Evaluating Pension Reform

111 Evaluating Pension ReformNFT 2/2004 Introduction Recognition of the “demographic time bomb” is relatively new. With the exception of the Greenspan Commission in the USA at the start of the 1980s, until the end of the 1980s few economists or social policy analysts con- sidered demographic change. Interestingly, some of the earlier discussion of what such change might imply looked not at old people but at young people – how the passing through of the baby-boom cohort would mean a dra- matic reduction in the number of young peo- ple entering the labour market, and what the consequences of this might be for employers in sectors that had made extensive use of young workers (NEDO, 1998).1 The World Bank’s “Averting the Old-Age Crisis” publication was important as much for being a first contribution to a debate about what an increase in the share of the population that was old would mean as for the appropri- ateness of its analysis or its recommendations. It provoked a series of emulations from inter- national organisations such as the ILO, the OECD and the European Commission, each of which coloured its analysis and prescrip- tion with its own philosophy and understand- ing of the needs of its particular constituency. Evaluating Pension Reform by Bernard H Casey Bernard H Casey b.casey@lse.ac.uk Bernard H Casey is a research associate of The Pensions Institute, the Cass Business School, City University, London and a senior research fellow at London School of Economics. He was previously a senior economist at the OECD, and he has been writing about the employment of older workers and on retirement and retirement income systems for over 20 years. The intention of this article is not – pace Shakespeare – to praise the new Swedish pension system – but also it is not to bury it. Rather it is to lay out some of the basic premises on which retirement income systems are predicated. It tries to make explicit notions that are often implicit and sometimes not addressed at all. Discussion of the relative merits of “NDC-type” systems, or of pension reform more generally, cannot take place without this being done. It is necessary to consider not merely public pension systems, and as importantly, private pensions systems that interact with them, but also employment systems that provide work for older people and social service systems that provide care to those unable to help themselves. It is also necessary to make clear what objectives particular reforms are trying to achieve. Normally, these are multiple and, in so far as they are, there are ordered. The hierarchy in question need to be made clear as much as do the objectives themselves. The article argues that only if reform is sufficiently encompassing, and only if it is sufficiently transparent, will it be sustainable. 112 Evaluating Pension Reform Pensions policy and policy for older people What is interesting about the debate that has been provoked is how fragmented it has been. Much of the discussion has centred upon pensions, some of it has been about health and long term care and some of it has been about employment. The few attempts to integrate these themes have been presented under the umbrella of analyses of fiscal costs of societal ageing – and in constructing these, the Euro- pean Commission and the OECD took the lead (OECD, 2001). However, these attempts were highly restricted. The principal concern was government expenditure, of which pen- sion expenditure was the most easily identifi- able and least problematic to project. Much uncertainty surrounded projections of health and care expenditure, partly because costs were less easy to identify, partly because assumptions about future morbidity were fraught with difficulty, and partly because technological developments were highly un- certain. Employment entered into the analysis in a yet more subsidiary fashion. Employees were considered in terms of being contribu- tors and, subsequently beneficiaries of pen- sions. Pension systems were seen as being potentially more sustainable if fewer people retired early and more people retired late. If pension reform – the subject of this series of articles – is to be understood seriously, the first step must be to cease studying pensions by themselves. It has been shown that, were the incidence of early retirement to be reduced and the effective retirement age were to rise by one or two years, a not inconsiderable share of the “pensions crisis” might be allevi- ated (OECD, 2003). However, other than by proposing that early pension opportunities be cut back or made less generous, those who conduct such simulations give little thought to how this raising of the effective retirement age might be realised. In many countries, the workforce is going to grow more slowly than in the past, and in some, it is even going to shrink. Even this does not mean that there will be an increase in demand for older workers. Institutions, practices and attitudes toward the latter, amongst labour unions as much as amongst employers, have been developed over a time when early retirement provided a con- venient and socially acceptable way of down- sizing and of restructuring to enhance com- petitiveness and flexibility, and this led to the development of attitudes and expectations amongst employees themselves. None of these practices and attitudes will be easy to change in the medium term. Moreover, there might be no need for them to change. Europe is enlarg- ing. There are reserves of what are recognised as well qualified labour waiting on the borders of “old Europe”. For employers, their availa- bility might provide a solution to the shortage of labour as satisfactory as one that involved the reorganisation of work for older employees. Ageing societies do not produce only a greater share of the population that is, poten- tially, financially dependent, it also tends to produce a greater share of those who are physically dependent – the “very old”. How much physical dependency (and included here should also be “mental dependency” as gener- ated by senile dementia) will increase de- pends upon whether ageing goes hand-in- hand with an extension or compression of morbidity. Here the realm of uncertainty is large. However, if there is an increased need for care services, this has its own implica- tions. Care provision for the very old might be given informally, but, if it is, this is likely to have employment implications. The care giv- ers – who often fall into the same age group as older workers, early retirees or even slightly delayed retirees – are likely to require some kind of adaptation of their working schedules analogous to the changes that have been de- manded, and to some extent achieved, by younger workers who have responsibilities 113 Evaluating Pension Reform caring for children. This has its own work- organisation consequences. If such reorgani- sation is not forthcoming, the requirement of the pension system for more contributors in amongst one of the groups where labour force participation is currently relatively low is likely to be frustrated. Of course, the alternative would be to rely upon formal care provision. Yet if this is provided by the state, it has its own fiscal implications, offsetting some of the gains that the increase in activity rates is supposed to generate (Casey, 2003a). Were, instead, the costs of formal care to be met privately, this would have its own conse- quences. If costs were to be met by the other- wise informal care givers, this might discour- age them from going out to work, since the care costs thus incurred could dig heavily into their incomes. If the costs were to be met by those who receive the care, this would raise distributional questions – who would obtain what care and of what quality – and adequacy questions – would the sharp reduction in dis- posable income be tolerable. The discussion of who pays raises a further dimension that has to be taken into account in analysing pension reform. The preoccupation of many policy makers has been with reduc- ing the state’s involvement in providing pen- sions because this reduces both the share of public expenditure in GDP – with its implica- tions for interest rates and currency stability – and the “tax wedge” – with its negative impact on labour demand and labour supply. There are calls for a greater role for personal respon- sibility and for people to save to make provi- sions for old age. However, if privatisation merely takes the form of mandatory savings, it is unclear whether the tax wedge is reduced. Moreover, although contributions to manda- tory private insurance systems are not normal- ly taken into account when disposable income is being assessed, they need to be taken into account when pre- and post-reform well-being is measured. Equally, whatever protestations might be offered to the contrary, governments might find themselves under political, if not judicial, pressure to meet liabilities of any scheme that has mandatory membership and is subject to standards and rules of behaviour that the government has approved (Casey, 1998). The fact that pension policy does not exist in isolation from other areas of social policy – broadly defined – makes evaluation of reform a complex process. It is clearly insufficient to concentrate solely upon a single measure. Reform is only sustainable if it can be shown to be backed by appropriate, complementary policies, and if it can be shown that negative spill-overs have been foreseen and minimised. Whether reform is indeed “successful”, however, depends upon who is measuring the outcome. Those who do the measuring either tend to use their own yardsticks, or, when they use a common yardstick, they often fail to agree on what the individual unit of measure- ment means. This can be seen when the terms “security”, “equity” and “efficiency” – the traditional calibrators employed in debate – are analysed more closely. Understanding “security”, “equity”, “efficiency” and “solidarity” Discussions with academics, policy makers and practitioners from a large number of coun- tries have made clear in how many ways these terms can be used. The table below summa- rises the interpretations of the first three advanced at the most recent conference of ISSA’s European section.2 Here, there were participants from central and eastern Europe as well as from western Europe, functionaries of public schemes of “partitarian” schemes and of intergovernmental and international organisations, as well as academics and repre- sentatives of NGOs. The individuals offering definitions came from countries with widely differing pension systems, they had greater or 114 Evaluating Pension Reform lesser experience of pension reform and/or often deeply opposing views of how pension systems should be constructed and what their role was. Accordingly, the heterogeneity of responses, if illuminating, was scarcely sur- prising. The table does not show the frequency or intensity of any response merely whether a particular definition was advanced at all. Per- haps because the audience was European, the notion of “solidarity” was frequently advanced, but even then, more than one meaning was placed upon it.3 Thus, solidarity could be used in each of the following situations. Again, applying each type of solidarity has its own implications. Traditional solidarity tends to mean that children support their par- ents, whereas egalitarian solidarity means ei- ther that current workers support current reti- rees or that citizens support one another. But it could also be applied to describe a situation where current cohorts should respect later cohorts. Moreover, however much one is told one ought to be solidaristic, there is always room for debate about how solidaristic one should be. It is not only a question of whether one should support but at what level. Associ- ated with this is the question of the extent to which one should redistribute. A solidaristic system may contain a more or less substantial element of redistribution. Continental Euro- pean, public pension systems, which empha- sise proportionality, are per se less redistribu- tive than Bevridgian ones, but both would claim to be solidaristic. Given that there are such questions to an- swer, it follows that there is also much scope to discuss the trade-offs between the other Table 2: Types of solidarity Benefactor Beneficiary Exemplified by… Traditional adult child elderly parent family-based support Narrow egalitarian active workers retired workers statutory employees insurance Broad egalitarian younger generation older generation citizens insurance Mixed earlier/current cohort current/later cohort public NDC-type system Actuarial individual individual individual accounts Table 1: With respect to pensions, what was meant by…. Security Effectiveness Equity adequacy of benefit assured minimum level of benefit assured benefit level kept stable costs kept under control ability to pay “promised” benefits assured benefit doing as intended benefits well targeted well and economically administered vulnerable groups protected inequality reduced poverty alleviated solidarity maintained proportionality ensured actuarial neutrality assured 115 Evaluating Pension Reform objectives of security, efficiency and equity. Indeed, it is precisely because “solidarity” is not a neutral term that the prima facie unusual equation of “equity” and “actuarial neutrali- ty” could be made and the new term “actuarial solidarity” could be introduced.4 That this was done by some of the advocates of NDC- type systems shows that concepts are not immutable and that, in subtle ways, it is pos- sible to recast the debate. A concept that is implicit in the definitions of security, efficiency and equity is the notion of risk. A key objective of a pension scheme, and one to which almost all commentators would subscribe, is to assist people to mini- mise the “risks” associated with old age. How- ever, in discussion of pensions it is not unusu- al to find the term “risk” being used as s synonym for “uncertainty”. This is not cor- rect. For example, societal ageing is not a risk. Within the reliability of current population projections, it is a certainty. Individual ageing is not a certainty; it is an aspiration. However, whilst an individual can aspire to live long, and can plan for a period of retirement, he or has little information on how long he or she will be old. He or she is faced with a longevity risk.5 Only an efficient annuity market, which scarcely exists, or a public pension scheme would allow this risk to be managed. Individuals face risks of sickness and “risks” of maternity, care giving and studying. Pri- vate markets almost always fail to cover these risks, which explains why most countries make some provision for them in their social protec- tion systems. Moreover, some of these “risks” – child rearing and studying – are usually considered to be to the benefit of the collectiv- ity, and this, too, explains why most public pension systems provide some form of “cred- its” for time out of the labour market. Over and above these individual risks, there are risks that affect all people simultaneously. Unemployment is one of these, since it is not randomly experienced but, at any one time, experienced by more, or by fewer, people. The unemployment risk is uninsurable for this reason – not only because it is difficult for a private company to control morally hazard- ous behaviour – and thus it is covered by “social” insurance or its equivalent. Equally, inflation is experienced by all people simulta- neously, so it, too, is a risk that cannot be met, or cannot be met well, by private insurance contracts – index-linked annuities are availa- ble only at a very high cost. Public, pay-as- you-go pension schemes are better able to adjust benefits to cope with price rises. When assessing any one form of pension provision, account should be taken less of the extent to which it enables actors to manage risk and more of the extent to which it as a whole is vulnerable. By “vulnerable” is meant the extent to which its performance, relative to expectations, is damaged by events that have an impact either upon the individual or upon the collectivity. Vulnerability, thus, refers to exposure to risks and to changes in the envi- ronment. Whilst no pension system is invul- nerable, particular systems might be more vulnerable than others. For instance, they might leave the participant less exposed on some sides to factors they cannot control, or they might contain mechanisms whereby unantic- ipated outcomes can be mitigated and where- by the impact of collective risk is smoothed if not eliminated. However, systems that are less vulnerable in certain dimensions might be more vulnerable in others. Indeed, it is perfectly possible that the price of reducing exposure in one area increases exposure in another. Table 3 (next page) summarises the vulnerability of different systems. The analysis in this table is, however, at best indicative. With respect to certain risks, it indicates that different systems place this risk on different actors. With respect to certainties, it makes clear that all systems are vulnerable to societal ageing. There is an implication that it is better not to have all eggs in one basket – 116 Evaluating Pension Reform Table 3: Vulnerability, by type of pension system DB private DB public DC private NDC public Multi-pillar Societal ageing Exposed become too expensive to fund, so pensions cut Exposed PAYGO contribution rate becomes too high, so pensions cut Exposed more shares to be sold to fewer buyers, so pension capital falls Less exposed but pension levels fall for any given retirement age Exposed cannot escape demography Macro- economic slowdown Exposed Exposed Exposed Exposed Exposed Financial market risk Exposed borne, in first instance, by employer and share holders Unexposed except in extremis Exposed borne, in first instance, by employee/retiree Unexposed except in extremis Less exposed Moral hazard Exposed actuarial non- neutrality, “back- loading”, vesting Heavily exposed actuarial non- neutrality, especially with respect to early pensions and their equivalents Unexposed Unexposed Less exposed Political risk Exposed tax privileges can make them vulnerable Heavily exposed Exposed tax privileges can make them vulnerable Less exposed but can be reformed (e.g., Italy) Exposed Life-time/ life-course risks (including low income) and longevity risk Exposed seldom credits for time not worked, at best disability component Less exposed credits for time unemployed, studying, maternity, caring, usually a disability component, usually minimum pension (i.e. redistributive) Exposed disability separate, longevity only if annuity market perfect Less exposed credits for time unemployed, studying, maternity, caring, also depends if separate minimum pension operates Less exposed, but depends on mix 117 Evaluating Pension Reform and thus the “multi-pillar” approach appears to be preferable. However, exactly what the mix should be is not particularly clear. The World Bank, in some of its more thoughtful writings (for example, James, 1998), has been able to classify a large number of countries as having such a system – it could be applied to that of Canada, Switzerland, the UK and the USA but also Japan (with a state pension, extensive company pensions and extensive private/personal schemes), Denmark, Finland, the Netherlands and pre-reform, as well as post-reform, Sweden (with a state system and extensive, collectively-agreed industry or oc- cupational pensions). With respect to compa- ny provision, this might be mandatory, as in Switzerland, or voluntary, as in the UK. The same might apply with respect to collectively- agreed schemes. Moreover, the “public-pri- vate mix” varies. For example, the public component is much higher in Sweden than in Denmark. And the argument can be taken further. If it is a mixed portfolio that is to be recommended, a system that “invests” in land, labour and capital might be preferred to one that “invests” only in capital. Prima facie, a classic, public, defined-benefit system does the former and a private, funded, defined- contribution system does the latter. The question, then, becomes one of how the components of vulnerability are appraised. Which is more important, and which system best allows responses to be made when the system is subject to challenge? Reforming pension systems Pension reforms that involve substantial cuts in benefits or entitlements are scarcely likely to be popular; indeed they can excite enor- mous passion. One has only to observe the events in Italy and France, but also in Austria, in the past year to recognise this. This has led governments to tread the road of reform with care. The first major attempt to trim back the German pension system – announced on the day the Berlin Wall was breached in 1989, but referred to as the “1992 Reform” – was pre- ceded by the deliberations, over many years, of a special commission that had sought to involve all relevant interests (particularly “the social partners”) and had taken extensive sci- entific advice. The feeling was that only with consensus could acceptable and lasting re- form be made. However, this approach went out of fashion in Germany. By the mid-1990s, the German government reformed again, ef- fectively by fiat and in what was described as a most “un-German” fashion. Non-consensu- al reform, admittedly justified by a commis- sion of “experts”, has been perpetuated by the succeeding government.6 Acceptability of reform, even potentially painful reform, can be enhanced simply by giving adequate notice. Thus, when the Amer- ican government decided to rise the “normal” retirement age from 65 to 67, it phased the change so that it did not start until 20 years later and would not be completed until 2027. Acquiescence can also be achieved by hiding changes as changes in technical detail that are so complex and esoteric that few notice their implications. A good example of this was the UK where changes to the calculation formula of the complementary public pension (SERPs) in 1995 effectively cut its value in half by lowering the accumulation rate and redefin- ing the earnings to which this applied. Swedish social democratic tradition has meant that pension reform is the subject of “politics” rather than “social partnership”. Responsibility was in the hands of political parties and, having accepted the need for action, these sanctioned the basic principles on which the reform of the last decade was made. The principle of linking benefit levels more closely to contributions, to retirement age and to expected length of retirement was attractive. What is more, whether or not this was a conscious part of the strategy at the 118 Evaluating Pension Reform time, those who advanced the NDC solution were able to take advantage of what is now a widely recognised fact. It might be easier to overcome political blocks to retrenchment by engaging in “paradigmatic reform” – thor- oughly recasting the system and introducing new principles such as the express link be- tween contributions and benefits and benefits and longevity – than by proposing major “parametric reform” – changing parameters of the system such as accumulation rates, benefit levels or retirement ages.7 Paradigmatic reform that involved the es- tablishment of an NDC system is not peculiar to Sweden. The Dini reform in Italy had many common features but was conceived entirely independently – there was no “policy learn- ing”. However, despite suggestions made in both countries at the time, and whatever the other attractions the approach has, one of the things that an NDC system does not do is remove uncertainty. Even if the pension cal- culation formulae, or the basic pension pa- rameters, were set in stone, the extent to which any one individual or cohort can asses his, her or its pension rights ex ante is limited. The benefit that will be received depends upon developments, economic and demographic, collective and individual, that cannot be fore- seen (Disney, 1999). This conclusion is scarce- ly surprising. This is not to say that the uncer- tainties associated with an NDC system are “worse” than those associated with a tradi- tional, public, defined-benefit system, merely to say that they are different. What is false is to propose that adoption of such an approach abolishes uncertainty. Even in the Swedish case, this has been acknowledged. The “balancing mechanism”, the impact of which might be profound, was introduced as a component of the reformed system only at a late stage. It is a highly sophisticated device, and it does the job that was set for it. But it is a highly technical instrument. It was not publicly debated and, indeed, it is doubtful if it could have been debated, had it been presented as what it was. Indeed, if NDC-like systems have the politi- cal advantage that they “obfuscate … reality” (Williamson, 2004, p.54), the “balancing mech- anism” takes the process of obfuscation to new heights. Moreover, the contrast between what happened in Sweden and what happened in Italy is remarkable. In the latter country, it became clear that the initial reforms were insufficient to reduce expected future costs, or to reduce them sufficiently fast. Reform be- came a “never ending process” (Franco, 2002) in which successive governments sought to bring forward implementation dates, to in- crease the minimum number of contributions necessary to achieve a full pension, or to raise the minimum age at which that pension could be drawn. It is not the intention, here, to comment upon the appropriateness of the Italian NDC system, rather to point out that the public in Italy are deeply aware of the fact that more changes are being made, that these will impact upon pension age and pension level, and that some groups will be more affected than oth- ers. That there are general strikes, and that the government is resorting to somewhat non- conventional routes in taking the legislation through parliament, is symptomatic of the fact that major reform is a highly political issue. It is not to be expected that political actors in Sweden behave exactly as their Italian coun- terparts. However, if the implications of a technical mechanism such as the “balancing mechanism”, which has potentially as far- reaching consequences as some of the compo- nents of the bill currently before the Italian parliament, were appreciated, one might ex- pect some response. Proposals to reform pensions led to the fall of the Juppé government in France in 1997and contributed to the demise of the first Berlus- coni government in Italy in 1994. The unwill- ingness of governments to enter into debate, 119 Evaluating Pension Reform and to rely upon disguising reform as techni- cal adjustment, is understandable. And on this basis, it is also understandable why European governments might be interested in ceding some responsibility for decisions elsewhere. Although social protection is a national com- petence, in recent years there have been at- tempts, via the Open Method of Coordination (OMC) and the Basic Economic Policy Guide- lines, to give reform a European dimension. Mutual target-setting, benchmarking, and implicit, if not explicit, “naming and sham- ing” are ways to nudge governments of mem- ber states to take actions that they might otherwise have sought to avoid. Moreover, in taking these actions, they are able to hide behind the argument that “Brussels expects” or “Brussels requires”. There is nothing new in this. The establishment, at Maastricht in 1992, of “convergence criteria” that would smooth the way toward monetary union facil- itated governments taking, and “selling”, pos- sibly painful fiscal decisions and enabled them to reinforce demands for wage moderation. However, the OMC itself is highly un- democratic. The intention at Lisbon, where it was launched, whilst not intended to subvert any conventional political processes in mem- ber states, was that the process would involve, as actively as possible, a wide range of inter- ests and actors – social partners and NGOs.8 In practice, something very different hap- pened. The National Strategy Reports on pen- sions were, to all intents and purposes, bu- reaucratic exercises carried out by national civil servants. Although most of the reports, in their introduction, refer to some sort of ex- change between those drafting the document and civil society, this overstates what often happened – a draft was passed for information and, possibly, for comment, but with no com- mitment to integrate the latter. However, two countries – Sweden and Greece – stand out as mentioning no involvement of outsiders at all, whilst some countries do no more than refer to the involvement of social partners in system administration – Denmark and Portugal, for example. In the case of Sweden, the absence of any meaningful dialogue with outside interests might explain how the concept of “actuarial solidarity” could be invented, and then used, without being questioned. And yet the OMC process could also show the way forward. One of the objectives con- tained in the pension strategy guidelines is that governments should: • make pension systems more transparent and adaptable to changing circumstances, so that citizens can continue to have confi- dence in them; • develop reliable and easy-to-understand information on the long-term perspectives of pension systems, notably with regard to the likely evolution of benefit levels and contribution rates; • promote the broadest possible consensus regarding pension policies and reforms; • improve the methodological basis for effi- cient monitoring of pension reforms and policies; and • promote the broadest possible consensus regarding pension policies and reforms. In other words, the guidelines invite govern- ments to make the reform processes as open as possible and, implicitly, they concede that reform will be possible only if governments do heed this recommendation. The guidelines encourage debate about security, efficiency and equity, and about the meaning of solidar- ity: Only if policy makers are prepared to engage in such discussions will reform be sustainable. Nevertheless, the OMC concept has a wider lesson than this. The OMC with respect to pensions is but one of several OMC and OMC-like processes. In the past year or so, 120 Evaluating Pension Reform those who have been discussing the way for- ward for pensions policy have come to recog- nise that the latter cannot be treated in isola- tion. The Commission, itself, proposed the “streamlining” of the three OMCs for social protection – for pensions, for health and long- term care and for social exclusion. Inter alia, this is an acknowledgement that, as far as issues of societal ageing are concerned, there is a need to consider questions not only of financial dependency (pensions) but also of physical dependency (care) and to recognise that elderly people might suffer from poverty (exclusion). Standing back further, however, it is possi- ble to see that the OMC-like process of the European Employment Strategy (EES) is rel- evant. The strategy does not merely set overall employment targets; it sets targets for the employment of people in their late 50s and early sixties. It also contains guidelines that encourage the development of policies and practices that will maintain the skills and the working abilities of older people, and im- prove the quality of jobs so that older people can stay in work. Logically, the streamlining process should seek to integrate the EES with the social protection OMCs (Casey, 2003b). If it were to, it would underline the importance of taking a comprehensive view of what was involved in reform, and it would leave all better able to evaluate what that reform brought and how successful it was. References Börsch-Supan, A. (2004), What are NDC Pension Systems? What Do They Bring to Reform Strat- egies? Paper presented to the World Bank and RFV Conference “Notional Defined Contribu- tion Pensions”, held in Sandhamn, Sweden, Sep- tember 2003 (http://www.rfv.se/konferens/docs/ Borsch_Supan_20040306.pdf). Casey, B. (1998), Mandating private pensions – the public finance implications, International Social Security Review, 1998 No. 4, pp. 57-70. Casey, B. (2003a), How will we provide and pay for long-term care?, European Journal of Social Security, Vol.5 No. 1, pp. 67-89. Casey, B. (2003b), Coordinating Coordination: Beyond Streamlining. Paper presented to the VDR/BM/MPI conference “The Open Method of Coordination: Quo Vadis?”, held in Berlin, Germany, March 2003 (http://www. futuregovernance.ac.uk). Casey, B. (2004), “Security, effectiveness and eq- uity in social protection”: where is ISSA going, and where should it go? Geneva: International Association of Social Security ISAA European Series (forthcoming). Disney, R. (1999), Notional Accounts as a Pension Reform Strategy: An Evaluation. Washington DC: World Bank (Social Protection discussion Paper No. 9928) (http://www.worldbank.org/ wbi/pensions/otherevents/june2003/readings/ w1/disney.pdf). Franco, D. (2002), Italy: A Never-Ending Pension Reform, in Feldstein, M. and Siebert H. (eds.), Coping with the Pension Crisis – Where Does Europe Stand?, Chicago University Press. Freeman, R. and Wise, D. (eds.) (1981), The Youth Labor Market Problem: Its Nature, Causes and Consequences. Chicago: University of Chicago Press. James, E. (1998), The political economy of social security reform: a cross-country review. Annals of Public and Cooperative Economics, Vol. 69, No. 4, pp. 451-482. Ministry (2002), National Strategy Report on the Future of Pension Systems – Sweden. Stockholm: Ministry of Social Affairs (http:// www.europa.eu.int/comm/employment_social/ soc-prot/pensions/sv_pensionreport_en.pdf). NEDO (1998), Defusing the Demographic Time Bomb. London: National Economic Develop- ment Office OECD (2001), Fiscal Implications of Ageing: Projections of Age-Related Spending. Paris: Organisation for Economic Cooperation and Development (Economics Department Work- ing Papers No. 305) (http://www.olis.oecd.org/ olis/2001doc.nsf/linkto/eco-wkp(2001)31). 121 Evaluating Pension Reform OECD (2003), Policies for an Ageing Society: Recent Measures and Areas for Further Reform. Paris: Organisation for Economic Cooperation and Development (Economics Department Working Papers No. 369) (http://www.olis.oecd. org/olis/2003doc.nsf/linkto/eco-wkp(2003)23). Scharpf, F. (2002), The European Social Model: coping with the challenge of diversity, Journal of Common Market Studies, Vol. 40. No. 4, pp. 645-670. Settergren, O. (2003), Financial and inter-genera- tional balance? An introduction to how the new Swedish pension system manages conflicting ambitions. Scandinavian Insurance Quarterly, 2003, No. 2, pp. 99-114. Williamson, J. (2004), Assessing the pension re- form potential of a notional defined contribution pillar. International Social Security Review, 2004, No. 1, pp. 47-64. Notes 1 Related to this had been an earlier discussion, conducted mainly in the USA, of the implica- tions of the baby-boom itself, whether this would lead to a reduction in the relative earnings of members of the baby-boom cohorts, and wheth- er the reduction would be a temporary or a permanent one (Freeman and Wise, 1981). 2 This was the 2002 regional conference held in Budapest in November 2002 where the theme was “Security, equity and efficiency in social protection”. The author was “rapporteur”. See, Casey 2004. 3 However vague the notion of a “European social model” is, and however much “accession” might contribute to its redefinition (for example, Scharpf, 2002), almost all definitions recognise that “solidarity” is one of its essential compo- nents. 4 The term occurs in the Swedish National Strat- egy Report submitted under the pensions OMC process where it is contrasted with “egalitarian solidarity” (see Ministry, 2002). 5 Settergren (2003) refers to “uninsurable risk” of longevity – which he calls “actuarial risk”. How- ever, this is a case of conflation of terms – since the way in which an individual experiences longevity may not be the same as the way in which society does. 6 On the day this text was being completed (02- 04-04), there were half a million people protest- ing on the streets of Köln, Stuttgart and Berlin against the “dismantling” of the German welfare state – including proposed changes to pensions. 7 That this is an “advantage” of paradigmatic change is noted in writings emanating from, on the one hand, the World Bank and, on the other, the International Labour Organisation. The situ- ation is summed up well by one of the advocates of a shift towards funding, privatisation and increased use of actuarial principles who argues: NDC systems also create a new set of “para- meters”, another aspect of the new rhetoric, which may make reform more palatable. It per- mits redefinitions and readjustments and changes the focus of debate from parametric reform to the introduction of “a new system” (while this is not the case with respect to economic sub- stance), thus enabling parameter change. This point is not without some irony, and even more so, because we claimed that NDC systems strengthen credibility through transparency, while we now use it as a device to deflect emo- tional opposition by using a new rhetoric. (Börsch-Supan, 2004, p26) 8 Thus paragraph 38 of the Lisbon Council Con- clusions read: A fully decentralised approach will be applied in line with the principle of subsidiarity in which the Union, the Member States, the regional and local levels, as well as the social partners and civil society, will be actively involved, using variable forms of part- nership. A method of benchmarking best prac- tices on managing change will be devised by the European Commission networking with differ- ent providers and users, namely the social part- ners, companies and NGOs.