Population Trends Spring 2000 - Office for National Statistics

Population Trends 99

Spring 2000

Looking back to look forward: lessons from four birth cohorts for ageing in the 21st Century Maria Evandrou, Kings College London and Jane Falkingham, London School of Economics

Government projections indicate that by 2026 the number persons aged 60 and over living in Britain will reach 17.1 million, giving rise to concerns about the social and economic costs of an ageing population. Little attention, however, has been paid to the likely socio-economic characteristics of those retiring in the twenty-first century, and how they may differ from previous cohorts of elderly persons. This research analyses data from the General Household Survey (1974–1996) to examine the social and economic experiences of four birth cohorts. Findings indicate significant differences in the living arrangements, health and access to resources amongst the cohorts, which will have implications for the retirement prospects of the post-war baby boom generations in 2020 and beyond.

INTRODUCTION Official government projections indicate that by 2026 the number persons aged 60 and over living in Britain will reach 17.1 million1. This is an increase of nearly fifty percent on the 11.7 million persons in the same age category in 1996. The changing demographic dependency ratios, and predicted increased pressure on the social security and welfare budget as the large post-war birth cohorts enter retirement age, has given rise to a burgeoning literature on the social and economic costs of an ageing population2 3. Little attention, however, has been paid to the likely socioeconomic characteristics of the post-war ‘baby boom’ generations in the first few decades of the twenty-first century, and how they may differ from previous cohorts of elderly persons. Will greater membership of occupational and private pension schemes and higher home ownership mean that tomorrow’s older population will be wealthier, no longer dependent on the state for income in retirement? Will they be healthier, making fewer demands on health and community services? What new calls will be placed on services and the family, and how can policy makers begin to plan for appropriate policy interventions? We cannot look into a crystal ball to see into the future. We can however, look back into the past and see how the experiences of the baby boomer generations to date are likely to shape that future. Wadsworth using data from the 1946 birth cohort argues that “a degree of anticipation is possible within lifetimes ... through those things which make an imprint on life at one point, and which are carried on into later life” 4 . This paper examines the social and economic experiences of four different birth cohorts over the period 1974–1996. The experiences of two cohorts currently in old and latemiddle age (born in the late 1910s and early 1930s) are compared with those of two younger cohorts, taken to represent those entering retirement in 2020–30 (born in the late 1940s and early 1960s). The analysis focuses on three key areas - living arrangements, health, and

27

Office for for National National Statistics Statistics


Spring 2000

access to resources - for each of the four cohorts, looking at how their experiences differ at the same chronological point in their life course. By adopting such an approach, it is hoped to shed light on what the baby boom cohorts might look like in 30 years time and how they may differ from older people today. The paper discusses the findings, drawing out the implications for the future demand for social care.

DATA AND METHODOLOGY There is no single source of longitudinal data in Britain that currently provides demographic and socio-economic information both across an extended time-period and for a range of different birth cohorts. The British Household Panel Study (BHPS) at Essex University is beginning to yield exciting and useful data, but it is restricted in that data collection only commenced in 1991. The ONS Longitudinal Study (LS) covers the period since 1971, but the range of socio-economic data included is limited. This research therefore uses a pseudo-cohort approach to examine the experience of four birth cohorts at the same chronological age and to compare the differences between them. The analysis uses data from two main sources. Firstly, vital registration data on family formation (marriages, births and divorce) is examined by birth cohort. Secondly, data from 23 years of the General Household Survey, a cross-sectional survey of individuals living in private households carried out annually by ONS, is used to generate pseudo-cohorts. The experience of representatives of each birth cohort is tracked in successive years of the GHS. For example, those respondents aged 55–59 in the 1975 GHS are taken as being representative of the birth cohort born in 1916–20. The same birth cohort is then represented by those aged 56–60 in 1976, 57-61 in 1977, and finally those aged 76–81 in 1996. This constitutes a pseudo-cohort or quasi-cohort approach since the individuals are not the same from year to year. Rather than tracking individuals per se, it is the group means or proportions that are taken. Thus the unit of analysis is the cohort and what is measured is the average experience for the cohort. The main advantages of using a pseudo-cohort approach are that it is less expensive in time and money as the data already exists; covers a long time-period, and there is no problem of sample attrition, as with panel data. Since a fresh sample is drawn from the surviving population each year, the cohort means remain representative of that population. However there are limitations to employing such a method. Firstly, it is not possible to examine the duration of events, such as marriage or unemployment, for individuals but only the average duration for the cohort as a whole. Secondly, as with any analysis of cross-sectional data over time, there may be problems with consistency of data. For example, the question on limiting longstanding illness changed in 1977 and 1978. This meant that these two years were excluded from the analysis of chronic illness in order not to jeopardise the overall findings. Thirdly, there is the problem of sampling error in the successive cross-sectional samples. The results shown do not indicate the size of the confidence intervals. However, the large sample size of the GHS means that the sample errors around the estimates are relatively small5. The middle response rate has fluctuated between 76 per cent and 85 per cent over the period 1974 to 1996. The sample size of each five year birth cohort varies over time as both the overall sample size of the GHS has changed and the share of the cohort in the sample population alters with age. Finally, the results are affected by both age and period effects which are often difficult to disentangle. Nevertheless, despite these limitations, pseudo-cohort analysis can provide useful insights into inter-cohort differences and help to inform policy making.

THE COHORTS The four cohorts selected are people born in the years 1916–20, 1931–35, 1946–50 and 1961–65. There is always debate as to arbitrary nature of cohort boundaries. The choice here reflects a set of people born in same

Office for National Statistics

28

five year time-period who consequently have many historical experiences in common. ●

Those born in 1916–20 represent today’s ‘middle aged’ elderly, being aged 80–84 in 2000. Twenty-five years earlier, in 1975, they were aged 55–59 and approaching retirement.

●

The cohort born in 1931–35 are today’s ‘younger’ elderly, aged 65–69 in 2000. In 1975 there were entering mid-life, constituting the group aged 40–44.

●

The cohort born in 1946–50 represents the ‘first baby boom generation’. They were aged 25–29 in 1975. In 2000 they are aged 50–54 and starting to plan for retirement.

●

The cohort born in 1961–65 represents the ‘second baby boom generation’. In 1975, aged 10–14, they were still at school. In 2000 however, they have reached their mid-to-late thirties. Education and training has largely been completed, and occupational pathways are well established, along with family life.

Marked differences are to be expected between the cohorts, as they were born into and grew up in very different economic, technological and social climates. The differences between the experiences of the post-war and pre-war cohorts are considerable, with neither of the 1940s and 1960s cohorts having experienced World Wars, and both growing up under the umbrella of the Welfare State, reaping the benefits of the 1944 Education Act and expecting to receive health and social care ‘from the cradle to the grave’ under the NHS and social services (Table 1). However, the experiences of the two post-war cohorts also differ from each other. The first baby boomers (1946–50) were born in a period of post-war austerity, experiencing rationing and selective education. However, when they entered the labour market the economy was entering a period of relative prosperity. Not only was the job market buoyant, but the rapid expansion of higher education in the 1960s also meant that a growing number stayed on at school and entered university. In addition to new opportunities in education and work, the introduction of the pill heralded a new sexual freedom. In contrast, the second baby boomers (1961–65) were born in a period of prosperity, experiencing the consumer spending boom of the 1960s, and comprehensive secondary education. But by the time this generation came to enter the labour market at the end of the 1970s, the economy was entering a recession, resulting in sharp rises in unemployment. Individuals born during the peak birth year of 1964 reached schoolleaving age in 1980, at the depth of recession. Some may have never had a permanent full-time job. The spirit of radicalism and freedom, enjoyed by the preceding baby boom cohort, was missing; not only in terms of employment and income, but also in other spheres of life with 1984 seeing the first person diagnosed with AIDS. These very different economic and social environments will have affected the respective life chances of the different cohorts, as well as cohort members’ expectations of employment, the welfare state and life in general. As the paper illustrates, the trend towards greater inequality is a persistent theme throughout the lives of the second baby boomer cohort.

CHANGING LIVING ARRANGEMENTS: THE RISE OF SOLO LIVING The overwhelming majority of social care is provided by family members 6. Current policy recognises this role and explicitly relies upon its continuance. But to what extent will it be able to continue to do so in 2020 and beyond? What are the likely changes in the family and in living arrangements? It is likely that ‘solo living’ will take on greater relevance in the new millennium. A number of factors will contribute to a rise in the


Table 1

Spring 2000

Age of birth cohorts at the time of selected political, economic and cultural events

EVENT

Birth cohort 1916–20

1931–35

1946–50

1961–65

1929 Wall Street Crash

9–13

1939 Hitler invades Poland

19–23

1942 Publication of Beveridge Report

22–26

7–11

1944 The Education Act establishes free secondary education as a universal right

24–28

9–13

1945 World’s first atom bomb (Hiroshima); end of World War II

25–29

10–14

1947 National Health Service launched

27–31

12–16

0–1

1951 Korean War; Listen with Mother launched

31–35

16–20

1–5

1961 First man in space; Cold War deepens with Bay of Pigs standoff and rise of the Berlin wall

41–45

26–30

11–15

0

1963 Nelson Mandela jailed; JF Kennedy assassinated

43–47

28–32

13–17

0–2

1967 Homosexual acts between consenting adult males and abortion legalised; NHS makes contraception available on ‘social grounds’ (to married and unmarried women)

47–51

32–36

17–21

2–6

4–8

1971 Divorce Reform Act (1969) implemented

51–55

36–40

21–25

6–10

1973 Britain joins the Common Market

53–57

38–42

23–27

8–12

1974 Wilson elected; Nixon impeached; inflation tops 20%; comprehensive education introduced

54–58

39–43

24–28

9–13

1975 Equal Pay Act (1970) and Sex Discrimination Act; SERPS established

55–59

40–44

25–29

10–14

1979 Margaret Thatcher elected; Iran takes US hostages; USSR invades Afghanistan

59–63

44–48

29–33

14–18

1982 Unemployment tops 3 million; Falklands War

62–66

47–51

32–36

17–21

1985 Fowler social security review reforms SERPS and freezes Child Benefit; Live Aid

65–69

50–54

35–39

20–24

1987 Margaret Thatcher elected for 3rd term; Black Monday

67–71

52–56

37–41

22–26

1990 Mandela freed; riots over poll tax; invasion of Kuwait; Germany reunified; end of Thatcher era as she resigns

70–74

55–59

40–44

25–29

1991 Communism ends and Soviet Union dismantled; GP fundholders and NHS trusts introduced

71–75

56–60

41–45

26–30

1993 Child Support Agency begins work; Maastrict Treaty ratified

73–77

58–62

43–47

28–32

1996 Job seekers allowance comes into force; Big Issue Celebrates 5th birthday

75–79

60–64

45–49

30–34

1997 Tony Blair and ‘New Labour’ elected

76–80

61–65

46–50

31–35

1998 DSS Green Paper A new contract for welfare: Partnership in Pensions published

77–81

62–66

47–51

32–36

78–82

63–67

48–52

33–37

1999 Lowest inflation for 35 years (1.1%); Royal Commission on Long Term Care reports

Source: Adapted from Evandrou, M. 1997 Baby Boomers: Ageing in the 21st Century. London: Age Concern.

29



Spring 2000

proportion living alone. Firstly, it is likely that fewer future elderly persons will be married or living in a co-habiting union than is the case today. The proportion of women remaining ‘never married’ amongst the four cohorts is shown in Figure 1. A higher proportion of women from the 1960s cohort remain unmarried in their early 30s than in previous cohorts at the same age. Similar trends are found for men. These trends partly reflect a shift towards later marriage. The key question here is whether a higher proportion of the 1964 cohort will never marry or whether they will eventually catch up? Looking at the position and slope of the line to date (Figure 1), it seems unlikely that this will be the case. If current trends continue, ONS projections indicate that over 10 per cent of women and 16 per cent of men from the 1960s cohort will not have formed a marital or be in a permanent co-habiting union by the time they reach the age of 50 7. This compares with 4 per cent of women and 8 per cent of men born in Figure 1

Proportion of women remaining ‘never married’ by birth cohort, England and Wales

Figure 2

Proportion of women who ever divoeced by birth cohort, England and Wales

Percentage 30 1946 25

20

1931

1964

15

10

5

Percentage 100

0 25

30

35

40

80

45 Age

50

55

60

65

Source: ONS 1999, Marriage and Divorce Statistics. 60

40

The proportion living alone at any given age rises across the cohorts (Figure 4). It is forecast that 41 per cent of the 1940s cohort will be living alone by age 75 compared with 38 per cent of the 1930s cohort and 37 per cent of the 1916–20 cohort. Furthermore, a quarter (26 per cent) of the 1960s cohort will already be lone householders by age 60. Given that the proportion living alone increases with age, particularly between ages 60 and 75, it is likely that close to a half of the 1960s boomers will be living solo by age 75. These trends will have important implications for social care in 2020 and beyond, particularly with regard to the availability of co-residential kin, and in particular spouses.

1964

1916

20 1931 1946 0 15

20

25

30

35

40

45

50

Age Source: ONS 1999, Marriage and Divorce Statistics.

1946. Secondly, among those who have married, a greater proportion will also have experienced divorce or separation. In fact, nearly one in six (16 per cent) women born in 1964 have already witnessed the break-up of their marriage by their 33rd birthday (Figure 2) 8. On the other hand, with increasing longevity by the time they retire fewer of this cohort will have experienced the break-up of their union through death. This will not, however, be sufficient to offset the rise in divorce. Thus the marked shift to later and fewer marriages interacts with rising divorce to increase the propensity to live alone. Will the 1960s cohort in old age experience similar, or higher, proportions of solo living than today’s elders? It appears that the 1960s cohort has already embarked upon a different living arrangements trajectory compared with earlier cohorts, with much higher proportions living in single person households at younger ages (Figure 3). Applying current proportions of people in different living arrangements within each marital status from the 1996 GHS to the 1996 based projections of marital status, it is possible to estimate the proportion of each cohort that will be living alone at ages 60 and 75 for the younger cohorts9 (Figure 4). This approach involves a ‘steady state’ assumption regarding the propensity within marital status groups to live alone, and given that in future more people may in fact choose to live alone than today (for example today’s never married may live with siblings whereas tomorrow’s never married may be less likely to do so), these figures may be taken as ‘lower boundary’ estimates.


30

Of course, spouses are not the only providers of care. Children are also an important source of social care and not just in later life. Rising divorce rates may impact upon the availability of adult children to care Figure 3

Proportion living alone by birth cohort, England and Wales

Percentage 50

1916–20

40

30

20

1961–65

10

1946–50 1931–35 0 15

20

25

30

Source: GHS 1974–96

35

40

45 50 Age

55

60

65

70

75


Figure 4

Proportion living alone at age 60 and age 75 by birth cohort, England and Wales

Figure 5

Percentage

Percentage

50

50

Age 60

Age 75

40

30

30

20

20

10

10

0 1916–20

1931–36

Percentage of women childless at age 30 and 40 by birth cohort, England and Wales

Age 30

40

1946–50

1961–64

Spring 2000

Age 40

0 1920

1931

Cohort Derived from ONS 1996-based, Marital status projection.

for their elderly parents and parents-in-law. On the one hand divorce and subsequent re-partnering may mean that there are more children per parent, i.e. both natural and step-children. But on the other hand little is known about how divorce may affect family ties and feelings of intergenerational obligation. Will children of divorced parents, and in particular of fathers who failed to maintain regular contact with them, show the same willingness to care? Similarly, will ex-daughter-in-laws be prepared to offer care to their ex-mother-and father-in-laws? Evidence from America found that divorced fathers were less likely to view their children as a potential source of support in later life 10. Adult children of divorced parents were also less inclined to view co-residing with their elderly parents as an option11. Since 1964 the overall trend in fertility in Britain has been downwards, suggesting that the baby boom cohorts will have, on average, fewer children than their predecessors. However, this does not necessarily mean that there will be fewer children to provide support for the baby boom cohorts in later life. In terms of the potential support for elderly parents from their children, it is whether you have any children at all rather than the number of children that is of central importance12 . Thus the proportion of each cohort remaining childless into old age is the critical variable. The proportion of women remaining childless by their 30th and 40th birthdays for the different cohorts is shown in Figure 5. It is predicted that by age 40 over a fifth (21 per cent) of women from the 1964 cohort will remain childless, compared to 14 per cent of women born in 1931 (today’s young elders)13. Such a level of childlessness is not unprecedented historically: women born in the 1920s also experienced rates of childlessness by age 40 at around 20 per cent. However, what makes it particularly challenging with the 1964 cohort is the size of the cohort and their longevity, with a greater proportion surviving into old age. It may be that the above projections over-estimate the degree of childless amongst the second baby boomers. Many women in this cohort have already postponed childbearing until their 30s. With changes in medical technologies by 2020 these women in their 50s and 60s may have additional options available to them (for example, IVF). Medical technologies may well extend fecundity as well as providing

1946

1964

Cohort Source: ONS 1999, Birth Statistics.

‘post menopause’ childbearing opportunities. It is already the case that a woman in her 60s has given birth to a healthy child in Italy. Given rising divorce and increased childlessness, relations with other relatives such as grandparents, grandchildren, nephews and nieces may also assume greater importance. Similarly, friendship networks will be more important among 21 st century elders, be they neighbours or ‘virtual’ friendships on the internet. Certainly, we are seeing the emergence of increasingly complex family structures, and new forms of family ties. Although there is no doubt that there will be more persons, in terms of absolute numbers, requiring support in 2020 than today, it is less clear who will provide it and what mix of care will be available and whether this will match the preferences of older people themselves. Demand for social care will also depend upon future trends in health and dependency, as well as people’s expectations.

H EALTH AND DEPENDENCY What is the evidence on the health prospects of older people in the next two or three decades? Will younger cohorts be healthier than their parents or grandparents? What do patterns in chronic ill health, health risk behaviour and work stress suggest? And what are the implications for demand upon health and social care? It could be argued that, having benefited from the NHS and having experienced improved living conditions in their youth and middle age, the post-war baby boom cohorts will go on to ‘age healthily’. Recent reviews of trends in mortality, morbidity and health risk behaviour, indicate that although improvements in mortality rates and increasing life expectancy suggest that we are indeed healthier14, trends in selfreported health status suggest there is no comparable general improvement in health15 . Life expectancy in old age has risen, however the extent to which these gains are disability-free is the subject of ongoing debate 16. Research indicates that as life expectancy increases, there will be a rise in the proportion of people experiencing light to moderate disabilities, but a fall in those with severe disabilities17. Evidence from the GHS on the pseudo-cohorts is mixed. A widely used indicator of health status is long-standing illness and whether this limits

31



Spring 2000

people’s activity. Figures 6a and 6b presents the proportion reporting a limiting long-standing illness (LLSI) using 3 year moving averages plotted against the mid-point of the age group for the cohort. As expected, the proportion of each cohort reporting a LLSI rises with age. The shape of the trajectory of the 1916–20 cohort peaks and troughs at older ages. This may be a selection effect, reflecting the fact that ‘less healthy’ people are more likely to have left the sample either through death or entry into institutional care (the GHS does not include the institutional population). More interestingly, successive younger cohorts appear to report slightly higher levels of LLSI than the previous cohort at the same age, and women report slightly higher levels than men. For example, by age 29, 11 per cent of women from the 1960s cohort report LLSI compared to 8 per cent of women in the 1940s

Figure 6

cohort. This may reflect a general increase in the propensity to report ill health regardless of age or cohort (i.e. a period effect); alternatively it may reflect real differences in the health status between cohorts. A further indication of future trends in health can be obtained by looking at the trends in health risk behaviour between the different cohorts. The GHS has collected information on smoking in alternate years since 1972. Overall, smoking among both men and women has decreased steadily over the last twenty years18. This decline in the propensity to smoke is reflected in the proportion of each cohort who currently smoke (Figures 7a and 7b) by age. Interestingly the downward gradient is strongest for men in all cohorts, demonstrating a higher proportion ‘giving up’ amongst men than women. By age 40, a

Percentage reporting limiting long-standing illness, by birth cohort (3 year moving average), Great Britain

Percentage 50

Percentage 50

Women

Men

1916–20

1916–20 40

40

30

30 1931–35

20

1931–35 20

1946–50

1946–50

1961–65

1961–65

10

10

0

0 15

20

25

30

35

40

45 50 Age

55

60

65

70

75

15

20

25

30

35

40

45 50 Age

55

60

65

70

75


Figure 7

Percentage currenty smoking by birth cohort, Great Britain

Percentage 60

Percentage 60

Women

50

50

1931–35

1931–35

1946–50

1946–50

1916–20

1961–65

1916–20 40

Men

40

1961–65

30

30

20

20

10

10

0

0 15

20

25

30


35

40

45 50 Age


55

32

60

65

70

75

15

20

25

30

35

40

45 50 Age

55

60

65

70

75


higher proportion of women smoke than men amongst the 1946 cohort, similarly by age 56 more women smoke than men amongst the 1931 cohort. In terms of this indicator of health risk behaviour, younger cohorts show lower levels of smoking than earlier cohorts. Given that smoking is related to a wide variety of illnesses and health states19, evidence of decreased prevalence amongst members of the younger cohorts in this study suggests that elderly people in 2030 will be healthier than today’s generation of elders at the same age. Health in later life may also be affected by quality of life during working years, such as the level of stress faced within the workplace20. Workers in Britain work the longest hours in Europe21. Furthermore, the British Social Attitudes Survey found that 30 per cent of British workers reported that their work was stressful and 44 per cent ‘often or always’ came home from work ‘exhausted’22. Evidence from the GHS indicates that men from the 1960s cohort have experienced longer working weeks at earlier stages in their life course than preceding cohorts (Figure 8). By age 30, 20 per cent of employed men in the 1961–65 cohort reported working 50 or more hours per week compared with 13 per cent of men in the1946–50 cohort, at the same age. Longer hours worked at younger ages may adversely affect health, directly or indirectly, both now and in later life. Figure 8

Proportion of men in employment working 50 or more hours a week by birth cohort (3 year moving average), Great Britain

Figure 9

Spring 2000

Proportion of men currently unemployed by birth cohort (3 year moving average), Great Britain

Percentage 20 1961–65

15

10 1931–35 1946–50

5 1916–20

0 15

20

25

30

35

40 Age

45

50

55

60

65

Source: GHS 1974–96 Percentage 30

ACCESS 1946–50

In addition to the health of elderly people and the availability of potential sources of informal care, their access to financial resources will also be a key determinant of the level of their demand on the state and their ability to enjoy an active life. Income in retirement, particularly pension income, is crucially dependent on past work histories. Post-war industrial change has resulted in the emergence of alternative working practices and the notion of a lifelong career has all but disappeared.

25 1961–65 1931–35

20

TO RESOURCES IN LATER LIFE

15 1916–20 10

5

Changes in the labour market have affected men and women differently. The proportion of men in employment at selected ages by birth cohort is shown in Figure 10. The combined employment

Figure 10

Proportion of men in employment by birth cohort (3 year moving average), Great Britain

0 15

20

25

30

35

40 Age

45

50

55

60

65

Percentage 100

1946–50


1931–35 1916–20

1961–65 80

In addition to longer hours of work, the 1960s cohort has experienced higher levels of unemployment, at earlier stages in their life course than the 1946–50 cohort (Figure 9). Unemployment is also related to ill health23, although the direction of causation is not clear cut. Increased levels of work-related stress may result in greater risks of physical and mental ill health in later life for the 1960s baby boomers, with consequences for the demand for social care.

60

40

In short, the evidence on the health of future elders is somewhat mixed. Thus it is premature to assume that tomorrow’s elderly will be healthier than today’s. Health status is complex and cannot easily be represented by a single indicator. Trends in smoking provide one indication that younger generations will be healthier in later life than current generations of older people. However, the proportion reporting limiting long-standing illness suggests the opposite; with younger cohorts reporting slightly higher levels than previous cohorts at the same age. Whether future elders will be healthier than today’s elders may also depend upon the impact of factors as yet unknown, such as pollution and advances in food technology.

20

0 15

20

25

30

35

40

45 50 Age

55

60

65

70

75


33



Figure 11a

Spring 2000

Proportion of women in full-time employment by birth cohort (3 year moving average), Great Britain

Percentage 70

Proportion of women in part-time employment by birth cohort (3 year moving average), Great Britain

Percentage 40 35

1961–65

60

Figure 11b

1946–50

30

1946–50

50

1931–35

25 40 20

1931–35 30

1916–20 15

20 10 10

1961–65 1916–20

5

0

0 15

20

25


30

35

40 45 Age

50

55

60

65

70

20

25

30

35

40 45 Age


trajectories of the different cohorts broadly outline the typical employment profile of men, rising at young ages, then levelling out between the late 20s and 50s and then declining as people progressively retire. However, it is clear that successive cohorts of men both enter the labour market later and leave earlier, and have lower overall participation rates at any given age. By age 30 the proportion of men in employment amongst the 1960s cohort is 84 per cent compared with 94 per cent amongst the 1940s cohort. There is also a 10 per cent difference at age 45 between the 1940s cohort and the pre-war cohort (i.e. 86 per cent as against 95 per cent). Interestingly, when comparing the two older cohorts, by age 60 only 53 per cent of the 1931–5 cohort remained in employment compared to 83 per cent of the 1916–20 cohort at the same age. This reflects the trend towards earlier withdrawal by men from the labour market. In contrast, women from the 1960s cohort experience higher rates of participation at any given age than the earlier cohorts, although their participation rates are still lower than those of men (Figures 11a and 11b). They are much more likely to work full-time than women in earlier cohorts; at age 26, 54 per cent of the 1961–65 cohort women are in full-time employment compared to 35 per cent of the1946–50 cohort women. The extension of women’s retirement age to 65 may well shift the employment trajectory out, which will mean that women will have longer to build up their pension entitlement. These trends have implications for the level of resources in later life. More women will have longer working lives enhancing their potential for accumulating pension entitlement in their own right. The recent legislation on pension splitting on divorce will also benefit women. The shorter working lives of men, however, will tend to have the opposite impact reducing overall pension entitlements. This may be exacerbated by the fact that the proportion of the 1960s baby boomers that had experienced unemployment by the age of 35 was very much higher than for any previous generation. The rise in structural unemployment during the 1980s, coupled with changes in the labour market, such as part-time, temporary and transitional jobs, has meant that the notion of a lifelong career has all but disappeared.


15

34

50

55

60

65

70

Both 1960s generation men and women are more likely to have private pensions (occupational and personal pensions) than current elders 24. In addition, more of the 1960s cohort will have access to housing wealth, both through their own tenure and through inheritance, than previous generations25 . The GHS indicates that between three-quarters and four-fifths of the 1940s cohort will enter retirement owning their own homes and an even greater proportion of the 1960s baby boomers will do so (Figure 12). Thus, rising membership of pension schemes and greater levels of owner occupation both point to the baby boomers being better off in retirement than today’s older people.

Figure 12

Proportion living in owner-occupied housing by birth cohort (3 year moving average), Great Britain

Percentage 100

90 1946–50

80

1931–35 70

1961–65 1916–20

60

50

40

30 15

20

25


30

35

40

45 50 Age

55

60

65

70

75


Spring 2000

DISCUSSION

Key findings

The demands on both the welfare state and the family in the 21st century will depend not only upon the living arrangements, health and access to resources of the baby boomers in retirement but also on individual attitudes and expectations. The 1960s baby boomers are better educated than previous generations and are likely to be more discerning consumers both of leisure activities and of health and welfare services. Social care agencies, purchasing authorities and Trusts in 2030 will be faced with clients as customers who will expect to be involved in the development of services. Many of the elderly boomers in the next century will be well-resourced in old age and able to purchase health and social care services outside the public sector, and as such will have many of their expectations met. However, a significant minority will not be so fortunate, as they will not have accumulated adequate resources across their working life. Any complacency concerning the retirement incomes of the new millennium elders may be premature. With the increasing emphasis on earnings related private pensions, inequalities in working life are perpetuated into retirement. In addition to those with better pensions, there will also be a group of individuals who have been unable to accumulate rights to a second tier pension, either because of interrupted earnings histories (e.g. carers) or low income. Increasing polarisation within future cohorts of elderly people, will be a key facet of 21st century ageing. Inequalities will persist between those who will experience full work histories, have acquired pension rights and housing wealth, and those who will not. The 1960s baby boomers, in particular, faced high unemployment levels when they first entered the labour market. Some of them have never had a full-time job, whereas others benefited from the Thatcher era, giving rise to the 1980s ‘Yuppie’. Inheritance of housing wealth primarily goes to individuals who are already owner-occupiers26, further reinforcing the trend towards polarisation and Richard Titmuss’ notion of ‘two nations in retirement’. The pseudo-cohort analysis presented here has been useful in two main ways. Firstly, it has served to refocus attention on the life course, and to highlight the fact that economic and social experiences earlier in the life course impact upon prospects in later life. Certain characteristics are carried through time (e.g. sex), others are accumulated over time (resources), many change over time (age and health) and some are also affected by the time period in which one lives (e.g. the welfare system experienced). Secondly, it has emphasised the importance of taking cohort differences into account within projections for social policy planning. In order to be better placed to plan for an ageing population in the 21st Century, further research is needed which investigates differences both between and within cohorts, exploiting the rich sources of birth cohort data within Britain.

3

4

5

6

ACKNOWLEDGEMENTS We are grateful to the Office of National Statistics, The Data Archive and MIMAS computer services for access to the GHS data 1974–1996, which has been used by kind permission. The analysis and interpretation of the data is the responsibility of the authors alone. The research forms part of the work of the ESRC SAGE (Simulating Policy for an Ageing Society) Research Group funded under grant number M565–28–1001.

7

8

9

10

REFERENCES 1

2

ONS. 1996- based National Population Projections. Series PP2 no. 21. London: TSO (1999). Johnson, P. and Falkingham, J. Ageing and Economic Welfare. London: Sage (1992).

11

12

●

Demographic trends have implications for the continued reliance on the family as a source of social care for tomorrow’s elders. – ‘Solo living’ will take on greater relevance in the new millennium. It is likely that a quarter (26 per cent) of the 1960s cohor t will be lone householders by age 60 and that close to a half of the 1960s Boomers will be living solo by age 75. – Over a fifth (21 per cent) of women from the 1960s cohort will remain childless, compared with 14 per cent of women born in 1931 (today’s young elders).

●

The evidence from the GHS on the health of future elders is mixed: – trends in smoking indicate that, other factors being equal, younger generations will be healthier in later life than current generations of older people; – the proportion reporting limiting long-standing illness suggests the opposite, with a higher proportion of the 1960s cohort reporting ill health than previous cohorts at the same age; – younger cohorts are also facing higher levels of stress in the workplace (for example, in terms of longer work hours) and higher levels of unemployment.

●

Successive cohorts of men both enter the labour market later and leave earlier, and have lower overall participation rates at any given age.

●

Women from the 1960s cohort experience higher rates of participation at any given age than the earlier cohorts and longer working lives enhancing their potential for accumulating pension entitlement in their own right.

OECD. Ageing in OECD Countries: A Critical Policy Challenge. Paris: OECD (1996). Wadsworth, M. The Imprint of Time: Childhood, History and Adult Life Oxford: Clarendon Press (1991). Breeze, E. Reporting on sampling error. Series GHS no. 18, OPCS. London : HMSO (1990). Allen, I. And Perkins, E. (eds). The Future of Family Care for Older People. London: HMSO (1995). These figures are derived using data from the Government Actuary’s Department 1996-based marital status projections. See Shaw, C. ‘1996-based population projections by legal marital status for England and Wales’ Population Trends 95 (1999). ONS. Marriage and Divorce Statistics, Series FM2 no. 24. London: TSO (1999). Hancock, R. Jarvis, C. and Mueller, G. The Outlook for Incomes in Retirement. London: ACIOG (1995). Goldscheider, F. ‘Divorce and remarriage: effects on the elderly population’ Reviews in Clinical Gerontology, 4, pp 253–9 (1994). Goldscheider, F. and Lawton, L. ‘Family experiences and the erosion of support for intergenerational coresidence’ Journal of Marriage and the Family, 60, pp.623–32 (1998). Knodel, J., Napaporn, C. and Siriboon, S. ‘The Impact of Fertility decline on Familial Support for the Elderly: an illustration from Thailand’ Population and Development Review 18 pp.79–103 (1992).

35



13 14

15

16

17

18

19

20

21

22

23

24

25

Spring 2000

ONS. Birth Statistics, Series FM1 no. 26. London: TSO (1999). Drever, F. and Whitehead, M. (eds) Health inequalities. London: TSO (1997). Dunnell, K. 1997. ‘Are we healthier?’ In Charlton, J. (ed) The Health of Adult Britain 1841–1994, ONS Series DS nos 12 & 13. London: TSO (1997). Robine, J.M. Mathers, C. Bone, M. and Romieu, I. Calculation of health expectancies: harmonisation, consensus achieved and future perspectives. Colloque INSERM/ John Libbey Eurotext Ltd (1993). Bone, M. et al. Health Expectancy and its uses. London: HMSO (1995). OPCS. Living in Britain. Results from the 1994 General Household Survey, London: HMSO (1996). The Carnegie Inquiry Into the Third Age. Health: Abilities and Well-being London: Carnegie Inquiry (1993). Theorell, T. ‘Stress at work’ in Proceedings: stress and coronary heart disease. Coronary Prevention Group. London (1986). Central Statistical Office. Social Trends 1996. London: HMSO. (1996). Curtice, J. ‘Satisfying work - if you can get it’ in R. Jowell et al (eds) International Social Attitudes: the 10th British Social Attitudes Report. London: SCPR (1993). Bartley, M. ‘unemployment and ill health: understanding the relationship’ Journal of Epidemiology and Community Health. Vol. 48: 333–7 (1994). Hancock, R. ‘Financial Resources in Later life’ in Evandrou, M. (ed) Baby Boomers: Ageing in the 21st Century. London: Age Concern (1997). Holmans, A. and Frosztega, M. House Property and Inheritance in the UK. London: HMSO (1994).


36