ARPA: Tales of Robin Hood (part 3): The long view – social policies and the life cycle

John Hills Good Times, Bad Times: The Welfare Myth of Them and Us, Policy Press, Bristol, 2014 (336 pp). ISBN 9781447320036 (paperback) RRP $45.99.

Social policies shift resources between individuals and families, through taxes, cash payments and in-kind services, and to understand the impact of these measures fully, we need to look at several dimensions. One relates to policy design: the first article in this series discussed how social policy can have ‘Robin Hood’ goals (redistribution from rich to poor) or ‘piggy bank’ goals (redistribution between stages of life, often from the middle to the young and old) (Whiteford 2015a). Another relates to time, which complicates the picture considerably. The measures I discussed in the second article of this series were based on data collected at a point in time, which enabled a ‘snapshot’ comparison of what different kinds of households received and contributed in a given year (Whiteford 2015b). But household and individual circumstances change over time. Some changes happen as life takes its course: for example, children begin life as receivers, but grow up, leave school, start work and become contributors. Other changes are changes of fortune: a person in work may lose their job or become sick and need income support for a period of time.

This third article in the series engaging with John Hills’ Good Times, Bad Times: The Welfare Myth of Them and Us examines how social policies in the United Kingdom and Australia redistribute across the life course and between rich and poor over time. The mix between redistribution to the poor and redistribution across the life cycle cannot be observed directly in annual data on incomes or social spending, since annual data cannot identify the extent to which households have already paid for benefits in past years, or the extent to which they will do so in the future, or the extent to which current net taxpayers will be future beneficiaries. Different strategies are needed, and they are my focus here.

In Chapter 3, Hills goes back to the original calculations made for the fictional Osborne and Ackroyd families in 1989 for the Spongers documentary (see Whiteford 2015b) and points out that while the lower income Ackroyds get much more in cash benefits than the higher income Osbornes over their working life, the Osbornes would get about 50 per cent more in their lifetime from pensions and about 14 per cent more from the National Health Service, simply because they could be expected to live longer. They would also have received 80 per cent more in education benefits, because their children stayed at school longer and went on to university (which was free at the time). Overall, he estimates that the total amount of welfare state spending received by the Osbornes was only 5 per cent less than spending received by the Ackroyds. However, the Osbornes did pay more than twice as much in lifetime taxes out of lifetime earnings that were 2.7 times higher.

Hills then asks us to consider aggregate taxes paid and benefits received by age groups as if there was a representative ‘average’ household stuck in a kind of ‘Groundhog Day’—where the average household did grow older but the taxation and spending systems stayed exactly the same (at their 2005 parameters) for all of the household’s life. Over the course of this assumed average life, he calculates that by the age of 59 the ‘average household’ would have received £117,000 in cash benefits and £218,000 in services, but paid £536,000 in direct and indirect taxes, leaving them £202,000 out of pocket. Nevertheless, they would have already ‘got back’ 63 per cent of their taxes paid. Following them through to age 85, their total cash benefits would grow to £315,000 their services to £342,000 and their taxes to £692,000. As very old people, they are now only £35,000 out of pocket.

If we replicate these calculations for Australia in 2009–10 we get a picture that is similar in some regards, but very different in important details. By the age of 65, an Australian household ‘on average’ would have received $216,000 in cash benefits and $565,000 in services, but paid total taxes of $859,000 leaving them $79,000 out of pocket (Australian Bureau of Statistics 2012). In contrast to the UK, they would have got back out of the system roughly 91 per cent of the taxes they paid. After the age of 65 they would receive a further $158,000 in cash benefits and $188,000 in services, but pay $72,000 in direct and indirect taxes. In contrast to the UK, where the benefits they received after 60 were twice as much as the taxes they paid, for Australians over 65 the benefits are nearly five times their taxes. What makes the Australian situation different is the very low level of taxes paid by older households. Looked at over the life course, Australian households on average would have received over a $1.1 million in benefits ($374,000 in cash benefits and $753,000 in services), but paid $931,000 in taxes. The difference between average lifetime taxes and social spending simply reflects the fact that social spending is a smaller proportion of total spending in Australia than in the UK.

Calculations of this sort are complicated by the fact that the level and distribution of benefits and taxes do change over time as policies change, and so the volume of redistribution to different age groups can change quite significantly. Hills (p. 56) points out that the UK Labour Government (1997–2010) focused on spending on pensioners and on families with children, while spending cuts under the Conservatives since 2010 have tended to protect older people at the expense of people of working age and those with children. Correspondingly, the Australian Labor Government in Australia from 2007 to 2013 significantly increased age pensions (Whiteford 2014), but reduced spending on families with children (Redmond, Whiteford & Adamson 2011).

Because the real world is complicated, various ways of modelling the lifetime distribution of benefits and taxes are required. In a comparative study looking only at cash transfers and income taxes, Falkingham and Harding (1996) compared Australia and the UK and estimated that in the mid-1980s in Australia, 38 per cent of lifetime benefits received by individuals, on average, were financed through taxes they paid at another stage in their life cycle, and the remaining 62 per cent of lifetime benefits involved redistribution between rich and poor; in the UK these shares were reversed, with 38 per cent of lifetime benefits involving redistribution between individuals and 62 per cent involving redistribution over different phases of the life cycle of the same individual.

Hills (pp. 65–67) reports on a more recent calculation of this sort simulating lifetime benefits and taxes for 4,000 different households using the tax and social security systems in place in 1991. The total level of spending received by each decile of lifetime income was found to be remarkably similar across income groups—ranging from a little over £300,000 for the poorest lifetime income decile to a little under £300,000 for the richest lifetime income group. What differed most were the taxes paid by lifetime income groups—£140,000 for the poorest and £500,000 for the richest. Overall, he concluded that nearly three-quarters of what the British welfare state was doing at the time was the ‘piggy bank’ rather than ‘Robin Hood’ role. The reason for this increased share of redistribution across the life course appears to be that the more recent study included non-cash benefits and indirect taxes.

Just this month, a new study by the UK Institute of Fiscal Studies (IFS) (Levell, Roantree & Shaw 2015) called ‘Redistribution from a lifetime perspective’, updates and deepens this type of analysis. As with other studies, this work involves modelling the lifetimes from age sixteen to the end of life of a cohort of 5,000 individuals born in the first part of the ‘baby boom’ (between 1945 and 1954), assuming they lived all their life under the current (2015–16) UK tax and benefit systems (although simulations using the systems in force between 1978–79 and 2016–17 are also presented). The simulations model direct and indirect taxes and cash benefits, but not non-cash benefits, such as healthcare and education.

It should be noted that these types of estimates are also artificial: outcomes may differ for different age cohorts and no one lives their life under an unchanging tax and benefit system. But these simulations provide important information on the split between intrapersonal redistribution across the life cycle and interpersonal redistribution between rich and poor.

The starting point for Levell, Roantree and Shaw’s (2015) results flows from the difference between incomes in cross section and longitudinally. For example, they find that, while only 22 per cent of the cross-section lowest income decile are currently employed, on average persons in the lowest lifetime income decile spend about two-thirds of their working life in paid employment. As a result while more than 80 per cent of the poorest 30 per cent are entitled to means-tested benefits in at a single point in time (cross-section), for the poorest 30 per cent of the lifetime distribution (longitudinal) the figure falls to around 50 per cent. Strikingly, the richest 10 per cent in lifetime terms spend almost one-fifth of their lifetimes entitled to one of the main means-tested benefits.

Income persistence (having a similar income over time) is strongest at the top and bottom of the income scale, but even so nearly 80 per cent of those whose lifetime income falls in the poorest decile spends time in other deciles along the way, with even a small minority (2–3 per cent) spending periods in the richest 10 per cent. While 35 per cent of the richest lifetime decile never leaves that group, nearly two-thirds spend periods in other groups, with around 5 per cent having been in the poorest decile at some stage.

Because of these substantial income fluctuations across people’s lives, the extent of redistribution between rich and poor is considerably less than estimated from the kind of snapshot analysis discussed in detail in the second article in this series (Whiteford 2015b). For example, at a point in time 36 per cent of individuals are net welfare state recipients—they receive more in cash benefits than they pay in direct and indirect taxes, but over the course of a lifetime, only 7 per cent of people receive more than they paid. Depending on the counterfactual assumed between 57 per cent and 60 per cent of total redistribution involves transfers across the life course, with 43 to 40% being redistribution between lifetime rich and poor.

We do not appear at the moment to have any more recent Australian study that models the lifetime distribution of benefits and taxes here, although the Productivity Commission (2015) will be publishing such a study later this year. Rothman (2009) looked at the distribution of cash benefits for older people (age and service pensions) as well as superannuation tax concessions by deciles of lifetime income, assessing the impact of a range of initiatives made in the 2009 Budget. While this analysis shows the overall progressive effects of the substantial pension increase and related changes made in 2009, it is not as comprehensive as Hills’ work, because it does not include benefits paid to working age households or indirect taxes. It seems very likely that if it were possible to replicate Hills or the IFS analysis in Australia that the share of redistribution across the life course would increase, primarily due to the inclusion of health care spending. But it also seems very likely that Australia would still have a larger share of spending that is redistribution between income groups, given the extreme ‘Robin Hood’ profile of our spending on cash transfers.

It is also worth noting that part of the reason for the relatively flat profile of UK spending across lifetime income groups is related to the higher life expectancy of the better-off, meaning that they potentially benefit from more years of pensions and health care. In this context, research by the Office of National Statistics (2011) for England and Wales found that at age 65, the life expectancy of males classified by occupation as ‘higher managerial and professional’ was 18.8 years, compared with 15.3 years for those assigned to occupations classified as ‘routine’. At the same age, the life expectancy of females classified by occupation as ‘higher managerial and professional’ was 21.7 years, compared with 18.5 years for those assigned to occupations classified as ‘routine’. Levell and colleagues (2015), in contrast, estimate differences in age at death for deciles of lifetime income, with a particularly large gap for those in the poorest lifetime decile for whom average age at death is 60 years, while for the second poorest decile, average age at death is closer to 80 years. People in the richest lifetime income decile live, on average, to over 85 years.

A recent Australian study by Philip Clarke and Andrew Leigh (2011) found that at age 60 the difference between life expectancy for those in the highest income quintile and the lowest income quintile was five years for men and 5.4 years for women, with men in the lowest quintile expected to live to just over 78 years and women in the same income group to 83 years. Comparing life expectancy by income group is not the same as classification by occupation, nor is current income group the same as lifetime income group, but a broader claim about socio-economic difference can be supported. Moreover, there is one factor—unique to Australia among rich countries—that would offset the socio-economic gradient in life expectancy: we are the only OECD country where the public pension is income-tested and excludes most of the highest-earning 20 per cent of people over 65 years of age. Richer Australians will be likely to benefit from health care spending for longer but not from pensions. Offsetting this to some extent is that older Australians pay relatively very low taxes, so we boost their disposable incomes in this way rather than through pension spending, with the distribution of these tax expenditures favouring better-off older people. A future instalment in this series will discuss the distributional implications of pension tax expenditures in the UK and Australia.

In summary, because of our uniquely income-tested age pension, Australia redistributes less across the life course than does the UK, with the higher share of people over pension age in the UK reinforcing the distribution of resources to this age group. Where Australia does offset this effect is by clawing back much less in direct and indirect taxes from older households.

The life course perspective offers a powerful corrective to many of the myths of the welfare state, both in the UK and Australia. Aspects of these myths will be explored in future articles in this series, but the next instalment will consider a further time lens on people’s lives: the risks that individuals face at different times in their lives and the consequent variations that arise within a year or from year to year in working life and retirement.

REFERENCES

Australian Bureau of Statistics 2012, Government Benefits, Taxes and Household Income, Australia, 2009–10, Cat. no. 6537.0, Australian Bureau of Statistics, Canberra.

Clarke, P. & Leigh, A. 2011, ‘Death, dollars and degrees: Socio-economic status and longevity in Australia’, Economic Papers: A Journal of Applied Economics and Policy, vol. 30, no. 3, pp. 348–355.

Falkingham, J. & Harding, A. 1996, Poverty alleviation versus social insurance systems: A comparison of lifetime redistribution, Discussion Paper No. 12, NATSEM, University of Canberra, Canberra [Online], Available: http://www.natsem.canberra.edu.au/storage/dp12.pdf [2015, Sep 28].

Redmond, G., Whiteford, P. & Adamson, E. 2011, ‘Middle class welfare in Australia: How has the distribution of cash benefits changed since the 1980s?’, Australian Journal of Labour Economics, vol. 14, no. 2, pp. 81–102.

Whiteford, P. 2015a, ‘Tales of Robin Hood (part 1): Comparing welfare myths and realities in the United Kingdom and Australia’, Australian Review of Public Affairs [Online], Available: http://www.australianreview.net/digest/2015/09/whiteford.html [2015, Sep 28].

Whiteford, P. 2015b, ‘Tales of Robin Hood (part 2): Are the poor too expensive? Redistribution and the welfare state’, Australian Review of Public Affairs [Online], Available: http://www.australianreview.net/digest/2015/09/whiteford2.html [2015, Sep 28].

Peter Whiteford is a Professor in the Crawford School of Public Policy at The Australian National University, Canberra. He has researched child poverty, family assistance policies, welfare reform, and other aspects of social policy, particularly ways of supporting the balance between work and family life over several decades.

View other articles in this series by Peter Whiteford:

Tales of Robin Hood (Part 4): Social Security and Risk Over the Short and Medium Terms (Digest, Oct 2015)
Tales of Robin Hood (Part 2): Are the Poor Too Expensive? Redistribution and the Welfare State (Digest, Sep 2015)
Tales of Robin Hood (Part 1): Welfare Myths and Realities in the United Kingdom and Australia (Digest, Sep 2015)