Longevity is increasing – what about the retirement age?

26 Longevity is increasing – what about the retirement age? NFT 1/2004 However, worldwide experience shows that usually calculations concerning the popula- tion projections have failed. Life expectancy has in most countries been growing faster than projected even if life expectancy in some developing countries has decreased rapidly due to AIDS. According to United Nations statistics, the life expectancy at birth grew during the past five decades in the world by 20 years, in Europe by 8 years and in Finland by 12 years. When designing or analysing a pension scheme, more important than the life expect- ancy at birth is the life expectancy at the entrance to the labour market, say 25 years, and the life expectancy at the entrance to retirement, say 65 years. The difference be- tween the life expectancy at birth and the life expectancy for a certain age is visualized using statistics concerning life expectancy in Finland. During the first half of the last centu- ry, the mortality rates for young people in Finland decreased very rapidly, as shown in the graph on the left in chart 1. As the total life expectancy at birth increased by 19 years, three thirds of this increase was due to de- creased mortality rates for young people aged less than 25, and only one tenth of the increase was due to decreased mortality rates among elderly people aged 65+. For the last half of the century, the situation changed significant- ly. As the life expectancy at birth increased by 12 years, only one third was due to decreased mortality rates among young people under 25 Longevity is increasing – what about the retirement age? by Christina Lindell Pension schemes and life expectancy In most countries the continuously decreasing mortality trend and thus the corresponding increasing life expectancy has continued for a very long time. When looking at the pension expenditure in the long term, it is essential whether this trend will continue as in the past or whether it is slowing down or stopping. When making long-term projections concerning the cost of a pension scheme, the fluctuation and uncertain- ty concerning the economical assumptions are well-known but the demographic and especially the mortality fluctuation has often been considered to be under control.Christina Lindell christina.lindell@etk.fi Christina Lindell is Head of Department at the Planning and Actuarial Department of the Finnish Centre for Pensions. This article is based on a report on the debate on the retirement age which preceded the pension reform in Finland and an article written for the research conference in 2003 arranged by the International Social Security Association (ISSA). 27 Longevity is increasing – what about the retirement age? and almost half of the change was due to decreased mortality rates among elderly peo- ple aged 65+. The life expectancy was 11 years until 1940s for people aged 65 as is shown on the right hand graph of chart 1. The life expectancy began to grow in the 1940s, but not very rapidly. Thus, when designing the pension schemes in Finland in the 1930s and 1950s there was not much use for past statistics when projecting future life expectancy and thus the future pension costs. A rough estimate shows that with fixed retirement ages and an average time of 20 years in retirement (including early retire- ment) the pension expenditure grows by 5 per cent for each year the longevity of pensioners aged 60+ is growing. Thus, using the present falling mortality trend, the life expectancy for old-age pensioners may continue to increase by one year per decade and thus the pension expenditure by 25 per cent during the next 50 years. As the costs of the pension schemes in most countries are increasing also due to many other reasons than longevity (e.g. the baby- boomers born after the Second World War are reaching the retirement age and the pension schemes reach maturity), the question of ad- justing the retirement age to the changes in life expectancy will in most countries arise sooner or later. There are of course a lot of reasons why people retire early, such as labour market reasons, poor health, burnout, stress and other problems influencing the atmosphere at the workplaces. Therefore a successful postpon- ing of the effective retirement age also re- quires a co-operation between social, health and labour authorities and between employ- ees and employers. The aim of this paper is to sort out technical alternatives of adjusting the retirement age to the changes in life expectan- cy. The main attention is paid to old-age pensions and a brief review of the situation in selected countries is included. Adjusting the pension scheme to increasing longevity by gradually raising the set retirement age The traditional way of adjusting the pension scheme to increasing longevity is to raise the set retirement age. The raise can be done by raising the retirement age once and later de- cide if there are need for additional raises, or by agreeing on a plan of successive raises of the retirement age, or it can be tied to a suitable indicator, like keeping the proportion between working years and years in retirement un- changed. The first is probably the most com- mon alternative but it is actually a special case of the second one, which we shall look at in more detail in the following. The last alterna- 10 13 16 19 22 25 1870 1890 1910 1930 1950 1970 1990 2010 2030 2050 R ema in in g l on ge vit y StatFin,2002 Eurostat,1998 SII, Fin, 1998 tre nd R ema in in g l on ge vit y 76 78 83 00000000000000000000000000000000000 00000000000000000000000000000000000 00000000000000000000000000000000000 00000000000000000000000000000000000 00000000000000000000000000000000000 00000000000000000000000000000000000 00000000000000000000000000000000000 00000000000000000000000000000000000 000000000000000000000000000000000 000000000000000000000000000000000 000000000000000000000000000000000 000000000000000000000000000000000 000000000000000000000000000000000 000000000000000000000000000000000 000000000000000000000000000000000 000000000000000000000000000000000 000000000000000000000000000000000 000000000000000000000000000000000 00000000000000000000000000000 00000000000000000000000000000 00000000000000000000000000000 00000000000000000000000000000 00000000000000000000000000000 00000000000000000000000000000 00000000000000000000000000000 00000000000000000000000000000 00000000000000000000000000000 00000000000000000000000000000 00000000000000000000000000000 00000000000000000000000000000 79 65 70 000000000000000000000000000000000000 000000000000000000000000000000000000 000000000000000000000000000000000000 000000000000000000000000000000000000 000000000000000000000000000000000000 000000000000000000000000000000000000 000000000000000000000000000000000000 000000000000000000000000000000000000 000000000000000000000000000000000000 000000000000000000000000000000000000 000000000000000000000000000000000000 000000000000000000000000000000000000 000000000000000000000000000000000000 000000000000000000000000000000000000 000000000000000000000000000000000000 000000000000000000000000000000000000 000000000000000000000000000000000000 000000000000000000000000000000000000 000000000000000000000000000000000000 000000000000000000000000000000000000 000000000000000000000000000000000000 000000000000000000000000000000000000 000000000000000000000000000000000000 000000000000000000000000000000000000 000000000000000000000000000000000000 000000000000000000000000000000000000 000000000000000000000000000000000000 000000000000000000000000000000000000 000000000000000000000000000000000000 0000000000000000000000000000000000 0000000000000000000000000000000000 0000000000000000000000000000000000 0000000000000000000000000000000000 0000000000000000000000000000000000 0000000000000000000000000000000000 0000000000000000000000000000000000 0000000000000000000000000000000000 0000000000000000000000000000000000 0000000000000000000000000000000000 0000000000000000000000000000000000 0000000000000000000000000000000000 0000000000000000000000000000000000 0000000000000000000000000000000000 0000000000000000000000000000000000 0000000000000000000000000000000000 0000000000000000000000000000000000 0000000000000000000000000000000000 0000000000000000000000000000000000 0000000000000000000000000000000000 0000000000000000000000000000000000 0000000000000000000000000000000000 0000000000000000000000000000000000 0000000000000000000000000000000000 0000000000000000000000000000000000 0000000000000000000000000000000000 0000000000000000000000000000000000000 0000000000000000000000000000000000000 0000000000000000000000000000000000000 0000000000000000000000000000000000000 0000000000000000000000000000000000000 0000000000000000000000000000000000000 0000000000000000000000000000000000000 0000000000000000000000000000000000000 0000000000000000000000000000000000000 0000000000000000000000000000000000000 0000000000000000000000000000000000000 0000000000000000000000000000000000000 0000000000000000000000000000000000000 0000000000000000000000000000000000000 0000000000000000000000000000000000000 0000000000000000000000000000000000000 0000000000000000000000000000000000000 0000000000000000000000000000000000000 0000000000000000000000000000000000000 0000000000000000000000000000000000000 0000000000000000000000000000000000000 0000000000000000000000000000000000000 0000000000000000000000000000000000000 0000000000000000000000000000000000000 0000000000000000000000000000000000000 0000000000000000000000000000000000000 0000000000000000000000000000000000000 0000000000000000000000000000000000000 0000000000000000000000000000000000000 0000000000000000000000000000000000000 0000000000000000000000000000000000000 0000000000000000000000000000000000000 0000000000000000000000000000000000000 0000000000000000000000000000000000000 0000000000000000000000000000000000000 78 47 66 25 0 45 65 75 85 1901-1910 1951-55 2001 A ge Life expectancy at birth, at ages 25 and 65 Life expectancy at age 65 including projections Chart 1. Changes in life expectancy in Finland 28 Longevity is increasing – what about the retirement age? tive is rather close to the use of an adjustment indicator, which will be discussed later. When raising the retirement age succes- sively according to a predetermined plan, it may be carried out by raising the retirement age by a whole year. The problem connected with this type of raise is that the difference in retirement age may be a whole year for people born in December and January, even if the difference in age is as small as one day. To avoid too big differences in retirement ages for two successive cohorts the retirement age could be tied to a monthly level for each cohort. As the life expectancy for elderly has increased by 4 years since the 1960s, a suita- ble raise could be to raise the retirement age by one year per decade. A more moderate crite- rion would be to raise the retirement age by one month per cohort. The greatest problem with raising the set retirement age according to a predetermined plan is connected to the uncertainty related to the changes in mortality rates. The target is that the retirement age is set already before a person enters the labour market and main- tained unchanged until retirement age. When looking at the past 40 years the changes in mortality rates have been much greater than expected. If mortality rates in the future de- crease more than expected, the retirement age need to be adjusted further and if mortality rates decrease less than expected, the raises already carried out may be oversized. Raising the retirement age is closely related to the accrual rates in a defined benefit scheme. When designing a pension scheme, usually a target level of the pension is fixed. If the accrual rate remains unchanged it means that the target level of the pension increases. Paid pension contributions may however justify the increase. Probably the greatest problem concerning unchanged accrual rates is con- nected with early retirement. Disability and other pensions, including projected pensiona- ble service, would increase with unchanged accrual rates, and that is hardly the intention. An alternative would be to decrease the accrual rates in proportion to the raise of the retirement age, but cutting the past accrual rates would be difficult to justify. A further alternative would be to maintain the past ac- crual rates unchanged but change the future accrual rate with the intension to maintain the target level unchanged. This alternative may however result in separate accrual rates for each cohort. Changes in retirement ages in selected countries In most countries the retirement ages in the statutory pension schemes are preset. Previ- ously changes in retirement ages were not so common. The trend was more to offer path- ways for early retirement besides the standard retirement age. Since the 1990s the trend has changed. Countries are looking for solutions to decrease the effect of increased longevity on pension costs. Only in three countries, Denmark, Iceland and Norway, the retire- ment age is over 65 and only one country, the USA, has decided to raise the retirement age beyond 65. In 2003 it was also proposed in Germany that the standard retirement age would be raised gradually from 65 to 67, but this proposal was rejected by the Govern- ment. The most common changes are to equal- ise the retirement age for men and women and to raise the retirement age up to 65. Excep- tions are Denmark and Finland1 . In both countries the retirement age is lowered to- gether with tightened conditions for early retirement. The aim of lowering the retire- ment age is, however, to raise the average effective retirement age. Table 1 gives an overview of selected countries where the re- tirement age is higher than 65 years and coun- tries where the retirement age is changing. Changes in early or deferred retirement ages are not included. 29 Longevity is increasing – what about the retirement age? Table 1: The retirement age in selected countries Country Current Changes in retirement ages retirement age Austria 65(M) 60(F) 60 -> 65 (2024-2033) (F) Denmark 67 65 (1.7.2004) (national pension) Finland 65 63-68 (2005) (earnings-related pensions) Greece 65 60->65(2007 -> ) (F old system) Great Britain 65(M), 60(F) 60->65 (2010-2020) (F) Hungary 62(M), 59(F) 55(1996) ->62 (2009) (F) Iceland 67 - Italy 65(M), 60(F) 1) 57-65 (gradually phased out in 19 years) (gradually implemented in 19 years) Japan 60 60->65 (2013-2025) 2) Norway 67 - Poland 65(M), 60(F) - 3) Switzerland 65(M), 63(F) 63->65 (2009) (F) United States 65y 4m 65->67 (2003) (2027) 1) In the current system one can retire after 35 years of coverage at age 57, regardless of age after 38 years of coverage. The years of coverage rise to 39 years in 2006 and to 40 years in 2008. 2) The retirement age for basic pensions in Japan is already raised to 65 by 2013. 3) According to a government bill in Poland the retirement age for women will in 2009 start to raise gradually to 65. The raise is suggested to be 6-9 month per year. Adjusting the pension scheme to increased longevity by a factor Also other solutions than changing the retire- ment age are looked for in order to keep the pension costs in check. One method recently spread is to adjust pensions to increased lon- gevity instead of raising the set retirement age. Often the adjustment is combined with flexible retirement ages and forces the insured to make a choice: retire at the same age as earlier cohorts with a slightly reduced pension or receive an unreduced pension by continu- ing to work a little bit longer. A factor suitable for adjusting both defined benefit and defined contribution schemes to increased longevity can be found using actu- arial mathematics. Define the probable present value of a pension at retirement age as the value of a lump sum sufficient to finance the future pension expenditure, taking into ac- count the life expectancy and a supposed yield from investing the lump sum. The present value thus depends on two main components, the mortality rates and the discount rate. The discount rate is in the long term reflecting the difference between the average yield of the lump sum and the average index used for adjusting the accrued pension rights. The adjustment indicator, called a longevi- ty factor, is now the probable present value of a unit pension (eg. a pension of one euro per year), which is regularly recalculated using new information on mortality rates. Mortality rates and life expectancy are often calculated separately for males and females. However, only one shared factor for both genders is developed, because in a statutory pension scheme the benefits are not allowed to be determined gender-specifically. If the discount rate used is zero, the longevity factor is equal to life expectancy. Thus the value of the longevity factor changes with the age of cal- culation: the higher the calculation age, the smaller the factor. The theoretical background and formulas of the adjustment factor is de- scribed in Lindell [2004]. 30 Longevity is increasing – what about the retirement age? The longevity factor is used in a different way depending on whether the pension scheme is a (notional) defined contribution ((N)DC) scheme or a defined benefit (DB) scheme. Roughly the two schemes may be described as follows. In (N)DC pension schemes an accu- mulated (notional) pension capital is changed into a series of payments or life annuities by dividing the capital with the probable present value of a unit pension. In DB schemes the payments (accrued pensions) or life annuities are known, while the probable present value is calculated by multiplying the accrued pen- sions by the probable present value of a unit pension. Adjusting the pensions to increased longev- ity in DC and NDC schemes is simply achieved when changing the accumulated (notional) capital into life annuities by dividing the cap- ital with the longevity factor (or equivalently by multiplying with the inverse value of this factor) calculated at the effective age of retire- ment. In (N)DC pension schemes the longev- ity factor automatically takes into account early pension reductions and deferred pension increases. In DB schemes adjusting pensions to increased longevity is achieved by multi- plying the accrued pension by a coefficient, which is the quotient of two longevity factors calculated at the set retirement age at a base year t0 , which is the year when the longevity factor was introduced, and at year t, which is the year when the insured reaches the set retirement age ( or equivalently by dividing by the inverse value of this quotient)2 . In DB pension schemes, where the pension may be taken earlier or postponed, the longevity coef- ficient may further be developed to automat- ically take into account also early pension reductions and deferred pension increases. In these calculations the longevity coefficient is calculated as a quotient of two longevity fac- tors at the effective retirement age in year t given the set retirement age and year t0. Recalculating the longevity factor using statistical or projected mor- tality rates? An adjustment factor actually should take into account the past, present and future changes in mortality rates. But as a consequence of the unexpectedly rapidly falling mortality rates it is very difficult to make reliable projections concerning life expectancy. It seems therefore to be unfair if possible fails in mortality pro- jections would affect the pension level. An alternative is to recalculate the longevity fac- tor only with observed mortality statistics. In practice it is easy and transparent to recalcu- late a longevity factor based on statistics, because the statistical office in each country already produces life and mortality tables. But statistics always describe the past, with the consequence that the longevity factor follows the actual mortality trend with some lag. By recalculating the pensions each year accord- ing to new observed changes in mortality rates, the lag could be minimized. An easier and more obvious way is to make the changes only once when the pension is granted. Even if such a factor reflects the changes in mortal- ity with some portion of lag, it does not play a very significant role in a PAYG scheme. Mortality has decreased for decades but the adjustment factor is used from a certain year onwards. As a consequence of updating the longevity factor by statistics and applying it only once per person, increased longevity influences not only increased working years (or a reduced pension) but to some extent also increased years in retirement. Applying the longevity factor on disability, unemployment and other early pensions As a starting point the longevity factor is applied on the whole population when reach- ing the retirement age. The disability pension could also be adjusted already when granted, 31 Longevity is increasing – what about the retirement age? but would such a pension benefit be enough especially for breadwinners with children? The situation for these early pensioners are on, one hand, to some extent contradictory, because their possibilities of increasing their old-age pension by continuing to work a little bit longer are very limited. On the other hand, it is difficult to leave the early pensioners outside the adjustment system at least when they reach the retirement age. The incentive for the active population to work longer de- creases if there is a possibility of avoiding pension adjustment by receiving some type of early benefit before the old-age pension. The nearer the old-age pension, the more difficult it also is to distinguish between who is disa- bled and who is not. Adjusting the pension scheme to increasing longevity in selected countries The pension schemes using the adjustment method are all defined contribution (DC) or notional defined contribution (NDC) schemes. But the adjustment method may as well be applied in a defined benefit (DB) scheme. In Finland, the laws including an adjustment method passed Parliament in February 2003 and in Norway a similar adjustment method is in January 2004 proposed by the pension committee. The adjustment method can be either automatic, the adjustment factor can be set once or a decision may be made to recalcu- late the factor at certain intervals. An example of the first method is Sweden and of the latter is Italy. Sweden reformed its statutory pension scheme in 1999 and it will be gradually imple- mented. Sweden changed its scheme from a defined benefit scheme to a mainly notional defined contribution scheme. In this new scheme, the main part is financed using PAYG and the pension rights accrue according to paid contributions. The contributions accu- mulate during the working life to a notional pension capital. The pension can be with- drawn beginning from age 61. The notional pension capital is changed to monthly life annuity payments by dividing it by the lon- gevity factor. This factor is determined sepa- rately for each cohort upon retirement using the latest available statistics on mortality rates and a discount rate of 1.6%. Similar methods are also used in Poland and Lithuania. In Italy pension rights are accruing accord- ing to a notional contribution into a notional pension capital. At retirement age the notional capital is changed to life annuity payments by multiplying the pension capital with a factor which equals the inverse value of a longevity factor. The factor is fixed for each retirement age (57-65). Unlike the Swedish scheme, the factor may be changed only every ten years by a decision of the Ministry of Labour and Social Policy in order to take into account changes in longevity and GDP. In Switzerland a reform is underway, where the mandatory occupational DC pension is changed. Today the accrued pension capital is multiplied by 7.2%, which equals the inverse value of a longevity factor. The plan is to lower the percentage due to increased longevity. Beginning in 2005, statutory earnings-re- lated pensions in Germany will be adjusted according to a new sustainability factor. This factor will take into account the relationship between the number of pensioners and the number of contributors to the system. The factor will have the effect of reducing the annual pension adjustment if the ratio of pen- sioners to contribution payers changes to the detriment of the contribution payers. The new sustainability factor will thus in addition to the life expectancy take into account the birth rate, immigration and emigration and the la- bour force participation rate. Compared to the longevity factor, this sustainability factor is not cohort-specified. 32 Longevity is increasing – what about the retirement age? In Norway the pension committee proposed in January 2004 a flexible (62-70) retirement age and an automatic adjustment of the Nor- wegian defined benefit pension scheme to increased longevity. The longevity coeffi- cient is suggested to be calculated as the inverse value of the quotient of two longevity factors described earlier and thus adjusting pensions to longevity is achieved by dividing the pensions with this coefficient. The pro- posal also includes an automatic adjustment to longevity of early retirement reductions and deferred retirement increases. The Finnish pension reform 2005 and adjusting the scheme to changing longevity The main goals of the reform are to postpone the average effective retirement age by 2-3 years, to adjust the pension scheme to in- creased life expectancy, to minimize the need to raise the contributions and to support the ageing population’s well-being at work. The retirement age will become flexible between ages 62 and 68. The accrued old-age pension will be granted without reduction between the ages 63 and 68. Only pensions granted at the age of 62 will be reduced and in case retirement is postponed past the age of 68, an increment of 0.4% per month will be granted. At the same time pathways to early retirement will be blocked. Working an extra year today implies foregoing one year of pension and paying additional contributions, with often little or no increase in fu- ture pensions. Therefore the ac- crual rate is raised to 4.5% per year if a person continues work- ing beyond the age of 63, while the normal accrual rate will be 1.5% per year. The high accrual rate also justifies the lack of deferred coefficients between the ages 63 and 68. Additional working years will also be made financially worthwhile by abolishing the 60% ceiling of the accrued pension. The pension scheme will be further adjust- ed to increased life expectancy by introducing a longevity coefficient. The aim of the coeffi- cient is to reflect part of the increase in life expectancy in the number of working years. This means that, starting from the year 2010, the amount of new old-age pensions will depend on the development of life expectancy compared to the year 2009. Only one longev- ity coefficient is determined for each year. It is always calculated for the cohort which turns 62 and it will be fixed for this cohort irrespec- tive of the retirement age. Also disability pensions will be adjusted by the longevity coefficient at age 63. The discount rate will be 2%. The decision to leave early pension re- ductions and deferred pension increases out- side the longevity adjustment as well as the decision to fix only one coefficient per cohort irrespective of retirement age was made on purpose. The adjustment to longevity is per- haps a bit inexact, but the adjustment system will be easy to achieve and transparent. Chart 2 shows that using projected mortal- ity rates from either Eurostat or StatFin the longevity coefficient will in the future de- Chart 2. Longevity coefficient using projected values of mortality by Eurostat and StatFin. 0,85 0,87 0,89 0,91 0,93 0,95 0,97 0,99 2009 2019 2029 2039 2049 Eurostat, 1998 StatFin, 2002 33 Longevity is increasing – what about the retirement age? crease by 11-15%. But there is a large portion of uncertainty related to the mortality projec- tions. The longevity coefficient is however decreasing very slowly. For a person with average income, the pension for each new cohort is decreasing by 4-6 euros per month. The additional working time needed to com- pensate for the decrease caused by the longev- ity coefficient is about 2-3 weeks for every new cohort. Table 2 shows that the additional working time needed to compensate for the longevity coefficient in the long term, around 2050, is not more than one and a half years. However, according to the Eurostat popula- tion projection, the life expectancy at age 62 increases by 3-4 years during the same period. The short additional working time compared to the increase in life expectancy is explained by the triple accrual coefficient for people working in the age bracket 63 to 68. Table 2. The longevity coefficient for selected cohorts and the additional working time needed to compensate for the coefficient Additional working time compensating for the coefficient Year of birth Year of retire- ment Longe- vity coeff. Accrued pension 50% wages Accrued pension 60% wages 1957 2020 0.956 5 m 6 m 1967 2030 0.917 11 m 1 y 1 m 1977 2040 0.892 1 y 2 m 1 y 5 m 1987 2050 0.880 1 y 4 m 1 y 7 m Long-term pension projections made by the Finnish Centre for Pensions use a population projection based on the one made by Eurostat. Chart 3 shows that in the year 2050 the pen- sion expenditure without a reform increases to about 36% of the wage sum, while the reform decreases the expenditure by 4.3 per- centage points. The effect of the longevity coefficient is about 2.5 percentage points. At this stage it is very difficult to assess how the insured will react to the flexible retirement age. Will everyone retire at 63 or will the triple accrual rate induce people to continue work- ing? The sensitivity of the pension expendi- ture as a percentage of the wage sum was tested in relation to the choice of retirement age. For the test, the assumption was that everybody will retire at 63, at 68, or between the ages 63 and 68. As seen from the right-hand figure of chart 3 the result was that the choice of retirement age was not very significant with regard to pension expenditure as a percentage of the wages. If everyone retires at 63, the pension expenditures rise in the beginning but de- crease later, compared to the alternative of everyone retiring at 68. In general it is possi- ble to say that the flexible retirement age is cost neutral as regards the pension expendi- tures in relation to the wage sum. Chart 3. Pension expenditures in per cent of the wage sum for private-sector employees and the sensitivity in regard to the choice of retirement age 0 5 10 15 20 25 30 35 40 45 50 2000 2010 2020 2030 2040 2050 % of wage bill Current legislation Reform 15 17 19 21 23 25 27 29 31 33 35 2000 2010 2020 2030 2040 2050 Year % of w ag e bi ll 63 Estimate 68 34 Longevity is increasing – what about the retirement age? Notes 1 The Finnish reform is described later in this paper. 2 The purpose is to stress that the terms dividing and multiplying are used in both DC and DB pension schemes. The difference between the adjusting methods are that in a DB scheme the starting value is 1 (the quotient of two identical longevity factors) and this value either increases or decreases in time depending on whether the adjusting is carried out by dividing or multiply- ing the pension with the longevity coefficient. In a (N)DC scheme there is no starting value. Dividing the accumulated capital with the lon- gevity factor, say 15, is equivalent to multiply- ing the capital with the inverse value 0.067 or 6.7%. References Introduction to the Hungarian Pension System. Central administration of national pension in- surance (ONYF)- Hungary. Kannisto V. – Nieminen M., Revised Life Tables for Finland 1881-1990, Population 1996:2, Sta- tistics Finland, 1996. Laitinen-Kuikka S.-Bach J.-Vidlund M., Eläketur- va Länsi-Euroopassa. Finnish Centre for Pen- sions, 2002. Lag om inkomstgrundad Ålderspension, Govern- ment Proposal to Parliament 1997/98:151. Swe- den. Life expectancy, statistics and projections, United Nations, 2001. www.un.org. Lindell Christina, Longevity is increasing – what about the retirement age? Finnish Centre for Pensions. Working Papers 6. Finland, 2004. Lindell Christina, Elinaika pitenee – miten käy eläkeiän? Finnish Centre for Pensions. Report 18. Finland, 1999. Modernisert folketrygd - bærekraftig pensjon for framtida. Norges offentlige utredninger, NOU. 2004:1. GAN Grafisk AS, Oslo (www.pensjonsreform.no). Mortality rates, statistics and projections, The Social Insurance Institute. Finland, 1998. Mortality rates, statistics and projections , Statis- tics Finland, 2002. Mortality rates, projections, Eurostat, 1998. Nachhaltigkeit in der finanzierung der sozialen sicherungssysteme, Bericht der Kommission, Germany, August 2003, www.Soziale-sicher- ungssysteme.de/index.html. USA: Social Security Administration, SSA. www.ssa.gov. ZUS, Social Insurance institution. Social insur- ance in Poland. Information, facts. Warsaw 2002. www.zus.pl.