The Swedish pension reform: a good model for other countries?

304 The Swedish pension reform: a good model for other countries? In this article I will discuss the Swedish Pension reform in the light of the international debate. My main point is that the core issue in a pension reform is to uphold an open and transparent generational contract, with a fair balance between active and retired today and in the future, and to maintain a political responsibility for monitoring the contract. The new Swedish pension system has transferred all financial and demographic risks onto the individual and is meant to function automatically for an indefinite future. Hence, it does not fulfil this basic task of a public pension system and it has simply become the latest example of an attempt to avoid the realities of the dilemma facing aging societies. It is not a good model for other countries. NFT 4/2003 Introduction The Swedish pension reform has been much observed in the international debate. Among features of the reform that have been especial- ly praised are the financial stability of the PAYG component, the new way of “double entry accounting” for such a system and the attempts to keep individuals continuously in- formed about their pensions. Representatives of the European Commission have described the Swedish model as the only really sustain- able approach to pension reform. The head of the European Central Bank, Mr. Duisburg, has time and again said the same. And the World Bank is also positive about the model. Following such attitudes from influential play- ers in the field, pressure seems to be mounting on countries needing to make reform to consi- der the Swedish model as an option. In such a situation, it is urgent that the model is thor- oughly scrutinized and that its weaknesses as well as its strengths are clearly understood. This journal has embarked upon a promising avenue of clarification of the true nature of the Swedish pension model. In issue nr 4/2002 Mssr Hagberg&Wohlner described their view of the new system and what they had wanted to see instead. In issue nr 2/2003 Mr. Könberg, as the leading politician behind the reform, The Swedish pension reform: a good model for other countries? KG Scherman kg.scherman@swipnet.se by KG Scherman Mr Scherman is Honorary President of the International Social Security Association (ISSA). He has had a long career in the Swedish public sector. In 1976 he was appointed State Secretary of the Ministry of Housing and Physical planning. In 1979 he became the Managing Director of the Karolinska Hospital in Stockholm. He was Director General of the Swedish National Social Insurance Board 1981-1996. His international engage- ment includes participation in the activities of the ISSA; between 1993 and 1998 he was the President of the Association. Mr Scherman holds degrees as Master of Engineering from The Royal Technical High School in Stockholm and as Master of Business Administration from The Stock- holm School of Economics and Business Administration. 305 The Swedish pension reform: a good model for other countries? gave his view, and Mr. Settergren, the leading expert, explained the technicalities as well as the ideology behind the PAYG part of the new system. This article is intended to carry the discussion a bit further. It is hoped that Swed- ish and international experts and policy mak- ers will follow and contribute their views. Reform needs and the new system; a short repetition The reform needs were similar to those pre- vailing in most industrializes countries, i.e.: A normal pension age that had been un- changed for decades, in spite of an increas- ing life expectancy; A “baby boom” generation that is approach- ing retirement; A benefit system that was overgenerous – in this case one that required only 30 years employment for a full pension and that based benefits on average earnings during the 15 best years. Many of these problems originated from the fact that the system was designed at a time when expectations about economic growth were much more optimistic than today. The new old age pension system contains an earnings-related part and, in addition, it offers protection to those who have no or only a low earnings-related pension. It contains a minimum pension, guaran- teed by the state for all residents in Sweden. The level of that minimum pension, today, is quite high. In addition, there are various sup- plements available for those who have no, or only a low, pension. The guarantee is indexed according to the cost of living, regardless of the development of wages. Hence, in the long run, its relative value will diminish in the face of growth of wages. This is the stated policy of the government. The public earnings-related scheme con- sists of two parts: a fully funded, premium reserve scheme and a pay-as-you-go scheme. A core idea in the new system is to retain a stable contribution rate of 18,5% of covered earnings into the indefinite future. The contri- bution is split between the premium reserve scheme and the pay-as-you-go scheme. Certain periods (social security benefits, child care, military service, higher education) give pen- sion rights for which the individual and the state pay the contributions in full. This is an important feature for creating social justice with-out overburdening the pension systems finances. The premium reserve scheme is new. The contribution to that part is 2.5 percentage points and it pays for life annuities based on insurance principles. It is administered separat- ely from the pay-as-you-go earnings-related scheme. Private and public fund managers are available. The rest of the administration and the insurance function of this sub-system is a public responsibility. The pay-as-you-go scheme is completely redesigned. It has become a notional defined- contribution (NDC) scheme. This redesign has been much commented upon in the inter- national debate. One of its principal intentions is to maintain a stable contribution rate into the indefinite future. This scheme has the following features. The benefit formula is tightened up and benefits are based on all earnings over an individual’s full working career. Indexation rules are linked to average wage development: o pension rights being indexed to average wage growth, o pensions in payment being indexed to average wage growth reduced by 1.6 percent per year (“flexible indexation”). Benefits are made dependent on life expect- ancy, meaning that a benefit drawn at a certain age by an individual belonging to one cohort will be lower than that for the preceding cohort, if life expectancy has increased. The PAYG part is financed by a contribution of 16.0 percent. 306 The Swedish pension reform: a good model for other countries? The PAYG system contains two other im- portant features. The first is an automatic balancing mechanism. New calculation meth- ods have been established to make it possible to estimate the assets and liabilities of the PAYG scheme. If the estimated liabilities of the system exceed its assets, the yearly revalu- ation of pension rights and pensions in pay- ment will be reduced enough to enable pen- sion liabilities to grow at the same rate as the system’s assets. Obviously, such a mecha- nism makes the system financially stable. Whatever happens, it reduces current and future pensions by as much as needed in order to restore financial equilibrium to the system. The second special feature is a special fund, called the buffer fund. All contributions are paid into the fund and all pensions are paid out of this fund. As a consequence, the buffer fund accumulates capital in certain periods, for example if large cohorts reach working age or if labour force participation increases. The surplus generated under such periods will be used to counter financial strains on the system in other periods. Such strains will emerge when the baby boom generation reaches pen- sion age. At the outset of the new system, most of the pension fund that had been accumulated under the old pension system was transferred to the buffer fund, where it serves as a sort of “start up capital”. The result: Work more, much more, or accept a lower pension The whole working career is the basis for the pension and the benefit drawn at a given age becomes lower for later cohorts when life expectancy increases. Hence, a basic implica- tion of the new system is that people will have to work longer or save more – considerably longer and more than many realize – or to accept a lower annual pension. Already for those approaching retirement the benefits will be reduced as compared with the old system. Younger persons will be sub- ject to a further reduction in pension benefits as a result of increases in life expectancy – theoretically speaking calling for postpone- ment of retirement with around one year for those born in 1954, nearly two years for a person born around 1975 in order to restore a certain level of the pension. According to estimates made by the National Social Insur- ance Board, a person who extends his or her working life accordingly, the replacement rate will stabilise around 60 percent of average lifetime earnings. For a “model person” with steady earnings over more than 40 years, earnings that increase along with general wage trends, this outcome will be equivalent to 60 percent of final earnings, a fairly high replace- ment rate. But reality is not likely to result in such a favourable outcome as this. There are several reasons why. In the Board’s calculations, the estimates of return on investment in the funded part of the earnings-related scheme are fairly optimistic. Moreover, people rarely work with steady earnings over more than 40 years. With more conservative estimates of what the financial market can produce, and when applying the calculations to “real human beings” the out- come is rather different. It turns out that to reach a replacement rate of 60 percent a per- son born in the 1940s might have to work until 67; one born in the 1950s until around 68; and one born in the 1960s until around 68 and a half. Of course, following the fact that the amount of the pension is based on average lifetime earnings, the result differs considera- bly dependent upon the lifetime work pattern of an individual. Nevertheless, there is a sub- stantial increase in the age at which a certain target replacement rate can be obtained. To these calculations should be added the effect of the automatic stabilizing mecha- nism. According to recent estimates, the risk of this mechanism being activated sometimes in the future is around 30% by the turn of 307 The Swedish pension reform: a good model for other countries? 2002/2003. According to one scenario, the result could be a reduction of the PAYG part of the pension of around 15%. In the case of the example above, such a development would add another two years work to reach a target replacement rate, whatever that target was. Which incentive effect is the greatest –to work and/or save more or to accept a lower pension- is much debated. There are many who believe that the incentive to work longer or to save more is weak and that what will generally happen is that people will draw their pension as soon as it is possible and will accept a poor standard of living. The same will be the result if work opportunities are not available. Others think that the design of the pension in the long run will strongly affect attitudes and opportunities in society. The official language is clearly in line with the latter belief. There is neither any reference to a pension age nor to a replacement rate. The key words are “flexibility“, “freedom of choice” and “abolition of the fixed retirement age“. The ideology of the reform lies hidden behind these words. A “paradigm shift” has occurred To understand the new Swedish Model, it is necessary to realise that the new system is completely different from the old one. It is an actuarially-based system and it has thorough- ly overturned the distribution of risks between the individual and the state. A “paradigm shift”, to use the World Bank vocabulary, has occurred. Consider the discussion about the merits of a PAYG versus a funded system. That discus- sion has been intense over more than 20 years now, beginning with the World Bank promo- tion of the so called Chilean Model, and taken further by the publication “Averting the Old Age Crises” by Estelle James et al1 . In the discussion, certain features have usu- ally been attributed to each of the two models, although as the following discussion shows, not all of them are essential attributes. How- ever, for the sake of understanding the Swed- ish model this dichotomy can help to clarify the matters at hand. PAYG systems are often said to be character- ized by the following features: contributions flowing into the system one year are used to finance the same year’s pension payments; the system is defined-benefit; the system is publicly administered; and the political process is supposed to ensure that the necessary steps will be taken to ensure a balance of social goals and finan- cial constraints, both today and in the future. It is only the first of these bullet points that define a PAYG system in the strict sense, the rest of them are features that go together with traditional public PAYG systems. The strength of a system with these attri- butes is that it takes care of some problems facing the individual when he or she wants to plan for retirement, problems that explain public involvement in pensions that goes be- yond the desire to alleviate poverty. The wish to take care of poverty alleviation is a vital part of the reason for public involvement with retirement programs. By itself, however, it does not explain the scope of this involvement. The near universality of comprehensive pub- lic actions with respect to pensions suggests a general consensus that individual decisions and free markets can not be counted on to produce a desirable level or pattern of savings for retirement. There are several reasons for this. These include the wish to avoid myopic behaviour, to reward the prudent, and to pro- tect people from insurance market failures. Myopic behaviour means that some indi- viduals give too little weight to the utility of future consumption, resulting in them saving too little, and realising this only when they are 308 The Swedish pension reform: a good model for other countries? already old and unable to do anything to cure their previous mistakes. There is an obvious need to reward the prudent by ensuring that they are not exploit- ed by those who do not want to take responsi- bility for themselves, but equally, to ensure that those who do try to look after themselves receive an adequate benefit. Most societies establish some minimum level of consump- tion below which they do not want their mem- bers to fall. Insurance market failures are a reality. Among the problems that a complete reliance on private markets causes for the individual are the insurmountable difficulties in estimat- ing: future economic growth rate and future returns on investments; future trends in average mortality; changes in price and wage levels after re- tirement; and his or her own longevity, relative to that of the cohort as a whole. There are weaknesses, too, in such a PAYG system in the real world. One is that the existence of a minimum pension creates a moral hazard, in that some individuals will decide to rely on the minimum benefit instead of making their own provision for retirement. Another is that politicians have tended to promise too much, and that people tend to believe that someone else is paying for their benefits. Funded systems are often said to be charac- terized by the following features: contributions paid in by, or on behalf of, an individual, are accumulated and the accu- mulated savings are used to finance pension payments; the system is defined-contribution; the system is privately managed; and there is no political responsibility for bal- ancing social goals and financial constraints, either today or in the future. Here to, it is only the first of these attributes that define the system in the strict sense, the rest of them are features that go together with a general notion of funded systems. The debate on the potential impact of a funded pension system on the economy, and whether it should be privately managed, is intense. One line of argument concerns ques- tions about the impact of funded pensions on savings and investments, and on growth of the economy. The debate is not conclusive, but the argument that there should be any signif- icant impact is losing ground. What is clear, however, is that, like Profes- sor Barr2 , many economists have now under- stood that the idea that funded systems mean “that people take care of themselves in old age”, while PAYG systems mean that they “leave the responsibility to their children”, contains a misunderstanding of how the ag- gregate economy functions. At the aggregate level, consumption goods cannot be stored; it is always today’s production that is distribut- ed between active and non active. Another line of argument concerns the desirability of letting private entities admin- ister pension schemes and manage the in- vestment of funds, and of “getting the politi- cians out”. Given the fact that the effect on aggregate economy might not be all that great, the insistence on the merits of funded solu- tions probably rests with mistrust in the poli- ticians’ ability to cope with long range prob- lems. This is forcibly advocated in “Averting the Old Age Crises” and many other publica- tions. About all these matters, about PAYG vs. funded pension schemes, a wealth of in- formation can be found in ISSA3 and ILO publications, for instance in the ISSA Re- view4; in the documentation of the ISSA conference on the Future of Social Security, held in Stockholm 19985;in the ILO survey ”Social Security Pensions; Development and reform”6, and in Mr Thompsson’s ground- 309 The Swedish pension reform: a good model for other countries? breaking book “Older and Wiser; The Eco- nomics of Public Pensions”7 . The new Swedish NDC scheme does not fit into either of these two broad types of models. Following the introduction of the automatic balancing mechanism, it has the following characteristics: contributions flowing in in one year are used to finance the same year’s pension payments, with a buffer fund countering variations in the inflow of contributions relative to the outflow of pension payments; the system is strictly defined-contribution; the system is publicly administered; but there is no political responsibility to bal- ance social goals and financial constraints, either today or in the future. Thus, although it is a PAYG system, it lacks most of the strengths such systems tradition- ally have. But, still, the commitment to retain a mandatory PAYG scheme into the future is there. Much of the confusion around the Swedish model stems from the fact that the power of traditional thinking is so great. “A PAYG system, and, of course, especially a PAYG system in Sweden, must retain the basic fea- tures of social responsibility traditionally as- sociated with such systems”. That is how many people think. But they are incorrect. In particular as a consequence of the automatic balancing mechanism, a truly defined-contri- bution system has been created – benefits become totally dependant on the contribu- tions and on internal rules in the pensions system. Combined with the provisions where- by every amount of contribution creates cor- responding pension rights, and with a pledge not to increase contributions in the future, there is no room for manoeuvre left. There is no way to monitor the generational contract and no way of adjusting the system in the face of changes in external conditions to attain a fair balance between social goals and finan- cial constraints in the future. Hence, one of the prime aims for transferring to a funded sys- tem, to “get the politicians out” has been achieved in the new Swedish PAYG system! All risks stemming from external disturbanc- es on the financial performance of the system are automatically transferred into reduced benefit levels. The paradigm shift is thereby realised. The design of the guarantee pension is another feature in the new model that contrib- utes to the picture of shifting of risks to the individual. The guaranteed level is price-in- dexed. If wages grow in real terms, and the guaranteed level is not adjusted accordingly, the level that the guarantee provides succes- sively decreases compared to wages. The Swedish government has stated that it finds it appropriate to allow its relative importance to diminish in the face of a real average wage growth. Today the guarantee is around 80 000 SEK per year and the ceiling for a pension from the public system is around 180 000 SEK. At a growth in real wages of 2%, the ceiling will become 360 000 SEK in 35 year’s time, while the guarantee level will remain the same: 80 000 SEK. This development is con- trary to what usually has been considered to be the basic interest in most public pension sys- tems. At least, this has clearly been the case in the Swedish context. It is the sharing of risks that is the most important feature of an insurance system. This has nothing to do with whether a partic- ular scenario has a high or a low probability of materialising in the future. And, as a matter of fact, whatever the quality of prognoses, we end up in not knowing much about such probabilities. The automatic adjustment mechanism in the PAYG component sees to it that all risks of an imbalance in the finances of the earnings- related schemes are borne by the individual. As already mentioned, the automatic balanc- 310 The Swedish pension reform: a good model for other countries? ing mechanism, according to one scenario, could result in pensions being as much as 15% lower. The design of the guarantee pension transfers successively greater poverty risks onto the individual. Moreover, it is worth pointing out that, because the effect of a raising life expectancy is to be borne on the benefit side, all risks of not being able to find suitable work at an advanced age rest with the individual. And the social insurance schemes covering unemployment or inability to work due to sickness or disability cover only people up to the age of 65. A consequence of these restrictive features of the new system is that people have to resort to private arrangements to a higher degree than before. This transfer of risks from the public to the individual exposes the individual to the dilemmas of myopic behaviour and insurance market failures. These are risks that traditionally are seen as to be borne collective- ly. Their existence was the reason for public involvement in the first place. The paradigm shift has emerged only gradually One reason why it has been possible to make this complete overhaul of the pension system might be that it has occurred behind closed doors. It was the result of deliberations by a group of politicians in charge of implement- ing a reform of which, back in 1994 only the general principles had been agreed. The para- digm shift, itself, was neither agreed in 1994 nor was it ever presented subsequently as a change of those principles. Mr Settergren rightly draws the following conclusion from the situation that gradually emerged: “As Swedish pension reformers had set out to create a (notional) defined-contribu- tion scheme it was necessary to make sure that the system was financially stable. Otherwise it would have been logically inconsistent”. But the content of the 1994 principles was quite different. There, it was generally accept- ed that keeping a balance between social goals and financial constraints was to be a leading principle. Goals were formulated with respect to re- placement rates that reflected what was socially acceptable A wish for a stable contribution rate was clearly formulated, but the switch from a defined-benefit to a defined-contribution system was portrayed as the result of the introduction of a full working career as the basis for the pension rather than as an over- riding principle. The financial constraints, or – more accu- rately – the absence of financial constraints, were formulated in the discussion about the buffer fund that became the backbone de- termining the financial performance of the system. In this context, the terms of the original documents that proposed the pay- ment to the state budget of monies to com- pensate for some of the extra burdens that it would incur as a result of the reform are worth citing. After having described the proposed compensation, the text reads: “Of course these proposals affect, as has been described above, only the financial side. Neither the successive phasing in of the contributions, nor the transitional use of the buffer fund for other then old-age pension payments, affect the benefit side, that is obvious”. In 1994, those responsible for the reform thought they could guarantee that the new rules could be kept in place for the foreseeable future, even if the level of contributions was to remain constant. It was claimed that the re- serves accumulated in the old pensions fund would ensure this. There would be enough left over to compensate the national treasury for the extra burdens that the reform place on the state budget. As a matter of fact, the new pension system as designed 1994 was still a traditional PAYG system, although with some 311 The Swedish pension reform: a good model for other countries? interesting new features, as Mr Chichon points out in his article of 19998 . This was also the way in which the reform was presented to the general public, especially by the Social Dem- ocratic party that was endeavouring to get its members to accept the reform. Later on it was discovered that the financial situation was not as favourable as believed in 1994. Reserves were not sufficient to both cover pension obligations and compensate the national treasury. This was because it had become apparent that the demographic pro- jections initially used were out of date. Adults were living longer, and fewer children were being born. Despite these altered conditions, large sums have been transferred from the fund to the national treasury. So far, SEK 258 billion has been transferred, which is roughly one-third of the fund’s reserves. And more is intended to follow. What happened was that the idea of the contribution rate being kept unchanged indef- initely was allowed to become a cornerstone of the reform, and that the wish to transfer funds to the state budget in the very same process was transformed from a result of projections showing that there was money left over in the buffer fund into one of the leading principles of the reform. Following this, the automatic balancing mechanism was invent- ed and put in place. But this was a matter of choice. Obviously, there were alternatives to this set of leading principles. As a matter of fact, when the projections and assumptions behind the initial reform proved unsustaina- ble, the whole project should have been recon- sidered and subject to open debate. The collapse of the projections behind the 1994 principles was never brought into the open and no public debate occurred. The result of the decisions made was, as Mr Setter- gren points out in his article, that social justice became the same as inter-generational bal- ance defined as “having a constant ratio of present value of pension benefits over present value of contributions for all birth cohorts”. This is the guiding principle behind the final design of the new PAYG system, with its automatic balancing mechanism. There is no room left for any other social goals or for a political monitoring of the generational con- tract in the future. The impression of a gradual shift of focus, and of a gradual retreat from political respon- sibility for the social outcome of the pension system is aggravated by a series of other features of the present situation. Among these is the government’s stated opinion that the value of the minimum pension shall diminish in the face of real wage growth. Another is the fact that the social safety net is not extended to higher age groups as the de facto pension age is increased. A third is the complete change of the funded component. In 1994, it was stated that the funded scheme should include a guar- anteed minimum yield and that the wish to provide for diversity in the management of funds should not be allowed to take prece- dence over the wish for security. The rules governing life insurance companies were mentioned as good examples in this respect. Ultimately a completely different model was designed, with 700 funds for the individual to choose between and with no minimum guar- antee. This complete overhaul was presented neither to the parliament nor to the general public as a change of principle. The paradigm shift has created great confusion Those advocating a “paradigm shift” often say that a complete overhaul of the vocabulary should make it easier for the general public to understand and accept necessary changes in the pension system. Most probably, this is a false hope. Instead what happens, or at least, has happened in Sweden, is that people do not understand anything at all. Even among ex- perts, the debate has become confused. 312 The Swedish pension reform: a good model for other countries? The changes necessary to make the system financially sustainable are mixed up in the changes of principles, and the reform leaves the general public behind. Four examples may illustrate this situation. the increase in the pension age that is brought about by introducing a factor dependent on remaining life expectancy in the pension calculation formula and by changing the rules for flexible retirement. On the basis of this, it is claimed that there is “free choice” and “flexibility”, when, in fact, what is happening is that the retirement age, as that concept is conceived today, will be gradual- ly raised. Should the regular measures built into the system prove insufficient, the auto- matic balancing mechanism will take care of the need for an extra reduction in bene- fits, forcing people to try and postpone retirement yet further. the reduction in the replacement rate that is brought about by not merely increasing the number of years taken into account in calcu- lating the benefit, but also by switching to a lifetime perspective, introducing a couple of non-contributory periods into the basis for the pension, and changing the indexation method from the price index to the wage index. A comparison of the new and the old system show “winners” as well as “losers” instead of only “losers”. This obscures the fact that an important result of the reform is the requirement for people to work longer under the new than under the old system to obtain a pension of a given level. the lack of clarity surrounding the worth of the funded component of the pension that is brought about by the design of the funded component with its confusing range of funds and with pensions solely dependent on whichever market return on investments that the individual can obtain. Every discus- sion on the merits of this component una- voidably ends up in complete uncertainty, since no one knows what the development will be in the future. A good illustration can be found by comparing what Mr Könberg and Mssr Hagberg&Wohlner wrote in this respect. the fundamental change in welfare policy in the long run that is implied by some ele- ments of the reform. Particular mention should be made of the effective cut in the level of the minimum pension, since its level is indexed to prices and, relative to wages it will fall if there is an increase in average real wages. This change in welfare policy has never been discussed openly. A consequence of this approach is that the “losers” only gradually realize what has hap- pened and this hampers the political process. No one knows what part of the public re- sponse is caused by ignorance and what part is caused by acceptance. The true nature of the new system is poorly understood. Mssr Hagberg&Wohlner advo- cate successive reforms, with political re- sponsibility retained in order that the genera- tional contract can be monitored. This is the approach applied by Germany, France and the US – countries that, to date have introduced “mere parametric reforms”. Mr Könberg com- ments upon Mssr Hagberg&Wohlner’s argu- ment by stating that there is no major differ- ence between their proposal and the system that has been implemented. But this is incor- rect. In the world of “parametric reforms”, the traditional focus on balancing social goals and financial constrains is retained. In Mr Kön- berg’s world, politicians have withdrawn from taking on such a responsibility. Mr Könberg also makes the point that his various oppo- nents (there are more critics than Mssr Hag- berg&Wohlner) have differing views about what the best alternative should be. But all his opponents set social goals at the centre. In that respect, they have much in common, and they all differ from Mr Könberg. There is no room 313 The Swedish pension reform: a good model for other countries? for an automatic balancing mechanism of any sort in their alternatives. It is also worth ob- serving that none of the other countries that have reformed their pension system in a fash- ion that is said to resemble the Swedish NDC model (for example, Latvia, Poland and Italy) have included a component like the automatic balancing mechanism. The outcome of the system is not known. When a traditional PAYG system is said to be financially sustainable, this is normally taken to mean that it is sustainable over all. But, as a consequence of the design of the balancing mechanism, the emphasis must switch to the benefit side and the social consequences of the new system. Traditionally, when reforming a mature public pension system, it is projections show- ing a need for an increase in the contribution rate that causes alarm. It is important to realise that this follows from the design of a defined- benefit system as such. As, by definition, the benefit rules are established in advance, every disturbance in the system emerges as a distur- bance of its financial sustainability. In the new, earnings-related Swedish system, it is the financial rules that are defined in advance. These rules establish the financial scope for the total amount of benefits. Individual rights and pensions are adjusted accordingly. If the wish is still that the system should offer a fair balance between financial and social goals, the “alarm system” has to be redesigned. Instead of it registering the need to reset contributions, since such resetting is not per- mitted, there is a need for it to register well- being and its future development. This re- quires the establishment of social indicators and the development of means to make projec- tions of these into the future. Among such indicators are likely to be those showing the relationship between pensions and wages and income disparities among pensioners – in each case illustrating how these would stand given differing economic and demographic assumptions. No such indicators are available. Even Mr Könberg is subject to misinter- preting the outcome of the new system. He states that, given 2% economic growth and certain other conditions, pensions under the new system will, on average, be the same as under the old one. In one scenario, already described above, the automatic balancing mechanism could lower pensions by as much as 15 %. This is a significant difference. Confusion also affects the information to the general public. In Sweden, the National Social Insurance Board provides an individu- al report on a yearly basis, indicating the amount of the benefits that can be envisaged under different scenarios. Surprisingly enough, the information does not contain any calcula- tion of replacement rates, in spite of the wide- spread knowledge that individuals have great difficulties in making comparisons between absolute amounts, one based on the present value of money and real wage level, another reflecting some future situation. It is also worth pointing out that the assumptions about returns on investment used in these individual reports seem to be rather optimistic, giving the individual an impression that the pension will be quite high. A more conservative assump- tion would have produced a considerably lower pension. Another part of the endeavour to make the pension system as transparent as possible is the annual reports on the system’s assets and liabilities. The first of these annual reports was compiled and presented in 2002. It is very likely that this attempt to present the financial balance of the PAYG system needs further analysis, by actuaries and other experts, be- fore any conclusions can be drawn about how well it describes the real situation. It is also to be noted that the annual report contains no indicators of the adequacy of pensions in the future for “real life people” or, as a matter of fact, not even for “model persons”. 314 The Swedish pension reform: a good model for other countries? Item Present situation Future situation Finances of earnings-related pensions Stable Stable Contributions from the state budget for social security benefits High ?? Amount of minimum pension Fairly high Successively lower in the face of economic growth Costs of minimum pension Modest Diminishing Adequacy of pensions with respect to the full set of services and benefits available for older people Acceptable ?? Availability of means tested supplements (especially housing allowance) Good ?? Income disparities among pensioners and between them and the active population Small Most probably increasing Employment opportunities for people aged 55-67 Comparatively good but much too low to meet the need Probably improving Employment opportunities for people aged 67-75 Non existent ?? Availability of in-kind benefits Comprehensive ?? Cost-sharing for in-kind benefits Low ?? Public costs of in-kind benefits High ?? Availability of adequate health care Good ?? Cost-sharing for health care Low ?? Situation for the elderly Good ?? The arrangements for old age contain much more than pensions, and employment is the key factor The potential consequences of the operation of what is a mechanistic pension system raise a number of further questions about the future well-being of older people. These questions relate to the availability of various kinds of services and whether those services are subsi- dized or not. It is worth asking whether a pension system can function automatically and offer an adequate level of well-being without reference to what happens in these fields. They also relate to the labour market. It is worth asking whether there will be suffi- cient employment opportunities for people, not only when they are older but at all stages of their life, given that the whole working career is the basis for the individual pension. A table may summarise the state of knowl- edge. See below. As can be seen, the only thing we do know about the future is that the finances of the new Swedish earnings-related pension scheme are secured. But we do not know anything about the adequacy of the pensions that are offered. We know very little about the replacement rates that can be obtained, and we know even less what the environment in which the pen- sions’ adequacy should be measured will be. The pension system is meant to function auto- matically, but, not only do we not know what this means for pensions, we also do not know how a large number of factors that determine the ability to accumulate a pension or to have a satisfactory level of well-being in old age are going to look. This simply adds to risk. Pensions Work Other Over- all 315 The Swedish pension reform: a good model for other countries? The dilemma just described cannot be solved by trying to formulate policies and make forecasts of costs and social outcomes in all dimensions. Instead, this analysis makes it abundantly clear that the future is uncertain and will remain so. The present can not pre- scribe what the future will be. Consequently, a totally automatic public pension system cannot, realistically, prevail. The de facto increase in pension age places the focus on employment opportunities for older people. It is necessary to make it possi- ble to continue working after age 65; other- wise the reform becomes merely a new way of reducing pensions. That requires laws or col- lective agreements that prohibit mandatory retirement (or at least mandatory retirement before at least the age of 70). However, not even such provisions can guarantee longer working lives. Work environments and em- ployment conditions must be adapted for old- er workers, and there must be a change in attitude regarding their rights, and those of all employees, to develop their occupational skills and knowledge. There must also be a change in attitudes in the labour market, among em- ployers, labour unions and older workers, themselves, regarding older people’s ability and potential. Even that would not be sufficient. Those who would like to work a few years past age 65, but who, due to illness or lack of employ- ment, cannot work must have access to the general welfare system as well as social insur- ance under the same terms as younger people. These are the most pressing needs in every aging society and they can not be avoided simply by reforming the pension system. The reform needs to be reformed The point of departure for the Swedish reform is a realistic one. In the face of growing life expectancy, there is no viable alternative to raising retirement ages and encouraging peo- ple to work longer. The general principles approved by the Swedish parliament in 1994 were supposed to satisfy the need to lower costs and also to provide for greater flexibility in response to future changes in the popula- tion and the economy. The level of political agreement was regarded as especially signif- icant, as were the unequivocal assurances that the new system would be financed in such a way that it could be sustained without any restrictions on pension levels, beyond those specified in the general principles that were set down then. Many years have passed since 1994 and the system that has emerged differs in many re- spects from that originally envisaged. More- over, even the original reform plan was charac- terised by a lack of concern for the social goals that should be taken into account in the design of a pension system. Hence, a whole range of measures are needed in order for it to be possible to say that the Swedish pension poli- cy offers benefits that are secure, adequate and equitable. Action is needed too make it possible to continue working after age 65. Moreover, people aged over 65 needs to have access to the general-welfare system as well as social insurance under the same terms as younger people. Otherwise, all the talk about “flexibil- ity“ and “people being allowed to decide for themselves when to retire“ will lead merely to further reductions of pension benefits in the future. “Freedom of choice“ will exist only in theory, not in reality. Steps must be taken to revise the policy governing the minimum pension. The size of the guaranteed level for the pension is tied to inflation. If earnings increase at a faster rate than inflation, the guarantee will become less and less important with the passage of time. Such a development is not acceptable and it lacks fairness and political credibility. The relationship of the minimum pension to aver- 316 The Swedish pension reform: a good model for other countries? age income must be reviewed and adjusted from time to time. The system must be changed so that the automatic balancing mechanism in the PAYG earnings-related scheme is rescinded. This mechanism is unacceptable. Social goals and the basic concept of social insurance imply a political responsibility to monitor the pension system and take necessary measures if its financial and/or social stability is threatened. This was the case under the old system, but it must also be the case under any new system. Most probably, the consequence of this obser- vation is that the NDC-concept, as it operates at present, does not work properly. In an NDC-scheme, it is the contributions that con- stitute the pension rights. Hence financial imbalances can not be remedied by raising contributions, as this would give rise to addi- tional pension rights, causing new financial strains in the future. It is not possible to foresee what will happen in the future, hence it is not possible to foresee which measures need to be taken in order to arrive at a proper balance between social goals and financial constrains. Therefore, it is not advisable to rule out the possibility of raising contribu- tions in order to restore the financial balance in a way that does not, automatically, add pension obligations to the system. It is essential that the monies that have been transferred to the state budget be returned to the buffer fund. To date, these transfers have totalled SEK 258 billion. Such a depletion of the buffer fund will have unacceptable conse- quences for pensioners. It increases the risk that the automatic balancing mechanism will be activated and, consequently, leads to fur- ther decreases in the level of pensions. Turning to the premium pension, changes should be made so that a minimum guaranteed yield on investment is introduced. Such a redesign would provide for a risk sharing among the participants in the scheme. It would also offer a realistic basis for the information that is provided annually to the individuals about the pension they have accumulated to date. Last, but by no means least, the actual results of the new pension system for real-life people must be evaluated in relation to the goals declared in 1994. Simulations for a “model person” are not adequate. Assumptions about the economy, demography and invest- ment returns must be safe, simple and explicit. Only in this way will it be possible to provide the population with clear and transparent information that will assist them in decision making and in their appraisal of the reform. The Swedish pension reform was charac- terised by an ambition to achieve a broad political consensus about the design of the pension system. It is likely that this broad consensus was a prerequisite for breaking the deadlock that had prevailed in the political system for more than 10 years. More prob- lematic is the fact that this consensus is still there, eight years after the decision on the principles for the reform. In the long run, this could have a perverse effect. If the political system and the other representatives of the individual citizens, notably the unions, fail in their task of examining, questioning and pro- posing alternatives, it is perfectly possible that people will become frustrated and disap- pointed. Ultimately, it is a question of what the Swedish people as a whole want from a pension system. We are still waiting to see what will happen once the political debate on these issues is revived, as it surely will be when some of the results become apparent. It is to be hoped that people will start to call for a greater element of solidarity, “solidarity” in the traditional sense of the word, to be instated into the system. 317 The Swedish pension reform: a good model for other countries? A model for other countries? As to whether or not the Swedish reform could be a model for other countries the an- swer is inherent in the conclusions drawn above. The new Swedish pension system is not a good model for other countries. Instead, it has simply become the latest example of an attempt to avoid the realities of the dilemma facing aging societies. These realities can be enumerated in the following way, with obvi- ous conclusions attached to each of them: We know very little about the future Hence, forget about the automatic pension system, which means o do not introduce an ”automatic adjust- ment mechanism” o do not preclude the need to raise contri- butions without granting new pension rights, i.e. avoid an NDC model that excludes such a possibility. There is no way to avoid the fact that it is the active population that provides for the in- active Hence, in the face of a growing life expect- ancy, the alternatives are: o to increase the pension age o to increase contributions o to decrease pensions, or o to do some combination of these three Do what has to be done openly, frankly and transparently. The baby boom “problem” needs special attention Remember that the liabilities are already there, so o the problem of societal ageing is too complex to solve merely by manipulat- ing the pension system. Recognise that action must be taken now if a collapse of the pension system is to be avoided, and therefore o most probably it is wise to begin to build up a substantial fund o but other measures are also needed to strengthen the economy. Core issues are Improve employment opportunities and conditions for all, including the elderly. Face the realities of a raising life expectan- cy, and make necessary changes openly, including raising the pension age. Strengthen the national economy in order to make it possible for fewer active people to provide for the baby boom generation when this retires. Uphold an open and transparent genera- tional contract, with a fair balance between active and retired today and in the future. Sweden has gradually, and without an open debate, designed a pension system that rests on the false presumption that it is possible to avoid political responsibility for upholding an implicit intergenerational contract or for moni- toring and evaluating that contract. The model that has been constructed is profoundly un- democratic, and, for this reason, if no other, it will ultimately fail. So, too, will any other model that denies the basic prerequisites for a solution to the demographic dilemma that faces most countries in the world today. Notes 1 World Bank; Averting the Old-Age Crises: Pol- icies to protect the old and promote growth, published by Oxford university Press, Oxford, in 1994 2 See Nicholas Barr; The Economics of the Welfare State, published by Stanford University Press in 1987, and many subsequent publications, for instance Reforming pensions: Myths, truths, and policy choices, International Social Security Re- view, vol. 55, 2/2002, pp3–36. 3 ISSA (the International Social Security Associa- tion) is a non-governmental organisation (NGO) with the status of consultative member of the Economic and Social Council of the United Na- 318 The Swedish pension reform: a good model for other countries? tions. Members are institutions administering Social Security benefits or working in related areas. For the moment, the ISSA consists of nearly 400 member organisations from 150 coun- tries. 4 See in the International Social Security Review, published in English, French, German and Span- ish (in order of appearance to reflect the develop- ment of the debate) Beattie, R; McGillivray,W; A risky strategy: Reflections on the World Bank report “Avert- ing the old age crises” vol. 48 3-4/1995, pp. 5–22. Quisser, M; Chile and beyond: The second- generation pension reforms in Latin Amer- ica vol. 48 3-4/1995, pp. 23–39. Reynaud, E; Financing retirement pensions: Pay-as-you-go and funded systems in the European Union vol. 48 3-4/1995, pp. 41– 57. James, E; Providing better protection and pro- moting growth: A defence of “Averting the old age crises” vol. 49 3/1996, pp 3–17. Beattie, R; McGillivray,W; Rejoinder vol. 49 3/1996, pp 17–20. Singh, A; Pension reform, the stock market, capital formation and economic growth: A critical commentary on the World Banks proposals vol. 49 3/1996, pp 21-43. McGillivray,W; Pension reform: Where are we now? vol. 53 1/2000, pp 3–10. Holzmann, R; The World Bank approach to pension reform vol. 53 1/2000, pp 11–34. Gillion, C; The development and reform of social security pensions: The approach of the International Labour Office vol. 53 1/ 2000, pp 35-63. Scherman, K.G.; A new social security consen- sus? The ISSA’s “Stockholm Inintiative” vol. 53 1/2000, pp 64–82. Ross, S.G.; Doctrine and practice in social security pension reforms vol. 53 2/2000, pp 3–29. Quisser, M; Pension reform and international organisations vol. 53 2/2000, pp 31-45. 5 The conference proceedings are published in English, French, German and Spanish. See the publication The future of social security, pub- lished in Stockholm by the Federation of Social Insurance Offices, in 1998 in English and in 1999 in French, German and Spanish. 6 ILO; Social Security Pensions; Development and reform, published by the International La- bour office (ILO), Geneva, in 2000. 7 Lawrence Thompsson; Older and wiser: The economics of public pensions published in Eng- lish by Urban Institute Press, Washington, DC, 1998, in Spanish by CISS, the Inter-American Conference on Social Security, 1999 and in Por- tuguese by the Brazilian Ministry of Social Secu- rity, 1999. 8 Michael Cichon; Notional defined-contribution schemes: Old wine in new bottles?, International Social Security Review, vol. 52 4/1999, pp. 87– 105.