Social gerontology and changing approaches to work and retirement

The development of social gerontology as a discipline ran parallel with the emergence of retirement, both coming into prominence from the late 1940s (Atchley Reference Atchley1982; Fennell, Phillipson and Evers Reference Fennell, Phillipson and Evers1988). Many of the pioneer social scientists working in the field of ageing focused on retirement as a significant area for research (Burgess Reference Burgess1960; Havighurst Reference Havighurst1954). The period from the 1950s to the early 1990s was characterised by the development of a range of theories examining transitions from work and adjustment to retirement (Wang, Henkens and van Soling Reference Wang, Henkens and van Solinge2011). Retirement was variously theorised as raising issues for securing new social roles (Havighurst Reference Havighurst1954); as a process enabling succession in the workplace (Cumming and Henry Reference Cumming and Henry1961); as creating ‘structured dependence’ (Townsend Reference Townsend1981); and as a process shaped by differential access to physical and material resources (Wang Reference Wang2007). These and other approaches (see also Phillipson Reference Phillipson2013) viewed an extended period of retirement, together with a diminished role for paid work, as part of a new and distinctive stage within the lifecourse.

The 1990s and early 2000s brought a significant change to debates around work and retirement. On the one side, came increasing pressure on individuals to remain in some form of work for as long as possible; on the other side, a decline in the institutional supports associated with the welfare state, with emphasis placed on the responsibility of individuals to manage and plan their retirement (Vickerstaff and Cox Reference Vickerstaff and Cox2005). Following the move out of economic recession in the mid-1990s, the pattern of early exit from work went into reverse with increases in economic activity for men and (more pronounced) women in their fifties and sixties (Léime et al. Reference Léime, Street, Vickerstaff, Krekula and Loretto2017). Governments became concerned about the economic consequences of ageing populations and the associated costs of pensions and care services (e.g. OECD 2006). Delaying retirement became a central policy theme, with a shift from promoting early exit/early retirement to identifying pathways into work or helping people to remain at work, combating age discrimination and encouraging self-employment (Phillipson and Smith Reference Phillipson and Smith2005; Taylor and Earl Reference Taylor and Earl2016).

The changes identified led McDonald and Donahue (Reference McDonald and Donahue2011: 414) to argue that in contrast with previous levels of support, ‘retirement had become a “lost” institution’, abandoned in the move to encouraging longer working underpinned by the raising of pension ages (see also Macnicol Reference Macnicol2015; Polivka and Luo Reference Polivka and Luo2015; Weller Reference Weller2016). But it might also be argued that retirement rather than being ‘lost’ was ‘evolving’ in new and distinctive ways, with the EWL narrative an important influence in re-positioning its place within the lifecourse. This argument is developed by Cahill, Giandrea and Quinn (Reference Cahill, Giandrea and Quinn2015) in their analysis of the diverse pathways from career employment taken by different cohorts of older workers. They take the view that:

However, it might be argued that this type of analysis itself reflects the influence of the EWL narrative: the idea of older workers as a generally healthier group, released from demanding occupations, able to control technology and experiencing more options for work-ending. Such arguments are put forward as statements of fact rather than hypotheses to be tested. But what is the evidence for such views? Is there an alternative reading of the issue and literature which might indicate problems with the narrative used to justify EWL? The next section of this paper assesses some of the problems with the narrative used to justify an extended working life. The discussion groups these into five areas: attitudes, demography, work, training and social inequality.

The attitude problem

The raising of pension ages – though viewed as inevitable (and desirable) by governments – has provoked considerable opposition across Europe (for relevant studies, see e.g. Hokema Reference Hokema, Ní Léime, Street, Vickerstaff, Krekula and Loretto2017; Macnicol Reference Macnicol2015; Van Dyk Reference Van Dyk and Scherger2015). Research by Naumann (Reference Naumann, Kumlin and Stadelmann-Steffe2014), reviewing a range of European countries, indicated the limited room for manoeuvre which governments had on this issue, with approval for the initial raising of pension ages offset by resistance to further increases. Studies of retirement preferences across a variety of European countries also confirmed the extent to which most workers still intended to retire before 65 – notwithstanding the drive to raise pension ages to 67 and beyond (Hofäcker Reference Hofäcker2015). EWL policies also raise concerns for groups such as older women, given the failure to tackle gender inequalities at work and in the home (Lewis Reference Lewis2006). Street (Reference Street, Léime, Street, Vickerstaff, Krekula and Loretto2017: 22) makes the point that women earn less for equivalent jobs, are more likely to take a career break to care for children, and are more likely to care for a disabled adult child and older family members. Such factors underline the dangers of raising pension ages in the absence of a broader set of policies designed to support women (and men) at different phases of the lifecourse (see below). Implementation of EWL, however, would suggest that public concerns about the policy have been viewed as secondary to the broader issue of responding to pressures arising from population ageing. This raises the issue of the extent to which population change can be used as a justification for a policy such as EWL. The next section of this paper examines some of the demographic arguments used to support this approach.

The demography problem

Population pressures arising from ageing societies are invariably cited as supporting the case for EWL. This section examines this issue in three main ways: first, through highlighting demographic variations across different countries, notably in relation to the rise of the ‘baby-boom’ generation; second, issues raised by variations in healthy life expectancy; third, questions raised by projections regarding global migration. On the first of these, applying EWL policies may be an important consideration for some countries, less relevant or necessary for others. Germany, where the overall population is projected to decline by 18.8 per cent over the period 2015–2060 (Berg et al. Reference Berg, Hamman, Piszcezk and Ruhm2015), would be an example of the former; Sweden, where the population is set to increase by 30 per cent over the same period (Statistics Sweden 2015), an illustration of the latter. The existence of a ‘large’ cohort of baby-boomers entering retirement is often presented as a justification for EWL (Koltikoff and Burns Reference Koltikoff and Burns2004; Willetts Reference Willetts2010). The demography of the baby-boom refers to the increase in the birth rate across industrialised countries from the mid-1940s through to the mid-1960s. As with the issue of demographic change, sharp contrasts exist between countries. Some (e.g. Finland) had a relatively compressed surge in birth rates after the Second World War, this coming to an end at the beginning of the 1950s (Karisto Reference Karisto2007). Others (notably Australia and the USA) experienced a long period of increasing birth rates – from the mid-1940s through to the mid-1960s. The UK had a distinctive pattern of two separate peaks – in 1947 and 1964 – creating ‘first-wave’ (1945–1954) and ‘second-wave’ (1961–1965) baby-boomers. In terms of cohort size, the Finnish and UK examples provide limited grounds for using the baby-boomer argument as a justification for raising pension ages. Even in those countries where there was a sustained expansion in birth rates, other factors, notably those relating to changing patterns of employment, may operate to weaken the existence of the baby-boomer generation as a rationale for EWL (see below).

An additional argument made in support of EWL concerns the gains in life expectancy over the past 40 years. However, the problem here is eliding the substantial gains in life expectancy at birth with the more modest increases after 65. Indeed, the limited growth in ‘healthy life expectancy’ after 65 for particular ethnic and social groups raises difficulties in implementing EWL (see e.g. Olshansky, Goldman and Rowe Reference Olshansky, Goldman and Rowe2015). Variations in healthy life expectancy between countries are another significant factor, e.g. in the case of Europe: from a healthy life expectancy after 65 of just 4.2 years in Slovakia, 6.2 years for Hungary, 7 years in Germany to 11.6 years in Denmark and 12.6 years in Sweden (OECD 2014).

The demographic case for EWL (especially moving through the 21st century) might also be questioned given the increase in global migration and projections for population growth for Africa and Asia. The migration issue has been an important component of the EWL narrative, with the view that increased in-migration is unlikely to meet the needs of Europe's ageing population (Sinclair, Watson and Beech Reference Sinclair, Watson and Beech2013). However, this view may come to be questioned as economic and political upheavals generate a substantial increase in migration to Europe, in general, and Germany and the Nordic countries, in particular (Kingsley Reference Kingsley2016; United Nations High Commissioner for Refugees 2015). The implications of population projections for Africa must also be noted, with an increase of 1.3 billion people over the period 2015–2050 representing more than half of global population growth between 2015 and 2050 (United Nations 2015). Porter (Reference Porter2015) makes the point that:

Given this context, policies to extend working life may become marginal compared with the need to integrate migrant populations into Western economies and benefit from the range of skills which they possess (see also Feldman Reference Feldman2015; King Reference King2017). Of course, EWL and drawing migrant groups into employment are not mutually exclusive polices. However, they rely upon assumptions about future patterns of working which may themselves come under scrutiny, an issue examined in the next section of this paper.

The work problem

A key rationale for EWL policies is linked to ensuring that sufficient numbers of workers are available to meet potential shortages of labour. However, from an employment perspective three issues need to be considered: (a) what type of work is EWL likely to support?; (b) will future changes to employment increase or decrease the demand for older workers?; and (c) does the term ‘older worker’ aid or obscure some of the challenges associated with EWL?

Taking the first question, the promotion of EWL has come at a time when ‘good quality’ employment appears to be in decline because of technological and related changes (Avent Reference Avent2016; Ford Reference Ford2015; Frey and Osborne Reference Frey and Osborne2015; Haldane Reference Haldane2015; Srnicek and Williams Reference Srnicek and Williams2015). This may be less an issue for the existing cohort of employed older workers, notably those in long-term employment or self-employment – the main groups presently contributing to EWL. However, EWL is being implemented when opportunities for full-time employment are contracting – illustrated by the decline in countries such as the USA in labour-force participation among prime-age workers (Dvorkin and Shell Reference Dvorkin and Shell2015). Instead, there is the growth of employment with limited security in respect of pensions, health insurance and guaranteed hours (OECD 2015; Srnicek and Williams Reference Srnicek and Williams2015; Weller Reference Weller2016). This type of employment – associated with what Standing (Reference Standing2011) has termed the ‘precariat’ – may be increasingly targeted at older workers displaced from their lifetime employment (e.g. 24% of those on ‘zero hours’ contracts are in the 50–64 age group; Office for National Statistics 2017). In the UK, much of the job growth amongst older workers has consisted of self-employment (accounting for nearly one in two workers aged 65+), but with much of this producing low levels of income for those entering or remaining in this type of work. In the USA, a survey by Koenig, Trawinski and Rix (Reference Koenig, Trawinski and Rix2015) found that those aged 45–70 who had been unemployed in the past five years and who were re-employed were twice as likely to be working part-time as the total workforce in this age group: 34 per cent compared with 16 per cent. Equally significant was the finding that 48 per cent of those who had found new jobs said that they were now earning less when compared with the job they had before becoming unemployed.

Second, to what extent will changes to employment increase or decrease the demand for older workers? One neglected area in EWL policy is the degree to which technological change is likely to transform the terms of the debate about the extent to which later-life employment can be achieved. Frey and Osborne (Reference Frey and Osborne2013) have predicted in the case of the USA that 47 per cent of jobs are at ‘high risk’ of being automated in the next decade, with similar estimates being made for Europe (Frey and Osborne Reference Frey and Osborne2015; see also Avent Reference Avent2016; Haldane Reference Haldane2015; Lawrence, Roberts and King Reference Lawrence, Roberts and King2017). These authors make the point that while new technology can both generate as well as destroy jobs, the former tend to be rather fewer than the latter, noting that: ‘In 2010 only about 0.5 per cent of the US workforce was employed in new industries that did not exist a decade earlier’ (Frey and Osborne Reference Frey and Osborne2015: 63). Importantly, most of these jobs were at the high skill/high education/high wage end, suggesting limited job growth in areas which might fit an increased supply of older workers.

Third, the category ‘older worker’, whilst useful in policy terms, also obscures gender, ethnic, class and industrial divisions affecting people grouped in the 50+ age category (see also Taylor et al. Reference Taylor, Loretto, Marshall, Earl and Phillipson2016). Thus, taking gender as an example, in the case of men, EWL policies may be relevant for some countries, less for others: Sweden, for example, may have reached ‘saturation’ point with labour-force participation in 2015 of 82 per cent for men aged 55–64; the UK has seen a modest expansion in labour-force participation: rising from 65 per cent for men aged 55–64 in 2005 to 71 per cent in 2015 (Eurostat 2015). Greece illustrates the impact of austerity policies with a decline in labour-force participation for men aged 55–64: from 59 per cent in 2005 to 55 per cent in 2015. Women, on the other hand, represent a contrasting picture with substantial growth in labour-force participation, albeit from a very low basis, in countries such as Hungary and Poland. However, these increases demonstrate the tensions with the EWL narrative, with the expansion in labour-force participation for women ‘disproportionately over-represented in the flexible forms of work [part-time working, temporary contracts, self-employment] emerging within the globalisation process’ (Buchholz et al. Reference Buchholz, Hofäcker, Mills, Blossfeld, Kurz and Hofmeister2009: 61). Buchholz et al. observe that:

The training problem

One response to the problems facing the implementation of EWL is to call for a closer alignment between education and training and the workplace. Rowe, for example, suggests that:

In the UK, Altmann (Reference Altmann2015: 36) has called for a ‘national strategy to address adult skills gaps and ongoing re-training requirements’. However, the evidence suggests that meeting this objective will be a major challenge, requiring substantial investment in further and higher education.

Hyde and Phillipson (Reference Hyde and Phillipson2015), analysing data from the English Longitudinal Study of Ageing (Waves 1–5), demonstrate that participation in formal education has been in steep decline amongst the 50–59 age group since 2008. Analysis of the 2014 Wave of the UK Understanding Society Survey (a longitudinal study with data collected on an annual basis) confirmed that only a minority of older workers are likely to have access to regular training and updating of skills: 23 per cent of men and 26 per cent of women aged 50–59 reported receiving work-related training in the past 12 months; the figure for those aged 60–69 was 11 per cent. Data from the UK Labour Force Survey also reveal important socio-economic differences in the likelihood of older workers’ participation in training programmes (Department for Work and Pensions 2014). The evidence suggests a decline in training affecting all occupational groups but with older workers in low-skilled occupations the least likely to receive training (see also König, Hess and Hofäcker Reference König, Hess, Hofäcker, Hofäcker, Hess and König2016). Part-time workers and the self-employed, both of whom are more likely to be drawn from the 50+ age group, are also likely to receive substantially less training than their full-time counterparts (Vickerstaff, Phillipson and Loretto Reference Vickerstaff, Phillipson and Loretto2015).

Similar trends have been reported by researchers in the case of other countries. Cummins, Harootyan and Kunkel (Reference Cummins, Harootyan and Kunkel2015: 153), drawing on national data-sets in the USA, found that: ‘Older workers [were] less likely to participate in both publicly and employer-sponsored training compared to younger workers’. In Sweden, an employment rate of 77 per cent for men (72% for women; 2014 figures), has been achieved despite the virtual absence of people aged 50–64 in formal education – 4 per cent compared with 13 per cent of those aged 35–49 (Krekula and Engström Reference Krekula and Engström2015). An investigation set up by the Swedish government concluded that: ‘many people [are being] denied further training after 50 … something which in turn creates difficulties for older workers to maintain or gain employment, or get new work tasks’ (Krekula and Engström Reference Krekula and Engström2015: 127). Conen, Henkens and Schoopers (Reference Conen, Henkens and Schoopers2014), in a survey of employers’ actions across several European countries to support EWL, found some of the lowest levels of training amongst countries with the highest proportions of older people – Germany and Italy are notable examples (see also Carmichael and Ercolani Reference Carmichael and Ercolani2014).

The evidence suggests that implementation of EWL policies has coincided with a decline in the availability of training and skill development for older workers. This raises concern that sections of this group (especially those with limited education and obsolete skills) are being targeted as a new ‘reserve of labour’ to fill the ‘short-term’, ‘precarious’ jobs which have emerged with global capitalism (Standing Reference Standing2011, Reference Standing2014). Indeed, it might be argued that much of the expansion of ‘self-employment’ amongst those 50 and over can be read as responding to increased demands for flexible but insecure and unstable forms of employment. The implications of this for generating new forms of inequality and division are considered in the next section of this paper.

The inequality problem

A further issue regarding EWL concerns whether it accentuates processes associated with cumulative advantage/disadvantage which accrue over the lifecourse (Dannefer Reference Dannefer2003). This might reflect the emergence of a ‘do-it-yourself’ style of retirement (Cahill, Giandrea and Quinn Reference Cahill, Giandrea and Quinn2015), with emphasis placed upon individual responsibility and risk taking (Macnicol Reference Macnicol2015; Polivka and Luo Reference Polivka and Luo2015). The extent of inequality arising from EWL and the rapid reversal from early retirement has been documented in a series of studies, notably in Germany, but also other European countries, for example those by Rinklake and Buchholz (Reference Rinklake, Buchholz, Blossfeld, Buchholz and Kurz2011), Buchholz, Rinklake and Blossfeld (Reference Buchholz, Rinklake and Blossfeld2013), Hofäcker and Naumann (Reference Hofäcker and Naumann2015) and Hofäcker, Hess and Naumann (Reference Hofäcker, Hess, Naumann and Torp2015). Hofäcker and Naumann (Reference Hofäcker and Naumann2015) make the point that early exit pathways were an important route for workers with low or limited skills, reducing the financial penalty associated with the loss of work. Closure of these pathways, however, has forced many to continue working longer – but mainly driven by financial need rather than any active choice to remain in the labour market. Hofäcker, Hess and Naumann conclude from their survey of the labour market situation of older workers in 30 countries (drawing on data from the European Social Survey), that:

But it is precisely this ‘critical labour market situation’ which few countries or employers are willing to address. Rinklake and Buchholz summarise the situation as follows for Germany, but the point has relevance for many industrial countries:

Achieving a fuller working life

Based on the above argument and others in this paper, a contrasting approach would be to place greater stress on what the Department for Work and Pensions (2014) has termed a ‘fuller working life’, but linking this less to EWL per se and more to improving work quality and security as a precondition of any policy for encouraging working in later life. This broader objective indicates challenges for research and public policy towards older workers which may be summarised as follows.

First, achieving the goal of a ‘fuller working life’ will require attention to changes in employment and in the labour market. Evidence for the deteriorating quality of the workplace environment – driven by globalisation and technological change – is now substantial (for UK evidence on this issue, see e.g. Smeaton and White Reference Smeaton and White2016); and the question of whether sufficient jobs will be available in the future – following the analysis from Frey and Osborne (Reference Frey and Osborne2015) and others – is a critical issue to consider when considering the implications of changes to pension ages. Based on evidence cited in this paper, advancing expected retirement ages in the absence of secure forms of employment may result in an extension of inequalities from the middle to later phases of the lifecourse. One solution may be to recognise that EWL will almost certainly contribute to the expansion of ‘precarious’ forms of working but to match this with corresponding rights for vulnerable groups of older workers (especially those displaced from their lifetime employment), through, for example, ensuring access to trade unions, guaranteed hours and regular training (for a development of this theme, see Standing Reference Standing2014).

Second, and following the above argument, greater attention needs to be given to the way in which the raising of pension ages is generating inequalities linked to contrasting experiences in terms of health, finances and social relationships (Holley-Moore, Hochlaf and Riaz Reference Holley-Moore, Hochlaf and Riaz2017). Many groups of workers will not benefit from a period of retirement (dying prematurely) or will have insufficient income to be able to leave what may be ‘precarious work’ in their sixties/early seventies. Others may choose to leave early or have a minimal EWL (because they have substantial pensions and savings) and will experience a lengthy retirement supported by many years of ‘healthy life expectancy’. This suggests new inequalities arising from EWL, with a substantial re-distribution from the poor to the wealthy, a process reinforced by economic austerity and cuts to the welfare state (Cooper and Whyte Reference Cooper and Whyte2017; Toynbee and Walker Reference Toynbee and Walker2017). Indeed, EWL appears to be reinforcing the ‘social division of welfare’ identified by Richard Titmuss in the 1950s, when he observed that:

The argument is for a work and retirement policy which recognises processes of cumulative advantage and disadvantage operating over the lifecourse (Dannefer Reference Dannefer2003; Kendig and Nazroo Reference Kendig and Nazroo2016). This point has been made by Berkman, Boersch-Supan and Avendano, where they argue that given divisions between groups (especially in terms of health):

This is an important argument which if adopted could do much to resolve many of the problems associated with the EWL approach. The available evidence, however, suggests that the reverse is happening in many countries, with those in poor health often forced to remain in work to secure health insurance or retirement benefits (Benjamin, Pransky and Savageau Reference Benjamin, Pransky and Savageau2008); or those with a lifetime of low incomes needing to work later in contrast to those with the benefit of secure occupational pensions and/or access to income from property who can take early retirement (Phillipson Reference Phillipson2013).

Third, developing research and policy in the field of training for older workers should also become central to achieving a fuller working life (Cummins, Harootyan and Kunkel Reference Cummins, Harootyan and Kunkel2015). The focus here should be on support for those in ‘precarious’ forms of employment who are often the least likely to have access to training and continuing education. Some options might include: more imaginative use of computer-based training or ‘e-learning’ to assist those working from home or those juggling work and care-giving responsibilities (Czaja and Sharit Reference Czaja and Sharit2009); specific obligations placed upon employers to expand training and learning as a pre-condition for creating non-standard forms of employment; and making further training a legal entitlement for those changing careers or moving into bridge forms of employment. Encouraging a prominent role for universities and community colleges in workforce development should also be considered. The number of older people wishing to take advantage of higher and college education will almost certainly grow, with the ‘baby-boom’ generation (and the cohorts behind them) likely to fuel demands for new types of vocational and non-vocational courses. Reflecting this, a variety of initiatives might be followed by the university and college sector: experimenting with new types of career development courses aimed at those entering mid-life; launching research programmes testing new types of training for a more diverse workforce; developing courses aimed at assisting those who experienced educational disadvantage throughout their lives but who face a period of extended working (Phillipson and Ogg Reference Phillipson and Ogg2010).

Fourth, if retirement is now viewed as a 20th-century institution, the same might be said of the term ‘older workers’. Women, men, different ethnic sub-groups, those in growing/declining cities/industries, may have the category 50+ in common, but very little else which carries sociological meaning. The ‘fragmentation’ of the label ‘older worker’ is the reverse side of the fragmentation of retirement: for some groups, EWL will be a positive addition to ‘active ageing’; for the ‘losers from globalization’ (Buchholz et al. Reference Buchholz, Hofäcker, Mills, Blossfeld, Kurz and Hofmeister2009) (men with limited skills, some minority ethnic groups, those living in post-industrial cities), EWL has limited salience as a way of managing ‘work-ending’. In addition, it is possible the term ‘older worker’ itself compounds the problems facing people struggling to remain in work, with research indicating that the moment someone becomes categorised as an ‘older worker’ they are potential targets for prejudice and discrimination (Desmette and Gaillard Reference Desmette and Gaillard2008; Riach Reference Riach2009; Taylor et al. Reference Taylor, Loretto, Marshall, Earl and Phillipson2016).

Finally, whether viewed as a ‘fuller’ or an ‘extended’ working life, considerable opportunities exist for re-thinking theories concerning the transition from work to retirement. For example, the evolving work–retirement landscape challenges the popularity of rational choice theories which emphasise the power of individuals to choose amongst different options about their future (e.g. Becker Reference Becker and Elster1986; Hofäcker, Hess and König Reference Hofäcker, Hess and König2016). Vickerstaff and Cox (Reference Vickerstaff and Cox2005: 92) view this as part of the ‘individualization of retirement’, the result of which has been ‘less to increase the majority of people's range of alternatives and choices over when and how to retire and more to enlarge the range of risks they [have] to cope with’. Similarly, Wang and Schultz (Reference Wang and Schultz2010: 186) emphasise the extent to which: ‘retirement decisions are often made in the face of incomplete and imperfect information, which renders a sense of uncertainty in the decision-making process’. They further note that: ‘So far, no theoretical framework has incorporated this notion of uncertainty, which could have important theoretical implications for understanding the psychological aspect of the retirement decision’. However, as the trends surveyed in this article suggest, ‘uncertainty’ and ‘risk’ are now a key feature of transitions from mid-life onwards (Vickerstaff Reference Vickerstaff2006, Reference Vickerstaff, Twigg and Martin2015). The changes associated with this development reinforce the need for theories and policies alert to the unstable foundations on which the institutions of work and retirement now appear to rest.