Pensions and other age-related expenditures in Europe Is ageing too expensive?

1 Pensions and other age-related expenditures in Europe Is ageing too expensive? Bo Magnusson bo.magnusson@his.se Bernd-Joachim Schuller bernd-joachim.schuller@his.se University of Skövde Box 408 S-541 28 Skövde Sweden Abstract Because of low fertility rates and a rising life expectancy at birth and at the age of 65, the European population is becoming older. According to the projections, presented by the European Commission, the consequence is a stagnating or even falling total population. Regarding the working age population (aged 15 64) a quite large fall is projected. Because the older part of the population (above 65 years) will grow both in relative and absolute terms, the dependency rate defined as people 65 years and older relative to the working age or active population will rise strongly. This means that fewer active people have to support economically more old persons. In this paper we will based on the projections of the European Commission for population and Gross Domestic Product investigate how pensions and other age related expenditures will develop up to 2050. Here we are looking closer on the public expenditures. We will even present calculations how GDP per head of population and age related expenditures based on the projections from the Commission probably will develop. Age-related expenditures are both shown as a percentage of GDP and in absolute terms. Furthermore, we will even discuss some developments, which can level out the negative population effect of the ageing process on the active population, like net immigration, rising labour force participation rates, and falling unemployment rates. In the projections of the European Commission, no policy changes, besides of the already implemented ones are included. Keywords: Population, fertility, life expectancy, migration, age structure, labour supply, employment, economic growth, age-related expenditures

2 Pensions and other age related expenditures in Europe Is ageing too expensive? 1. Introduction According to projections, presented by the European Union (EU), size and age structure of the European population in the future will change considerably. This is a development, which is not unique to Europe. Three factors are commonly mentioned as explanations of size and age structure of populations 1 : Fertility, life expectancy and net migration. While fertility is below the size which leads to a stable population, life expectancy both at birth and at the age of 65 will continue to rise. These factors will lead to at least - a stagnating population, which on average will become older and older. Though net migration - as projected - will be much larger in the future than in the past, it will improve the situation, but not level out the negative effects on population and age structure. Furthermore migrants usually are younger and have a higher proportion of people in working ages than the average population 2. With other words, both population size and age structure will be influenced by net immigration. The mentioned demographic changes will have some important economic and social consequences 3, of which the following can be mentioned. When population stagnates and becomes older, this will ceteris paribus have impacts on production and growth, because the labour supply and employment will be negatively influenced. Labour, capital and goods and service markets will be influenced. Probably the relative prices of labour and capital will change. The structure of consumption, savings and investment probably will change because of population ageing. Fewer active people have to economically support a larger number of older people, at least in relative terms. This means that the old age dependency rate will increase considerably. Age-related public expenditures will increase, both in absolute terms and relative to GDP. The following items are included in the age-related expenditures 4 : Pensions, health care, long-term care, unemployment benefits and education. In this paper we will investigate the following aspects of population and economic development. How has the population developed in the past? Here we have a look at total population and net migration from 1960 to 2000. Furthermore, we present the population and migration projections 5 of the EU from 2004 to 2050. We are presenting projections for the total population, the population between 0 14 years (the young age group), the one between 15 and 64 years (the working age population) and the population 65 and older, the labour force or labour supply and the total employment. 1 See e.g. Lutz (2006), Carone & Costello (2006), Zamac (2007), Jonsson & Pettersson (2002) 2 E.g. Rybakovskii (2006) 3 E.g. Oliveira Martins et al. (2005) 4 European Economy, Special Report 1/2006 5 European Economy, Special Report no 1/2006 and No 4/2005: The projections are not forecasts and assume no policy changes during the projection period.

3 Independent of pension systems 6, maintenance of the retired people is based on the production of goods and services. Therefore we even will have a look at the projections of GDP. Furthermore, we will present the age-related expenditures relative to GDP and in absolute terms. This paper is mainly based on the projections of population and production, presented by the Commission of the European Union 7. As mentioned above, these projections are not seen as forecast. Furthermore, no policy changes are assumed, except the ones already decided and implemented. The projections start with the year 2004, sometimes even 2003 and have 2050 as the final year. Among the projections, presented by the EU, we are choosing the ones for the EU15 8, the EU10 9, the EU25, which means EU15 plus EU10, and the ones for Sweden. Bulgaria and Romania, who joined the EU in 2007, are not included. The paper is organized in the following way. After the introduction, in section 2 we present figures for total population and migration for the period from 1960 to 2000 and the projections for 2004 to 2050. In section 3, fertility and the age structure are described. Section 4 is about working age population, labour supply and employment. In section 5 we show the age-related public expenditures. Section 6 discusses GDP, economic growth and growth theory. In section 7 the EU-projections of the growth rates for GDP, GDP pc and GDP per employed person are shown. Section 8 presents figures for GDP, GDP pc and GDP per employed person in constant prices. Even calculations for age-related expenditures are presented here. Section 9 consists of a summary and the conclusions and in section 10 the references are mentioned. 2. Total population and migration In this section, we are presenting figures for population and migration between 1960 and 2050. In table 1, we show the total population. The figures for 1960 to 2000 are from population statistics, while the ones for 2004 to 2050 are projections. 6 i.e. whether they are funded systems or pay-as-you-go systems, or with other words whether pension payments are depending on savings or payments from the active population to the old one 7 European Economy, Special Report No 4/2005 and Special Report No 1/2006 8 EU15 : Belgium, Denmark, Germany, Greece, Spain, France, Ireland, Italy, Luxembourg, Netherlands, Austria, Portugal, Finland, Sweden, United Kingdom 9 Cyprus, Czech Republic, Estonia, Hungary, Lithuania, Latvia, Malta, Poland, Slovak Republic, Slovenia. These countries joined the EU in 2004 and are sometimes called the new member countries.

4 Table 1: Total population in the EU15, the EU25, the EU10 and Sweden, 1960 2050, in millions Population, millions EU15 EU25 EU10 Sverige 1960 299,7 361,5 61,7 7,5 1990 349,5 424,5 75,0 8,6 2000 378,1 452,9 74,8 8,9 2004 382,7 456,8 74,1 9,0 2010 390,8 464,2 73,4 9,2 2020 398,4 470,2 71,8 9,6 2030 400,6 471,2 70,6 9,9 2040 397,5 465,9 68,4 10,0 2050 388,3 453,8 65,5 10,2 Observe that the figures for the EU15 and the EU25 between 1960 and 1990 include only Western Germany. After 1990 even former GDR is included. The population of the former GDR in 1990 was approximately 16 million with stagnating tendencies since 1960. 1960-2000: Figures from population statistics, 2004 2050: projections Sources: European Economy, Special Report no 4/2005, European Economy, 6/2004 As table 1 illustrates, between 1960 and 2000 the population in the EU and in Sweden was rising. Yet for the EU10, we can observe, that population since 1990 was falling. Regarding the period 2004 to 2050, the projections show the following. The population of Sweden will continue to rise, while the one for the EU10 is projected to continue to fall. Population in the EU15 and the EU25, which is dominated by the EU15, population will rise until 2030. After that year, population will fall. In the next table (table 2), we present figures for net immigration. Table 2: Net immigration, EU15, EU25, EU10 and Sweden, 1960 2050, 1000 persons Net immigration, 1000 persons 1960-2000 2004-2050 EU15 14539 37123 EU25 12802 39710 EU10-1737 2586 Sweden 679 1069 The figures for net immigration are cumulated over the entire period of 1960 to 2000 and 2004 to 2050, respectively Source: European Economy, Special Report No 4/2005 As table 2 illustrates, the net immigration in the EU15 and in Sweden was positive between 1960 and 2000, while for the EU10 net emigration could be observed. The projections are showing that while the EU15 and Sweden will continue to have net immigration as between 1960 and 2000, even the new member countries will experience net immigration. Let us finish this section by comparing net immigration with total population (table 3).

5 Table 3: Net immigration, related to total population, 2000 and 2050, EU15, EU25, EU10 and Sweden, per cent Net immigration/total population, % 2000 2050 EU15 3,84% 9,57% EU25 2,83% 8,75% EU10 (net emigration) 3,95% Sweden 7,63% 10,48% Sources: see table 2; authors calculations As table 3 illustrates, net immigration will have a larger importance for the size of total population in the future, compared with the past. We can observe, that Sweden in the past had relatively seen a much larger immigration than the EU. While the net immigration to the EU increases quite substantially, even for Sweden we can observe rising net immigration. To summarize this section, the following can be mentioned. While population between 2004 and 2050 in the EU15 will continue to rise and the one of the EU25 falls slightly, the projections show that the population in the EU is at maximum in 2030 and falling after that year. In the EU10, a negative population development already can be observed since 1990. Sweden is somewhat unique: the projections illustrate, that the population will increase even until 2050. A possible explanation is related to net immigration. According to the projections, net immigration to Sweden is in relative terms higher than in the EU. Furthermore, if migrants are younger and more active on the labour market than the average, then this can improve the Swedish population situation furthermore 10. 3. Fertility and age structure After having presented figures for population and migration, in this section we will have a look at the other two explanations of demographic changes, i.e. fertility and life expectancy. In the next table (table 4), we have a look at fertility, which is defined as children per woman. Table 4: Projections of fertility, children per woman, EU15, EU25, EU10 and Sweden, 2004-2005 Fertility: children per woman EU15 EU25 EU10 Sweden 2004 1,53 1,58 1,23 1,74 2010 1,57 1,52 1,24 1,84 2020 1,60 1,57 1,40 1,85 2030 1,60 1,59 1,56 1,85 2040 1,60 1,60 1,58 1,85 2050 1,61 1,60 1,58 1,85 A fertility rate of 2,1 is assumed to be necessary for a stable population Source: European Economy, Special Report no 4/2005 10 See footnote 2

6 Table 4 shows that the fertility rates will be too low to assure a stable population. Fertility rates are - especially in the new member countries - projected to rise, but from very low levels. The somewhat rising fertility rate in Sweden is above the European average. The next two tables (table 5 and 6) are illustrating life expectancy at birth and at 65, separated for males and females. Table 5: Life expectancy at birth for males and females, EU15, EU25, EU10 and Sweden, 2004 and 2050, years Life expectancy at birth, years Males Females 2004 2050 change 2004 2050 change EU15 76,4 82,3 5,9 81,5 87,4 5,2 EU25 75,4 81,8 6,4 82,2 86,9 5,4 EU10 70,1 78,7 8,6 78,2 84,1 5,9 Sweden 78,1 83,3 5,2 82,4 86,5 4,1 Source: European Economy, Special Report no 4/2005 Table 5 illustrates the following. Life expectancy at birth will continue to rise. Though life expectancy increases more for males than for females, females can expect a longer life than males even in 2050. Seen from a Swedish perspective, it is remarkable that while Swedish males have a higher life expectancy than the average European male both in 2004 and in 2050, Swedish females are in 2004 above the EU average, but in 2050 below. The increase of life expectance is especially large between 2004 and 2050 for males in the new member countries. Even life expectancy for females in the EU10 is projected to rise somewhat more than in the total EU. The next table illustrates life expectancy at 65 for males and females (table 6). Table 6: Life expectancy at 65 for males and females, EU15, EU25, EU10 and Sweden, 2004 and 2050 Life expectancy at 65, years Males Females 2004 2050 change 2004 2050 change EU15 16,3 20,8 4,4 19,9 24,3 4,4 EU25 15,9 20,5 4,6 19,5 23,9 4,4 EU10 13,5 18,5 5,0 17,2 21,2 4,1 Sweden 16,7 20,0 3,3 19,8 22,8 3,0 Source: European Economy, Special Report no 4/2005 As table 6 shows, life expectancy at 65 is projected to rise a little bit more for males than for females, who yet even in 2050 can expect to have a higher life expectancy than males. The next step in the description of the population situation is to have a look at the age structure. We will present projections for the young population (0 14 years), the working age population (15 64 years) and the elderly population (65+ years).

7 Table 7: The total population, divided in three age groups (0 14, 15 64, 65+) in the EU15, the EU25, the EU10 and Sweden, 2004 and 2050, millions millions 0-14 15-64 65+ of which 80+ 2004 2050 2004 2050 2004 2050 2004 2050 EU15 62,4 52,7 255,1 221,3 65,2 114,2 16,3 44,2 EU25 74,8 60,4 306,8 259,1 75,3 133,3 18,2 49,9 EU10 12,4 8,6 51,7 37,8 10,1 19,1 1,9 5,7 Sweden 1,6 1,7 5,8 6,0 1,5 2,5 0,5 0,9 Source: European Economy, Special Report no 4/2005 Table 7 illustrates that in the EU15, EU25 and EU10 the young population and the working age population are falling, while population 65 and older is growing. We can even observe that the very old population (80+) is growing considerably. Even here the Swedish population development is somewhat different from the European one. Both the young population and the working age population are increasing somewhat. Of course even in Sweden the old population is growing considerably, but relatively not as much as in the EU totally. In the next table we will present figures for the different age groups in relation to the total population (table 8). Table 8: Age groups (0 14, 15 64, 65+) as percentage of total population, in the EU15, the EU25, the EU10 and Sweden, 2004 and 2050 Per cent 0-14 15-64 65+ 2004 2050 2004 2050 2004 2050 EU15 16,3 13,6 66,7 57,0 17,0 29,4 EU25 16,4 13,3 67,2 57,1 16,5 29,4 EU10 16,7 13,1 69,8 57,7 13,6 29,2 Sweden 17,8 16,7 64,4 58,8 16,7 24,5 Source: European Economy, Special Report no 4/2005; authors s calculations As table 8 shows, in relative terms the young population (0 14) will fall in the EU15, EU25 and EU10, respectively with approximately 3 percentage units, while the Swedish young population only falls with 1 percentage unit. Both in 2004 and in 2050 the proportion of young to total population is larger in Sweden than in the EU on average. Regarding the working age population (15 64), we can observe in the EU a fall of this part of the population with approximately 10 percentage units. Even here the situation in Sweden is somewhat different: while Sweden in 2004 has relatively seen a somewhat smaller working age population, because of the relatively small Swedish fall, in 2050 the situation in Swedish is somewhat better than in the EU totally. Finally, regarding the population 65+, in 2004 Sweden is on the EU15 and EU25 level, while in 2050 in the EU15 and EU25 this age group comprises close to 30 percent of total population. In Sweden this relation is approximately 25 percent.

8 Having shown figures for the working population and the old population, the next step is to present old age dependency rates, defined as the population 65 and older (65+), relative to the working age population (15 64) (table 9). This is only one of several possible old age dependency rates. We even could have chosen a dependency rate which relates 65+ to employed persons (aged 15 64). The second way of defining dependency rates gives a somewhat larger result. Table 9: Old age dependency rate, defined as 65+ to 15 64 population, EU15, the EU25, the EU10 and Sweden, 2004 and 2050 2004 2050 EU15 25,5 51,6 EU25 24,5 51,4 EU10 19,6 50,4 Sweden 26,4 40,9 Source: European Economy, Special Report no 4/2005 As table 9 shows, the old age dependency rate is projected to increase considerably. While it is expected that the rate in the EU15 and the EU25 will be slightly more than twice as large in 2050, compared to 2004, in the EU10 this rate will rise from approximately 20 percent which is below the EU15 and EU25 level - to 50 percent. Even here we can observe a more favourable development in Sweden. While in 2004 the old age dependency rate in Sweden was somewhat above the EU average, in 2050 it is projected that this rate in Sweden will be 10 percentage points below the European average. In absolute figures we can express the development in the following way. While in 2004 the relation between the population in the working ages to the one 65+ was 4:1 in EU15, 5:1 in EU10 and somewhat lower than 4:1 in Sweden, in 2050 2 persons in the working ages have to support one person of 65+. In Sweden, this relation will be approximately 2,5:1. In the future, probably more people above 64 will work, perhaps often in part time employment. This could influence the old age dependency rates somewhat downwards. We will yet not try to calculate possible effects. As a summary for this section, we can mention the following regarding the projections for 2004 and 2050. Fertility in Europe is rising, but is too low to assure a stable population. This is valid even for Sweden, which has a higher fertility rate than the EU totally. Life expectancy both at birth and at 65 years is rising, but somewhat more in the EU than in Sweden, which has yet high life expectancy figures, above the European average for males and below for females. Young (0 14) and working age population (15 64) is falling in the EU, but not in Sweden. The consequence is that old age population (65+) is increasing much more in the EU totally than in Sweden. Relative to total population, yet even in Sweden the group of people between 15 and 64 is falling, but much less than in the EU. A consequence of the change of the age structure is increasing

9 old age dependency rates. The projections show that the Swedish dependency rates are growing considerably less than the ones in the EU totally. 4. Working age population, labour supply and employment In this section we will have a look at the development of some population variables which are crucial for the production possibilities of the total economy: - Population in the working ages (15 64) - Labour supply - Employment While the working age population probably is strongly related to total population, the labour supply is - besides of population - influenced by many factors: some people are too sick to work, others do not wish to work, a third group may try to look for a job unsuccessfully, etc. Regarding the young part of the population, it is possible that socioeconomically it is preferred that these groups are not working, but participating in education, training and instruction, which probably influences their future labour productivity in a positive way. Even employment is probably influenced by many factors. Here we can mention the demographic situation and development, but even the participation rates. Furthermore, (falling) unemployment can be mentioned as an important source of rising employment. In the next table (table 10) we will have a look at the projections for the working age population (15 64) and compare them with the ones for the total population. Table 10: Working age population (15 64) and total population, EU15, the EU25, the EU10 and Sweden, 2004 and 2050, millions Millions Population 15-64 Total Population EU15 EU25 EU10 Sweden EU15 EU25 EU10 Sweden 2004 255,1 306,8 51,7 5,8 382,7 456,8 74,1 9,0 2050 221,3 259,1 37,8 6,0 388,3 453,8 65,5 10,2 Difference 2004-2050 -33,8-47,7-13,9 0,2 5,6-3 -8,6 1,2 Difference percent -13,2-15,5-26,9 3,4 1,5-0,7-11,6 13,3 Source: European Economy, Special Report no 4/2005 As table 10 illustrates, the changes of the working populations are much larger than the ones of the total population. While the EU15 in 2050 according to the projections will have a somewhat larger total population, the one between 15 and 64 will fall with 13 percent. The population change in the EU10 is much more drastic: here according to projections, the total population will fall with close to 12 percent, while the working population will fall with 27 percent. The consequence for the EU25 is the following one:

10 while population between 2004 and 2050 is projected to stagnate (a fall of 0,7 percent), the working population will fall with more than 15 percent. Obviously, the population situation in Sweden is more favourable; while population is projected to rise (with 13 percent), even the population between 15 and 64 will increase (with approximately 3 percent). Before presenting the projections for the labour supply, we will have a look firstly at the labour market participation rates and secondly at the unemployment rates. Table 11 shows the projected labour market participation rates for the population between 15 and 65 for 2003 11 and 2050. Table 11: Projected labour market participation rates for the population between 15 and 64, EU15, the EU25, the EU10 and Sweden, 2003 and 2050 Participation rates, age group 15 64, % 2003 2050 Change EU15 70,4 76,1 5,7 EU25 69,6 75,5 5,9 EU10 65,4 71,8 6,4 Sweden 77,5 81,1 3,6 Source: European Economy, Special Report no 4/2005 As table 11 illustrates, the projected participation rates are highest in Sweden both 2003 and 2050, followed by the EU15 and EU25. The participation rates are lowest in EU10. Yet it can be observed, that the increase of participation rates is lowest in Sweden and highest in the EU10. In the next table we have a look at the change of projected labour market participation rates for different age groups (table 12). Table 12: Projected changes of labour market participation rates, 15 24, 25 54, 55 64, EU15, the EU25, the EU10 and Sweden, 2003 and 2050 Changes of participation rates, % 15-24 25 54 55-64 EU15 1,4 5,1 17,8 EU25 2,2 5,3 17,7 EU10 1,7 6,2 18,3 Sweden 3,7 3,5 6,9 Source: European Economy, Special Report no 4/2005 Table 12 shows that the labour market participation rates are increasing in all age groups. In the EU, participation rates are rising especially much in the age group 55 64. In Sweden, the participation rates for the group aged 15 24 are increasing more than in the EU, while the opposite is true for the other age groups. As we can observe, the increase of the participation rate for the older worker group is quite low in Sweden. On the other 11 The Commission chooses for most of the projections 2004 as a start year. For some of the population projections, 2003 is chosen as a start year.

11 hand, labour market participation and as we will see later - employment is quite high in Sweden, compared to the EU totally. Having a such large projected increase of participation in the group of 55 64 years old people, it can probably expected, that even the participation rates for the group 65+ will rise in the future. For this group, we yet do not present participation rates. The European Commission has shown 12 that already in the quite short period from 2000 to 2004, labour market activity rates for the age group 55 64 has grown substantially. A third factor besides of the working age population and the participation rates which influences total employment is unemployment. In the next table (Table 13) we will present the projections for unemployment. Table 13: Projected unemployment as per cent of the Labour force, EU15, the EU25, the EU10 and Sweden, 2003 and 2050 Unemployment rates, % of labour force 2003 2050 change EU15 8,2 6-2,2 EU25 9,3 6,1-3,2 EU10 14,8 6,6-8,2 Sweden 5,7 4,3-1,4 Source: European Economy, Special Report no 4/2005 As table 13 shows, according to the projections unemployment will fall substantially. The largest fall can be observed in the new member countries, while the smallest fall is projected in Sweden. Yet, while the EU10 even in 2050 is projected to have higher unemployment rates than the EU, for Sweden the unemployment rates are projected to be below the European average. The next step is to present figures for labour supply (table 14). Table 14: Labour supply projections, EU15, the EU25, the EU10 and Sweden, 2003 and 2050, 1000 persons Labour supply, 1000 persons 2003 2050 change change % EU15 178659 168393-10266 -5,7 EU25 212306 195527-16779 -7,9 EU10 33647 27134-6513 -19,4 Sweden 4491 4905 414 9,2 Source: European Economy, Special Report No 4/2005 According to table 14, labour supply in the EU is projected to fall between 2003 and 2050. Even in this aspect, the situation in Sweden is unique; in this country labour supply is projected to increase with more than 9 per cent. 12 European Economy, Special Report No 3/2005

12 If the Labour force is approximated with Labour supply, a fall of the unemployment rate with 1 percent unit could implicate a rise of employment in the EU25 with close to 1 million persons. It is yet possible that the projections for the labour supply are somewhat low. They are related to the population between 15 and 64. Expecting a relatively large increase of labour market participation rates for workers between 55 and 64 years, probably even the number of people 65+, who want and have the right to participate in the labour market will rise. Before having a look at the according to us perhaps most important variable total employment we will present the employment rates for the working age population (15 64) and the group of worker between 55 and 64. Already in table 12 we could illustrate that this group is projected to have a high increase of participation rates. Table 15 shows the employment rates for the total working population and the ones for group of 55 64. Table 15: Employment rates for 15 64 and for 55 64 old persons, EU15, EU25, EU10 and Sweden, 2003 and 2050, percent Employment rates, % 2003 2050 change % 15-64 EU15 64,6 71,5 6,9 EU25 63,1 70,9 7,8 EU10 55,7 67,1 11,4 Sweden 73,1 77,6 4,5 55-64 EU15 41,4 60,2 18,8 EU25 39,9 58,9 19 EU10 31,7 52,9 20,2 Sweden 68,8 76,6 7,8 Source: European Economy, Special Report No 4/2005 Table 15 illustrates that the employment rates are increasing, especially in the group of 55 64 years old persons. We can even observe (compare with tables 11 and 12), that the employment rates are increasing more than the participation rates. This can be explained with falling unemployment. The increase of Swedish employment rates is not unexpected lower than the one in the EU. Sweden has both in 2003 and in 2050 higher employment rates than the EU totally. It is projected, that Sweden already in 2003 has higher employment rates for the group 55 64 than the EU on average in 2050. The EU10 is projected to have the largest increase of employment rates. Yet this group of countries is projected to have the lowest employment rates both in 2003 and in 2050. We are now ready to have a look at the employment projections (table 16).

13 Table 16: Employment projections, EU15, EU25, EU10 and Sweden, 2003 and 2050, 1000 persons Employment, 1000 persons 2003 2050 change change % EU15 163984 158270-5714 -3,5 EU25 192638 183625-9012 -4,7 EU10 28653 25355-3298 -11,5 Sweden 4234 4694 460 10,9 Source: European Economy, Special Report No 4/2005 Table 16 shows that while employment in the EU as projected will fall between 2003 and 2050, the opposite is projected to happen in Sweden: employment is rising with more than 10 percent. According to us, four population variables are of special interest: - Total population - Working age population (15 64) - Labour supply - Employment In table 17 we gather the projections for these variables from the tables above to summarize this section and to make it easier to compare them. Table 17: Projections for population, population between 15 64, labour supply and employment, EU15, EU25, EU10 and Sweden, 2003 and 2050, millions of persons Changes 2004-2050 EU15 EU25 EU10 Sweden Population, millions 5,6-3,0-8,6 1,2 Population, % 1,5-0,7-11,6 13,3 Population 15-64, millions -33,8-47,7-13,9 0,2 Population 15-64, % -13,2-15,5-26,9 3,4 Changes 2003-2050 Labour Supply, millions -10,3-16,8-6,5 0,41 Labour Supply, % -5,7-7,9-19,4 9,2 Employment, millions -5,7-9,0-3,3 0,46 Employment, % -3,5-4,7-11,5 10,9 Source: European Economy, Special Report No 4/2005, authors calculations Table 17 illustrates the following. While population in the EU15 and EU25 more or less is stagnating, in the new member countries a relatively large fall is projected. Sweden is one of the few countries in Europe with positive population projections. Regarding the working age population (15 64), the development in the EU is quite dramatic. Large falls are projected in the EU15, the EU25 and especially in the EU10. The opposite development is projected for Sweden; here a rise of the working population is expected.

14 Even the labour supply in the EU is projected to fall, but not as much as the working population. For Sweden it is projected that the labour supply will rise: less than population, but more than the working population. Finally, we can observe that employment is projected to fall in the EU totally, but not as much as the working population and the labour supply. For Sweden we can observe a rise of employment. Obviously, though the projections for the working population in the EU are quite gloomy, employment will not fall as much as working population. We can mention at least the following explanations. Unemployment is projected to fall. Labour market participation is projected to rise, especially for the groups of older worker (55 64). Without presenting any figures, we would expect that even labour market participation for people of the ages 65+ will rise. Finally, probably even net immigration will level out some of the negative population developments. 5. Age-related public expenditures As could be shown above, while the total population in the EU is stagnating, the age structure is changing: fewer young and working age people, more people 65+. The older part of the population will not only grow in relative terms, but even absolutely. Especially the group of very old people (80+) is projected to increase considerably. In this section we will investigate the consequences of the projected change of the age structure for the public expenditure on pensions, healthcare, long-term care, education and unemployment transfers 13. To start with, let us discuss shortly how public expenditure in relation to GDP could be influenced by ageing. Starting with pensions, because the increase of the population 65+ both in relative and in absolute figures, it is expected that expenditures will rise. Average retirement age, which today is below 65 years, seem to rise, which could have some moderating effects. As shown above, life expectancy of the people 65+ is rising. With other words, in the future there are not only more people in these ages, they even will live longer. The question is, whether they will live longer in good health or in bad health. But even if the health status of older people will improve, probably this means because of the rising number of people, larger expenditures for health care. While the number of people 65+ is increasing, the group 80+ is increasing much more. Probably these people need quite much of long-term care, either in institutions or at home. Therefore it is expected that expenditures for long-term care will rise. While the group of older people is increasing, the development regarding younger people is the opposite one. It is therefore possible that expenditures for education are decreasing. On the other hand, it is even possible that expenditures per student are increased to improve quality of education. Therefore, probably, expenditure savings are not very 13 The section is based mostly on European Economy, Special Report 1/2006

15 large. Finally, falling unemployment rates probably will lead to falling expenditure for unemployment benefits, at least relative to GDP. Let us now have a look at the projections in age-related public expenditures (table 18). Table 18: Projections of age-related public expenditures, EU15, EU25, EU10 and Sweden, 2004 and 2050, percent of GDP Public expenditures, % of GDP EU15 EU25 EU10 Sweden Pensions 2004 10,6 10,6 10,9 10,6 2050 12,9 12,8 11,2 11,2 Change 2,3 2,2 0,3 0,6 Health care 2004 6,4 6,4 4,9 6,7 2050 8 8 6,2 7,7 Change 1,6 1,6 1,3 1 Long-term care 2004 0,9 0,9 0,2 3,8 2050 1,6 1,5 0,4 5,5 Change 0,7 0,6 0,2 1,7 Unemployment benefits 2004 0,9 0,9 0,4 1,1 2050 0,7 0,6 0,2 0,9 Change -0,2-0,3-0,2-0,2 Education 2004 4,6 4,6 4,7 7,3 2050 4 4 3,4 6,4 Change -0,6-0,6-1,3-0,9 Total 2004 23,5 23,4 21,1 29,6 2050 27,2 26,8 21,3 31,8 Change 3,7 3,4 0,2 2,2 Source: European Economy, Special Report No 1/2006 As table 18 illustrates, the level of age-related public expenditures is projected both in 2004 and 2050 to be higher in Sweden than in the total EU, where the change of expenditure is larger than in Sweden. Furthermore, the level of expenditures in the EU10 is projected to be lower than in the EU15 and in Sweden and close to constant. The table shows furthermore, that pensions represent the largest part of the expenditures in the EU, followed by health care, education, long-term care and unemployment. We can even observe that the decrease of expenditures for unemployment benefits and education is not large enough to level out the increase of expenditures for pensions, health care and longterm care. To summarize this section, the following can be mentioned. Relative to GDP, the agerelated public expenditures in the EU are rising. Compared with the projected changes of the number of people in the ages 65+, the projected changes of expenditures seem to be quite moderate.

16 6. GDP, Economic growth and Growth theory: a short discussion Whether pensions are financed by a pay-as-you-go system (PAYG) or financed by returns on capital funds, total production and its growth is the base of the material standard of living, both for the working population and the retired one. Therefore we will in this section have a look at total production, expressed by potential GDP, and its growth. We will even have a look GDP per head of population (GDP pc), which can be seen as an average measure of living standard, and at GDP per employed person, which often is used to express (labour) productivity. To project the future level and growth of production, i.e. of GDP, the Commission report 14 uses a production function approach of the following type: (1) Y = TFP x L(exponent a) x K(exponent 1-a) where TFP (total factor productivity) embeds the technological level. L is equal to total employment 15 and K shows the capital stock 16. Furthermore, constant returns to scale are assumed. The production function is slightly changed to: (2) Y = (TFP (exponent 1/a) x L) (exponent a) x K(exponent 1-a) or (3) Y = E x L (exponent a) x K (exponent 1 a) With other words, in this type of production function, technological progress is labouraugmenting and E x L can be seen as employed labour in efficiency units 17. Finally, the exponent a can be seen as the share of labour costs (= 0,65) in GDP 18. As mentioned before, we are interested in three different growth rates: - Potential GDP - GDP pc - GDP per employed person (GDPEMP) According to the production functions (1) (3), potential GDP can increase, when the technological level, expressed by TFP or E, increases, when L increases and when K increases. GDPEMP can increase, when the technological level increases and when K grows relative to L (capital deepening). Finally, GDP pc can increase because of technological progress, capital deepening and rising participation rates. 14 European Economy, Special Report No 4/2005, chapter 3: Labour productivity and GDP 15 When the number of worked hours per person is constant 16 According to our interpretation, K is the physical capital stock 17 European Economy, Special Report No 4/2005 18 European Economy, Special Report No 4/2005

17 As we could show above, total L expressed as the number of employed persons - will fall, even though the labour market participation rates are projected to rise. According to the Commission, employment will fall with 0,1 percent a year 19. The Commission even assumes, that TFP will increase (with more than 1 percent a year) 20. This implicates that employment in efficiency units (E x L) will rise. What about capital deepening? There are two arguments for rising K/L: - if labour becomes relatively scarce, a relative increase of wages could be expected. This can lead to an increasing use of other factors, e.g. capital - the change of the pension systems from pay-as-you-go, i.e. redistribution via taxes and other public fees, to funded systems could lead to increased saving and investment with a rising capital stock as a consequence. To summarize this section, the following can be mentioned. Though a small decrease of employment is projected, TFP will continue to increase. Therefore, it is projected that employment in efficiency units (E x L) will rise. Furthermore, the Commission projects that capital deepening will continue. The conclusion is therefore, that GDP can continue to rise. With a stagnant or even falling population, even GDP pc can rise. Finally, rising technological level and rising capital deepening make it possible, that even GDP per employed person (GDPEMP) can rise. 7. Projections for GDP, GDP pc and GDPEMP: The Growth Rates In this section we will present the projections. To start with, we present the projections for potential GDP (table 19). Table 19: The projections for the growth rates of GDP, EU15, EU25, EU10 and Sweden, 2004 2050, percent per year Growth rates for potential GDP, % EU15 EU25 EU10 Sweden 2004-2010 2,2 2,4 4,5 2,7 2011-2020 2,1 2,2 3,5 2,7 2021-2030 1,4 1,5 2,5 2,1 2031-2040 1,2 1,2 1,2 1,8 2041-2050 1,3 1,2 0,6 1,9 2004-2050 1,6 1,7 2,3 2,2 The growth rates show the yearly average during the given period; 2,2 for EU15 in 2004 2010 means that the yearly GDP growth rate on average during this period is projected to be 2,2 % Source: European Economy, Special Report No 4/2005 Table 19 shows that on average growth rates of potential GDP are projected to fall, especially after 2020 21. The fall of growth rates is relatively large in the EU10. Yet, growth rates are positive, which implicates that GDP continues to increase. 19 European Economy, Special Report No 4/2005 20 European Economy, Special Report No 4/2005 21 While total employment is projected to rise until approximately 2020, after that year employment will fall (see European Economy, Special Report 4/2005; Carone & Costello, 2006)

18 The next growth variable to be presented is the one for GDP pc (table 20). Table 20: The projections for the growth rates of GDP pc, EU15, EU25, EU10 and Sweden, 2004 2050, percent per year Growth rates for GDP pc, % EU15 EU25 EU10 Sweden 2004-2010 1,9 2,2 4,6 2,3 2011-2020 1,9 2,1 3,8 2,3 2021-2030 1,4 1,6 2,7 1,7 2031-2040 1,3 1,4 1,5 1,6 2041-2050 1,6 1,5 1,1 1,8 2004-2050 1,6 1,7 2,6 1,9 The growth rates show the yearly average during the given period; 1,9 for EU15 in 2004 2010 means that the yearly GDP pc growth rate on average during this period is projected to be 1,9 % Source: European Economy, Special Report No 4/2005 Even for the growth rates of GDP pc, a fall is projected, as table 20 illustrates. Compared with the growth rates of GDP, the projected fall of growth rates of GDP pc is somewhat smaller. As the growth rates are positive, GDP pc will continue to increase. The authors have shown in other papers 22 - compared with the periods from 1820 to 2000 and 1980 to 2000 that growth rate projections between 2004 and 2050 are not lower than in the past (1820 2000). Finally in this section we will present the projections for the growth rates of GDP per employed person (GDPEMP) (table 21) Table 21: The projections for the growth rates of GDP per employed person (GDPEMP), EU15, EU25, EU10 and Sweden, 2004 2050, percent per year Growth rates for GDP per employed person, % EU15 EU25 EU10 Sweden 2004-2010 1,3 1,5 3,6 2,2 2011-2020 1,9 2 3,4 2,5 2021-2030 1,8 1,9 2,9 2 2031-2040 1,7 1,8 2,1 1,7 2041-2050 1,7 1,7 1,8 1,7 2004-2050 1,7 1,8 2,7 2 The growth rates show the yearly average during the given period; 1,3 for EU15 in 2004 2010 means that the yearly GDPEMP growth rate on average during this period is projected to be 1,3 % Source: European Economy, Special Report No 4/2005 Table 21 illustrates that the projected growth rates for GDPEMP on average are somewhat higher than the ones for GDP pc. For the EU10, a considerable fall in GDPEMP growth rates is projected. Even for Sweden a fall but much smaller can be observed. For the EU15 the variations in growth rates for GDPEMP are quite small. 22 Magnusson & Schuller, 2006a and 2006b

19 To summarize this section, we can mention the following. Though growth rates for GDP are projected to fall somewhat, even in the future, the total economy of the European Union will probably be quite large compared with the rest of the world. Even for Sweden we can observe a fall in growth rates for GDP. As the projections show, growth rates in Sweden are expected to be larger than in the total EU. Regarding the projected growth rates for GDP pc, we can observe a small fall for the EU. Even for GDP pc, growth rates for Sweden are projected to be higher than in the EU. The strong fall of growth rates in the EU10 leads to the conclusion that these countries probably will have problems to converge to the average GDP pc level in Europe. Living standard in all parts of Europe is projected to rise. Finally, it seems that the projections for the growth rates of GDP per employed person (GDPEMP) are much more stable, than the other growth rates. GDPEMP is not only describing productivity. The level and increase of these variables is sometimes seen as an indicator of international competitiveness 23. 8. The Development of GDP, GDP pc and GDPEMP according to the projections and Age-related expenditures Though the future development of the European population seems to be not very favourable for production and growth, at least for the period from 2004 to 2050, substantial growth of GDP, GDP pc and GDPEMP is projected. All growth rates are positive, even if falling tendencies can be observed. In this section we firstly illustrate the implications of positive growth rates for GDP, GDP pc and GDPEMP in monetary terms and constant prices. Secondly, we compare the level of these variables with the agerelated public expenditures. We start with the projections for GDP (table 22). Table 22: GDP in prices of 2004, EU15, EU25, EU10 and Sweden, billions of Euro, 2004 and 2050 Billions of Euro, prices for 2004 2004 2050 2050/2004 Change average year, % EU15 9772 20719 212,0 1,65 EU25 10246 22068 215,4 1,68 EU10 474 1349 284,6 2,30 Sweden 278 756 271,9 2,20 To adjust for different price levels in different countries, GDP usually is expressed in Purchasing Power Parities (PPP) or Purchasing Power Standard (PPS). In our source, obviously, the figures are not in PPP or PPS, but in (constant prices) Euro Source: European Economy, Special Report No 1/2006, authors calculations As table 22 illustrates, according to the projections, GDP in constant prices will in 2050 be more than twice as large as in 2004. Both in the EU10 and in Sweden, growth of GDP is projected to be larger than in the EU totally. 23 E.g. Porter, 1998

20 In the next table, we will present the projections for GDP pc (table 23). Table 23: GDP pc in prices of 2004, EU15, EU25, EU10 and Sweden, Euro, 2004 and 2050 Euro, prices of 2004 2004 2050 2050/2004 Change average year, % EU15 25534 53358 209,0 1,62 EU25 22430 48629 216,8 1,69 EU10 6397 20595 321,9 2,57 Sweden 30889 74118 239,9 1,92 According to the authors, the figures are not adjusted for different price levels, i.e. they are not expressed in PPP Source: European Economy, Special Report No 1/2006, authors calculations As table 23 shows, it is projected that GDP pc between 2004 and 2050 in the EU and in Sweden will grow with more than 100 percent, in the EU10 with more than 200 percent. In the next table, we will present figures for the GDP per employed person (table 24) Table 24: GDP per employed person (GDPEMP) in prices of 2004, EU15, EU25, EU10 and Sweden, Euro, 2004 and 2050 Euro, prices of 2004 2004 2050 2050/2004 Change average year, % EU15 59295 130909 220,8 1,74 EU25 52976 120180 226,9 1,80 EU10 16571 53205 321,1 2,57 Sweden 65986 161057 244,1 1,96 According to the authors, the figures are not adjusted for different price levels, i.e. they are not expressed in PPP Source: European Economy, Special Report No 1/2006, authors calculations Even for GDPEMP, a large rise between 2004 and 2050 is projected. As a comparison of tables 23 and 24 shows, with the exception of the EU10, GDPEMP is projected to grow slightly more than GDP pc. We are now reading to address the main question of this paper: Is ageing in Europe too expensive? Or with other words: Can Europe afford that the age structure of the population is changing in the way projected? To answer these questions, we will calculate the total expenditure for pensions and other age-related expenditures both by using GDP and GDP pc as bases. Table 25 shows the calculations for total expenditures, compared with GDP.

21 Table 25: GDP and age-related expenditures in prices of 2004, EU15, EU25, EU10 and Sweden, Euro, billions, 2004 and 2050 EU15 EU25 EU10 Sweden GDP, Euro, billions 2004 9772 10246 474 278 2050 20719 22068 1349 756 change 10947 11822 875 478 change % 112,0 115,4 184,6 171,9 Age-related expenditures, % GDP 2004 23,5 23,4 21,1 29,6 2050 27,2 26,8 21,3 31,8 Age-related expenditures, Euro, billions 2004 2296 2398 100 82 2050 5636 5914 287 240 change 3340 3516 187 158 change % 145,5 146,6 187,0 193,0 Source: European Economy, Special Report No 1/2006, authors calculations According to the European Commission, age-related expenditures represent between 21 percent (EU10, 2004) and 32 percent (Sweden 2050) of projected potential GDP. In table 25 we have transformed these relative figures into absolute ones. We can make several observations. Age-related expenditures will rise with enormous amounts. The relative rise of these expenditures is obviously larger than the one of GDP. Yet, in absolute terms, the rise of GDP is between close to three times to four times larger than the one of the agerelated expenditures. Therefore, the conclusion is that though ageing in Europe is expensive, it seems that because of the large projected increase of potential output the Europeans can afford to become older and older, at least up to 2050. Though absolute figures are often easier to understand than relative ones, they can be so large that they are not anymore comprehensible 24. Therefore we will even describe the age-related expenditures in per head of population (pc) terms. In our last table (table 26), we present the projections of GDP pc and compare them with the ones for the age-related expenditures per head of population. 24 One of the authors mother told, that she gave up, when the price of 1 US$ in Germany 1923 was 6 billion (European!) Reichsmark, or 1 US$ = 6 000 000 000 000

22 Table 26: GDP pc and age-related expenditures, per head of population, in prices of 2004, EU15, EU25, EU10 and Sweden, Euro, 2004 and 2050 EU15 EU25 EU10 Sweden GDP pc 2004 25534 22430 6397 30889 2050 53358 48629 20595 74118 change 27824 26199 14198 43229 change % 109 116,8 221,9 139,9 Age-related expenditures pc (ARE pc), Euro 2004 5999 5250 1350 9111 2050 14515 13032 4382 23529 change 8516 7782 3032 14418 change % 142 148,2 224,6 158,2 GDP Age-related expenditures (ARE), pc, Euro 2004 19535 17180 5047 21778 2050 38843 35597 16213 50583 change 19308 18417 11166 28805 change % 98,8 107,2 221,2 132,3 Source: European Economy, Special Report No 1/2006, authors calculations As already shown above, according to the projections, GDP pc and therefore even average standard of living will rise considerably between 2004 and 2050. Table 26 illustrates that the age-related expenditures per head of population (ARE pc) are rising even more. Yet, even when GDP pc is adjusted with ARE pc (i.e. GDP - ARE, pc) we can observe, that the standard of living can rise considerably. To summarize this section, the following can be mentioned. Age-related expenditures (ARE) are projected to rise considerably between 2004 and 2050. In relative terms they are increasing more than GDP and - per head of population - even more than GDP pc. But as our on the projections based calculations show, in absolute terms GDP in billions of Euro will rise three to four times as much as ARE. Even GDP pc will in absolute terms increase considerably more than ARE pc. Therefore we would state that ageing is not too expensive and that the Europeans can afford the economic consequences of an ageing population. 9. Summary and Conclusions The size, development and age structure of the population is determined by fertility rates, life expectance and migration. According to population projections for the period of 2004 to 2050, in the EU, though fertility rates are somewhat rising, they are below the ones which lead to a stable population (2,1 children per woman). Though Sweden has higher fertility rates than the EU on average, even here they are too low to stabilize population. Furthermore, life expectancy both at birth and at the age of 65 is projected to rise somewhat more in the total EU than in Sweden. Finally, compared with the period from 1960 to 2000, it is projected that net migration to the EU will increase considerably. It can be mentioned, that the new member countries (EU10) will change their position from a net emigration region to a net immigration one. Sweden has relatively seen a higher net immigration than the EU. This is projected to continue even in the future.