseline projection, because this projection is build on the most realistic assumptions given current knowledge. We show the high growth- & low growth-variants to indicate what will happen if all assumptions work together in population growth or population decrease. We use the no-migration variant to indicate the implicit effects of migration in the baseline projection. We do not use the other three variants as they do not provide additional information, being merely different combinations of the basic assumptions base, low & high.

12 CHAPTER 1 Figure 1.8 Baseline projections of the EU-25 youth, working age and elderly population Age <15 90 Age 15-64 320 80 300 70 60 million 280 260 million 50 240 2005 2010 2015 2020 2025 2030 2035 2040 2045 2050 High growth/low growth Base projection No-migration 40 2005 2010 2015 2020 2025 2030 2035 2040 2045 2050 High growth/low growth Base projection No-migration 220 Age 65+ 160 All ages 550 140 500 120 million 450 million 100 400 2005 2010 2015 2020 2025 2030 2035 2040 2045 2050 High growth/low growth Base projection No-migration 80 2005 2010 2015 2020 2025 2030 2035 2040 2045 2050 High growth/low growth Base projection No-migration 350 Source: Eurostat (EUROPOP2004). Figure 1.9 Projection of the future EU-25 population by agegroup, in four scenario s 500 500 400 400 300 200 Base projection 300 200 No-migration 100 100 <15 15-64 65+ 0 <15 15-64 65+ 0 500 500 400 400 300 200 Low growth 300 200 High growth 100 100 <15 15-64 65+ 0 <15 15-64 65+ 0 Source: Eurostat (EUROPOP2004). The above tables clearly show that the number of elderly will grow while the number of workingage persons will decline, even in the high growth scenario! In the baseline scenario the working age group will decline from over 300 million to just over 250 million in the next 45 years. Figure 1.10 compares the detailed age distribution of 2005 with that of 2050, in all four scenarios. The height of the bars represents the relative size of the age group in the total population, the blue

IDEAL 2006 13 bars for the 2005 population and the green bars for the 2050 population. All the blue bars together sum to 100%, as do all the green bars together. In most scenarios, the bulk which is in the middle in 2005 will shift to the right. In the high growth scenario the 2050 age distribution will become the most equal regarding all people under the age of 80. The share of older elderly (meaning 80+) however will triple in this scenario, and more than double in the other scenario s. Figure 1.10 Age distribution of the future EU-25 labour force, in four scenario s Base projection No migration 0-4 10-14 20-24 30-34 40-44 50-54 60-64 70-74 80+ 2005 base Baseline 2050 0-4 10-14 20-24 30-34 40-44 50-54 60-64 70-74 80+ 2005 nomig No migration 2050 Low growth High growth 0-4 10-14 20-24 30-34 40-44 50-54 60-64 70-74 80+ 2005 low Low growth 2050 0-4 10-14 20-24 30-34 40-44 50-54 60-64 70-74 80+ 2005 high High growth 2050 Source: SEO calculations based on Eurostat (EUROPOP 2004). Consequences for employment The ageing of our society implies that the elderly population will grow faster than the working age population. Thereby, ageing leads to lower employment because nearly all persons stop working at the age of 65, most people even earlier. That means that with a constant employment rate (the share of people in the working age who are employed), employment in persons will decline because the share of people in the working age ( potential employment ) will decline. This will happen in each of the four scenario s, as is shown in Figure 1.11. The working age rate will decline from 67% to somewhere between 55-57%. One of the consequences of this decline working age rate is the extra burden on the welfare state. Not only will taxes and social security contribution be collected by less people, there will also be a greater demand for health care etc. Of course this problem will be greater in some countries than in others, depending on the institutional design of the welfare state and on the national share of elderly. Figure 1.12 shows the present grey rate for European and other countries. It becomes clear that in Japan, Italy, Germany, Sweden and Belgium ageing will be a problem much bigger (and sooner!) than in the non-western or Anglo-Saxon countries.

14 CHAPTER 1 Figure 1.11 Working age as % of total population, EU-25 in four scenario s, 2005-2050 67 64 61 percent 58 2005 2010 2015 2020 2025 2030 2035 2040 2045 2050 Base projection No-migration Low growth High growth 55 Source: SEO calculations based on Eurostat (EUROPOP 2004). Figure 1.12 Grey rate per country in 2005 (population 65+ as % of population 15-64) Turkey India China Ireland Canada United States Poland Czech Republic Netherlands Luxembourg Hungary Denmark Switzerland Austria EU-25 Spain EU-15 France Portugal Belgium Sweden Germany Italy Japan 6 8 11 16 18 18 19 20 21 21 22 23 23 23 24 24 24 25 25 26 26 28 29 30 0 5 10 15 20 25 30 Source: Eurostat (pjan, 2006), CIA Factbook (2006), World Development Indicators (2005). The trends mentioned in this paragraph have serious impacts on the future labour market. In Chapter 3 we will explore the effects of ageing and population growth in more detail.

IDEAL 2006 15 1.5 Unemployment The unemployment target of the Lisbon agenda looks even more ambitious than the employment rate target. In the year the target was set out, only four countries had an unemployment rate below or just at the target rate of 4%, while the EU average was at 7.7%. In 2004 unemployment was even higher, and only recently the quarterly figures show signs of recovery thereby finally bringing down unemployment. When looking at unemployment rates over the last two decades, three countries draw immediate attention: Ireland, Spain and Poland (see Table 1.4 and Figure 1.13). Where most unemployment rates are within the range of 4-12%, in these countries the unemployment peak is much higher. Table 1.4 Harmonized unemployment rates 1983 1990 1995 2000 2001 2002 2003 2004 2005 2006Q2 Austria 3.9 3.6 3.6 4.2 4.3 4.8 5.2 4.9 Belgium 10.7 6.6 9.7 6.9 6.6 7.5 8.2 8.4 8.4 8.6 Czech Republic 8.7 8.0 7.3 7.8 8.3 7.9 7.4 Denmark 8.4 7.2 6.7 4.3 4.5 4.6 5.4 5.5 4.8 3.9 EU15 10.0 8.2 10.0 7.7 7.3 7.6 8.0 8.1 7.9 7.4 EU25 8.6 8.4 8.8 9.0 9.1 8.7 8.2 France 7.7 8.5 11.1 9.1 8.4 8.9 9.5 9.6 9.5 8.8 Germany 8.0 7.2 7.4 8.2 9.0 9.5 9.5 8.2 Hungary 6.4 5.7 5.8 5.9 6.1 7.2 7.3 Ireland 13.9 13.4 12.3 4.3 4.0 4.5 4.7 4.5 4.3 4.4 Italy 7.4 8.9 11.2 10.1 9.1 8.6 8.4 8.0 7.7 Luxembourg 1.7 2.9 2.3 2.1 2.8 3.7 5.1 4.5 4.7 Netherlands 9.2 5.8 6.6 2.8 2.2 2.8 3.7 4.6 4.7 3.9 Poland 13.3 16.1 18.2 19.9 19.6 19.0 17.7 16.3 Portugal 8.2 4.8 7.3 4.0 4.0 5.0 6.3 6.7 7.6 7.4 Sweden 1.7 8.8 5.6 4.9 4.9 5.6 6.3 7.8 Spain 14.1 13.0 18.4 11.1 10.3 11.1 11.1 10.7 9.2 8.3 United Kingdom 10.8 6.9 8.5 5.4 5.0 5.1 4.9 4.7 4.7 Canada 12.0 8.2 9.6 6.8 7.2 7.7 7.6 7.2 6.8 Japan 2.6 2.1 3.1 4.7 5.0 5.4 5.3 4.7 4.4 4.1 Turkey 7.7 8.0 7.6 6.5 8.3 10.3 10.5 10.3 10.3 United States 9.6 5.5 5.6 4.0 4.8 5.8 6.0 5.5 5.1 4.6 China* 2.3 2.5 2.9 3.1 3.6 4.0 Source: Eurostat LFS (une_rt_a, 2006); OECD LFS (2006), *World Development Indicators (2005). But there are marked differences between these countries as well. In Spain unemployment was by far the highest of all countries in our study during most of the nineties, despite a spectacular fall from 20% unemployment in 1994 to 11% in 2000. Irelands unemployment rate was still high in 1993 (nearly 16%) but from that moment on the Irish employment miracle brought down unemployment a spectacular 12 percentage points in the following 7 years. In less than a decade Ireland made the transition from highly problematic to a nearly full employment country. Poland showed above average unemployment rates during the whole period, but a decreasing trend radically turned into rising unemployment trend after 1998. It reached its peak of nearly