| HOME | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|---|
| ||||||||||||||||||||||||
RESEARCH ARTICLE |
1 Centre for Behavioural and Social Sciences in Medicine, University College London, U.K.
2 City University, London, U.K.
3 Imperial College, University of London, U.K.
Address correspondence to Chris Gilleard, Centre for Behavioural and Social Sciences in Medicine, University College London, Wolfson Building, Riding House Street, London, U.K. E-mail: CGilleard{at}aol.com
 |
Abstract |
|---|
 TOP Abstract MethodResultsDiscussionReferences |
|---|
Methods. With use of self-reported data from the retired members of a nationally representative survey, an index of consumption was constructed. Its internal reliability was analysed and analyses of variance performed to examine the impact of class of origin, cohort, and class at exit on levels of self-reported "consumerism."
Results. The index demonstrated sufficient internal reliability to provide an operational measurement of "consumerism" within this retired population. Analyses of variance confirmed that class at exit and cohort but not class of origin contributed significantly to variation in levels of consumerism. These effects were not mediated by cohort differences in health.
Discussion. Occupational standing immediately before retirement was associated with "consumerism" after retirement. This is unsurprising. The results also show that birth cohort exercised a stronger historical influence on current consumption in later life than did class background (i.e., paternal social class). This supports our thesis that the limited but increasing immersion in mass consumer society of successive cohorts who were born and grew up earlier in the 20th century continues to be reflected in levels of "consumerism" in retirement.
Retirement in Britain has changed dramatically over the last 50 years. Though originally viewed as a social and personal "tragedy" (Townsend, 1963
, p. 157), retirement has become something to look forward to, a good to invest in, and, increasingly, a resource to struggle for. Retirement is represented as the acquisition of leisure rather than the loss of employment, as people's position in the productive process no longer provides the core of their social and cultural identity (Gilleard & Higgs, 2000, pp. 38–42). What was created as a form of state-sanctioned unemployment has become a "free-floating" position within society, a less determinate social category with increasingly permeable boundaries.
Identity and lifestyle are fashioned as much by the practices of consuming as they are by the experiences of the workplace. As a result, consumption has become a focus for a number of different disciplinary and research paradigms over the last few decades. These range from the abstract theorisation of Baudrillard (1998) and Featherstone (1991) through the history of consumer society (Cohen, 2003) to the anthropology of shopping (Miller, 1998). What all these writers acknowledge is that consumption has symbolic as well as utility value. Increasingly, the good life is constructed out of consumer choice (Bauman, 1998).
At each stage of life, patterns of consumption facilitate processes of social differentiation. In the decade following the Second World War, household consumption in the United Kingdom was dominated by the necessities of life. During this period, pensioners' household expenditure was devoted to rent, fuel, light and power, and, most importantly, food (U.K. Ministry of Labour, 1957, p. 246). By 2001, the largest source of expenditure for British retired couples was leisure goods and services. Food and drink, housing, and fuel costs represented a little over one third of such households' average total weekly expenditure: £107.90 out of a total expenditure of £316.40 (Office of National Statistics, 2001, Table 4.1).
Despite this cultural shift in the nature and style of life after retirement, relatively little attention has been paid to the role that consumption plays in fashioning what Peter Laslett (1989) termed "the third age," when people who are healthy and free from the responsibilities of work and child care can begin to realise a greater degree of self-fulfilment. Gilleard and Higgs (2005) have argued that the third age can be viewed not so much as a chronological age group but as a cultural "field," a social space delineated by the opportunities for continuing participation in mass consumer society. Viewed in this light, the period when the third age emerges becomes particularly important. The third age is predicated upon the material conditions and the associated experiences that affected particular birth cohorts (Gilleard, 2004; Gilleard & Higgs, 2002). We have argued that a key element that frames participation in the third age comprises the extent and nature of household consumption. Over the last few decades, the significance of consumption as a determining or a defining characteristic of social life has grown (Edwards, 2000). The acquisition and use of goods and services have produced a cultural environment that encourages the construction of lifestyles and identities based upon consumer-driven distinctions (Slater, 1997).
The significance of cultures determined by social class membership, which were common features of life in the first half of the 20th century in Britain (McKibbin, 1998), has declined, and identities based upon "lifestyle" have become more common (Chaney, 1993). This is a product of rising affluence in the United Kingdom together with the increasing availability of non-production-based sources of identity, such as music, the media, fashion, vacations, and home ownership. The salience of cohort over class became particularly apparent in Britain with the cultural revolution of the 1960s, which fractured the cultural and social stability of postwar British society (Rosen, 2003). Although this was true of many Western industrial nations, the United Kingdom experienced this combined affluence and cultural change later than the United States but before many continental European countries. Consequently, engagement with, or distance from, the cultural frameworks of consumption has become an important aspect of contemporary ageing (Carder & Hernandez, 2004; Ekerdt, 2004; Vincent, 2003). This is marked by the cultural demise of the old age pensioner as emblematic of old age and its replacement by a "third ager" defined by a focus on self-realisation and an ongoing engagement with lifestyle, shopping, vacations, and lifelong learning. What determines the degree and nature of that participation is both the level of economic resources that members of particular cohorts of retirees command and the extent of their socialisation into the cultural practices of a mass consumer society.
What turns a cohort into a generation is a complex issue. Karl Mannheim argued that a generation is the product of temporal location (e.g., a birth cohort) and a set of experiences that are distinct from those of other earlier or later cohorts (Mannheim, 1997, pp. 33–35). Consumption in later life can be seen as the continuation of a pre-existing engagement with the consumerism of postwar society (Gilleard, 2004). Whilst occupational history influences retirees' current disposable income, generational location (birth cohort) determines successive cohorts' exposure to and experience of the rising affluence of the postwar period. Cohorts born in the early decades of the 20th century have been affected by the Depression of the 1930s and the austerity of the Second World War and the period of postwar reconstruction. As consumption has come to dominate production as the vehicle for the expression of lifestyle identities, socioeconomic position or class has become less culturally significant (Bottero, 2005). Class less and less projects forward as the historical context for contemporary consumption. But despite the decline of class as a dominant cultural form, it continues to act as a material factor in the determination of discretionary spending income. Thus, although its social impact is undeniable, it operates within the material present rather than from the enculturation of the past.
The current study has made use of recent data from a national longitudinal study of British men and women in their 60s and 70s to examine this thesis. It is designed to address two questions. The first is whether patterns of consumption amongst the retired population do possess sufficient internal unity to justify treating "consumerism" as a "sociocultural unity," which can be measured reliably and meaningfully within these cohorts. If this proves to be so, the second question we examine is whether such measures of consumerism vary more by contemporary income and historical cohort (treated here as a proxy for generational location) rather than by the historical class position of the household in which individuals have grown up.
In Britain, since 1911, membership of social class has been derived from the Registrar General's fivefold classification of occupations (Halsey, 2000, p. 16). In contrast to the United States, where education and occupation have jointly formed the basis for studies of social stratification, the relative lack of secondary and tertiary educational opportunity in Britain—at least until the 1960s—has limited its utility as a source of social differentiation (Heath & Clifford, 1996). Consequently, our analyses are confined to this class-based stratification system. Our hypotheses are that "third age consumerism" does possess significant internal unity and second that generational location rather than "class of origin" exercises the greater influence on "consumerism" in retirement. Our final hypothesis is that people's economic circumstances at retirement tend to overshadow the cultural influences of the past, whether these are framed as arising from a particular generational or class position.
|
METHOD |
|---|
|
TOPAbstractMethodResultsDiscussionReferences |
|---|
). The study population is highly representative of their age group in the British population. As children, their fathers had a social class distribution matching that of men aged 25 to 45 years in the 1931 National Decennial Census. In later life, their sociodemographic characteristics were similar to those of men and women aged 65 to 75 years in the 1991 Census. Finally, their state of health matched closely that of people the same age in the 1995 Health Survey for England (Blane et al., 1999). To operationalise the impact of generational and class position, each of the subjects was assigned to a particular birth cohort (born in the 1920s versus born in the 1930s) a "class of origin" (paternal occupational status), and a "class at exit" (last job before leaving the labour market). Class of origin was based upon their fathers' occupation categorised into the British Registrar General's standard fivefold occupational classification. The five categories were reduced into a threefold division: Classes I and II (professionals, executives, and managers), Classes IIIn and IIIm (skilled manual and nonmanual occupations and owners of businesses), and Classes IV and V (semi- and unskilled manual occupations). Class at exit was based upon the same standard occupational classification applied to the last job before retirement, reduced to the same tripartite division.
Procedure and Measures
The subjects were sent a self-completion questionnaire in spring 2000. The sample received two mail-outs of the survey questionnaires 14 days apart. Those who responded to neither mailing were telephoned and, whenever possible, interviewed over the phone. Just over 93% of those approached either completed the mailed-out questionnaires or completed the interview over the telephone. There were 140 female and 123 male respondents. The mail-outs comprised an explanatory letter together with a self-report booklet containing various self-report measures of health, life satisfaction, social engagement, and quality of life and a stamped, addressed envelope.
The questionnaire included a 26-item scale about consumption and lifestyle that was derived from an earlier survey investigating the utility of different bases for social categorisation in market research (O'Brien & Ford, 1997). We chose this scale as the template for our third age "Index of Consumerism" as it had been developed explicitly to evaluate the merits of consumer-oriented alternatives to occupationally based measures of social class. The original study had been conducted in 1987, employing a list of 18 items chosen to reflect either the ownership of material possessions, the nature of leisure activities, or the sources of household income and wealth. We added a further eight items reflecting social engagement (buying a daily paper, voting in elections) and ownership of "newer" consumer durables that have become more widespread in Britain in the last 15 years, such as mobile (cell) phones, digital TV, and internet access. The items in the original and subsequently revised scales are shown in Table 1.
|
). Full details of the sample design and the response rates are available elsewhere (Taylor, Conway, Calderwood, & Lessof, 2003). For the purposes of the current study, all items concerning finances, consumer durables, and leisure from the survey instrument were reviewed to identify data that would yield binary information similar to that of the Boyd–Orr study. Although the interview format was different from the self-report booklet used in the latter study, it was possible to identify 12 items that were common to both samples. These items were extracted from the responses of men and women from the same two cohorts used in the Boyd–Orr study: those born in the 1920s and those born in the 1930s. The ELSA Study is a much larger and hence more reliably representative sample of the English population. It was, however, conducted 2 years after the Boyd–Orr study. Consumption of home technology items such as mobile (cell) phones and personal computers has undergone exponential growth over the last decade, and consequently we anticipated that there would be slightly higher rates of personal computer ownership, access to the internet, and ownership of mobile phones in the ELSA sample compared with the Boyd–Orr sample.
Our analyses are confined to men and women from two birth cohorts for whom there are complete data on the 12 consumer items. Table 2 presents the relative frequency of owning various consumer goods, experiencing various forms of leisure, and owning various financial means of consumption amongst the two samples. All data were treated as if they formed binary responses to the question of whether or not the respondent owned the item or had "consumed" it.
|
|
RESULTS |
|---|
|
TOPAbstractMethodResultsDiscussionReferences |
|---|
To create a more homogeneous scale, we eliminated items with a loading of less than 0.20, reducing the "scale" from 26 to 19 items. One item unrelated to "consumption" (voting in elections) was also removed and the reduced 18-item scale subjected to a further internal reliability analysis. The corrected item-whole correlations ranged from 0.15 (owns a home/house abroad) to 0.53 (owns a credit card). The alpha coefficient was 0.77. This 18-item scale became our index of third age consumerism and principal dependent variable in subsequent analyses that examined the effects of class of origin, class at exit, and birth cohort.
Class and Cohort Effects on Third Age Consumerism
A three-way analysis of variance was performed examining the effects of class of origin, class at exit, and cohort on our Index of Third Age Consumerism. The results are presented in Table 3.
|
Although gender differences were not central to our hypothesis, we did examine whether or not gender mediated the relative effects of class at origin versus cohort on contemporary "third age" consumerism. Because the study's focus was more upon household possessions than individualised consumption, gender was not expected to have a large effect especially as, within a British population in their 60s and 70s, most men and women live with their spouse and share ownership of household goods, as well as sharing leisure activities like holidays (Office for National Statistics, 2004).
As the numbers in the cells were too small to permit a four-way analysis of variance, we conducted two separate analyses of variance: one examining class at origin, cohort, and gender, and the second examining class at exit, cohort, and gender. In the first analysis, as in the analyses reported earlier, class at origin had no significant main effects or any interaction effects. Gender exercised a significant effect, and it also interacted significantly with cohort (but not class). Whereas men and women from the 1930s cohort had similar mean levels of consumerist participation (mean score for men was 9.1, for women 8.9; n = 51 and 57, respectively), men born in the 1920s had significantly higher scores than had women born in the same decade (mean score for men 9.5, for women 7.2; n = 72 and 83, respectively). In the second analysis, class at exit exerted a significant effect, but in that analysis, gender did not. As in the original analysis, there was a significant interaction between class at exit and cohort.
Finally, to exclude the possibility that cohort effects were confounded by age-associated variations in health, we examined the relationship between (a) the presence versus absence of self-reported long-term illness, (b) the presence versus absence of a long-term illness limiting everyday activities, and (c) the presence of one or more serious illnesses (specifically cancer, diabetes, heart disease, hernia, high blood pressure, joint disease, lung disease, pneumonia, stroke, and/or thyroid disease) and scores on our consumer index. Forty-seven percent of those born in the 1920s (68/144) and 36% of those born in the 1930s (36/101) reported having a long-term illness. This trend toward more members of the older cohort having a long-term illness did not reach statistical significance (
2 = 3.4, df 1, p <.10 >.05). There was no significant difference in score on the consumer index between those with (scale mean 8.3, SEm = 0.35) and those without (scale mean 9.1, SEm = 0.27) such a disability (F = 3.2, df 1, 242, NS).
There was no statistically significant difference in the prevalence of limiting long-term illnesses between the two cohorts (2 = 1.3, df 1, NS), nor did having a limiting long-term illness significantly influence participation in consumerism (F = 3.2, df 1, 242, NS). Significantly more people in the younger cohort (35/108) were reportedly free of any serious illnesses than were those (28/155) in the older cohort (2 = 7.2, df 2, p <.05). Nevertheless, the number of serious illnesses people reported did not significantly influence levels of consumerist participation. Those reporting no serious illness (n = 63) had a mean score of 8.8 (SEm = 0.44), those reporting one serious illness (n = 101) a mean score of 8.9 (SEm = 0.33), and those with more than one (n = 99) had a mean score of 8.2 (SEm = 0.36) differences that were not statistically significant (F = 1.2, df 2, 260, NS).
|
DISCUSSION |
|---|
|
TOPAbstractMethodResultsDiscussionReferences |
|---|
But how satisfactory is it to load so much upon a cohort difference that represents no more than a mere decade? How acceptable is it to privilege cohort over age in our analyses? Have we, in fact, demonstrated little more than that poorer older people consume less than richer younger people? The differences in age between our cohorts were quite small (mean age of the 1920s cohort = 70.8 years, SEm = 0.10, range 68–75 years; mean age of the 1930s cohort = 65.1 years, SEm = 0.13 years, range 60–67 years). To find within such a narrow age band significant cohort effects is arguably a more powerful argument favouring our thesis as the implied periodisation difference is quite small. It is nevertheless an important difference, as Marwick has pointed out, because it was from amongst the cohort born in the 1930s that the pioneers of Britain's cultural transformations of the 1960s emerged (Marwick, 1998, pp. 116–118). We would predict that in 10 years' time when we can include a further cohort of retirees born in the 1940s, differences in consumer participation in retirement will be more profound and class of origin differences even weaker.
Privileging ageing over cohort, we believe, is the more erroneous strategy. Cross-sectional and longitudinal studies of savings and consumption in retirement do not support the assumption that ageing per se leads to any large drop in expenditure beyond those that are associated directly or indirectly with retirement from work (see Ameriks, Caplin, & Leahy, 2002; Hurd & Rohwedder, 2003), nor does age per se influence retirees' reports of a postwork decline in spending (Hurd & Rohweder, 2003). Over time (when age is held constant), retirees' "average propensity to consume" has increased markedly (Nieswiadomy & Rubin, 1995). Longitudinal studies do suggest that in the United States, though not necessarily in Europe, there is some postretirement decline in disposable income (Hungerford, 2003), but even then it is not clear that this leads to a progressive reduction in consumption. Figures from the U.K. Family Spending National Survey, conducted closest in time to the Boyd–Orr Resurvey (1999), indicate that discretionary spending—as reflected in the proportion of expenditure devoted to leisure goods and services—rose by successive (adult) age groups. Representing only 15% of the expenditure of households aged under 30 years, spending on leisure goods and services rose to 16% for those over 30 and under 50 years, 18% for those aged 50–64 years, and reached 20% of the total expenditure for those 65–74 years (Office of National Statistics, 1999, Table 2.2). In short, to view "age" rather than "cohort" as the engine driving third age consumerism risks mystifying rather than illuminating the structural processes underlying people's patterns of consumption pre and post working life.
Of course, it is possible that by choosing to analyse our data by cohort rather than age group, we have failed to consider other, more immediate explanations of so-called "cohort" differences in consumption, such as the impact of age-associated health impairments. People in older cohorts did show a tendency, in the current sample, to report more long-term illness, and they were also more likely to report one or more serious illnesses (significantly so). But neither illness nor limiting long-term disability affected our measures of consumer participation. We must conclude, therefore, that the cohort effects we found were not the result of age-associated health differences. The national health service system in the United Kingdom, with its "free at the point of delivery" universal health care model, significantly limits the personal costs of illness and disability compared with countries like the United States with more individualised systems for health care provision. As a result, health care costs in the U.K. account for very little "discretionary" expenditure at either an individual or a household level.
The results from both the Boyd–Orr and ELSA Studies indicate that participation in consumer culture has increased in successive cohorts of retirees in the United Kingdom. This is not because ageing leads to a decline in consumerist participation, but, we would argue, because there is a secular trend toward the integration of successive cohorts into mass consumer society. This trend is likely to continue as consumption continues to unite working and postworking populations. The physical separation between people of working age and retired people has grown throughout the last century, as a result of declines in intergenerational living, retirement age, and number of people continuing to work in later life. Participation in consumer culture suggests the emergence of new, if not unproblematic, bonds between working and postworking people. Whereas neither workers nor retirees are defined solely by their consumption, their common participation in consumer culture is an area of emerging interest for both theory and research in social gerontology. Such linkages are likely to become more relevant as the baby boomer generation heads into retirement.
Common ownership of, or access to, the media and other sources of entertainment as well as shared access to the means of communication, in the form of mobile (cell) phones and the internet, ensures that both working and nonworking people remain bound together as active citizen–consumers. The class divisions of the interwar period that separated our subjects during their youth no longer differentiate them later in life, whereas retirement from the productive processes carries with it less personal social and economic disadvantages than was the case for earlier cohorts. Whether this represents the transformation of a gendered, class-dominated "tragedy" of retirement into a leisured lifestyle that continues to draw upon the consumerism of mass culture will become clearer in the future when the baby boomers retire "en masse." This study along with many others suggests that such a transformation has begun, though some fear it may soon be aborted (Schier, 1997). Whatever the future may bring, our findings suggest that consumerism forms a consistent and coherent strand in the contemporary cultures of the third age. Acknowledging this does not mean that we should underplay other older continuities represented by family ties and patterns of intergenerational exchange.
|
Acknowledgments |
|---|
The authors thank Lee Berney, Maria Evandrou, Katrina Hilari, Aubrey McKennell, and Roger Thomas for their invaluable assistance as members of the Advisory Group. They are also grateful for the comments of three anonymous referees who helped make this a better paper than it was when first submitted to the Journal. Finally, thanks to all those members of the Boyd–Orr cohort whose participation made this study possible.
|
Footnotes |
|---|
Received for publication September 14, 2004. Accepted for publication May 24, 2005.
|
References |
|---|
|
TOPAbstractMethodResultsDiscussionReferences |
|---|
| ||||||||||||||||||||||||
| HOME | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|---|
