Number 5: Structural ageing, labour market adjustment and the tax/transfer system

This report was published by the former Department of Families, Community Services (FaCS).

Executive Summary

This paper is a revised version of ‘Older workers, disability and early retirement in Australia’ presented at the FaCS/ANU joint conference ‘Income support, labour markets and behaviour: a research agenda’ held in Canberra in November 1998. Note that the paper was finalised in 1999, and the figures and tables are only current up to that date. In particular, any trends arising from the introduction of the GST are not covered.

The paper raises the issues of population ageing and fiscal sustainability, the early retirement phenomenon and social security receipt among older retired people. It examines government policy initiatives introduced to reduce early retirement incentives and provides some alternative policy responses as a basis for further departmental work and public discussion.

In the 21st century, population ageing will place pressures on social expenditure in many countries including Australia. In Australia, the proportion of the population aged over 65—12 per cent in 1998—is projected to more than double to 24–26 per cent in 2051. The ageing of the population reflects the maturing of the baby boom generation, low fertility and rising life expectancies.

Early retirement has an important bearing on the extent and timing of projected social expenditure rises. These pressures are starting; the first of the baby boom cohort is now passing beyond age 50, and this is an age at which the incidence of social security receipt starts to increase dramatically.

There are features of the Australian retirement income system that appear to create an early retirement incentive. For example, the social security system (in marked contrast to the active labour market policy enforced for younger workers) tends not to treat many older customers as potential labour market participants. The apparent accessibility of Disability Support Pension (DSP), possibly as a form of early retirement pension, is a related issue. In addition, there are opportunities for ‘double dipping’ in the occupational superannuation system—retirees can obtain a superannuation lump sum prior to pensionable age and then spend or invest it in such a way as to then obtain social security benefits.

The Australian government has taken a number of initiatives aimed at reducing incentives to early retirement. These include: the standardisation of the retirement age for men and women by phasing in a higher age for women; the increase in the compulsory preservation age for occupational superannuation from 55 to 60; the phasing out of certain provisions that mainly benefit older women; the introduction of a larger reasonable benefits limit (RBL) for retirement benefits taken at least half in annuity form; and the introduction of the Deferred Pension Bonus Plan.

Further policy changes that may be used to reduce early retirement include: reducing inflows to DSP; restricting access to lump sums prior to age pension age; and raising the general age of pension eligibility.

The paper concludes that although fiscal sustainability is not a major concern, it is nonetheless an important issue. While parts of the costs of ageing are unavoidable, there is a policy concern if, as seems likely, there are found to be systemic incentives to early retirement. It seems inevitable that attention will need to be directed to all areas where there is potential for expenditure saving. Governments will necessarily be attracted to measures that raise, not lower, the average age of retirement.

[ Return to Top Return to Section ]

1 Introduction

Concerns about early retirement need to be understood in the context of population ageing and the associated costs it will impose. In Australia the proportion of the population aged over 65—12 per cent in 1998—is projected to more than double to 24–26 per cent in 2051. The number of the very old (those 80 and over) is projected to rise even more rapidly, from 0.5 to 2.2 million.

Earlier retirement appears to have been the result of a conjunction of influences including greater levels of personal wealth, economic restructuring, changing social attitudes and, possibly, the presence of more easily available income support provisions. It has been a trend observed in all Organisation for Economic Cooperation and Development (OECD) countries, and indeed Australia’s experience has not been greatly different from that of the OECD area as a whole.

Early retirement accelerates the rise in age dependency. The proportion of the population aged 50 and over is beginning to rise now, as the first of the ‘baby boom’ cohort pass into this age group. The incidence of social security usage in the 50–64 age group is 33.4 per cent. The World Bank and the OECD have both argued for discouraging early retirement and raising retirement ages as a means of coping with the fiscal stress caused by ageing populations.

In Australia, the trend to early retirement for older male workers has been evident for a considerable time, though this trend may have levelled off in recent years. While early retirement trends have received considerable attention, the recent reversal of this trend—which has also been observed in some other OECD countries—has received rather less attention. Older females in Australia have had relatively low labour force participation rates but these, unlike those for males, have been rising. Overall, aggregate labour force participation rates among older Australians are steady or even rising slowly.

This presents a paradox, since Australian Bureau of Statistics (ABS) and OECD measures of the average retirement age suggest Australians are retiring earlier. This paper suggests that retirement age statistics should be treated with some scepticism. Labour force participation rates, possibly adjusted for unemployment and average hours worked, are a better measure of the contribution of older workers to aggregate production and taxpaying capacity. Similarly, rates of social security dependency are the best measure of the fiscal burden of early retirement.

There is no ‘right’ retirement age, since the diversity of individual circumstances is such that flexibility will need to be the keynote in future policy development. However, it is important to recognise the effects of the retirement income system (in particular the Age Pension age as well as the age of compulsory preservation) on public perceptions of what age people should retire.

Community attitudes have been broadly supportive of early retirement for many years, and this support has been reflected in the provisions governing access to superannuation, social security payments and the like, and in employer behaviour and worker expectations. These attitudes

may have aided the extensive structural adjustments we have witnessed in recent decades, but they will no longer be appropriate to the demographic situation projected for the 21st century and beyond.

The current relative stability or rise in participation rates among the older working age group is not a reason for complacency: it has not been matched by any decrease in the proportion of this population receiving social security benefits. Given cost pressures projected to arise from population ageing and past and prospective increases in life expectancy, policy initiatives should aim at increasing the average age of retirement.

Early retirement is a concern both from the perspective of the fiscal sustainability of the tax/transfer system, and also from the perspective of the workers themselves and their standard of living in retirement. For the system, early retirement will add to the fiscal burden of an already rising age dependency ratio.

Early retirement is for many people involuntary, but there is also plenty of evidence that government policies can have a marked impact. A case in point is New Zealand, where the pension age fell to 60 but is now rising back toward 65. This has had a marked impact on actual retirement ages. It appears that, while retirement may not be voluntary for some people, many others do have some choice over when they retire.

Aspects of the Australian system that may be of particular concern are the widespread use of superannuation lump sums and the gap between the preservation age (now 55, but to become 60 between 2015 and 2025) and the Age Pension age. Another concern is the effect of the Age Pension means test, so that using up superannuation assets prior to pensionable age (sometimes termed ‘double dipping’) is relatively costless at certain asset levels.

The retirement income system comprises not only social security but also occupational superannuation. It is possible that interactions between components, such as the availability of preserved superannuation benefits prior to pensionable age, are contributing to early retirement decisions.

Recent initiatives such as the deferred pension bonus scheme and the ‘over-55’ rule (taking into account superannuation assets from that age) may have some impact, but the extent of this is yet to be evaluated.

The use of the social security system to support early retirement is also a concern, as reflected for example in the rapid rise in Disability Support Pension (DSP) paid to people over 50. Though tightening eligibility could be one response, this approach has difficulties; nor is DSP the only payment potentially available to those who want to cease work.

Another issue is the wide variety of non-work tested payments available to women in the 50–61 (and younger) age group. Raising the pension age for women may have little impact if prospective pensioners have access to alternative, similar payments.

One general approach is to reduce differences between payments and to extend ‘mutual obligation’ requirements. One option along these lines is a single workforce age payment (SWAP), which would reduce or abolish differences in the provisions for the various categorical groups in the system. A less radical alternative is a single payment for older clients; for example, the over-50s payment option.

There is only limited evidence about the effectiveness of various forms of labour market assistance for older customers. Evaluation of a pilot labour market assistance strategy implemented in 1990 indicates that wage subsidy schemes had the best outcomes; that outcomes for women were better than for men; and that it was possible and desirable to effect change in community and employer attitudes to older unemployed people. Information is not yet available about the methods tested by Job Network Members under the Employment Services Market Reforms.

Resumption of the trend to early retirement could have serious negative consequences for the efficiency and flexibility of the labour market. Reductions in the level and quality of labour would generally reduce the level and rate of growth of economic activity below the economy’s potential. Proposals to reduce unemployment by facilitating early retirement not only have major difficulties, but also usually do not achieve their aim. There was a tendency in some OECD countries in the 1970s and 1980s to implement such proposals; however, there is now a general view among OECD countries that early retirement does not lead to the creation of additional labour market opportunities for the unemployed.

[ Return to Top Return to Section ]

2 Population ageing and fiscal sustainability

2.1 Social expenditure projections

The ageing of the population will cause pressures on the pension and health systems in many countries in the 21st century. In Australia, the number of those aged 65 and over will rise from 2.2 million in 1996 to around 4 million in 2021 and 6 million in 2051. The number aged 80 or over is expected to rise from 500,000 to around 2.15 million. More importantly, the number of aged will be rising relative to the number of those working and paying taxes to support them. The age dependency ratio (the ratio of people over 64 to those of working age) will increase from 19.2 per cent in 1991 to a projected 32.2 per cent in 2051 (ABS 3222.0, 1998).

The child dependency ratio will fall, so it is sometimes argued that the total burden of dependency will not rise greatly. However, Retirement Income Modelling (RIM) Taskforce (1998) data show that from 2014 the total dependency ratio will rise sharply, then plateau at a new higher level around 2041.

The cost to all governments of an aged person is 2.3 times that of a young person and, for the Commonwealth Government, 4.1 times (Gallagher 1995). Not all aged persons can be considered dependent, and many in the working age group are also dependent. For these reasons, it is not appropriate to simply add the aged and child dependency ratios together.

Two alternative dependency ratios are the labour force and the retired dependency ratio. The labour force dependency ratio is simply the ratio of persons not working to those who are working. The retired dependency ratio is the ratio of those retired to those working. Both ratios can be further adjusted to a full-time equivalent work measure, thus taking account of the trend to part-time work. Bacon estimates that the retired dependency ratio, adjusted for part-time work, is currently around 48 per cent and will rise to more than 70 per cent in 30 years time (Bacon 1997, p. 38).

The ageing of the population is in part the result of the maturing of the baby boom generation, and in part the result of rising life expectancies. A male born in 1994 has an average life expectancy of 75 years; a woman, 81 years. At age 65 life expectancy for males is almost 16 years, and for females 20 years,1 and these expectancies are expected to continue rising at the rate of about one year a decade. Other important factors affecting dependency ratios are fertility and migration; in general it is possible to project such ratios with some degree of confidence, barring war or major catastrophe.

Clare and Tupule (1994, p. 17) demonstrated that within feasible limits, variations in the rate of immigration have a limited impact on the age structure of the population:‘Even the most ambitious migration program, by historical standards, would not eliminate a substantial increase in age dependency ratios’. McDonald and Kippen (1999, p. 3) show that ‘levels of immigration above 100,000 per annum add large numbers of people to the population with little impact on the age structure’.

ABS population projections indicate the number of people in older age groups is expected to increase rapidly until 2011 and continue increasing at a slower rate to 2051.

Projections of the ‘burden of ageing’ also need to take account of the proportion of the aged that will be receiving government benefits, and the proportion of the working age population that will actually be earning.

Little change is projected by the RIM Task Force in the proportion of the aged population that will be receiving pension. This is currently around 80 per cent, and may drop to 75 per cent under increased and better superannuation coverage. However, the average pension received will drop; some 30 per cent of age pensioners now receive a part rate, resulting in an average payment of about 90 per cent of the maximum. This latter percentage is expected to decline in the medium and long term, though the exact proportion will depend on prevailing interest rates.

Current projections are for continuing increases in women’s participation rates but continuing declines for men, particularly older men. In the short to medium term the overall balance is slightly favourable. Bacon (p. 37) notes that the early retirement dependency ratio ‘falls slightly into the next century reflecting the fact that female re-entry to the labour force far outweighs the increases arising from early retirement’.

2.2 Budget cost of population ageing

There have been a number of projections of the budget cost of ageing. Two important ones are Economic Planning and Advisory Council (EPAC)2 (Clare and Tupule 1994) and National Commission of Audit (NCA 1996). The EPAC paper was fairly sanguine about the costs of an ageing society, projecting a total increase in social expenditures of around 2.5 per cent of GDP over the period 1990–2051 (Table 1) after taking account of the 1994 retirement income reforms that introduced the superannuation guarantee (SG).3 More than half this increase was attributable to the health budget.4

EPAC assumed that per capita expenditure on welfare in each age group (apart from the age pensioner group) would rise at 1 per cent a year. This figure implies that the real level of benefits would keep pace with projected productivity growth of 1 per cent a year, so EPAC’s projections take account of the changing composition of the non-aged group but not the possibility of an increasing incidence of welfare receipt among some age groups.

The NCA stated that EPAC’s estimates were optimistic because:

• each new projection of the population seems to show an increase in the number of aged persons;

• the underlying growth in health costs due to increased use of technology has been underestimated (NCA 1996, p. 133).

The NCA used labour force projections by the RIM Task Force that expected continuation of the upward trend in female workforce participation and a fall in male participation in all age groups, but especially the 55–59 age group (see Figures 3 and 4 in NCA, pp. 130–131). However, projections for social expenditures, such as EPAC’s, assume constant age-specific transfer expenditures.

Projections of health and pension outlays provided for the NCA by the RIM Taskforce are also shown in Table 1. These projections are quite similar to those of Creedy and Taylor (1993), who calculated an increase in social expenditure from 20.4 per cent of GDP in 1988 to 27.4 per cent in 2031. However, these authors also showed how sensitive these projections were to the economic assumptions (such as rate of productivity growth) used.

Several points can be made about the NCA projections. First, they relate only to the aged, not the near-aged. Second, the growth in expenditure on pensions for the aged is very moderate, at 1.1 per cent of GDP over the projection period. Third, the bulk of the cost blowout—6.1 per cent of GDP—occurs in the health area. Such projections are particularly problematical.

Projections of health expenditure are subject to more than the usual uncertainties, and in contrast to projections of social security costs, there appears to be little consensus about likely trends in health costs. RIM has projected a rise in health expenditures by about 2 percentage points of GDP by 2041, which translates into an increase in government spending of about 1.5 per cent of GDP (OECD 1998).

This projection assumes that growth in age-adjusted per capita health expenditure slows to 1 per cent per annum (the rate since 1996) and that the long-run growth of labour productivity remains at around 1.25 per cent. Considerably larger cost increases would occur if age-adjusted health expenditure per capita were to continue growing at 2 per cent a year, the long-term average rate. On the other hand, continuation of current improved growth rates for economy- wide productivity will increase the financing (tax) base.

Table 1: Expected increase in costs to the year 2031 as a percentage of GDP

  EPAC central projection RIM/revised estimates
Aged Pensions +1.4 +1.1(a)
Health (public and private)(b) +2.1 +6.1
Education −1.2 0.0(c)
Net Impact +2.3 +7.2

(a) The EPAC projection refers to income support and services. However, the RIM projections refer only to age and veterans pensions.
(b) EPAC’s estimate was based on 1 per cent per capita real growth in health expenditure. RIM’s projection is based on 2 per cent per capita growth, which is more consistent with recent trends. A RIM projection based on 1 per cent per capita growth in real health costs would result in a cost increase of 1.3 per cent of GDP by 2031.
(c) Education costs have not been revised, but EPAC considered that it is possible that education and training expenditure may need to remain at around 6 per cent of GDP if the needs of the mature aged are to be met.

Source: EPAC, 1994 and Retirement Income Modelling Task Force. From NCA 1994 p. 134, Table 6.2.

In contrast to the NCA, the Australian Institute of Health and Welfare has projected a slow decline in total health expenditures, to only 6.8 per cent of GDP in 2031.5 All these projections ‘should be seen against the backdrop of relatively stable total health expenditures over the 1990s in most OECD countries including Australia, reflecting the impact of active policy interventions on both the demand and supply side. The long term perspective is also encouraging in Australia, with our total health expenditure within the range of 7–8 per cent of GDP for the last 25 years’ (Kalisch 1999, p. 8).

In general, the SG is projected to have only a modest downward impact on welfare outlays. This issue is discussed in detail in Appendix 3.

2.3 Can we sustain projected social expenditure increases?

EPAC (C&T 1994) concluded that a total rise in social expenditure of 2.3 per cent of GDP by 2051 (of which 1.4 per cent was attributable to social security transfers) was manageable by historic standards. However, the NCA ‘does not agree with EPAC’s optimistic appraisal of the feasibility or desirability of attempting to raise taxation in the future to cover the higher social expenditure burden’ (NCA 1996, p. 135). NCA’s reasons were:

  • a lack of confidence that income tax burdens could increase as a proportion of GDP on current policy settings, especially in any durable sense;
  • an assessment that indirect tax collections would be a falling proportion of GDP on the basis of (then) current taxation policy. (NCA pp. 135–137).

The C&T projections do not take into account the prospective increase to 65 of the Age Pension age for women (C&T, p. 32). On the other hand C&T assumes a SG 6 per cent higher than is current policy (see below), which will tend to create an offsetting error in terms of the impact on total pension outlays.

There is a more general issue here about the sustainability of the whole social expenditure system. While there is little systematic evidence of the extent of work or savings disincentive effects, it is possible that effective marginal tax rates are so high in some places that we are running into ‘Laffer curve’-type effects. Attempts to ease effective marginal tax rates (EMTRs) on some sectors of the population inevitably push these problems into other sectors.

A related issue is the decline in the PAYE tax base, with more workers using contractor/self- employee status. If these sources of pressure on the tax/transfer system are not adequately addressed, fiscal pressures from population ageing will not necessarily be easy to accommodate.

We do have a certain amount of latitude in relation to this sobering equation. In the immediate future the labour force is projected to increase at a faster rate than the total population, mainly due to continued rises in female participation rates.‘Over the next decade or so Australia will have an available workforce which will be relatively large, and dependency ratios which are relatively low by historical and world standards’ (Clare and Tupule 1994, p. 26).

Other points can be made about the budgetary implications of ageing and early retirement:

  • It is often assumed that because Australia’s spending on transfers is low by OECD standards, we have considerable flexibility to raise spending as need arises. In fact high direct tax rates have created circumstances where it may be quite hard to raise the government share of GDP;
  • International comparisons understate the cost of Australia’s retirement income system, which should really include the cost of mandatory superannuation and (more arguably) associated tax concessions. A compulsory contribution of 9 per cent of payroll will amount to almost 4 per cent of GDP, assuming the payroll share of GDP is 45 per cent (the long run average). This means the real cost of our retirement income system will become roughly 8 per cent of GDP;
  • Both the EPAC and the NCA projections pre-date the Government’s legislation to index pension rates to 25 per cent of male total average weekly earnings (MTAWE). However, since these projections incorporate pension rates that rise in line with projected productivity-cum-wage movements, this should not be an additional source of error.

2.4 Impact of early retirement

Projections of the age dependency ‘burden’ have tended to be based on the benign view that the relevant dependent group are those aged 65 and over; the first of the baby boomers do not reach this age until about 2011. But the average retirement age, as well as the average age at which people become dependent on social security, is nearer to 60 than 65.

Receipt of social security benefits is strongly age-related. Table 2 shows that incidence of social security benefits rises markedly in each age group, beginning at 22.1 per cent for those 50–54 and reaching 54.8 per cent among those 60–64. For those over 64 (not shown), 82.6 per cent receive either a full or part-rate payment under FaCS or DVA schemes.

Figure 1 compares the 60+ dependency ratio (relative to the population over 15) with that for those 65+. The former is markedly higher, by about 6 per cent, and also begins to move up quite quickly around 2003, not 2011.

Table 2: Actual DSS and estimated DVA customers aged 50–64 at June 1997

Age groups Total no. of persons in age group % DSS or DVA customers
50–54 1,094,228 22.10
55–59 852,604 33.40
60–64 722,718 54.80
Total 2,669,950 34.60
% of prime working age(a) 21.60  

(a) The prime working age is the population aged 15 to 64.

Sources: ABS Catalogue No. 3101, DSS Customers: A Statistical Overview 1997 and DVA Benefits Statistics Summary,June1997. From Perry 1998 Table 8.

Because of the very marked trend to male early retirement in the past 20 years, part of the baby boom generation is beginning to retire now. The oldest of the baby boom cohort (born after 1945) turned 50 in 1996, and will be 55 in 2001. From 1997–2010 the 50–64 age group will increase by 46 per cent. Growth will then level off as this group move into the 65-plus bracket.

Between 1997 and 2010 the changing age distribution of the population is projected to give rise to a 24 per cent increase in social security numbers compared to a 13 per cent increase in the total population (Perry and Whiteford 1998).

Changes in labour force participation could have a marked effect on the dependency ratio and associated costs of ageing. McDonald and Kippen (1999) have simulated the impact on the labour force dependency ratio (LFDR) of a return, over 20 years, to 1973 participation rates for men aged 35–64 years. The result of this projection is that the LFDR actually falls from 1998 to 2018, after which it rises again but only returns to its current level by 2048:‘That is, throughout the next 50 years, dependency measured in this way would be lower than it is now … This is a much more favourable result than would be the case in European countries’ (McDonald and Kippen 1999, p. 8). Australia is in a more favourable situation because its fertility rate has been higher and because we have sustained a moderate flow of immigration.

Figure 1: People aged 60 or more and people aged 65 or more as percentage of population aged 15 and over

Graph showing upward trend and projection for people aged 60 or more and people aged 65 or more as percentage of population aged 15 and over

Sources: ABS 3201.0 and ABS 3222.0 Series 1, 2, and 3.
Series 1: Assumes fertility declining to 1.75 births per woman by 2005–06 and net annual overseas migration of 90,000 per annum.
Series 2: Assumes fertility declining to 1.75 births per woman by 2005–06 and net annual overseas migration of 70,000 per annum.
Series 3: Assumes fertility declining to 1.60 births per woman by 2005–06 and net annual overseas migration of 70,000 per annum.

[ Return to Top Return to Section ]

3 The early retirement phenomenon

3.1 Trends in early retirement and participation

According to the OECD, the average retirement age for Australian males fell from 66.0 in 1950 to 61.8 in 1995. For females there was a fall from 63.6 to 57.2. Figures for other OECD countries show similar trends (see OECD 1998f).6 The decline for men and the current average age of retirement are the same as the OECD average; but the decline for women is greater than the OECD average and the average age of retirement is significantly lower.

The average age of retirement is, however, a somewhat ambiguous concept, especially given the marked changes in patterns of women’s labour force participation over the past 40 years. One suspects that if early ‘retirement’ to raise children is taken out of the picture, the average retirement age for females may be rising, not falling.

The ABS (1998a) states that in November 1997, the average age of retirement from full-time work was 48: that is, 58 for males and 41 for females. These ABS figures include many females who left the labour force early for family reasons and never returned. For both males and females, these ABS figures indicate a greater incidence of retirement in the ages 45–64 compared with 1992 and 1994.

In the light of this, it is perhaps surprising that we find, in the aggregate labour force participation rates for Australia over the past 20 years, no consistent evidence that people have been retiring earlier (ABS various, Cat No. 6203.0).

Aggregate labour force participation among 45–54 year olds has been rising more or less steadily, from fewer than 70 per cent in the late 1970s to more than 78.5 per cent currently. For those 55–59, it has been fairly steady at around 55 per cent and has, if anything, risen slightly after a dip in the 1980s: it now stands at 58.4 per cent. For 60–64 year olds, participation has been fairly steady at around 30 per cent (currently 31.8 per cent), though there was something of a dip in the early 1980s—ascribed to the passing through of the service pension cohort (which had access to pensions at age 60), which peaked in 1983.

Labour force participation figures can be misleading in several dimensions as indicators or early retirement trends. Bacon (1997, p. 32) defines ‘early retirement’ as retirement from career employment before a person reaches the pension age. It follows that ‘being classified as not in the labour force does not necessarily mean retired. Our estimates indicate that the retired only make up some 60 per cent of those classified as Not in the Labour Force’ (p. 33). Second, movements in participation rates and in retirement rates can be in opposite directions, especially for females.7

Notwithstanding Bacon’s comments—which cannot be faulted in a conceptual sense—statistics on the average age of retirement are not especially useful in establishing trends either in the costs of supporting early retirees, or in the communities’ ability to sustain those costs (in particular, the ABS figures, cited above, on women’s retirement age are of little use). What we need to know from the perspective of fiscal sustainability is what proportion in each age group is working, and what their aggregate hours and earnings are.

Labour force participation is an aggregation of three things—full-time work, part-time work and looking for work. The rising participation rate series disguises three separate trends:

  • a rising incidence of women’s labour force participation;
  • a rising incidence of part-time work;8
  • a rising incidence of unemployment.

It is important not to confuse the age-participation rate curve (which is quite steep) with a time trend—which it is not. While men’s full-time work has been falling at all ages (and particularly among those 45–65), there has been some tendency for participation to stabilise in recent years. Women’s full-time work has been rising, but less steeply).

The decline in full-time employment among older workers reflects trends in male employment at all age groups. A secular downward trend can be clearly discerned in male employment ratios over the period 1972 to 1997, amounting to about 10–15 per cent at lower age ranges but rising sharply to around 30 per cent for those 60–64. However, the employment ratios for 1991 and 1997 are virtually identical, indicating that this trend may now be slowing (abstracting from the full-time/part-time issue).

Quinn (1998, p. 5) argues that a similar breakout in early retirement trends has taken place in the United States since 1985, and in the Netherlands, Norway and (perhaps) Portugal. In contrast, these trends have continued in many other countries. Clearly, different conditions in different countries are playing a part. In particular, the different financial incentives to early retirement seem to be playing an important role.

Data from Jordan (2000, forthcoming) show very sharp upward shifts in employment ratios for females over this period, (about 20 per cent in the middle age ranges), but the upward shift is less marked for older women, at 10–15 per cent. In aggregate the upward shift for older women has not balanced the downward shift for men. Because about half of older women work part- time, total hours worked by older people have declined as a percentage of aggregate hours except post-1985.

The picture of labour supplied by older workers is not as comforting as the aggregate participation statistics appear to indicate. However, neither is it alarming, particularly given the indications that earlier retirement trends among men are levelling out.

Early retirement developments in Australia match trends in many other OECD countries, reflecting underlying changes in economic conditions and social values permeating the post- industrial economy. As such, it must be asked if they are amenable to policy intervention. My answer is a qualified yes, for reasons and on the evidence set out below.

3.2 Trends in the labour market experience of older working age people

It would be a mistake to be pessimistic about the labour market prospects of people aged 45–64 years. Between 1979 and 1998, the proportion of this age group in full-time employment has been relatively stable; it declined in the late 1970s and early 1980s, but then recovered. Overall, the proportion of this age group employed full- or part-time has increased from 56 per cent to 62 per cent.

The proportion of people aged 45 to 54 who are employed has increased from 67.5 per cent to 74.0 per cent, and the proportion employed full-time has been relatively stable at around 57.0 per cent. People aged 55 to 64 were adversely affected by labour market change during the late 1970s and early 1980s (mainly men), but the proportion in employment has increased from 37.3 per cent in 1985 to 43.7 per cent in 1998. The proportion of this age group employed full-time has also risen marginally, from 31.0 per cent in 1985 to 32.1 per cent in 1998.

Overall, the proportion of people aged 45 to 64 years participating in the labour market increased from 57.8 per cent in 1979 to 65.7 per cent in 1998. This increase has outweighed the increase in employment, so the unemployment rate for this age group has increased from 2.8 per cent in 1979 to 5.6 per cent in 1998; nonetheless, the proportion actually in work has risen from 55.0 to 60.1 per cent.

While there is evidence of higher than average labour market discouragement in this age group, mainly among 55- to 64-year-olds, the extent of ‘underemployment’ among older people should not be exaggerated. The unemployment rate of older people remains significantly lower than for people under 35. When account is taken of discouraged workers, older people’s ‘unemployment’ would be greater than for people of prime working age but would not be worse than that for people under 30 (see Department of Family and Community Services 1999, Attachment C).

An inverse measure of participation is the ABS measure of those ‘not in the labour force’ (NILF). The proportion of people aged 45 to 64 in this category increased from about 4 per cent in 1979 to about 44 per cent in 1985. Since 1985 there has been a continuous decline in the proportion of this age group who were not in the labour force; a trend driven by increasing women’s participation in the labour force over a period of stable men’s participation. Overall, the proportion of 45- to 64-year-olds not in the labour force was 34 per cent in 1998, down from 42 per cent in 1979.

The aggregate participation and NILF trends mask a more complex underlying picture that can reveal significant labour market disadvantage affecting some mature age people. United States research indicates that a job loss results in large and lasting effects on future employment probabilities, and these effects vary with the age of the worker.‘Displaced workers in their fifties are estimated to have a three in four chance of returning to work within two years after a job loss, whereas for a 62-year-old job loser, the probability is less than a third.

Once re-employed, men 50 and over face significantly higher probabilities of exiting the workforce than do workers who have not experienced a recent job loss; however the direction of this effect gradually reverses over time. The net outcome of these entry and exit rates is a substantial gap between the employment rates of men who have and have not lost jobs, that lasts at least seven years’ (Chan and Stevens 1999). While we cannot be sure how closely the Australian experience in this regard mirrors that in the United States, it does remind us that older workers are far from being a homogeneous group.

3.3 Labour market trends by gender

These generally favourable labour market trends—summarised below—mask some significant differences in trends for men and women. The labour market situation for women has improved quite dramatically over the past 20 years, and while the position of older men in the labour market clearly deteriorated in the 1970s and early 1980s, there have since been encouraging signs of stability. (This section is an edited version of that in FaCS 1999. More detailed information on trends for older men and women is provided in Attachment D to that paper.)

Labour force participation

Among men aged 45 to 54 there has been a small increase in the proportion not in the labour force (that is, neither employed or looking for work) from about 9 to 13 per cent. The proportion of 55- to 64-year-old men who are not in the labour force increased from 30 to 40 per cent between 1979 and 1985, but has been stable since.

This contrasts strongly with the situation for women. The proportion of older women who are not in the labour force has declined substantially over the past 20 years—from more than 50 per cent to 30 per cent for women aged 45 to 54, and from 80 per cent to fewer than 70 per cent for women aged 55 to 64.


Over the past 20 years the proportion of older men who are employed has declined. This decline occurred first and most dramatically for men aged 55 to 64 from the 1970s to the mid-1980s; however, the trend reversed around 1993 and there has since been an increasing proportion in work. The proportion of men aged 45 to 54 who are employed has declined by a comparatively small amount. For both age groups the decline in full-time employment has been only slightly greater than the decline in total employment.

In contrast, the proportion of women aged 45 to 54 who are employed has risen by more than 20 per cent to around 66 per cent, and the proportion employed full-time has risen by more than 10 per cent to just under 40 per cent. The proportion of women aged 55 to 64 years who are employed has also risen by more than 10 per cent to just over 30 per cent, with the proportion employed full-time rising from 12 to 15 per cent.


Unemployment for both older men and older women has increased over the past 20 years. Most of the increase occurred during the 1982–83 and 1990–91 recessions, with the latter having a more adverse impact. Older male unemployment rose more than did older women’s unemployment: slightly more than 5 per cent of men aged 45 to 54 and 4 per cent of men aged 55 to 64 are now unemployed, while just over 3 per cent of women aged 45 to 54 and about 1.4 per cent of women aged 55 to 64 are unemployed.

Though the unemployment rate for those 50–64 appears low relative to the average, this needs to be qualified to the extent that unemployed older workers have simply dropped out of the labour market. Measures of hidden unemployment and under-employment show that older workers do have difficulty in obtaining jobs. VandenHeuval (1999, Table 3) calculates a ‘revised’ unemployment rate taking into account discouraged jobseekers, and finds on this measure that unemployment may actually be more severe for older than for prime-age people.

In addition, older workers who become unemployed are more likely to remain so. They are disproportionately represented among the long term (52 weeks or more) and very long term (104 weeks or more) unemployed (VandenHeuval 1999, Table 4).

The increasing share of jobs held by older workers

Evidence that older workers are improving their position in the labour market can be found in their share of total jobs. People aged 45 to 64 were 24.4 per cent of all employees in August 1984 and 29.8 per cent in August 1998—at the same time as their share of the working age population increased only from 25.3 to 27.7 per cent. Put a different way, people aged 45 to 64 years received 46.6 per cent of the increase in employment from 1984 to 1998.

A closer look at what has happened to the 45–54 and 55–64 age groups shows that the labour market performance of the baby boom generation is markedly stronger than that of previous cohorts of workers. People aged 45 to 54 held 15.8 per cent of all jobs in 1994, when they were 13.0 per cent of the population. Now that the majority of this age group consists of baby boomers, they hold 21.5 per cent of all jobs but constitute only 16.7 per cent of the population.

From 1984 to 1998, people aged 55–64 have also performed strongly in holding jobs. Their share of jobs declined only slightly from 8.5 per cent to 8.3 per cent, while their share of the population declined from 12.3 per cent to 10.9 per cent.

Older jobseekers: Australian Social Trends, ABS 1999

This ABS publication reports on findings from the Survey of Employment and Unemployment Patterns (SEUP), a longitudinal survey conducted over 1995–97.

The relevant group is older jobseekers aged 45 to 59: that is, a group we know from existing analysis to be disadvantaged in the labour market. The story from SEUP adds to the existing picture from cross sectional data sources.

Some 19 per cent of all jobseekers were in the relevant age group—168,000 out of 875,000 in May 1995. During the period surveyed about a quarter found jobs (albeit not always stable or well-paying ones); another quarter moved out of the labour market, and the balance had not worked. Jobseekers aged 55–59 were particularly unsuccessful in finding work. Older jobseekers came disproportionately from less skilled occupations and had relatively few educational qualifications. Being considered too old by employers was the major reason given for not finding work.

The experience of older jobseekers should not be confused with the experience of older workers in general. As the section above makes clear, older workers have managed to secure their full share of jobs created over the past decade. Moreover, their unemployment rates are not higher than the average, though these rates conceal the effect of labour force withdrawal.

3.4 Future trends in older workers’ labour force participation and retirement patterns

It seems clear that the strong trends toward early retirement evident among males in the 1980s are now moderating. Bacon (1997, p. 32) suggests that ‘the increase in early retirement has slowed and … it might even have stabilised’. Bacon shows age-specific retirement rates from full-time work. The S-shaped curves for both males and females in each of the 45–49, 50–54 and 55–59 age groups appear to indicate we have been through a transition to a new equilibrium that may prove to be sustained.

There are several factors that might be expected to slow the tendency toward early retirement, including:

  • the passing out of the system of the service pension cohort that was eligible for pension at age 60. Numbers in this group peaked in 1983;9
  • increasing labour force participation by women (though this has been less marked among older women);
  • a slowing of the pace of economic restructuring after the early 1990s recession—both the availability and need for lump-sum inducements for older workers to retire can be expected to decrease;
  • attitudes to older workers can be expected to change as the population ages; State and Territory legislation also now inhibits discrimination against them.

On the other hand, factors such as the spread of occupational superannuation will tend to make early retirement more affordable for some. Superannuation coverage of the workforce has increased from about 40 per cent in 1983 to 92 per cent (NCA 1996, p. 323). Increased usage of some social security benefits such as Disability Support Pension (DSP) will also support earlier retirement for those with some level of disability.

The RIM Unit has produced projections of labour force participation by age and gender from 1995 to 2056 (see Perry 1998, p. 25). RIM anticipates continued strong increases in female participation rates until 2026, with a more modest rate of increase thereafter and small declines (by at most a couple of per cent over the entire period) in male participation rates. There will be virtually no change in participation rates of 60- to 64-year-old males.

Of some interest is the projection that females aged 45 to 54 will attain the same participation rate as their male counterparts by 2051. These projections are reasonably consistent with shorter-term projections by ABS (1995) and NCA (1994, pp. 130–31)—the latter using earlier RIM projections.

Irrespective of whether age-specific retirement rates are stabilising, the number of early retirees will increase due to the impact of the baby boom cohort passing into the 50-plus bracket. To this extent early retirement will create spending pressures in the social security system. What we can say, on the basis of current workforce trends, is that these do not appear likely—barring major recession or other catastrophe—to be in excess of the capacity of a growing tax base to finance them. Further, we have a five- to 10-year ‘window’ before demographic effects begin to overwhelm growing labour force participation.

[ Return to Top Return to Section ]

4 Reasons for early retirement Some hypotheses

4.1 Some hypotheses

Explanations for early retirement include the observation that the generation of workers now retiring has enjoyed relatively secure economic conditions for much of their working lives and have accumulated wealth, homes and the like, allowing them the opportunity to choose leisure over further work. Other factors adduced are the increased incidence of second income earners in families, and the rising spread of occupational superannuation. Finally, the availability of social security benefits may be important.

Quinn (1998, p.6) notes that ‘the simplest explanation for the long-run decline in the labour supply of older American workers is wealth … For many current retirees, this general national prosperity has been augmented by windfall gains from two sources—a strong real estate market and large and unexpected increases in social security benefits’. While the Australian Age Pension is very different from United States social insurance, the same general comment could well apply here. Some researchers have attributed most of the decline in the participation of older United States men to rising social security wealth; others about a third—still a substantial fraction (Quinn 1998, p. 7).

Also important have been employer pension plans—especially those of the defined benefit variety—that contain incentives to early retirement. Anderson, Gustman and Steinmeier (1997), focusing on the 1969–89 period, attribute about a quarter of the reduction in men’s full-time work to changes in employer pensions and social security together.

So why has this changed in recent years? One United States hypothesis is that changes in incentives to retire have had and will continue to have a permanent impact on work by older Americans. The other is that temporary cyclical factors are responsible, and that when the current upswing ends the long-run decline will resume. Quinn (1998) attempts to disentangle long term and cyclical influences using regression techniques and concludes that, even allowing for a strong economy, there has been a clear break-out from previous trends.

‘Although the strong economy is important, I believe that there is also a new attitude to work late in life, encouraged by public policy initiatives that have changed the incentives facing older Americans. These incentives work’ [Quinn 1998, p. 9, emphasis added]. While the choices and incentives facing retiring workers are very complex,‘considerable research using the [Retirement History Study]10 has established that workers do behave as though they understand and respond to both Social Security and [employer] pension incentives’ (Quinn 1998, p. 16).

In Australia, as retirement savings increase there will be greater potential for people to finance early retirement through a combination of investment income and, later, access to social security benefits. There may not be adequate safeguards in the system to prevent this.

The idea that accumulated wealth and the increased incidence of second income earners in the family has enabled many to choose early retirement seems inconsistent with the observation that many older Australians are on means-tested social security payments related to unemployment and disability. There is also evidence from surveys that some early ‘retirement’ is involuntary: those involved do not realise that they have in fact retired, like it or not, until they have been unsuccessfully seeking work for some time.

Some of the contemporary employment problems experienced by older workers may have their origins in relatively low average education and skill levels (OECD 1998c, p. 50). There is a cohort of older workers in their 50s and 60s who are relatively poorly educated compared to younger people, and who have a relatively poor capacity to adapt to changing technology and industrial structures. Individuals in this generation have on average comparatively low levels of educational attainment and have tended to work in relatively unskilled and blue-collar jobs.

In 1995, less than half of workers aged 50–64 had attained post-secondary educational qualifications, and more than 40 per cent had not completed the equivalent of year 12 education. The educational profile of older workers is expected to improve markedly over time (Table 3).

Data on computer use by age also indicate that all age groups under 55 have vastly greater levels of exposure to computer technologies than do older people (FaCS 1999, Table 2). In the future, older working age people should be less disadvantaged in these respects than previous generations and this, together with their higher level of education, is likely to have significant positive implications for their employment prospects.

We might therefore expect to see a diminution of current trends to early retirement (particularly the involuntary component) as better educated and more adaptable younger people pass through the relevant age brackets. Also, greater mechanisation and reduced requirements for hard physical effort should reduce the impact of age disadvantage in employment, though a conflicting trend is the tendency for some new technologies to make life more difficult for some older workers.

Employers may have negative perceptions about the abilities and skills of older workers; for example, that they are more likely to use sick leave and workers’ compensation. Employers may also be reluctant to invest in the human capital of older workers. Further, older workers may be less mobile, as reflected for example in their higher rates of home ownership. Older workers, especially males, may tend to concentrate their job search activity within localities, occupations and industries in which they have had work experience and feel most comfortable and which may not be in growing areas of the economy.

Table 3: Percentage of Australians by age that had attained at least upper secondary education (1993)

Age group Percentage
25–34 57
35–44 56
45–54 51
55–64 42

Source: OECD education database.

Some researchers have claimed that there is no hard evidence to support negative employer stereotypes of older workers. Others have argued that there is some such evidence, but it is far from comprehensive (OECD 1997). This issue is pursued in Section 4.3.

In a 1993 ABS survey of the unemployed, 44 per cent of respondents aged 45 and over identified age as the greatest single obstacle to obtaining employment. At age 55, this jumped to 65 per cent. A Sydney Morning Herald opinion survey found that almost 50 per cent of those aged 55 and over felt uncertain about their chances of finding a job if they were sacked. Optimism about being able to continue in employment receded with rising age (Encel 1998, pp. 2, 6).

In a like vein, an ABS panel study of unemployment over 1994–97 found, of the 168,000 job-seekers aged 45–59 in May 1995, by September 1997 only 22 per cent were in stable work, 13 per cent were in unstable work, and 43 per cent had not worked at all. Thirty-two per cent were absent from the labour market at September 1997. Of those who had not worked and were still looking for work, 50 per cent reported that their main difficulty in finding work was that they were considered too old by employers (ABS 1998b, p. 20).

4.2 Voluntary or involuntary?

RIM modelling suggests that there are three distinct retirement peaks for men, the initial one at age 55 (the preservation age); the second at 60 (the retirement age in many superannuation schemes), and the largest at age 65 (pension age). These peaks—which are also found in most countries—are strongly suggestive of at least some policy impact on retirement decisions.

Involuntary retirement is normally the result of retrenchment followed by a period of unemployment and active job-seeking and a gradual realisation that a new job will not be forthcoming. Voluntary retirement is related to lifestyle choices either not to continue working or not to actively seek work following redundancy or retrenchment. In practice there is likely to be a considerable grey area between these extremes. Older workers may agree, for example, to accept a redundancy package and be prepared to take the chance that they may or may not be able to find other work thereafter.

In examining labour market statistics, older workers appear not to suffer from any marked disadvantage in the labour market. However, the average duration of unemployment is greater for older age groups:‘this suggests a hard core of unemployed older males with an average duration of unemployment at close to two years’ (O’Brien 1998, p. 14).

Many older workers face barriers to their continuing employment or re-employment. The availability of lump sum termination payments at age 55 has made this a de-facto early retirement age. In one survey, most employers thought of an ‘older worker’ as one about 50 or, in some cases, as low as 45. The majority of 55- to 64-year-olds believed that age discrimination begins at or below 45 (Unikowski 1996, p. 27). Certainly, older workers are now more likely to be made redundant or retrenched than are younger ones, in marked contrast to earlier ‘last on, first off’ policies.

4.3 Barriers to employment of older workers

A recent OECD study has examined perceptions and evidence with regard to the productivity, adaptability and trainability of older workers. Common perceptions are that such workers are:

  1. more costly than younger workers, due to age/seniority based payment systems;
  2. less productive, because of poorer health or fewer and less up-to-date qualifications;
  3. less willing/able to learn new skills;
  4. having shorter expected stays with the company, thus not making it worth investing in their skills (OECD 1997, p. 4).

The OECD examined each of these perceptions. It found that the cost argument is ambiguous; seniority and age-based pay systems are seen to be disappearing, and it may be that ‘unobserved’ characteristics increasing the productivity of long service workers are being rewarded rather than seniority per se. However, the OECD conceded that the pension costs of older workers are generally higher, noting evidence that sickness absence is higher among older workers but that casual absence is generally felt to be more prevalent among young people.

On the productivity question, the OECD survey found little systematic evidence of difference between the job performance of older and younger workers, though this may depend on the nature of the task, with a negative relationship between age and performance where the job content is changing rapidly. It is likely that there are certain jobs for which older workers are well suited.

As for qualifications, it is self-evident that older people are less highly qualified than young people. However, this situation will change: the older workforce of the future will be much better qualified, and ‘certification’ has meant that many who in previous times would not have acquired a formal qualification do so now, though this does not necessarily bring greater skills than the on-the-job training their older colleagues have had.

With respect to adaptability/trainability, the findings of industrial gerontological and occupational psychology research tend to decry the common view of older workers as less flexible and less easy to train than younger workers, though they do suggest that training regimes might be different for different types of worker. The OECD (1997, p. 12) concedes that older people have less experience in training and, for this reason alone, might be less ‘trainable’. Another problem is that older people are—or at least feel themselves to be—less likely to be rewarded for undertaking training.

The last argument against training older workers is that it is not worthwhile because there is relatively less time for the employer to earn a return on the cost of providing that training. Older workers may similarly be less likely to invest personally in training:‘prima facie, this argument is odd. It is only necessary to look at survival rates by age to see that older workers have the longest tenures, except insofar as these are truncated by retirement … Only among the very oldest age group, those close to retirement age, does the proportion staying on for at least five years tail off’ (OECD 1997, p. 13). It is the younger people who are more likely to receive training, yet these also are much more likely to leave.

The OECD survey concludes that if employers do not value older workers, it could be because they are acting on the basis of wrong information or wrong understanding.

4.4 Evidence on ‘push vs pull’

Once unemployed, older workers face a high risk of remaining unemployed because of negative employer perceptions regarding their efficiency. Disability, sickness or lack of skills may be further obstacles, and re-training opportunities are often felt to be the preserve of younger workers. The trend towards part-time and casual work may also mean there may be fewer opportunities to obtain work based on the traditional 40-hour week.

A 1996 ABS survey appeared to show that the extent of induced or involuntary early retirement is quite marked compared with voluntary early retirement (Cornish 1997). In total, 81 per cent of male retirees aged 45 or over and 64 per cent of women reported ‘induced’ retirements motivated by constraints such as employment problems, health problems (both of self and others) and compulsory retirement or redundancy.

Unikowski (1996, p. 26) reports:‘around one-fifth of male early retirees and one quarter of females gave their preferences for more leisure as the reason for their retirement … Ill health or injury is of far greater significance as a cause of male early retirement than any other voluntary or involuntary factor, accounting for almost half of male early retirements. The reasons for women’s early retirement are more evenly divided between ill health and preference for leisure and family reasons, in that order’.

However, the respondent’s health status is self-assessed: ill health may be a more socially accepted reason for early retirement than the preference for leisure, and is a prerequisite for receipt of DSP (Ruhn 1994, cited in Unikowski 1996, p. 26).

In terms of the contribution of adverse employment conditions to early retirement, Cornish (1997, p. 14) suggests that ‘the male discouraged job seekers who had retired early over the period from 1992 to 1996, typically retired after a sustained struggle to find full-time work. On average these men reported that they had stopped looking for full-time work approximately three years after they had finished their last full-time jobs’. However, discouraged jobseekers represented less than a quarter of male early retirees for this period, and less than 10 per cent for women.

Cornish concludes that there are two distinct early retirement patterns. The first is ‘voluntary’; the second and more common pattern was ‘induced’ in the sense that while some element of choice may have existed, the retirement was actually initiated by factors beyond the individual’s control:‘in total, approximately four in five men and two in three women who retired early over this period reported retirements consistent with the concept of induced early retirement’ (1997, p. 14).

There are difficulties in interpreting this survey evidence. There is some concern in the literature, for example, about the utility of self-reported health status as a measure of actual health problems. Poor health may be used to rationalise early retirement, being seen as more socially acceptable than a simple preference for leisure. Another problem is that the results of these surveys may reflect the particular economic conditions prevailing after the recession of the early 1990s rather than any more general tendency.

In the United States, self-reported survey evidence on reasons for retirement is now regarded with a great deal of scepticism (see Quinn et al. 1990, Chapters 2 and 3). A person in poor (but not disabling) health is likely to give health as a reason for retirement, whereas the real reason was the interaction of poor health and the availability and amount of social security and/or retirement benefits. The two sets of forces interact deeply and extensively. Answers to these sorts of questions have also been found to depend on the exact manner in which they are asked. (Quinn et al. 1990, p. 47).

Modern research using longitudinal data sets and sophisticated measures of financial incentives tends to give answers to these questions in direct contradiction of the earlier tradition of self- reported reasons for retirement. Such data ‘[permit] researchers to study the objective circumstances that existed prior to the behaviour under study and, thereby, to supplement the explanations given by the respondents themselves … Early evidence suggested that the overwhelming reasons for retirement were health and labour market constraints, and that very few people retired voluntarily. Modern research suggests the opposite—that many Americans do leave their career jobs voluntarily’ (Quinn et al. 1990, pp. 122–23). It is also possible there has been a real trend over time, with an increasing proportion of voluntary quits.

For Australia, Cornish found induced retirement to be associated with reliance on government pensions/benefits. It is interesting that around one-quarter of voluntary early retirees in Cornish’s survey also reported pensions/benefits as their main source of income, consistent with the view that social security is being used to support some voluntary early retirement in Australia.

There is further evidence for this in the retirement intentions data. Among those intending to retire early, lifestyle and financial reasons predominated across all income groups. Whereas almost all the men in the high income group expected to be self-supporting, this was less true among low and middle income groups, with about 40 per cent of these men placing emphasis on government pensions/benefits and part-time work as post-retirement income sources (Cornish, Table 17, p. 25). Further, some of those expecting to be self-supporting may find that these expectations are not realistic.

There was negligible evidence of substitute retirement, which involves the wife working instead of the husband. Instead, retirement was typically a joint decision (p. 17).

Around two-thirds of the men retiring early in this sample reported receiving a superannuation lump sum, and this figure applied both to voluntary and induced retirements (Cornish, p. 18). However, the size of lump sum payments was significantly greater in the voluntary group.

The ‘push vs pull’ issue may be important from the viewpoint of public policy. If the system is seen as inducing people to retire early, this would appear to create a stronger case for policy change than if people are being forced out reluctantly and with little element of choice. Further, it may make little sense to retain older people on work-tested payments if work is not in fact available. However, this black-and-white dichotomy of choice is not really sustainable in an environment where, for example, employer behaviour may also be conditioned on the availability of certain social security and superannuation benefits for early retirees.

The debate here is about supply vs demand theories of employment. Those attuned to the view that the real problem is inadequate demand for older workers will argue for sympathetic government policies and avoiding coercion. The other view is that greater supply will induce greater demand—through, for example, falls in relative wage rates, pressures on employers to retrench less selectively, and financial pressures on retirees not to leave jobs.

There is plenty of evidence that demand for older workers is not a fixed quantum, and that social security policies impacting on labour supply can have a very marked effect on the average retirement age. However, the demand side is also important and the two sorts of policies will need to work in tandem. For example, any measures to increase activity testing for older clients are likely to be dependent for their success and community acceptance on favourable employment conditions11 and, possibly, complementary labour market and training policies.

4.5 Sources of funds for early retirement

Table 4 shows the main income support payments received by people aged 50 to 64. The increase in the number of 50- to 64-year-olds receiving income support since 1978 now costs an extra $2.6 billion a year. About half of this is due to the ageing of the relevant part of the population, the other half to the rising incidence of social security dependence. This is despite the rise in labour force participation over the same period.

Table 4: Main income support payments received by people aged 50 to 64

  Men Women All persons
Disability Support Pension 204,600 79,200 283,900
Unemployment payments 81,000 28,600 109,600
Mature Age Allowance 52,800 600 53,400
Partner Allowance 4,100 50,300 54,400
Widow B Pension/Widow Allowance N/A 34,600 34,600
Wife Pension/Mature Age Partner Allowance N/A 93,500 93,500
Age Pension N/A 208,300 208,300
Other minor payments 29,000 26,200 55,200
TOTAL 371,500 521,300 892,900

Source: DSS and DVA data for June 1997. From FaCS 1999 Table 4.

For those aged 50 to 64, aggregate dependence on social security in 1997 was 34 per cent, up from 27 per cent in 1978 and 31 per cent in 1988 (Table 5). The percentage of those not in the labour force (NILF) or unemployed peaked at 54 per cent in 1983–85, and has since declined to 46 per cent.

Subtracting the percent of the 50–64 population on payment from the percent NILF and unemployed gives a residual that is the presumed incidence of self-support.12 This peaked at 22 per cent in 1985 and has since fallen to 12 per cent—which would appear to be anomalous in the face of the expectation that more rather than fewer early retirees would now be self-supporting.

Disaggregating these figures by sex provides an insight into this phenomenon (see Tables 6 and 7). For older males, the presumed percentage on self-support is now only 7 per cent, similar to the figure prevailing in the late 1970s, after reaching a high point of 12 per cent in 1991. For females, however, there has been a steady fall in this percentage, from around 35 per cent at the beginning of the period to 18 per cent today.

Table 5: DSS, DVA and DEETYA payments for persons in the age range 50 to 64 years, 1978 to 1997 (20 years)

Year Total DSS, DVA and DEETYA customers Population 50–64 yrs (ABS) % of 50–64 yr olds on payment % of 50–64 not in labour force or unemployed (ABS) Presumed % of self or spousal support No. of payment types
1978 553,970 2,066,318 26.8 47.5 20.7 15
1979 570,042 2,083,984 27.4 48.5 21.1 15
1980 584,007 2,102,091 27.8 48.6 20.8 15
1981 594,863 2,128,345 27.9 49.6 21.7 15
1982 614,607 2,148,521 28.6 50.7 22.1 15
1983 670,999 2,169,291 30.9 53.5 22.6 15
1984 694,189 2,190,880 31.7 51.9 20.2 16
1985 696,747 2,200,612 31.7 54.2 22.5 16
1986 693,837 2,211,820 31.4 51.9 20.5 16
1987 673,797 2,223,997 30.3 52.1 21.8 16
1988 688,042 2,241,301 30.7 51.2 20.5 16
1989 675,984 2,262,585 29.9 49.5 19.6 16
1990 666,827 2,285,676 29.2 48.6 19.4 16
1991 714,517 2,309,752 30.9 49.2 18.3 16
1992 737,074 2,338,897 31.5 50.2 18.7 17
1993 794,035 2,366,817 33.5 50.4 16.9 17
1994 830,595 2,419,573 34.3 48.8 14.5 17
1995 871,781 2,484,770 35.1 46.9 11.8 20
1996 874,177 2,552,814 34.2 46.9 12.7 22
1997 893,154 2,669,550 33.5 45.9 12.4 22
Change over 20 yrs 339,184 603,232 6.6 p.p −1.6 p.p −8.2 p.p 7

Source: Derived from FaCS 1999, Attachment E, Table 3.

By contrast, the percentage of the older female population receiving income support has been almost steady, at around 40 per cent. The substantial (14 per cent) rise in older female workforce participation over the period appears to have offset previous private transfers (down 17 per cent) and to have had virtually no effect on public transfers. This is consistent with the labour market providing opportunities for older married women.

Part of the story here concerns the distribution of employment among and within families. Women who have moved into jobs have tended to have partners who are also employed. Jobless men, however, tend to have jobless wives. The increase in the number of older working age families where neither partner works has resulted in a consequent increase in dependence on income support. It is not clear whether this phenomenon reflects the growth of an ‘underclass’ in Australian society, or whether it reflects the adverse work incentives that arise once one member of a couple becomes dependent on the social security system.

Table 6: DSS, DVA and DEETYA payments for males in the age range of 50 to 64 years, from 1978 to 1997 (20 years)

Year Total DSS, DVA and DEETYA customers Population 50–64 yrs (ABS) % of 50–64 yr olds on payment % of 50–64 not in labour force or unemployed (ABS) Presumed % of self or spousal support No. of payment types
1978 174,119 1,025,266 17.0 23.2 6.2 10
1979 187,651 1,035,032 18.1 24.4 6.3 10
1980 206,079 1,044,656 19.7 25.3 5.6 10
1981 210,652 1,057,533 19.9 27.1 7.2 10
1982 226,856 1,070,855 21.2 28.6 7.4 10
1983 261,617 1,084,294 24.1 33.7 9.6 10
1984 278,165 1,097,913 25.3 32.2 6.9 11
1985 281,847 1,104,774 25.5 35.0 9.5 11
1986 278,381 1,113,432 25.0 32.3 7.3 11
1987 254,796 1,120,408 22.7 34.1 11.4 11
1988 266,344 1,130,325 23.6 33.7 10.1 11
1989 252,469 1,141,815 22.1 32.4 10.3 11
1990 243,223 1,155,006 21.1 32.1 11.0 11
1991 256,557 1,167,843 22.0 34.1 12.1 11
1992 281,250 1,183,002 23.8 35.4 11.6 12
1993 308,928 1,197,486 25.8 37.0 11.2 12
1994 333,944 1,223,928 27.3 35.6 8.3 12
1995 346,834 1,256,483 27.6 34.2 6.6 14
1996 359,634 1,291,206 27.9 34.4 6.5 15
1997 371,431 1,350,409 27.5 34.4 6.9 15
Change over 20 years 197,312 325,143 10.5 p.p 11.2 p.p 0.7 p.p 5

Source: FaCS 1999 Attachment E, Table 1.

With an aggregate 6.6 per cent rise in social security dependency for those 50–64 over the period (to 33.5 per cent), and virtually no rise since the mid 1980s the figures are not alarming. However, even if this percentage were to hold steady there would be implications for social security expenditure because of population ageing, and it would be a source of concern were it to rise further.

Data from the ABS Retirement and Retirement Intentions surveys (1994, 1990 and 1986) indicate that 52 per cent of all males and 36 per cent of all females who retired at or over the age of 45 rely mainly on a pension or benefit. The other main sources of income for males are superannuation (16 per cent) and investments (12 per cent). The most important source of income for females (38 per cent) is ‘someone else’s income’. This decreases in importance for older females as they become eligible for pension in their own right.

Table 7: DSS, DVA and DEETYA payments for females in the age range 50 to 64 years, 1978 to 1997 (20 years)

Year Total DSS, DVA and DEETYA customers Population 50–64 yrs (ABS) % of 50–64 yr olds on payment % of 50–64 not in labour force or unemployed (ABS) Presumed % of self or spousal support No. of payment types
1978 380,279 1,041,052 36.5 71.4 34.9 14
1979 382,694 1,048,952 36.5 72.2 35.7 14
1980 384,197 1,057,435 36.3 71.6 35.3 14
1981 384,677 1,070,812 35.9 71.9 36.0 14
1982 387,972 1,077,666 36.0 72.6 36.6 14
1983 409,873 1,084,997 37.8 73.3 35.5 14
1984 416,400 1,092,967 38.1 71.8 33.7 14
1985 415,281 1,095,838 37.9 73.5 35.6 14
1986 415,837 1,098,388 37.9 71.7 33.8 15
1987 417,134 1,103,589 37.8 70.4 32.6 15
1988 422,185 1,110,976 38.0 69.1 31.1 15
1989 423,943 1,120,770 37.8 66.9 29.1 15
1990 423,905 1,130,670 37.5 65.4 27.9 15
1991 446,273 1,141,909 39.1 64.6 25.5 15
1992 455,824 1,155,895 39.4 65.4 26.0 16
1993 485,107 1,169,331 41.5 64.1 22.6 16
1994 496,651 1,195,645 41.5 62.2 20.7 16
1995 525,038 1,228,287 42.7 59.9 17.2 19
1996 514,543 1,261,608 40.8 59.7 18.9 21
1997 521,623 1,319,141 39.5 57.7 18.2 21
Change over 20 yrs 141,344 278,089 3.0 p.p −13.7 p.p −16.7 p.p 7

Source: FaCS 1999 Attachment E, Table 2.

In comparing survey results from 1986 to 1994, the relative importance of service pension is declining whereas reliance on superannuation (for men) and investments and part-time work (for women) appears to have increased. Work-tested social security benefits such as Newstart Allowance play a relatively small role in funding early retirement for males once they are over 55, while DSP becomes much more important at 28 per cent.

Unikowski notes that, moving from surveys of those already retired to those not yet retired, there is a greater expectation about the role of superannuation in funding retirement (1996 p. 22, Table 5). Increasing expectations of superannuation are accompanied by decreasing expectations of pension receipt. This may indicate a potential shift toward increased self- reliance in the future, though this is certainly not reflected in current data.

4.6 Does the social security system induce early retirement?

There is an argument that the incentive to retire early is embedded in the ratio of income from employment to income from superannuation plus social security. Simple examination of the retirement age trend graph and the graph of real social security payment rates in Australia appears to show a correlation between the two series—prima facie evidence of an induced retirement effect.

This would be reinforced by calculation of earnings replacement rates when we take account of the relative stagnation of male wages in the lowest quartile of the income distribution. However, such an exercise can be deceptive. While Age Pension replacement rates (that is, income in retirement as a percentage of pre-retirement earnings) are one determinant of retirement behaviour, there are many others. Replacement rates in Australia are very much at the low end of the OECD average except for those on low incomes (Table 8).

A study by Johnson (1998) similarly indicated that retirement incomes in Australia are very much at the low end of the OECD spectrum, though Whiteford and Bond (1999) showed that this conclusion might be modified by a fuller definition of income taking into account housing and other assets, as well as government social expenditures.

In addition to the real level of publicly provided retirement benefits, the OECD argue that implicit tax rates on pre-retirement earnings are an important determinant of the retirement age.13 In many countries continued work at older ages is heavily taxed either because the Age Pension benefit formulae provide diminishing returns or because of the benefits foregone, such as those for unemployment or disability. According to the OECD’s calculations, the implicit tax rate on work from 55 to 64 is, in Australia, 21 per cent (Table 9): at the low end of the international spectrum.

Gruber and Wise (1998) call this implicit tax ‘the tax force to retire’. They report a ‘strong correspondence between the tax force to retire and unused labour capacity’ (pp. 161–2). (Labour capacity is measured in relation to the difference between labour force participation rates among older workers in the 1960s, and now.) A regression of unused capacity against tax force explained 82 per cent of the country variation in unused labour capacity, and Gruber and Wise observed that changes in early retirement provisions precede changes in labour force participation.

Table 8: Old-age public and mandatory pension system: change since 1960s

  Standard age of entitlement(a) Gross replacement ratios Summary indicator(b) Pension accrual rate(c)
1961 1995 1961 1995 1967 1995
Australia 65 (60) 65 (60) 19.1 40.9 0 0
Austria 65 (60) 65 (60) 79.5 79.5 13 12
Belgium 65 (60) 65 72.6 67.5 32 15
Canada 70 65 31.3 51.6 23 0
Czech Republic 60 (57)   53.2   1
Denmark 67 67 35.9 56.2 2 1
Finland 65 65 34.9 60.0 10 4
France 65 60 50.0 64.8 25 12
Germany 65 65 60.2 55.0 13 11
Greece 65 (60) 62 (57)   120.0   25
Hungary   60 (56)   54.6   1
Iceland 67 67   93.0   10
Ireland 70 66 38.6 39.7 0 0
Italy 60 (55) 62 (57) 60.0 80.0 24 10
Japan 60 (55) 60 (58) 24.6 52.1 5 3
Luxembourg 65 65   93.2   19
Netherlands 65 65 29.0 41.2 9 9
New Zealand 65 62 32.0 61.3 0 0
Norway 70 67 25.3 60.0 17 9
Poland   65 (60)   53.7   9
Portugal 65 65 (62.5) 85.0 82.6 15 10
Spain 65 65   100.0 0 0
Sweden 67 65 53.8 74.4 21 0
Switzerland 65 (63) 65 (62) 28.4 49.3 12 11
United Kingdom 65 (60) 65 (60) 33.4 49.8 0 10
United States 65 65 39.1 56.0 0 0

(a) Standard age of retirement for women in parenthesis.
(b) The summary indicator of gross replacement rates is an average of four cases: Two earnings levels (that is, average and two-thirds of average) and two household compositions (that is, a single worker and a worker with a dependent spouse). For all cases it is assumed that the employee starts work at the age of 20 and that he has uninterrupted work until the standard age of entitlement to public pensions. The earnings profile over the working life is assumed to be flat and earnings re-valued in line with changes in average earnings.
(c) Increase in old-age pensions for a 55 year-old male by working for 10 more years, in percentage points of the gross replacement rate.

Source: OECD 1998b, p. 184.

Simple correlation appears to corroborate the OECD’s previous findings that ‘incentives to retire early have a potentially strong effect on activity rates of older people’ (OECD 1998b, p. 187 and Figure 18). Econometric results suggest that a 10 per cent increase in the implicit tax rate for those aged 55–64 leads to a drop in male participation of 3.5 per cent. On this basis, the implicit tax rate in Australia would have reduced male participation in this age range by more than 7 per cent. If other non-employment benefits such as DSP are included, the implicit tax rate and hence implied employment impact might be even larger.

Table 9: Implicit average tax rate on work from 55 to 64, 1967–95(a), per cent

  1967 Old-age pension 1995 Old-age pension 1995 Overall(b)
Australia 0 0 21
Austria 31 34 34
Belgium −2 23 37
Canada −15 6 6
Denmark 0 0 51
Finland 0 22 42
France 2 14 49
Germany 4 10 32
Iceland .. 1 1
Ireland 5 14 32
Italy 30 79 79
Japan 10 28 28
Luxembourg .. 29 65
Netherlands 5 8 57
New Zealand 0 9 27
Norway 3 15 15
Portugal 5 4 33
Spain 6 18 45
Sweden −9 18 18
Switzerland −2 0 0
United Kingdom 6 5 15
United States 8 12 12

(a) The implicit tax (or subsidy) on continued work (see text) is the average annual variation in the social security wealth relative to gross earnings as a result of postponing retirement from 55 to 64 years of age. The social security wealth is the sum of the discounted value of expected benefits (either pensions or other non- employment benefits) minus the discounted cost of obtaining these benefits. See BIondal and Scarpetta (1998). Figures are relative to annual earnings and refer to a single individual at average wage.
(b) Old-age pension and unemployment-related benefits.

Source: OECD 1998b p. 186.

However, the implicit tax on older workers cannot simply be removed, since in the Australian context it would probably need to involve a drastic tightening of eligibility conditions for benefits to such workers, many of whom might not easily find jobs. Tightening eligibility for DSP, for example, is likely to lead to increased claims for unemployment and sickness benefits. The OECD (1998b, p. 192) notes that ‘the removal of disincentives to work may need to be accompanied by labour market reforms that promote job opportunities for older workers’.

The Government’s tax package may reduce financial incentives to early retirement, both by lowering marginal tax rates on wages, and also by reducing the pension taper rate to 40 per cent.

Atkinson and Creedy (1996) simulated individual’s work-to-retirement decisions using an optimal choice model to calculate an individual’s best route through the ‘retirement maze’. While they had to make many simplifying assumptions, the authors found that:

  • current policies provide a significant incentive to retire early;
  • the pension means test—at least hypothetically—has a substantial impact on the choice of the retirement age and the allocation of assets at retirement.

An alternative view is that public policy may have little impact on retirement decisions. Several United States studies have found that, despite the marked trend toward earlier male retirement in the 1970s and 1980s, changes to pension plans and social security entitlements probably explained only a third or a quarter of these trends, with the remainder accounted for by other factors, including rising real wages, changes in disability insurance, changes in tastes for retirement and changes in the industrial distribution of jobs (Anderson et al. 1997, p. 4; Quinn et al., 1990).

Several studies predict that changes to the United States Old Age, Survivor’s and Disability Insurance (OASDI) system, designed to make the system actuarially fair to older retirees, will have relatively little impact on the average retirement age. The normal retirement age in OASDI will rise from 65 to 67 in gradual steps beginning in 2000, and the delayed retirement credit from 3 per cent a year to 8 per cent. This is predicted to cause the number of individuals working full-time between the ages of 65 and 68 to be 4–6 per cent higher than under the pre-1983 rules (Gustman and Steinmeier 1985).

The recent New Zealand experience provides a rather different picture. Beginning in 1992, the age of eligibility for the New Zealand superannuation scheme is being raised from 60 in three- month increments every six months, and will reach 65 in 2001. According to one researcher, the estimated behavioural effect is for a staggering 30.5 per cent increase in participation among those 60–64, though with the policy affecting only 55 per cent of this group it translates into a predicted 16.8 per cent overall rise (Maloney 1997, p. 29 and see also McCormack 1996, p. 128).

Most surprisingly, this policy has not been accompanied by greater unemployment among the 60–64 group. Much of the increase in participation appears to have been accommodated by people simply staying longer in their existing jobs.

Such a result is not simply transferable to Australia: there is considerable difference between the New Zealand superannuation pension and other social security benefits for those below pensionable age, and a considerable difference in means tests (New Zealand super is universal).14

In the past it has sometimes been argued that early retirement can be beneficial in freeing jobs for younger people. Indeed, there was a tendency in the 1970s and early 1980s in a number of OECD countries to promote early retirement for this reason. Since then, however, there has been a marked shift in thinking on these issues—see Section 6.7.2—with governments becoming more aware that early retirement exacerbates age dependency burdens and that the demand for labour is not fixed but depends on factors such as price and skills. In other words, getting rid of skilled, experienced older workers by no means guarantees that their places will be filled by less experienced younger workers.

Another issue is whether jobs are actually available for people over, say, 55. Though the job prospects for people unemployed at this age are poor, it is not true that they have no choices. Typically they do have some choice about whether to stay on or not if they are in an existing job. The New Zealand experience described above definitely indicates that incentives to retire later can increase employment among older people, through just such a mechanism.

[ Return to Top Return to Section ]

5 Social security receipt among older retired people Numbers receiving FaCS and DVA payments

5.1 Numbers receiving FaCS and Department of Veterans’ Affairs (DVA) payments

Many early retirees end up on the income support system. Among those aged 50–64 (about 2.67 million people), 912,000 or 34.2 per cent were on support in 1997—28.3 per cent of males and 40.2 per cent of females. There are significant differences between men and women in terms of the predominant type of payment.

Older retired people (those aged 50–64) have access to a bewildering array of possible social security and DVA payments (Table 10 and Figure 2). The more important of these include Disability Support Pension (284,000), Age Pension for women (208,000), Newstart Allowance (110,000), Wife Pension (86,000), Partner Allowance (54,000), Mature Age Pension/Allowance (53,000), DVA disability (12,000), Widow Allowance (17,000), and Widows B Pension (17,000). There is a total of 22 possible payments (including Austudy); only a minority (12 per cent) of these clients are on Newstart Allowance, the only payment that is work tested.

Six of the payments listed in this table are in the process of being phased out. This includes Age Pension for women 60–64, Wife Pension, Widow Allowance, Partner Allowance, Mature Age Partner Allowance, and Widow B Pension (Mature Age Allowance [MAA] remains but is now subject to allowance, not pension, conditions, and we can disregard the four DVA payments, which are in the nature of war service compensation). However, this still leaves 10 payment types remaining, five of which—Newstart Allowance, Carer Pension, DSP, Sickness Allowance and MAA—are significant in terms of numbers; and five of which—Parenting Payment (single or partnered), Sickness Allowance, Special Benefit and Drought Relief Payment—will ‘attract’ relatively few people.

Of these, MAA is the only payment designed specifically for older people. It follows that some of the complexity of payments for older people is a transitional problem that will gradually disappear, and much of the remaining problem is a reflection of the general complexity of the payment structure.

MAA is available to those 60 and over who have been on income support for at least nine months. MAA recipients are paid under allowance conditions, but are not required to submit to activity testing. Currently MAA recipients are mainly men, since women 61 and over would normally be eligible for Age Pension if residentially qualified. This will change as the earlier pension for women is phased out.

Some 383,000 males aged 50–64, comprising 28.4 per cent of the relevant population, are on income support. Of these, 217,000 are on Disability Support Pension (including 12,000 DVA pensioners); 81,000 on Newstart Allowance (long and short-term) and 53,000 on MAA. The remainder are on various other payments.

Among females, receipt of disability pension is far less prevalent, with 79,000 on this payment compared with 208,000 on Age Pension and 147,000 on Wife Pension or Partner Allowance. In all, 533,000 females aged 50–64 (40.4 per cent of the total) receive some form of income support.

Receipt of social security benefits is strongly age-related. Table 7 shows that incidence of social security benefits rises markedly in each age group, beginning at 22.1 per cent for those 50–54 and reaching 54.8 per cent among those 60–64. For those over 64 (not shown), 82.6 per cent receive a full or part-rate payment under FaCS or DVA schemes.

Table 10: DSS, DVA and DEETYA payments for those aged 50–64, 1997





Payment type Number Per cent Number Per cent Number Per cent
Age Pension 208,299 22.7     208,299 39.1
AUSTUDY/ABSTUDY (DEETYA) 4,502 0.5 1,807 0.5 2,695 0.5
Carers Pension 14,965 1.6 7,758 2.0 7,207 1.4
Disability Support/ Invalid Pension 283,860 31.0 204,608 53.3 79,252 14.9
Disability (DVA) 12,117 1.3 12,117 3.2   0.0
Drought Relief 3,156 0.3 2,607 0.7 549 0.1
Mature Age 53,400 5.8 52,759 13.8 641 0.1
Mature Age Partner 7,293 0.8   0.0 7,293 1.4
Newstart: Long-term 62,562 6.8 46,519 12.1 16,043 3.0
Newstart: Short-term 47,036 5.1 34,438 9.0 12,598 2.4
Parenting Allowance 4,256 0.5 3,214 0.8 1,042 0.2
Partner Allowance 54,447 5.9 4,141 1.1 50,306 9.4
Service Pensions (DVA) 9,106 1.0 9,066 2.4 40 0.0
Sickness Allowance/Benefits 2,948 0.3 2,133 0.6 815 0.2
Sole Parent/Supporting Parent/ Widow A Pension 9,827 1.1 2,063 0.5 7,764 1.5
Special Benefits 1,506 0.2 318 0.1 1,188 0.2
War Widows (DVA) 4,880 0.5   0.0 4,880 0.9
Widows B Pension 17,360 1.9   0.0 17,360 3.3
Widow Allowance 17,271 1.9   0.0 17,271 3.2
Wife Pension 86,183 9.4   0.0 86,183 16.2
Wives/Widows (DVA) 11,935 1.3   0.0 11,935 2.2
Total DSS, DVA and DEETYA customers 916,909 100.0 383,548 100.0 533,361 100.0
Population 50–64 yrs (ABS) 2,669,550   1,350,409   1,319,141  
% of 50–64 yr olds on payment 34.3%   28.4%   40.4%  

Source: DSS, DVA and DEETYA data for June 1997.

It is notable that the over-55 group is substantially relieved of the obligation to look for paid work. This group’s eligibility for DSP is more generous than that for other groups as labour market conditions can be taken into account. The activity test for Newstart Allowance is more generous than for other groups because of reduced reporting and greater scope for voluntary work. Mature Age, Widow and Partner allowances are designed to remove various groups from the obligation to look for work on the basis of age and long-term unemployment. Thus while activity testing has become more stringent for younger people, the opposite response has been taken for this group.

Figure 2: DSS and DVA payments for people 50–64, 1997, percentage

Graph showing DSS and DVA payments for people 50–64, 1997, percentage (calculated from data in Table 10)

Source: Calculated from Table 10.

5.2 Trends in the incidence of Disability Support Pension

In the United States, several studies have argued persuasively that earlier accommodation of health problems through the growth of disability benefits has influenced retirement patterns.

Bound and Waidman (1992) for example, estimate that as much as a third of the decline in labour force participation among American men aged 60–64 may be due to disability payments, though they also accept that most of those whose departure from the labour force was facilitated in this way did suffer from disabling conditions. What public policy changes did was to change the margin at which those in less than good health could afford to leave the labour force.

Bound and Burkhauser (1998) summarise their review of the disability literature by noting that changes in access to and the generosity of disability programs ‘have had significant effects on both the economic well-being and the work force attachment of those individuals whose health limits their capacity for work’, but that ‘there remains a tremendous amount of uncertainty regarding the behavioural effects of disability insurance programs’ (p. 59).

The anticipated increase in numbers on disability pension15 in Australia, based on past trends, is of concern. At current growth rates, annual new grants of DSP are expected to rise from 73,000 in 1995 to between 80,000 and 140,000 a year within 10 years (Walsh 1997). The lower figure assumes that the age-specific grant rate is pegged at 1995 levels, whereas the higher figure assumes that it continues to rise by 5 per cent a year. Under the Walsh projections, the number of DSP recipients could rise from just over half a million at present to more than 900,000 by 2006, with much of the increase in the 50–64 age group. The Minister for Family and Community Services (Newman 1999) recently stated that the figure,‘unless we do something’, would likely be over 750,000.

As noted above, such trends are not unique to Australia. Lonsdale and Seddon (1994) found that the population in receipt of disability benefits had increased in virtually all OECD countries. Gruber and Wise (1997) report that the incidence of pre-retirement disability or unemployment- related benefits is about a fifth in Belgium and France, a quarter in the Netherlands and Sweden, a third in the United Kingdom and 37 per cent in Germany. In the United States it is only 12 per cent.

Lonsdale and Seddon (1994, p. 149) state that ‘there is no clear evidence that these increases are accounted for by changes in the incidence of disability. But they are often associated with economic and demographic changes such as increases in unemployment and an ageing population’.

This is consistent with Australian research by Morrow (1998), who finds DSP numbers and unemployment rates are strongly correlated. However, demographic transition has not yet had an effect on aggregate disability numbers in Australia. Rather, Jackson (1999) finds that almost all the rise in numbers to now has been caused by rises in age-specific incidence rates.

The incidence of new DSP grants rises steeply with age, especially among men. More than 10 per cent of the male working age population aged 50–64 was in receipt of Invalid Pension (IP) in 1987, and about 4 per cent of females. By 1995 these rates had risen to 14 per cent and 6 per cent respectively.16 These compare with population average prevalence rates of 4.1 per cent among males and 1.5 per cent for females. Currently, about 38 per cent of female claimants and 57 per cent of male claimants are aged 50 and over, and this group accounts for more than half of all new DSP grants (Walsh 1997, p. 17).

As a proportion of the total DSP population, the number aged over 55 is currently 39 per cent and the proportion aged over 50 is 55 per cent. Among males, incidence of IP/DSP among the 50–64 age group rose from 8 per cent in 1978 to 15 per cent in 1993, though it appears to have stabilised at that level in recent years and was still 15 per cent in 1998 (see Figure 3). Incidence is highest at 25 per cent among those 60–64 (McCormack, 1996, Chart 5). Total male reliance on income support rose from 1978 to 1998 from 21 per cent to 28 per cent.

DSP incidence among those 60–64 roughly doubled from 1983 to 1993, which may reflect the declining number of men who are qualified to receive service pension, the availability of which would have tended to depress DSP numbers in the past.

For women, DSP incidence has risen from 3 per cent to 7 per cent since 1978. By contrast, aggregate dependence on income support has been relatively steady, at around 41 per cent (Figure 3).17 The increasing numbers of single women (who cannot claim Wife pension) and the phasing out of Widow B and Wife Pension may be factors contributing to increased take-up. In the future the phased increase in the pension age for women can be similarly expected to have an impact.

There have been a number of initiatives since the early 1970s affecting basic eligibility criteria for DSP. These have aimed to increase the importance of the level of impairment (physical, intellectual or psychiatric) in the assessment of eligibility and to reduce the extent to which socioeconomic factors (age, skill level or labour market conditions) are taken into account.18 Despite such initiatives, there continues to be strong growth in the number of people in receipt of DSP.

5.3 Is there a rising incidence of sickness and disability?

The rising trend in disability numbers would appear to indicate an increasing incidence of sickness and disability among older workers, which contrasts with the observed tendency toward greater life expectancy among both men and women. We do not know enough about levels of morbidity, well-being and disability in the population; two national surveys have been conducted by ABS, but ‘are limited in use since they rely on self-assessment and are subject to considerable error’ (Gatiss 1997, p. 31).

Figure 3: Percentage of older persons 50–64 on IP/DSP versus all payments

Graph showing percentage of males aged 50–64 on IP/DSP versus all paymentsGraph showing percentage of females aged 50–64 on IP/DSP versus all payments

One hypothesis has been that people are living longer as a result of advances in medicine and so on, but are not necessarily healthier: that is, some people whose conditions might in earlier times have caused death are being kept alive, but at reduced levels of functionality. This hypothesis has been the subject of some debate in the literature.

An alternative theory is that people are generally healthier and that, together with greater life expectancy, the onset of chronic illness is being delayed. Increased longevity may shift the period of high costs to later ages, without increasing the period over which high costs are incurred (McCallum et al. 1994).

In a review of the international literature, the OECD (1996) concluded that the extension of life tended to prolong both independent and dependent periods. In relation to aged care, it concluded that the need for care was most likely to increase in line with demographic developments; however, more recent evidence has tended to suggest that disability rates among the elderly are falling in most industrialised countries (see Kalisch 1999, p. 17).

This evidence has been summarised in Waidman and Manton (1998) and Jacobzone et al. (1999). The general conclusion from Waidman and Manton was that, of the countries surveyed, only three (Australia, the UK and Canada) did not show consistent evidence of declining disability—though neither did these three show any worsening. Jacobzone et al’s work provides broadly consistent findings, but suggested that improvements in the incidence of severe disability were more significant for men than for women, and that the greatest improvement was in the younger age groups (65–74 and 75–79) within the aged population:‘The broad finding from the overseas evidence is that increased longevity does not lead to increasing incidence of disability or chronic poor health … More research is required, especially to introduce greater comparability in the methodology and analysis used to produce these international results’ (Kalisch 1999). Another question is whether these results for the pension age population can safely be projected backwards to those in the older working age group.

The upward trend in DSP incidence among older men is in marked contrast to aggregate measures of health status such as life expectancy and disability rates, and some self-reported health status. For example, the ABS Disability Surveys (ABS 1993) show a slight decline between the 1988 and 1993 surveys in the percentage of males aged 60 to 64 reporting a disability or handicap.

Death rates (all causes) for men aged 55–59 and 60–64 declined dramatically from 1970 onwards—that is, at the same time as their labour force participation began to seriously deteriorate (McCormack 1996, p. 132). It has been suggested that the two trends are linked—that early retirement helps men live longer. On the other hand, higher levels of unemployment tend to have a negative impact on the health of those affected.

The proportion of early retirees citing ill health or injury as the main cause of their early retirement has remained remarkably stable over the three retirement surveys run since 1989, dropping from 41 per cent to 39 per cent. Compared with the upward trend in DSP usage rates, this appears to suggest at least some usage of DSP as a source of income support for labour force exit.

In the United States, a substantial number of studies point to disabling health as a basic cause of retirement for at least 25 per cent of the older population (Stewman 1986, cited in McCormack, p. 33). The theory here is that ‘as mortality rates drop, some people are “rescued” from death and remain among the living. Most of those “new survivors” are ill with chronic diseases (but not so ill that they will die). By staying alive for a while, they have more years for their illnesses to advance in severity and more time to develop other chronic conditions’ (Verbrugge 1984, p. 507; in Gatiss, p. 33).

Verbrugge’s theory does not appear to be supported by Australian data—and even if we accept it, it is likely that it falls a long way short of explaining the rapid rise in the incidence of DSP.

One of the largest components of rising social expenditure associated with population ageing is health costs (section 2.2). However, most analysts agree that the bulk of these costs are associated with secular upward trends due to technological change and similar factors; ageing per se typically accounts for only a quarter or even a fifth of the total projected rise in health costs.19 This is consistent with the notion that increased longevity delays the onset of chronic illness, but does not affect its incidence.

[ Return to Top Return to Section ]

6 Possible policy changes to reduce early retirement Recent policy initiatives

Recent OECD analysis (OECD 1998a) suggests that reversing early retirement is one of the most effective means of countries adjusting for ageing populations. This is consistent with the analysis of McDonald and Kippen in the Australian context, cited earlier.

6.1 Recent policy initiatives

The Australian Government, like others in the OECD, has taken a number of recent initiatives that will reduce incentives to early retirement, including:

  • the standardisation of the retirement age for men and women by phasing in a higher age for women over 20 years (this process will finish in 2013);
  • full preservation of all SG contributions from 1 July 1993;
  • compulsory preservation of all superannuation benefits from 1 July 1999;20
  • the increase in the compulsory preservation age for occupational superannuation from 55 to 6021 (although this is fairly slow, and will still leave opportunities to ‘double dip’ prior to retirement age);
  • the phasing out of certain provisions that mainly benefit older women, including Wife’s Pension and Widow B Pension;22
  • allowing a larger reasonable benefits limit (RBL) for retirement benefits taken at least half in annuity form;
  • the introduction of the Deferred Pension Bonus Plan;23
  • the provision in the Workplace Relations Act 1996 that an employer must not terminate a person’s employment because of age;
  • the removal of a compulsory retirement age in most States and Territories (except Tasmania) over 1990–96, and the foreshadowed Commonwealth Public Service Legislation abolishing compulsory retirement at age 65.

There is no specific age discrimination act in Australia. However, age discrimination in employment is specified as one of the grounds of discrimination covered in the Human Rights And Equal Opportunity Commission (HREOC) Act 1996. However, many complaints are unable to be conciliated because the Act does not make discrimination on grounds of age unlawful (HREOC 1998). Age discrimination is also covered in various degrees in various States and Territory Acts.

A change that may also have had some effect is the inclusion of superannuation assets in the means test for those aged between 55 and the pension age, if they had received income support for at least 39 weeks after age 55 (the ‘over 55 change’). This change may act as a disincentive

to voluntary early retirement by reducing entitlement prior to Age Pension age; on the other hand, it may increase entitlements on reaching pensionable age. The impact of this change has yet to be evaluated.

It is not clear to what extent these changes will influence the general retirement age, or how quickly this will happen.

The Government is currently formulating a policy response to the International Year of Older Persons, to be called A National Strategy for an Ageing Australia.24

6.2 Reducing inflows to Disability Support Pension

An option that many commentators advance for reducing the inflow into DSP as an ‘early retirement’ option would be to tighten the eligibility criteria by, for example, adopting a higher minimum impairment level and/or a less generous ‘inability to work’ criterion.

However, previous attempts to control growth in DSP numbers by changing eligibility criteria have proved contentious and/or unsuccessful. Anecdotal evidence suggests there is a widespread perception (for example, among Centrelink staff, medical practitioners and individual claimants) that DSP is an appropriate income support payment for older long-term unemployed people who have some level of disability. Any new attempt to tighten eligibility criteria may not be effective in practice if this attitude prevails.

Tightening of DSP criteria would displace a number of people with a range of workforce barriers. Many of these people would flow to Newstart Allowance (NSA), though some would be eligible for other payments. The NSA activity test could be made more flexible to accommodate the wide variations in individual capacities, circumstances and needs likely to be found among this group; however, it is probable that many people would continue to see NSA as unsuitable for older people with poor labour market prospects.

In the future, automatic flow-on of MTAWE indexation in age pensions to workforce Age Pension-type payments could exacerbate the situation by further increasing the differentials between pension and allowance entitlements.

Sorensen (1998) notes the experience of Norway in attempting to tighten disability criteria, showing that ‘a large proportion of those whose request for a pension is turned down due to a stricter application of the disability concept re-apply within a short period of time and are subsequently awarded a pension. Thus in many cases the pension award is postponed rather than avoided. In the mean time most of the applicants received one or another alternative social security benefit … Such observations underline the point that the success of such strategies rely heavily on the success and the effectiveness of rehabilitation and activation measures’ (Sorensen, p. 6–7).

6.3 Is dissipation of lump sums a problem?

Some early (and other) retirees are able to ‘double dip’ the retirement income system by taking their substantially tax-financed lump sum and dissipating it prior to pension age, and/or structure their affairs to prevent it affecting their Age Pension entitlement. About 60 per cent of retiring superannuants receive a lump sum only; 25 per cent both a lump sum and a pension, and fewer than 15 per cent receive a pension only (Kalisch, 1992, p. 75). Other survey evidence indicates that about 75 per cent of superannuation payouts are lump sums (ISC Quarterly Bulletin, 1998).25

The evidence on the extent of deliberate ‘double dipping’ does not suggest this occurs to any marked extent. Kalisch (1992, p. 83) suggested from survey evidence that in fact there was relatively little dissipation of lump sums on items such as overseas trips and other forms of consumption:‘just over 85 per cent of the lump sum value was directed towards financial or equity investments, property or businesses’. While some limited expenditure on holidays was relatively common it normally accounted for only a small percentage of the lump sum received.

Kalisch (1992, p. 86) concludes that ‘the evidence … suggests that there is not widespread double-dipping and superannuation entitlements are not generally dissipated with the intention of establishing Age Pension eligibility or increasing the size of the pension payment’. The problem with this conclusion is that Kalisch’s definition of double dipping is fairly restrictive. He defines ‘double dipping’ as ‘the practice of dissipating rapidly or before pension age a superannuation lump sum which has received significant tax concessions … with the express intention to receive a greater Age Pension than might otherwise have occurred’ (p. 77).

At the time of Kalisch’s article, the average lump sum payment of $75,000 was just under half the $150,000 asset test threshold for home owners, and was also less than the $100,000 estimated as enough to provide a life annuity equal to the means test free area. Further, the median lump sum was less than $40,000. Less than a quarter of payments were over $100,000. Other surveys indicate even higher concentrations of lump sums at the lower end of the range (Kalisch, p. 82). On the face of it and adopting Kalisch’s test of intent, many people won’t have to bother with ‘double dipping’—the system already allows it explicitly.

A survey commissioned by the Department of Social Security from Reark Research suggested that ‘diversion of lump sums to income generating investments is not as common as indicated by other information sources’ (Kalisch 1992, p. 86). Those who received a smaller lump sum were more likely to use it to clear debts, for home improvements and for general living expenses. This was particularly true of age and service pensioners.

One interesting finding in Kalisch’s survey is that unpublished ABS data show ‘relatively little difference in the main disbursement of lump sums comparing those who receive a lump sum only with those who receive both a lump sum and a regular payment from superannuation’ (p. 84). This implies that if there is to be policy change directed to increasing the use of annuities, it will need to be fairly comprehensive.

Kalisch’s survey does not establish that greater lump sums will reduce pension expenditure. On the other hand, ABS data on intended disbursements of superannuation lump sums appear to indicate that as lump sums increase in average size, a greater proportion will be spent on financial investments (p. 87). Projections of lump sum payments based on 40 years of full-time superannuation coverage indicate lump sums could be around $160,000 (at 1992 values) for those with income at the level of three-quarters Average Weekly Earnings (AWE) (p. 88), a figure close to the median earnings level. But trends toward shorter working lives and the broken work patterns typically associated with raising children suggest such projections could be atypically high.

These projections also indicate that only about a quarter of retirees will have payouts sufficient to remove them from receipt of Age Pension entirely. In practice this means that many more people will be subjected to the very high effective marginal tax rates (EMTRs) implicit in the Age Pension means test.26

RIM’s methodology in projecting investment income of retirees assumes an aggregate 25 per cent leakage from superannuation lump sums. There are two possible views about this degree of leakage: first, that it shows that the bulk of superannuation benefits are used productively; and second, is that a large and rising sum of superannuation money is not being applied in ways that reduce government pension expenditure.

Should we be concerned at all about ‘double dipping’? One answer might be that savings on pension expenditure are not the primary objective of mandatory superannuation, and that the system is achieving its intended effect of improving the adequacy of retirement incomes. The other is that there are limits to the extent government can or should support retirement incomes, such that the system ought to recoup, via the means test and taxation, tax expenditures on occupational superannuation.

6.4 Reducing dissipation of lump sums

One option suggested by many commentators is to raise the age of compulsory preservation further to align with Age Pension age—that is, 65 rather than 60.

However, this would appear to disadvantage many people who might legitimately expect to retire early and who indeed might have substantial superannuation assets for just such an eventuality.

The ‘let-out’ is that current provision for raising the preservation age to 60 includes provision for early release of benefits where a non-commutable life pension is taken. This might encourage annuity take-up, but it creates inequity between those who can and cannot hang on until the compulsory preservation age.

Another option that has been suggested to address this inequity would be to require that superannuation benefits be taken almost wholly in the form of a life annuity.

This could involve penal taxation of lump sums over a certain limit, and/or tightening the rules under which the higher rate of RBL applies by requiring the income stream be a life annuity; requiring that all of the benefit in excess of a certain amount be in annuity form, (rather than half of the total, as at present); and reducing the lower RBL for lump sums.

The effect would be that while early retirees could reduce their post-age 64 incomes, they could only do so to a limited extent. However, the annuity option may not be attractive to many people.

6.5 Tightening superannuation preservation rules

Effective from 1 July 1999, the preservation arrangements changed so that all future contributions (including member undeducted contributions) are required to be preserved until retirement on or after age 55 (Treasurer and Minister for Social Security 1997). This will simplify fund administration and help ensure that superannuation savings are in fact used to enhance retirement incomes.27 Further, the preservation age is to be raised to 60 beginning in 2015, affecting anyone born after 30 June 1960 (by 2025 this group will have a preservation age of 60).

This change will indicate to employers that the government does not believe that workers should normally retire as early as 55. Individuals affected by the higher preservation age will continue to be able to gain early access to preserved benefits where the benefits are taken as a non-commutable life pension or lifetime annuity on termination of employment—which, in conjunction with other incentives described in the previous section, should encourage the use of income streams.

Superannuation legislation allows superannuation benefits (including preserved benefits) to be released from a fund in cases of ‘severe financial hardship’. From December 1997, this test became a more objective one that applies to:

  • people who have been in receipt of a specified Commonwealth income support payment for a continuous period of at least six months and who satisfy their superannuation fund trustee that they are unable to meet living expenses; or
  • people aged over 55 who have been in receipt of such a payment for a cumulative period of 39 weeks or more after reaching age 55.28

While the objective of the hardship test is laudable, the effect of such provisions may be to undermine the system’s ability to discourage early retirement for our customers. Currently, $10,000 a year may be drawn down under the hardship rules.

There appears to be a prima facie case for aligning the age of compulsory preservation with the Age Pension age: the prospective raising of the preservation age to 60 will ameliorate but not resolve this problem.

The disadvantage of a full alignment is that it would adversely affect those whose early retirement is involuntary, and who might have to live on severely reduced incomes before they were able to gain access to their superannuation. However, the people’s ability to gain access to superannuation before the preservation age—so long as it is in annuity form—along with the specific hardship provisions might be regarded as adequate safeguard against hardship.

6.6 Increasing the age of pension eligibility

This has been an issue in several overseas pension schemes, and could be used alone or in conjunction with the restrictions on lump sum access canvassed above. Its justification would be that it helps to redress the blowout in age dependency ratios (on both sides of the equation) and that a pension age initially set at the turn of the century, when life expectancies were much lower, is no longer appropriate.

OECD estimates suggest that should the average retirement age rise to 70, most of the adverse consequences of ageing on living standards could be avoided (Blondal and Scarpetta p. 13). While this would be a large rise, even a lesser increase could still provide substantial benefits. The OECD has a rule-of-thumb that a one-year decrease in the average retirement age causes GDP to decline by about 2 per cent and pension contributions by those still of working age to rise by 7 per cent. (This rule may not, however, fit the Australian situation.)

One difficulty with the option of raising the pension age is that its success in this country would depend on the extent to which we were able to reduce recourse to other payments as early retirement options.

In the United States, legislation in place will see a slow phasing in of age 67 (by 2027) as the pensionable age under OASDI. In Sweden, pension rates have been negatively indexed to changes in longevity so that the system is actuarially neutral with respect to rising life expectancies (however, the pension age remains at 65). Other OECD countries are considering similar steps as a response to the serious age dependency burdens—some might say crisis— they are anticipating.

However, 65 continues to be the most common pensionable age in the OECD area. In Germany, for example, the retirement age for men is rising from 63 to 65 by 2009, and for women from 60 to 65 in 2018. Many countries allow early retirement on a reduced pension from age 60 and deferred retirement, typically up to 70, at higher rates. There has been a tendency to make pension adjustments for early and late retirement actuarially more neutral in order to discourage early retirement.

In the United States, the increase in the ‘normal’ retirement age will not prevent continuing early retirement from 62, though early retirees will get less than under current law (reductions for early retirement are, in the United States scheme, now almost actuarially neutral). It is predicted that many workers will continue to claim benefits from 62 (Congressional Budget Office 1999, p. 1). Currently two thirds of workers taking Social Security retirement benefits decide that the reduction in their monthly cheque is a price worth paying to begin collecting benefits before 65; indeed, most of those awards are made to workers at 62. This and other evidence suggest that the age of earliest eligibility for benefits is an important factor in many people’s retirement decisions (CBO p. 12).

The United States National Commission on Retirement Policy recently suggested gradually raising the earliest age of pension eligibility to 65. The CBO’s analysis of this is of some interest in considering possible increases to the pension age in Australia. It notes that ‘an important issue in the debate over such proposals is whether raising the earliest eligibility age would result in serious hardship’ (p. 2). It finds that ‘most of the workers who took early retirement benefits in the early 1990s had [incomes] sufficient to at least keep them out of poverty had they not been able to start collecting benefits until age 65. Moreover, most of those workers did not appear to have health limitations or other problems that would have prevented them working to a later age if necessary’ (p. 3).

Of those whose other incomes were sub-poverty-line (roughly a quarter to a third of early beneficiaries), most appeared not to have disabilities or health problems preventing them or their spouses from delaying retirement, working more hours or saving more.

The CBO notes on the other hand that a significant minority (roughly 10 per cent) of those who took early benefits had both a very low outside income and might also have serious problems working until a later age:‘other combinations of income-related and work-related problems produce similar estimates’ (p. 22), and that if the minimum retirement age were raised, the cost of expanding related programs (Supplemental Security Income and Disability Insurance) could be substantial.

6.7 Encouraging gradual retirement

In the past, the traditional model of retirement has been an abrupt and total withdrawal from the labour force. Greater longevity is calling this model into question, with many experts advocating a more gradual scaling down of activity rather than abrupt cessation, both from the point of view of fiscal sustainability and the financial and emotional health of the people concerned. Indeed, there is evidence of such behaviour emerging in the United States (Quinn et al. 1990, p. 243) and to a lesser extent in Australia (Bacon 1997, p. 28).

In Australia in 1998, 50 per cent of people aged 65 and over who were working did so part-time, compared with 27 per cent of those aged 55–64 and 22 per cent for those 45–54 (ABS 6203.0). However, this tends to be in part a female phenomenon: the proportion of older working-age men engaged in part-time work is still less than 10 per cent. Those most likely to continue working after pensionable age are the self-employed, who presumably have the flexibility to time their own, gradual withdrawal from the labour force.

Some countries are using a mixture of part-time work and income support to make the transition from work to leisure less abrupt, though it is difficult to avoid such schemes becoming an invitation to early partial retirement. If phased withdrawal from the workforce is to be the desired pattern for the future, it may have implications for the appropriateness of current social security payments.

Gradual or partial retirement has many benefits ascribed to it, including:

  • helping overcome the supposed ‘pension shock’ associated with rapid transition from working to not working;
  • allowing the demands of work to be better adjusted to any decline in capabilities or stamina;
  • enabling older people to stay in work longer than might otherwise be possible.

Overseas experience with gradual retirement schemes

Despite these attractions, the incidence of gradual retirement in OECD member countries is extremely limited (OECD 1997 p. 22). The problems identified are fivefold:

  1. swamping by early retirement provisions;
  2. negative consequences under pension schemes;
  3. difficulty for the employer in organising work;
  4. older worker’s own attitudes;
  5. over-regulation.

‘Swamping’ refers to the phenomenon—observed in many OECD member countries—that company level initiatives to allow phased retirement are overtaken by public early retirement schemes that are attractive to employers because the costs are shouldered externally. They are also attractive to the affected workers themselves, who might find their income from early retirement superior to the income available from continued part-time work. Also, few company pension schemes allow simultaneous work with the sponsoring company.

In relation to the difficulty of organising part-time work, the OECD noted that while senior managers praised the idea of partial retirement, their subordinates—who had to organise it— were much less enthusiastic, pointing to the difficulties of coordination and continuity, and the possible lack of commitment of part-timers.

Using older workers as ‘mentors’ to train their successors, seems to have proved less easy than was hoped. One case study concluded that ‘in practice … it was skill rather than age that determined the choice of tutor … In general, the tutors [were] young because they had the technical competencies necessary to perform this role’ (Schmidt 1995, cited in OECD 1997, p. 24).

On the issue of older workers’ attitudes and desires, a wealth of surveys shows that older workers react favourably to the idea of being able to retire gradually. However, actual behaviour is not always in accord with these responses. Relative income replacement rates are one explanation, and there is evidence that older workers do not fear ‘pension shock’ and seek full

exit where it is available. In some cases where partial retirement is embraced, it is seen more as an opportunity to leave the labour force than one to prolong working life. There may also be a rejection of part-time working itself, which men may see as a ‘women’s option’.

Finally, there is evidence that partial retirement schemes in Europe have suffered from over- regulation. Particularly onerous is the replacement condition, under which partial retirement would be allowed only if the job were filled by one of the unemployed. Recruitment was thereby confined to a limited population and vetted by labour market authorities. The OECD (1997, p. 25) concludes that ‘the Swedish scheme, which had no replacement condition and which placed no obligation on the employing organisation to offer part-time work to older workers seeking a partial pension, stands out as the scheme with the highest rate of take-up’.

In the United States, observers have emphasised the importance of ‘bridge jobs’ as a step between career jobs and full retirement. Depending on the definition of career job, a quarter to half of older workers in the United States remain in the labour force after they leave their career jobs. However, by no means all these transitions have been voluntary: in the case of the self- employed in particular (many of whom lack access to contributory pensions), longer working may be a product of necessity rather than choice.

Overall, the OECD concludes that ‘gradual retirement appears to be an ideal which faces many impediments to be realised’ (1997, p. 26).

Overseas experience with early retirement schemes

In the past, it had been argued that encouraging early retirement would be beneficial in freeing jobs for younger unemployed people. Indeed, there was a tendency in a number of OECD countries in the 1970s and early 1980s to promote early retirement for just this reason. However, these expectations were not met in practice.

The OECD points out that ‘a few countries have operated special early-retirement schemes designed specifically to facilitate early withdrawal from the labour market. They were originally intended to help reduce unemployment or to ease the social cost of major structural changes in the economy. Benefit levels tended to be relatively generous. Such schemes have fallen very substantially in importance over the last decade’ (OECD 1998, p. 41). The increased cost of encouraging early retirement has been a significant factor in the demise of such schemes.

In a like manner the International Labour Office (ILO 1995) reports on schemes designed to replace older workers with younger ones:‘the total impact of such measures on the problem of youth unemployment was negligible. New entrants to the labour market often lacked the experience and know-how necessary for the jobs vacated by older workers. More often, however, the jobs just disappeared’.

‘The results of some studies show that the replacement of older workers by younger unemployed workers in such schemes was relatively successful, ranging from 95 per cent in France, where replacement of retirees by unemployed workers (not necessarily young workers) was obligatory, to 67 per cent in Belgium. A West German study carried out in 1984 to evaluate early retirement schemes found that as a result of the ‘solidarity contracts’ in France, the number

of unemployed at the end of 1983 was reduced by between 89,000 and 116,000 (out of a total of 2.23 million). In the United Kingdom, the total unemployment figure of 3.14 million was about 88,000 lower than it would have been without the ‘Job Release Scheme’. The cost of these schemes, however, was estimated to be about 15 per cent higher than the average cost of other labour market policy measures and about 175 per cent higher per retiree than the net cost of an unemployed worker in the case of France and about 10 per cent and 112 per cent respectively in the case of the United Kingdom, where the compensation paid to the retiree was lower’ (ILO 1995, p. 40).

The ILO also found that industries were taking advantage of early retirement schemes to transfer some of the costs of restructuring their businesses and workforce to government. In effect, government was subsidising the cost of retrenching older workers—many of whom would have been made redundant regardless of the operation of the early retirement schemes— in unprofitable or outdated sectors of industry.

There is now an apparent consensus among OECD countries that early retirement does not lead to the creation of additional labour market opportunities for the unemployed, and that it is highly undesirable in the context of ageing populations and the threat this poses to prospective growth in living standards.

Governments have become more aware that early retirement exacerbates age dependency burdens, and that the demand for labour is not fixed but depends on factors such as price and skills. Getting rid of skilled, experienced older workers by no means guarantees that their places will be filled by less experienced younger workers.

6.8 Increasing labour market assistance

Soerensen’s (1998, p. 12) paper noted the possibility that reforms to social security provisions will be of limited effectiveness unless they are supported by a ‘reorientation of labour market policies and of the personnel management of companies and employers’.

Effectiveness of earlier labour market programs29

In the late 1980s, Australia adopted the ‘active society’ approach to income support arrangements that had been (and continues to be) advocated strongly by the OECD. This gave rise to the Newstart and Jobs Education and Training (JET) programs, and the Disability Reform Package, which sought to ensure that people receiving government income support payments were encouraged to remain economically active, look for work, undertake training or accept (in some cases) subsidised jobs.

Part of this approach included pilot arrangements, implemented in the 1990s, aimed at helping older unemployed people (55 and over).

This pilot, the ‘Older Unemployed Labour Market and Income Support Strategy’, was evaluated in May 1992 and found no measurable improvement in the employment or education outcomes of unemployed 55- to 64-year-olds. A major caveat is that the 1990–91 recession occurred

immediately after implementation of the Older Unemployed Strategy, and its impact casts considerable doubt over the extent to which the evaluation findings can be generalised.

There are a number of findings that might have relevance to the forms of assistance provided to older unemployed people, the potential for campaigns to influence attitudes to employment outcomes for older people and the potential for reforms to social security arrangements to counter the increasing dependence of older people on social security. Findings include:

  • older unemployed people appear to have better outcomes from wage subsidy assistance (JOBSTART) than from training through Skillshare, JOBTRAIN and Job Search Training;
  • there has been some success in changing community/employer attitudes to older unemployed people;
  • outcomes for women, despite their being less skilled, have been better than for men. This has been attributed to their being more motivated, more receptive to the idea of training and less selective in the type of work they are seeking;
  • there has been resistance to accepting older workers by training providers, apparently based on the belief that older workers have less chance of achieving a positive outcome;
  • both employers and older people believe the Government supports early retirement and feel the retirement ages of 60 for women and 65 for men are compulsory.

The evaluation recommendation to the Government to undertake a marketing campaign aimed specifically at changing attitudes was not implemented.

The 1993 evaluation of the JOBSTART program appeared to confirm that wage subsidies are an effective form of employment assistance for older people (DEET 1993). There has been very little new Australian information on the relative effectiveness of labour market programs for different age groups since 1993. It is still too early to know the results of innovative approaches by Job Network Members under Employment Services Market Reforms.

Information on Job Network Services as they relate to mature age workers is at Appendix 4.

7 Conclusion

Ross (1997, p. 9) notes that it is very difficult to devise a flexible retirement system with a means-tested Age Pension. He further states that ‘as long as benefits are available in lump sum form years before entitlement to the Age Pension, the stated objective of a significant reduction in Age Pension payments will not be achieved’ (p. 12).

This appears to be supported by RIM’s projections showing only a modest impact of the Superannuation Guarantee (SG) on net spending on age pensions (RIM Taskforce 1998). This is not to deny that the SG and associated reforms are projected to achieve a very substantial boost to retirement incomes, which has also been a stated objective of superannuation reforms.

A problem for analysing such criticisms of the current system is that we do not have good tools for quantifying the likely behavioural effects of the current system and possible changes to it. Projections using tools such as the RIM model typically use assumptions about behaviour (for example, that 25 per cent of lump sums are dissipated), but would not be able to model policy deliberately aimed at changing that behaviour without the introduction of further, similar assumptions.

It is impossible for the existence of what amounts to a guaranteed minimum income in Australia not to have an impact on behaviour. The system is very likely to reduce the average retirement age below what it otherwise would be.

Since this is part of the price we are prepared to pay for keeping people out of poverty, the real question is not whether the system induces early retirement but whether it does so more than is necessary to achieve its anti-poverty objectives. In this respect, certain aspects of the system appear undesirable:

  • the fact that we tend to write off people’s labour market potential after they reach 50, a trend that runs counter to the ‘active employment strategy’ that has been a major direction of social security policy since the mid-1980s;
  • that there may be a false perception among employers that older workers are not productive or are expected to retire relatively early;
  • the high and rising incidence of DSP, which appears to have taken on the form of an early retirement pension for some who in earlier times might have been able to continue participation in the paid workforce;
  • the opportunities for ‘double dipping’ prevalent in the occupational superannuation system—retirees can obtain and use up a superannuation lump sum funded in part by substantial superannuation tax concessions, and having done so can then obtain social security benefits.

Effort needs to go into refining policies and improving the research base on which policies are to be made. That said, Australia appears to be relatively well placed to weather the demographic ageing predicted to occur in the 21st century.

[ Return to Top Return to Section ]

Appendix 1: Taxation arrangements for superannuation

The retirement income system in Australia comprises the social security system and occupational superannuation. An important part of the latter is the concessional tax arrangements that apply to it, and which are estimated to impose substantial annual revenue costs on revenue.30 Tax affecting superannuation has been changed often in the past 15 years, and is astonishingly complicated. Since 1992, superannuation has been taxed at three stages:

  1. when contributions are made to a fund (at 15 per cent and between 15 and 30 per cent for incomes over $78,000 a year). However, members’ undeducted contributions are not taxed;
  2. when income is earned by a fund (at 15 per cent);
  3. when benefits are paid out—lump sums are taxed on payout according to a rising schedule and income streams (annuities) as income over the retirement period. Benefits accumulated since 1988 are taxed to the extent that they exceed $93,731 (indexed) at the rate of 15 per cent.31 Penalty tax is charged on amounts in excess of the ‘reasonable benefit limits’ (RBL), now $471,088 (lump sum) and $942,175 (if at least half in pension form).

In consequence, the tax treatment of superannuation accords with neither comprehensive income nor consumption tax principles, and in fact can be shown to be an amalgam of the two. Effective tax rates are much less than the comprehensive income ideal would imply, but significantly greater than zero.

Effective tax rate calculations are complicated if we take into account the pension means test as an implicit tax. This involves very high marginal rates over certain income ranges for certain types of investments, but low effective taxes when lump sums are invested to yield income streams less than the free areas or in owner-occupied housing (or are otherwise consumed). There are partial concessions for annuities.

The higher RBL applies if at least half the benefit is used to provide an income stream. A tax rebate of 15 per cent has been introduced as a further carrot to take benefits as income,32 and the return of the ‘undeducted purchase price is not taxable’. FaCS also provides incentives to take up income streams, under new rules that effectively exempt the parent capital from the asset test.33

[ Return to Top Return to Section ]

Appendix 2: Income support provisions

(from FaCS 1999, Attachments)

Australia’s income support and welfare system provides a range of income support payments for people aged 45 and over that are out of the workforce. The particular payment depends on individual circumstances such as age, marital status, gender, health and workforce experience. The payments available to unemployed people aged 45 or over but under Age Pension age are:

  • Newstart Allowance (NSA)
  • Mature Age Allowance (MAA)
  • Partner Allowance (PA)
  • Parenting Payment (PP)
  • Widow Allowance (WA)
  • Disability Support Pension (DSP)
  • Austudy Payment
  • Carer Payment

Newstart Allowance

Newstart Allowance (NSA) is payable to people aged 21 or over but under Age Pension age. To qualify for NSA, people must among other things be unemployed and must either satisfy the activity test or be exempt. They satisfy the activity test by undertaking job-search activity, by participating in another activity designed to improve their chances of finding work, or by complying with the terms of an Activity Agreement.

The types of activities that may be undertaken as an alternative to job search include voluntary work, self-employment development, rehabilitation and training. All activities other than job search require Centrelink’s approval.

Centrelink can also require people to enter into a Newstart Activity Agreement under which customers may agree to undertake a specific activity or a combination of activities that are likely to improve their employment prospects. An activity agreement may include activities such as job search, vocational training, job search training, paid work experience, voluntary work, work for the dole and participation in a rehabilitation program.

There are special arrangements for people living in remote localities, who can meet their activity test obligations by participating in any activity they have nominated and of which Centrelink approves. Activities likely to be approved include participation in activities in a resource centre, community adult education classes, helping with parks and gardens and some cultural activities.

Customers 50 or over can satisfy the activity test by undertaking 32 hours or more of voluntary work per fortnight or a combination of voluntary work and paid work of 40 hours or more per fortnight for an unlimited period each year, without an activity agreement.

Mature Age Allowance

Mature Age Allowance (MAA) was introduced in 1994 in response to the difficulties that older people were experiencing in finding employment in a tight labour market. MAA had a sunset clause of 30 June 1996, but was to be reviewed prior to that date to decide if a continuing payment was warranted.

MAA was introduced as a pension payment subject to the pensions means test, and provided relief from the ‘activity test’ for people who had tested their employment prospects in the labour market and had been unable to gain employment. A review of MAA in 1994–95 resulted in the Government’s decision to extend a modified MAA permanently from July 1996. People receiving MAA on 30 June 1996 were ‘saved’ and continue to be paid under pension conditions.

Currently, to qualify for MAA, a person must be at least 60 (and below Age Pension age), either have been in receipt of an unemployment payment for nine months prior to the claim or be transferring from a non-activity tested payment, and have no recent workforce experience. From July 1996, people claiming MAA are paid under the allowance income and assets tests.

Mature Age allowees are unable to gain access to Job Network services.

Partner Allowance

Partner Allowance (PA) was introduced in 1994 to ensure that married Job Search Allowance, Newstart Allowance, Sickness Allowance and Special Benefit recipients did not have a tax liability because of the replacement of the dependent spouse rebate for people with children by the Home Child Care Allowance.

A secondary objective of PA was to extend the principle underpinning the integration of family payments—to direct payments to the partners of allowees, usually women and the primary care-givers in families, in their own right. Prior to this, one member of a couple, usually the male, received payment at the married rate, which included an amount in respect of the dependent partner. The effect of PA was to split the married rate of payment so each partner received a payment in his or her own right. There was no activity test requirement and no regard to the person’s workforce experience.

The 1994 White Paper on Employment and Growth limited eligibility for PA to those born on or before 1 July 1955, with no dependent children and little or no recent workforce experience. At the same time, Wife Pension was abolished except for those currently receiving it and qualification for PA was extended to partners of pensioners.

Currently, to qualify for PA, a person must be the partner of a pensioner or allowee, born on or before 1 July 1955, have no dependent children and have little or no recent workforce experience. Recipients of PA can gain access to Job Network services.

Parenting Payment

Parenting Payment (PP) provides income support and opportunities for financial independence to the primary carers of children.

In general terms, to qualify for Parenting Payment a claimant must have the care of a dependent child/ren aged under 16.

PP (single)—Parenting Payment for sole parents—is paid under the pension income and assets tests, whereas PP (partnered)—Parenting Payment for partnered parents—is paid under allowance income and assets tests (with some modifications).

Widow Allowance

Widow Allowance (WA) was introduced from 1 January 1995 to provide an income support option for older women who have lost the support of a partner later in life and who have no recent workforce experience. When first introduced, WA was available only to women aged 50 and over who became widowed, divorced or separated after turning 50. Since March 1997, qualification has extended to include women aged 50 and over who became widowed, divorced or separated since turning 40.

Widow Allowance recognises the particular labour market difficulties faced by older women who have spent long periods out of the workforce caring for a family and who lose the support of their partner. It recognises that these women made career and life choices under more traditional expectations of their labour market participation and are now particularly disadvantaged in the labour market.

Women in receipt of WA have access to the Jobs, Education and Training (JET) program as well as a range of Job Network services.

Disability Support Pension

In 1991, a package of reforms was introduced to target disability-related payments more effectively and to enable greater participation of people with disabilities in the labour market. As part of this active strategy, invalid pension was re-titled Disability Support Pension (DSP). Payment of DSP is not inconsistent with rehabilitation or supported employment, so Sheltered Employment Allowance, Incentive Allowance and Rehabilitation Allowance were abolished.

The basic qualification for DSP is that the person be aged between 16 years and Age Pension age, have a physical, intellectual or psychiatric impairment and have a continuing inability to work. DSP recipients have access to the current range of Job Network services.

Austudy Payment

A person can qualify for Austudy Payment if he or she is 25 or over and satisfies the activity test by undertaking qualifying study. The person must be undertaking full-time study at an approved institution and must satisfy progress rules.

Carer Payment

Carer Payment is a means-tested pension payment for those who are providing constant care for a person who has a physical, intellectual or psychiatric disability. The care must be provided in the home of the person being cared for and must be required for an indefinite period or a minimum of six months. In most cases, the person receiving care must be receiving a social security or DVA pension or a social security allowance.

[ Return to Top Return to Section ]

Appendix 3: Impact of the Superannuation Guarantee (SG) and associated superannuation reforms

Under compulsory superannuation in force since 1992, employers must contribute to a superannuation fund a fixed percentage of the earnings of each employee who earns more than $450 (now $900) a month. The minimum contribution rate is currently 7 per cent and rises to 9 per cent by 2002. Many employers provide contributions in excess of the minimum.34

There are different projections of the impact of the SG, partly explained by the inclusion in earlier projections of employee and government co-contributions jointly amounting to 6 per cent of earnings. These are not now government policy.

EPAC’s projected costs of welfare spending took account of likely increases in superannuation wealth among the aged consequent on the SG and associated retirement income reforms announced in 1994. The likely impact was that though there would be little change in the proportion of the aged population receiving Age Pension—that is, about three quarters—the average payment would reduce from about 90 per cent of the current maximum to about 50 per cent (Clare and Tupule 1994, p. 63).

In the absence of the SG, welfare expenditure is projected by C&T to rise from 6.9 per cent of GDP in 1990 to 9.3 per cent in 2051. However, the SG and associated reforms reduces this latter figure by 1.4 per cent to 7.9 per cent (pp. 33–34). Age pension expenditure is projected to rise from 3.3 per cent of GDP in 1989 to 5 per cent in 2041, dropping to 4.5 per cent in 2051. Overall,‘the effects of retirement income reform are projected to have a substantial but significantly delayed impact on government expenditures’ (p. 35), amounting to some 1.5 per cent of GDP by 2051 (C&T p. 42, Table 3.16).

Some caveats should be made about this projection. First, it was previously envisaged that there would be a compulsory employee co-contribution of 3 per cent and a further 3 per cent government co-payment, but is not current government policy. Second, it does not take account of the cost of tax concessions for occupational superannuation (pp. 33–34). Brown (1993) has projected that there will be a net cost as a result until around 2025, beyond which net savings will begin to materialise.

In summary EPAC projects a total 60-year rise in social expenditures of 4 per cent of GDP before retirement income reform and 2.5 per cent after it (p. 41). The NCA’s projections for social security transfer expenditure are actually quite similar to EPAC’s (Table 6); the main difference between the two relates to health expenditures.

The RIM Task Force’s more recent (1998) projections indicate a much more modest impact of the SG, amounting to only 0.3–4 per cent of GDP (Figure 20), though the projected aggregate rise in pension expenditure is similar to other projections. One reason the SG has a modest effect is that for many retirees it ‘fills in’ the free area, without having an impact on the pension received. Another reason is that RIM assumes a 25 per cent loss of SG-financed lump sums to non-financial investments or other spending.

One possible negative effect of the SG and of better occupational superannuation in general will be that many more older people may have sufficient superannuation savings to induce earlier retirement. This effect is augmented by the availability of preserved superannuation after 55, and by the very slow phasing in of the higher (age 60) preservation age (details in S6.2).

This worry is exacerbated by the possibility of double dipping, which means that superannuants can dissipate part of their superannuation wealth prior to Age Pension age with very little negative impact on their retirement incomes, taking into account the high effective tax rates they can face over some ranges of income.35 In effect the social security system will subsidise some part of their pre-Age Pension age consumption in this manner.

On the face of it the simple solution is to require occupational superannuation to be preserved until pension age. However, this will have a negative impact on many people who are not necessarily in a very good situation and so is not the neat solution we might hope.

[ Return to Top Return to Section ]

Appendix 4: Job Network services

(This is an edited version of the relevant section in FaCS 1999)

The ‘Job Network’ was a major reform to the provision of labour market assistance introduced from May 1998. There are currently approximately 300 Job Network Members in more than 1400 sites. Around 2.2 per cent of Job Network Members specialise in services for ‘mature workers’.

The range of services available through the Job Network includes:

  • Job Matching—gathering vacancies from employers and matching unemployed people to jobs;
  • Job Search Training—providing training in job search techniques such as resume and interview preparation to secure employment;
  • Intensive Assistance—individually tailored help for jobseekers assessed as requiring intensive and personalised assistance to secure employment.

Other forms of assistance include the Community Support Program, New Enterprise Incentive Scheme and Entry Level Training Assistance.

The most significant determinant of the type and level of Job Network assistance provided to older people is the Job Seeker Classification Instrument.

The Job Seeker Classification Instrument score is calculated by collecting information on the particular characteristics of a jobseeker. Older people can accumulate Job Seeker Classification Instrument points quite quickly and, while this can help a person gain early access to Intensive Assistance, it can also work to preclude older people from accessing Job Search training—a simple and cost-effective form of employment assistance.

This may explain why there are noticeably smaller proportions of Newstart Allowance recipients aged 50 and over receiving this form of assistance. Administrative data for November 1998 indicate that the proportion of each age group gaining access to job search training increases until around 50 years of age but then drops rapidly for older age groups.

The second way rapid accumulation of Job Seeker Classification Instrument points can restrict access to Job Network assistance is that a high score, as well as specific responses to some questions on the instrument, may indicate that a jobseeker has special needs triggering the requirement for a Special Needs Assessment.

The fundamental purpose of this assessment is to decide whether a person would benefit from Intensive Assistance. A negative assessment means the applicant cannot be referred for Intensive Assistance or Job Search Training. Neither jobseekers nor job network members can have such decisions reviewed. The person can be helped only through Community Support Program (a small program), FaCS specialist employment services for people with disabilities, or other services they find themselves.

Older unemployed people do appear to have a high level of access to Intensive Assistance. In November 1998, about 16 per cent of unemployed customers (120,800) had recorded on their income support record that they were receiving Intensive Assistance. The proportion of each age group being assisted increased with age until around 60. Between 20 and 25 per cent of unemployed customers aged 40 to 59 were recorded as receiving Intensive Assistance.36

Access to Intensive Assistance drops dramatically only for Newstart Allowance (NSA) recipients aged 60 and over. Only about 13 per cent of this age group were receiving assistance. While Mature Age Allowees (60 and over) cannot get access to Job Network services, people aged 60 and over are eligible if they remain on NSA; a finding at odds with the picture of access increasing with age. It probably reflects a greater level of assessed inability to benefit, agreement between providers and customers on ‘opting out’ of assistance, and widespread confusion about Job Network eligibility for customers aged 60 and over.

Caveats have to be placed on the apparent high level of access to Intensive Assistance by older unemployed people. The figures above, for November 1998, are very early and at this early stage little is known about the level or forms of assistance actually being provided to different groups of jobseekers receiving Intensive Assistance by Job Network members.

[ Return to Top Return to Section ]


  1. ABS ‘Deaths’, 3302.0, 1996, life tables.
  2. In fact, this EPAC Background Paper is not an official Office of EPAC product:‘the views expressed in this paper are those of the individual authors only …’.
  3. As is well known, population ageing will require increased transfers from the economically active to the economically inactive population, an economic truth that holds irrespective of the financing mechanism adopted. The advantage of private saving is that these transfers come out of capital income, rather than taxation, and so appear to place less stress on the taxation system. However, this proposition needs to be modified if—as is the case with the private superannuation system—part of private savings is financed by public tax expenditures at the time those savings are made and/or interest income accrues.
  4. Goss (1994, cited in Saunders, 1996, p. 19) has calculated that 40 per cent of the increase in health expenditure projected by Clare and Tupule is due to total population growth and a further 41 per cent to the assumed increase in health expenditure at all ages, leaving only 19 per cent arising from the ageing of the population alone.
  5. See Podger (1998) for a description of the factors underlying these various projections and their differences.
  6. The figures quoted are from ILO Economically Active Population 1950–2010, December 1996. Similar figures are quoted in Blondal and Scarpetta (1998, Table 11.1), and are based on calculations of retirement probabilities as defined in the Notes to that table (p. 54).
  7. Bacon (1997, p. 33) estimates that since 1978, about 25 per cent of the fall in participation rates for males aged 45–54 and around 75 per cent of that for males aged 55–59 was due to early retirement.
  8. Currently about 10 per cent of employed workers among males 45–64 and 4 per cent of females. In aggregate, part-time workers comprise about a quarter of all employed people.
  9. However, while the participation rate for male workers age 60–64 subsequently stabilised, it has rebounded only slightly if at all from its 1983 lows and still rests at about 50 per cent. Unikowski (1996) speculates that the availability of the service pension may have contributed to the expectation that the unofficial retirement age for men is around 60 years (p. 10).
  10. A 10-year longitudinal study of about 8000 older Americans sponsored by the US Social Security administration.
  11. Assuming such testing has an employment focus. Activity testing can also encompass participation in socially valued activities such as voluntary work, in which case the condition of the labour market becomes less important.
  12. It is only a rough guide; some of those on payment and not unemployed may also be in the labour force.
  13. Based on the Working Paper by Blondal and Scarpetta (1998).
  14. It is taxable so to this extent is reduced with rising personal income. Until recently, a special surcharge also applied.
  15. Disability Support Pension is payable to a person who has a physical, intellectual or psychiatric impairment of 20 points or more under the impairment tables and, because of the impairment, has a continuing inability to work (in the next two years) at award wages or participate in education or training for work for at least 30 hours a week.
  16. This increase partly reflects the merging of SEA and Invalid Pension in forming DSP, and the increased restrictions on Sickness Allowance—Walsh, p. 14.
  17. I am indebted to Julia Perry for providing these figures.
  18. The major categories of impairment for DSP are musculo-skeletal/connective tissue and psychiatric/psychological; these categories account for more than 60 per cent of DSP recipients. Factors limiting employment for people from the first group are generally related to age (65 per cent are over 50) and education (a large proportion has little formal education and a work history in manual labour) rather than the disability itself (DSS 1997 p. 3).
  19. OECD, 1988, p. 62; Australian Institute of Health, 1997, p. 3.
  20. There has been compulsory preservation of all new superannuation benefits from 1 July 1996, less an amount that equates broadly to what people can now take out of a fund were they to resign from their employment.
  21. The preservation age will be increased such that by 2025 it will be 60 for anyone born after 30 June
  22. 1964. However, it will remain at 55 for anyone born before 1 July 1960.
  23. Since 1 July 1995, no new grants of wife pension have been made. Partners of age pensioners must claim another form of income support that reflects their individual circumstances.
  24. This scheme, introduced in July 1998, provides a bonus to people who remain in full-time employment and defer taking up the age pension. The bonus is weighted toward retirement at age 70, at which age it reaches its maximum value ($21,831 single, and $36,428 couple). The bonus is indexed to movements in the age pension.
  25. For information about this, see
  26. Almost half of lump sums are transferred to other superannuation funds or rolled over to the same fund; this most often occurs with larger lump sums. Transfers and rollovers are mostly used to by retirement income streams known as allocated pensions; these entail regular drawings from an account, subject to legislated annual limits. The maximum drawing rate would liquidate the account by age 80, whereas the minimum drawing rate would never liquidate the account. Allocated pensions are more attractive than annuities because of their greater flexibility; because the purchase price is not inflated by adverse selection problems; and because beneficiaries can maximise age pension receipts later in life.
  27. EMTRs over the income range where pension is withdrawn are typically in the order of 65–70 per cent. The Government’s tax package announcement that the pension taper will be reduced to 40 per cent should ameliorate this somewhat, though not for those subject to the separate assets test. Over the asset test range the effective rate of imputation can be calculated as 15.6 per cent, implying that an asset yielding 5 per cent a year is subject to an effective tax rate of around 155 per cent. Under a 40 per cent taper the implied asset test imputation rate becomes 19.5 per cent. However, fewer than 5 per cent of pensioners are currently subject to the assets test (DSS Customers: A Statistical Overview, 1997, p. 5).
  28. The broad principle underlying the current preservation rules is that benefits that have received concessional taxation treatment should be preserved until retirement. However, this principle is not fully reflected in the rules (for details see Rothman, 1997, p. 6) which, in addition, are enormously complex for funds to administer. RIM Task Force calculated that in 1995 about 65 per cent of account balances were not preserved, with this proportion falling over time as a greater proportion of contributions are subject to the SG and are therefore preserved (Figure 21). SG assets currently account for 12 per cent of total superannuation assets; this will grow to about 23 per cent in 2020.
  29. Previously, release was approved by the Insurance and Superannuation Commission (ISC). This body has been abolished and its functions split between the Australian Securities and Insurance Commission (ASIC), and the Australian Prudential Regulation Authority (APRA).
  30. This section borrows heavily from FaCS 1999.
  31. Though there is no consensus on what these costs are, or how they should be measured. There has been substantial criticism of the cost estimates contained in the Treasury’s annual Tax Expenditure Statement, currently running at $9.3 billion a year.
  32. Benefits accumulated prior to 1983 are substantially untaxed. In 1983 tax on lump sums was increased to 15 per cent up to a threshold (then $50,000) and to 30 per cent above the threshold. In 1988 tax was introduced on employer contributions and fund earnings, so the lump sum tax was reduced in consequence. Lump sums in excess of the RBLs are taxed at the highest marginal rate. These limits are indexed to average weekly ordinary time earnings (AWOTE).
  33. Pensions and annuities purchased with superannuation monies, up to the RBL, are taxed as ordinary income less a 15 per cent rebate. Any excess to the RBL is taxed as ordinary income with no rebate.
  34. Income streams are means tested on the basis of their characteristics. Products that provide an income stream for life or for a fixed term of life expectancy (or at least 15 years where life expectancy exceeds 15 years) and which meet other requirements (for example, are non- commutable and have no residual capital value) are exempt under the asset test. All other income stream products are asset tested. Under the income test, the actual income paid from income stream products with a term in excess of five years is assessed, with a deduction allowed based on the purchase price. Assessable income from other income streams is deemed at standard rates.
  35. Those employers failing to contribute are assessed for a payroll tax called the Superannuation Guarantee Levy, which is credited to the relevant employees. In fact very few employers pay the levy, probably because it is not tax deductible whereas superannuation contributions on behalf of employees are.
  36. This argument is complicated by the impact of superannuation assets on other social security payments, which are generally more tightly means tested than the Age Pension (DSP is an exception). Current legislation requires that superannuation assets be taken into account for means test purposes from age 55; that is, the preservation age.
  37. The figures presented in this paper on customers receiving Intensive Assistance are an undercount.

[ Return to Top Return to Section ]


Anderson P., Gustman, A. & Steinmeier, T. 1997,‘Trends in male labour force participation and retirement: some evidence on the role of pensions and social security in the 1970s and 1980s’, NBER Working paper 6208, October.

Atkinson, M. & Creedy, J. 1996,‘The choice of early retirement age and the Australian superannuation system’, in DSS 1996, pp. 161–194.

Australian Bureau of Statistics (ABS) 1995, Australian Social Trends, Cat 4102.0. Australian Bureau of Statistics (ABS) 1998, Population Projections, Cat 3222.0 (see p. 9).

Australian Bureau of Statistics (ABS) 1998a, Retirement and Retirement Intentions, Cat 6238.0. (see p. 13).

Australian Bureau of Statistics (ABS) 1998b, Australian Employment and Unemployment Patterns, Cat 6286.0.

Australian Institute of Health and Welfare 1997,‘The impact of population ageing on health expenditures in Australia in the 21st century’—chapter in Deeble, J. 1997, Long Term Care Insurance, National Centre for Epidemiology and Population Health.

Bacon, B. 1997,‘Work, retirement and dependency’, People and Place 5(2), pp. 26–39.

Bacon, P. & Gallagher, P. 1996,‘Early retirees—trends and their use of superannuation benefits and social security payments’, in DSS 1996, pp. 41–84.

Baldwin S, & Falkingham, J. (eds) 1994, Social Security and Social Change: New Challenges to the Beveridge Model, Harvester Wheatsheaf.

Blondal, S. & Scarpetta, S. 1998,‘The retirement decision in OECD countries’, OECD Ageing WorkingPapers1.4.

Bound, J. & Waidman, T. 1992,‘Disability transfers, self-reported health and the labour force attachment of older men: evidence from the historical record’, Quarterly Journal of Economics, November, pp. 1393–1419.

Bound, J. & Burkhauser, R.V. 1998,‘Economic analysis of transfer programs targeted on people with disabilities’, Card and Ashenfelter, eds, Handbook of Labour Economics, Elsevier Science, Amsterdam.

Brown, R. 1993,‘Integration of public and private provision for retirement’, in Disney and Briggs, eds, pp. 143–154.

Chan, S. & Stevens, A. 1999,‘Job loss and retirement behaviour of older men’, NBER Working PaperNo 6920, February.

Clare, R. & Tupule, A. 1994,‘Australia’s ageing society’, EPAC Background Paper No 37. Congressional Budget Office 1999,‘Raising the earliest eligibility age for Social Security benefits’, CBO Papers, Washington DC.

Cornish, R. 1997,‘An investigation of the retirement behaviour of older Australians’, Fifth Annual Colloquium of Superannuation Researchers, University of Melbourne.

Creedy, J. 1994,‘Means-tested versus universal benefits: social welfare and poverty alleviation with endogenous earnings’, Department of Economics Working Paper, University of Melbourne.

Creedy, J. & Taylor, P 1993,‘Population ageing and social expenditure in Australia’, Australian Economic Review, 3rd quarter.

Deeble, J. 1997, Long Term Care Insurance, National Centre for Epidemiology and Population Health.

Department of Employment, Education and Training (DEET) 1993, An Evaluation of JOBSTART, November.

Department of Family and Community Services (FaCS) 1999,‘Submission to the House of Representatives Inquiry into Issues Specific to Older Workers Seeking Employment, or Establishing a Business, Following Unemployment’, mimeo.

Department of Social Security (DSS) 1996,‘Early retirement seminar’, 14 December 1995, AGPS, Canberra.

Department of Social Security (DSS) 1997,‘National forum on future directions for disability policy, 24–25 July 1997’, Outcomes and Forum Papers, DSS, Canberra.

Department of Social Security (DSS) 1998, Reform of Income Support for People with Disabilities: Discussion Paper, April.

Disney, J. & Briggs, L. 1994,‘Social security policy: issues and options’, Social Security Policy: the Future Conference, November 1993, AGPS, Canberra.

Disney, R. 1997, Can We Afford to Grow Older? A Perspective on the Economics of Ageing, MIT Press, Cambridge, Mass.

Doyle, S. & Piggott, J. 1998,‘Mandatory annuity design: a preliminary study’, Society of Actuaries Conference: Retirement Needs FrameworkOrlando, Florida (draft—mimeo).

Encel, S. 1998,‘Downsizing, rightsizing and capsizing’, Fifth National Conference On Unemployment, 1–2 October, RMIT, Melbourne.

Economic Planning Advisory Council (EPAC) 1988,‘Economic effects of an ageing population’, Council PaperNo 29, October, AGPS.

Economic Planning Advisory Council (EPAC) 1992,‘Economic and social consequences of Australia’s ageing population: preparing for the 21st century’, Background Paper No 23, December.

Economic Planning Advisory Council (EPAC) 1993,‘Retirement income perspectives’, Background PaperNo 30, July, AGPS, Canberra.

Gallagher, P. 1995,‘The policy use of the products of the retirement income modelling task force’, Third Annual Colloquium of Superannuation Researchers, University of Melbourne.

Gatiss, R. 1997,‘Of life, death and superannuation’, Fifth Annual Colloquium of Superannuation Researchers, University of Melbourne.

Gruber, J. & Wise, D. 1997,‘Social security programs and retirement around the world’, NBER Working Paper 6134, August.

Gruber, J. & Wise, D. 1998,‘Social security and declining labour force participation’, American Economic Review: Papersand Proceedings 88(2), May, pp. 158–163.

Gustman, A. & Steinmeier, T. 1985,‘The 1983 Social Security reforms and labour supply adjustments of older individuals in the long run’, Journal of Labor Economics 3(2), pp. 237–253.

Human Rights and Equal Opportunities Commission 1999, Age Matters? A Discussion Paper on AgeDiscrimination,

Ingles, D., Jackson, W., Podger, A., and Raymond, J. 1982,‘Taxation Expenditures’, Research Paper No 17, Development Division, Department of Social Security, March 1982.

Ingles, D. 1998a,‘Older workers, disability and retirement in Australia’, paper presented to a conference ‘Income Support, Labour Markets and Behaviour: a Research Agenda’, FaCS and CEPR, ANU, November.

Ingles, D. 1998b,‘Universal and selective benefits revisited: reflections on the optimal rate of taper’, Social Security Journal, 1998/2.

International Labour Office 1995, World Labour Report, ILO, Geneva (see Ch 2:‘Ageing societies: problems and prospects for older workers’).

Jackson, N. 1999,‘Understanding Population Ageing: A Background Paper’, Australian Social Policy, 1999(1), pp. 203–24.

Jacobzone, S., Cambois E., Chaplain, E. & Robine, J. 1998,‘The health of older persons in OECD countries: is it improving fast enough to compensate for population ageing?’, Labour Market and Social Policy Occasional Papers No 37, OECD. Paris.

Johnson, P. 1998, Older Getting Wiser, Report to the Institute of Chartered Accountants, Australia.

Kalisch, D. 1992,‘Superannuation and the Age Pension: double dip or top-up’, EPAC 1992, pp. 73–93.

Kalisch, D. 1999,‘Maintaining prosperity in an ageing Australia: an Australian perspective with an international flavour’, Keynote address, Aged Care Australia and Australian Association of Gerontology Joint National Conference ‘The Age of Celebration and Expectation’, Sydney, 5–8 September.

Lonsdale, S. & Seddon, J. 1994,‘The growth in disability benefits: an international comparison’, in Baldwin and Falkingham, (eds) 1994, Social Security and Social Change: New Challenges to the Beveridge Model, Harvester Wheatsheaf.

Maloney, T. 1997,‘The impact of recent welfare reforms on labor supply behaviour in New Zealand’, IRP Discussion Paper No 1149–97.

McCallum, J. et al. 1994, Hospital and Home: A Longitudinal Study of Hospital, Residential and Community Service Use by Older People Living in Dubbo NSW, Office of Ageing, NSW Government Printer, Sydney.

McCormack, J. 1996,‘Labour market exit and re-entry of male workers 55+’, DSS 1996, pp. 121–159.

McDonald, P. & Kippen, R. 1999,‘Ageing: the social and demographic dimensions’, Productivity Commission Conference ‘Policy Implications of the Ageing of Australia’s Population’, Melbourne, March 18–19.

Morrow, I. 1998,‘Determinants of male Disability Support Pension receipt 1971–96’ FaCS, mimeo based on paper presented at Economists Conference, Sydney 1998.

National Commission of Audit (NCA) 1996, Report to the Commonwealth Government, June, AGPS, Canberra.

New Zealand Periodic Report Group 1997, 1997 Retirement Income Report: Building Stability, NZ Government Printer, Wellington.

Newman, J. 1999,‘The future of welfare in the 21st century’, address to the National Press Club, 29 September.

O’Brien, M, 1998,‘Old enough for the scrapheap? Employment issues for the baby boomer generation’, Fifth National Conference on Unemployment, 1–2 October, RMIT, Melbourne.

OECD 1994, OECD Jobs Study: Evidence and Explanations Part II—the Adjustment Potential of the Labour Market, Paris.

OECD 1996, Caring for Frail Elderly People: Policies in Evolution, Paris. OECD 1996, Employment Outlook, Paris.

OECD 1997, Incentives and Disincentives to Early and Late Retirement: Joint ILO–OECD Workshop on Development and Reform of Pension Schemes, December 15–17, Paris.

OECD 1998a,‘The labour market and older workers’, Social Policy Studies No 17. OECD 1998b, Economic Outlook, Paris.

OECD 1998c,‘Maintaining prosperity in an ageing society: background report’, Report to Ministers on AgeingPopulationsC/MIN(98)9/ANN, Paris.

OECD 1998d, Employment Outlook, June (Ch4—’Workforce ageing in OECD countries’). OECD 1998e, Ageing Populations: Social Policy Implications, Paris.

Perry, J. & Whiteford, P. 1998, Social Security Trends and Issues, FaCS (mimeo).

Podger, A. 1998,‘Where are we heading along the ageing public/private pathway? Social, health and economic challenge’, CEDA/National Ageing Research Institute Forum on Our Ageing Population: Implications for Policy Planning, 21 April, Melbourne.

Quinn, J., Burkhauser, R. & Myers, D. 1990, Passing the Torch: the Influence of Economic Incentives on Workand Retirement, Upjohn Institute, Michigan, USA.

Quinn, J. 1998,‘The labour market, retirement and disability’, Conference on Income Support, Labour Markets and Behaviour: a Research Agenda, FaCS/CEPR, ANU, November.

RIM Taskforce 1998, Presentation to FaCS seminar, July (unpublished).

Ross, J. 1997,‘The gap between objectives and policy outcomes in Australian retirement incomes’, Fifth Annual Colloquium of Superannuation Researchers, University of Melbourne.

Rothman, G. 1997,‘Aggregate analysis of policies for accessing superannuation accumulations’, Fifth Annual Colloquium of Superannuation Researchers, University of Melbourne.

Saunders, P. 1996,‘Dawning of a new age? The extent, causes and consequences of ageing in Australia’, SPRC Discussion paper No 75, December.

Soerensen, O.B. 1998,‘Variability of retirement practices: and appropriate response to labour market developments’, Report to 26th General Assembly of the International Social Security Association, Morocco, 25–31 October.

Treasurer and Minister for Social Security 1997, Savings: Choice and Incentive, AGPS, Canberra.

Unikowski, I. 1996,‘Departing for a life of sunshine: background to the seminar on early retirement in Australia’, DSS 1996.

VandenHeuval, A. 1999,‘Mature age workers: are they a disadvantaged group in the labour market?’, Australian Bulletin of Labour 25(1), March, pp. 11–12.

Waidman, T. & Manton, K. 1998, International Evidence on Disability Trends Among the Elderly, Office of Disability, Aging and Long-Term Care Policy, Office of The Assistant Secretary For Planning And Evaluation, US Dept of Health And Human Services, Washington DC.

Walker, A. 1998,‘Australia’s ageing population’, NATSEM Discussion Paper No 27, February. Walsh, J. 1997,‘Trends and characteristics of disability support pensioners’, DSS 1997, pp. 11–36.

Whiteford, P. & Bond, K. 1999,‘Income support, retirement incomes and the living standards of older people in Australia: trends and comparisons’, paper for the Productivity Commission Conference ‘Policy Implications of the Ageing of Australia’s Population’, Melbourne, March 18–19.

World Bank 1994,‘Averting the old age crisis’, Policy Research Report, Oxford University Press.

[ Return to Top Return to Section ]

Last updated: