The Post-war Welfare State

Revised version of Address in the University of Auckland Winter Lecture Series, 20 June 1995, published in Social Policy Journal of New Zealand, Issue 7, December 1996, p. 17-28.[1]

The modern welfare state developed in most rich countries in the post-war era. It was a response to three major trauma. First, there was the interwar depression, in which the brutality of unconstrained capitalism generated a response of a kinder, gentler way of organizing society. Second the war itself led to upheaval in many European countries. The welfare state was seen as a means of integrating the people back into the nations struggling with the chaos of their recent past, in order to generate a degree of social cohesiveness. Third, but by no means least, industrial society was sweeping away the old forms of community provision.

It is now two generations since the war, three since the depression, and the welfare state they generated is under considerable pressure to change. This article looks at the pressures, arguing that many of the contentions favouring reform are spurious. On the other hand, there is no doubt that some of the assumptions which implicitly underpinned the first fifty years of the post-war welfare state no longer hold. Identifying them, undermines the politics of nostalgia – of the belief we can go back to the old approach – which is probably the surest way to undermine the viability welfare state.

The Welfare State in Industrial Society

That welfare states replaced older ways of social protection is perhaps harder to understand in New Zealand because, excepting the Maori, there were no existing institutions, however antiquated, for the settlers in a new country. That is why New Zealand was early in the welfare state development, for there was no older social structure to fall back upon. New Zealand could not use the parish based provision of Britain, because it had no suitable parishes, while the commitment to a secular state – a state without an established church – meant that parishes could not be artificially created. In the nineteenth century European New Zealand even lacked sufficient of that fundamental unit of victorian virtue, the family. Most of the old and indigent at the end of the century had no children or, if they did, their children did not live near them. Thus New Zealand had to create means of community support in a system outside the family or the locality.

Although New Zealand was among the earliest of welfare states, elsewhere industrial society was undermining the traditional systems of the more established societies. They also began to develop state administered systems of support. This development reached a new level in the post-war era, as the turmoil of the 1930s and the 1940s exacerbated the long term trends towards social incoherence in industrial society.

Yet by the early 1990s there is a sense in which the welfare state is in retreat, or at least under severe pressure to modify. This is not just a statement about New Zealand. It applies to most. New Zealand’s may be under greater pressure because its economy has stagnated over the last decade while OECD GDP has grown on average around 20 percent. Even so their welfare states are under pressure too.

It would be easy to explain this pressure as a consequence of the transition from industrial to post-industrial society, but that is not particularly helpful because it does not explain which characteristics of one form of economic organization were once relevant but are no longer. With hindsight it is easy to describe how industrial society changed social conditions and old means of support, but the fundamental changes in social and structure between industrial and post-industrial society are less clear. Consequently there is little agreement as to what is the appropriate reform of the welfare state.

Some advocates of the reform of the welfare state tend to want to go back to the social structures of pre-industrial society, or have no account of social structure in their theories. An example of the first are those who want to make the idealized middle class Victorian family, perhaps with a Christian basis, as the foundation of our social policies, without any real consideration as to whether such families were as pervasive in the past as implicitly assume, whether those families functioned as ideally as is assumed, and why that family form has become less common.

On the other hand the radical right proposals for reforms, which attempt to look forward, barely involve any social structure at all. Their account – from a limited neo-classical economics perspective – is society as a collection of isolated individuals able to function in a social – and a moral – vacuum in which each looks after himself (or herself, on rare occasions when gender issues are acknowledged). In the New Right account of the world the affluence of post-industrial means there is no longer a widespread need for collective provision for health care, retirement, and education. Ironically the affluence which their approach assumes, is the result of compulsory collective structures which were central to industrial society, especially in its latter stages of greatest affluence.

There are other contradictions. There is not much in common between the past looking right of fundamentalist Christians and the radical right of structure-less libertarian secularism. This tension is very evident in the United States’ Republican Party. While its two wings may agree the present order of things is wrong, their solutions pull in quite different directions. The fundamentalists require the centralized imposition of a particular social structure, which they summarize in the expression “Christian family values”, while the libertarians want to withdraw the state from imposing any structures at all.

Yet as much as we may criticize these proposed reformers, there is a core of truth in their argument that there are some deep problems in the modern welfare state which are not being addressed by those who want to return to the institutions of a generation ago. Some defenders of the welfare state are even more out of touch than are its critics.

The perspective of this paper is somewhat different. The welfare state is seen as a response to deteriorating social coherence in industrial society, especially in regard to those aspects concerned with economic protection. There is no new coherence in any post-industrial society. In some respects there is less, for affluence has added to the heterogeneity of society, which is both an advantage insofar as it liberates personal differences, and a disadvantage insofar as it reduces social cohesion. The human demand for social coherence is such that we will continue to seek means of expressing it, probably in a modified welfare state. The evidence from the surveys is that the vast majority of New Zealanders desire a welfare state in some form, and are deeply bitter over the changes made in the last decade.[2]

Indeed one of the effects of MMP is that the desires of the polity will be better transmitted into the governance of the nation, in contrast to the way those desires have been ignored over the last decade. The transformation will not happen over night, not only because many of our current politicians bring attitudes from the old FPP era, but also because every political system works only imperfectly. Nevertheless the likelihood is that over the next decade the polity will demand and the politicians will supply a more, rather than less, extensive welfare state – but it will be a different one.

Patterns of Convergence

But that broad prediction leaves unstated what kind of welfare state. There is a growing international literature of comparative studies.[3] What strikes the reader is that there is both an overall pattern of welfare state provision, and many variations in the pattern between countries. As Einar Overbye has persuasively argued there is for some convergence of outcome of provision for the elderly, although not of delivery mechanisms. He suggests that there will be a tendency towards a two tier public or compulsory system, with a bottom tier of a tax-financed (usually income-tested) minimum, plus a second tier of contribution-based/earmarked taxes-based (usually defined-) benefit scheme providing earnings related benefits.[4] Thus there will be a combination of a minimum floor for those with marginal labour market experience, plus a second tier top up to reflect the inequalities in labour market experience (although often there will be an ceiling on the top up). Outside the state system there will be further non-comprehensive private and voluntary tiers (in which in Australasia home ownership will be important).

New Zealand is moving down this path for retirement provision, with the first tier firmly in place as New Zealand Superannuation. It has been more timid about the second tier, with the Todd report on retirement incomes arguing for voluntary provision in the second tier.[5] However Australia has moved towards a compulsory contribution based second tier, and one will not be surprised if New Zealand follows within the decade. If so it will only be catching up the pattern in most Western countries.

Two terms which may be unfamiliar: “contribution-defined” and “benefit-defined” may be used to describe the convergence. New Zealanders are very conversant with benefit-defined income maintenance schemes for the elderly for that is the basis of today’s New Zealand Superannuation. For a long time there has been a guaranteed retirement income related to the overall incomes of the community.

The notion of a contribution-defined scheme is less practically familiar. In it the amount the individual contributes is set, the contribution is invested, and the pension the individual receives after retirement is determined by the return on the investment, so there is no guaranteed benefit level. An example is the sort of scheme of which Roger Douglas is a proponent.[6] Its only certainty is the size of the contribution. The future benefit depends future returns on investment, rates of inflation, and longevity. Whatever the forecasts, the outcome may well be different. Contribution-defined schemes leave one’s income a hostage to events outside the individual’s control.

The experience of post-war welfare states has been to move away from income maintenance based solely on contribution-defined schemes to ones which are benefit-defined or where there is a mixture of the two.

Risk Management and Risk Shifting

Underlying the difference between benefit-defined and contribution-defined schemes is a fundamental question of welfare states – the allocation of risk between the individual and the community. In a contribution-defined retirement scheme any variations from the assumptions, for better or worse, end up as variations of the individual’s retirement income, for better or worse. Thus future uncertainty is the individual’s responsibility. In the case of the benefit-defined system the consequences of any variations in the assumptions of the funding underlying the scheme (be they explicit or implicit), are borne by the state. Ultimately the state funds by levies on its taxpayers, so the effect of a benefit-defined scheme is to shift the risk of a contribution-defined scheme from one generation, across all generations.

Historically this shifting of risk from the individual to the whole of society has been a central feature of the welfare state. Recently, especially in New Zealand but also elsewhere, there has been a tendency for the risk to be shifted back to the individual. Most obviously the stand-down period shifts the risk of unemployment to the unemployed, user charges shift the risk to the sick, tertiary student charges shift the risk to the student, and so on. Understanding risk shifting is central to explaining why the welfare state is in difficulties today.

Consider the drafting of the sheep, with the sheep rushing down a race, and someone switching a gate at its end to sort each sheep into one of two yards. Welfare states have a parallel element of this drafting in order to determine entitlements. Either you are elderly or not, you are sick or not, you are unemployed or not, you are a widow or not, and so on. This image is especially appropriate for New Zealand because more than most welfare states the benefit entitlements were based on membership of such categories.

One reason sheep drafting works is because the choice is relatively unambiguous – at least to the person making the decision. A second more subtle reason is the sheep’s inability to affect the decision. If the animals knew they were being selected for slaughter, days before the shrewder would go on a diet, appear thin to the drafter and go into the yard for further feeding rather than the one to the freezing works.

The technical term for altering circumstances to get such advantageous treatment is “moral hazard”. The term comes from insurance, as when someone learning they are sick takes out insurance to pay for the costs of the treatment, thus avoiding paying the costs themselves. Since moral hazard can be disastrous to the profitability of insurer, various measures are taken to minimize it.

Ambiguity and moral hazard are closely related. efficiently. neither was a major problem in the welfare state. When there is little unemployment it is not difficult to determine whether one can get a job; at earlier levels of medical technology it was not difficult to determine whether one needed treatment; a marriage certificate and a death certificate proved one was a widow.

Today, the selection process is much more complex. With high unemployment how to tell someone has assiduously looked for work and proved unemployable, or whether they are a shirker? Medical technology has moved from dealing with urgent limited – often life critical – events, to ones where there is much more judgement. (The expression “elective” surgery indicates that increasingly medical treatment has an element of choice.) And the widow is less common today than the solo mother where the father does not support the child.

Such ambiguities were not conceived as important by the creators of the post-war welfare state. Once they arise, there is the opportunity for the individual to adapt, or present, their circumstances to increase support from the state. The state can respond by introducing more and more complex rules, and administering them more and more assiduously. As it does so, the simplicity of the drafting yard gets undermined, and the cost of managing the system becomes increasingly expensive. It is easier to withdraw the entitlement, to the detriment of people with justified needs. Thus the risks of the social uncertainty get transferred from the state back to the individual, and the scope of the welfare state gets reduced.

The Welfare State and the Labour Market

Crucial to the effective functioning of the welfare state is the labour market. Until recently, the New Zealand welfare state was founded on full employment. New Zealand and Australia have traditionally been the two of the world’s welfare states which have delivered welfare most through the labour market.[7] When the fourth Labour government abandoned full employment as a primary economic objective, as it did formally in its 1987 election manifesto, it paved the way for the abandoning of the New Zealand welfare state as it had been traditionally organized.

What exactly constitutes full employment is not easy to describe. Our conventional account of the labour market has more relevance to an historical reality than the current one. It pictures employment as a male phenomenon of full time for an entire working life, ignoring the differences that arise from women’s work experience, from part-time and part-year work patterns, and the consequences of the constant demand for upskilling. The dynamics of the labour market experience is almost totally ignored, or simplified beyond recognition.

The full employment of New Zealand’s early post-war history is captured in the 1950s’ joke about the Minister of Labour knowing the name of every unemployed worker in the country – both of them. It had an element of truth, because often the total registered unemployed was often less than a hundred, and until the mid 1960s, usually less than two thousand. However there were unregistered unemployed in somewhat greater numbers – up to 3 percent of the labour force.[8] But unemployment was not such a threat these people needed to report. Shortly they would find a job.

Full employment underpinned the welfare state. It made drafting so much easier. Almost everyone who thought of applying for an unemployment benefit could be found a job. Households in need of additional income would have the main earner work more hours, or have the spouse go out to work.

Today such full employment no longer exists. Typically recorded unemployment rates are two and three times the level they were in the 1950s and 1960s. Even that ratio underestimates the qualitative magnitude of the change. In the first two decades after the war unemployment levels were typically in a comfort zone where the labour market was not stressful to the unemployed.

It belongs to another occasion to describe all the consequences of this changed state of the labour market – on firms, on households, on lives. Policy has not always responded. For example since in the past a person either had a job or could no get one, the treatment of any additional earnings of the unemployed could be clumsy. At not very high additional income, all the additional earnings were abated away. That situation still exists today, where beneficiaries still face effective marginal tax rates on some income of almost 100 percent. This reduces the incentive to an unemployed person to build up a work experience by part-time jobs, while easing oneself into a fuller work rate. Thus people on benefits get trapped into dependence. In effect the simple drafting gate used for full employment – either you were in or out of work – is too crude in today’s unemployment environment.

Under-full employment also means reduced output, and reduced government revenue. Yet the smaller economy is expected to carry a larger burden of dependency from the unemployed, including those who appear on sickness, invalids, domestic purposes, and superannuation benefits but are really unemployed because they would prefer to earn their living.

There is an irony here. The welfare state is presented as a mean of protection against adverse circumstances of the individual. Yet when the economy persistently experiences adverse circumstances, the welfare state is unable to function properly. Quality macroeconomic management is an integral part of the success of the welfare state. During the 1980s and the early 1990s New Zealand’s economic performance was poor, compared to other rich countries. That is why our welfare state is more tattered than its overseas’ equivalents.

The Cost of the Welfare State

This is not to argue the welfare state is unaffordable. This is a popular argument among those who want to abolish it, or reduce it in size to some minimalist role, like in their idealization of the United States. Frequently the minimalists cite data to suggest that the cost of the welfare state is rising, relative to total production and our capacity to pay. However this partly depends upon the exact measures that are used. Often their total cost of benefits includes income taxation paid on benefits, the surcharge on New Zealand Superannuation, while Family Support it treated as a benefit, rather than a negative income tax, all of which exaggerate the cost of the welfare state. A further complication is that many countries have public sector welfare programs which are administered via the private sector, and do not appear in the public sector statistics. New Zealand’s accident compensation system prior to ACC was such a scheme. When ACC was introduced in 1975, there was not the massive increase in public provided welfare, the statistics appeared to indicate. Finally an adjustment is necessary for the cyclical state of the economy. The costs of support are going to be higher during the economic trough, when there is less work available for the poor. Unless these adjustments are made comparing costs of welfare provision over time are misleading.

The relative cost of the welfare state is the number of dependents multiplied by the relative costs of the average resources provided to each dependent. The number of dependants are partly the result of factors over which the nation has little control, like sickness, invalidity, and demographics.

Even here the numbers need not be as frightening as they are sometimes presented. It is true that populations are getting older, with proportionally more aged 60 and over. In a quarter of a century New Zealand will have a proportion which is already typical in Western Europe.[9] Not only are these countries already coping with this population pressures, but the same proportions will be easer for us, because the age for potential retirement is rising.

The assumption of everyone being retired at the age of 60 is outmoded because work is not so physically demanding, and the elderly are able to acquire the new technological skills. It was absolutely right to plan the raising the age of entitlement to New Zealand state retirement benefit in steps to 65. That reduces the long run dependency ratio, and increases the potential workforce. One may not be surprised if the age of state entitlement is raised even higher – say to 67 – in the next decade or so.

Moreover part of the future rise in the age dependency ratio is because the proportion of the young is falling, which reduces the demands for public and private resources to support them, including increasing the number of adults in the paid workforce because they are not as involved in child caring.

However there are two components of the welfare state dependency ratio which are more problematic. One is that which arises from unemployment, which impacts directly by the cost of the unemployment benefit, and indirectly since the unemployed are not contributing to the government by taxation on earnings. Note however, that a rising dependency ratio means there are not enough workers, and so unemployment may fall.

A second source of rising dependency has been solo parents. This involves deep emotions, because of the sexual element, and because solo motherhood seems to threaten men. However, technically the rise of the cost of solo parenthood is the consequence of New Zealand underfunding two parent families with dependent children. The state support for a child in a one parent family is considerably higher than a two parent one, so family breakdown increases the burden to the state. If it had better supported two parent families, the cost of any change in the balance to one parent families would not have been so great. (Perhaps if New Zealand had funded two parent families more adequately, there would be less change in the balance, because financial stress leads to family break ups.)

More generally, and ironically, if the welfare state had adequately supported dependent children, it would not have the same worries about a rising dependency ratio, because the increases in funding the elderly would be offset by reductions in dependent children and non-working parents. Another irony is that children now become a greater burden on the state as they move from dependency on the family into tertiary education. There would not be the same tangle over tertiary student fees and support if had better supported them when they were younger. Some of the problems of rising welfare costs today, reflect demographic changes which expose the past underfunding of children.

Overall then, rises in the dependency ratio may be a tractable problem, unless society completely falls apart with permanent unemployment and family breakdown far in excess of what has been experienced in the postwar era thus far. If western economies cannot deal with unemployment, they venture into a territory so different from the post-war experience that it is hard to predict sensibly the long run outcome. The immediate worry, however, is the costs per dependent.

Costing Support

Defining what is meant by cost is complex. Simply looking at the direct cost to the public purse is not enough. Suppose all retirement income was funded by private investment rather than from general taxation. Paying that income would still be a burden on younger generations. Consider the thought experiment where all the retired died over night. Irrespective of whether they were state or privately funded, the material income of younger generations would rise the following morning.

Neither is the problem merely one of income adequacy, although deciding what is a reasonable minimum income is by no means simple. Once a satisfactory benefit level structure has been attained, the level rises with long run changes in average income, so the burden per dependent does not become more onerous over time. However there are other expenditures which are more pressing. The most notable are spending on education and health. For reasons of space the focus here is on health expenditure.

Once there seemed only a limited demand for medical resources. Subsequent advances in medical technology mean practically today there is an unlimited demand for medical care. How to constrain this demand to a practical level, consistent with our national productivity and our desires for other material goods and services?

One option would be to leave the decision to the private market. That suffers from at least two defects. First, reducing public intervention in medical markets can raise costs with little improvement in health, as the United States experience shows. Second, the private market still has to deal with the erratic incidence of illness. The usual solution, private medical insurance, is subject to the same problems of inefficiency, inequity, and expense because the funder is still a third party in the transaction between the doctor and patient just as occurs in a public system. So although private market delivery, and with it private responsibility, has a role in health care, the community is still faced with the same technical problems, whether it chooses a public or private delivery mechanism, or some balance between.

(Another problem is that because private systems favour adverse selection – picking the easy cases and clients – there has to remain a public system to cover the rest of the community, and it is very difficult to construct an efficient system in such circumstances. If moral hazard is the sheep avoiding drafting by their behaviour, adverse selection is the drafter selecting sheep in the drafter’s interest, even though society may have some other criteria for entitlement.)

It might be thought that a merit of private delivery is that it reduces tax levels, but even that is not obvious. Tax is a burden, but so is paying medical insurance, or paying user charges. It is not obvious that one is more weighty than another especially if, as the US experience suggests, in comparison to more socialized systems, private delivery can double the cost of care for no evident improvement in health outcomes.

More generally is it possible for an economy sustain tax levels a decent welfare state requires? For all the confidence expressed in the financial pages, there is little empirical evidence that high taxation is necessarily damaging to economic growth. (As it happens the greatest successes of the welfare state were in a period of much higher levels of taxation than today, and economic performance was superior too.) There are clearly more complex mechanisms than the crudities which are sometimes presented in the public debate.

The Future of the Welfare State

Despite the tendency to present the welfare state as though it was a decorative jewel on the body politic, to be sold off because of a diminished state of affluence, the welfare state is an integral part of the modern nation economy, softening the harshness of unconstrained capitalist state, and addressing the need to provide some coherence in an increasingly fragmented society. Indeed one might argue that the post-war era for the rich economies was the time of the welfare state, in all its various forms.

But are we still in an era of welfare states? Certainly there is a view that just as the economies of Eastern Europe have collapsed, we are seeing a crumbling of the traditional welfare state of the West, although the event is slower and less dramatic. New Zealanders could point to the undermining of the foundations of the system in the 1980s, followed by the destruction of key arrangements and institutions in the early 1990s. They might point to similar, but not as extensive, changes in welfare states overseas.

Discerning the tides of history is by no means easy for those in the boat. One is, however, struck by New Zealanders unwillingness to abandon the principles of the welfare state. The may reluctantly accept a need for change of the means, but there is no acceptance of an abandoning of the ends. Perhaps even more importantly, under MMP such popular preferences are more likely to be followed.

But what is a viable alternative? If those with the nostalgia for the welfare state of a couple of decades ago are not addressing the problems it faces today, the majority of the critics are not addressing the problems that it resolved. It is probably impossible to impose the sort of morality over a nation which those who want to return to “victorian values” promote. Equally the goal of social anomy the New Right wants to pursue seems implausible. Yet we cannot go back to where the nation was a generation ago. The best guess is that we will muddle on, radical solutions ruled out by the polity, and intelligent solutions ruled out by a lack of systematic analysis. At some stage in the future the welfare state will evolve into a new institution for regulating our social and economic life. But that seems likely to be generations away.

1. I am grateful to Einar Overbye and Susan St John for comments on an earlier version.
2. e.g. J. Vowles, & P. Aimer (1993) Voters’ Vengeance, Auckland University Press; and J. Vowles, P. Aimer, H. Catt, J. Lamare, & R. Miller (1995) Towards Consensus: The 1993 Election in New Zealand and the Transition to Proportional Representation, Auckland University Press.
3. e.g. F.G. Castles, (1985) The Working Class and Welfare, Allen & Unwin, Sydney, and (1993) “The Wage Earners’ State Revisited: Refurbishing the Established Model of Australian Social Protection, 1983-1993”, Australian Journal of Social Issues, Vol 29, no 2, p.120-145; G. Espinger-Andersen, (1990) Three Worlds of Capitalism, Polity, Oxford, and (ed) (1995) After the Golden Age: Welfare and Employment in Open Economies, UN, Copenhagen; D. Mitchell, (1992) Welfare States and Welfare Outcomes in the 1980s, Research School of Social Sciences, Australian National University; E. Overbye, (1995) “Convergence in Policy Outcomes: Social Security Schemes in Perspective”, Journal of Public Policy, Vol 15, No 2, p.147-174, and (1995) Different Countries on a Similar Path: Comparing Pension Policies in Scandinavia and Australia, Paper to the International Sociological Association Research Committee no 19 Seminar on Comparative Research on Welfare State Reforms, Pavia, Italy, September 1995.
4. Overbye op. cit.
5. Task Force on Private Provision for Retirement (1992) Private Provision for Retirement: The Options, Wellington. (Todd Report)
6. R.O. Douglas (1993) Unfinished Business Random House, Auckland.
7. Castles op. cit.
8. R. Braae & J.Gallacher (1983) “Labour Supply in New Zealand; 1976 to 1981” in B.H.Easton (Ed) Studies in the Labour Market, NZIER Research Paper No 29, Wellington, p.15-48.
9. T. Ashton, & S. St John (1993) Private Pensions in New Zealand: Can They Avert the Crisis? Institute of Policy Studies, Wellington.