The Future of Social Protection in a Unifying Europe[1]
A B Atkinson, Nuffield College, OXFORD
Introduction
1 Origins of the Modern Welfare State
2 The Economic Challenge to Social Protection
3 Ways Forward
Conclusions
Introduction
When I began
studying economics in the early 1960s, very few economists were interested in
the welfare state. This puzzled me, as it was my interest in social issues,
indeed a period spent working on the front line of the welfare state, that led
me to switch my university course from mathematics to economics. At that time, however, the subject of the welfare
state rarely appeared in the economics literature; articles on social security
were seldom to be found in the major academic journals; the study of poverty was
regarded as largely a matter for sociologists or political scientists. The
welfare state was about redistribution, and economists did not feel happy about
discussing distributional issues.
This
situation has changed greatly since then. Many economists are now studying the
Welfare State, writing about such matters as unemployment insurance, invalidity
benefit, and the funding of pensions.
Policy towards to income maintenance is seen to have implications for
the labour market; policy towards retirement savings is seen to have
consequences for the capital market. The
Welfare State is taken as a critical element in explaining differences in
economic performance.
Yet
I remain puzzled. I am puzzled first by
the strength of the views held by many economists regarding the welfare state,
and secondly by the extent to which the pendulum appears to have swung. An
institution that was previously ignored by economists is now held to be one of
the major determinants of macro-economic economic outcomes; an institution that
was felt to be largely redistributive is now being dismissed on efficiency
grounds, without regard to the functions it was set up to perform. Many economists and
policymakers agree with Buchanan that “the
“social model” that many Europeans hold as superior to the somewhat more
limited welfare states elsewhere is not economically viable for the
twenty-first century” (1998, p 14). It is with the challenge to the European
welfare state that I am principally concerned in this lecture. Strong policy conclusions
are being drawn, and we need to be clear about their basis. Is social
protection so costly that we have to retreat to a residual welfare state? Do
European governments retain any freedom of choice?
A natural reply to my puzzlement is that “times have changed”. The welfare state of the 1960s was different in type and scale, and the world was very different. Economists allocate their time according to the seriousness of the issue, and they have turned their spotlight on the welfare state because it is now a serious problem. My own views are rather different, but I fully agree that a historical perspective is essential. I start therefore in Section 1 with the birth of the welfare state, before turning to the issues of today’s Europe in Section 2, and exploring possible ways forward in Section 3.
1 Origins of the Modern Welfare State
When did the modern welfare state begin? In my view, the formative period for the modern welfare state was the latter half of the nineteenth century. This timing is of interest, since the later decades of the nineteenth century, like those of the twentieth, saw a marked phase of growth in the world economy. According to Williamson, the period 1870-1914 was one of “rapid globalization: capital and labour flowed across national frontiers in unprecedented quantities, and commodity trade boomed as transport costs declined sharply” (Williamson, 1996, p 277). Leaving aside the important difference with regard to labour mobility, the period had considerable similarity to the period from 1970 to today, with closer integration of commodity and capital markets, and reduced barriers to trade and capital movements.
It has however been argued that this earlier period of globalisation in the nineteenth century was one marked by the absence of the welfare state, and that modern social security systems only came later. On this view, social protection grew alongside economic protection: the retreat from an integrated world economy after 1918. In a recent paper, Tanzi argues that before that date “public spending for social protection was very small” (2004, p 5). He points to the “many activities undertaken by churches, civil societies, or extended families that provided considerable protection to individuals against various economic risks” (2004, p 5). In his analysis, the welfare state came into being in the autarchic period after the First World War, when protection and barriers to mobility were erected.
In terms of public expenditure, it is clearly the case that spending in the nineteenth century was small. The scale on Figure 1 (based on the data of Lindert, 1994) would not accommodate current levels of spending: the largest figure in 1910 is less than 2% of national income. But spending was building up over the period from 1880 to 1914. There may have been an acceleration after 1918, but it was the development of a trend already in place.
Spending levels are only one index. Also important are the dates of first introduction of social protection measures, and the shift in public opinion that preceded the legislation. Figure 2 (taken from Atkinson, 2002) shows the date of first introduction of different welfare state programmes in 24 OECD countries from 1880 to 1939. Unemployment benefit and family allowances were mostly introduced after 1918, but many of the other schemes predated 1914. As it was put rather colourfully by an American commentator in 1913, "From the frozen shores of Norway down to the sunny clime of Italy, from the furthest East and up to Spain, all Europe, whether Germanic, Saxon, Latin or Slav, follows the same path. Some countries have made greater advance than others, but none have remained outside of the procession, unless it be a few of the more insignificant principalities of the Balkan peninsula. The movement for social insurance is one of the most important world movements of our times." (Rubinow, 1913, page 26).
Growth of Modern Employment Relation
I believe therefore that it is wrong to associate the birth of the welfare state with the protectionist interwar period. The pressures for its introduction predated this protectionist period, and, even if spending was slow to grow, an important part of the welfare state was in place before 1918. In my view, the creation of the modern welfare state was, at least in part, a response to the development of the world economy. It was concern about the distributional impact of expanding trade and factor mobility that contributed to the setting in place of social security.
At the same time, internationalisation of the economy was not the
only, nor indeed the major, reason for the emergence of the welfare state at
the end of the nineteenth century. I would give priority to two other factors.
The first is the role of social insurance in the light of the development of
the modern employment relationship. This is most evident in the case of
unemployment. We may note that unemployment is a relatively recent
concept. According to the Oxford
English Dictionary, the word came into common use in Britain around
1895. Its recent origins are well
captured in the title of the book by Salais et al (1986) about unemployment in
France called L'invention du chômage, in which they see the emergence of
unemployment as associated with a particular form of labour market
situation. As described by Piore in his
review, "the modern concept of unemployment derives from one particular
employment relationship, that of the large, permanent manufacturing
establishment. Employment in such
institutions involves a radical separation in time and space from family and
leisure time activity ... When employment ties ... are severed, there is an
empty space in the worker's life which is sharply defined and that space is
what is meant by unemployment" (1987, p 1836). On this view, as the economy industrialised and became
urbanised, employment became all-or-nothing.
A depression could leave industrial workers totally without resources,
unable to fall back on home production or on the mixture of part-time
employment and self-employment that might be found in rural society. While the insecurity of wage employment was
not confined to the industrialised, urbanised ("modern") sector,
growth in this latter employment posed particular problems for the traditional
forms of income support. According to
Garraty, "methods of dealing with poverty and unemployment devised for
pre-industrial societies broke down under the new conditions" (Garraty,
1978, p 86).
At the same time that Salais was writing on the
invention of unemployment, Hannah produced his study of the development of
occupational pensions entitled Inventing Retirement (1986). The parallel extends beyond the title. Hannah emphasises that "for a proper
understanding of this largely new phenomenon, we have to look to the employment
relationship" (Hannah, 1986, p 21). Retirement, like unemployment, became
a discrete event, in contrast to the earlier more gradual dimunition of
productivity. In 1913, Rubinow described
how "the economic conditions of the wage-contract accentuate the economic
disability of old age. Under normal
physiological conditions, old age, unless preceded by a definite ailment,
should lead to a gradual failing of the productive powers. As the medieval independent worker became
old, he worked less and produced less, but he went on working as long as he
could produce something" (1913, p 304). In this respect there is a close
similarity with unemployment; there is however an important difference, which
is that along with compulsory retirement came also the provision of
occupational pensions. As has been
described by Hannah, "mandatory retirement, legitimised by more generous
pensioning, conveniently allowed bureaucracies to buy themselves out of the
expensive bias towards overpaying older workers implicit in their reward
systems. In the twentieth century, such
motives were commonly behind the expansion of existing pension schemes or the
founding of new ones" (Hannah, 1986, pp 135-6).
Social Objectives
The historical origins bring out the relation between the welfare state and the working of the labour market, and this in turn will help our understanding of the impact of today’s welfare state and of the possibilities for reform. But I do not believe that the economic factors are the only ones. The second element in my interpretation of the birth of the welfare state is its role in achieving social and political objectives. The introduction of social insurance was a response to the perceived shortcomings of the industrialised market economy. In the UK, it had become increasingly clear that the traditional means of support – family, charity, and poor law assistance – were failing to achieve a national minimum. Mayhew in his London Labour and the London Poor, written in 1856, had evoked public disquiet; and statistical underpinnings were later provided by the researches of Booth and Rowntree. The degree of public concern was evidenced by successive Royal Commissions on the Housing of the Working Classes in 1884, on the Aged Poor in 1895, and on the Poor Laws and the Relief of Distress in 1909. We talk today of a “crisis” in the welfare state, but, as it was put by Ashford, “the true `crisis’ of the welfare state ran from roughly 1850 to 1900 when liberal leaders in all countries were searching for new answers” (1986, page 34). The legitimacy of the free market economy was in question.
If legitimising the
market economy was part of the project, so too was the building of national
identity. It was no accident that the move
in the UK from the Poor Law to National Insurance was a move from local, parish
jurisdiction to national administration. It was no accident that the pioneer
countries in the field of social security were relatively recently unified such
as Bismarckian Germany or relatively recent states, such as New Zealand. The
new schemes were building solidarity not just between the employed and the
unemployed, or between workers and the elderly, but between the citizens of
relatively new political units. This observation has obvious relevance to the
EU of today.
Conclusions
The creation of the modern welfare state at the end of the nineteenth century was, at least in part, a response to the development of the world economy. At the same time, internationalisation of the economy was not the only reason for the emergence of the welfare state. The introduction of social insurance has to be seen in the light of the development of the modern employment relationship, with the associated risks of unemployment and retirement. Introduction of social insurance was a response to the perceived shortcomings of the industrialised market economy. It both served to legitimise the market economy and contributed to the building of national identity.
2 The Economic Challenge to Social Protection
One
hundred years later, this positive view of the welfare state is being
challenged. It is being seen as dysfunctional rather than functional. There are
vocal demands for social protection to be scaled back. I turn now to these
arguments. In my view, it is very important to be clear as to what exactly is
being asserted. Economists are drawing
strong policy conclusions, and we need to be sure how fully they are describing
the menu of choice open to governments. This is essential, since their
proposals threaten one of the most successful social innovations of recent
centuries. To quote Ashford again, “the transformation
of the nineteenth century liberal state, and its diverse manifestations
throughout Europe and North America, into the contemporary welfare state is
perhaps the most remarkable accomplishment of democratic governance” (1986, p
1). The
retrenchment of the welfare state is a threat to the incomes of many people.
While the existing social transfers may spill over to those not in need, and
there may be abuse, it is hard to deny that for many of the recipients these
transfers are essential to maintain a decent standard of living. We need
therefore to have a clear view as to how far the conclusions are inevitable and
how far we have a choice. The main theme of this section is – what choices are
there?
The Tax Cost Argument
A clear statement of the economic challenge is that of Tanzi: “high tax countries, and especially the so-called welfare states, will need to cope with their needs with progressively lower tax revenue. This scaling down of tax revenue would leave the countries with two options. First, to reduce the generosity of the benefits that individuals receive from the welfare states or to make these benefits less universal and thus better targeted. Second, to force citizens to buy at least parts of their insurances from the private sector” (Tanzi, 2004).
This is a valuable statement, because he is very clear about the nature of the cost. He is not talking about the excess burden of the welfare state, compared to private or other alternatives. He focuses on the total cost in terms of government financing. In the fanciful example that I have used elsewhere (Atkinson, 1999), in the unlikely event that a Martian offered to finance the welfare state, then Tanzi’s concern would disappear.
It is helpful to distinguish two versions of this argument. The first is a pure Laffer curve argument. It is posited that total revenue reaches a maximum at some tax rate less than 100%. The effect of globalisation is, according to Tanzi and others, to shift the Laffer curve downwards and move its peak to the left. If we were not already at the maximum of revenue, then we will soon be overtaken by events. As put by Tanzi, “tax competition among jurisdictions, ballooning electronic commerce, and increased mobility of the factors of production will likely cause significant falls in tax revenue in future years” (2002, p 116). This is a real set of economic constraints, on which economists are well placed professionally to advise. We should though distinguish the argument from a second argument, which is that the tax rate is reaching the politically acceptable maximum. This argument lies outside the sphere of competence of economists. Electoral constraints are not the same as economic constraints. Voters should be given the opportunity to make choices; their choices should not be pre-judged.
So it is on the Laffer argument that we should focus. It is evidently an important argument. At the same time, as Tanzi recognises, it applies to all government spending. It is a threat not just to the welfare state but to all activities financed by tax revenue. Of course, the welfare state is a large element, and its significance has been greatly increased by demographic and macro-economic developments. In simple terms, the required tax rate is
Replacement rate x Dependency rate
+ (Other government spending + Cost of tax expenditures + Cost of debt interest) / Total income
Reducing the
replacement rate can evidently have a large effect. With a dependency ratio of
a half (i.e. 1 pensioner or unemployed for each 2 workers), a 5% cut in the
replacement rate reduces the required tax rate by 2½ percentage points. At the
same time, it is clear that, while the
Welfare State may represent a particularly large item in the Budget, the tax
cost is the same Euro for Euro as if the spending were on roads or military
defence. By the same token, tax
expenditures have an identical impact to that of direct spending.
Allowances against income taxation play the same role as cash transfers, in
that both increase the tax rate necessary.
A higher tax exemption for the elderly reduces the overall tax receipts,
as do child tax allowances. This is important since the option of private
provision for retirement is often encouraged by tax concessions. The two –
welfare state spending and tax expenditures – do not usually appear in the same
equation, but they both raise the required tax rate.
But this is not all.
The equation also brings out the tax burden of debt finance. When people talk
about debt finance transferring the burden to future generations, they usually
have in mind the real burden in terms of goods and services. As Samuelson and
others pointed out many years ago, the resources used to fight the Second World
War were produced during that war; the burden in this sense could not be
transferred. The national debt is, in this sense, simply a re-arrangement of
claims – we owe it to ourselves. But the equation reminds us that there is a
potential financing burden. To pay these claims, we have to raise taxes,
and if we are at the taxable limit, then other government spending has to be
displaced. Conversely, if the state has acquired income-yielding assets, then
this reduces the need for taxation. It has always struck me as absurd that
governments sound alarms about the future burden of state pensions while
simultaneously divesting themselves of state assets. In the UK, the net worth
of the state has fallen very considerably in recent decades. This is the legacy
of privatisation. But the whole discussion of privatisation focused on the
control of assets and ignored the benefits of ownership. It would be quite
possible for the state to have retained a minority stake or to have used the
proceeds to reduce debt, rather than to finance current deficits.
Retreat to a Residual Welfare State?
We have seen that there are alternative ways of reducing the overall tax burden. Suppose, however, that we are forced to reduce social protection expenditure. Does this imply moving to a residual welfare state? As far as retirement is concerned, should we rely on private savings, supported by some form of minimum income guarantee? Such a guarantee would provide a safety net. No modern government is going to leave old people totally destitute. Given that savings are risky, provision has to be made for those who invested their life savings in Enron. The minimum guarantee would taper benefit, either with a 100% marginal tax rate, or by a rate lower than 100% but still high by the standards of top income tax rate: e.g. 75% or 80%. Short-term benefits such as those for the unemployed would also be provided subject to an income and assets test.
Such a residual welfare state is not a hypothetical situation. In the UK, successive governments have taken significant steps towards such a situation. In the 1980s, for example, unemployment insurance was largely dismantled, through restrictions on coverage and benefit entitlement. Individual unemployment insurance, providing earnings-replacement on the basis of contribution record, is now limited to a flat rate benefit, payable for a maximum of 6 months, for which many people do not qualify. It has been replaced by a means-tested Jobseekers’ Allowance. In the case of state pensions, we have again a largely flat rate system. The state earnings related pension has been scaled back, leaving the basic flat-rate state pension as a crucial element. This was in the past indexed to average earnings, so that pensioners shared in rising national prosperity. However, in the early 1980s the UK government abandoned the link to earnings and guaranteed only price indexation. The obvious outcome was that the basic pension began to fall steadily as a proportion of average earnings. It is not therefore surprising that Figure 3 shows that the UK was already an outlier in the 2000 figures for public pension spending in EU15 countries and that the total pension spending is forecast to fall over the next half century.
Such fiscal prudence might be thought matter for self-congratulation. In my view, however, this retreat to a residual welfare state is flawed – for two important reasons. The first is that experience with means-tested benefits in the UK has demonstrated that a significant proportion of those entitled to these benefits do not claim their entitlement. The means tests implicit in income targeting deter claimants. The latest estimates suggest that at least a quarter of those entitled to Pension Credit in the UK do not claim. Incomplete take-up has been a persistent feature of social assistance. Unless take-up can be seriously increased, the means-tested route cannot deliver income security.
The second flaw in the residual welfare state is that the overall tax rate is reduced at the expense of very marginal high tax rates on those receiving or potentially receiving means-tested benefits. The replacement of individual-based unemployment insurance by family-based jobseekers’ allowance means that the partners of the unemployed now face a strong disincentive to work. Under the insurance system, the unemployment of one member did not affect the work incentives of the other, but with joint income assessment there is little incentive to stay in work. This has led two earner couples to become no-earner couples. For pensioners, the implicit tax on savings and private pensions is a severe disincentive to make private provision: today, 40% of UK pensioners face tax and benefit withdrawal rates of over 50%, and this number is likely to rise.
The Distortion Argument
This brings me to incentives in general. I have concentrated on the tax cost argument, but we must carefully distinguish this from a second challenge to the welfare state: the distortion argument. On this line of reasoning, social protection is different from other tax outlays because it not only costs money but also distorts key economic decisions. The welfare state is not only too expensive but also is the cause of Europe’s economic malaise. To see that this argument is different, one has only to go back to my hypothetical Martian who offers to pay for the welfare state. On the distortion argument, we might well reject the offer on account of the economic damage caused.
This argument is sometimes made
along the lines that any interference with the market economy will distort
decisions; it will impose marginal tax rates different from zero; it will cause
us to depart from a level playing field. The trouble with this position is that
it assumes a world of perfectly competitive
and perfectly clearing markets, but in such a theoretical framework we find
none of the contingencies for which the welfare state exists. There is no
involuntary unemployment. There are assumed to be a full set of capital and
insurance markets. I can buy today personal care in the event that I need it in
x years time. However, in order to examine the economics of social
protection, we have to move away from an assumed world of perfectly competitive and perfectly clearing markets. We have
to allow for at least some of the contingencies for which the welfare state was
created.
To illustrate this point, I take the
non-competitive model used in the influential article
by Alesina and Perotti (1997) entitled “The Welfare State and Competitiveness”.
They study a two-country world where in the home country wages are bargained by
trade unions, generating unemployment in
that country. In the other (foreign) country, there is a competitive labour
market with full employment. Homogeneous workers have an alternative use for
their time, valued at R, which is assumed to be less than the full employment
wage. If there is unemployment benefit B, then the resulting reservation wage
is (R+B). In a right to manage model where firms determine employment, unions
bargain over wages as a mark-up (1+m) over this reservation wage. The gross
wage is increased by the payroll tax at rate t (assumed here to be the method
by which social transfers are financed). The resulting total labour cost is W
≡ (1+t)(1+m)(R+B). This means that, as shown in Figure 4, the welfare
state affects employment via both the fiscal burden (t) and the behavioural
effect (B).
In such a situation, it does indeed appear that, even if there were no tax cost, cutting the welfare state, reducing B, would indeed raise employment and improve economic performance. Two points should however be noted. The first is that we have to remember that unemployment insurance was introduced as a means of improving the functioning of the labour market. The creators of National Insurance were not unaware of the possible negative side, and designed the scheme to take these into account. We have therefore to take account of the institutional features of unemployment insurance. It is not just a benefit paid unconditionally to the unemployed. In Atkinson (1999), I formulate a model where in addition to the unionised sector there is a non-union sector in which labour is paid an efficiency wage to prevent ‘shirking’. In this model, it is possible that unemployment insurance benefits (as distinct from the taxes necessary to finance them) raise employment. A key reason for the difference in prediction is that this alternative (second) model takes account of the institutional conditions under which benefits are paid. On the one hand, the contribution conditions of unemployment insurance (usually ignored in economic models) mean that benefits enter the compensation package negotiated by unions and that there is differential coverage of the two sectors. On the other hand, the provisions, found in most unemployment insurance schemes, which disqualify people who have been dismissed for industrial misconduct, mean that the benefit does not enter the non-shirking condition that determines wages in the non-union sector.
The second point to
be stressed concerns the relation between labour market policy and retrenchment
of the welfare state, represented in Figure 4 by m and B respectively. There is
a view that these measures have to be used in conjunction. The OECD study by Elmeskov, Martin and Scarpetta
of "Key lessons for labour market reforms" concluded that
"comprehensiveness seems indeed to be a crucial feature of any successful
strategy to reduce unemployment because reforms in different areas can
reinforce each other's effects" (1998, p 223). The article by Coe and
Snower in IMF Staff Papers called "Policy Complementarities",
argued "an important group of labor market policies are complementary in
the sense that the effect of each policy is greater when implemented in
conjunction with the other policies than in isolation. [What is required] is deeper labor market
reforms across a broader range of complementary policies" (1997, p 1).
These authors are careful in their definition of complementarity, which
concerns the reinforcing effects of policies, but the statements have been
interpreted as saying that countries cannot choose to concentrate on one arm of
the strategy. They cannot, for example, succeed by making labour markets more
flexible but retaining generous social protection. It is not however
evident that governments have to both
make labour markets more flexible and cut back on social protection. The
model just described is one in which a reduction in either m or B
reduces unemployment. Put differently, the more flexible the labour market, the
less the cost in terms of benefit distortion.
Conclusion
In this section, I have suggested that the residual welfare state is not a promising route to follow. Fortunately, there is scope for choice. There is no inevitability about scaling back social protection. If our concern is with the total tax cost, then there are several other routes to tax reduction. If our concern is with disincentives, then there are ways in which social protection can be designed to offset the negative effects, and increased labour market flexibility offers an alternative to cutting social protection. This leads me to the final section, which deals with possible ways forward, with particular reference to the enlarged EU.
3. Ways Forward
My starting point in this section is that we are not going to abandon the social objectives of the welfare state. In the case of the EU, a commitment to social inclusion has been embodied in the Lisbon Agenda set in 2000. Attention has focused on the economic goals, but the social agenda has been taken forward in an impressive way.
Social Inclusion and the Lisbon Agenda
At the Nice Summit in
December 2000, it was agreed to advance social policy on the basis of an open
method of coordination, modelled on that already adopted for employment in the
“Luxembourg process”. The process of open co-ordination involves fixing
guidelines for the Union, establishing quantitative and qualitative indicators
to be applied in each Member State, and periodic monitoring in a process of
peer review. The open method of coordination has been controversial, not least
in the debate about the European Constitution, but my own view is that it is a
rather clever invention. It is tied to subsidiarity, under which social policy
is the responsibility of Member States. At the Nice Summit it was agreed that
each Member State should implement national action plans for combating poverty
and social exclusion. These National Action Plans on Social Inclusion are like
the National Action Plans on employment, and their purpose is to set out
national strategies and detailed policies.
This process recognises that action is the
responsibility of Member States under the principle of subsidiarity. But this
does not mean that the objectives of social policy are set by Member States. The objectives are set at an EU level.
Concretely, they are embodied in the set of social indicators agreed by
Heads of State and Government at the Laeken European Council in December 2001.
Member States – rather remarkably - have succeeded in defining an agreed set of
European social indicators. The primary indicators encompass financial poverty, income inequality, regional variation in
employment rates, long-term unemployment, joblessness, low educational
qualifications, life expectancy and poor health. In each case there are breakdowns, showing
for example poverty among men and women, or breakdowns by age groups.
The significance of
this EU agreement about the dimensions
along which social performance is to be measured may be seen if we return to
the issue of the limits to taxation. Fiscal competition is often evoked as a
source of such a limit. Where countries compete to attract capital, businesses
and skilled labour by lowering tax rates, there may be downward pressure on
social spending and hence reduced social performance. Insofar as the EU
monitors social performance, and social performance is worsened by cuts in
social protection, then this wing of the Lisbon Agenda provides a counterweight.
The EU as an institution serves to internalise some of the costs of
globalisation, and to moderate fiscal competition. It is important in this
regard that the social inclusion process forms part of the acquis communautaire
for new Member States.
The
statement of social performance criteria is a step forward, but what does this
mean in terms of policy choices?
Structural
Reform by Member States
I
begin with the actions of individual Member States. As I have indicated earlier
I feel that a retreat to a residual welfare state is a flawed prescription.
Instead, I believe that we should seek to reform social protection in a way
that works with, rather than against the grain of economic policy.
To
exemplify this point, I take one single, very salient policy area. It is widely
argued that Europe needs to raise the age of retirement and to end mandatory
retirement. Such steps have been justified on the tax cost argument; it is also
frequently argued that the pension systems and early retirement programmes tilt
the playing field against work. At the same time, it is important to be quite
clear about the nature of the argument. A level playing field means that the
decision whether to retire or to continue working for another year would be
unaffected by pension considerations or the tax system. Public policy would be
neutral. On the other hand, a tax cost perspective may lead us to want to
actively encourage people to stay at work. In this case, the pension system can
be biased in favour of postponing retirement by offering more than an
actuarially fair increase in the postponed pension to those who stay on at
work. There are two different policy stances: level playing field or reduce
tax cost.
From
the standpoint of the tax cost, the key variable is the ratio of retired to
workers. The proportion of the latter, however, depends on both the age of
retirement and the age of entry to the labour force. We talk a lot about the
latter but the former has also been a significant cause of rising dependency.
We have lost workers at both ends of the age spectrum – see Figure 5 (source:
European Commission, 2004a, p 31). Of course, there are good reasons why people
spend longer in education, but it is hard to believe that this is the only
cause. People delay their studies; they take gap years; recruitment procedures
are lengthy. There is a great deal that countries could do to raise the
participation rate of those aged under 30.
There are big differences across countries. From the standpoint of
social protection, these considerations suggest that we should focus on the
length of working life and not on the retirement age. A manual worker who
started at 15 has worked for 45 years when he reaches the age of 60; the person
starting at 25 has to continue to 70 to match that period of contributions. The
pension system should be constructed on the basis of the length of working
life.
EU-level
action
EU policy with regard to employment and pensions is pursued through the open method of coordination. An outside observer may well ask what reason there is to expect this to have any effect? Why should Member States agree voluntarily to be bound by a policy objective that departs from their own priorities? The social agenda is either irrelevant or powerless. However, one can imagine reasons why the EU Member States have adopted a set of common indicators to assess performance and contemplate setting EU targets, where these will cause them to behave differently. An important element is commitment. Governments, recognizing their limited tenure, are keen to commit their successors; governments, recognizing their own frailties, are disposed to committing themselves not to deviate in the face of changed circumstances. Both Madame de Pompadour (“après nous le déluge”) and Ulysses (“bound to the mast”) come into play. Re-writing the Lisbon agreement would require the agreement of a substantial majority of EU Heads of State and Government. There are also issues of credibility. In Atkinson (1996), I have argued that, in electoral competition, it may be in the interests of the more conservative party to commit itself to a poverty target, since this will increase the credibility of a promise not to move too far away from policies of redistribution.
Ex ante there are therefore some reasons to explain why countries have entered into a binding policy framework. Nonetheless, the constraints are not tight; and we have seen with the Stability and Growth Pact that much more formal rules can be broken. Peer pressure on its own may be insufficient to ensure that Member States individually set in place policies that will reduce the EU poverty rate or reduce other dimensions of social exclusion. It may therefore be the case that we have to conclude that “the objectives of the proposed action cannot be sufficiently achieved by the member states”, to quote from the Maastricht Treaty, and therefore that EU action is justified.
What form, however, should this action take? In keeping with the spirit of subsidiarity, it seems reasonable to consider the least interventionist form of coordination, which is to
- Set only a minimum standard.
- For this standard to be defined relative to
the circumstances of the Member State.
- For Member States to have, within these
constraints, the maximum possible latitude in the design of the policy.
To be more concrete, let
us address the issue of financial poverty among families with children. Child
poverty is receiving increasing attention. As is noted in the 2004 Joint Report
on Social Inclusion “in most countries children experience levels of income
poverty that are higher than those among adults. Material deprivation among
children must be matter of serious concern, as it is generally recognised to
affect their development and future opportunities.” (European Commission, 2004,
pp 19 and 21). The EU could address this
issue through setting a minimum income for children, defined as a percentage of
the Member State median equivalised income for each child (and possibly
age-related). Implementation would be left to Member States, who could employ
different instruments. The minimum could be provided via child benefit, via tax
credits, via benefits in kind, or via employer-mandated benefits.
Such a policy would be
an investment in the future of Europe’s children; it would also be an
investment in the future of Europe. In Section 1, I suggested that the origins
of the modern welfare state were to be found, at least in part, in the ambition
to build and unify new nations. The same applies today at the level of the EU.
We are seeking to create a new supra-national polity. The identity of this
polity has been established as a free trade area, a common market, and a common
currency, but the social identity has yet to be firmly constructed. The
creation of a common policy of guaranteeing a minimum income for children could
be a significant building block.
Conclusions
My conclusions may be summarised as follows:
Looking Back, we have to remember the origins of the welfare state and the functions it was established to perform. It grew out of the needs of the modern employment relationship, and the invention of retirement and unemployment. It was created to meet the concerns about the social legitimacy of the liberal market economy.
Today, there are important economic challenges to the welfare state, but there are choices that we can make. There is nothing inevitable about a retreat to a residual welfare state.
Looking Forward, in a unifying Europe, the EU social agenda is a restraint on fiscal competition within the EU, now covering the new Member States. The welfare state needs to be reformed at the national level, and I have given the example of lengthening the working life, but action is also necessary at the EU-level. The introduction of a EU minimum income for children would address growing concerns about child poverty, and would be a powerful way of forging the social identity of the EU.
Atkinson, A B, 2002, “Globalisation
and the European Welfare State at the Opening and Closing of the Twentieth
Century”, in H Kierzkowski, editor, Europe and Globalization, Palgrave
Macmillan, London.
Buchanan, J M, 1998, “The Fiscal Crises
in Welfare Democracies with Some Implications for Public Investment” in H
Shibata and T Ihori, editors, The Welfare State, Public Investment, and
Growth, Springer Verlag, Berlin.
Coe, D T and Snower, D J, 1997,
"Policy complementarities: The case for fundamental labour market
reform", IMF Staff Papers, vol 44: 1-35.
Elmeskov, J, Martin, J P, and
Scarpetta, S, 1998, "Key lessons for labour market reforms: evidence from
OECD countries' experiences", Swedish Economic Policy Review, vol
5: 207-252.
European Commission, 2004, Joint
Report on Social Inclusion, European Commission, Brussels.
European Commission, 2004a, Report
of the High-Level Group on the future of social policy in an enlarged European
Union, European Commission, Brussels.
Garraty, J A, 1978, Unemployment in
History, Harper, New York.
Hannah, L, 1986, Inventing
Retirement, Cambridge University Press, Cambridge.
Lindert, P H, 1994, “The Rise of
Social Spending, 1880-1930”, Explorations in Economic History, vol 31:
1-37.
Piore, M J, 1987, “Historical
perspective and the interpretation of unemployment”, Journal of Economic
Literature, vol 25: 1834-1850.
Rubinow, I M, 1913, Social Insurance,
Williams and Norgate, New York.
Salais, R, Baverez, N and Reynaud,B,
1986, L’Invention du Chômage, Presses Universitaires de France, Paris.
Tanzi, V, 2002, “Globalization and the
Future of Social Protection”, Scottish Journal of Political Economy, vol
49.
Tanzi, V, 2004, “Globalization, Social
Protection and Public Finance”, paper prepared for United Nations Development
Program.
Williamson, J G, 1996, “Globalization, Convergence, and History”, Journal of Economic History, vol 56: 277-306.
