Introduction
Many contemporary proposals for welfare state reform include raising the retirement age and promoting the wider participation of older people in volunteering. Such proposals refer not to all older people but, rather, target those who are past retirement age but healthy and without a legal responsibility for a dependant in need of care (cf. Laslett Reference Laslett1996). This group has been called the young old (cf. Neugarten Reference Neugarten1974). Their paradoxical name captures their specificity: the combination of characteristics of young and older people. Like young adults, they are capable of carrying out everyday activities. Like older persons, on the other hand, they are not obliged or expected to be involved in paid and unpaid work. For the welfare state, the ‘young old’ constitute a resource whose activities are hard to steer.
The ‘young old’ rose in importance when reform strategies were developed to counter the perceived ‘crisis’ of the welfare state. One basic idea was to activate yet unused resources for welfare production, namely the young old. Early strategies assumed that the welfare production of individuals was synonymous with their participation in paid work and them being young adults or middle-aged (cf. Townsend Reference Townsend1981). With retirement, people thus stopped being productive and were considered as old. An increasingly favoured approach to increase welfare production has consequently been to delay retirement, e.g. through a higher retirement age and lower pensions for early retirees. More recent reform strategies, however, have emphasised the contribution of volunteering and informal care to welfare production (e.g. Evers Reference Evers, Evers and Svetlik1993; Orloff Reference Orloff1993). As both activities can also be carried out after retirement, the label old no longer separates unproductive from productive citizens, but rather denotes the absence of an obligation for welfare production: viz. retirement from the labour market and the absence of a legal responsibility for dependent persons in need of care, e.g. one's children, spouse or parents. Approaches to increase welfare production consequently turned to strengthening the obligation to be involved in informal care, e.g. in the framework of long-term care insurance, and to increase older persons' involvement in volunteering. While the first approach aimed at decreasing the number of young old persons, the second suggested a more positive view of the young old, as a resource for instead of a burden on the welfare state (Carr Reference Carr2005; Warnes Reference Warnes1993).
Not only are the young old of substantial interest to the welfare state, their existence inextricably derives from it. It was only with the public health measures and health-care systems adopted in developed welfare states that average ‘healthy life expectancy’ increased considerably. Likewise, it was only with the introduction of old-age pensions, disability and unemployment benefits that people could cease paid work at increasingly younger ages (Guillemard Reference Guillemard, Johnson, Conrad and Thomson1989). In the case of continental European countries, early retirement was even encouraged by the welfare state to counterbalance jobless growth (Bussemaker and van Kersbergen Reference Bussemaker, van Kersbergen and Sainsbury1999). Both developments helped distinguish the young old as a socio-demographic group. Their existence, however, has put pressure on the welfare state, above all on pay-as-you-go financed pension schemes (Organisation for Economic Co-operation and Development (OECD) 2006). The relationship between the young old and the welfare state can therefore be characterised as a feedback loop. Put another way, because of the reciprocal influence, associations between characteristics of the welfare state and the number of young old persons can be expected; the relative number of young old persons is a characteristic of welfare states.
This paper develops the idea that the welfare state influences the number of young old people. There is a growing body of literature on the young old, which generally recognises their relationship to the welfare state (e.g. Carr Reference Carr2005; Gilleard and Higgs Reference Gilleard and Higgs2002), but empirical studies have not yet examined the influence of the welfare state on their relative number. Instead, most have investigated either country-differences in older persons' health status and involvement in informal care, or the influence of retirement regulations on the number of retirees (e.g. Ebbinghaus Reference Ebbinghaus, Scharpf and Schmidt2000; Hank and Stuck Reference Hank, Stuck, Erlinghagen and Hank2008; Minicuci et al. Reference Minicuci, Noale, Pluijm, Zunzunegui, Blumstein, Deeg, Bardage and Jylhä2004). Whether variations in retirement regulations are sufficient to explain the size of the young old population has not been established. In fact, previous studies have not even produced robust estimates of the number of young old persons.
Before his death in 2000, Peter Laslett asked Eric Midwinter how many persons there were in the third age (Laslett used the term third age instead of the term young old but they are equivalent). The answer Midwinter (Reference Midwinter2005) provided to this question is only partially satisfactory, however, as his calculations did not take into account people's health status or involvement in unpaid work. Laslett (Reference Laslett1996) had used both aspects for the definition of third-agers in his widely acclaimed study. Most other empirical studies of the young old either identify the young old as retirees, like Midwinter, or define them by chronological age (e.g. Chou and Chi Reference Chou and Chi2002; James and Wink Reference James and Wink2007). The present paper has, therefore, to determine the number of young old people before it can assess the influence of the welfare state on the total.
In sum, this paper seeks to sharpen our understanding of the influence of the welfare state on the number of young old persons. In this context, two questions are of special interest. First, how many young old people are there? Country-differences in the demographic structure and the welfare state regime point to country variations in the relative number of young old people; while gender differences in the lifecourse suggest that there are substantial gender differences in the relative size of the male and female ‘young old’ populations. We have therefore used survey data from 10 European countries to estimate the number of young old men and women for each country. The paper also examines whether retirement regulations explain variations in the relative number of young old people. Retirement regulations are the most often discussed and utilised tool for influencing the number of young old persons, but they are not the only instrumental aspect of the welfare state. Whether their influence overrides that of other welfare state attributes will be studied for the 10 countries.
How does the welfare state influence the number of young old persons?
The welfare state influences the number of young old people in various ways, some more directly than others. Adopting a broad perspective, this paper conceives the welfare state as responsible for the welfare of its citizens and, for this purpose, as intervening in social and market processes (Kaufmann Reference Kaufmann2003; Orloff Reference Orloff1993). The pertinent instruments of the welfare state are not restricted to the much-discussed social insurance and benefits, but also comprise inter alia social services and fertility policies (see Pfau-Effinger Reference Pfau-Effinger2004). Assuming that in modern western societies almost everyone will reach the ‘young old’ life stage, the relative size of the young old population at a given time depends largely on how long people remain young old (cf. Gilleard and Higgs Reference Gilleard and Higgs2002). The welfare state therefore influences the number of young old people by conditioning how long people remain young old, which depends on the ages at which they become and cease to be young old. The principal influences are summarised on Figure 1.
Figure 1. The influence of the welfare state on the number of young old persons.
Aspects of the welfare state influencing when a person becomes young old
People become ‘young old’ when they retire from paid work or when their legal responsibility for a dependent person in need of care ends, whichever occurs later. A legal responsibility for a dependent person in need of care is taken here as the responsibility for either a minor or a person in need of long-term care (LTC), e.g. a spouse or parent. The welfare state thus influences when a person becomes young old by influencing the timing of three life-events: the retirement transition, when a person's youngest child reaches the age of consent, and the end of the legal responsibility for a dependent person in need of LTC.
The welfare state influences the timing of the retirement transition through the eligibility ages for old-age pensions, especially the threshold ages for early, regular and late retirement, and by the real value of pensions' payments (Schils Reference Schils2005). Several authors have noted that in recent decades, disability and long-term unemployment schemes have often been used as alternative pathways to early retirement (Ebbinghaus Reference Ebbinghaus, Scharpf and Schmidt2000; Guillemard Reference Guillemard, Johnson, Conrad and Thomson1989). Their eligibility rules therefore also influence the ages of retirement transitions. It should be noted, however, that retirement regulations are more important for men than for women, especially in continental and southern European countries where it is still common for women to refrain from paid work and instead to be engaged in household and family tasks. Many women are therefore entitled to only meagre pensions from the official retirement age, and are little affected by early and late retirement options (Pfau-Effinger Reference Pfau-Effinger2004; Sainsbury Reference Sainsbury and Sainsbury1999).
Turning to the timing of the second mentioned life-event, the age at which a person's youngest child reaches the age of consent, we find that the welfare state has both a direct and an indirect influence. The state indirectly influences a woman's age at the birth of her last child through fertility policies, and it directly legislates the age at which a child reaches the age of consent (Krause Reference Krause2007). It has to be noted, though, that reaching the age of consent does not necessarily coincide with becoming materially independent of the parental household. The third mentioned life event, the end of the legal responsibility for a dependent person in need of LTC, in most cases derives from the person in need of LTC, not the legally responsible person. The welfare state only marginally influences the timing of these life events, but it does create and shape the legal obligation for dependent persons in need of LTC. Some countries, such as Germany and Austria, have created the responsibility through a comprehensive LTC insurance, while most other countries only deal with it as a side issue in their social and health legislations or do not legislate for the LTC responsibility at all (European Commission (EC) 2006).
Aspects of the welfare state that influence when a person ceases to be young old
A person ceases to be ‘young old’ when they become ill (in which case they become ‘old old’), or when a legal responsibility for a dependent person in need of care begins (in which case they revert to being ‘middle aged’), whichever occurs earlier. In these cases, the dependent person is most often an older person in need of LTC. The influence of the welfare state on this life-event has just been discussed. The welfare state influences the timing of ill-health directly through the health-care system and indirectly through public-health programmes. Vigorous public-health programmes promote the health of an entire population, as by intervening in health-related behaviour such as tobacco smoking (cf. Childress et al. Reference Childress, Faden, Gaare, Gostin, Kahn, Bonnie, Kass, Mastroianni, Moreno and Nieburg2002).
Can retirement regulations explain the number of young old persons?
Retirement regulations are the most common instrument for influencing the number of young old people, though their influence might be limited. Their instrumentality derives from the fact that they explicitly specify a chronological age as a threshold for becoming young old. This threshold, however, is also regulated by other factors, as discussed above, and it influences the number of young old persons only in conjunction with the threshold for ceasing to be young old and with the demographic structure. Retirement regulations can therefore only be an effective instrument for influencing the number of young old persons when their influence is sufficiently strong to override the influence of the other mentioned factors. Retirement regulations have often been used to increase the number of young old people. During the 1970s and 1980s in continental European countries, stagnant economies led to high unemployment rates. To reduce the number of unemployed people, opportunities for early exits from paid work were developed. The intention was that older workers would withdraw from the labour market and that (younger) unemployed persons would fill their positions. Older workers did indeed withdraw from the labour market at younger ages, which increased both the number of retirees and the number of the young old. The desired effect on the unemployment rate, however, did not come about (Bussemaker and van Kersbergen Reference Bussemaker, van Kersbergen and Sainsbury1999; Ebbinghaus Reference Ebbinghaus, Scharpf and Schmidt2000).
Recently, retirement regulations have also been changed to lower the number of young old persons, as in Germany and Italy (OECD 2006). Complaints have long been voiced about the pressure exerted by the growing number of retirees on public pension schemes. The policy goal has consequently been to reduce the number of retirees, particularly the young old, and the usual means has been to raise the official retirement age (Esping-Andersen et al. Reference Esping-Andersen, Gallie, Hemerick and Myles2002). Whether the decline in the overall number of retirees will be in line with the policy makers' expectations has yet to be seen. Another strategy to reduce the pressure on public pensions has been to promote private pension schemes, which allow retirement according to individual preferences, and may have the opposite effect (Schils Reference Schils2005). Further to investigate the influence of retirement regulations on the number of young old people, we examined the retirement age, the average value of pensions and the possible use of disability and long-term unemployment pensions to enable early exits from paid work in 10 European countries. The selected countries represent the social democratic (Denmark, Sweden, The Netherlands), the liberal (England as a part of the United Kingdom), and the conservative (Austria, France, Germany, Italy) welfare regimes identified by Esping-Andersen (Reference Esping-Andersen1990), as well as the rudimentary regime (Greece, Spain) described by Leibfried (Reference Leibfried, Ferge and Kolberg1992).
Retirement regulations
Table 1 presents the key retirement regulations around 2004, and shows that there were many gender and country differences. As the survey included people aged 50 or more years, we will probably be able to observe the effects of developments over the last 15 years. While the official retirement age for men was 65 years in 2004 in almost all countries, there was some variation in the official retirement age for women, and considerable variation in the possibilities for early and late retirement for both genders. The average pension level ranged from about 40 per cent of average earnings in the United Kingdom (UK) to around 80 per cent in Greece, and in many countries disability and long-term unemployment pensions were used to promote early exits from paid work.
Table 1. Retirement regulations in 10 European countries, circa 2004
Concerning the retirement age, it should be noted that the age given for Greece applied to persons insured before 31st December 1992; for those insured after that date it was 65 years. Likewise, the retirement age given for Denmark applied to those who had reached the age of 60 years by 1st July 1999. For younger people, the retirement age was lowered to 65 years (EC 2005). In southern European countries, the pension system was fragmented and there were and are numerous possibilities for early retirement (Guillén and Matsaganis Reference Guillén and Matsaganis2000). It is therefore not surprising that in Italy in 2004, the earliest eligibility age for pensions was 57 years – and that it had previously been even lower. In France also, several widely used early retirement schemes were previously available (OECD 2005c, 2006). In Sweden, alongside the official (normal) and early retirement age, a late retirement age has existed since 1999. In that year the possibility of postponing retirement until aged 70 years was replaced with an entitlement to stay until 67 years (OECD 2006). In 2004, this was a unique arrangement among the 10 countries.
The pension level reported in the table is the average pension expressed as a percentage of economy-wide average earnings (OECD 2005f). As many women's working biographies are irregular or interrupted, the pension they receive is usually lower than men's. This is especially true in continental and southern European countries (Trifiletti Reference Trifiletti1999; Sainsbury Reference Sainsbury and Sainsbury1999). With regard to the use of disability and long-term unemployment pensions for an early exit from paid work, there were more opportunities in some countries than others. Early exits through disability pensions have been comparatively common in The Netherlands and Austria. In 2000 in The Netherlands, before reforms restrained this practice, disability schemes were even named as the main source of income for unemployed Dutch men younger than 60-years-of-age (OECD 2005d). In Austria between 1993 and 2000, early retirement because of reduced work capacity had been possible from the age of 55 years, and disability schemes have been and still are easily accessible for people aged 57 or more years. While the general age of eligibility was raised from 55 to 57 years in 1997 for men and in 2000 for women, there is still no age threshold for public employees (OECD 2005a). Early exits from paid work through long-term unemployment pensions are also relatively common in Germany and Denmark. In both countries, old-age pensions are available for long-term unemployed people from the age of 60 years (OECD 2005b, 2006).
Calculating the number of young old persons
For calculating the number of young old persons, data from the Survey of Health, Ageing and Retirement in Europe (SHARE) (Börsch-Supan and Jürges Reference Börsch-Supan and Jürges2005) and from the English Longitudinal Study of Ageing (ELSA) were obtained. Both surveys are of people aged 50 or more years, and as the structure and approach of ELSA was a model for SHARE, the datasets are comparable. For this study, the data collected in 2004 and 2005 on people aged 50–90 years in 2004 were selected. The upper age limit of 90 years was introduced to avoid bias from selection effects in the higher age groups. ELSA has data on 9,032 people meeting the inclusion criteria, and SHARE has information on 23,718 people. The data were weighted by age and gender, and in Italy and Denmark also by region.
The cases in both datasets were divided into three groups: the ‘middle-aged’, the ‘young old’ and the ‘old old’. The middle-aged were people below retirement age and/or legally responsible for a dependant in need of care. Being below retirement age was operationalised as being employed, self-employed, unemployed, or being a homemaker younger than the official retirement age. Being legally responsible for a dependent person in need of care was operationalised as being the parent of a minor and being responsible for another person in the framework of long-term care regulations. In Germany and Austria, information on the coverage of persons by long-term care (LTC) insurance was explicitly collected. In the other countries, having a care responsibility for a person under the terms of a LTC regulation was covered when it was reported as employment or self-employment. A study by the OECD (2005e) noted that in many countries LTC-givers receive contracts and payments.
People not identified as ‘middle-aged’ were categorised as either ‘young old’ or ‘old old’, depending on their ability to carry out everyday activities. This capability was operationalised for the young old as the absence of limitations in any instrumental activity of daily living (iADL), whereas the old old had such limitations. The iADLs were those specified by Katz (Reference Katz1983) : dressing, including putting on shoes and socks; walking across a room; preparing a hot meal; shopping for groceries; making telephone calls; taking medications; doing work around the house or garden; managing money, such as paying bills and keeping track of expenses. For each iADL, the respondents had been asked whether they had had difficulties over at least three months because of physical, mental, emotional or memory problems (answer categories ‘yes’ or ‘no’). We calculated the number of limitations per person (Cronbach's alpha was 0.85 in SHARE and 0.78 in ELSA), and defined ‘1’ as the cut-off value.
The number of young old persons
The percentages of those aged 50–90 years in each of the three life stage groups in 2004 were calculated for men and women separately for each country, and are displayed in Figures 2 and 3. The charts show that the ‘young old’ percentages ranged, for men, between 36 (UK) and 49 (Austria), and for women, between 35 (Greece) and 52 (Austria). In absolute numbers, the country totals of ‘young old’ men and women ranged from 750,000 (Denmark) to 13.8 million (Germany) (Eurostat 2008). The range of the young old country percentages was greater for women (17 points) than for men (13 points), and in most countries the share of the female population that was young old was higher than that for men. The gender differential exceeded five per cent in just two countries, the UK (6 points) and Greece (10 points). Greece was the only country in which the share of young old people was lower among women than men. The rank orders of the countries by the young old share were identical for men and women. In both cases, six of the 10 countries were ordered in the sequence DK⩽S⩽F⩽D⩽NL⩽A (the abbreviations are explained in Table 1). The countries with different positions in the two rankings are the UK and the southern European countries, Italy, Spain and Greece. The UK had the lowest share of young old men and a comparatively low share of young old women. The southern European countries had low shares of young old men and intermediate shares (Italy, Spain) or the highest share (in Greece) of young old women.
Figure 2. Life stage groups of men aged 50–90 years, 10 European countries 2004.
Figure 3. Life stage groups of women aged 50–90 years, 10 European countries 2004.
Retirement regulations and the number of young old people
When one considers the information on retirement regulations displayed in Table 1, some relationships are apparent with the relative number of young old people. The clearest feature is that in most countries the female young old share was higher than that for men, consistent with women's lower retirement age. Secondly, most countries with comparatively high shares of young older people had well developed opportunities for early retirement (Austria, Italy, Netherlands, Germany) and/or a high pension level (Austria, Italy, Spain), whereas most countries with a low share of young old people either had low pensions (UK, Denmark) or a high (possible) retirement age (Sweden, Denmark).
When using the retirement regulations to predict the number of young old persons, several inconsistencies arise. Greece, Italy and Spain, for example, had an intermediate to high share of young old men, but only a low to intermediate share of young old women. Another example is that the high pension level and plentiful early retirement opportunities in Greece did not lead to a high number of young old persons. To explain these inconsistencies, one must turn to other factors. To explain the considerable gender differences in the relative number of young old persons in the southern European countries, for example, one can draw on information about the typical lifecourse. In these countries, comparatively few women participate in the labour force (Trifiletti Reference Trifiletti1999), and therefore the majority cannot profit from early retirement regulations and can become ‘young old’ only after reaching the official retirement age. To explain the exceptionally low share of young old people in Greece, as another example, information on health status is pertinent. Among the 10 study countries, Greece had the highest share of ‘old old’ men and the second highest share of ‘old old’ women. If Greece had the average percentage in the old old category, its share of young old persons would be much higher. Among men, indeed, the share of young old persons would rise to a level consistent with its retirement regulations. Those examples show that retirement regulations can help but are insufficient to explain the number of young old persons in a country.
Discussion
This article has sought to develop our understanding of the influence of the welfare state on the number of young old persons. To this end, the number of young old persons was estimated for 10 European countries and compared with the regulations concerning the ages of normal, early and late retirements from the labour force. The estimates were that in 2004 the percentage of ‘young old’ persons among those aged 50–90 years ranged from 36 (UK) to 49 (Austria) for men, and from 35 (Greece) to 52 (Austria) for women. These high figures justify the attention that the young old receive by policy-makers as well as by social scientists interested in reformulations of the lifecourse. The marked country and gender differences in the estimated shares of the young old, moreover, establish that this concept refers to attributes other than mere ‘age group’. The concept of the young old differs from an age group in meaning and reference, and in the number of people it embraces.
An attempt to explain variations in the relative number of young old people by reference to the countries' retirement regulations was only partially successful. This contradicts the common belief that retirement regulations are an appropriate instrument for influencing the number of the young old. To explain the relative number even approximately, we had to draw additionally on information about health status and the lifecourse. For a more meticulous explanation, one probably needs to add information on other influences of the welfare state, e.g. fertility policies, LTC regulations and the health-care system, and on factors not directly related to the welfare state, e.g. culture and the demographic structure. With the youngest post-1945 baby boomers already having turned young old, the number of young old persons will drastically increase over the next years and decades, and the influence of the demographic structure will become more prominent. So even though retirement regulations are the most direct of the welfare state's influences on the number of young old persons, their value for this purpose might be more effective in conjunction with, for example, fertility policies and LTC regulations. This is especially true in southern European countries and for women.
Taken together, the findings underline that the relative number of young old persons is a function of welfare state provision, which upholds the idea that comparatively small and large shares characterise different types of welfare states. One can indeed find a pattern similar to that proposed by Esping-Andersen (Reference Esping-Andersen1990) and Leibfried (Reference Leibfried, Ferge and Kolberg1992) : continental European countries have a relatively high number of young old people; Scandinavian countries have relatively few; southern European countries have intermediate to relatively low numbers and marked gender differentials; and the UK has a relatively low number and a high gender differential. This pattern breaks down, however, when the mechanisms by which the welfare state influences the number of young old persons are examined, as illustrated by Greece.
To develop these ideas, the next step might be to compare the influence of different factors on the number of young old persons. This study has focused on the influence of retirement regulations and has used health status and the typical lifecourse in a country as supplementary explanations, but we have not been able to quantify the influence. To fill the gap, one could conduct either detailed case studies of change over time in selected countries or a multivariate analysis at the country level, but the latter will be possible only when appropriate data for more countries becomes available. For future analyses, it would also be helpful to have longitudinal data and a variable on LTC regulation coverage for all countries. With longitudinal data, one could determine the likelihood of an ‘old old’ person later recovering from iADL limitations. In that case, people would have been misclassified on the basis of cross-sectional data. Summing up, the influence of the welfare state on the relative number of young old people is not entirely a matter of a country's retirement regulations but also depends on LTC regulations, the health-care system and fertility policies. Unravelling the influence and comparing the number of young old persons in different countries, one finds characteristic patterns for different welfare states.
Acknowledgements
This paper has used data from Release 2 of SHARE 2004. The SHARE data collection has been primarily funded by the European Commission through the Fifth Framework Programme (project QLK6-CT-2001-00360 in the thematic programme Quality of Life). Additional funding came from the US National Institute on Ageing (U01 AG09740-13S2, P01 AG005842, P01 AG08291, P30 AG12815, Y1-AG-4553-01 and OGHA 04-064). Data collection in Austria (through the Austrian Science Foundation, FWF), Belgium (through the Belgian Science Policy Office) and Switzerland (through BBW/OFES/UFES) was nationally funded. The SHARE data collection in Israel was funded by the US National Institute on Aging (R21 AG025169), by the German-Israeli Foundation for Scientific Research and Development (GIF), and by the National Insurance Institute of Israel. Further support by the European Commission through the Sixth Framework Programme (projects SHARE-I3, RII-CT-2006-062193, and COMPARE, CIT5-CT-2005-028857) is gratefully acknowledged. For methodological details, see Börsch-Supan and Jürges (Reference Börsch-Supan and Jürges2005). Furthermore, we thank the Office for National Statistics of the United Kingdom for providing us with population estimates for England for 2004. Those estimates are the basis of the weights used for England. Earlier versions of this paper were presented at the conferences Welfare State Feedbacks: From Outcomes to Inputs in Aalborg, Denmark (31 January–2 February 2007), Forum for Social Policy Research in Rotterdam, The Netherlands (9 March 2007), and Engendering Policy and Politics in Manchester, UK (21–22 June 2007). We thank the participants at those conferences, in particular Trudie Schils, Romke van der Veen and Kees van Kersbergen, for their useful comments and suggestions. Finally, thanks to Tony Warnes and two anonymous reviewers for their suggestions that helped much to improve the paper.
