The Portability of Pension Rights Within Europe1

Free movement? The portability of pension rights within Europe1

“The Community shall have as its task, by establishing a common market and progressively approximating the economic policies of Member States, to promote throughout the Community a harmonious development of economic activities, a continuous and balanced expansion, an increase in stability, and accelerated raising of the standard of living and closer relations between the States belonging to it. …the activities of the Community shall include… (c) the abolition, as between Member States, of obstacles to freedom of movement for persons, services and capital”

“Such freedom of movement shall entail the abolition of any discrimination based on nationality between workers of the Member States as regards employment, remuneration and other conditions of work and employment.”

Articles 2-3, 48.2 Treaty establishing The European Economic Community, 1957

When the European Economic Community was founded in 1957 its purpose was a peaceful and stable Europe, and the creation of a common market was considered to be central to achieving this aim. Mobile citizens were constituents of this common market from the start and in order to enable mobility the Community pledged to remove obstacles workers might face. Among these were the treatment of part-time employees, equal pay and social rights connected to employment, such as parental leave regulations and pension rights. This was the theory. In practice it took a long time until more specific rights were introduced and up until today mobile workers might well be penalised for moving between countries. This paper will explore the mobility of social rights in Europe from the perspective of pension rights, among the most significant and long-lasting for citizens. In order not to be disadvantaged in comparison with a stationary employee, a mobile worker needs to be able to carry accrued pension rights with her or him to the new place, no matter whether this is in the same county or in another European one. If such mobility is not supported, citizens who move risk losing their pension benefits and could face poverty in retirement. It is therefore important to know how the broad guarantees given by the European Communities in 1957 look from this more detailed perspective. In the following the paper will first give an overview of the evolution of European law regarding public and occupational pensions. This

1 This paper is part of the ESRC project “Pension rights and mobility in Europe" (Paul Bridgen, Traute Meyer, Caroline Andow), based within the Centre for Population Change at the University of Southampton that began in November 2011. The data and analysis presented here is based on the collaboration of the research team. The main aim of the project is to explore whether movement within the EU is really ‘free’ or whether mobile workers might incur pension right losses on account of their mobility. This paper is a work in progress; please do not quote without permission.

1 is followed by a discussion of the reasons for the EU’s slowness in passing encompassing legislation. Indeed during the first decades the common market did not really lead to much mobility of labour, and therefore policy-makers could afford to stay passive. However, with the collapse of the Soviet bloc this situation changed and migration increased quite substantially after 1989, as part three will show. We will discuss what types of citizens are moving, where most of them have moved to and why. The sudden increase makes reform of social rights more pressing, and it also makes it more important to know what impediments to mobility there are for current migrants within the European Union. In part four we discuss the literature that has dealt with the question of how mobility might affect pension rights. This suggests there are good reasons to assume mobile workers are penalised. Investigations are yet lacking into what such principles might mean in everyday life and how they affect public as well as occupational rights. In part five the paper offers a first empirical investigation into the impact of mobility on pension levels of workers who move between different types of pension regime, and in particular from East to West.

1. Portability of pension rights in the EU – The regulatory framework

Pension rights of mobile workers fall under two types of legislation: the mobility of public pensions between member states is governed by European law; the mobility of non-state pension rights is governed by national legislation.

Public pensions With the Treaty of Rome in 1957 EU member states committed themselves to creating an internal market that would include the free movement of workers. This notwithstanding it took 14 more years until the first legislation about the portability of statutory social security rights for mobile workers came into force in 1971 (Regulation (EEC) 1408/71 14 June 1971). This remained unchanged until 2010 when the Council simplified it (Regulation 883/2004, 1 May 2010) leaving its fundamental principles intact (Verscheuren, 2009). Thus, a mobile worker should be treated like a national in the Member State in which he or she chooses to work (Art. 4, 883/2004). In addition, the rules of only one Member State apply to this worker at any one time (Art. 11, 883/2004). Moreover, statutory benefits are exportable within the union, therefore, as a pensioner the individual is free to move (Art. 7, 883/2004). EU governments were aware that national systems were based on the assumption of lifelong membership, and that mobile workers might therefore be penalised. The methods of calculations they put in place were designed to avoid such effects. Thus, legislation stipulates that to work out someone’s pension after a pan-European career, the pension administrators in each Member State in which a migrant has worked have to make two calculations on their

2 behalf: firstly they determine the individual’s statutory entitlements only for the years worked in this Member State, using national legislation (“independent benefit”). Secondly, in order to avoid that rules favouring long term members like a minimum period of insurance penalise the mobile worker, national pension administrators also work out the pension entitlement of each mobile individual assuming that they spent their whole working career in this Member State; on this basis they apportion the individual’s entitlement in relation to the actual time spent in this particular Member State (“pro-rata benefit”) (Art. 52, 883/2004). The individual will receive whichever of these amounts is higher. If the pension benefit of a member state is based on a funded defined contribution scheme only the first calculation is necessary (Art. 52(5)). This change in the more recent legislation was a response to the different social security schemes in the expanded EU (Verschueren, 2009).

Non-state pensions European regulation for non-statutory supplementary pension rights, i.e. entitlements acquired under occupational or personal pension schemes is much weaker, because it does not force member states to avoid penalties for short term membership of schemes. The EU allows member states to voluntarily apply the EU legislation in place for state pensions (i.e. Reg. 883/2004) to their compulsory non-state systems, but to our knowledge only France has yet asked for this legislation to be extended in this way (OJ C135/17, 5.5.2011). In our view it is not surprising that national governments refrain from asking for a voluntary tightening of their legislative freedom. The mobility of non-state pension rights is protected in some ways. According to EU legislation workers who move between countries must be governed by national legislation for the times they spend in one country. This means that they must be treated exactly like workers who move within one Member State (Directive 98/49/EC (29 June 1998) Art. 4). Moreover, providers of supplementary pension schemes must inform members who move to other Member States about their rights (Art. 7) and after the worker’s retirement they must pay the pension to another Member State, if required, net of taxes and transaction charges (Art. 5). However, no rule addressing possible disadvantages that short term membership brings exist for non state pensions. The European Commission recognised that the existing legal framework needs improving to protect the supplementary rights of mobile workers more effectively. This is why in 2005 it proposed a portability Directive designed to remove obstacles to the acquisition, preservation and transferability of supplementary pension rights. However, this Directive has yet to pass the unanimity hurdle in the European Council – its scope has been reduced already, leading to less transferability of rights than in its original version (Oliver, 2009). Nevertheless, EU governments have not been ready to agree so far.

In summary, one important problem of pension right mobility, penalties for short-term membership, has been addressed by EU-legislation, but only for public pensions. Reliance on one type of benefit only is rare in Europe, and many employees have additional non-state provision, but the mobility of occupational pensions much depends on the quality of national legislation. This varies between countries, creating unequal opportunities for workers from different countries.

2. Mobility of workers in the European Communities before 1989

As we have just seen, from the perspective of pension rights, important obstacles to the mobility of workers remain. One important reason for this is that after the treaty of Rome was agreed in 1957 the market for goods integrated much faster than the labour market. Between 1960 and 1970 the inter-EU trade in goods increased rapidly, from 34% of all imports and exports of the six founding member states to 48% and to 60% of all twelve members by 1990 (Tsoukalis 1997: 19). During the same period only a small and in fact decreasing share of the EU population was mobile within Europe. Indeed in the average of the period 1958 to 1980 trade within the EC as a proportion of all EC trade increased by 0.8 per cent per year while the movement of workers from one EC country to another as a proportion of the total number of labour migrants in the EC decreased by 1.3 per cent per year (Straubhaar 1988: 51). The movement of workers with EC citizenship as a share of all migrant workers in the EC fell from around 60% at the beginning of the 1960s to 20% at the beginning of the 1970s and it remained at this level until at least the late 1980s (Straubhaar 1988:51). It is generally recognised in the literature that the intra EU labour market is characterised by “free movement and little mobility” (Tsoukalis 1997: 117). Migration within Europe as a share of the total population has been low, in absolute terms and in comparison with the United States, Australia, Canada and Japan (Puhani 2001: 129). This relative immobility is not very surprising if we consider the difficulties a person who moves away from the country he or she was born in encounters: they leave behind their families and friends, must use another language, they have to live in a new cultural context. In the literature on migration this is why it is assumed that one common reason why people become migrants is that they are “pushed” away from their countries of origin. They move when the economic or political problems in their home countries become so strong that they prefer the upheaval of leaving to the difficulties of staying. This is why before 1989 main movements within the internal market took place from South to North, i.e. from the poorer regions, Southern Italy, Greece, Portugal, Spain, to the richer countries (Tsoukalis 1997: 118). This is one explanation as to why European institutions and legislators could afford to focus much more on rules and regulations for cheese, car tires and other goods than on removing

4 the obstacles for persons. Another, perhaps more important reason is that the 1960s and early 1970s are generally regarded as the “golden age” of the welfare state, i.e. as a time when unprecedented economic growth strengthened trade unions and enabled governments to expand their national welfare states. This means they focused on national protection, rather than European mobility during that time (Tsoukalis 1997: 22). In 1989 the Soviet Bloc disintegrated, leading to an extension of the European Union more than a decade later; in 2004 Poland, Hungary, the Czech Republic, Slovenia, Slovakia, Estonia, Lithuania, Latvia joined, Bulgaria and Romania in 2007. The economies of Continental and Eastern Europe were much poorer than those of the long-standing members, even if compared to the Southern states of the EU15 and as a consequence, migration from these countries to Western Europe increased rapidly, as we will show below.

3. Mobility of workers in the European Union after 2000

As shown, liberalisation of movement within the Community did not prompt mass migration and intra-EU movement remained low for a long time, but data from before 2000 is patchy so it is difficult to show exactly how migration developed. The academic literature acknowledges that even recent international migration statistics can be unreliable, and that there are problems with the quality and comparability of data2 (for example Lemaitre, 2005). Kupiszewska and Nowok (2005) identify Eurostat as the most comprehensive source on the movement of migrants between EU member states. Both Eurostat and the OECD provide fairly consistent data on annual inflows and the size of the foreign population in EU states from 2000 onwards. We use this data in our analysis, triangulating both sources where possible to increase confidence in the figures3. We show that in recent years intra-EU migration has been increasing and that the social rights issues connected to mobility have thus become relevant for more European citizens.

2 See appendix 1

3In some cases both data sources report the same figures; divergent data suggests that exact measures of the size of inflows or foreign population cannot be given.

3.1 Size of foreign population

Table 3.1: Overview of foreign nationality populations by country

% pop of EU %change EU %change % of pop of foreign foreign EU+OECD foreign EU+OECD foreign nationality nationality pop nationality pop nationality (2009) (2009) 2000-2009 2000-2009 Spain 4.96-5.11 5.43 329-464 302 UK 2.91-3.1 3.7 70-131 99 Germany 2.88-3.09 5.21 -3 - 5 -9 Italy 1.88-2.06 2.18 338-359 241 France 1.98-2.01* 2.51 0** :: Netherlands 1.76-1.88 2.66 34-40 21 Poland 0.04 0.04 25-33*** 0**** Source: Eurostat, Population and Social Conditions, International Migration and Asylum and OECD, International Migration Database and Population Statistics, own calculations Where a range of values is shown, data has been drawn from Eurostat and the OECD :: missing data *size of population in 2007, data appears incomplete ** Eurostat data only, % change 2003-2009 ***125 based on %change 2001-2009 (Eurostat), 133 based on %change 2002-2009 (OECD) **** % change 2002-2009 (OECD)

Table 3.1 provides an overview of the EU and EU+OECD4 foreign nationality populations in eight EU member states5 drawing on data from Eurostat and the OECD. We acknowledge that looking at the foreign population by nationality6 might capture differences in naturalization practice as opposed to the absolute size of the population (Bonin et al., 2008) but we use these figures to indicate broad trends. This table demonstrates that in 2009 the size of the EU foreign population was particularly large in Spain and very small in Poland. Further it demonstrates that the size of the EU and EU+OECD foreign populations in the western European countries of Spain, the UK, Italy, and the Netherlands have increased between 2000 and 2009. The increases have been dramatic in Spain and Italy suggesting that these two countries in particular have become more attractive to migrants. The data is unreliable or

4 See appendix 2

5 See appendix 3

6 According to the OECD, European countries have traditionally produced data on foreign residents (ie residents with a foreign nationality) rather than the number of foreign-born. The data is more comprehensive for foreign nationality than it is for foreign-born citizens. See note on OECD data definitions: http://stats.oecd.org/OECDStat_Metadata/ShowMetadata.ashx?Dataset=MIG&Lang=en&Coords=[VA R].[B15]

6 missing for France and Poland and in Germany the EU foreign population has not increased and may have even declined slightly. The small difference between the size of the EU foreign populations and the EU+OECD foreign populations show that migrants are mostly from within the EU. Only in Germany there is a much larger EU+OECD than EU foreign population. Along the same lines it is interesting to note that generally the EU+OECD foreign populations are increasing at a slower rate than the EU foreign populations, with the UK as the only potential exception. This emphasises the increasing importance of intra-EU migrants in the foreign nationality populations of these countries.

3.2 Annual migration flows

Tables 3.2 and 3.3 show the annual immigration flows into these countries since 2004, i.e. the years following the two rounds of enlargement. Between 2004 and 2007 there was an increase of migrants into all of these countries7, and the OECD data (table 3.3) enables us to see that most of these migrants to the Western EU countries were from within the EU. In 2007 Spain and Italy were receiving similar numbers of EU immigrants as Germany. These two countries were traditionally mass emigration countries as they were poorer than the Northern European member states. However, it appears that following the inclusion of poorer Eastern European countries, Spain and Italy were receiving unprecedented numbers of EU immigrants.

Comparing the percentage change in annual inflows of immigrants between 2004 to 2007 and 2004 to 2009 allows us to see that there have not been linear increases in the number of inward migrants to these countries since 2004. There has been a decline in numbers in all countries at some point since 2007, with Germany and the Netherlands being the least affected and Spain and Italy seeing the most drastic declines; between 2007 and 2009 the flows into Spain and Italy more than halved. The reason for these dips is the global financial crisis and ensuing economic recession which was felt more sharply in different countries at different times. However, as we have seen above, in spite of these declines, the size of the foreign populations in these countries still saw steady increases during this time period.

7 Only the EU+OECD migrants to France show a small decline between 2004-2007 but we are cautious with this data. France is a difficult case because the data is only available from the OECD and it appears incomplete. For example, the data suggests that in 2009 France received just 3,108 EU immigrants and that these migrants only came from the new member states; there are no recorded immigrants from the original EU-15 countries. We only use the data on France to draw tentative conclusions

Table 3.2: Total inward EU migrants by country, 2004-2009, Eurostat data

Germany Spain Italy UK The Netherlands Poland

2010 :: 145,353 118,611 :: :: ::

2009 :: 144,867 136,133 167,415 47,312 ::

2008 335,914 193,329 212,862 197,719 55,413 3,060

2007 343,851 389,203 324,801 171,863 43,228 196

2006 320,727 304,349 67,838 :: 31,921 409

2005 318,378 257,767 72,470 120,557 27,519 233

2004 301,486 249,278 95,300 111,488 26,351 164

2007 as a % of 2004 114 156 341 154 164 120

2009 as a % of 2004 111* 58 143 150 180 1866*

Source: Eurostat, International Migration and Asylum, International migration flows, own calculations Missing data for France; :: data not available *Figure refers to 2008 as a % of 2004 because of missing data

Table 3.3: Total inward EU and EU+OECD migrants by country, 2004-2009, OECD data

Germany Spain France The Netherlands Poland EU EU+OECD EU EU+OECD EU EU+OECD EU EU+OECD EU EU+OCED 2009 348,090 417,932 144,867 163,844 3,108 15,542 53,797 65,097 6,270 10,679

2008 335,914 405,028 193,329 216,508 5,918 19,760 55,247 67,036 8,234 12,697

2007 346,089 414,769 389,203 415,697 5,453 18,636 43,111 53,242 13,047 16,774

2006 322,653 392,210 304,349 330,715 3,684 17,809 31,814 42,272 8,056 11,183

2005 318,751 391,604 257,767 282,100 3,307 18,338 27,437 37,286 12,481 15,257

2004 301,960 381,562 249,278 271,729 3,575 19,153 26,283 36,765 9,271 11,929

2007 as a % of 2004 115 109 156 153 153 97 164 145 141 141 2009 as a % of 115 110 58 60 87 81 205 177 67 90 2004 Source: OECD, International Migration Database, own calculations Missing data for the UK and Italy

3.3 Migratory movements in the EU

Despite some recent declines in annual inflows, the sizes of the EU foreign populations in the western European countries we have looked at have been increasing since 2000. Where do these migrants come from? Immediately two patterns are obvious: 1/ there is a general movement from east to west; and 2/ there is a divide between migrant movements in northern and southern Europe. Map 1 illustrates these movements.

Map 1: Main migration flows by nationality between our selected EU countries

SOURCE: OECD, International Migration Database and Eurostat, Population and Social conditions, own calculations France is not included because of a lack of reliable data

> (more than) 100,000 immigrants per year ≈ (approximately equal to) 50,000 immigrants per year < (less than) 12,000 immigrants per year

Calculation of flows depends on available data as to nationality of immigrants; Romanians to Spain (2009); average based on OECD and Eurostat data Romanians to Italy (2009); Eurostat data only Polish to Netherlands (2009); average based on OECD and Eurostat data Polish to Germany (2008); average based on OECD and Eurostat data Polish to UK (2005); OECD data only Germans to Poland (2008) average based on OECD and Eurostat data

The pattern of east to west migration in Europe has long been recognised (for example Black et al. 2010). There has however been little recognition of the north/south divide; in the north we see large movement of Polish migrants to the UK and Germany, and in the south we see large streams of Romanians moving into Italy and Spain. There are smaller, yet still significant streams of Polish migrants to the Netherlands, and Romanian migrants to France8. Therefore it does seem that the recent enlargement of the EU has significantly impacted on the migratory patterns we see, especially with regard to Italy and Spain changing from traditional emigration to immigration countries. In the past, Southern Europeans migrated to the richer north, but now emigrants from the poorer Eastern European countries are migrating to the west, along the latitudinal lines identified.

3.4 Who is moving and why?

East to West migration: Romania to Italy and Spain We have seen above that many emigrants were still going to Spain and Italy in 2010 despite dramatic falls since 2007. A large number were from Romania. In 2007, nearly 200,000 Romanians entered Spain, accounting for half of all EU inward migrants into Spain. In 2009, the number of Romanian immigrants had declined rapidly to just over 52,000 increasing again to just over 60,000 in 2010 (Eurostat). In 2010 Romanians accounted for 41% of inward EU migrants to Spain, still the largest share, and they made up 36% of EU foreign nationality citizens in Spain (Eurostat). Italy reported even more Romanian immigrants in 2007, over 270,000 which accounted for 84% of all EU migrants into Italy. By 2010 this number had fallen sharply, but still just over 92,000 Romanian immigrants entered Italy, 78% of all inward EU migrants into that country (Eurostat). In Italy, Romanians accounted for an even larger share of EU foreign citizens, 72% in 2010 (Eurostat). Who are these Romanian migrants?9 In 2008/9 76% of Romanians moving to Spain and 80% moving to Italy were between 20 and 64 years old, and in this group the larger share was younger than 34 (Eurostat; Stanek 2009). Eurostat data also suggests that in 2008 and 2009 more women than men came from Romania to Spain and Italy, but this is contested by other studies (Stanek 2009), and therefore we leave this question open. Moreover, the Romanians moving to Spain, Italy and France are mostly educated to secondary or tertiary level and almost half of those in Spain had a degree (table 3.4, see also Ahonen et al. 2009).

8 The data for France is not considered reliable and is used only to make tentative suggestions

9 See appendix 4

Table 3.4: Educational level of Romanian migrants to Spain, Italy and France

Primary Secondary Tertiary Romanians to Spain 24% 13% 62% Romanians to Italy 35% 55% 10% Romanians to France 24% 29% 47% Source: Database on Immigrants in OECD countries, OECD 2000 census, own calculations10 Data refers to educational level in country of origin.

Their high level of education notwithstanding, Romanians in Spain and Italy are likely to be employed in unskilled or semi-skilled occupations (table 3.5, see also Alexandru, 2007; Ahonen et al. 2009). The number of skilled Romanians in France is an exception.

Table 3.5: Occupational category of Romanian migrants to Spain, Italy and France

Mode occupational category Elementary Occupations (36%) Romanians to Spain Crafts and related trades workers (27%) Elementary Occupations (25%) Romanians to Italy Crafts and related trades workers (20%) Romanians to France Professionals (35%) Source: Database on Immigrants in OECD countries, OECD 2000 census, own calculations

The OECD data suggests that Romanian immigrants in Spain and Italy are likely to stay for less than five years (table 3.6), but Romanian immigrants in France stay longer. Generally, there are many difficulties in recording length of stay (Engbersen et al. 2010, Pollard et al. 2008, Stanek 2009), and therefore we cannot be sure how long these migrants will stay.

Table 3.6: Duration of stay of Romanian migrants in Spain, Italy and France Up to 5 5 to 10 years More than 10 years Unknown years Romanians to Spain 76% 7% 17% 1%

Romanians to Italy 53% 25% 8% 14%

Romanians to France 19% 21% 40% 19% Source: Database on Immigrants in OECD countries, own calculations

10 The International Standard Classification of Education (ISCED; cf. UNESCO 1997) was used as a baseline, but groups have been aggregated as follows: Primary level: ISCED 0/1/2; Secondary level: ISCED 3/4; Tertiary level 1: ISCED 5A/5B.; Tertiary level 2: ISCED 6.

Explaining migration to southern Europe Spain and Italy have turned from emigration to immigration countries because Romanians left their home country after the enlargement of the European Union. Spain and Italy are relatively poor compared to member states in the north, but they are relatively rich in relation to those in the east. In the academic literature the situation in Romania after communism is considered to have been a major push factor for migration. High unemployment and unstable economic and social conditions led to mass emigration (Alexandru, 2007). At the same time, in the late 1980s and early 1990s, there was rapid economic development in Italy and particularly Spain (Stanek, 2009). Migrants often choose their destination because they hope to find employment (Potot, 2010; Van Dalen and Henkens, 2012). A large informal economy and a demand for low skilled labour, especially in agriculture made Spain an attractive destination for Romanian migrants (Arango and Jachimowicz 2005; Culic, 2008; Potot 2010), while in Italy an underground labour market offered temporary work for illegal migrants (Ban 2009; Uccellini, 2010; Culic, 2008). An ineffective immigration control system in both countries made such movements easier (Arango and Jachimowicz 2005; Culic, 2008; Ban, 2009; Potot 2010, Uccellini, 2010). These are clear push and pull factors but they do not explain why other migrants, such as the Polish, were not drawn to these countries but instead moved elsewhere within the EU. Additional reasons for the predominance of Romanian migrants in the south are the similarity between the Romance languages Italian, Spanish, Romanian and French (Culic 2008; Pollard et al., 2008; Ban 2009; Uccellini, 2010; OECD 2012:22-23) as well as social and cultural networks. Linguistic similarity coupled with geographical proximity made Italy a particularly favoured destination for Romanians (Culic, 2008; Uccellini 2010). In both countries, social ties formed after first migrants settled which encouraged further migration (Arango and Jachimowicz, 2005; Culic, 2008; OECD 2012:22). Indeed, recent ethnographic research about Romanian immigrants in Italy identified well- established migrant communities in particular areas (Culic, 2008) and social networks are known to be a significant pull factor for migrants generally (see for example Van Dalen and Henkens, 2012:34).

East to West migration in Northern Europe Like Spain and Italy, Germany and the UK have a large number of annual immigrants, but because of their more wealthy economies they have always been immigration destinations within the EU. In both countries the largest flows are from Poland. Germany received just over 150,000 Polish immigrants in 2006, 39% of all EU immigrants to the country in that year (both data sources). This number had decreased to just over 112,000 by 2009 but this was still 32% of all inward EU migrants to Poland (OECD data). Interestingly, despite the fact that the Polish have made up the largest group of migrants since 1995, the largest share of the stock of

EU foreign nationality population in Germany since 1985 has been Italian, a result of earlier migration into Germany (22%, 2009, both sources). Further, the large EU+OECD share of foreign population in 2009 highlighted above was predominantly accounted for by Turkish citizens, at least in part due to the fact that second generation Turks born in Germany prior to 2000 were given Turkish citizenship. However, with the large inflows of Polish migrants, we expect that in the future Poles will make up the largest share of the foreign nationality population in Germany. In the UK we also see a large inflow of Polish migrants, although the most recent data is from 2005 (Eurostat). In this year, just shy of 50,000 Poles arrived in the EU which account for 41% of all inward EU migrants into the country. This was a huge increase from fewer than 4,000 Poles in 2003, 5% of all EU immigrants into the UK. By 2007, Poles had taken over the dominant share of the EU foreign population in the UK from the Irish and in 2009 Polish immigrants accounted for just under a third of the foreign nationality population (OECD). Poles also make up the largest group of immigrants to the Netherlands, although the numbers are much smaller. The data for 2009 diverges, with Eurostat recording just over 11,000 and the OECD recording nearer 13,000 Polish immigrants to the Netherlands. In either case, the data shows that the Polish accounted for 24% of all EU immigrants in this year. The largest share of the EU foreign population is of German nationality, and this has been the case since 1985. However, as in Germany, the EU+OECD foreign nationality stock is larger, because of a significant Turkish population. Who are these Poles moving in Northern Europe? Again, the large majority are of working age: in 2008/9 92% of the Poles moving to Germany and 86% of those going to the Netherlands were between 20 and 64 years old, and again a substantial share were younger than 34 (Eurostat). There are no equivalent figures for Britain, but the literature suggests they too are young (Drinkwater et al. 2010; OECD 2012). This time our data suggests that these migrants are more likely to be male: in 2008/9 about 10% more men than women moved to the Netherlands and about 50% more moved to Germany. Engbersen et al. (2010) found that Polish migrants to the Netherlands prior to 2004 were predominantly women but that there was a substantial increase in male Polish immigrants following enlargement. There is no data with regard to Polish migrants in the UK, however the literature suggests that they too are more likely to be male (Drinkwater et al. 2010; OECD 2012). As in Southern Europe, Polish migrants to Germany, the UK and the Netherlands are most likely to be educated above primary level. This is especially the case for Polish migrants to Germany and the Netherlands, but there is a large amount of missing data with regard to the UK (table 3.7).

Table 3.7: Educational level of Polish migrants to Germany, the UK and the Netherlands

Primary Secondary Tertiary Polish to Germany 30% 53% 17% Polish to UK* 20% 16% 22% Polish to Netherlands 38% 42% 21% Source: Database on Immigrants in OECD countries, OECD 2000 census, own calculations Data refers to educational level in country of origin; *missing data for Polish migrants to the UK, 42% unknown education level

However, the gap between educational qualification and sector of employment is smaller in Northern than in Southern Europe; Table 3.8 shows that workers moving to Germany tend to be employed in semi-skilled or skilled occupations and the Polish immigrants in the UK are most likely to be in skilled, professional occupations. However, in the UK the next largest group are ‘service workers’ and ‘shop and market sales workers’ and in the northern European countries only a minority works in skilled occupations. We do not have any data on the occupational categories of the Polish working in the Netherlands but the literature suggests that they are most likely to be working in the low-skilled sectors. Overall we see a deskilling of migrants moving from east to west in Northern Europe, a pattern also noted by Ciupijus (2011), but this is less pronounced than in the south.

Table 3.8: Occupational category of Polish migrants to Germany, the UK and the Netherlands

Mode occupational category Crafts/Related Trades (23%)/ Polish to Germany Technicians and associated professionals (20%) Professionals (18%) Polish to UK Service workers and shop and market sales workers (17%) Polish to Netherlands :: Source: Database on Immigrants in OECD countries, OECD 2000 census, own calculations; :: no data available

Finally, the data suggests that Polish immigrants in the UK and Germany are most likely to stay for more than ten years, but we do not have any data with regard to Polish immigrants in

14 the Netherlands (table 3.9)11. This implies that Northern migrants are more likely to settle for the long term than those in the South.

Table 3.9: Duration of stay of Polish migrants to Germany, the UK and the Netherlands

Up to 5 years 5 to 10 years More than 10 years Unknown Polish to Germany 1% 12% 87% Polish to UK 12% 87% Polish to Netherlands :: :: :: :: Source: Database on Immigrants in OECD countries, own calculations; :: missing data

In contrast to this data, Pollard et al., (2008) reported temporary, circular migration on the basis of their survey of Poles who had returned from England; three quarters stayed for less than a year and a fifth stayed for between one and three years. Of those who planned to return to the UK, 40% anticipated staying for less than two years. Engbersen et al. (2010), like Stanek (2009), used family reunification and presence of children in the host county to indicate how long migrants intended to stay. On this basis they suggested that because the immigration of Polish children to the Netherlands was low, the Polish were not intending to stay in the Netherlands for a long period. Overall, therefore it is difficult for us to know how long migrants will stay, and in fact it is possible that the migrants might not know themselves.

Explaining migration to northern Europe As in Romania, difficult economic conditions encouraged emigration from Poland (Steinhardt, 2009). However, the pull into the Northern hemisphere cannot be explained by linguistic similarity. German, English and Dutch do not have the same linguistic roots as Polish. Instead geographical proximity, economic difference and social networks, established by a history of regular and irregular migration, encourage migratory flows. Germany has been a traditional destination for Polish migrants (Steinhardt, 2009), even though in 2009 Germany still restricted access to its labour market for migrants from the new member states (Bonin et al., 2008). Despite this there was a very large flow of Polish migrants encouraged by the common border and by the fact that the German economy is much stronger (Pollard et al., 2008). Steinhardt (2009) used data from the Federal Labour Agency in Germany to show that many seasonal workers in Germany come from Poland. It

11 In Germany, the UK and the Netherlands the census does not cover duration of stay and so labour force surveys have been used. However the discrepancy between sources is thought to be minimal and to guarantee comparability across sources only three categories of duration of stay have been used (‘up to 5 years’, ‘5 to 10 years’ and ‘more than 10 years’).. See methodological note http://www.oecd.org/dataoecd/32/44/40136955.pdf

15 will be interesting to see, when the data becomes available, how the lifting of the restrictions for Polish migrants which happened on 1st May 2011 has affected the patterns of migration, and whether the Polish move in greater numbers to Germany as opposed to the UK because of the geographic proximity. In the UK, the main reason for the surge in Polish migrants was the lack of restrictions immediately following Poland’s accession to the EU. Only the UK, Ireland and Sweden permitted free access to their labour markets following the inclusion of the eight new member states in 2004 (Duvell, 2007; Pollard et al.2008; Bonin et al. 2008; OECD 2012). In addition, there has been a history of Polish migration to the UK since WW2 (Pollard et al., 2008; Ciupijus, 2011) and a large number of Polish workers were already working illegally in the UK prior to enlargement (Ciupijus, 2011; Duvell, 2007; Drinkwater et al. 2010). Therefore long established links and social networks were already in place to further attract Polish migrants who had gained the right of free movement (Duvell, 2007). Polish migrants interviewed as part of qualitative work carried out by Pollard et al. (2008) also highlighted that the desire to learn English was a strong pull factor and that young Polish migrants want to go to London, as it is seen as a vibrant, exciting and culturally diverse city. Unlike the UK, the Netherlands did not fully open its labour market to Poland until 2007 however there were agreements in place from the early 2000s which allowed Polish migrants to work legally in the country and by 2006 temporary work permits introduced as a transition measure after enlargement were issued relatively easily to Polish migrants. In addition there were already Polish migrants working illegally (Engbersen et al. 2010) which suggests that there may have been established social networks, again acting as a pull factor.

West to East migration: Poland Poland receives a comparatively small number of immigrants annually and it has a tiny foreign population. We can only suggest tentatively the dominant nationality of migrants entering Poland because of the discrepancies between the OECD and Eurostat figures. Both sources suggest that Germans constitute the largest immigration flow in Poland since 2001. According to the more comprehensive OECD data, in 2007, there were 6,700 immigrants, accounting for 51% of all EU inward migrants into Poland. By 2009, the number of German immigrants decreased to 1,700 representing 27% of EU immigrants to Poland but this still represented the largest flow by nationality. We can be more confident about the dominant nationality of the foreign population in Poland because we have similar data from both sources. Germans have consistently accounted for the largest share of the EU foreign population in Poland since 2001 and in 2009 between 31% (OECD) and 33% (Eurostat) of the EU foreign population were of German nationality. Whilst this represents a significant

16 share of the foreign population, overall the numbers are low, and Poland is a mass emigration rather than an immigration country. Eurostat does not provide any data about the age of Polish immigrants into Germany, but it does suggest that they are almost four times more likely to be male than female (Eurostat, International Migration and Asylum, International Migration flows, own calculations). This fits with the pattern we identified in our data of female migrants dominating the flows in Southern Europe and male migrants dominating the flows in the north. However, due to the inconsistencies between our findings and the literature we treat our data with regard to the sex of migrants with caution. Further, contrary to the pattern we have seen with regard to the Polish and Romanian immigrants moving east, German migrants moving west to Poland are almost just as likely to be educated to primary level (52%) as they are secondary or tertiary (48%) (OECD 2000 census, Database on Immigrants in OECD countries, own calculations12). This might indicate that those with lower educational qualifications are less inhibited to move from a richer to a poorer country, although more data would be needed to substantiate such a claim. Interestingly, in Poland these German migrants were most likely to be employed as ‘skilled agricultural and fishery workers’ (17%), ‘crafts and related trades’ (13%) and ‘professionals’ (13%) (OECD 2000 census, Database on Immigrants in OECD countries, own calculations13) suggesting that even though they were less likely to be highly educated, they were perhaps more likely to find skilled work that the migrants we have seen moving from east to west. There is no data as to how long German migrants stay in Poland, but we have shown that we are cautious about relying on this data, and it is in fact possible that the migrants might not know themselves. We are not surprised that we cannot find any literature to explain this movement because it is so small. We can only speculate that we see this flow because Germany is Poland’s largest EU neighbour.

Falls in annual migration flows in context We noted above the decline in annual migration flows in all of our selected countries because of the global financial crisis. Now that we know more about these flows, we can contextualise these declines. For example, where we saw sharp declines in Spain, we can attribute this at

12 The International Standard Classification of Education (ISCED; cf. UNESCO 1997) was used as a baseline, but groups have been aggregated as follows:

 Primary level: ISCED 0/1/2.  Secondary level: ISCED 3/4.  Tertiary level 1: ISCED 5A/5B.  Tertiary level 2: ISCED 6.

17 least in part to a fall in the number of Romanian males migrating to work in the Spanish construction sector because the property market in Spain has suffered considerably as a result of the recession (Stanek, 2009; Elrick and Ciobanu, 2009). At the same time, whilst most EU countries have suffered from the global crises, some eastern European countries, including Romania and Poland, have overall seen their economies improve since their accession to the EU (Pollard et al., 2008, based on 2007 Eurostat data; see also OECD, 2012:79). Improved prospects and job opportunities in Romania and Poland, set against high levels of unemployment in western EU countries including Spain, Italy, France, Germany, the UK and the Netherlands, may well have discouraged migration, especially as the chance of finding employment is known to be a key factor in determining whether an individual will choose to migrate (Pollard et al. 2008; Potot, 2010). We do not know how economies in the EU will perform and indeed compare to one another in the coming years, but the pull factor for migration from the east is likely return as the more developed western countries recover.

3.3 Migration after EU enlargement – a summary After the fall of the Berlin wall migration in Europe increased. The deep economic crisis that started in 2008 dampened migration and annual migration numbers into our chosen countries fell, particularly steeply in Spain and Italy. Yet over the same period in all of these member states except in Germany there has been an increase in the size of the EU foreign nationality population indicating that whilst annual inflows have declined, the downturn has not led to mass emigration. There are still a significant number of European citizens living within the EU but outside of their country of origin for which the mobility of pension rights will be important. In addition we have seen that EU citizens dominate these foreign nationality populations and that the EU foreign populations are increasing more quickly than the EU+OECD foreign populations which points to the increasing relative importance of intra- EU migration. Secondly, the largest migratory flows in the EU are from east to west, which is at the same time a movement from poor to rich. This finding is not new; but our identification of a latitudinal divide between flows in Northern and Southern Europe is. In the South, Romanians tend to migrate to Spain, Italy and France, while Poles are more likely to stay in the Northern hemisphere, moving to Germany, the Netherlands and the UK. We suggest that linguistic similarity is a key reason for the movement in the South, but in the North, geographical proximity draws Polish migrants to Germany and the Netherlands, and unrestricted access to the labour market immediately following Poland’s accession to the EU made the UK a particularly attractive destination. Our deliberate inclusion of Poland as an eastern European host country has shown that there is movement from rich to poor, albeit marginal in comparison to the movements from poor to rich. Overall these movements demonstrate there is a need to simulate movements between EU countries of differing

18 affluence. Thirdly, those that migrate are free and able to seek opportunities in the richer West: migrants tend to be young and they are highly educated, but they pay a price: eastern European workers in the west are likely to be over-qualified for the occupation they find employment in, especially the Romanians in Spain and Italy. The increase in migration means that the need for the protection of pension rights for workers has become more pressing. We will not deal with the question of how likely it is that further reforms on European level will take place in the near future. Instead, in our next step we will focus on what the academic literature tells us about the impact of European mobility on individual pension rights.

4. Portability of pension rights in the EU – the academic literature

Literature on the portability of pensions within the EU can be divided into the portability of public rights and of occupational rights offered by employers.

4. 1 Statutory pensions rights Substantial parts of the literature exploring the portability of public pension rights has presented these as unproblematic, assuming that the legislation described above does prevent losses on account of mobility (Schmähl 1993; Whiteford 1996; Andrietti 2001; Holzmann et al. 2005; Kalogeropoulou 2006; Forteza 2010). EC Regulations 1408/71 and 574/72 ensure that time spent working in other countries is taken into account when statutory pension claims are calculated and hold each country responsible for paying for pension rights earned in its system (Schmähl 1993:320-321; Holzmann et al. 2005:11). However, this claim has not been investigated empirically (see also Holzmann and Koettl 2011:49) and only one study has attempted to systematically explore whether features of national pension schemes might disadvantage mobile workers despite the EU regulation (Fenge and Weizsäcker 2010). These authors show that where the pension is not calculated on the basis of separate annual amounts, a mobile worker might still be penalised. This happens, for example, where the pension is based on an average annual amount using best years of earnings. Member states can share data with regard to the number of contribution years but not regarding earnings or contribution amounts in other countries and so the pension is based only on earnings within the national jurisdiction. The two calculations required by the EU regulation described above will not overcome the fact that the average salary taken into account is likely to be lower for a worker that moves prior to the end of his career and therefore the pension will be lower than for his non-mobile counterpart. The more varied the earnings history, the greater the potential losses. Countries are generally moving away from best years of earnings practices, or at least extending the amount of years taken into account so this problem will not be relevant for

19 many. However, France still bases the pension on the twenty-five best earnings years of the worker (European Commission, 2012)14). A long reference period such as this will reduce the potential losses but not prevent them entirely. The authors also highlight problems caused by caps on pensions or minimum income top-ups. Overall the design of national pension schemes can distort mobility despite the regulation, contradicting the dominant message in the literature. On the other hand, Coldron and Ackers (2009) showed that migrants might use system differences to their advantage by building up public pension rights in more generous systems and retiring in countries where prices are lower.

4.2 Occupational pension rights As shown, for almost two decades the literature generally accepted that legislation protected the statutory pension rights of mobile workers and therefore, where pension portability has been examined, the focus has been on supplementary pension rights, and in particular on defined benefit occupational pension schemes.

Portability of defined benefit schemes Employers offer supplementary pension schemes to reward employee loyalty and discourage mobility (Schmähl 1993; Whiteford, 1996). Therefore it is not surprising that schemes are designed in ways that lead to losses for early leavers. Until the early noughties, in many countries these schemes were defined benefit schemes. Kalogeroupoulo (2006) summarises the practices of such schemes which restrict the acquisition of rights and can penalise mobile workers: waiting periods before being allowed to join a scheme, a minimum age for scheme members, and vesting periods, i.e. minimum periods of pension scheme membership before being entitled to accrue rights (also Andrietti 2001). Whiteford (1996) highlights that in defined benefit schemes higher contributions made in the early years of membership allowing for a salary increase over the career of the employee can mean that an early leaver will have paid too much relative to the benefits they will receive. If the deferred pension they are granted on leaving is not then indexed linked, the value of the pension will similarly decline relative to price and wage increases. Indexation of deferred benefits is needed to decrease the losses that early leavers face. The penalties for early leavers of defined benefit schemes have been explored in more detail by Schmähl (1993) and Andrietti (2001). Both compared the retirement income of a hypothetical worker who spends his or her entire career working for one employer with an employee that moves between employers, where the defined benefit pension scheme is identical in all cases. They find that mobile workers could be penalised significantly for

20 mobility within a country even when all other scheme conditions are equal. The mobile worker could face vesting losses if he or she moves prior to the end of the vesting period, or pension annuity losses as a result of moving mid-career when the salary which will be used to calculate the pension from the first company will be lower than it would have been at the end of the career. Moves between countries are further complicated by different regulatory frameworks, scheme designs and terminology (Andrietti 2001). Moreover, whether early leavers’ accrued rights are preserved, transferable to a different scheme or can be cashed in, whether pensions transfers are free or taxable and the different statuses of migrant workers (posted workers, permanent migrants, cross-border workers or migrant workers who remain members of schemes in their country of origin) must also be taken into account when analysing pension right portability (Andrietti 2001). A more recent study of the Caribbean came to very similar results as those just described, presenting more evidence for how early leavers of defined benefit pension schemes are penalised in comparison with non-mobile workers (Forteza 2010).

Portability of defined contribution schemes In comparison, defined contribution schemes are more portable than defined benefit schemes. This is because after a vesting period early leavers can normally withdraw the actuarially fair value of their pension funds (Andrietti 2001). However, if an early leaver has only accrued few pension rights in a defined contribution scheme and if the earnings of the pension fund are not shared equally between active and former employees then, due to inflation, the value of the fund he or she has to invest in an annuity in retirement will be relatively reduced (Whiteford 1996). The treatment of dormant rights of early leavers in these schemes will determine whether the mobile worker will face losses. Pension schemes are moving towards defined contribution schemes but whilst they are more portable they are at the same time less generous than defined benefit schemes, thereby disadvantaging all workers.

4. 3 Impact of mobility on individual rights: a summary So far, research has focused on how the design of schemes, public or private, might disadvantage a mobile worker and it demonstrates clearly how certain scheme principles penalise mobility. However, the literature has generally but prematurely assumed that existing EU regulation protects the public pension rights of mobile workers, and an overview of the national legislation for early leavers of non-statutory schemes is missing, i.e. we know too little about how this is regulated and how rights are worked out in practice in each country. Consequently, a comparative overview is missing as well. Moreover, none of the studies summarized above review the impact of mobility on both public and occupational rights taken together.

Very little work has been done on another important aspect of mobility: the generosity of different systems (but see Coldron and Ackers 2009). The quality of the protection of rights is not the only criterion that decides whether migrants are penalised by moving. Let us imagine, for example, that the benefit amount accrued by an employee in Poland is lower than the amount the worker could have earned with the same wage and length of employment in Spain. If this employee had worked in Poland for a decade they might find that their overall pension after thirty further years in Spanish employment is too low to retire in Spain, simply because the Polish entitlements disadvantage her or him in comparison with a Spanish worker on similar lifetime wages. By the same token, she or he might find that they would live more comfortably than their fellow countrymen if they returned to Poland in retirement. In this example the differences are explained by the different generosities of the Polish and Spanish pension systems. However, in another case, possible pension differences between a migrant and non-migrant worker of otherwise similar careers might be explained by the fact that the migrant is exposed to different insurance principles dominant in the countries where they work – for example a citizenship based flat-rate system and a wage-replacement based insurance system. The non-migrant would only be affected by one of these principles. In summary when assessing the extent to which EU mobility is possible, and whether or not workers are penalised by it, we need to know more about how rights are protected, but we also have to pay attention to the question of how much workers are entitled to. Against this background our project will analyse regulation, but also the insurance principles and generosity of different pension regimes in Europe. In the next part we summarise the two types of public-occupational pension regime that exist in the EU and we ask what impact the different organising principles and generosities have on the entitlements of mobile workers with differing income levels.

5. Beveridgean and Bismarckian pension regimes – principles and outcomes

In the comparative literature on pension regimes in Europe it has become common to divide European countries into two types of pension regime: Beveridgean and Bismarckian models (eg Bonoli 2003; Ebbinghaus et al 2011). These classifications refer to the organisation of the public pension and the non-state, occupational systems. As we will show below, they also refer to the institutions that govern pensions. We examine pension outcomes for workers in the two systems, showing that on this measure, the regime types are not so different.

5.1 Beveridgean pension regimes – main principles The key feature of a Beveridgean pension regime is that it has a broad base, normally a public pension, which is universal and flat rate. This means that everyone or a very large share of the population is entitled to it, but benefits are not income-related, instead they are fixed at the

22 same level for all. This level can be at or above the poverty line to protect citizens from hardship. Poverty preventing universal pensions have long existed in most Beveridgean countries of Europe: Sweden, Norway, Denmark, Finland, the Netherlands and Switzerland, which base the public pension on citizenship. However, the level can also be below the poverty line, which was the case in the UK from 1945 to 2012. Universal, flat rate pensions do not provide sufficient benefits to those on average or above average incomes, no matter whether their level is at or below the poverty line. Higher earners expect to preserve their living standard in retirement and this is why we see strong occupational benefits in Beveridgean countries. These are income related and their function is to increase the replacement rates of pensions to earnings in retirement. In most Beveridgean countries occupational schemes were defined benefit schemes, but reforms since the mid- 1990s have changed them to career average or defined contribution schemes. The benefits are still income-related, but much less predictable for the individual, because their final outcome is now tied to market dynamics and in many cases to the pace of demographic change. In all Beveridgean countries except Britain occupational pensions have been a mandatory part of the pension regime, coverage was therefore high (Clasen et al. 2011: 292-3). In the UK voluntarism prevailed until 2012, and therefore this is the Beveridgean country with the lowest occupational coverage, never more than half of the working population in the private sector.

5.2 Bismarckian pension regimes – main principles The key feature of a Bismarckian pension regime is that public contributions and benefits are earnings-related. Only those who have paid contributions are entitled to a public pension and if the career has been short or earnings low pension entitlements will be low as well. This means that citizens who work part-time or intermittently will receive a lower pension. Consequently, women are more vulnerable in a Bismarckian system than men, even though in most Bismarckian countries care-related rights have been introduced into the public system. Because of the generosity of state pensions for full-time workers there is less need for them to invest in additional occupational benefits and therefore these are less developed in Bismarckian than in Beveridgean countries, even though we do see an occupational pension sector in Germany and Austria, too. France, Italy, Poland, and Belgium also have Bismarckian regimes.

5.3 Comparing the principles of Beveridgean and Bismarckian regimes The institutional structure of both systems is quite different. In the Beveridgean world the universal first tier, the flat rate pension is fully public; occupational pensions are mandatory, but they might be administered by employers or/and trade unions, or other non-state

23 institutions. In the Bismarckian world, in contrast, the full income-related pension is public, i.e. in contrast to the Beveridgean system there is only one tier governed by one set of rules. This difference also makes the calculation of mobile workers’s statutory rights less complex in Bismarckian systems than in Beveridgean. If we examine pension outcomes, however, the institutional differences lose their importance for higher and average earners and instead, it matters more how generous the entitlements are of public and mandatory occupational pensions combined in the first and of public pensions in the second group of countries. Benefits will be income related in both cases, but replacement rates can be quite different, as the next section will show. For lower earners however, institutional differences still matter. This is because the Beveridgean public tier is more redistributive and designed to address poverty risks. Lower earners are more likely to benefit from this system design than those on average or higher wages. Again, the next section will illustrate this point.

5.4 Portability of entitlements in Beveridgean and Bismarckian regimes The systems just characterised emerged and evolved in national contexts, i.e. they were shaped by national political interests, and reformed with citizens of nation states in mind. In contrast, our project focuses on citizens who work in one country and then move to employment in another – and perhaps move back again. They therefore have accrued pension rights under two different sets of rules. As we will show below, by moving it is possible that a person will leave a Bismarckian system and move to a Beveridgean one or the other way around. They might also move between two countries belonging to the same type of pension regime. Considering these systems were not meant to enhance such mobility, how do their principles affect it? Assuming that EU mobility regulations outlined above work as they should do, the answer differs according to income levels. In short, for those on average or higher incomes regime type does not matter because variation within each regime is high for such workers. Instead, the level of replacement, i.e. benefit generosity is more significant (see also Bridgen/Meyer 2011: 194; Myles 1998). In contrast, those on lower incomes fare much better in Beveridgean systems. Thus, they will benefit by moving from Bismarckian to Beveridgean countries. We have illustrated this situation in table 5.1 and graph 5.1. The table shows the pension outcomes of four hypothetical workers, two earn 2500, two 1500 Euro monthly throughout their working lives. We show their pensions if they work for 45 years, but also if all work for ten years, and if they work for 20 years, 10 of those in a Beveridgean, 10 in a Bismarckian country, in comparison with the 20 years in the same country. For their replacement and contribution rates we have calculated the average rates for the two groups of countries as provided by the OECD for 2009 (2011: p.121, 153, see tables 5. 2 and 5.3, appendix). The

Beveridgean countries on average offered a net replacement rate for compulsory pensions of 86.2% of last wages for lower and one of 62% for higher incomes. The Bismarckian countries grant replacement rates worth 68.3% of last wages after a full employment career for a lower earner and 60.9% for a higher earner. Thus the Bismarckian system is less generous on average than the Beveridgean, and even though it is less redistributive than the Beveridgean regime, some degree of redistribution can be seen there as well, possibly due to measures that dilute the strict insurance principle. For these lower outcomes, the Bismarckian workers also pay slightly lower average contributions than the Beveridgean workers.15 How do these results change if we differentiate by income groups? Table 5.1 shows that the worker with the higher income is only very slightly better off on average in the Beveridgean system. She receives a pension worth two per cent more of what she would be entitled to in the Bismarckian system, but her pension as a share of her overall contributions is slightly lower than it would be in a Bismarckian country. In contrast, the worker with the lower income is much better off in a Beveridgean regime. Her average pension in the Beveridgean country comes to 144% of its value in the Bismarckian country and her pension share of contributions is higher in the former than in the latter. This means that the best position of the lower earner is to stay in Beveridgean countries, but if she works in a Bismarckian one, she would gain by moving to a Beveridgean. Regarding the higher earner, do the system differences outlined above mean that a mobile worker on a higher income from a Bismarckian country could move wherever she likes, if protection of state pension rights is her main concern? The answer is no. While the Bismarckian and Beveridgean average outcomes are similar regarding these workers, they conceal substantial variation within each country group (graph 3.1). This means that this worker can afford to ignore the dominant insurance principle of a country, but they would be well advised to make their choice dependent on the national level of replacement. Thus, a migrant from Bismarckian Poland or Germany could be better off by moving, for example to the Beveridgean Netherlands or Denmark, but also by migrating to Bismarckian Spain or Austria because these countries offer much more generous replacement rates. Within group variation also means that while lower income workers are well advised to stay within the Beveridgean group from a pension perspective, they could still improve their situation by moving to the most generous countries of that group, namely the Netherlands or Denmark.

15 The main variation regarding contributions between both groups is caused by employers, for whom contributions in Bismarckian countries are much higher (14.1% on average) than in Beveridgean (5.6% on average – see table 5.3, appendix).

Table 5.1: Portability of flat-rate and earnings-related pension rights, 2009 Bis Bev Bis Bev Monthly income EURO 2500 2500 1500 1500 Contribution rate % of earnings, employee average 8.3 9.9 8.3 9.9 Contribution years 10 10 10 10 Pension after 45 years, net replacement 1522.5 1550 900 1293 Monthly pension after 10 years 338 344 200 287 Total contributions after 10 years 24900 29700 14940 17820 Monthly pension as % of total contributions after 10 years 1.36 1.16 1.34 1.61 Beveridge monthly pension as % of Bismarck , 10 years 102 144

Total contributions after 45 years 112050 133650 67230 80190 Monthly pension as % of total contributions after 45 years 1.36 1.16 1.34 1.61 Beveridge monthly pension as % of Bismarck , 45 years 102 144

Bev to Bis to Solid Solid Monthly pension after 20 years, 10 in each regime Bis Bev Bis Bev Higher Earners 683 683 677 689 Lower Earners 487 487 400 575 High Earners - % change in pension because of movement 99 101 Low Earners - % change in pension because of movement 85 122 Bis = Bismarckian, Bev = Beveridgean. Own calculations and OECD 2011

Graph 5.1: Beveridgean and Bismarckian countries grouped according to net replacement of low wage earners, 2009 140

120

118.2

100 94.8 88.6 86.8 87.4 85 80 75.75 Bis. average low performers: Pol, 72.1 Ger

64.5 Bev. average low UK, Sw 62.2 60 61.4 55.8 Bis. average medium: Fr, It, Belg 49.7 47.7 44 44.645.5 44.45

40 Bev .average med No, CH Average Average replacementcompulsoryrate system

Bis. average high: Sp, Aus

20 Bev. average high: NL, Den

0 1 Average wages of worker

Bis = Bismarckian, Bev = Beveridgean. Own calculations and OECD 2011

Considering that many migrants in Europe work in less qualified jobs, on lower wages (see above), the Netherlands, Britain and the Nordic countries would offer them relatively better compulsory pensions than Austria, Germany, Italy and other Bismarckian countries, and within this group Denmark and the Netherlands are best, while the UK and Sweden are worst. However, to take advantage of this situation, these workers would need to move at a relatively young age, otherwise they might run into the problem identified at the end of the previous section: i.e. that they would miss too great a period of rights accrual in their adopted country for which the less generous rights built up in the home country would be insufficient as compensation. In these circumstances, such workers could be at risk of relative poverty should they decide to remain in their adopted country.

For workers on higher wages, additional non-statutory benefits and their protection are vital if they live in or move between countries granting low mandatory replacement rates, no matter whether these are Bismarckian – for example Belgium, France, Germany, Poland, or Beveridgean such as Norway, Sweden, Switzerland or the UK (graph 5.1, table 5.1). In general mobile workers on higher wages receive relatively lower benefits than those on lower wages and therefore need voluntary occupational benefits more than lower waged workers if they want to protect their standard of living in retirement. Thus, the reliability of long-term protection of rights acquired with one employer is very important. Mobility in national labour markets has long been a challenge. Ever since occupational pension schemes expanded national policy-makers recognised that unless rights of early leavers were protected, occupational pension schemes would add to labour market rigidity because workers would not want to move employers and lose their entitlements. To see how well protected mobile workers are, it is therefore important to consider the national level as well as the question of whether rights are portable to another country within the EU. In any case, the quality of regulation of occupational schemes is very important for mobility in and from both types of regimes. Above, we have established that type of pension regime and generosity levels matter for migrants. However, pension regime type is not the main driver of real migration flows; indeed it is doubtful whether this is a factor at all when mobile workers decide where to go. Indeed our analysis in part three shows that low waged workers are moving in large numbers to Bismarckian Italy, Spain and Germany or the UK which has been identified as one of the least generous Beveridgean countries and on this basis it does not seem that the design or generosity of the pension system plays a part in the choice of destination. In fact, young migrant workers may well not be thinking about their pension outcomes at all.

6. Significance of this paper for our next steps

The aim of this paper has been to explore how mobile pension rights are within the EU and whether workers will be penalised for their mobility when they retire. We have shown that European legislation protects workers moving between public pension systems to some extent, although not entirely, and that occupational pension rights are regulated only at national level, exposing workers to the distortions that affect early leavers of schemes. For a long time this deficiency mattered less in practice because relatively few European citizens moved between countries. However, this has changed since the enlargement of the EU which has encouraged large numbers of migrants to move from east to west. On the academic side little work has been done in this area, and there has been no attempt to explore holistically the public-private pension outcomes of those who move within the EU.

Academics have examined the portability of public pension rights on the one hand, albeit to a limited extend, and more have looked at how migration might impact on occupational pension outcomes, but these calculations have been independent of each other, which is somewhat surprising given how intertwined public-pension rights are. In addition, there is no comparative overview of the national occupational rights of mobile workers, i.e. of early leavers or late joiners of schemes. Thus, more comprehensive assessments of how migrant workers fare, taking public and occupational benefits into account are needed if we want to know how ‘free’ movement within the EU is in practice. We also saw that different types of pension regime and generosity levels have rarely been part of research on mobility. Our research in this area shows that potential income losses occur when workers move not only between different types of pension systems but also between schemes of differing generosity. This demonstrates that the issue of the impact of migration on pension rights is about more than just regulation; regime differences and benefit levels matter too. Our next step will be to simulate pension outcomes based on hypothetical migrant trajectories in the EU. These simulations can be informed by what we now know about migratory movements over the last decade and pension regimes/generosities. We have shown above that many migrants move from poor, Bismarckian to richer countries that are either Bismarckian or Beveridgean. It seems sensible to include Poland and Romania as countries of origin as nationals from these countries have been moving in large numbers, and the UK and the Netherlands as host countries as these are among those with the highest shares of foreign nationals but which also represent the Beveridgean model. In addition, at least a Southern and a Continental European Bismarckian country should be included, Germany and Spain seem sensible because their share of foreign nationals is also high and because the generosity of their pension systems differ. Regarding the type of worker, the migration statistics suggest that our typical biographies could be male or female and that they tend to be young. They are likely to be well-educated in the east, but they are unlikely to find jobs that match their skill level and often they become low-wage workers in the west. Our preliminary calculations in part three suggest that those moving from the east to the Beveridgean west, i.e. to the UK, and the Netherlands will be better protected than those migrating to the Bismarckian countries such as Germany, France or Southern Europe. Those moving from the East to the Bismarckian countries on the continent or in Southern Europe are likely to be poorer in comparison with those moving North, but the calculation of their pension rights is going to be more straightforward considering the one-tiered nature of the Bismarckian statutory system. Nevertheless, because their entitlements are likely to be lower, in particular if they move to Germany, voluntary supplementary rights will (in theory) be very important for them. Therefore they may well

29 need to rely on national legislation not harmonised by the EU in which case it will be necessary for us to know how short term members and early leavers of supplementary pension schemes are treated under such law. Finally the question is still open for us whether to rely exclusively on the dominant migration streams when constructing our typical migrant biographies. For example, qualified workers moving within the richer parts of Europe might not be quantitatively dominant; however, they constitute a certain type of migrant, and are relevant for European knowledge economies. This is why they have already attracted the attention of researchers as we discussed in part two above. Moreover, an integrated European economy presupposes free labour mobility between all parts of the EU, rich to poorer as well as the reverse, so it is important to assess the impact of this type of migration even if it does not currently constitute a dominant stream. Our data has shown that intra EU migrants tend to be young but we could simulate different trajectories: a young migrant accruing pension rights in the host country and later migrating back to the country of origin, and an older worker migrating later during his or her career and retiring in the destination country. The pension outcomes of these workers cannot be judged apart from the relative costs of living in the countries in which they choose to retire.

References ACKERS, L. & OLIVER, L. 2008. Scientific Mobility and Pensions. Report to the Association of Certified and Chartered Accountants (ACCA). The Liverpool Law School, University of Liverpool: European Law and Policy Research Group.

AHONEN, E. Q., PORTHÉ, V., VÁZQUEZ, M. L., GARCÍA, A. M., LÓPEZ-JACOB, M. J., RUIZ-FRUTOS, C., RONDA-PÉREZ, E., BENACH, J. & BENAVIDES, F. G. 2009. A qualitative study about immigrant workers’ perceptions of their working conditions in Spain. Journal of Epidemiology and Community Health, 63, 936-942.

ALEXANDRU, M. 2007. Migration and Social Mobility. A New Perspective on Status Inconsistency. The Romanian Journal of European Studies, 5-6.

ANDRIETTI, V. 2001. Portability of Supplementary Pension Rights in the European Union. International Social Security Review, 54, 59-83.

ANDRIETTI, V. 2004. Pension choices and job mobility in the UK. Economics Working Papers. Universidad Carlos III, Madrid.

ARANGO, J. & JACHIMOWICZ, M. 2005. Regularizing Immigrants in Spain: A New Approach. Migration Policy Institute [Online], 1 September 2005. Available: http://www.migrationinformation.org/Feature/display.cfm?ID=331 [Accessed 10th May 2012].

BAN, C. 2009. Economic Transnationalism and its Ambiguities: The Case of Romanian Migration to Italy. International Migration, 1-31.

BLACK, R., ENGBERSEN, G., OKOLSKI, M. & PANTIRU, C. (eds.) 2010. A continent moving west? : EU enlargement and labour migration from Central and Eastern Europe. Amsterdam: Manchester: Amsterdam University Press; Manchester University Press distributor.

BONIN, H., EICHHORST, W., FLORMAN, C., HANSEN, M. O., SKIÖLD, L., STUHLER, J., TATSIRAMOS, K., THOMASEN, H. & ZIMMERMANN, K. F. 2008. Geographic Mobility in the European Union: Optimising its Economic and Social Benefits. IZA Research Report No. 19.

BONOLI, G. 2003. Two worlds of pension reform in Western Europe. Comparative politics 35(4): 399-416.

BRIDGEN, P. & MEYER, T. 2011. The impact of the new public and private pension settlements on Britain and Germany on citizens' income in old age. In: J. Clasen (ed.): Converging Worlds of Welfare? British and German Social Policy in the 21st Century. Oxford, Oxford University Press: 180-217.

CIUPIJUS, Z. 2011. Mobile central eastern Europeans in Britain: successful European Union citizens and disadvantaged labour migrants? Work, Employment & Society, 25, 540-550.

CLASEN, J., MAU, S., MEYER, T. & SEELEIB-KAISER, M. 2011 Conclusion: parallel paths, great similarities, remaining differences. In: CLASEN, J. (ed.): Converging Worlds of Welfare? British and German Social Policy in the 21st Century. Oxford, Oxford University Press: 180-217.

COLDRON, K. & ACKERS, L. 2009. European citizenship, individual agency and the challenge to social welfare systems: a case study of retirement migration in the European Union. Policy and Politics, 37, 573-589.

COUNCIL DIRECTIVE 98/49/EC of 29 June 1998 on safeguarding the supplementary pension rights of employed and self-employed persons moving within the Community

COUNCIL REGULATION (EC) No 1408/71 of 14 June 1971 on the Application of Social Security Schemes to Employed Persons, to Self-Employed Persons and to Members of their Families Moving Within The Community, OJ L 28, 30 January 1997

COUNCIL REGULATION (EC) No 883/2004 of the European Parliament and of the Council of 29 April 2004 on the Coordination of Social Security Systems, OJ L 200, 7 June 2004

CULIC, I. 2008. Eluding Exit and Entry Controls: Romanian and Moldovan Immigrants in the European Union. East European Politics & Societies, 22, 145-170.

DISNEY, R. & EMMERSON, C. 2002. Choice of pension scheme and job mobility in Britain. Institute for Fiscal Studies.

DRINKWATER, S., EADE, J. & GARAPICH, M. 2010. What's behind the figures? An investigation into recent Polish migration to the UK. In: BLACK, R., ENGBERSEN, G., OKOLSKI, M. & PANTIRU, C. (eds.) A continent moving west? : EU enlargement and labour migration from Central and Eastern Europe. Amsterdam: Manchester: Amsterdam University Press; Manchester University Press distributor.

DUVELL, F. 2007. EU Accession of Bulgaria and Romania: Migration Issues Oxford: COMPASS Available: http://www.compas.ox.ac.uk/fileadmin/files/Publications/Briefings/B- 07-16_Bulg-Rom-Briefing.pdf [Accessed 9th May 2012].

EBBINGHAUS, B. (ed.) 2011. Varieties of pension governance: The privatization of pensions in Europe. Oxford, Oxford University Press

ELRICK, T. I. M. & CIOBANU, O. 2009. Migration networks and policy impacts: insights from Romanian–Spanish migrations. Global Networks, 9, 100-116.

ENGBERSEN, G., SNEL, E. & DE BOOM, J. 2010. 'A van full of Poles': Liquid migration from Central and Eastern Europe. In: BLACK, R., ENGBERSEN, G., OKOLSKI, M. & PANTIRU, C. (eds.) A continent moving west? : EU enlargement and labour migration from Central and Eastern Europe. Amsterdam Manchester: Amsterdam University Press; Manchester University Press distributor.

EUROPEAN COMMISSION 2005. Proposal for a Directive of the European Parliament and of the Council on improving the portability of supplementary pension rights, COM (2005) 507 final, Brussels, 20 October

EUROPEAN COMMISSION. 2008. Evidence on the main factors inhibiting mobility and career development of researchers. Final report, Contract DG-RTD-2005-M-02-01. Luxembourg: Office for Official Publications of the European Communities.

EUROPEAN COMMISSION, 2010. Social security, supplementary pensions and new patterns of work and mobility: researchers' profiles Luxembourg: Office for Official Publications of the European Communities.

EUROPEAN COMMISSION, 2011. Notification by the French Government to the European Parliament and to the Council. Official Journal of the European Union, C 135/17, 5.5.2011, p.1.

EUROPEAN COMMISSION, 2012. Mutual Information System on Social Protection (MISSOC), Comparative Tables on Social Protection. Available at

FENGE, R. & VON WEIZSÄCKER, J. 2010. Public pension systems and distortions of intra- EU mobility: the Lodge Test. Journal of Pension Economics & Finance, 9, 263-275.

FORTEZA, A. 2010. The Portability of Pension Rights: General Principles and the Caribbean Case. Development Policy Review, 28, 237-255.

HEWITT ASSOCIATES. 2007. Quantitative overview on supplementary pension provision. Brussels: Hewitt Associates.

HOLZMANN, R., KOETTL, J. & CHERNETSKY, T. 2005. Portability Regimes of Pension and Health Care Benefits for International Migrants: An Analysis of Issues and Good Practices. Geneva: Global Commission on International Migration.

HOLZMANN, R. & KOETTL, J. 2011. Portability of Pension, Health, and other Social Benefits: Facts, Concepts, Issues. Social Protection Discussion Papers 62725 [Online]. Available: http://ideas.repec.org/p/wbk/hdnspu/62725.html [Accessed 6 December 2011].

KALOGEROPOULOU, K. 2006. Improving the Portability of Supplementary Pension Rights. Journal of Social Welfare and Family Law, 28, 95-104.

KUPISZEWSKA, D. & NOWOK, B. 2005 Comparability of statistics on international migration flows in the European Union. Central European Forum for migration research CEFMR working pair 7/2005, Wasssay: CEFMR

LLUBERAS, R. 2008. The Effect of Pensions on Job Mobility: Empirical Evidence for the UK. Towers Watson Technical Paper No. 1103825.

MCCORMICK, B. & HUGHES, G. 1984. The Influence of Pensions on Job Mobility. Journal of Public Economics, 23, 183-206.

MYLES J. 1998. How to design a ‘liberal’ welfare state. A comparison of the USA and Canada, Social Policy and Administration, 32 (4): 341-64

OECD 2011. Pensions at a Glance 2011. Retirement income systems in OECD and G20 countries. OECD: Paris.

OECD. 2012. Free Movement of Workers and Labour Market Adjustment: Recent Experiences from OECD Countries and the European Union [Online]. OECD Publishing. Available: http://dx.doi.org/10.1787/9789264177185-en [Accessed 12th July 2012].

OLIVER, E. 2009. From portability to acquisition and preservation: the challenge of legislating in the area of supplementary pensions. Journal of Social Welfare and Family Law, 31, 173-183.

POLLARD, N., LATORRE, M. & SRISKANDARAJAH, D. 2008. Floodgates or Turnstiles? Post-EU Enlargement Migration Flows to (and from) the UK. Institute for Public Policy Research: London.

POTOT, S. 2010. Transitioning strategies of economic survival: Romanian migration during the transition process. In: BLACK, R., ENGBERSEN, G., OKOLSKI, M. & PANTIRU, C. (eds.) A continent moving west? : EU enlargement and labour migration from Central and Eastern Europe. Amsterdam: Manchester: Amsterdam University Press; Manchester University Press distributor.

PUHANI, P. A. 2001. Labour Mobility: An Adjustment Mechanism in Euroland? Empirical Evidence for Western Germany, France and Italy. German Economic Review, 2, 127-140.

SCHMÄHL, W. 1993. Harmonization of Pension Schemes in Europe? A controversial issue in light of economics. In: REIN, M. & ATKINSON, A. B. (eds.) Age, work and social security Basingstoke: Macmillan.

STAN, S. 2009. Romanian Migration to Spain and Its Impact on the Romanian Labour Market. Working Paper No. 14 November 2009. Universitat Autònoma de Barcelona: QUIT Centre Sociològics sobre la Vida Quitidiana i el Treball.

STANEK, M. 2009. Patterns of Romanian and Bulgarian Migration to Spain. Europe-Asia Studies, 61, 1627-1644.

STEINHARDT, M. F. 2009. The Dimensions and Effects of EU Labour Migration in Germany. In: GALGÓCZI, B., LESCHKE, J. & WATT, A. (eds.) EU Labour Migration Since Enlargement : Trends, Impacts and Policies Abingdon, Oxon, , GBR: Ashgate Publishing Group.

STRAUBHAAR, T. 1988. International Labour Migration within a Common Market: Some Aspects of EC Experience. Journal of Common Market Studies 27, 45-62

TREATY ESTABLISHING THE EUROPEAN ECONOMIC COMMUNITY (Rome, 25 March 1957, T.S. 1 (1973): Cmnd 5179) (EC Treaty)

TSOUKALIS, T. 1997. The New European Economy revisited. Oxford: Oxford University Press

UCCELLINI, C. M. 2010. 'Outsiders' after Accession: The case of Romanian migrants in Italy, 1989-2009. Political Perspectives, 4, 70-85.

VAN DALEN, H. P. & HENKENS, K. 2012. Explaining low international labour mobility: the role of networks, personality, and perceived labour market opportunities. Population, Space and Place, 18, 31-44.

VERSCHUEREN, H. (2009) ‘Regulation 883/2004 and Invalidity and Old-Age Pensions’ European Journal of Social Security, 11(1-2): 143-162.

WHITEFORD, E. A. 1996. Adapting to change : occupational pension schemes, women and migrant workers : an examination of the extent to which occupational pension schemes in the UK, the Netherlands and Germany enable women and migrant workers to accrue adequate pensions, The Hague; London, Kluwer Law International.

Appendix 1 - There are differences in how countries collect data on migration: population registers (Belgium, Germany, the Netherlands, Sweden, the Czech Republic, Hungary, Slovenia, France and Portugal (latter two immigration only)), sample surveys (Cyprus, Ireland, UK, Portugal (emigration only)), statistical forms (Poland, Slovakia, Malta)

2 - We included data on migration from developed industrialised democracies outside the EU as well because we wanted to gain a broad sense of the political and administrative pressures on each state to deal with immigration issues from all developed countries, whose migrants are more likely to acquire pension rights than those from less developed countries. Not all EU countries are within the OECD, but we include all EU-27 countries and then add the additional OECD countries. Therefore in our EU+OECD analyses we include 40 countries in total (EU-27 plus Australia, Canada, Chile, Ireland, Israel, Japan, Korea, Mexico, New Zealand, Norway, Switzerland, Turkey and the United States). This data is only available from the OECD, and where the Eurostat and OECD data for EU migration (annual flows/ size of foreign population) are similar we can be confident about relying solely on the OECD figures to gain a sense of the patterns with regard to EU+OECD migration. However we are not always able to do this because of discrepancies between the sources.

3 - We used the 2000 OECD census data to select our countries. It is a snapshot of foreign born immigrants within different European member states during the year 2000, not a count of migrants entering the country in that year. In the 2000 OECD census data all EU-27 countries are included in ‘country of birth’, but Bulgaria, Cyprus, Estonia, Latvia, Lithuania, Malta, Romania and Slovenia are excluded from ‘country of residence’. These excluded countries are small and it is not thought that they would have recorded higher numbers of immigrants than the five countries selected for comparison. Where the census data is used, immigrants are categorised by ‘country of birth’ with all citizenship types included. The OECD census data looks at the foreign born only and does not provide data by foreign nationality. The data shows that in 2000 Germany reported the largest number of intra-EU immigrants, followed by France, the UK, Italy and Spain. Out of the new EU members the Czech Republic received the largest number of immigrants even in 2000, but these predominantly came from the Slovak Republic as both countries were formerly part of

Czechoslovakia. As this is an atypical situation we chose instead to include Poland which recorded the next highest number of immigrants. The majority of countries chosen above - Germany, Spain, France, Italy and Poland - represent Bismarckian regime types and only one, the UK, is based on the Beveridgean model. Therefore we added the Netherlands to our selection as another example of a Beveridgean regime. According to the census data, the Netherlands received just slightly more immigrants than Poland in 2000.

4- The OECD provides data on the socio-demographic characteristics of migrants, but it is drawn from the 2000 census so it is a snapshot of the immigrants in the country at that point in time and does not provide information about the migrants arriving each year. This is especially problematic for the age variable; comparing the OECD census data and the Eurostat annual data on the age of immigrants, the OECD suggests immigrants are much older. However this can be attributed to the data as the census will include those who have migrated at an earlier age and grown older in the host country. The Eurostat data is preferable because it provides annual records of immigrants, but only for the age and sex variables. Therefore, we use Eurostat data for information as to the age and sex of migrants, and the OECD data to provide an indication of their level of education (home country) and occupation (host country) and their duration of stay. With regard to the latter variables, we acknowledge that we are using data that is over ten years old and it may not be reliable; immigrants may have changed jobs in between arriving in the host country and the time of the census for example. We keep these considerations in mind and use this data only to give us an insight into typical migrant biographies. Where possible, we refer to the literature and compare the patterns we find.

Table 5.2

Net pension replacement rates of public pension & mandatory private where existent Income as % of average 0.5 1 1.5 Bismarck Austria 91.3 89.9 84.6 Belgium 74.9 52.1 42.5 France 69.4 60.4 53.1 Germany 54.8 56 55.6 Italy 72 71.7 71.8 Poland 33.2 33.2 33.3 Spain 82.3 84.9 85.4 Average 68.3 64.0 60.9 Std dev 17.7 18.3 18.8

Beveridge Denmark 131.9 89.8 80.8 Netherlands 104.5 99.8 96.4 Norway 72.9 60.3 49.2 Sweden 67 53.6 72.6 Switzerland 78.6 64.1 46.2 United Kingdom 62 37.4 26.8 Average 86.2 67.5 62.0 Std dev 24.5 21.2 23.5

Bismarck as % of Beveridge 79 95 98 Std dev Bismarck as % of Bev. 72.2 86.0 80.1 OECD 2011, Pensions at a Glance: 121, 153. Italics: Public and mandatory private

Table 5.3

Public pension contribution rates and revenues, 2009

Employee Employer Total Bismarck Austria 10.3 12.6 22.9 Belgium 7.5 8.9 16.4 France 6.8 9.9 16.7 Germany 10 10 20.0 Italy 9.2 23.8 33.0 Poland 9.8 9.8 19.6 Spain 4.7 23.6 28.3 Average 8.3 14.1 22.4 Std dev 1.9 6.2 5.7 Beveridge Denmark 0 n.a. Netherlands 17.9 0 17.9 Norway n.a. Sweden 7 11.9 18.9 Switzerland 4.9 4.9 9.8 United Kingdom n.a. Average 9.9 5.6 15.5 Std dev 5.7 4.9 7.8

Bismarck as % of Beveridge 84 252 144 Std dev Bismarck as % of Bev. 33.7 126.4 73.5 OECD 2011, Pensions at a Glance: 121, 153. Italics: Public and mandatory private