Selfish Generations by David Thomson

New Zealannd Sociology, May 1992, Volume 7, No 1, p. 98-107.

Keywords Distributional Economics; Social Policy

Inter-generational income distribution is a difficult area, and it is a brave social scientist who would set out to offer a comprehensive account given our state of knowledge. David Thomson appears to have no such qualms.

The policy problem is simple enough. What is there to stop the oldest generation from exploiting younger generations by seizing its income? There is an obvious restraint on a younger generation doing it to their elders, because they then set an example to those younger to repeat the exercise. But the restraint does not exist for the oldest of all, since they will be dead when the younger generations get their turn. When economists explored such issues the model equilibria prove to be technically unstable, and usually some ad hoc restraint has to be added to prevent this inter-generational exploitation. These models are not specifically welfare state ones for they can involve pure capitalist economies. But driving all of them is the assumption that individuals behave selfishly.

Do they though? And even if they do, are there social and political restraint which impede the inter-generational exploitation ? Does the empirical evidence about income transfers over time support the account of selfish generations?

The transfers occur in at least three ways. First, individuals invest their savings and, at a later time, they realise a return on the investment. Second, there family transfers. Parents (and more generally, older generations – grandparents among others – chip in) invest in their children without – as a rule – getting much financial return (except perhaps support in old age). Additionally, the children may inherit family wealth. The details of these two general processes are complicated, but are central to the modern economy. They are largely ignored in this book (but not in this review).

The third major means is that there may be an active government transfers of resources between generations. Most evidently, the young receive education and health services, and family support (and benefit) from the state, while the old receive national superannuation and health services. It follows that those in the middle must be carrying the burden of these additions to the welfare of the young and old (although since the young tend to live with those in the middle, the first transfer is less clear).

It is this third state-generated inter-generational transfer which Thomson’s book focuses on, and even then, it looks only at the transfer to the old. Indeed, Thomson is obsessive about this transfer to the exclusion of all others. He says: ‘the core of all modern welfare states is what I shall call the implicit welfare contract between generations’. He never defends the statement and thus has little to say about vertical transfers between the rich and poor, and horizontal transfers within generations including health services, unemployment, sickness, accident, and domestic purposes benefits. Neither does he write much about the transfer to children.

Let me make it clear that I am not arguing about whether there are welfare state transfers to the elderly. Despite Thomson being described on the book cover as ‘a leading writer on social history…in New Zealand’, he is surprisingly ignorant of the work that has been done in New Zealand on inter-generational issues. If he had, he would have extensively quoted the work of Suzanne Snively (e.g. 1988), and the more recent fiscal incidence study by the Department of Statistics (1990), which demonstrate clearly that the biggest net inter-household transfer in a specific year between the identified social groups is to the elderly (but note that parents and children are in the same household). Quoting this work would have saved the book many pages, improved the level of analysis and made it that much more readable.

That the transfers to the elderly are large is hardly to be contested. But Thomson has a stronger thesis for he argues that the older generations are benefiting more from the welfare state than younger generations.

‘… the big winners … have been … those born between about 1920 and 1945. Throughout their lives they will make contributions which cover only a fraction of the benefits. For their successors the reverse is true.’ (p.3)

Much of the book is a rambling attempt to justify this accusation, based on anecdotal rather than systematic argument. Virtually any thesis about income distribution can be supported by taking a couple of special cases and comparing them. The problem is to provide a comprehensive account. I shall resist the temptation to review anecdote by anecdote, and instead concentrate on the book’s one attempt to argue the thesis coherently. The points made usually apply to the anecdotes as well.

Chapter 5 contrasts the experience of “the Earlys, a ‘typical (sic) couple’ born around 1930, and the Lates, a ‘typical couple’ born 25 years later”. Note first that the Earlys are about 60 at the time the book was written, with another 20 odd years of life expectation, and the Lates are 35 with another 40 odd years left. Thomson’s analysis requires forecasting into a distant future. I would not trust an economist to do that with the required precision the analysis requires.

In fact, Thomson does not appear to be any great shakes at forecasting. Referring to the Labour proposal to lift, by steps, the age of entitlement for the pension to 65, he writes: ‘the choice of the year 2006 came as no surprise -it means that those born before 1945 are to be protected, those born after are not’. I take it that the writer was surprised by the decision announced in the 1991 budget (even before the book was out) to scale the year back to 1992. (The term ‘protected’ is insidious -presumably he means ‘supported between 60 and 65’).

The next problem the analyst faces is to offer a systematic measure of the phenomenon he is considering. Thomson rightly wants to discount the effects of inflation, but did he really mean to eliminate economic growth? In doing so, he ignores that the Lates are typically on a standard of living about 40 percent higher than the Earlys at the same stage in the life cycle. Could not the phenomenon that Thomson claims to exist be merely the Earlys sharing some of this growth?

In terms of the inflation and growth adjusted measure used in the book, Thomson is probably postulating that the Earlys obtain a greater share of GDP (or some such aggregate) than the Lates. However, the indicator used is the median gross earnings for men aged 35-45. Unfortunately, this unit of account is contaminated by the changing income levels of women. Women’s earnings are rising, which means that the Lates are going to have a higher income (in these standards units) compared to the Earlys, which distorts the book’s argument in favour of the Earlys having more resources relative to the Lates.

The study uses a quaint notion of a ‘standard family’, which assumes the couple get married, have children and live together until death do them part. Perhaps the analysis is intended to imply this is as likely for the Earlys as for the Lates. In any case, the assumption is not only unrealistic, but it ignores the fact that taxes are being used from the Earlys and, even more so, from the Lates to support solo parents. (The unemployed also do not appear in this book’s scheme of things.)

The choice of a median income. is another difficulty, since it does not allow changes in the distribution within cohorts. Vertical changes could have been neutralised by using the mean income. A further complication is that the figures used are median male incomes. As well as exacerbating the problem of women re-entering the workforce (Thomson may well exaggerate the financial contribution of the Early women), it obscures important changes going on in youth (the Lates are likely to stay in education longer), and at retirement (where market participation has been reducing probably for the Earlys, although we can be less sure of the pattern for the Lates in 2020).

Unfortunately Thomson does not give his calculations, so I cannot check or adjust them. (A not unimportant consideration given the changes in Census definitions over time, especially following the adding in of social security payments in recent years).

But there is a summary table, which I reproduce here in a slightly condensed form. Note that the income data for each couple is not given. The units are those of the median income for 35 to 44 year males. Suppose that in one year, the median is $10,000, and the spending (or tax or whatever) is $4,000 in that year. That comes to .4 in the standard units.

Table Thomson’s Estimates of ‘Life Time Balance Sheets (sic) for the Earlys and Lates.’

Contributions: Earlys, Lates
Total lifetime income tax contributions (‘more realistic estimate’): 6-7, 15
Total lifetime ‘other taxes’: 14; 14
Total possible life time contributions: 20-21; 29

Family benefit: 1.5, 0.5
Old age pension: 12.6; 8.7
Education: 3.8; 2.5
Housing: 1.1; 0.5
Health: 3, 3
General Government Services: 15, 12
Total Benefits: 37, 27

Source: Table 5.1, p.166.

The table postulates that the Earlys pay less tax than the Lates. That is probably correct, because the Late’s relative income is higher (the woman works more), and because income tax rates have been rising. Whether the table’s overall figures are relatively correct, I cannot say, but almost certainly the Lates paid more ‘other taxes’ than the Earlys because their total relative income was higher, and indirect tax levels have also been rising (and are unlikely to fall through the next 40 years).

The benefit side is more problematic. First, it claims that the Earlys received more family benefit than the Lates, which is surprising since the family benefit was not introduced in 1946 after the Earlys turned 16 and became ineligible for it. (There was a minuscule family allowance from 1927). What the figures appear to be saying is that the Earlys received more family benefit for their children than the Lates did. The family benefit was relatively more valuable in the earlier years, and the Earlys would have had more children than the Lates.

This raises (yet another fatal) problem for the analysis. The Lates are (within a year or so) the children of the Earlys. I shall come back to the wider issue of how children should be incorporated in the analysis, but at this stage, note that a benefit provided by the state for the Lates is attributed to the Earlys. The same thing happens with education. The state expenditure on the Lates appears in the Early column. Probably, the same thing happens for part of health spending. An added complication is that education and health have been rising proportions of GDP spending, so the Lates would have received more per capita than the Earlys. (The Earlys would have been lucky to have a tertiary education, the Lates would have been unlucky not to.)

The ‘old age pension’, introduced in 1898 and superseded in 1939, is a rather strange name for the current retirement provision. Thomson has had to make a series of assumptions about what the Earlys and Lates will receive. He thinks that the former will receive about a third more than the latter. He may be right. As I have said economists have no great claims to be reliable forecasters. But there are some problems. The median income assumption is especially dangerous, because there has been (and is likely to be) considerable income redistribution within the age cohort. (Consider the universal entitlement from 60 in 1976, replaced by a super surcharge in 1984, to be raised further in 1992). Thomson also assumes similar longevity of the two generations. Underpinning this is a question of retirement policy with which I deal below.

The social security benefits exclude unemployment, sickness and domestic purposes, all of which would be more beneficial to the Lates than the Earlys. (Note that there is now an early retirement benefit, which while not as generous as national superannuation, mitigates the raising of the difference between Earlys and Lates by the retirement age from 60 to 65.)

Where Thomson gets his general government services figures from is unclear . I suspect he is projecting a relative reduction in government services over the next two decades, but wonder whether he adjusted the tax side as well.

Housing is a muddle in this book. Again it is dealt with below.

Overall, Thomson argues that the Earlys put less into society than they take out, and the Lates put in more. Unfortunately, the data is built upon many peculiar and unreliable assumptions while the forecasts crucial to the analysis are not convincing. A footnote to the table says: ‘the procedure has been conservative, seeking to play down the contrasts between the experiences of the two couples’. In virtually every case -children, benefits, government spending, women working, longevity, capacity to work -the ‘conservative play down’ favours Thomson’s hypothesis of the Earlys doing better than the Lates.

The conclusion may be true, within the narrow frame which Thomson uses, although that is unproven. However, there are some wider issues which also need addressing, even if the thesis were true.

First, can we ignore the inter-generational (but intra-household) transfer between parents and children? The Earlys might reasonably argue that they raised more children than the Lates, so the study overestimates their relative standard of living. Moreover, the investment they put into the kids is returned in part in a better retirement provision. They could not get to university, but they paid for the next generation to do so, and their retirement provision partly reflects the additional economic capacity of the economy as a result of their sacrifice.

A study purporting to cover the inter-generational transfers of the welfare state would contribute little if it ignores the parent-child transfer, and the opposite and hence the offset -of the transfer from the worker to the retired.

Second, there is the problem of what exactly the state retirement provision is for . Thomson argues the welfare state is primarily a cohort social insurance scheme. That is not obvious.

Consider my neighbour, George, who is about a decade older than the Earlys, with myself, a decade older than the Lates. The Very Early left school at 14 and worked to 59 before being made redundant. (He lived on the redundancy pay until he turned 60). George is a good bloke but, frankly, his occupational skills had become redundant before he did. I did not finish university until I was 23, so I shall be 68 before I have done my 45 years. Moreover, it looks as if my occupational skills ( which appear to be predicting doom from incompetent economic management) will still be valuable at retirement. Should I automatically be entitled to a state provided retirement benefit from 60 or even 65? Raising the age of universal entitlement may reflect changing patterns of occupational preparation, skills, ability to work and longevity.

This anecdote shows the retirement age is not set in stone. Nor is it obvious that if George gets more retirement benefits than me, he is better off. It is possible that from the perspective of the university academic, where it is said that retirement begins on the day you get tenure, retirement is an attractive occupation. In my view, one should be able to contribute to society as long as one is capable, and so early retirement need not be beneficial. What we are unsure about is how many Earlys are being forced into an early retirement, which is merely a euphemism for ‘unemployment’.

That leads to the third point of the role of inflation and unemployment. They are more closely linked than may be apparent at first, for each destroys savings. Inflation does this by diminishing the value of fixed interest wealth. It has been especially cruel because income tax is levied on the nominal, rather than the real, return on investment. But unemployment also destroys savings as they are consumed to sustain life when earnings are insufficient. Historically, much of our state provision for retirement has been a response to the destruction of private provision by depression and speculation.

What the Earlys have experienced is the considerable destruction of their savings by inflation from about 1970 to the late 1980s. Now they are being destroyed by the forced unemployment of early retirement. Neither of these processes are addressed in any detail in the book.

The housing inflation nexus is a complex one which the book fails to understand. The estate duly evidence is that wealth plateaus in the early 30s, which seems inconsistent with the known observation that personal savings increase with age (Easton, 1983; NZPC, 1990). What seems to be happening is that the major form of savings is housing, which give significant returns to young adults from the capital gains (including from the fixed value of the mortgage). However, you can only own one owner-occupier housing, so that investment returns are lower later in life, and may well be negative in after tax real terms. Additional savings thus offset the loss of the easy return from housing.

The book simplifies this complexity to tile single aspect of trying to assess the cost of a house purchase (although it is unaware of affordability index studies). The calculations ignore the contribution of women to the purchase, and overlook the way that inflation impacts on the down payment, but wipes out the mortgage costs.

One investment area which the book does not touch upon is government debt. There is some truth in the view that one generation’s public borrowing is the next generation’s tax. The high borrowing in recent decades might well be argued to be pushing liabilities onto future generations. However, much of those liabilities have been reduced in real terms by inflation, once more illustrating the central role it plays in inter-generational transfers. The result of the inflation was that despite heavy public borrowing, the public debt to GDP ratio was relatively constant throughout the 1970s and only really took off (i.e. became less favourable to future generations) in 1984 (e.g. Dalziel and Lattimore, 1991:28).

Fourth, suppose the hypothesis is true (if unproven). If the Earlys can exploit the Lates, then in due time, the Lates can exploit the Very Lates, who in turn can exploit the Very Very Lates, and so on. This could have been built into the tabulation but was not. Moreover, if it were true, then perhaps it is no more than each generation reducing the inter-generational 40 percent inequality from economic growth. We would need a more subtle exposition than that offered in this book to argue that it was wrong or an injustice. (One could equally argue that it parallels the effect of private investment and retirement provision.)

This is a review, not a book, and space is at a premium. The one further point which needs to be impressed upon the reader is how little Thomson knows about New Zealand if his citation record is any indication. 1l1ere is a chapter entitled , A New Poor’ which does not cite a single piece of work on poverty in New Zealand. If it had, the study would not have had to ‘raise doubts about the relative poverty of the elderly’. That they tend to be better off than families has been known since the mid-1970s, and was a major driver of social policy in the 1970s (Easton, 1981). His statement that in New Zealand, ‘there is no ready answer to questions such as ‘how did old age pensioners fare compare with younger families?’ (p.41) is true, if one denies the vast literature which exists on such topics.

Thomson might defend himself by saying: ‘this interpretation is not bolstered by voluminous references at every turn: it must stand or fall as readers judge it to ‘ring true’ , (p.7). Perhaps, although the references seem assiduous enough of overseas sources, giving the ringing a pseudo-scientific plausibility .It is the New Zealand material that is omitted. To be fair to Thomson, as far as I can judge, he is quite indiscriminate, omitting both material which supports or contradicts his thesis and material which would have developed and added rigour to the chatty presentation.

This methodology of testing by ‘ringing true’ seems to me to be a dangerous one. A good number of books promoting racial prejudice and conspiracy theories rest on such insecure foundations. Not surprisingly, the response to this book has been strong public reactions from those who find it confirms or denies their prejudices. The underlying message of the world which has changed little -in terms of demography, women working, family patterns and general social change -but in which one group is now robbing another is likely to ring true to those who wish it were true. Many of those who would hotly deny the underlying thesis would nonetheless be sympathetic to the social assumptions on which it rests. Given its limited description of the role of the welfare state, it is not surprising to see the book being quoted favourably by the New Right (Gibbs, 1991).

This book by, we are told, a leading social historian thus sits uneasily in the category of the pseudo-academic which justifies unthinking prejudice. In doing so, it demeans the sterling work of those who have tried to look at these complex issues systematically, and adds heat to a public debate while pretending to shed light.

Selected Bibliography
Dalziel, P. and Lattimore, R. (1991) A Briefing on the New Zealand Economy: 1960-1990. Oxford University Press.:
Department of Statistics (1990) The Fiscal Impact on Income Distribution 1987/88, Wellington.
Easton, B.H. (1981) Pragmatism and Progress Social Security in the Seventies, University of Canterbury Press.
Easton, B.H. (1983) Income Distribution in New Zealand, NZIER Research Paper No.28, Wellington, New Zealand Institute of Economic Research.
Gibbs, A., 1991. ‘Does the Welfare State Have a Future?’, Economic Alert, 2(8),October
Snively, S.I. (1988) The Government Budget and Social Policy, Paper prepared for the Royal Commission on Social Policy, Wellington.