The New Swedish Pension System – a Fair and Sustainable Model

15 The New Swedish Pension System – a Fair and Sustainable ModelNFT 1/2004 Introduction In 1992 the Swedish parliamentary Working Group on Pensions presented ideas for a com- prehensive reform of the pension system that had existed since the beginning of the sixties. In the sketch1 two main principles were ar- gued for. Those were the life-income princi- ple, that is that the fees paid should determine the pension, and the adjustment principle, that is that pension rights and the pensions should follow the development of the economy and the life expectancy. Those principles became central parts of the agreement in the beginning of 1994 between The New Swedish Pension System – a Fair and Sustainable Model by Bo Könberg Bo Könberg bo.konberg@riksdagen.se In this article I comment on some of the many critical comments that KG Scherman made in his article (NFT 4/2003). I try to concentrate on points that I have not commented in my reply (NFT 2/2003) to Hagberg’s and Wohlner’s article (NFT 4/2002). I notice that Scherman agree with me that a reform was necessary, that in the future people will have to work longer or save more as life expectancy increases if they want the same relation of income and annual pension and that it is very important to improve employment opportunities, especially for older persons. I criticize him for presenting a misleading picture of many parts of the 1994 reform, for being very vague on what financing of his implied improvements would cost the active population, for advocating a system of recurrent negotiations between the political parties which most probable would lead to more abrupt and unpredictable changes than the system of rules that are an essential part of the new Swedish pension system. five parties representing some 85 % of the electorate. The new system consists of two parts, both are defined contributions (DC). The major part is a non-financial or notional system and the minor part is a financial or funded part. The former part is considered by many pension experts as an innovation and the first that has been approved in a legislative Mr Könberg, M.P., is Group Leader of the Liberal Party in the Swedish Parliament. He was Minister for Health and Social Insurance 1991-94. During these years he was also Chairman of the parliamentary Working Group on Pensions, which created the new Swedish pension system. Since 1994 he is a member of the parliamentary Implementation Group for the new system. 16 The New Swedish Pension System – a Fair and Sustainable Model body and implemented in any country. Some experts2 have argued that the new Swedish model rather than being a comple- tely new model can be seen as “a thoroughly reformed defined benefit scheme”. I will in this article not comment on that particular debate, but I am inclined to see it as a new model. Much more important than that is of course whether it is a good model or not. I mentioned in an earlier article in NFT3 that several countries since 1994 have enacted versions of the NDC (Notional Defined Con- tribution) model. Mr. Scherman confirms in his article4 this development and also adds that international experts from institutions like the European Commission, the European Central Bank and the World Bank are positive about the new Swedish model. More recently the government of the Czech Republic has announced that they intend to propose a NDC model and a large majority of the parliamentary Norwegian Commission on Pensions has proposed a similar model for their country. Among the many experts that has been positive about the model was as late as 1999 Mr. Scherman himself. He then concluded a report for ILO5 with “My concluding opinion is that Sweden is in the process of building a new pension system that is sustainable and, generally speaking, fair.” (Italics added here). Mr. Scherman played an important role in the eighties in arguing for the need of a reform and as head of Swedish National Social Insu- rance Board he supported the agreement of 1994. As the reader of his very critical article can see something drastic has happened since then. The question is whether this change has occurred with the pension system as such or with Mr. Scherman himself. I hope that I in this article will throw some light on the answer to that question by commenting on what chan- ges that really have been made in the reform since Mr. Scherman wrote his paper five years after the 1994 agreement. I will start this reply with just mention some important questions on which we – still – agree. It was necessary to reform the old Swedish system. It is very important for Swe- den and most other developed countries to improve the employment opportunities espe- cially for elder people. (He does not mention that an important part of the Swedish reform is that the right to stay in work now has been raised from 65 to 67 years and that the pension system itself has no ceiling at all.) It is neces- sary to work longer or save more when the life expectancy increases if you want the same replacement rate for your annual pension as before the increase. His main criticisms in the 15 pages long article seem to be that the system is too much governed by rules instead of political negotiations when needs of changes arise, examples of this are the automatic balancing mechanism and the price-indexing of the guarantee pension; the political consensus of 1994 has been upheld too long and the system is “pro- foundly undemocratic”, the transfer of an important part of the funds in the former system to the state budget , too much of the risks have been transferred from the state to the individuals, the replacement ratio is too low, the price-indexing of the Guarantee Pen- sion lacks in the long run fairness and credi- bility and the possibility to raise contributions wit- hout granting new pension rights should not be precluded. Is broad political consensus undemocratic? This is Mr. Scherman’s most important objec- tion, which arose with the introduction of the automatic balancing mechanism. 17 The New Swedish Pension System – a Fair and Sustainable Model The reform was reached through an agree- ment between five political parties represen- ting today more than 85 % of the voters. All changes have been made with the same major- ity. The two parties who disagree with the reform also disagree more with each other on their alternative solutions than with the re- form. When Mr. Scherman describes the new Swedish system as “undemocratic”, he is in- venting a new theory in the political science. What Mr. Scherman seems to mean is that it is time for the five parties to terminate the agreement that changes in the new pensions system can be made only when the five parties agree to do that. The idea seems to be that it is more democratic with political fights and maybe new temporary agreements than the decision to stick to the agreement of 1994 and make necessary changes in agreement. That position can of course be argued, but to descri- be what a large majority in Parliament has agreed upon as ”undemocratic” is of course extreme. Many countries including Sweden have tried – and still try – Mr. Scherman’s model for decision-making in this field. The old Swe- dish model, the ATP-system, was introduced in 1960 after the most dramatic political battle after the Second World War. That fight inclu- ded one referendum, the only extra parlia- mentary election since the victory of demo- cracy and a final decision in Parliament by the slim majority of one abstained vote. Mr. Scherman himself writes in his article that the original agreement of 1994 probably was necessary to break the deadlock that he himself criticized heavily during the eighties. The new system with agreed rules including the automatic balancing mechanism (“brake and gas pedals”) instead of ad hoc decisions is more likely to result in smooth rather than abrupt changes when changes are needed. What is now happening in many European countries when they belatedly decide on chan- ges in their pension systems illustrate this. The recent history of Sweden also points in the same direction. Even if Sweden was ear- lier than almost all other developed countries in changing the pension system, the decision of 1994 should of course have been taken much earlier. A pension system that gives a smooth ad- justment to significant economic fluctuations means far less of risks for unnecessarily pro- voking conflicts between generations. This is of no small value for a democratic society that, in a quickly changing world, needs to simul- taneously handle many complicated and con- troversial issues. With traditional pension sys- tems a great risk is that adjustments that are necessary are made too late and made under conflict. Is it unfair to transfer money from the pension funds to the state budget? In the public debate and in his article Mr. Scherman has heavily criticized the transfer of some 30 % of the old funds and the discus- sion of transferring a further 10 % to the budget. He is of course right when he claims that any transfer from the buffer funds weakens the economic basis of the system, but he usually “forgets” to mention that the ATP- funds were built up both for the old age pensions and for the disability pensions. And that the reform of 1994 as one of its features had the separation of the two systems. The economic responsibility for the invali- dity pensions has since some years been trans- ferred to the general budget. That budget also has the financial responsibility to pay fees for non-contributory pension rights for child care, military service and higher education. The institutions for sickness and unemployment insurance are also responsible for financing pension rights for insured time for sickness absenteeism and unemployment. 18 The New Swedish Pension System – a Fair and Sustainable Model The transfer of money that has been made so far is less than the transfer of economic responsibility. This will continue to be the case, even if a further – and final – transfer will be made as of the first of January next year. This is one of the two most important objec- tions Mr. Scherman has against the new sys- tem. If the two systems had not been separated – which was a good but not a necessary thing – would he then have argued in favour of transferring up to some 15 % of the Swedish GDP from the taxpayers to the pension funds? This is the economic equivalent of what Mr. Scherman argues. Are the risks unfairly divided bet- ween the state and the individuals? There are very many risks in the always uncer- tain future. And, I hope, many chances and opportunities. The idea of an insurance sys- tem is to spread the cost of risks among those insured. The risks in a mandatory system is of course much more widely spread that in voluntary systems. In the Swedish system all income- earners and all pensioners are included. The pension rights and the pensions in the PAYG are indexed by the change in the average wage, with the exception of the Guarantee Pension. (See further down.) A (minor) part of the fees are funded and can be used to buy equities, thereby spreading the risks with a system completely tied to what happens with the wages. The level of new annual pensions is “indexed” to changes in longevity. The automatic balancing mechanism takes into account all changes in society that affects the capacity to pay pensions. In sum, it can be argued that as the system has been implemen- ted and designed it spreads the risks very broadly among a very big part of the Swedish population. But what Mr. Scherman says that he wants is to have a different sharing of the risks between the insured and the state. But what is the state? Has the state different sources of income than the mandatory pension system has? Is the replacement ratio too low? The new pension system is a defined contribu- tion (DC) system. The higher the fees, the higher the replacement ratio. Mr. Scherman has – as far as I know – not demanded an increase of the fees from the present 18,5 %, but in his article his tone is critical about the replacement ratio. In a comment on calculations from the National Social Insurance Board he points to “one scenario” in which a reduction of around 15 % of the PAYG-pension could happen.. He forgets to mention that this scenario is the most pessimistic of the 72 (!) that was calcu- lated. Some objections can also be made to his way of describing the replacement ratio. First- ly he assumes that the income of an individual will increase all the way to retirement. That is not so for the average individual. The increase of wages between 55 and 65 years is lower than the average increase. Secondly he claims that the assumptions on the return on investments in the funded part of the system are “fairly optimistic”. He does not mention what they really are: 3,25 % in real terms. I would guess that the large majority of international experts would consider this assumption to be fairly pessimistic rather than fairly optimistic. More important than these rather technical points is whether the pension reform should have included higher fees and thereby a hig- her replacement ratio. First the facts about the contribution rate. The mandatory fee is 18,5 %. On top of that some 90 % of all Swedish wage- earners have an occupational pension that gives some 10 percentage points of the wage 19 The New Swedish Pension System – a Fair and Sustainable Model on top of the public pension. The average fee of that pension is at least 4 %. That means that the large majority of Swedish wage-earners pay 22-23% of their wage in order to get pensions when they stop working. The main costs of pensions are covered by those fairly high fees, but on top of them also come what is paid by general taxation, especi- ally for the guarantee pension and for pension rights for child care years. That cost is equal to an additional amount comparable to a contri- bution rate of 2 percentage points. It can also be argued that the old age pension fees for the sickness insurance, the unemployment insu- rance and the sickness pension should be considered as pension fees rather than as a part of the cost of those insurances. Summing up, an ordinary Swede pays every year some 25 % of his or her wage in order to get a pension. It is possible to argue that a political majority in Parliament ought to in- crease that level further, by taking money from the active years when people build fami- lies, raise their children and in most cases buy their own homes, to the years when they are retired. The calculation is rather easy to make. If we want to increase the replacement ratio by some 10 percentage points, the mandatory contribution rate has to be increased by some 3 percentage points every year of active life. That will mean that the level of mandatory or quasi-mandatory pension fees would increase to nearly 30 % of the wage. It can be done, but is it advisable? For the debate between Mr. Scherman and me, it would be important to hear whether Mr. Scherman proposes higher mandatory fees or just feel satisfied with criticizing the level of pensions. The latter is maybe acceptable for a pensioner who considers his or her pension too low, but maybe not for an Honorary Pre- sident of the International Social Security Association (ISSA). Is price-indexing the right method for the Guarantee Pension? Mr. Scherman points correctly to the fact that the new Swedish system is wage-indexed with the exception of the Guarantee Pension. He argues also correctly that the general wage- level and the level of Income Pension (“in- komstpension”) with a yearly increase of 2 % will be doubled in 35 years. And thereby will the relative value of the Guarantee Pension be considerably lowered. His own conclusion is that the level must be reviewed and adjusted from time to time. I agree with that, but he does not mention that the price-indexing is also a shelter for those pensioners with lower pensions during peri- ods of economic problems and lowered real wage levels, such as the first half of the nineties in Sweden. Through price indexation pensioners with the lowest pensions are gua- ranteed that the real value of their pension will not be lowered. Would Mr. Scherman like to change that? Should it be possible to raise fees without giving new pension rights? There are some misunderstandings about the NDC model. Among these are that it is inhe- rent in the model that the contribution rate can never be changed. That is not so. What is inherent is that fees paid always must result in pension rights. Most of the debate on this point, including Mr Scherman’s present remarks, has touched on the fact that financial lack of balance can- not be remedied by raising the fees, as they will result in an increase of the pension rights, that is of the liabilities of the system. This is correct even if it will be a rather long period before the increased pension is paid out. If there is a widespread feeling that the future replacement ratio will be too low – and that many people are too myopic or too poor 20 The New Swedish Pension System – a Fair and Sustainable Model to save more themselves – it is of course possible for Parliament to raise the contribu- tion rate from 18,5 % to – say – 20 %. For new entrants into the labour force this will increase the replacement ratio by some five percentage points. It is always possible for Parliament to in- crease the value of the Guarantee Pension or, for example, the value of pension rights for years of child care. Changes of this nature are entirely within the framework of the NDC system, and all Parliament has to do is pay for them with budget revenues. The advantage of the NDC framework is that these decisions are made explicit and transparent, and can be weighted against other needs. Mr. Scherman advocates that it should be possible to increase the fees without any in- crease of the pension rights, supposedly to remedy a lack of financial balance. Such a possibility would change the system from a NDC model to something different, maybe to a “reformed defined benefit scheme”. I am against that. The aim of Mr. Sherman on this point is obviously to create a possibility to increase the pensions. History is perhaps a guide. During the almost 40 years when Sweden had the old system the real value of the ATP-pensions was never raised by political decisions. During the short period when the new system has been in place, the real value has increased by 2,4 %. Concluding remark The main aims of the 1994 reform were to make the mandatory pension system, more sustainable, more fair – and to improve the incentives to work. I think that has been done. Many developed countries need to reform their pension systems. Some of them have used some ideas from the Swedish Pension Reform. Several others discuss features of that reform. The pronounced intention of Mr. Scherman’s article is to warn them from doing this. Among his reasons for that is that he considers the Swedish replacement ratio low and want Sweden to transfer large responsibi- lities to the state without any adequate finan- cial compensation. I doubt whether he is doing the countries now discussing pension reform a service. Notes 1 Pensionsarbetsgruppen: Ett reformerat pensions- system – Bakgrund, principer och skiss. Ds 1992:89. (Only in Swedish). 2 E.g. Michael Cichon: Notional defined-contri- bution schemes: Old wine in new bottles? ISSA Review, 4/1999. 3 NFT 2/2003 Könberg, B: Pensionsreform med sunda principer. (In Swedish. An English trans- lation can be ordered from the author.) 4 NFT 4/2003 Scherman, KG: The Swedish pen- sion reform: a good model for other countries? 5 Scherman, KG: The Swedish pension reform. ILO Discussion paper 7 (1999).