This chapter discusses policy options for Slovenian public pensions, based on the analyses developed in Chapters 1 to 3, and makes policy recommendations. It looks into how the rules of earnings-related pensions could be improved, making them more transparent and aligning them between various groups of workers. It shows how to improve pension financial sustainability and make the system more resilient to population ageing. The chapter also discusses options to improve the minimum pension and safety-net benefits for older people.
OECD Reviews of Pension Systems: Slovenia
OECD Reviews of Pension Systems
4. Reform options to improve public pensions
Abstract
4.1. Introduction
The pension system in Slovenia has evolved substantially over the last two decades through changes in pension parameters. The replacement rates for average earners are close to the OECD average while low-wage workers benefit from a strong redistribution. The contribution rate is higher than on average across OECD countries and has remained constant since 1996. In response to the COVID‑19 crisis in 2020, Slovenia introduced job-retention schemes and options to subsidise or defer pension contributions. As in all OECD countries, Slovenia had to deal with very volatile contribution revenues with a large slump in the second quarter of 2020. While it is still early to assess the long-term impact of the COVID‑19 crisis on the pension system, the analysis in Chapters 1, 2 and 3 highlighted different ways to improve public pensions in Slovenia.
The main weaknesses of the current system are threefold. First, eligibility conditions for earnings-related pensions are loose, whereas they are relatively restrictive for safety-net benefits. There is no planned legislation to change these conditions while pension expenditure is projected to rise strongly. Second, the calculation of contributory benefits is complicated and poorly co‑ordinated with safety-net benefits. Pension entitlements are unclear to workers before they are actually claimed, and even gross pensions depend on income tax rates. Third, population ageing will put a substantial pressure on pension finances, threatening sustainability. This chapter discusses reform options to improve the public pension system in Slovenia. The first section focuses on improving public earnings-related pensions, while the second section provides avenues to enhance financial sustainability. The third section deals with first-tier pensions. The last section summarises policy options.
4.2. Improving public earnings-related pensions
4.2.1. Simplifying the pension rules
The calculation of pension benefits is unnecessarily complicated. One reason is that the reference wage is based on the best consecutive 24 years of adjusted (see below) earnings. There are two issues with this method. First, it is complicated. Such a rule implies that workers do not know the pension entitlements they accrued, for example, in a given year. Moreover, workers may not know which consecutive 24 years in the whole career are the best. This makes it almost impossible for workers to know the pension level they can expect – beyond uncertainty related to the uprating of past earnings – before they actually retire, which is likely to undermine their understanding of how the pension system works.
The second issue is that, while using the best consecutive 24 years of earnings protects everyone by ignoring the remaining, less favourable years, it is more beneficial to people with strong career progressions, who also tend to have higher lifetime earnings. For a given level of spending, this rule is thus regressive, redistributing from low to high earners.
Basing pensions on the average lifetime earnings rather than the 24 best consecutive years would eliminate these unfavourable elements and greatly simplify the calculation of accrued entitlements and pension benefits. As the objective of this change is not to reduce pensions, it should be combined with raising accrual rates as needed, for example in a budget neutral way thus keeping the average pension unchanged (this would imply increasing the accrual rates by about 10%). As is the case today, the impact of career breaks on pensions should be cushioned by other instruments, i.e. granting pension entitlements for unemployment and childcare periods.
The large majority of OECD countries takes into account wages throughout the whole career for calculating the pension benefit. Recently, the Czech Republic, Greece and Norway joined this group. Exceptions are Austria (which will use lifetime earnings for people born from 1955), France, Portugal, Slovenia, Spain and the United States. France, Slovenia and Spain are the only countries using 25 years or less. France was planning to switch to lifetime earnings, but the reform plans were suspended due to the COVID‑19 crisis.
Another issue making pension calculation complicated is that the reference wage used to calculate gross pensions is based on gross earnings adjusted by a multiplication coefficient equal to the ratio of the net to the gross average wage. This means that gross pensions are calculated from wages that are expressed in a form that is close to net wages – although it is less than net wages for low earners and more than net wages for higher earners due to progressive income taxes. Hence, any adjustment to contribution and tax rates is likely to affect the coefficient, and thereby gross pensions. For example, an increase in personal taxes reduces gross pensions by lowering net wages, and then it might additionally reduce net pensions by increasing taxes paid on gross pensions. This unusual and undesirable calculation rule should be modified to ensure that, as in all OECD countries with defined benefit schemes except Hungary, the reference wage is calculated based on gross wages while adjusting accrual rates accordingly in a budget neutral way.
A third complication factor is due to the discretional annual allowance, which is granted once a year to all pensioners, but with a larger benefit for low pensions. This allowance increases old-age income and reduces old-age inequality, but it unnecessarily complicates the determination of total pension benefits. There is no need to add an additional instrument as there is already a variety of instruments (accrual rate, minimum reference wage, pension credits, etc.) that can be used to fulfil the objective pursued by the annual allowance.
Furthermore, different terms for period requirements are used to capture different aspects, but nuances are not always easy to understand. For retiring at age 60, 40 years of pensionable service without purchase are required. The pensionable service without purchase includes all work-related periods for which contributions have been paid, e.g. dependent employment, self-employment, agricultural activity, unemployment spells or parental leave. For retiring at age 65, 15 years of insurance period are needed. Insurance period is a broader term that includes all periods for which contributions have been paid, including purchase periods and voluntary contributions. Up to 5 years can be purchased at any time, which are included in the insurance period and in the period of pensionable service. When retiring based on the purchased periods, a permanent penalty of 0.3% per month missing before age 65 applies to pension benefits. For benefit calculation, all periods of pensionable service are used. These include pensionable service without purchase, purchase periods, but also other periods for which the due contributions were not paid.
There are various ways to drastically simplify the calculation of earnings-related pensions. Within a defined benefit (DB) scheme, it can be done based on a constant accrual rate applied to earnings during the whole career. This would generate the same replacement rate across all earnings levels. In order to achieve a given redistributive pattern of replacement rates – i.e. higher replacement rates for low earners – there are two main options. The first is to keep the minimum pension scheme – but adjusted to take into account lifetime earnings – currently working through the minimum reference wage. An alternative, which would facilitate the communication of pension entitlements and their understanding by workers, consists of complementing the constant-accrual-rate rule with a contribution-based basic pension. A contribution‑based basic pension is based on the contribution period but is not earnings-related. Nine OECD countries have such a scheme, including the Czech Republic and Estonia. The levels of the accrual rate and of the basic pension can be calibrated to achieve the replacement-rate pattern and therefore the redistribution across lifetime earnings that reflect social preferences (OECD, 2020, pp. 46-49[1]).
The 2016 White Paper on Pensions (MLFSAEQ, 2016[2]) suggested introducing a points scheme, which is another way to simplify the rules and improve transparency. In a points scheme, individuals earn points every year based on their pension contributions or their total earnings. This means that a generic points system results in a clear link between earnings and entitlements. Pension entitlements are computed by multiplying the number of acquired points by the point value, which is known at any point in time and follows a transparent valorisation rule. This framework provides a good basis for building confidence in the pension system as the acquired entitlements can be easily communicated to everyone at any age. Among OECD countries, public pension schemes are based on points in Estonia, Germany, Lithuania and the Slovak Republic whereas in France the planned introduction of a universal points scheme has been suspended due to the COVID‑19 crisis.
Here also the generic points scheme can be combined with a contribution-based basic pension to help achieve redistribution objectives. Alternatively, redistributive objectives can be achieved by setting the minimum and maximum number of points granted every year, similar to the minimum and maximum reference wage (called the minimum and maximum pension rating) in the Slovenian defined benefit scheme. Moreover, similar to the current situation in Slovenia, some additional points might be granted for periods of unemployment, childcare for one year or for periods of part-time work combined with a longer childcare period.
4.2.2. The minimum contribution base is high
Low earners pay a higher effective contribution rate than high earners because the minimum contribution base, at 60% of average wage, is higher than the minimum wage which was close to 50% of average wage in 2019.1 This results in those earning the minimum wage paying the same contribution amounts as those earning 60% of average wage. At these earning levels, workers will receive the same pensions as they will benefit from the minimum reference wage set at 76.5% of the average wage.
The higher effective contribution rate for low earners reduces their disposable income and might restrain work incentives, even though progressive income taxation partially offsets this effect. Lowering the minimum contribution base would be a step towards equalising effective contribution rates among earnings levels. However, this would also reduce contribution revenues, while low earners already benefit from much higher replacement rates. Hence, if the minimum contribution base were lowered, how to compensate revenue losses would be a normative question. For example, given that lowering the minimum contribution base would effectively increase the already high redistribution within the scheme – as low earners would pay lower contributions while still receiving benefits calculated on the minimum reference wage – lowering the minimum reference wage might be an option to accompany the measure. Lowering the minimum contribution base benefits low-wage workers while lowering the minimum reference wage used for pensions penalises low-income pensioners.
4.2.3. Providing more flexibility to combine work and pensions
In Slovenia, only 40% of the old-age pension can be claimed while working full time, which limits the flexibility to combine work and pensions. Thus, full-time work does not allow to access 60% of the benefits that are available to those with the same insurance history but who are not working. However, when combined with the 40% pension limit after 40 years of contributions, the accrual rate increases from 1.36% to 3% for new entitlements, hence providing financial incentives that are close to actuarial neutrality.
The access to pensions should be disconnected from whether working or not. There is no obvious reason why there should be any restriction to combine work and full pensions when combining is not costly for public finances, i.e. when total flexibility to combine work with a full pension does not deteriorate pension finances in the long run. In the past, some countries raised such obstacles based on arguments related to the “lump of labour” fallacy, or views, outdated especially given health improvements at older ages, which associated pension receipts with the inability to contribute to society through working. Most OECD countries have considerably relaxed the conditions allowing to combine work and pensions (OECD, 2017[3]; OECD, 2021[4]).
Likewise, terminating the employment contract is generally not used any more as an eligibility condition to receive a full pension. Only a few countries, including Finland, France, Italy and Poland impose such a constraint nowadays. The conditions allowing combining work and pensions should depend neither on the type of employment contract nor on employment history. In particular, they should not require terminating the employment contract nor impose restrictions when re‑entering employment after having retired.
Introducing one of the following options would improve flexibility in combining work and pensions. The first option would be to enable combining work and the receipt of 100% of pensions while limiting the accrual rate for extra years to the regular 1.36% that applies to prior years. The 3% higher accrual would apply only to those who withdraw 40% of their pension. Although this will render drawing full pensions while working consistent with actuarial principles, this solution would add to the already complex structure.
The second option would be to introduce a standard bonus scheme for deferrals and replace the 3% higher accrual rate by the regular accrual rate. For example, once eligibility conditions are met, deferring pensions on accumulated entitlements would lead to a higher pension of about 4%‑4.5% per year of deferral (about 1% per quarter).2 While this bonus would apply to deferred pensions, it should not apply to pensions combined with work.3 Individuals could then decide to either defer claiming the pension and benefit from the bonus or receive the pension without any bonus, while accruing additional pension entitlements from working but at the same rate as before drawing pensions.
The third option is similar to the second one, but with an element to nudge behaviours about working longer. For individuals who continue to work after having met the full-pension conditions, they would not pay employees’ pension contributions and employers’ pension contributions would be paid as additional income without generating any additional pension entitlements, whether working these extra years are combined with pensions or not. Thus, individuals would have the same choice as under the second option, but no additional pension entitlement would accrue, employees would not pay pension contributions and employers’ pension contributions would be paid as additional wages.
Whatever the option selected, it is important to ensure a high level of transparency in the communication of accrued entitlements. People should be in a position to easily assess the consequences of their decisions. Whether pensioners benefit from such a framework to combine work and pensions depends on their capacity to make well-informed choices, according to their individual situation and preferences, in order to avoid jeopardising their final retirement incomes.
4.2.4. Aligning pension rules for the self-employed and employees
The self-employed contribute and receive pension entitlements in a similar way as employees in Slovenia. The mandatory nominal contribution rates are harmonised and the self-employed pay the contribution rate equal to the sum of employees’ and employers’ shares.
Still, the income base to calculate contributions and benefits is lower for the self-employed than for employees with the same earnings, which does not ensure an equal treatment of workers in terms of both obligations (contributions) and entitlements (benefits). Harmonising contributions would require setting the contribution base at 86% of profits, rather than at 75% today. This would increase both the contributions paid by and the future pensions of the self-employed. In addition, profit calculation is favourable for the self-employed in the flat-rate cost regime, as they are allowed to deduct 80% of income as costs.
The self-employed, similarly to full-time employees, cannot pay contributions on less than 60% of the average wage, which is the highest level of the minimum contribution base in the OECD. While part-time employees pay pro-rated contributions, working time cannot typically be identified for the self-employed, which implies that the minimum contribution base cannot be pro-rated and that self-employed with low earnings face very high effective contribution rates. There may be good reasons to allow the self-employed with low earnings to pay less contributions than those based on the minimum contribution base, but this is not as straightforward as it might seem. Similar treatment of the self‑employed and employees implies that the self-employed with income lower than the full-time minimum wage should pay reduced contributions as is the case for the part-time employees. However, given the flexibility for many self-employed to report low profits, especially in the flat-rate tax regime, the possibility for reduced contributions might be misused, both putting at risk future pensions of the self-employed and deteriorating pension finances in the short run.
For high earners, there is no ceiling to the contribution base for employees while the self-employed do not pay contributions from earnings higher than 350% of the average wage. Such a ceiling implies that the self-employed with very high income do not pay their share of financing public pensions. Hence, there is a case for removing this ceiling and, if needed, additional measures to monitor profit reporting might be put in place.
4.2.5. Improving transparency of pension finances
Old-age and survivor pensions are financed together with disability pensions and some long-term care benefits. Securing adequate income when old is a different objective than providing income in case of disability or providing resources to finance long-term care needs. Entitlements to old‑age pensions are accrued throughout the career, allowing to shift resources from younger to older ages when labour income typically falls. By contrast, disability and long-term care schemes cover risks of high costs but low incidence. Separating the financing of disability insurance and old-age pensions would pave the way to managing separate budgets, which would improve transparency and facilitate the management of both schemes over the long term. In particular, the financial sustainability of old-age and survivor pensions cannot be precisely defined and monitored when financial management is blurred with other schemes covering risks that are different in nature.
To further improve the transparency of pension finances, the financial flows related to all redistributive elements within the pension scheme should be precisely identified and reported. Part of pension redistribution is financed by an earmarked subsidy from the state budget of 0.5% of GDP in 2019. The earmarked subsidy covers entitlements based on: periods of unemployment and childcare; some other reasons (e.g. for veterans); and, the part of the top-up from minimum pensions corresponding to the first 15 years of insurance. Neither the remaining cost of the minimum reference wage beyond the first 15 years nor the surplus generated by the maximum reference wage are reported by ZPIZ while the state budget covers the deficit of the ZPIZ budget, at 1.4% of GDP in 2019. The other part of redistribution thus takes place within the pension scheme – although this part is not reported – and/or is covered by transfers from the state budget in indeterminate proportions. It is not clear why only the earmarked subsidy is identified as part of expenses covered by the state budget, hence not entering the ZPIZ deficit (or surplus). Beyond improving transparency, estimating the cost of all redistributive measures could be the first step to justify using tax revenues to finance redistributive pension components.
The financing of health care contributions for pensioners is not transparent. Healthcare contributions for pensioners are paid by ZPIZ to the Health Care Institute, i.e. mainly from contributions levied on wages. This means that contributions on wages finance the health insurance of both workers (through health contributions) and pensioners (through pension contributions). The health insurance of pensioners should be financed either by health contributions paid by pensioners or general taxation.
Monitoring pension finances more transparently through better reporting of annual pension flows, both revenues and expenditures, would then allow a proper reporting of pension finances in terms of stock, i.e. in terms of cumulative balances. Such a monitoring of cumulative balances over the long term is crucial for the assessment of pension finances and for the detection of signs of sustainability issues. In particular, monitoring cumulative financial balances would allow separating the long-term structural mismatches between pension expenditure and pension revenues from the short-term fluctuation of contributions. A way of implementing such a monitoring would be to task an independent body with the calculation of the actuarial balance sheet of the pension scheme in order to better inform the society (Vidal-Meliá, Boado-Penas and Settergren, 2009[5]).
4.2.6. Lack of solid arguments for mandatory retirement in Slovenia
Mandatory retirement refers to terminating employment contracts when employees reach a certain age. Among other ways, this can be done through removing employment protection as of a certain age. There has been a push by EU institutions to restrict mandatory retirement as much as possible, as part of efforts to combat age discrimination, although with weak legal power. The rulings of the Court of Justice of the European Union (CJEU) have tried to circumscribe the conditions under which the practice of mandatory retirement might be considered lawful. According to the CJEU, a mandatory retirement age should be argued for with concrete occupation- or sector-specific evidence, for example related to health and safety concerns when working at an old age, and is only possible if the employees concerned have access to a full pension.
Mandatory retirement practices have been reduced in a number of countries. More than half of OECD countries do not allow for mandatory retirement in the private sector. Only in one in four countries does some form of mandatory retirement exist in the private sector before the age of 68 years. Although mandatory retirement is more common in the public sector, nine countries ban mandatory retirement even for civil servants. This also means, however, that it remains possible to lay off employees once they reach a certain age in many EU countries. Within its 2015 Recommendation that calls for governments to give people better choices and incentives to continue working at an older age and to respond to the challenges of rapid population ageing, the OECD recommends that countries seek to discourage mandatory retirement in close consultation and collaboration with employers’ and workers’ representatives.
Some economic arguments are sometimes put forward to justify mandatory retirement practices. These include the need to offset the impact of seniority in wage‑setting mechanisms and the strictness of employment protection against individual dismissals. Some even refer to the need to free jobs for young generations (“lump of labour” fallacy). None of these arguments is convincing to back a mandatory retirement age in Slovenia more than in other OECD countries.
The question of mandatory retirement should also be assessed within the current context of the Slovenian labour market. Slovenia has the second lowest labour force participation in the OECD among people older than 60. In line with early exits from the labour market in international comparison, eligibility to a full pension in Slovenia is based on loose conditions, from age 60 with 40 years of insurance. The introduction of mandatory retirement counteracts efforts to make people work longer, while pension spending is projected to increase substantially in Slovenia given fast population ageing. Since the employment protection of workers eligible to an old-age pension was lifted in December 2020 (Chapter 1), Slovenia is the only European OECD country allowing mandatory retirement before the statutory retirement age, at age 60, although effective implementation will depend on the decision of the Constitutional Court on the regulation. It is difficult to provide solid arguments for mandatory retirement in Slovenia based on international evidence, and in particular to allow mandatory retirement below 68 years, at least in the private sector, and certainly not before the statutory retirement age.
4.2.7. Aligning rules for civil servants and private‑sector employees
In Slovenia, the state budget finances the mandatory occupational pension scheme for civil servants. Based on the assumptions of the OECD pension model, the scheme will provide a pension top-up of 11% compared to full-career private‑sector employees with the same earnings. It is generally difficult to provide a good justification for higher replacement rates for civil servants. It is even more difficult today as a career spent totally in the public sector is much less common than in the past: granting specific treatment is even less suitable in the world of enhanced mobility between the civil service and the private sector. In addition, it is often argued that higher pensions for civil servants represent a form of deferred compensation for lower wages.
The special treatment of workers in some occupations or sectors raises fairness issues. Moreover, lower wages and higher pensions for civil servants generate, beyond the lack of transparency, significant inefficiencies (Whitehouse, 2016[6]). A large number of pension reforms in OECD countries over the past decades have integrated the pension schemes covering private‑sector and public-sector workers (OECD, 2016[7]). In Slovenia, this alignment should be achieved by adopting the same rules for mandatory pensions for all workers.
There are three options to align mandatory pensions across private‑sector and public-sector workers in Slovenia. The first would be to make the occupational scheme for civil servants voluntary, in which the employer might top up employees’ contributions. The second option would simply be to eliminate the occupational scheme for civil servants. This would create savings for the public purse or increase wages for civil servants, depending on how the savings from contributions that are currently paid are distributed. However, this would diminish the pensions of civil servants from mandatory schemes. The third option would be to extend the mandatory coverage to all workers. This would increase mandatory contributions paid by private‑sector employers with a positive impact on future pensions and potentially negative effects over time on international competitiveness, net wages and employment.
4.3. Addressing financial sustainability issues
Given current rules, pension finances will be subject to intense pressure in the forthcoming decades. This will come mostly from the shift towards an older population structure. Population ageing has started to accelerate from about the mid‑2010s and will continue at a fast pace until about 2050. Boosting employment rates, overall and especially at older age groups, could alleviate part of the financing pressure. However, significant pension reforms will also be required as the financial gap is projected to be very large. Although policy action might focus primarily on the spending side, additional revenues will be needed, beyond those that would be driven by higher employment.
More precisely, pension spending in Slovenia is projected to sharply increase from 10.0% of GDP in 2019 to 15.7% in 2050 and 16.0% in 2070, based on recent European Commission projections.4 This contrasts strongly with a modest increase of 1.1% of GDP on average among EU member countries by 2070; only in Luxembourg, the Slovak Republic and Slovenia are pension expenditures projected to increase by more than 5% of GDP. In 2050, based on these projections, Slovenia is the OECD country which will spend the most on pensions as a share of GDP except for Luxembourg.
Pensions in Slovenia are financed mainly from payroll contributions on a PAYGO basis. Revenues therefore tend to follow trends in the wage bill or GDP, while any deficit is automatically covered by the state budget. Hence, past and current pension parameters would imply that transfers from the state budget to finance pensions will need to sharply rise from around 2%5 to 8% of GDP6 between 2019 and 2050. In the absence of pension measures, this would result in substantial tax increases or spending cuts in other areas, or sharp increases in public debt.
Based on current rules, the pensions of almost all workers will need additional subsidies from the state budget. That is, given the internal rates of return of the Slovenian PAYGO scheme, workers’ contributions will be insufficient to finance pension promises (Box 4.1). While many countries subsidise low pensions within their public pension system or through taxes, in Slovenia contributions will be insufficient for almost all workers, even for those earning four times the average wage throughout their career. Additional top-ups from taxes or an increase in contributions paid by the next generations will be required to maintain current rules.
Box 4.1. Actuarially fair replacement rates in Slovenian public pensions
The OECD was asked by the government to compute replacement rates corresponding to actuarial fairness, i.e. the level of pensions that the PAYG system in Slovenia would deliver for different career patterns based on paid contributions compounded by the internal rates of return. By no means does this exercise provide any normative statement of what the replacement rates should be. It simply gives an indication of replacement rates that could be fully financed by contributions in the steady state, in particular disregarding other financial sources used to fund redistribution within the system.
To do this, different assumptions are needed. First, it is assumed that the pension contribution rate is equal to 21.9% corresponding to 90% of total social security contribution rate of 24.35% (Chapter 1). Second, the internal rate of return is assumed to be equal to the sum of employment growth and wage growth; the latter is in turn assumed to equal labour productivity growth (Chapter 2). Table 4.1 summarises the key results. Various scenarios are considered. Estimates in Panel A cover full-career private‑sector workers starting their career now at age 22 with three earnings levels: low earners at half the average wage during the whole career, average‑wage earners and high-wage earners at twice the average wage (Panel A). Panel B shows estimates for people with shorter careers, hence retiring at age 65. Projections also include self-employed workers, either paying contributions on total earnings or on the minimum contribution base (Panel C).
Table 4.1. Actuarially fair replacement rates in Slovenian public pensions
Calculations based on different assumptions
Note: All cases assume that claiming pensions is not combined with work, and that workers work without breaks until retirement (for shorter careers they start working later). The case with the higher contribution rate assumes that additional contributions are paid by employers to simplify the calculations. Based on current rules, the replacement rate does not depend on the productivity growth, on employment growth, the contribution rate, mortality assumptions. The average‑wage case for self-employed workers assumes that after deducting social security contributions, the profit of the self-employed equals the average wage net of social security contributions. Columns 4‑6 are not presented for Panels B and C to simplify the table. Projected changes in life expectancy are based on UN data for individuals born in 1996 while the no-change case relates to individuals born in 1960.
Source: OECD calculations.
Panel A
Column 2 shows the future net replacement rates based on current pension rules, consistent with what was shown in Chapter 1. Given these rules, replacement rates do not depend on productivity-growth assumptions, as productivity affects both the pension level when retiring and the wage level similarly (*). Column 2 indicates that for individuals retiring at age 62 after a 40‑year career, the future net replacement rates are equal to 95%, 63% and 59% for low earners, average earners and high earners, respectively.
Based on assumptions used in the 2021 Ageing Report (annual employment growth of ‑0.3% and productivity growth of 1.9%), the future replacement rates consistent with actuarial fairness (Column 3) are much lower than those currently promised (Column 2). For example, at the average wage, it is equal to 50% or 13 percentage points lower than based on current rules. Actuarially fair replacement rates are almost flat across different wage levels, the small pattern being explained by effective contribution rates, which are higher for low incomes (see more detail in the main text), and by progressive taxation. This implies that for low earners, who currently benefit from a strong redistribution, the actuarially fair replacement rate is drastically lower than based on current rules.
Columns 4‑6 show actuarially fair future replacement rates based on different assumptions than column 3. First, the exercise is replicated assuming the same future life expectancies as of now (Column 4), thereby neutralising the effect of expected health improvements. For example, at the average wage, the actuarially fair replacement rate is equal to 61% instead of 50% based on mortality projections. This is similar to the replacement rate promised by current rules: the future replacement rate is higher than what the PAYG system can deliver based on contributions and internal rates of return at the average‑wage level implying that expected improvements in life expectancy are not factored in current pension rules. Column 5 shows the impact of the less optimistic employment scenario, assuming an annual decline of 0.6% instead of 0.3% in column 3. Lower employment growth mechanically diminishes the internal rate of return, which translates into replacement rates being about 5 percentage points lower. Column 6 assumes annual productivity growth of 1.25% instead of 1.9%. When productivity growth is higher, both internal rates of return and wages are similarly affected upward. As pensions in payment are indexed to only part of wage growth, this increase in the rate of return allows to finance a higher replacement rate when retiring. Reciprocally, a lower productivity growth leads to a lower actuarially fair replacement rate at retirement, by about 2 percentage points compared with column 3 when productivity growth is lowered from 1.9% to 1.25% per year.
Columns 7‑8 show the two reform scenarios which, based on Figure 1.5 in Chapter 2, are more consistent with sustainable rates of returns at the average‑wage level. More precisely column 7 is based on increasing the contribution rate by 2 points and indexing pensions to prices. At the average wage, the actuarially fair replacement rate is equal in that case to 64%, very close to what the current rules promise. Column 8 is based on increasing both the retirement age and the contribution period by 3 years, switching to price indexation, 1.36% accrual rate up to the new retirement age with no possibility to combine work and pensions until that age. The actuarially fair future replacement rate increases to 68% at the average‑wage level. Tighter eligibility conditions and lower indexation allow to achieve a much higher replacement rate than in column 3, and at a level that is consistent with what current rules generate: based on current legislation (but assuming a 1.36% accrual rate after 40 years of contributions), the future net replacement rate for an average earner when retiring at 65 with 43 years of contributions would be 67%, a number that is not shown in the table.
Panel B
This panel shows the case of shorter contribution periods (15, 30 or 40 years) when retiring at age 65. Based on the Ageing Report assumptions, the future actuarially fair replacement rate is about 10 percentage points lower than what the current rules indicate. Compared to low earnings after full career, a short career is much less cushioned by the pension rules.
Panel C
This panel shows the case of self-employed workers. At the average‑wage level, for those who pay contributions on total earnings, the future net replacement rate based on legislated rules is at 54% slightly lower than for average‑wage employees (Chapter 1). The actuarially fair replacement rate is about 11 percentage points lower than implied by legislation, as for employees shown in Panel A. A much larger gap of 16 percentage points applies to the self-employed opting for the minimum contribution base (next row in the table) because their benefits are increased by the minimum reference wage. The last row shows low-earner self-employed, for whom the picture is similar to that of low-wage employees, as they also pay higher effective contribution rates.
(*): However, productivity growth affects indexation of pensions in payment.
A range of policy options should be considered to address financial sustainability issues. The main priority is to reign in pension spending, especially as both the tax wedge and the contribution rate are relatively high. However, given the projected financial gap, action will be needed to boost pension revenues as well. Box 4.2 summarises the expected impact of various policy options.
Box 4.2. Quantifying the impact of potential pension reforms on financial sustainability
Various scenarios have been simulated to limit the increase in pension expenditure. Table 4.2 summarises the expected impact of various policy options on pension expenditure, average pensions, the average age of claiming pension for the first time and the required transfer from the state budget to balance pensions, with details provided in Chapter 2.
For example, different options could separately generate pension savings of around 1% of GDP in 2050 compared with the no‑action baseline. One option is to increase the minimum conditions to access pensions from age 60 with 40 years of contributions to age 62 with 42 years of contributions by 2028 and then to link all eligibility conditions to life expectancy in a way that one year of life‑expectancy gains raises the eligibility conditions by 8 months. Similar savings can be achieved by lowering the indexation from a mix of 60% of wages and 40% of prices to 34% of wages and 66% of prices. In order to reduce spending by about 1% of GDP by 2050 through changes in the accrual rates applying to entitlements from 2027 will have to be reduced by 12%. On the revenue side, an additional 3 percentage points of contributions would be needed.
Table 4.2. Potential pension reforms to improve financial sustainability
Compared to the baseline of no policy adjustments in 2050
Notes: Net savings shows the reduction of pension expenditure for policy options 1‑5 and the increase in pension revenues for option 6.
Source: For details see Chapter 2.
Improving pension financial sustainability is never easy politically, but at least some of the options discussed below should be implemented as soon as possible to limit the economic and social costs from changing the rules in an ad hoc and abrupt manner when fiscal pressure becomes too tight. This could be done through introducing automatic adjustment mechanisms linking pension parameters with demographic or economic indicators (Box 4.3). Having an expert body – the greater the independence of this institution the better – in charge of the assessment of pension schemes, the pension expenditure projections and, more broadly, the evaluation of the impact of demographic changes would also be very helpful to strengthen the diagnosis and the acceptance of reforms, as well as to provide support for a sound management of the system (Fall and Bloch, 2014[8]). Several countries, including Belgium, Canada, France and Sweden, have such an independent body.
Box 4.3. Automatic adjustment mechanisms exist in more than half of OECD countries
Automatic adjustment mechanisms are present in about two‑thirds of OECD countries. Beyond the links between retirement age and life expectancy covered in Box 4.4, some mechanisms automatically adjust benefit levels. Automatic adjustment mechanisms help to both prevent the accumulation of large financial deficits and smooth needed adjustments thereby reducing political, social and economic disruptions. They are perceived as a remedy to the tendency of governments to procrastinate measures to address financial sustainability issues. Such mechanisms improve the rationality of sharing ageing costs (Börsch-Supan, 2007[9]) and improve transparency about how the adjustments will be made (Turner, 2007[10]).
In 15 countries mandatory pensions include funded or notionally defined contribution (FDC or NDC) schemes which directly account for changes in life expectancy in the calculation of annuities (*). Notional accounts within NDC schemes in Italy, Latvia and Poland additionally valorise pension entitlements with the growth rate of the wage bill or GDP through the notional interest rate. In Estonia, Germany, Japan, Lithuania, Luxembourg, the Netherlands and Sweden, benefits from mandatory pensions are linked to the pension financial balance, demographic ratios or the wage‑bill growth.
In Germany, the sustainability factor, which accounts for changes in the number of contributors relative to the number of pensioners, has been used to index the pension point value since 2005. The 2018 sustainability factor was positive, increasing pensions by 0.3%, but it is now projected to be negative, decreasing the adjustment of the pension point value by 0.5% per year until 2032. However, benefits cannot be reduced in nominal terms as a result of the adjustments. In that case, the downward adjustment from the sustainability factor is only applied if other factors in the pension point value (such as wage growth) are positive. Unapplied negative adjustments are, however, carried over to later years as it happened in the past.
In Lithuania, both the value of the pension point and of the basic pension are linked to changes in the wage bill. If the wage bill falls in nominal terms (which will cause a drop in contributions) the indexation of pension benefits and entitlements does not apply. In Estonia, the value of the pension point is also linked to contribution revenues.
In Sweden for the NDC scheme, the Swedish Pensions Agency calculates a solvency indicator, the balance ratio, by dividing the sum of the assets of the buffer fund and the approximate value of future contribution flows from current workers by pension liabilities (accrued notional pension entitlements) (Settergren and Mikula, 2005[11]). There is an automatic adjustment of NDC pensions to the balance ratio as there is no guarantee that the automatic link to life expectancy in the NDC formula, which computes NDC “annuities”, is enough to ensure financial sustainability. When a deficit is identified in the form of a balance ratio lower than one, a brake is activated, reducing the notional interest rate below the wage growth rate in order to help restore solvency, which both limits accumulation in notional accounts and reduces indexation of current pensions in payments. When rebalancing is achieved, any surplus is used to boost the notional interest and pension indexation during a catch-up phase. Sweden experienced some difficulties in applying the brake rule during the Great Recession, and revised it to avoid sharp adjustments. Overall, while the Swedish mechanism was put to the test, it proved resilient to such a huge economic shock, only requiring a small adjustment, with its broad principles remaining largely unchallenged.
Relatively recently, Finland introduced a sustainability factor in its DB pensions to ensure financial sustainability; Spain also introduced one, but suspended it before it came into effect. Portugal also has a sustainability factor, but it only applies to early retirement (OECD, 2019[12]). These sustainability factors are automatic adjustment mechanisms, linking pension benefits to life expectancy.
In Finland, since 2010 the initial level (at retirement) of PAYGO earnings-related pensions has been adjusted to take into account changes in life expectancy at age 62. The life expectancy coefficient lowers initial pensions by the ratio of average life expectancy at 62 in 2005‑09 to average life expectancy at 62 in the 5 years prior to retirement. The life expectancy coefficient is 0.957 for the cohort reaching the statutory retirement age in 2021, and is projected to be equal to 0.869 in 2066 (the year in which someone entering the labour market in 2020 will be allowed to retire). Additionally, the minimum retirement age is also linked to life expectancy (Box 4.4), which, along with actuarial adjustment of benefits, cushions the negative impact of the life expectancy coefficient on replacement rates.
In Japan, the adjustment mechanism of pension benefits, introduced in 2004, is based on changes in both the number of contributors and life expectancy, called macroeconomic indexation. The sustainability factor is the sum of two components: a fixed factor accounting for increases in life expectancy (currently ‑0.3%) and the average change in the number of contributors over the past 3 years (0.1% in 2019). However, this adjustment mechanism is not applied at times of negative inflation. Hence, a catch-up system was introduced in 2018, which carries over downward benefit revisions in years of negative inflation to later years. In 2019, as both prices and wages increased, the macroeconomic indexation was applied, and in addition the unrealised benefit reduction in the previous year was reflected through the carry over mechanism.
A comprehensive overview of automatic adjustment mechanisms in OECD countries is presented in OECD (2021[4]).
(*): When lump sums or programmed withdrawals are available, the assets are used throughout the retirement period, making the link with longevity implicit but similar to that prevailing with the pricing of annuities.
4.3.1. Increasing retirement ages and contribution periods
By allowing to access a full pension, i.e. without any penalty, from age 60 with 40 years of contributions, Slovenia offers loose conditions relative to other OECD countries. This is the case for people retiring now, but as there are no legislated changes to these conditions, differences with other countries will widen given reforms adopted elsewhere. Such loose conditions make it very challenging to finance adequate pension benefits in a sustainable way, a difficulty that will be exacerbated by longevity trends and relatively low fertility rates. With health improvements over the past decades, there is large scope to tighten eligibility conditions to access full pensions. By contrast, the option to retire at age 65 with a shorter contribution period is better aligned with the situation in other OECD countries. To reduce future spending, increasing the minimum retirement age and possibly the contribution period to get a full pension should therefore be at the top of the policy agenda given that effective retirement ages are low today.
In a second step, both the minimum and the statutory retirement ages should be linked to changes in life expectancy. Catching up with other countries in terms of current eligibility conditions in the short-to-medium term will not be enough to address future financial gaps (Box 4.2). Links to life expectancy reduce uncertainty about future pension rules by minimising the need for ad hoc adjustments. They improve credibility and help to build trust in the pension system. For example, transmitting two‑thirds of gains in life expectancy to the retirement age would broadly keep the share of working time (and of retirement period) in adult life constant across generations, thus contributing to equity. Yet, larger increases in the retirement age might be needed to ensure financial sustainability, given that the shift in the population structure goes beyond the impact of longer lives, and given that the starting point might not be balanced. Box 4.4 provides details about the links between retirement age and life expectancy in OECD countries. Tightening eligibility conditions will affect future pensioners.
One potential issue with linking the retirement age to life expectancy arises when inequality in life expectancy increases (OECD, 2019[13]). There is conflicting evidence about how socio‑economic differences in life expectancy have changed among OECD countries. In Slovenia, an increase in life‑expectancy inequality is overall not supported by the evidence. The life expectancy gap across education levels declined by 0.7 years among men between 2007 and 2017, while is increased slightly by 0.2 years among women.7 While automatically linking retirement age to life expectancy is one of the key policies to improve financial sustainability, it is important to monitor as closely as possible the medium-to-long-term trends in life‑expectancy inequality.
Without other changes in pension parameters, increasing the retirement age has a positive effect on pension levels and generates net savings from both a shorter retirement period and a longer contribution period. Depending on policy priorities, further savings may be achieved through additional measures, such as lowering the accrual rates to accompany the increase in retirement ages, for example in a way that keeps the target replacement rate at the minimum retirement age constant.
Childcare periods should not result in lowering the minimum retirement age. There are valid reasons to grant pension entitlements for periods of childcare and thereby to limit the impact of childcare-related breaks on pensions. However, it is far less obvious why parents should be able to retire earlier compared to childless people and only five OECD countries relax pension eligibility conditions based on having children. In Slovenia, mothers and fathers can retire four and two years below the statutory retirement age, respectively.
If perceived as socially desirable, increases in the retirement age might be cushioned by introducing an early retirement option, which would allow to retire a few years (two or three) before the statutory retirement age. This option should not be costly for pension finances, which implies, that at least actuarially neutral permanent penalties to benefits should apply – between 4% and 5% annually in Slovenia. Larger penalties would reduce incentives to retire early, but they would also increase the risk of people making mistakes in using this option and ending up with low benefits.
Consistent with the efforts to increase effective retirement ages, labour market policies should not be age‑specific. The current extended unemployment protection for older workers poses a risk of being used as a pathway to retire early. It might thus limit the impact of tightening eligibility conditions to old-age pensions. For older workers who cannot find employment due to health problems, disability benefits are the adequate policy instrument, while active labour market policies should help those without health incapacity who experience difficulties to find a job.
Box 4.4. Links of retirement age to life expectancy in OECD counties
Denmark, Estonia, Finland, Greece, Italy, the Netherlands and Portugal have linked retirement ages to life expectancy, although Italy has (temporarily) suspended the link for some occupations (*) (OECD, 2021[4]). The Slovak Republic had established such a link in 2012 but eliminated it in 2017; recent legislation foresees an increase of the retirement age to 64 by 2030, while a new mechanism of raising the retirement age beyond this age is to be established. Beyond pensions, such links lower the impact of ageing on total output and ultimately on the average standard of living of the whole population.
The exact link differs across countries. Denmark, Estonia, Greece and Italy increase the retirement age by one month for every month gained in life expectancy at age 65, except for Denmark which uses age 60. This might be needed to ensure financial sustainability, but a one‑to‑one link basically implies that all additional expected life years are spent working, while the length of the retirement period is constant, leading to a steady decline in the share of adult lives spent in retirement. In Denmark, the parliament has to vote every five years to ensure the link is maintained.
In Finland, the Netherlands and Portugal the statutory retirement age increases by two‑thirds of life expectancy at 65; Sweden plans to implement a similar link. In Finland, this is done with the expressed goal of keeping the ratio of expected time in retirement to time spent working constant. In addition in Portugal, someone with more than 40 years of contributions can retire four months earlier for each year over 40 years of contributions. This effectively implies that only half of life expectancy gains are reflected in the normal retirement age (OECD, 2019[13]). The Netherlands switched from a one‑to‑one to a two‑thirds link in 2020.
Not all links between retirement ages and life expectancy ensure by themselves the financial sustainability of PAYGO DB schemes, even if the pension system is based on sound finances initially i.e. notwithstanding the impact of demographic changes. First, whether it does depends of course on the extent to which changes in life expectancy are transmitted to changes in retirement ages. One‑third, two‑thirds or one‑to‑one links do not produce the same effects, even though they will all reduce the length of pension payments when longevity increases. Second, changes in the size of the working-age population driven by past fertility rates matter for pension finances irrespective of longevity gains. Third, in most countries additional years of work mean additional pension entitlements, and depending on how far current rules are from actuarial neutrality, this will generate more or less pension saving for the pension provider. As long as the pensioner-to-contributor ratio stays constant, a stable aggregate replacement rate can be financed by a stable contribution rate in a sustainable way when the initial parameters are also set in a sound way. This is why one objective of such links is to help stabilise the pensioner-to-contributor ratio, which tends to increase with longevity gains and retirement ages that do not adjust. Not raising the retirement age in line with improvements in life expectancy tends to deteriorate financial balances, which then need to be improved through lower replacement rates, reduced pension indexation, or higher contribution rates or additional tax resources.
Two aspects make the implementation of such a link attractive. First, it is conditional on health changes that are effectively taking place. If health improvements do not materialise then retirement ages do not increase. Second, such links limit the political cost to undertake such unpopular measures as raising the retirement age.
(*). Italy suspended the automatic links with life expectancy of both career-length eligibility conditions for early retirement (42.8 and 41.8 years for men and women, respectively), and the statutory retirement ages for some workers only, including those in arduous occupations until 2026.
4.3.2. Adjusting pension indexation
Reducing the indexation of pensions in payment is a powerful instrument to limit pension expenditure without lowering initial pension levels. There is no optimal indexation mechanism as, for a given level of spending, there is a clear trade‑off between higher initial benefit levels when retiring and a higher indexation. Price indexation maintains the purchasing power of pensions, while wage indexation ensures a stable relative income, but tends to be more costly. If the objective is to reduce pension spending beyond the impact of higher retirement ages, one option is therefore to either cut the initial pension (or the replacement rate at retirement) or to reduce indexation.
When comparing these two options, the first penalises in particular those with a shorter life expectancy while the second will lower the relative income of the oldest pensioners. From a financial point of view, one big advantage of a move to price indexation is that it generates savings even in the short term. In addition, it would affect both current and future pensioners, thus sharing the adjustment burden more broadly, which might be fairer if current pensioners have benefited from relatively favourable pension rules. In principle, there is no reason why current pensioners should not participate in improving financial sustainability provided that, consistent with price indexation, their purchasing power is not reduced. For example, changing pension indexation from today’s mix of 60% of wages and 40% of prices to full price indexation is projected to reduce pension expenditure by about 2% of GDP by 2050.8
4.3.3. Increasing contribution rates or expanding tax resources
Raising contribution rates is generally one key measure that could be used to improve pension finances. However, it is problematic because this tends to reduce the disposable income of current workers, deteriorate firms’ competitiveness and hamper employment. The combined impact of these effects depends on various factors, including the initial contribution rates, the stringency of employment protection legislation and the intensity of product market competition. Increasing the contribution rate shifts the cost of adjusting pension finances to workers and firms.
In Slovenia, the tax wedge – the difference between labour cost and take‑home pay as a percentage of gross wages – is relatively high, which suggests that the space to raise contribution rates might be limited. Based on a general equilibrium model which accounts for labour supply effects, an increase of about 3 percentage points in total contribution rates would be needed to increase contribution revenue by 1% of GDP. In Slovenia, as discussed above, the boundary between taxes and contributions to finance pension redistribution is blurred. The 2018 OECD Tax Policy Review of Slovenia provides policy options on how financing of public spending could be improved (OECD, 2018[14]).
Should the political decision be taken to improve long-term pension finances by raising contribution rates or using tax revenues, a buffer fund might be built up to smooth the increases. Indeed, instead of waiting to increase contribution rates when pension expenditure actually rises, accumulating the reserves earlier would enable more gradual increases. This would cause less disruptions in the labour and product markets, and would share the potential cost in terms of disposable income more evenly across cohorts. In that case, however, to avoid the buffer fund from being misused later on, it is crucial that objectives, dissolution rules and financing sources are set upfront in detail. In short, the design of the buffer fund should ensure that its use is restricted to fulfilling the initial mandate (or entail large political cost if this is not the case). As pension expenditures are projected to accelerate between 2030 and 2050, the buffer fund should start to accumulate resources as soon as possible and be gradually and partly withdrawn e.g. from the late 2030s. The more the future deficits are pre‑financed across generations, the lower the level that contribution and tax rates will need to reach.
Reserve funds can be used for various purposes: the better management of large temporary albeit long enough shocks, such as the impact of the baby boom generations retiring; the cushioning of short-term economic shocks affecting pension revenues (and benefits); and, the diversification of pension revenues. In Sweden, a reserve fund worth around 30% of GDP helps to smooth ageing-related adjustments and separate earnings-related pension finance from the state budget. Every year, the value of the reserve fund is added to the estimated value of the future contribution flows of current workers, the so-called implicit or contribution asset (Settergren and Mikula, 2005[11]). This makes for total assets, which are then compared with pension liabilities made of pension entitlements accrued so far.9 If the total assets are not enough to cover pension liabilities, both current pensions and pension entitlements are adjusted while a smoothing mechanism prevents abrupt changes in benefits. In Finland, the partial prefunding of pensions, with financial assets of mandatory schemes being around 85% of GDP, allows to diversify risks and to separate earnings-related pensions from the state budget. In the United States, the social security trust funds worth around 14% of GDP are not allowed to borrow and are strictly separated from the state budget. Unless the legislation is adjusted, the depletion of the trust funds, which is projected in the 2030s, will translate into benefit cuts.
The part of public pensions that is funded is negligible in Slovenia. This part comes from income coming from state‑owned assets managed by Kapitalska Druzba, which amounted to about 1% of total pension expenditure in 2019. In October 2020, the government proposed a bill to create the National Demographic Fund, which would pool state‑owned assets worth between 17% and 23% of GDP, with 40% of the dividends from these assets used to finance public pensions. Even with an optimistic assumption, this annual stream of income for pensions would not be larger than 0.5% of GDP,10 which would cover a small part of the expected financial gap, of around 8% of GDP by 2050. By comparison, implicit public pension liabilities are estimated to have increased sharply from 313% to 359% of GDP between 2015 and 2018.11 In order to improve transparency and better separate pension financing from the state budget, the National Demographic Fund might be used as a reserve fund for pensions, but this would require implementing clear rules about financing pension balances as for example in Sweden. In particular, it is crucial to limit the tendency of governments to interfere in the investment process of public pension reserve funds (Palacios, 2002[15]).
4.3.4. Adjusting replacement rates
Another policy option to further improve financial sustainability is, for a given career, to lower replacement for future cohorts depending on demographic factors, such as expected changes in life expectancy or contributor-to-pensioner ratios. For example, lowering replacement rates proportionally to changes in remaining life expectancy at age 65 from 2027 would lower pension expenditure by 0.7% of GDP in 2050.
This negative impact on future benefits might be offset by raising the retirement age and linking it to life expectancy. For example, with the two‑thirds link of both the retirement age and the contribution period to life expectancy, a decrease in accrual rates of around one‑third of changes in remaining life expectancy would stabilise the replacement rate at the normal retirement age. Such a benefit adjustment would affect only future retirees.
Depending on political objectives, redistributive features such as the minimum and maximum reference wages could also be adjusted to improve financial sustainability, affecting the replacement rates of low and high earners, respectively. For example, gradually lowering the minimum reference wage from 76.5% to 58.5% between 2027 and 2036 is expected to reduce expenditure by 0.4 percentage points of GDP in 2050. The net replacement rate for those earning half the average wage throughout career would decrease strongly from 90%, much above the OECD average, to the OECD average of 69%. Lowering the lowest pensions to such an extent, however, must be seen as a last resort option. As for high pensions, a gradual decrease of the maximal reference wage from 306% to 206% of average wages between 2027 and 2036 would lower pension expenditure by 0.3% of GDP in 2050, while the average pension would decrease by 2%. Those affected would see their future pensions lowered by up to one‑third.
4.4. Improving first-tier pensions
4.4.1. Revise eligibility criteria for social assistance benefits
The means test to access social assistance benefits should only apply to the individual or the couple requiring assistance, and should not include children. The obligation of adult children to provide financial support to individuals in need is likely to be a major obstacle for older people to access social assistance benefits. As a result, some older people might not be getting the financial support they need. Moreover, the current family obligation effectively acts as a tax on social mobility, as particularly children of low-income parents who have managed to grow out of their parents’ precarious situation would be obliged to support them.
Entitlements are also complicated because the financial social assistance eligibility thresholds depend on the number of working hours while the supplementary allowance is accessible to people who do not work. The level of social assistance benefits should depend on other income but neither on the employment status nor on working hours.
One interesting feature of the supplementary allowance is that it ensures that individuals who are older than the retirement age receive a higher social assistance benefit than working-age people, because after the retirement age there should be less concern about work disincentives. However, while the current eligibility age is at the statutory retirement age for men, it is earlier than the statutory retirement age for women. Moreover, there is no need to have the supplementary allowance as a distinct instrument. Rather it should be merged with financial social assistance, which would just include a higher eligibility threshold for people older than the retirement age.
4.4.2. Adopt a more integrated framework for first-tier pensions
For older people who do not work, one important issue arises as social assistance benefits are withdrawn at a rate of 100% against other income, meaning that for example one euro of contributory pensions results in the loss of one euro in social assistance benefits. Given that social assistance benefit levels have exceeded most minimum pensions – at least until 2021 –, such high withdrawal rates strongly diminish the advantages generated by pension contributions for low earners as extra contributions do not result in higher entitlements.
High withdrawal rates are attractive because they minimise cost, but this comes at the price of poor incentives. By contrast low withdrawal rates minimise crowding-out effects on labour supply and contributory pensions, but are more costly. Valdés-Prieto (2009[16]) suggests that it is optimal to opt for a scheme with a relatively low withdrawal rate, around 30%‑50%, which ensures that each amount or period of contributions results in higher total benefits. Moreover, withdrawing based on income allows for the elimination of the link with hours worked. By setting a lower withdrawal rate, receiving a higher income from other sources than social assistance always results in a higher total income – no matter whether this is the result of a higher pension, a higher wage or more hours worked.
The interaction between the minimum pension and safety-net benefits could take different forms depending on the withdrawal rate. Figure 4.1 shows possible interactions between social assistance (the merging discussed above between financial social assistance and the supplementary allowance for people older than the statutory retirement age) and the contributory pension – or any other source of income. While under current rules corresponding to a withdrawal rate of 100%, the level of total income of a person receiving social assistance does not increase with contributory pension benefits until the safety-net eligibility threshold, a withdrawal rate lower than 100% implies an increase of total income at a rate that is equal to 100% minus the withdrawal rate. As the withdrawal rate becomes smaller, people can combine a pension with social assistance benefits until higher pension benefit levels (Panel A). However, if withdrawal rates are to be implemented in a budget-neutral way, a lower withdrawal rate also means a lower benefit level for people with no other income (Panel B as an illustration). Alternatively, to guarantee that no one falls below a certain income level an initial withdrawal rate of 100% could be maintained, after which a lower withdrawal rate guarantees that higher contributory pension benefits also result in a higher income (Panel B, but with 100 as the minimum total income). Doing so would reduce the period for which having made more contributions does not affect the income level of people receiving social assistance. Box 4.5 provides examples of the withdrawal-rates structures used in four selected countries.
The guaranteed pension was raised in 2021 to ensure that a person with 40 years of pensionable service without purchase would have a pension income above the threshold to qualify for social assistance. The guaranteed pension generates discontinuities in pension build-up, with a steep increase in pension entitlements after 40 years of pensionable service without purchase followed by a period during which paying more contributions does not result in additional entitlements. As such, the scheme effectively eliminates work incentives generated by the increased accrual of 3% for low-income earners. Therefore, the scheme should be merged with the minimum pension so as to eliminate these discontinuities in a budget neutral way. Moreover, in order to ensure that having paid more contributions always generates higher incomes in old age, replacing the accrual rate of 29.5% after 15 years of employment by an accrual rate of 1.967% (= 29.5%/15) per year for the first 15 years of contributions could be considered.
Also, once a more integrated framework for first-tier pensions has been adopted, social assistance eligibility thresholds could be indexed to wages instead of prices. If both the minimum pension and safety-net thresholds follow the same indexation rule, then their ratio remains constant over time. However, there is a trade‑off between the level of the same indexation to both minimum pension and safety-net benefits, and the cost of the social assistance system. In any case, wage indexation of safety-net benefits should not be considered before a consistent first-tier pension framework is in place.
Figure 4.1. Total income and contributory pensions for different withdrawals of social assistance
Note: The withdrawal rate refers to the amount of social assistance benefit a person receives less for every euro of other income the person receives.
Source: OECD calculations.
Box 4.5. Withdrawal rates of targeted benefits in selected OECD countries
Canada currently has an initial withdrawal rate of 50%, after which the withdrawal rate increases to 75% meaning that an extra dollar in contributory pension benefits results in a loss of 0.75 dollars in targeted benefits. As contributory pension benefits increase further, the withdrawal rate is lowered to 50% again (Figure 4.2). Denmark and Finland provide large incentives for people with little contributory pension to build up a pension by not withdrawing the targeted benefit against initial contributory pension. While in Finland, only the first EUR 56 per month is not withdrawn, in Denmark this is almost tenfold the amount (DKK 7 475). At 30.9%, Denmark also employs a much lower withdrawal rate after this point than Finland (50%). By contrast, Sweden has an initial withdrawal rate of 100%, after which the targeted benefit is withdrawn at 48% against the contributory pension.
Figure 4.2. Withdrawal rates of targeted benefits in selected OECD countries
Source: Country profiles in OECD (2019[13]) Pensions at a Glance 2019: OECD and G20 Indicators, https://dx.doi.org/10.1787/b6d3dcfc-en and information provided by the countries.
Policy options
Improving public earnings-related pensions
Simplify the pension rules, while adjusting accrual rates as needed for example to stabilise pension levels on average, by: increasing the reference period from the best 24 years to lifetime earnings, using gross wages for the reference‑wage calculation; and, eliminating the annual discretionary allowance.
Improve the transparency of pension finances by: creating an independent expert body in charge of monitoring pensions to provide support for a sound management of the system; separating the financing of old-age and disability pensions as a first step to run separate budgets; improving the reporting of the net cost of minimum and maximum reference wages; and, explicitly recording the cumulative balance between contributions and entitlements over time.
Remove the restrictions to combine work and pensions once a worker is eligible for a full pension, provided that combining work and pensions does not deteriorate public finances in the long term.
Raise the contribution base of the self-employed from 75% of profits (86% of profits will harmonise contributions with employees).
Roll back the reform which removed the requirement to provide a justified reason when dismissing an employee who has met eligibility conditions to the old-age pension.
Align pension contributions and entitlements between civil servants and private‑sector workers.
Addressing financial sustainability issues
Tighten the minimum eligibility conditions to pensions (minimum retirement age and contribution-period condition for a full pension) and link retirement ages to life expectancy.
Remove the lowering of the minimum retirement age based on childcare periods.
Lower indexation of pensions in payment.
In addition, pension finances would be enhanced by combining some of the following options, with different impacts as discussed in the text:
Adjust benefits to life expectancy or to the ratio of contributors-to-pensioners, increase contribution rates, finance pension redistributive components from the state budget, and lower the minimum and/or the maximum reference wages.
Improving first-tier pensions
Remove the means-testing of social assistance benefits (both financial social assistance and supplementary allowance) to children of beneficiaries.
Eliminate the conditionality of financial social assistance and supplementary allowance on employment and hours worked; make the supplementary allowance eligible at the statutory retirement age for both men and women; and, merge the supplementary allowance with financial social assistance by granting a higher benefit level for people older than the retirement age relative to people below the retirement age.
Merge the guaranteed pension with the minimum pension in a budget-neutral way.
Adopt an integrated framework for old-age safety nets and contributory pensions by ensuring that contributions paid (at least from 15 years) result in higher total benefits through the withdrawal of safety-net benefits at a much lower rate than the current 100%.
References
[9] Börsch-Supan, A. (2007), “Rational Pension Reform”, The Geneva Papers on Risk and Insurance - Issues and Practice, Vol. 32/4, pp. 430-446, http://dx.doi.org/10.1057/palgrave.gpp.2510149.
[18] Čok, M., J. Sambt and B. Majcen (2010), Impact assessments of the proposed pension legislation, University of Ljubljana, Faculty of Economics.
[8] Fall, F. and D. Bloch (2014), “Overcoming Vulnerabilities of Pension Systems”, OECD Economics Department Working Papers, No. 1133, OECD Publishing, Paris, https://dx.doi.org/10.1787/5jz1591prxth-en.
[17] Mackenbach, J. et al. (2014), “Trends in inequalities in premature mortality: a study of 3.2 million deaths in 13 European countries”, Journal of Epidemiology and Community Health, Vol. 69/3, pp. 207-217, http://dx.doi.org/10.1136/jech-2014-204319.
[2] MLFSAEQ (2016), The White Paper on Pensions,, Ministry of Labour, Family, Social Affairs and Equal Opportunities, Ljubjana, Slovenia,, https://www.gov.si/assets/ministrstva/MDDSZ/pokojnine/Bela-knjiga-o-pokojninah.pdf.
[4] OECD (2021), Pensions at a Glance 2021: OECD and G20 Indicators, OECD Publishing, Paris, https://doi.org/10.1787/ca401ebd-en.
[1] OECD (2020), OECD Reviews of Pension Systems: Czech Republic, OECD Reviews of Pension Systems, OECD Publishing, Paris, http://dx.doi.org/10.1787/e6387738-en.
[12] OECD (2019), OECD Reviews of Pension Systems: Portugal, OECD Reviews of Pension Systems, OECD Publishing, Paris, https://dx.doi.org/10.1787/9789264313736-en.
[13] OECD (2019), Pensions at a Glance 2019: OECD and G20 Indicators, OECD Publishing, Paris, https://dx.doi.org/10.1787/b6d3dcfc-en.
[14] OECD (2018), OECD Tax Policy Reviews: Slovenia 2018, OECD Tax Policy Reviews, OECD Publishing, Paris, https://dx.doi.org/10.1787/9789264303898-en.
[3] OECD (2017), Pensions at a Glance 2017: OECD and G20 Indicators, OECD Publishing, Paris, http://dx.doi.org/10.1787/pension_glance-2017-en.
[7] OECD (2016), OECD Pensions Outlook 2016, OECD Publishing, Paris, https://dx.doi.org/10.1787/pens_outlook-2016-en.
[15] Palacios, R. (2002), “Managing public pension reserves Part II : lessons from five recent OECD initiatives : Managing public pension reserves - Part II : lessons from five recent OECD initiatives”, Social Protection discussion paper series, World Bank Group, http://documents.worldbank.org/curated/en/605351468182660681/Managing-public-pension-reserves-Part-II-lessons-from-five-recent-OECD-initiatives.
[11] Settergren, O. and B. Mikula (2005), “The rate of return of pay-as-you-go pension systems: a more exact consumption-loan model of interest”, Journal of Pension Economics and Finance, Vol. 4/2, pp. 115-138, http://dx.doi.org/10.1017/s1474747205002064.
[10] Turner, J. (2007), Autopilot: self-adjusting mechanisms for sustainable retirement systems, http://www.actuaries.org/Boston2008/Papers/IPT5_Turner.pdf.
[16] Valdés-Prieto, S. (2009), “The 2008 Chilean Reform to First-Pillar Pensions”.
[5] Vidal-Meliá, C., M. Boado-Penas and O. Settergren (2009), “Automatic Balance Mechanisms in Pay-As-You-Go Pension Systems”, The Geneva Papers on Risk and Insurance - Issues and Practice, Vol. 34/2, pp. 287-317, http://dx.doi.org/10.1057/gpp.2009.2.
[6] Whitehouse, E. (2016), “Pensions for Public-Sector Employees : Lessons from OECD Countries’ Experience”, Social Protection and Labor Discussion Paper, World Bank, Washington, DC, Vol. No. 1612, https://openknowledge.worldbank.org/handle/10986/25286.
Notes
← 1. The minimum wage increased by almost 9% in 2021, which is expected to be a stronger growth than the average wage’s, but the minimum wage will remain lower than 60% of the average wage.
← 2. On top of the increase stemming from the valorisation of past wages and before accruing additional entitlements.
← 3. When combined with work, pensions should just increase based on additional accruals without any bonus.
← 4. The 2019 figure of 10.0% of GDP is lower than the ZPIZ budget of 11.5% of GDP as the ZPIZ budget covers also benefits other than old-age and survivor pensions.
← 5. In 2019, all ZPIZ expenditure, which include also some long-term care benefits, were 2.2% of GDP larger than contributions. This difference was covered mainly by a transfer from the state budget, of which 0.5 percentage points was paid to cover some redistributive elements of pensions and 1.4 percentage points was paid to cover the ZPIZ deficit. When pro-rating the allocation of tax revenues based on expenditures by category, total transfers from the state budget to finance old-age and survivor pensions would stand at 1.8% of GDP in 2019, of which 1.2% of GDP would cover the “deficit”.
← 6. The 2018 EC projections show that the pension contributions as a share in GDP are expected to decrease from 8.9% to 8.6% between 2020 and 2050. The actual share for 2019 stood at 9.3% of GDP.
← 7. Source: http://appsso.eurostat.ec.europa.eu/nui/show.do?dataset=demo_mlexpecedu&lang=en. Mackenbach et al. (2014[17]) did not find significant changes among women, and an absolute decrease and relative increase in inequality among men.
← 8. Simulations show that changing pension indexation from a mix a 60%‑40% of wages and prices to 34%‑66% lowers pension expenditure by 1% of GDP in 2050. Reduction in pension indexation only to inflation, i.e. twice larger reduction in real indexation, would translate into approximately twice higher reduction in pension expenditure. Indeed, Čok, Sambt and Majcen (2010[18]) paper shows that reducing pension indexation from 100% of wages to 50% of wages and 50% of prices would have half the effect on pension expenditure compared to reducing pension indexation to prices.
← 9. Pension liabilities equal to the value of notional accounts in the Swedish NDC scheme.
← 10. Optimistically assuming annual dividends equal to 5% of assets means that 40% of the dividends would be around 0.5% of GDP (=23% (asset value in GDP) * 5% (annual dividend) * 40% (part of dividend to finance pensions)=0.46%).
← 11. Data based on releases of the Slovenian Statistical Office data: https://www.stat.si/StatWeb/en/News/Index/9322 and https://www.stat.si/StatWeb/en/News/Index/7179.