This chapter looks into pension developments over the past two years. It presents an overview of pension reforms introduced in OECD countries between September 2023 and September 2025. The chapter also describes recent demographic trends and ageing projections. The section on employment at older ages provides an overview of bonus/penalty pension schemes, of combining work and pension practices and of mandatory retirement ages in OECD countries.
Pensions at a Glance 2025
OECD and G20 Indicators
Abstract
Over the next 25 years, populations in OECD countries will age almost twice as fast as over the last 25 years. Past projections have systematically overestimated total fertility rates, and even the most recent projections are built on the assumption that total fertility rates will stabilise at current levels on average. However, long-term projections of life expectancy at older ages have been little affected by COVID‑19 and life‑expectancy gains are still projected to be lower than between the mid‑1990s and the early 2010s when they were exceptionally strong.
Pensioners who want to work, still face obstacles to do so in many OECD countries. Half of OECD countries have at least some restrictions to work while receiving a contributory pension after the normal retirement age, and two‑thirds have such restrictions before that. Moreover, half of OECD countries allow or require mandatory retirement practices for private‑sector workers and over two‑thirds do so for public-sector workers or civil servants.
Chile and Mexico undertook systemic reforms in their pension systems over the last two years. Chile has boosted its earnings-related pensions through a sharp increase in the mandatory contribution rate. It also increased redistribution in its pension system by adding several components, including a contribution-based basic pension, a pension supplement for women and higher targeted benefits. Mexico has introduced a huge earnings-related top-up to the mandatory funded defined contribution (FDC) scheme, which changes the nature of its earnings-related pensions. In addition, the Slovak Republic substantially increased its minimum contributory pensions, and both the Slovak Republic and Switzerland increased pensions overall by introducing a 13th month payment.
Increasing retirement ages remains a common strategy to improve financial sustainability of pension systems without reducing pension levels. Alternatively, financial sustainability can be pursued through raising contributions paid or reducing benefit levels. More than half of OECD countries will increase the normal retirement age for future retirees based on current legislation. Only Czechia and Slovenia decided to increase their statutory retirement ages since September 2023, from 65 to 67, and access to early retirement was tightened in the Slovak Republic. Ireland and Korea have increased contribution rates and Japan has raised the contribution ceiling to mobilise more resources for the pension system. Czechia has improved pension finances by reducing future pension benefits. Furthermore, seven countries have made it easier or financially more interesting to combine work and pensions.
Finally, several countries expanded the coverage of certain pension schemes. Ireland has introduced automatic enrolment in FDC pensions, but Lithuania abolished it. Japan, Korea and Mexico have expanded coverage to include one or more types of non-standard workers. The expansion of childcare credits in Korea has significantly increased the pensions of parents taking childcare breaks.
This chapter is structured as follows. The first section looks into population ageing and takes stock of past and projected evolutions in fertility, life expectancy and migration, and their implications for the development of the old-age to working-age ratio. The second section presents employment at older ages and provides an overview of bonus/penalty schemes, combining work and pension practices and mandatory retirement ages in OECD countries. The chapter then turns to pension reforms legislated in OECD countries since the previous edition of Pensions at a Glance.
Key findings
Population ageing
Population ageing will be fast over the next 25 years. On average across the OECD, the number of people aged 65+ per 100 people aged 20‑64 is projected to increase from 33 in 2025 to 52 in 2050 while it was 22 in 2000. The projected increase over this period is particularly strong in Korea, by almost 50 points, and in Greece, Italy, Poland, the Slovak Republic and Spain by more than 25 points.
The working-age population (20‑64) is projected to decrease by over 30% in the next four decades in Estonia, Greece, Japan, the Slovak Republic and Spain and even over 35% in Italy, Korea, Latvia, Lithuania and Poland.
Fertility rates continue to decline in many countries, while past population projections have systematically overestimated the evolution of the total fertility rate. If countries do seek to boost fertility, they should create conditions that help adults have the number of children they desire at the time of their choosing.
Fertility declines threaten the financial sustainability of pay-as-you-go pension systems. As the effectiveness of policies to uphold or increase fertility levels is uncertain, it would be prudent to prepare for a low-fertility future. This could be achieved through parametric reforms or through introducing automatic adjustment mechanisms adapting pensions to total contributions or a proxy thereof, such as growth of the wage bill, GDP or the number of contributors.
Improvements in life expectancy at age 65 have slowed significantly for both men and women compared to the period between the mid‑1990s and the early 2010s. The COVID‑19 pandemic has not affected the long-term projections of life expectancy at age 65.
UN population projections are based on net migration rates over the next 30 years that are two‑thirds of their levels between 1990 and 2020 in the OECD on average.
Working longer
On average across the OECD, 65.5% of people aged 55‑64 and 25.7% of those aged 65‑69 are in employment, compared to 82.5% of those aged 25‑54. In Denmark, Estonia, Iceland, Israel, Japan, Korea, New Zealand and Sweden, the gap in employment rates between people aged 55‑64 and those aged 25‑54 is 10 percentage points (p.p.) or less. That gap is between 25 and 30 p.p. in Austria, Poland and Türkiye, and it is even larger in Luxembourg and Slovenia.
The annual bonus and penalty rates in contributory pension schemes are 4.8% and 4.4%, respectively, on average among OECD countries, close to actuarial neutrality. Within contributory basic, defined benefit or points schemes, Belgium and Luxembourg, as well as Hungary for women, are the only countries that do not apply penalties. Disincentives to work after the normal retirement age are large in Belgium, Costa Rica, Greece, Luxembourg and Türkiye as bonuses to defer pensions are low or do not exist.
There are no restrictions on combining work and pension receipt beyond the normal retirement age in half of OECD countries, and one‑third of countries have no such restrictions before the normal retirement age. Moreover, in Belgium, France, Germany, Greece, Luxembourg, Mexico, the Slovak Republic, Slovenia, Spain and Türkiye, pension contributions are generally paid when pension recipients work beyond the normal retirement age while no or reduced pension entitlements are built up.
Eleven OECD countries do not apply any form of mandatory retirement to either public or private‑sector workers. Half of OECD countries, by contrast, have mandatory retirement practices for both public- and private‑sector workers. In the remaining eight countries, mandatory retirement exists solely for public-sector workers or statutory civil servants.
Current income of pensioners
The average income of people over 65 is equal to 87% of that of the total population on average across OECD countries. Those aged over 65 currently receive 70% or less of economy-wide average disposable income in Estonia, Korea, Latvia and Lithuania on average, and about 100% or more in Israel, Italy, Luxembourg and Mexico.
Recent pension policy measures
Retirement ages and incentives to work longer
The average normal retirement age among OECD countries will increase from 64.7 and 63.9 years for men and women retiring in 2024 to 66.4 and 65.9 years, respectively, for those starting their career in 2024. The normal retirement age will increase in more than half of OECD countries based on current legislation. Future ages range from 62 in Colombia (for men, 57 for women), Luxembourg and Slovenia to 70 years or more in Denmark, Estonia, Italy, the Netherlands and Sweden.
Czechia and Slovenia have raised the statutory retirement age from 65 to 67, to be reached in 2056 and 2035, respectively. Moreover, in Slovenia, the retirement age without penalty with 40 years of contributions will go from 60 to 62.
The Slovak Republic has linked early-retirement conditions to life expectancy. Italy has extended further multiple early-retirement schemes although conditions have been tightened for several of these. Czechia has introduced the option for workers in arduous or hazardous jobs to retire without penalty between 15 and 30 months earlier, and Spain now determines the arduousness or hazardousness of occupations based on occupational accident and sickness-leave statistics.
Czechia, Greece, Japan, Lithuania, Spain and Switzerland have made it easier or financially more interesting for pension recipients to work, and Denmark has increased its tax incentive for working beyond the statutory retirement age.
Benefits and contributions
Chile undertook a systemic reform strengthening the pension systems, improving earnings-related pensions as well as pension protection for low earners. Chile has raised pension benefits for both current and future pensioners and increased contribution rates significantly.
Mexico has introduced a large earnings-related top-up to the mandatory FDC scheme, changing the nature of its earnings-related pensions. It guarantees that old-age pensioners receive 100% of their last monthly salaries, up to the average monthly salary of social security participants, and even after only 20 years of contributions. As the residence‑based basic pension is paid on top of that, the replacement rate for low earners is well over 100%. How this reform will be financed over time is unclear.
Several countries have taken measures to boost women’s pensions. Chile has introduced a benefit compensating women for their lower retirement income due to their higher life expectancy, given that Chile applies sex-specific mortality tables. Mexico has introduced a new residence‑based basic pension specifically for women before the statutory retirement age.
Chile has increased targeted benefits significantly and the Slovak Republic raised the levels of minimum contributory pensions.
Slovenia legislated a comprehensive pension reform, which will improve both the financial sustainability and the equity of the system. Beyond the increase in the retirement age, the reference wage period for the calculation of benefits has been extended from the best 24 to the best 35 years, benefit accrual rates have been increased and pension indexation has been lowered.
To improve the financial sustainability of public pensions, Ireland and Korea have raised contribution rates, Japan has increased its contribution ceiling and Czechia has reduced future benefit levels.
Beyond Chile and Mexico, Korea, the Slovak Republic and Switzerland have increased benefits from mandatory earnings-related pensions.
Taking into account all legislated measures, full-career average‑wage workers starting their career at age 22 in 2024 will receive on average a net pension at 63% of net wages. Future net replacement rates are below 40% in Estonia, Ireland, Korea and Lithuania. The future net replacement rate of full-career workers earning half the average wage is higher at 76% on average.
Coverage
Ireland has introduced automatic enrolment in occupational pensions, while Lithuania abolished it.
Japan, Korea and Mexico have expanded coverage to include one or more types of non-standard workers.
Korea expanded childcare credits for parents, which will significantly increase their pensions.
Population ageing is driven by changes in three factors: fertility, life expectancy and migration. This section briefly looks into past trends and future projections of each of these factors, and of the resulting old-age to working-age ratio. As the relative importance of these three factors in population ageing can differ across countries, the last part of this section provides a decomposition of changes in the old-age to working-age ratio over the past 10 years by driver of population ageing.
Declining fertility
Total fertility rates (TFRs) halved on average across OECD countries since the 1960s. Increased educational attainment among women, improved access to effective contraceptive measures, a growing predominance of dual-earner households often grappling to reconcile work and family commitments, and increased economic, labour market and housing insecurities especially among younger people have all contributed to declining birth rates (OECD, 2024[1]). This trend may further have been spurred by changes in attitudes towards parenthood. Indeed, men and women increasingly find meaning outside of parenthood, while more intensive parenting norms emerged. More gender equality in households has exposed more fathers to the need to better balance time between work and family life. At the same time, family and care policies such as paid leave and formal early childhood education and care services have been strengthened to support families and help working parents balance work and family responsibilities.
Low fertility challenges the financial sustainability of pay-as-you-go pension systems. A total fertility rate below the population replacement level of 2.1 children per woman results in each future generation being smaller than the previous one, and thus a higher old-age to working-age ratio. While a low fertility rate entails a higher pressure on working-age people, pension systems’ parameters (retirement age, pension level and contribution rate) can be set in a financially sustainable way. Declining fertility requires regular reassessment of these parameters. Keeping a pension system financially sustainable in a context of low fertility is politically challenging in particular in the absence of automatic adjustment mechanisms; such an absence makes pension systems especially sensitive to the uncertainty around fertility-rate projections.
Projections have systematically overestimated the total fertility rates, and have therefore underestimated the pace of population ageing. Invariably, projections have assumed that the decline in the total fertility rate would stop around the time the projections were published and start increasing again soon after, only for the next edition to reveal that the trend reversal did not happen – except for a brief period between 2005 and 2010 (Figure 1.1). Estimates of the total fertility rate in 2025 have been corrected downward with almost every new edition: while the 1994 edition still foresaw a total fertility rate of 2.01 in 2025 on average across OECD countries, by the 2024 edition the estimate had decreased to 1.46.
The most recent projections still display a trend reversal around the time of the projection, but do not assume a substantial rebound in fertility levels. Projections in the 1990s assumed a quick return to the replacement level of 2.1 live births per woman by the end of the projection horizon in 2050, although editions since 2012 project a milder increase over the rest of the century. Under the 2024 projections, the average total fertility rate across OECD countries is projected to reach its lowest point in 2025, at 1.46, after which it would slightly increase.
Evolution of the OECD-average total fertility rate in different projections, 1980‑2070
Note: The lines refer to estimates and medium-variant projections for the 1994, 1996, 2002, 2006, 2012, 2015, 2017, 2019, 2022 and 2024 editions of the World Population Prospects. As data are only available for five‑year periods before 2022, the data are smoothed over a five‑year period to produce annual estimates.
Source: United Nations, Department of Economic and Social Affairs. World Population Prospects 1994‑2024: http://population.un.org/wpp/.
As much as possible, pension systems should be resilient to low fertility, which is a challenge for policymakers. The impact of the decline in fertility on the number of people contributing to the pension system can to some extent be mitigated by higher employment rates, in particular of women and older people (OECD, 2025[2]). Yet, given the uncertainty around the evolution of both fertility and employment rates in the future, it would be prudent to prepare for a low-fertility future (OECD, 2024[1]). For pension policy, this could be achieved through parametric reforms or through automatic adjustment mechanisms adjusting pensions to total contributions (Box 1.1). Adjusting pensions to total contributions not only accounts for changes in the size of the working-age population, but also for changes in productivity reflected in wage growth. If countries do seek to boost fertility, they should create conditions that help adults have the number of children they desire at the time of their choosing. Falling teenage fertility rates, rising female education levels and rising female employment rates are major accomplishments, which improve women’s well-being and reduce their old-age poverty risks. In modern societies, countries that are concerned about fertility rates should promote more gender equality and fairer sharing of work and childrearing. This involves providing family policies that help the reconciliation of work and family life, but policy must also have a greater focus on the costs of children, especially housing costs (OECD, 2024[1]). However, it is unlikely that such policies will enable countries to approach replacement fertility rates again.
In the face of declining fertility, the financial sustainability of the pension system can be improved through adjusting pensions to changes in total contributions. For pay-as-you-go pension schemes to be sustainably financed from contributions, the effective rates of return they generate on contributions should be equal to the system’s internal rate of return. When redistributive instruments are financed by external sources (i.e. not by pension contributions), a pay-as-you-go pension system provides an internal rate of return equal to the growth rate of total contribution receipts. In a system with a constant contribution rate, total wage‑bill growth is a good proxy for the growth rate of total contributions. In turn, the total wage‑bill growth is equal to the sum of the growth rates of the average wage and of total employment.
This is why in a generic NDC scheme the notional interest rate is equal to the growth rate of the contribution base: with such a notional rate, the scheme does not become financially unsustainable when fertility declines. Latvia and Poland use the growth rate of the total wage bill and Italy uses GDP growth as the notional interest rate applied to NDC accounts, all proxies of the growth rate of the contribution base. The notional interest rate in the NDC scheme that is being phased out in Greece is the growth rate of total contributions. Norway and Sweden, in contrast, use average‑wage growth as the notional interest rate, and therefore do not account for the evolutions in the size of the working-age population – although this is less of an issue for these countries as Sweden’s working-age population is projected to remain stable over the next 40 years and Norway’s to shrink to a much smaller extent than in other NDC countries (Chapter 6, Figure 6.6).
Some countries adjust to growth in total contributions or a proxy thereof in DB or points systems. In Estonia, the value of a pension point is adjusted for 80% to total contributions and for 20% to price growth, affecting both new pensions and pensions in payment. Lithuania adjusts the point value fully to wage‑bill growth. Japan corrects for declines in the number of contributors to public pensions. Finally, Greece and Portugal partially index pensions in payment to GDP, depending on economic circumstances.
Slowing life‑expectancy gains
After a period of much faster longevity growth between the mid‑1990s and the early 2010s than before, improvements in life expectancy at age 65 have slowed significantly for both men and women. On average in all 38 current OECD countries, the estimated trend in life expectancy at age 65 shows an increase at a pace of around 1.6 years for men per decade and 1.4 years for women during that period of faster life‑expectancy increases (Figure 1.2). Since about 2012, this pace has almost halved at 0.9 and 0.8 years per decade for men and women, respectively.
Annual change in the trend of remaining life expectancy at age 65 in the OECD on average, in years
Note: The breaks are significant at the 99% confidence level. To limit interferences from short-term fluctuations in change in period life expectancy, the breaks are estimated on the Hodrick-Prescott filtered trend series (lambda=100).
Source: See Chapter 6, Figure 6.4, https://stat.link/gkc90x.
These life‑expectancy gains have mostly been in good health. According to WHO data, the share of life expectancy at age 60 spent in good health has remained constant in OECD countries since 2000, around three‑quarters of life expectancy at that age (OECD, 2023[5]; 2025[2]). Hence, people not only live longer, they largely do so in good health as well. This illustrates that the relationship between age and health evolves over time. For instance, people in the United States have become biologically “younger” at any given chronological age since the 1980s (Levine and Crimmins, 2018[6]).
Despite COVID‑19, long-term projections of life expectancy at age 65 have been fairly consistent over the last decade. While UN Population Prospects in its 2002 and 2006 editions underestimated improvements in life expectancy in the 2000s, later projections are more consistent across editions (Figure 1.3). Although life expectancy at age 65 fell sharply from 2020 due to COVID‑19, the impact is projected to be temporary and future life‑expectancy levels would resume their pre‑COVID trend. For the OECD on average, life expectancy at age 65 is projected to increase by 1.0 year over the next decade, slowing slightly to 0.9 years per decade around 2050.
OECD-average remaining period life expectancy at age 65, in years, in different editions of the UN World Population Prospects, 1980‑2070
Note: The lines refer to estimates and medium-variant projections for the 2002, 2006, 2012, 2015, 2017, 2019, 2022 and 2024 editions of the World Population Prospects. As data are only available for five‑year periods between before 2022, the data are smoothed over a five‑year period to produce annual estimates.
Source: United Nations, Department of Economic and Social Affairs. World Population Prospects 1994‑2024: http://population.un.org/wpp/.
Trends in migration
In the OECD on average, over the next 30 years, the net migration rate is projected to be well below the rate observed between 1990 and 2020 based on UN population projections (Figure 1.4). The OECD-average net annual migration rate is projected to be 1.6 migrants per 1 000 inhabitants per year between 2025 and 2055, whereas it was 2.5 per year between 1990 and 2020. Between 2000 and 2020, the net migration rate has consistently exceeded 1.6 migrants per 1 000 inhabitants except in 2010, in the wake of the 2008 financial crisis. The peak in 2022 is to a large extent driven by an influx of people fleeing Russia’s war of aggression against Ukraine, with net migration rates in most European countries exceeding the 2019 rate.1 The net migration rate increased particularly sharply in Czechia, Estonia, Lithuania and Poland, where the 2022 rate exceeded the 2019 level by more than 20 points.
OECD-average net migration rate per 1 000 population
Source: United Nations, Department of Economic and Social Affairs. World Population Prospects 1994‑2024: http://population.un.org/wpp/.
Old-age to working-age ratios will be increasing at a fast pace by 2050
Trends in population ageing differ depending on the chosen demographic ageing indicator. The old-age to working-age ratio is the most commonly used demographic measure in relation to pension systems as its changes provide a proxy for changes in the number of potential beneficiaries relative to the number of potential contributors at stable retirement ages. The median age of the total population is one direct measure splitting, by definition, the entire population equally between those younger and those older than the median age.
The old-age to working-age ratio will increase fast over the next 25 years. On average across the OECD, the number of people aged 65+ per 100 people aged 20‑64 has increased from 22 in 2000 to 33 in 2025, and is projected to reach 52 in 2050 (Figure 1.5). Fast population ageing is partly driven by the baby-boom generation moving from the working-age into the old-age side of the fraction. The projected working-age population will decrease by 13% in the OECD on average over the next four decades, and by over 30% in Estonia, Greece, Japan, the Slovak Republic and Spain and even over 35% in Italy, Korea, Latvia, Lithuania and Poland (Chapter 6, Figure 6.5). As more people in that generation will die, population ageing will eventually slow. The increase in the old-age to working-age ratio over 2025‑2050 is projected to be particularly strong in Korea, about +50 points, that will overtake Japan as the OECD country with the highest ratio from around 2050. In Greece, Italy, Poland, the Slovak Republic and Spain, this ratio is projected to increase by at least 25 points over this period, while it would increase least in Israel (+5 points) and in Finland, Sweden and the United States (less than +10 points).
When assessed using the median age as an indicator, population ageing accelerated earlier and will slow down earlier as well. In contrast to the old-age to working-age ratio, which accelerated around 2010, the median age has been increasing faster since the 1980s. The increase in the median age will start slowing down earlier as well, around 2040. This is the result of the fall in fertility rates having an immediate impact on the median age as there are fewer children, but it takes one generation to affect the old-age to working-age ratio as the latter does not consider those under age 20.
Number of people older than 65 years per 100 people of working age (20‑64), 1950‑2100
Source: United Nations, Department of Economic and Social Affairs (2024). World Population Prospects 2024: http://population.un.org/wpp/.
Beyond the OECD average, in most countries, the old-age to working-age ratio is projected to increase faster over the next 25 years than over the previous 25 years. All OECD countries saw their old-age to working-age ratio increase, but by less than 20 points between 2000 and 2025, except Japan and Korea where the increase was higher (Figure 1.6). Between 2025 and 2050, the ratio is projected to grow by between 10 and 30 points in most countries. The change over this period is only projected to be lower in Israel, Finland, Sweden and the United States, while it would be higher in Italy, Korea and Spain. The old-age to working-age ratio will accelerate particularly fast in Luxembourg, Mexico, Spain and Türkiye, where it is projected to grow over three times faster between 2025 and 2050 than it did since 2000. In Finland, by contrast, the ratio is expected to grow at a significantly smaller pace. Overall, the ratio is projected to grow at a slower pace again after the middle of the century, and even decline in a few countries, in particular Portugal and Slovenia. In Chile, Colombia, Costa Rica, Iceland, Lithuania, Mexico and Türkiye, however, the ratio is expected to grow the fastest between 2050 and 2075.
Change in old-age to working-age ratio per 25‑year period, percentage points (p.p.)
Source: United Nations, Department of Economic and Social Affairs (2024). World Population Prospects 2024: http://population.un.org/wpp/.
The relative roles of fertility, life expectancy and migration in population ageing
In some countries, population ageing has mainly been driven by declines in fertility whereas in others rising life expectancy has been the more important factor in recent years. Decomposing the average annual change in old-age to working-age ratio following the method outlined in Box 1.2, fertility and life expectancy on average have had a similar impact on the old-age to working-age ratio over the last decade (Figure 1.7). In particular in Canada and Iceland, but also in Finland and the Netherlands, the impact of falling fertility well outweighed that of rising life expectancy. In Italy and to some extent also in Denmark and Spain, the reverse is the case, with life‑expectancy having been a much more important driver of population ageing than fertility decline in recent years. While the temporary reduction in life expectancy due to COVID‑19 may have resulted in a reduced importance of the life‑expectancy component, the results are very similar to those of Scott and Canudas-Romo (2024[7]) based on population data until 2019. Finally, migration has mitigated the impact of fertility and life expectancy to some extent. In particular in Canada and Iceland, the increase in the old-age to working-age ratio has been significantly lowered by immigration. In France and the Netherlands, by contrast, past migration is estimated to have had little or no impact.2
The decomposition of the old-age to working-age ratio follows the method proposed by Scott and Canudas-Romo (2024[7]), and the results presented here are based on an adjusted version of the code shared by the authors. Following the method, variable‑r decomposition, age‑specific population growth rates are expressed as the sum of the growth rates in births, survivorship and net migration. It is based on cohort data, tracking the size and mortality for each cohort from birth. Migration is treated as a residual: changes in cohort size that are neither the result of changes in birth rates nor in mortality rates, are attributed to migration.
As cohorts are followed from birth, long and uninterrupted data series on births, age‑specific population size and mortality are required: to determine the relative importance of these three drivers in the change in old-age to working-age ratio between 2013 and 2023, and assuming a maximum age of 100 years, data have to cover the full lives of each cohort from the 1912 birth cohort onward. For 10 OECD countries in the Human Mortality Database, data are available to decompose the old-age to working-age ratio over the period 2013‑2023, and assuming a maximum age of 100. In addition, Belgium and Canada are included by setting a maximum age of 90 and Australia and the United Kingdom with a maximum age of 89. The lower maximum age does mean that changes in mortality over age 90 are not taken into account, resulting in an underestimation of the life‑expectancy component in the decomposition. Among countries for which full data are available, the life‑expectancy coefficient is 16% lower if a maximum age of 90 instead of 100 years is applied. Hence, for the four countries with data only available to 89 or 90 years only, the life‑expectancy component is increased to compensate for the underestimation based on this 16% estimate, keeping the total change in old-age to working-age ratio constant.
Source: Scott and Canudas-Romo, (2024[7]), “Decomposing the Drivers of Population Aging: A Research Note”.
The average annual change in old-age to working-age ratio over a ten‑year period, decomposed by driver, 2013‑2023 or latest available
Note: Data for Denmark refer to 2014‑2024, for France, Italy, the Netherlands and the United Kingdom to 2012‑2022, and for Australia to 2011‑2021. * While for other countries, the results are based on population data until age 100, for Belgium and Canada population data this is limited to age 90 and for Australia and the United Kingdom to age 89 due to limitations in data availability. As this this means that gains in life expectancy over age 90 are not taken into account, the life‑expectancy component for these countries is increased with the average of the difference in the component when applying the 90‑year cutoff to the countries for which population data until age 100 are available (‑16%), keeping the total change in old-age to working-age ratio constant. Data with the 90‑year cutoff are available in the StatLink.
If current projections become reality and life expectancy continues to rise while fertility remains stable, gains in life expectancy will become the most important driver of population ageing. Based on United Nations population projections, Lee and Zhou (2017[9]) estimate that improvements in mortality will become the main driver of population ageing in advanced economies over the next decades. This marks a break with the past, as they estimate that population ageing over the last century was mostly driven by declining fertility. The picture is different in emerging economies, where fertility would remain the main driver of population ageing until the end of the 21st century, in particular in Sub-Saharan Africa (Lee and Zhou, 2017[9]). Nonetheless, given the importance of fertility declines in population ageing until now, the systematic overestimation of future fertility rates in previous projections means that there is a real risk that current projections underestimate the speed of population ageing over the coming decades.
These shifts in the drivers of population ageing may have important implications for pension policy. As mortality improvements become the more prominent force behind demographic change, adjustments to life expectancy will gain greater importance in efforts to maintain financial sustainability in the pension system. While adjustments both to evolutions in the size of the working-age population and in life expectancy will continue to be needed to maintain sustainability, the increasing importance of life expectancy in population ageing means that automatic adjustments to life expectancy will become more effective tools to maintain financial sustainability in the future. Unlike the cost of ageing due to lower fertility, which is difficult to allocate to any specific cohort as there is no clear beneficiary, it is fair to allocate the cost of higher life expectancy to the cohort that can expect to live longer (Schokkaert and Van Parijs, 2003[10]). This can be achieved through automatically adjusting the retirement age or the pension benefit level to life expectancy (OECD, 2021[3]). Finally, immigration can delay population ageing or slow its pace, but permanently lowering the old-age to working-age ratio would require an ever-increasing net migration rate across cohorts. Hence, immigration could be an effective strategy to “buy time” for countries to adjust to a new demographic reality, but it is not a permanent solution to population ageing.
The employment gap between prime‑age and older workers remains substantial
The employment rate of older age groups remains well below that of prime‑age workers. On average across the OECD, 65.5% of people aged 55‑64 and 25.7% of those aged 65‑69 are in employment, compared to 82.5% of those aged 25‑54 (Figure 1.8). Less than half of people in the age group 55‑64 are in employment in Luxembourg and Türkiye, compared with more than three‑quarters in Estonia, Iceland, Japan, New Zealand and Sweden. In the age group 65‑69, fewer than one in ten are employed in Belgium, Luxembourg and Slovenia against around half in Iceland, Japan, Korea and New Zealand. Moreover, in Denmark, Estonia, Iceland, Israel, Japan, Korea, New Zealand and Sweden, the gap in employment rates between people aged 55‑64 and those aged 25‑54 is 10 p.p. or less. That gap is between 25 and 30 p.p. in Austria, Poland and Türkiye, and it is even larger in Luxembourg and Slovenia.
Employment rates by age group, 2023
Source: OECD Labour Force Statistics; Australian Bureau of Statistics table LM9 for Australian employment rates 65‑69.
Pension policy is an effective tool to increase employment at older ages, as raising normal and early retirement ages triggers large employment increases. While not everyone affected by increases in retirement ages continues working for the extended period, there is little evidence of more people seeking access to disability or unemployment insurance in response to pension reforms (OECD, 2025[2]). Increases in the number of disability or unemployment beneficiaries due to pension reforms are largely the result of mechanical substitution: people who were receiving these benefits before remain longer in these schemes. In contrast, evidence of behavioural substitution, referring to people seeking access to disability or unemployment benefits in response to a retirement-age increase because they think they cannot continue working until the new retirement age, is limited (OECD, 2025[2]).
Various aspects of retirement and pension policies beyond normal and early retirement ages can affect employment at older ages. Three sets of policies can incentivise, facilitate or impede working longer. First, adequate penalties for early retirement and bonuses for deferral of pension uptake can provide financial incentives to work longer. Second, by making it possible to combine work and pensions, countries can avoid that people leave the labour market when they take up their pension. And third, mandatory retirement practices can stop older workers who want to stay in their jobs after a certain age from doing so. This section provides an overview of these policies in OECD countries.
Incentivising later retirement through bonuses and penalties
Early retirement can be discouraged through high enough minimum retirement ages and penalties before the normal retirement age, while late retirement can be encouraged through bonuses after the normal retirement age. Such penalties and bonuses are typically part of contributory public pension schemes, while residence‑based basic or targeted benefits are generally only available at the normal retirement age (although Canada, Denmark, Finland, and Iceland also increase non-contributory benefits in case of deferral). The higher the bonuses and penalties, the higher the incentives to work longer. Actuarial neutrality defines the bonus and penalty levels that are neutral for pension finances over time. Hence, actuarially neutral bonuses and penalties provide flexibility in retirement timing without affecting pension finances: higher (lower) than actuarially neutral penalties (bonuses) generate savings for public finances, and encourage (discourage) working longer (Box 1.3). Bonuses and penalties below actuarially neutral rates are effectively an implicit tax on employment of people around the retirement age, as an extra year worked results in a decline in pension wealth (Blöndal and Scarpetta, 1999[11]).
On average across contributory basic, DB and points-based pension systems in OECD countries, the actuarially neutral rate for anticipating or deferring pension by one year is 4.8%, ranging from below 4% in Luxembourg and Slovenia to around 6% in Estonia and the Slovak Republic (Figure 1.9, Panel A). This among others reflects differences in remaining life expectancy at the future normal retirement age. Estonia and the Slovak Republic currently have a relatively low remaining life expectancy at age 65 and their retirement age will increase at the same pace as life expectancy. By contrast, the normal retirement age is set to remain at 62 in Luxembourg and Slovenia.
All countries except Colombia, Costa Rica, Greece,3 Ireland, Israel, Türkiye and the United Kingdom allow for early retirement before the normal retirement age in their contributory basic, DB or points schemes. Deferring the uptake of contributory pensions is possible in all countries except Colombia.
Actuarial neutrality is a central indicator for the assessment of the size of this bonus or penalty and thus for the assessment of work incentives around retirement ages. When individuals defer their pensions and work past the retirement age, they should not only build up new entitlements but also receive a higher pension benefit from previously built-up entitlements as they will receive the benefits for a shorter period. Conversely, when retiring earlier, pensions should be lower. Actuarially neutral pension schemes ensure that at a given age (e.g. at the normal retirement age) a worker is overall financially indifferent in terms of contributing to and receiving pensions between retiring and working an extra year – that is, taking up the pension one year later does not change the total amount of already accumulated pensions the person can expect to receive in its life. A bonus on accumulated entitlements for deferring pension receipt that is larger than implied by actuarial neutrality provides financial incentives to work longer but is costly for the pension provider; a bonus that is lower than would be consistent with actuarial neutrality effectively is a disincentive to continue working. Similarly, penalties exceeding the actuarially neutral rate disincentivise early retirement whereas penalties falling short of the actuarially neutral rate make it financially more interesting to retire early.
The calculation of actuarially neutral rates for bonuses and penalties in a given pension scheme depends on four key determinants: the retirement age, mortality rates, pension indexation and discount rates. They do not depend on the other parameters used to compute pension benefits. Country-specific rates decrease with remaining life expectancy at the normal retirement age and with shifting for example from price to wage indexation as a lower bonus is needed to incentivise working longer if remaining life expectancy is longer and pensions grow at a faster rate during retirement. Therefore, part of the cross-country variation in actuarially neutral rates relates directly to differences in the retirement age as rates are low in case of a long period of pension receipt and high in case of a short period of receipt.
Source: OECD, (2017[12]), Pensions at a Glance 2017.
In several OECD countries, bonus and penalty rates within these schemes deviate significantly from actuarially neutrality. The average effective bonus is at the actuarially neutral rate, 4.8%, and the average effective penalty is slightly below at 4.4% (Panel A). Belgium and Luxembourg as well as Hungary for women are the only countries that do not apply penalties within such schemes in case of retirement one year before the normal retirement age – although Belgium is in the process of legislating a bonus-penalty scheme (see Recent pension reforms). By contrast in Canada, the penalty is over 2 p.p. above the actuarially neutral rate, generating strong disincentives to retire early. This is also the case in the occupational scheme in Switzerland, although to some extent this is offset by a penalty below the actuarially neutral rate in the public scheme.
Belgium, Greece, Luxembourg and Türkiye currently do not provide a bonus for deferring pension benefits, and the bonus in Costa Rica’s DB scheme is 3 p.p. below the actuarially neutral rate: this provides disincentives to delay pensions beyond the normal retirement age. In France, the lack of a bonus in the mandatory occupational scheme diminishes the incentives to work longer provided by the 5% annual bonus in the main public mandatory scheme (régime général). Korea, Lithuania and the United States provide a bonus of 2.5 p.p. above the actuarially neutral rate; in Canada and Japan it is 3.5‑4 p.p. above that level; and, Portugal’s bonus is even double the actuarially neutral rate. Bonuses well in excess of the actuarially neutral rate can provide strong incentives to delay claiming a pension but can also generate significant financial costs to the pension system.
DC pensions do not have explicit bonus and penalty rates, but they have built-in adjustments of benefits that can be received every month to the length of the retirement period. In FDC, the adjustments are actuarially fair by construction whether through lump sums or annuities. In NDC, the annuity conversion factor used to turn the notional capital into an annuity takes into account remaining life expectancy at the time of claiming the pension.
Early retirement is generally not possible in residence‑based basic and targeted pension schemes, but some countries do apply a deferral bonus in these schemes (Panel B). While deferral of non-contributory benefits is possible in most countries, only some provide a deferral bonus. Canada and Denmark have a deferral bonus in their residence‑based basic schemes, respectively, at about 2 and 1 p.p. above the actuarially neutral rate. Targeted benefits are only increased for deferral in Denmark, Finland, Iceland and Norway. Unique in allowing the early take‑up of a targeted benefit, following a recent reform, Iceland now calculates actuarially neutral bonus and penalty rates for each combination of cohort and age (see Recent pension reforms).
Actuarially neutral rate versus bonus/penalty rates applying when retiring one year after/before the normal retirement age, by type of scheme
Note: Bonuses and penalties applying to a person entering the labour market at age 22 in 2024, and retiring one year after and before the normal retirement age, respectively. No mark for bonus/penalty means that early/deferred retirement is generally not possible in the scheme. The actuarially neutral rate presented is the average of the rates for a bonus and penalty for retiring one year after/before the normal retirement age, for men and women combined. The actuarially neutral rates are on average about 0.2 p.p. higher/lower if calculated specifically for a one‑year deferral/anticipation. For France and Switzerland, the mandatory occupational scheme (O) is included separately from the public DB scheme (P for Switzerland, RG, Régime Général, for France). 1. Belgium does have a flat-rate incentive to work beyond becoming eligible to retirement, although the government has concrete plans to replace this with a bonus-penalty scheme starting initially at 2% and increasing to 5%. 2. The data for Czechia are the combined result for the contributory basic pension (0% bonus/penalty) and the earnings-related pension (6% bonus/penalty) for an average earner. 3. In Hungary, early retirement without penalty is only possible for women as men cannot claim a pension early.
Source: Table 3.6 and OECD calculations.
One attractive alternative to the traditional bonus, which increases the monthly pension until death, is a lump-sum benefit for deferring pension uptake. Spain introduced the option to have its 4% deferral bonus paid out as a lump sum to further incentivise delaying retirement in 2021. The lump-sum option might be a good tool to nudge some people into delaying retirement as survey research indicates that some people prefer receiving the lump sum over the 4% bonus (Ministerio de Inclusión, Seguridad Social y Migraciones, 2021[13]). However, it has been estimated (BBVA, 2022[14]) that the choice between both options is far from an actuarially neutral one as the lump sum would be well below what most people could expect to receive actuarially from the 4% increase in their monthly pensions. Following the pension wealth calculation (Chapter 4), a full-career average‑wage earner retiring in 2024 can expect to receive in actuarial terms around half the deferral benefit if taken out as a lump sum compared to the monthly bonus.4
In 2024, Belgium introduced a flat-rate deferral benefit that increases with the deferral period up to a maximum reached after three years of deferral. Flat-rate benefits mean that the incentive is relatively more meaningful for lower pensions. Moreover, the deferral benefit level depends on career length and can be taken up as a lump sum or monthly. The career-length conditions as well as a pro-rata reduction of the bonus in case of part-time employment during the period of deferral may undermine incentives to delay retirement: the bonus and penalty should adjust pension benefits for the expected duration of pension receipt, so a person’s labour market status should not matter. For a person working full-time during the deferral period, the lump sum is financially more attractive than the monthly benefit.5 The new government plans to replace the flat-rate deferral benefit with a bonus-penalty scheme of 5% per year from 2040.6
Another parametric alternative to a bonus, is an increased accrual rate for each year worked after fulfilling the career-length or age requirement to claim a pension. For instance in Hungary (OECD, 2024[15]) and Slovenia (OECD, 2022[4]), accrual rates are higher after 40 years worked. Accrual rates can be set in a way to mimic an actuarially neutral bonus for specific career profiles, although they may generate different incentives for people with different career profiles. Moreover, a bonus may be more visible than an increased accrual rate, and thus more effective to delay retirement.
Combining work and pensions
On average among European countries, about one‑fifth of pensioners who are younger than 70 years continue working during the first six months after first receiving a pension. Over 40% of recent pensioners in the Baltic States and Norway and around one‑third in Finland, Iceland and Sweden do so (Figure 1.10). Among European countries without restrictions on combining work and pensions before or after the normal retirement age in the OECD (see below), Denmark is the only one having a below-average rate of people continuing to work after retirement. Motivations are very different, however (Eurostat, 2023[16]): the majority of those working beyond retirement in the Baltic States indicate financial reasons, whereas in Norway the majority indicates to continue working out of joy for the work itself. In Finland and Sweden, motivations are more mixed, especially in Sweden where one‑quarter of those who continue to work say they primarily do so to remain socially integrated – a much higher rate than any other European country. On the other extreme, about one‑tenth of recent pensioners or less combine work and pensions in Belgium, France, Greece, Italy, Luxembourg, Slovenia and Spain.7
Share of recent pensioners (aged 50‑69) who continued working during the six months following the receipt of their first old-age pension in Europe, 2023
Source: Eurostat, table lfso_23pens06.
Restrictions on combining work and pensions may be harmful to efforts aiming at extending working lives beyond the normal retirement age. OECD countries vary strongly in how they regulate combining work and contributory pensions for private‑sector employees, and often apply stricter rules before compared to after the normal retirement age (Table 1.1).
Employment restrictions to combine work and pensions in contributory pension schemes, private sector
Note: No information for Iceland, Israel and Korea. Ireland and New Zealand do not have early-retirement options in their (quasi-)mandatory residence‑based basic pension schemes, and have no restrictions on combining work with that basic pension after the normal retirement age; in the Netherlands, conditions for early retirement in the quasi-mandatory occupational pension schemes are sector-specific. * In Hungary, early retirement is only possible for women. a. The data for Australia refer to the earnings-related Superannuation; the Work Bonus, which reduces the amount of earnings from work taken into account in the income test of the targeted Age Pension, was permanently set at AUD 300 per fortnight, currently 11% of average earnings. b. In Canada, the residence‑based basic pension is withdrawn at 15% against income (including earnings) exceeding 106% of economy-wide average earnings. The benefit cannot be taken up before the normal retirement age. c. In Finland, the employment contract does not have to be terminated to claim a pension when the upper age limit for taking up the pension, currently 68, is reached. d. In Spain, there is a requirement to defer pension uptake by at least one year before a pensioner can work after the normal retirement age.
Source: Information provided by the countries, and OECD, (2022[4]), OECD Reviews of Pension Systems: Slovenia.
There are no restrictions on combining work and pensions after the normal retirement age in half of OECD countries, and one‑third of countries do not restrict combining work and pensions before the normal retirement age either. There are no OECD countries that do not allow people to combine work and pension receipt at any time. In between those extremes, countries do allow pension recipients to receive earnings from work, but various conditions or limits apply. These include a requirement to terminate the employment contract to claim a pension, limits on hours worked or earnings above which pensions are reduced, or lower build-up of new pension entitlements given contributions paid.
Eight countries require that the employment contract is terminated to access pension benefit after the normal retirement age. Costa Rica, Finland, France, Italy, Mexico, Poland, Portugal and Türkiye only grant a pension after the employment contract has been terminated. The mandatory termination of the employment contract means that older workers are likely to be offered poorer working conditions when combining work and pensions compared to before claiming a pension. Finland only allows pension recipients to continue working for the same employer immediately after claiming a pension if the nature of the job is different, and France and Portugal have waiting periods for people to return to their old employer. The intention of these limitations mostly appears to be to avoid that people claim an old-age pension while planning to continue working.
Three countries reduce pensions when combining work and pensions after the normal retirement age, two of which only do so under some conditions, effectively serving as a labour tax on pensioners. Slovenia only pays out 40% (and even 20% after three years) of the pension if the person performs any kind of paid work.8 In France, the sum of pension income and earnings cannot exceed individual’s earnings before claiming a pension for people with an incomplete insurance record who retired before the normal retirement age (i.e. without a full pension). In Japan, if the sum of the earnings-related pension and earnings exceeds 111% of economy-wide average earnings, the earnings-related pension is reduced by half of the excess amount.
In ten countries, pension contributions are generally paid when pension recipients work beyond the normal retirement age while no or reduced pension entitlements are built up. This practice is de facto a tax on employment of pension recipients. This is the case in Belgium, France, Greece, Luxembourg, Mexico, the Slovak Republic, Slovenia and Spain, as well as in Germany and Türkiye where it only concerns employer contributions.9 In Belgium, Germany, Luxembourg and Türkiye, as well as in France for those without a full pension, pension contributions are paid on earnings beyond the retirement age, but no more pension entitlements are built up.10 Belgium and Germany do have special employment statutes with earnings limits (flexi-jobs and mini-jobs, respectively) accessible to pensioners through which workers can be exempted from paying pension contributions. In the Slovak Republic, contributions paid by working pension recipients only deliver half the normal amount of pension points. Greece and Spain levy a supplementary contribution of 10% and 9% of earnings, respectively, that does not result in a higher pension. Spain, moreover, is the only country that requires that pension uptake is deferred with at least one year before a person can combine work and pension receipt after the normal retirement age.11
Countries tend to apply stricter rules for combining work and pensions before the normal retirement age. Czechia, Latvia, Lithuania and Slovenia do not allow people to work while receiving early-retirement benefits. Another 13 countries require that employment contracts are terminated to claim a pension before the normal retirement age. In addition to the eight countries that require this to claim a pension after the normal retirement age (see above), it concerns Australia, Austria, Luxembourg, the Slovak Republic and Spain. Several countries provide exceptions to this rule in the case of partial retirement, so as to allow people to gradually reduce working hours in their current job and topping up their earnings with pension benefits.
Nine countries apply earnings limits to the amount of work a person can do while receiving an early-retirement benefit. France and Japan apply the same limits before as after the normal retirement age. Earnings limits tend to be much stricter before than after the normal retirement age: Austria, Belgium,12 Luxembourg and the Slovak Republic suspend early-retirement benefits above a very low earnings limit, below 20% of economy-wide average earnings, only allowing for small part-time or occasional employment. Thresholds are higher in Poland and the United States, at 70% and 33% of economy-wide average earnings, respectively. Italy has different limits depending on the early retirement scheme: there is no income limit under regular early retirement rules, but limits do apply to people retiring under special early-retirement schemes such as the Quota system.
Finally, in five countries, pension contributions have to be paid for working pension recipients before the normal retirement age while there is no or reduced build-up of pension entitlements. Belgium, France, Greece, Mexico and Türkiye apply the same rules on no or lower pension build-up given the contributions paid before the normal retirement age as after.
Obstacles to combine work and pensions after the normal retirement age should be removed. Such restrictions unduly constrain choices and therefore limit the well-being of workers. They are at odds with the emphasis on working longer given population ageing. Moreover, removing these obstacles is important as working longer raises individuals’ retirement income and generates positive aggregate effects beyond the pension system, e.g. through higher output and tax revenues. Rules to draw pensions should as much as possible not be linked to work status. Contributors have acquired pension entitlements which they should be able to draw once they meet eligibility conditions, irrespective of whether they work or not; and if they work, irrespective of their earnings, hours worked and employment contract. Likewise, older workers should be able to work irrespective of whether they receive their pension benefits. In addition, in order to efficiently promote more gradual forms of retirement, conditions to withdraw partial pensions should not depend on the amount of work and labour income after the normal retirement age (OECD, 2017[12]).
Mandatory retirement ages
Mandatory retirement rules end the employment of older workers, or allow employers to unilaterally change or terminate employment contracts from a certain age. In its strictest sense, mandatory retirement refers to the law prescribing that the employment relationship ceases when the employee reaches a certain age. The law can also allow employers to end the employment of workers from a certain age, but not oblige them, by including age limits in employment protection legislation or by easing restrictions on layoffs from a certain age. In its “softest” form, mandatory retirement practices can also include regulations that allow employers to unilaterally change employment conditions from a certain age. While such regulations do not necessarily result in the termination of the employment relationship, the lower earnings or job quality it implies, make it much less interesting for older people to remain in employment.
In order to promote longer working lives and give older people more choices, the OECD recommends tackling barriers to employment of older workers. Strictly speaking, mandatory retirement is a matter of labour market regulation and employment protection, although its impacts depend on the eligibility to pensions and the size of the benefits. One of the recommendations to achieve this goal, adopted by the Council of the OECD on Ageing and Employment Policies, is that countries seek to discourage mandatory retirement in close consultation and collaboration with employers’ and workers’ representatives. The OECD does acknowledge that “in a limited number of instances” mandatory retirement practices may be necessary (OECD, 2018[17]). Employers, in the public sector in particular, may struggle more without compulsory retirement in countries where employment protection rules are very rigid (OECD, 2017[18]).
Eleven OECD countries do not apply any form of mandatory retirement to either public or private‑sector workers (Figure 1.11). Half of OECD countries, by contrast, have mandatory retirement practices for both public- and private‑sector workers. In the remaining eight countries, mandatory retirement exists solely for public-sector workers or statutory civil servants. Hence, mandatory retirement is more common in the public than in the private sector in OECD countries.
Japan and Korea are the only OECD countries allowing for mandatory retirement before the normal retirement age in both the private and the public sector, and Ireland does so only in the private sector. Japan allows for private‑sector employers to terminate employment contracts from age 60, five years before the normal retirement age. The law does require companies to guarantee employment until age 65, although this typically includes less generous working conditions (Panel A). In the public sector, employment relationships currently end at 62 (Panel B), although Japan is in the process of increasing it to reach the normal retirement age of 65 in 2031. In Korea, the mandatory retirement age is 60 both in the private and the public sector, despite a current normal retirement age of 63. The age from which private‑sector employment can be terminated was increased from 55 to 60 as of 2017, but from age 55, employees’ wages can be reduced. To limit the impact of seniority wages, the “wage peak system” entails a wage cut for workers aged 55+ – partially compensated by government subsidies – in exchange for employment security until age 60 (OECD, 2018[19]; 2022[20]). Ireland currently still allows for private‑sector employers and employees to agree on a retirement age in employment contracts, most often at 65, although it is in the process of drafting a law that would prohibit mandatory retirement before the statutory retirement age of 66.
Ages from which different mandatory-retirement practices are allowed, and current normal retirement ages (men)
Note: No data are available for Iceland. * For employees, Ireland has no specific age from which it is allowed to end employment contracts, but it is commonplace for employment contracts to include a termination clause at age 65. For civil servants, Ireland has no mandatory retirement age for those who entered service between 2004 and 2012, but the mandatory retirement age of 70 applies to those who entered both before and after this period.
Source: Mandatory retirement ages based on information provided by the countries; normal retirement ages from Table 3.5.
Among OECD countries, mandatory retirement takes different forms as indicated above, with varying levels of strictness. First, mandatory retirement can apply in the strict sense: the legal obligation to terminate the employment relationship at a certain age. This is the type of mandatory retirement most commonly applied in the public sector.13 For private‑sector workers, a legal obligation to end the employment relationship at a certain age only exists in Luxembourg, where the employment agreement is automatically terminated at 65. Workers can be rehired again afterwards.
Second, as a common form of mandatory retirement, employers are allowed to terminate the employment relationship when employees reach a certain age, but they are not required to do so. This is by far the most common type of mandatory retirement for private‑sector workers in OECD countries. It can either be done through allowing clauses in employment contracts or collective agreements to terminate employment at a certain age, or though reducing employment protection at a certain age. Mandatory retirement clauses can for instance be included in contracts and collective agreements in Germany, the Netherlands, Spain and Switzerland. Reduced labour protection typically takes the form of shorter notice periods, limited severance pay and/or a relaxation of the rules on legal reasons for dismissals. This is among others the case in Austria, Belgium, France and Italy. Norway currently has both types of mandatory retirement in the private sector: employers are allowed to terminate the employment relationship when employees reach 72 years, and contracts and social agreements can include a clause automatically terminating employment from age 70 under some conditions. However, from 2026, the option to write a mandatory retirement at age 70 into contracts and collective agreements will be abolished, alongside an increase of the mandatory retirement age in the public sector from 70 to 72. Sweden similarly has both types of mandatory retirement practices, both available when “the right to remain in employment” expires, which is currently at age 69.
Finally, regulations can allow employers to change employment conditions from a certain age, which may result in lower earnings or job quality. This is for instance the case in the wage peak system in Korea, and with the possibility to terminate the employment contract and offer another contract at age 60 in Japan (see above). The conditions in the newly offered employment contract are typically less generous than those in the contract that expired when turning 60 (OECD, 2022[20]; 2024[21]). In Canada, moreover, collective agreements can specify that workers both from the public and the private sector are exempt from certain workplace benefits such as health insurance from the normal retirement age onward.14
Several countries have abolished mandatory retirement practices or increased mandatory retirement ages over the last decades (OECD, 2022[4]). The United States for instance abolished the mandatory retirement age in 1986, and in Denmark, it was abolished in the public sector in 2008 and in the private sector in 2016. In both countries, some exceptions remain for very specific occupations, often where there could be valid health and safety concerns such as air traffic controllers, but also in some other jobs such as judges, police and military personnel. Courts have been playing an important role in reducing mandatory retirement practices or preventing their introduction. In Estonia, the Supreme Court ruled in 2007 that mandatory retirement was unconstitutional. When Slovenia introduced mandatory retirement in 2020, the Constitutional Court initially suspended and subsequently annulled the regulation. Similarly, the Slovak Republic introduced an option for employers to give notice to employees when they turn 65 in 2022, which was suspended by the Constitutional Court, with a final decision yet to be taken. The Court of Justice of the European Union’s rulings offer a framework setting the boundaries within which the practice of mandatory retirement could be considered non-discriminatory and thus lawful (Oliveira, 2016[22]; Dewhurst, 2016[23]). First, the justification should be based on concrete evidence of age having a certain impact on job performance, not mere generalisations or assumptions. Second, any justification for a mandatory retirement age should be occupation- or sector-specific. Safety concerns could be a valid argument for mandatory retirement if there is international agreement that practicing a specific occupation above a certain age could endanger health and safety. And third, the availability of a pension is an important condition for mandatory retirement.
Mandatory retirement ages have been argued for on economic grounds in specific circumstances. A first argument concerns workers’ wages outgrowing their productivity when seniority is a substantial component in wage setting (Lazear, 1979[24]). When older workers cost more than they produce, mandatory retirement is a tool for firms to reduce wage costs without affecting their output (OECD, 2019[25]). There is some evidence that the low mandatory retirement age in France before 2003 was especially used against high-wage earners (Rabaté, 2019[26]). Increasing or abolishing the mandatory retirement age in such a context might reduce efficiency. A second argument is that mandatory retirement makes it possible to terminate employment contracts of less productive workers without facing (the risk of) high costs in countries or sectors where it is difficult or expensive for employers to dismiss such workers (OECD, 2019[25]; OECD, 2017[27]). Finally, some have argued that mandatory retirement leads to the redistribution of employment opportunities between generations, as older workers would free jobs for younger generations (OECD, 2022[4]). Even though there might be a trade‑off between the employment of older and younger workers in some very specific, well-protected sectors, in the economy as a whole job opportunities for younger people are not reduced when keeping older workers in employment longer (OECD, 2013[28]) – the idea that there is a trade‑off is the so-called lump of labour fallacy. To the extent that mandatory retirement in a given country is the consequence of employment and wage regulations, mandatory retirement is only a second-best instrument to deal with difficulties triggered by policies in other areas. The first-best solution would consist in addressing the employment and wage regulations mandatory retirement is meant to circumvent. This could be more difficult to implement in the public sector, however, as civil servants tend to have more stringent employment protection and as productivity generally is more difficult to assess, making a transition from seniority- to performance‑based wage setting more challenging.
This section summarises pension reforms introduced in OECD countries between September 2023 and September 2025. Annex 1.A provides more information about reforms passed during this period.
Changes in retirement ages and incentives to work longer
Normal retirement ages
The average normal retirement age is 64.7 years for men in OECD countries in 2024. The normal retirement age is defined as the age at which individuals permanently working full-time from age 22 are eligible for retirement benefits from all pension components without penalties. It ranges from 62 years in Colombia, Greece, Luxembourg and Slovenia – Türkiye is an absolute outlier with a current normal retirement age of 52 years – to 67 years in Australia, Denmark, Iceland, Israel, the Netherlands and Norway (Figure 1.12).15
Czechia and Slovenia legislated an increase in their statutory retirement ages by two years. In Czechia, the statutory retirement age was already increasing by two months per year until reaching 65 in 2030. Based on the 2024 legislation, the retirement age is set to increase further, but at a slower pace after 2030: it will go up by one month per year until it reaches 67 in 2056. At the same time, eligibility conditions have been relaxed for some people. Those with at least 20 but less than 35 years of coverage could previously only take up their pension five years after the statutory retirement age. This has been reduced to two years. Furthermore, an early retirement scheme for arduous and hazardous occupations has been introduced at the same time (see below).
As part of its substantial pension reform discussed below, Slovenia decided in September 2025 to increase its age thresholds in the pension system by two years between 2028 and 2035 while maintaining relatively short career-length conditions. The statutory retirement age will increase from 65 to 67 conditional on 15 years of contributions, and with 40 years of contributions retirement will be possible without penalty from age 62 instead of 60 previously. Retirement conditions for early starters increase accordingly: currently a person who started working before turning 18 can retire at age 58 provided they made 40 years of contributions, while in the future retirement will be possible from 60 for people who started working before 20. The reform does not change the normal retirement age for Slovenia, however, as a person with a full career from age of 22 can still retire without a penalty upon turning 62 years old.
Overall, based on already legislated measures, the average normal retirement age for men in the OECD will increase by almost two years to 66.4 years for men entering the labour market in 2024. Half of OECD countries will increase the normal retirement age based on current legislation for men. At the same time, cross-country differences are set to become starker: the normal retirement age will remain at 62 in Colombia (for men), Luxembourg and Slovenia, whereas it is expected to reach 70 years in Italy, the Netherlands and Sweden, 71 years in Estonia, and even 74 years in Denmark based on established links between the retirement age and life expectancy (Figure 1.12).16 However, after the Danish Parliament confirmed the increase in the statutory retirement age to age 70 from 204017 in May 2025, Denmark may soften the current one‑to‑one link between retirement age and life expectancy, in which case the projected future normal retirement age would be lower than 74. The eight countries with the highest future normal retirement age are all countries linking retirement age to life expectancy, including also Finland, Portugal and the Slovak Republic. The other OECD country with a retirement-age link to life expectancy is Greece, but the Greek normal retirement age is projected to be just below the OECD average in the future: this is because early retirement is accessible without penalty after a 40‑year career, hence it is the minimum age, which is set to increase from 62 to 66, that determines the future normal retirement age in Greece. Norway is expected to introduce a link and increase its retirement age by two‑thirds of life‑expectancy gains in the near future.
Nine OECD countries still allow single women to retire with a full pension at a lower age than men. Among them, Austria, Lithuania and Switzerland decided to close the gender gap in normal retirement ages by 2033, 2026 and 2028, respectively, while the gap will be reduced in Israel and Türkiye (Chapter 3). Costa Rica and Hungary will maintain a gender gap of two and three years, respectively, while it will remain five years in Colombia and Poland. In Chile, FDC pensions can be accessed by women at age 60 compared to 65 for men, but the targeted scheme (PGU) is only accessible as of 65 for both men and women, which determines the normal retirement ages for both men and women. Mexico has normal retirement age of 65 for both men and women, but is implementing a low, flat-rate benefit paid to women aged 60‑65 (see below). Among G20 countries, gender gaps in the normal retirement age exist in Argentina, Brazil and China and will be maintained in the future.
Normal retirement age for men entering the labour market at age 22 with a full career
Note: The normal retirement age is calculated for an individual with a full career from age 22. “Current” refers to people retiring in 2024. “Future” refers to the age from which someone is eligible to full retirement benefits from all mandatory components (without any reduction), assuming a full career from age 22 in 2024. Educational credits are not included. For better visibility, the scale of this chart excludes the lowest observed value of 52 for current normal retirement age in Türkiye.
Source: See Chapter 3, Figure 3.8, https://stat.link/pgr5v9.
Early retirement and incentives to work longer
The Slovak Republic has tightened eligibility conditions to early retirement, while Italy has extended further multiple early-retirement schemes although conditions have been tightened for several of these. The Slovak Republic has linked the career-length condition for early retirement to life expectancy. While previously, early retirement was possible with penalty after a 40‑year career, the career-length requirement now increases at the same pace as the statutory retirement age, which is linked to life expectancy.18 According to life expectancy projections by the UN, this means that for people entering the labour market now, the career-length condition that will apply when they retire will be 46 years. In addition, the penalty for early retirement based on career length has been increased from 0.3% to 0.5% per month, equalising it with the penalty for retirement two years prior to the statutory retirement age irrespective of career length.
In Italy, the so-called women’s option allowing women to retire early with a 35‑year career, has been extended for the period 2024‑2026, although it can now only be accessed from age 61 instead of 60 previously.19 The pensions of women retiring through this scheme are fully calculated based on notional defined contribution (NDC) rules, generally resulting in lower benefits than when calculated based on defined benefit (DB) rules in Italy. Also, the Quota 103 scheme has been extended for the period 2024‑2025, allowing for early retirement at age 62 with 41 years of contributions, whereas Quota 102 (retirement at 64 with 38 years of contributions) has been abolished. For people retiring through the Quota 103 scheme as of 2024, NDC rules are applied to their full pension. The early-retirement scheme for the unemployed, disabled people, caregivers or people in arduous occupations (Social APE) has also been extended for the period 2024‑2025, but the eligibility age has been increased from age 63 to 63 and five months. The scheme allowing for early retirement in case of restructuring of firms in crisis has been extended without changes, and remains accessible from age 58 with at least 35 years of contributions. Moreover, conditions for early retirement at 64 for people who are only covered by NDC pensions (i.e. people without contributions before 1996) have been tightened. Instead of 20 years of contributions previously, 25 years are needed to retire early (i.e. before age 67) from 2025, and 30 years from 2030.
Several other countries have made adjustments to penalties for early retirement or to bonuses for deferral. Austria has increased the deferral bonus for old-age pensions from 4.2% to 5.1% per full year of deferral, with a maximum of 15.3%. Czechia has halved its penalty for workers who acquired at least 45 years of contributions, from 1.5% per 90 days of early take‑up to 0.75%. Iceland now allows people to defer the uptake of the targeted pension and of the targeted supplement for single pensioners until age 80 against 72 previously. It has also replaced the fixed 6.0% bonus and 6.6% penalty per year with a bonus and penalty specific to each combination of age and birth cohort so as to be actuarially neutral. Ireland has introduced the option to defer claiming the contributory basic pension by up to four years, from age 66 to 70. The annual deferral bonus will regularly be reassessed according to actuarial principles and is bigger for longer deferral: in 2025, the bonus ranges from 4.7% for the first year of deferral to 5.3% for the fourth. Spain has provided some more flexibility in retirement timing in its deferral bonus. Previously, the bonus of 4% per year only accumulated per full year of deferral. Since 2025, the bonus instead accumulates at 2% per six months in case of a deferral of at least 18 months. Finally, while Belgium has introduced a flat-rate deferral benefit in July 2024, increasing with each day of deferral up to a maximum of three years, the new government announced at the beginning of 2025 that it plans to replace the flat-rate deferral incentive with a 5% bonus and penalty conditional on career length.
Denmark and Finland have also made adjustments to incentives to stimulate working beyond the normal retirement age. Denmark has increased its untaxed flat-rate benefit paid annually to people working in the first two years after reaching the statutory retirement age. The benefit is paid as a lump sum to people who on average work at least 30 hours per week over the year, irrespective of whether the public pension’s uptake is deferred. The benefit, which is currently 9.2% of economy-wide gross average earnings for the first year and 5.5% for the second, is set to increase by 30% on top of regular indexation between 2026 and 2029. Finland has increased the age threshold above which earned income is taxed preferentially from 60 to 65 years.
In addition, Czechia and Spain changed rules around early retirement for arduous or hazardous work. Czechia has introduced the option for workers to retire without penalty 15 months before the statutory retirement age if they have worked at least about 10 years (more precisely, 2 200 shifts) in jobs deemed arduous or hazardous, or 30 months before with at least about 20 years (4 400 shifts). Czechia plans on expanding the early-retirement scheme for arduous and hazardous jobs further. Currently, there is a supplementary 2% employer contribution for miners, paramedics and firefighters, giving them access to retirement five years before the statutory retirement age. This contribution could be increased to 5% and the list of occupations expanded to include among others specialised nurses, foresters, blacksmiths and foundry workers, and bricklayers and pavers. The Spanish Government, together with social partners, developed a standardised procedure to determine arduousness or hazardousness of occupations and, connected to that, early-retirement entitlements. Occupation-specific arduousness or hazardousness coefficients are based on the rate of occupational accidents by gender and age, the seriousness of these accidents, and the number and duration of sickness leaves in these occupations. The coefficients are supposed to be reviewed every 10 years.
Combining work and pensions
Several countries have recently made it easier or more interesting for pension recipients to work. Countries have moved in different directions to make it easier for people to combine work and pensions, with some countries introducing rules that others are moving away from. Japan’s earnings limit for combining work and pensions above which the pension is suspended, around the level of gross average earnings, will increase by 24% in 2026. In 2024, Lithuania removed earnings limits for people receiving social-assistance old-age pensions, which are paid to people without the 15‑year career required for the contributory pension. Spain has made the rules for combining work and pension receipt more flexible in 2025. Previously, combining work and pension was only possible for people with a full career (36.5 years in 2024) who deferred uptake by at least one year beyond the statutory retirement age. Only half of the pension benefit and no deferral bonus were paid out during the period of employment, irrespective of working time or earnings. After the reform, combining work and pension is open to anyone who is entitled to a pension, irrespective of having a complete career. It is still required that pension uptake is first deferred for at least one year, although a deferral bonus is now paid out. The pension amount that can then be taken up while working depends on the duration of the deferral: after one year of deferral, 45% of the pension can be taken up; combining work with a full pension requires a five‑year deferral (see above). Greece has replaced the 30% reduction in pension for working pension recipients by a supplementary social contribution of 10% of earnings, for which no supplementary entitlements are built up. Czechia took a different approach by exempting working pension recipients from having to pay the 6.5% employee pension contribution rate, while the employer contribution rate has remained unchanged. For the self-employed combining work and pensions, the contribution rate has also been reduced by 6.5 p.p., from 28% to 21.5%. This 6.5 p.p. reduction replaces the 0.4% pension increase for each year of combining work and pensions. Austria retains relatively strict earnings limits for pension recipients but has added some flexibility for 2024‑2025 that allows people to slightly exceed the limit during some months in the year.20
Spain and Switzerland have increased flexibility in transitioning from working life to retirement through partial retirement, and France has dissociated the age to access partial retirement from the minimum retirement age, which is increasing. In Spain, partial retirement is now allowed while reducing working time between 25% and 75% against only 50% before 2025.21 Switzerland has created the possibility to reduce working time between 20% and 80% and complement it with an inversely proportionate part of the public pension. Partial pension is accessible from two years before the statutory retirement age until age 70, and people can gradually expand pension uptake in up to three phases during this period (i.e. an initial working-time reduction, a bigger working-time reduction and a complete termination of employment). The usual penalty applies to the part of the pension taken up early, or the bonus to the part of the pension that is deferred. France has fixed the minimum age for partial retirement at age 60 in 2025. Previously accessible two years before the minimum retirement age, the accessibility age for partial retirement would have increased from 60 to 62 due to the increase in the minimum retirement age from 62 to 64 that was decided on in 2023.
Adjustments to benefits and contributions
The average income of people over 65 was equal to 87% of that of the total population on average across OECD countries in the latest year available. Older people fare best in Israel, Italy, Luxembourg and Mexico in relative terms, as incomes for the over‑65s were about the same or slightly higher than for the total population (Chapter 7). Older people also had high relative incomes on average in Canada, Costa Rica, France, Iceland, Portugal, the Slovak Republic, Spain and the United States in international comparison. In Estonia, Korea Latvia and Lithuania, by contrast, the income of older people was about one‑third lower.
Systemic reforms strengthening old-age income protection
Chile and Mexico undertook systemic reforms in their pension systems, as did Colombia although the reform has been suspended by the Constitutional Court. Chile has boosted its FDC earnings-related pensions through a sharp increase in the mandatory contribution rate and has added several redistributive components, including a contribution-based basic pension and a pension supplement for women. Colombia passed a reform removing the choice between contributing to the public DB or a private FDC scheme and increasing targeted benefits, although its implementation is uncertain after the Constitutional Court suspended it awaiting substantive review. Mexico has introduced a large earnings-related top-up to the mandatory FDC scheme, which changes the nature of its earnings-related pensions by severing the link between contributions and benefits for a large part of the population. Previously, Mexico had introduced a residence‑based basic pension in 2019 and Chile made its targeted benefit quasi‑universal in 2022. These new reforms add to the trend of Latin American countries increasingly seeking to tackle high old-age poverty, among others resulting from a large informal sector in combination with weak protection for the most vulnerable older people. Furthermore, in Costa Rica, the parliament is currently discussing a law proposal that would introduce a residence‑based basic pension by March 2027.
More specifically, Chile has strengthened its FDC pensions and has introduced three new benefits as part of the early 2025 pension reform. The employer contribution rate will be increased from 1.5% to 8.5% by 2034. Of the 8.5% contribution rate, 4.5 p.p. will flow into individuals’ FDC accounts, raising future pensions, which will generate a sharp increase in FDC entitlements. An additional 1.5 p.p. initially finances a new contribution-based basic pension and guaranteed bonds. Between 2044 and 2056, however, it will gradually be reallocated to FDC accounts and the contribution-based basic pension is supposed to cease to exist and guaranteed bonds will no longer be issued. The remaining part of the new employer contribution rate, 2.5 p.p., flows to existing disability and survivor’s insurance (currently 1.5 p.p.) and to a compensation for women for the part of their lower FDC annuities that is due to their longer life expectancy given the use of sex-specific mortality tables. More precisely, the women’s life expectancy compensation, paid from September 2025, tops up a woman’s annuity so that she would receive the same pension as a man of the same age and with the same amount of FDC savings, provided she retires at 65.22
The new contribution-based basic pension and guaranteed bonds will be financed by state subsidies in addition to the contributions of 1.5 p.p. The contribution-based basic pension will start to be paid to both current and future pensioners from 2026. The eligibility age is 65 for both men and women, although men will need at least 20 years of contributions to qualify whereas women will initially only need 10 years of contributions, increasing to 15 years for new pensioners from 2036, based on the same annual entitlement as for men. The maximum benefit is reached after 25 years of contributions and equals 2.5 UF23 or 8% of the gross average wage. The guaranteed bonds are given to people for contributions paid from March 2025 until 2055, and hence mostly benefit future retirees. They receive an interest rate on their contributions that is tied to that of government bonds. Upon reaching the statutory retirement age in the FDC scheme (65 for men, 60 for women), people can choose whether to use these bonds to increase the life annuity or programmed withdrawal from the FDC scheme, or whether to simply turn them into 240 monthly payments. Guaranteed bonds can be paid out earliest in September 2026.
In addition, Chile has increased the targeted Universal Pension Guarantee (PGU) by 11.6% on top of regular price indexation. The increase is first applied to people aged 82+ receiving up to CLP 250 000 – equivalent to about 21% of gross average earnings or 95% of the average FDC old-age pension – from September 2025. The increase will be applied to those 75+ one‑year later, and to all other old-age pensioners the year after that. The benefit is withdrawn at 56% against the pension received from the FDC scheme instead of 50% previously.24 Based on the OECD pension model, the increase of the targeted benefit will result in the total pension being 3.2% higher for an average earner and 5.2% higher for a low earner with a full career from age 22 in 2024 (Chapter 4).
Chile has furthermore changed some rules in the governance of the FDC pension funds to reduce fees and increase investment choice options. Every two years from 2027, 10% of individual accounts will be auctioned to the administrator that offers the lowest fee. People will have the option to opt out from the procedure, and can switch to another administrator at any time. While currently only five investment options are available, each with a different risk level, in 2027 Chile will move to a system of at least ten target-date funds with cohort-specific investment policies, gradually shifting funds from higher- to lower-risk investments as members approach retirement.
Colombia passed a reform removing the choice between building up earnings-related pensions in a public DB or a private FDC scheme, but its implementation is uncertain after the Constitutional Court suspended the reform in June 2025, awaiting substantive review. Following the reform, pension contributions for earnings up to the threshold of 2.3 times the legal monthly minimum wage (COP 2 990 000 in 2024) would be used to finance the public DB component; contributions paid from earnings above that threshold would flow into the individual FDC accounts up to a ceiling of 25 times the minimum wage.25 The elimination of competition between the public and private pension schemes is welcome as it resolves the issue of inequality in pension benefits for workers with the same career history, while reducing the related administrative complexity. Moreover, while the DB pension currently is only accessible after 25 years of contributions (1 300 weeks), this would be reduced to about 19 years (1 000 weeks); 25 years would still be required for a full pension.
In addition, Colombia would significantly increase the level of its targeted benefit, although it will remain low compared to other OECD countries. The reform would almost triple the targeted benefit to the level of the extreme poverty line, currently COP 223 000, although at 9% of gross average earnings it would remain one of the lowest targeted benefits for older people in the OECD (Chapter 3). Furthermore, while currently people lose the contributions they made if they did not qualify for the earnings-related pension, this would no longer be the case after the reform. For example, individuals who paid fewer than 300 weeks of contributions, would receive the contributions made with a 3% annual interest rate as a lump sum upon retirement. Those with more than 300 weeks but less than the 1 000 weeks required to qualify for the public DB pension would receive an annuity calculated on their contributions paid plus a tax-financed top-up.26 To finance the higher targeted pension, an additional contribution into the Pension Solidarity Fund would be paid on higher earnings as well as on higher pensions.27
The Colombian reform introduces several gender-related changes. The tax-financed top-up to the annuity paid to people not qualifying for a full pension would be higher for women than for men: women’s annuities would be topped up by 30% compared to 20% for men. The contribution requirement for women to qualify for a full contributory pension would gradually be reduced from 1 300 weeks in 2025 to 1 000 weeks. This reduction was included in the reform in response to a ruling by the Constitutional Court in 2023 that having the same contribution requirement for a full pension in combination with a statutory retirement age for women five years earlier than for men is an unconstitutional discrimination based on sex as women would have to attain the same amount of contributions over a shorter period.
Mexico has introduced a large earnings-related top-up to the mandatory FDC scheme, which changes the nature of its earnings-related pensions. The top-up guarantees that old-age pensioners receive 100% of their last monthly salaries, up to the average monthly salary of social security participants at the time of the top-up’s introduction. That ceiling is adjusted to price inflation, so over time the top-up will erode in relative terms, first for average and high earners, and subsequently also for people making below-average earnings. The new benefit was created because pensions are currently low as the pension system is still maturing – the FDC scheme was only set up in 1997, leaving generations in or close to retirement with only partial contribution records – and as the contribution rate is low on top of large informality. This new guarantee applies since July 2024 to everyone aged 65 or over receiving an FDC pension, which requires a contribution period of 825 weeks in 2024, increasing to 1 000 weeks in 2031. This means that this will generate very high pension even for workers with short contribution periods. The scheme is financed from a variety of resources, several of which are one‑time transfers. Hence, it is unclear how the financing measures foreseen for this top-up can cover the promises made in the longer term. The scheme would partially be financed from sleeper accounts – i.e. unclaimed accounts of which the owner cannot be contacted –, which can temporarily raise money but is unlikely to provide a sustainable source of funding.28 Moreover, the primary way to deal with sleeper accounts should be for the government and the pension regulator to make efforts to identify the owners of those accounts and move their funds to their main accounts. As the residence‑based basic pension is paid on top of that, replacement rates for low earners are well over 100% (see below). Effectively, the scheme overrules the proper functioning of the FDC scheme and creates a partially pre‑funded DB entitlement at 100% of last earnings.
Improving pension protection of low earners
Older people are more likely to fall below the relative income poverty threshold than the total population. Across all OECD countries, 13.0% of people aged 66‑75 and 17.5% of those aged 76+ are in relative income poverty, meaning that they have an equivalised disposable income below 50% of the median, compared to 11.4% of the total population (Figure 1.13). The income poverty rate is below 5% for the 66‑75 age group in Denmark, Finland, Iceland, the Netherlands and Norway, and only in Iceland among people aged 76+. By contrast, the Baltic states, Korea and New Zealand have poverty rates above 25% in the age group 66‑74 and even above 40% (except Lithuania) in the age group 76+. Australia, Costa Rica and the United States also face elevated relative‑poverty levels among older people. The poverty rate among people aged 65+ in New Zealand has doubled since the 2023 edition of Pensions at a Glance. This is due to the benefit level of the residence‑based basic pension, the only mandatory pension scheme in the country, falling just below the relative poverty line.
Percentage with income lower than 50% of median equivalised household disposable income
Note: Most recent data are for 2022 except for the following countries: Canada, Costa Rica, Finland, Latvia, the Netherlands, Sweden, the United Kingdom and the United States (2023), Germany and Japan (2021), Australia (2020) and Iceland (2017). Data for Colombia are unavailable.
Source: See Chapter 7, Table 7.2, https://stat.link/2sqwtk.
Australia, Chile, Iceland and Norway have increased targeted benefits. Moreover, Mexico has implemented a new residence‑based basic pension specifically eligible to women before the statutory retirement age. First-tier pensions play a key role in protecting older people against poverty. In particular, non-contributory first-tier pensions (targeted benefits and residence‑based basic pensions) are a primary tool to tackle old-age poverty. Contributory first-tier pensions (contribution-based basic pensions and minimum pensions) are redistributive as well, as eligibility depends on paying contributions but not on the amounts of contributions paid. Chile has introduced a contributory basic pension, and the Slovak Republic has increased the levels of minimum contributory pensions. By contrast, Belgium and Finland have restricted access to some first-tier benefits out of budgetary concerns.
More precisely, the structural reform passed in Chile and described above includes an increase in the level of targeted benefits by 11.6% on top of regular price indexation. In Australia, the housing benefit for renting in the private market which can be accessed by Age Pension recipients has increased by 15% in September 2023 and by 10% in September 2024. Iceland has raised the threshold above which the targeted pension and the targeted supplement for single pensioners are withdrawn against earnings-related pension income from 2.5% to 3.7% of gross average earnings. In Norway, targeted benefits for singles born before 1954, who receive benefits from the old pension scheme, have been increased in 2025 by 2.3% on top of indexation.29 Finland, by contrast, has frozen its housing allowance for pensioners at the 2023 level for the period 2024‑2027 and has tightened its means test.30 The asset threshold above which the benefit is withdrawn, has been lowered and the withdrawal rate increased from 8% to 15%.31
Mexico has introduced a new residence‑based basic pension specifically for women and eligible before the statutory retirement age of 65 years. The benefit is MXN 3 000 paid every two months, i.e. equivalent on a yearly basis to 9% of gross average annual earnings or about half the residence‑based basic pension paid to all people 65+. It is initially paid to women aged 63‑64, and coverage is expanded to younger age groups during 2025 to include all women aged 60‑64. For women living in indigenous or Afro-Mexican communities, the benefit covers the age group 60‑64 from the moment of introduction. The benefit terminates when turning 65, when women receive the same basic pension as men. The benefit is supposed to recognise women’s unpaid work as well as improve their economic autonomy.
The Slovak Republic has increased the minimum pension benefit, linked to the minimum subsistence level: the benefit after 30 years of contributions has increased from 136% to 145% of the minimum subsistence level, and for each extra year of contributions, the rate further increases by 2.5 instead of 2.0 p.p. of the minimum subsistence level up to 39 years of contributions.32 For a person retiring at the current normal retirement age, the benefit increases from 28% to 30% of gross average earnings. Belgium has added a supplementary eligibility condition to access the minimum pension: in addition to the requirement of 30 years worked or credited, 5 000 days (about 16 years) of effective employment33 is now also required. The supplementary eligibility condition is likely to particularly affect women as, among 65‑year‑olds in 2019, credited periods on average made up 39% of women’s careers and 30% of men’s careers, and the average total period of effective employment in full-time equivalents was 14.6 years for women and 19.4 years for men (Schols et al., 2022[29]).34
Improving financial sustainability
Countries have mainly relied on increases in contribution rates to improve the financial sustainability of their public pension schemes. For countries with relatively low public pension benefits, increases in contribution rates can help avoid that pension benefits be reduced further in the future to cope with financial pressure. With gross replacement rates closer to the OECD average prior to the reform (OECD, 2023[30]), Czechia has improved pension finances by reducing future pension benefits. Japan has increased the contribution ceiling, which will reduce financial pressure in the short term but be offset by a corresponding increase in pensionable earnings in the long term. Finally, Slovenia legislated a comprehensive pension reform adjusting multiple parameters including retirement ages (as explained above) in September 2025, which is expected to improve both the financial sustainability and the equity of the system.
In greater detail, Ireland decided to increase the contribution rate for the contributory basic pension: the rate paid by employees and the self-employed will increase from 4.1% to 4.7% between 2025 and 2028, and that paid by employers will increase from 11.15% to 11.75% over this period. The increase in the contribution rate is meant to allow Ireland to retain the statutory retirement age at its current level of 66 years in the near future as a previous government proposal to raise it to 68 by 2039 was cancelled. Korea decided to increase the total contribution rate of 9%, split evenly between employers and employees, to 13% in 2033, in 0.5 p.p. annual increments. This reform is envisaged as a first step in aiming to make the pension system more financially sustainable, with further reforms expected to follow.
Czechia decided to reduce the reference wage and the accrual rate over the period 2026‑2035. Currently, earnings are fully taken into account in pension calculation up to a threshold at 42% of gross average earnings, after which only part of earnings are included. From 2026 onward, in steps of 1 p.p. per year, only 90% of earnings below that threshold will eventually (from 2036) be included. Over the same period, the accrual rate will be reduced from 1.5% to 1.45% per year, in 0.005 p.p. increments.
Japan will gradually increase the contribution ceiling on earnings between 2027 and 2029. The contribution ceiling on earnings is gradually increased from JPY 650 000 in 2027 (144% of gross average earnings in 2024) to JPY 750 000 in 2029. This raises both paid contributions and earnings-related pension entitlements (i.e. pensionable earnings). In the short term, the measure thus brings in more resources for the pension system, while the corresponding increase in expenditures due to higher pension entitlements will only grow gradually over time. At a time of fast-increasing pension expenditures, such a measure can help reduce financial pressure due to population ageing in the short term, but it does not reduce this pressure in the long term.
Slovenia’s comprehensive reform will result in higher baseline replacement rates upon retirement but in lower indexation of pensions in payment. Initial pensions are adjusted by increasing accrual rates, on the one hand, and by reducing the reference wage taken into account in the pension calculation, on the other. From 2028 to 2035, accrual rates are increased from 29.5% for the first 15 years of the career and 1.36% for each supplementary year, to 30% and 1.6%, respectively. As a result, total accrual over a 40‑year career increases from 63.5% to 70.0%. At the same time, the period considered to determine the reference wage is extended from the best 24 to the best 35 years, thereby moving towards lifetime earnings.35 This extension increases fairness in the pension system as pensions calculated on earnings made during only part of the career benefit people with steep earnings profiles throughout their career compared to those with stable earnings. Low-earners with patchy careers are unlikely to be substantially affected by this change because out-of-employment periods are excluded from the calculation of the reference wage, and there is a floor to the reference wage at 76.5% of the average wage. Colombia, Costa Rica, France and Spain are now the only OECD countries using less than 35 years to calculate the reference wage for their DB pensions. Furthermore, indexation of pensions in payment has been adjusted, which would lead to slower increases in pensions over time. Pensions in payment are currently adjusted to 60% of wage growth and 40% of price inflation. From 2026 onwards, these percentages are gradually adjusted each year until pensions are indexed to 20% of wage growth and 80% of price inflation by 2045. The combined impact of these reforms is expected to reduce total pension entitlements.36
In the United States, the depletion date of the Old-Age and Survivors Insurance (OASI) Trust Fund has moved forward by three‑quarters to the first quarter of 2033 (Board of Trustees of the Federal Old-Age and Survivors Insurance and Federal Disability Insurance Trust Funds, 2025[31]). Three factors have contributed to the depletion date moving forward. First, a reform in 2025 repealed two provisions which had reduced benefits for some people who receive pensions from jobs that are not covered by Social Security. This increases benefits for some people receiving pensions from certain state and local or federal government retirement systems and some people receiving pensions from work outside the United States. Second, the current spell of low fertility is now assumed to last 10 years longer than in previous projections. And third, the labour share as a percentage of GDP is now assumed to stabilise at a lower level, reducing pension contributions as a share of GDP. In Spain, the gap between pension spending and social security contributions will widen in the coming decades without further reforms, despite high contributions. AIReF (2025[32]) projects that, under current rules, pension spending will raise by 3.2 p.p. of GDP between 2023‑2050, reaching 16.1% of GDP in 2050. This will create a persistent funding gap and a growing stock of implicit liabilities that are not provisioned for today (OECD, 2025[33]). Given the widening gap between pension expenditures and social security contributions in AIReF’s projections, the IMF presses for further pension reform (International Monetary Fund. European Dept., 2025[34]).
Other changes in earnings-related benefits and taxation
Pensioners’ incomes can be adjusted through adjusting replacement rates, changing indexation of pensions in payment, changing contribution rates in DC schemes, or modifying tax rates. In addition to the changes in pension benefits in Czechia (see above), Korea, New Zealand, the Slovak Republic, Slovenia and Switzerland have increased benefits from earnings-related pensions to some extent. Latvia has furthermore doubled the amount of income that is tax-exempt for pensioners, resulting in the large majority of pension recipients being exempt from paying taxes.
Korea, New Zealand, the Slovak Republic and Switzerland have increased benefits from earnings-related pensions. Korea has reversed the scheduled decline in the target replacement rate of an average earner with a 40‑year career: it is now fixed at 43% from 2026 onward compared with 41.5% in 2025 while it was set to gradually decline to 40% in 2028. This is financed by part of the increase in the contribution rate (see above). New Zealand decided to increase the employee’s as well as the employer’s matched contribution rate to the auto‑enrolment FDC scheme from 3.0% to 3.5% in 2026 and to 4.0% in 2028. The government contribution is halved, from 50% to 25% of contributions, and removed for high earners.37 The Slovak Republic and Switzerland have both introduced a 13th-month pension payment. In the Slovak Republic, it consists of a flat-rate benefit equal to the average monthly pension benefit in the preceding year with a floor of EUR 300.38 To receive the full 13th-month payment, old-age pensioners need to have paid at least 10 years of contributions; a pro-rata adjustment is applied for people with a shorter insurance period. The Slovak Republic has increased the contribution rate to the public pension-point scheme by 1.5 p.p. while reducing the contributions to the automatic-enrolment FDC scheme from 5.5% to 4.0%.39 The scheduled increase in the contribution rate to the auto‑enrolment scheme to 6% by 2027 has been cancelled. Switzerland has introduced a 13th-month pension payment in the public earnings-related scheme following a referendum, which may be financed from an increase in VAT. In addition, Portugal decided to index pensions in payment from the first year after retirement, instead of the second year previously.40
Latvia has changed a number of parameters in its pension system as well, most notably it has reduced the taxation of pension benefits. From 2025, Latvia has doubled the amount of income that is tax-exempt for recipients of various pensions including old-age, disability and survivors’ pensions, from EUR 500 to EUR 1 000 per month. As a result, pensions are exempt from taxation up to the level equal to 58% of gross average earnings. In 2024, about 40% of old-age pensions were below EUR 500 and about 90% below EUR 1 000 (Central Statistical Bureau of Latvia, 2025[35]). Latvia has reinstated the pension supplement for years worked before 1996 that was abolished for new pensioners in 2012. The supplement is a flat-rate monthly benefit, equal to EUR 1.62 per year worked before 1996 in 2025, or 1% of gross average earnings for 10 years worked.41 Furthermore, contributions have temporarily been rebalanced between its NDC and FDC schemes. Since 2016, a contribution of 14% financed the NDC scheme and 6% went to the FDC scheme (OECD, 2018[36]); between 2025 and 2028, this will be 15% and 5%, respectively.
Future replacement rates
On average across the OECD based on already legislated measures, an average‑wage worker is projected to receive a net pension from mandatory schemes at 63% of net wages after a full career from age 22 in 2024. Future net replacement rates are below 40% in Estonia, Ireland, Korea and Lithuania (Figure 1.14). At the other extreme, they are above 85% in Austria, Greece, Luxembourg, Portugal and Spain, and over 95% in the Netherlands and Türkiye.
The future net replacement rate of full-career workers with low earnings (50% of the average wage) is 76% on average among OECD countries, or 12 p.p. above that for average earners. Replacement rates are generally higher for low earners due to redistributive features within pension systems. In Lithuania and Poland, the net replacement rate for low earners is very low, around 40%. On the other side of the spectrum, in Denmark, Mexico and Slovenia, it is more than 100%, meaning that net income is higher when moving from work to retirement, with Greece, Luxembourg and the Netherlands being close to 100%.
Measures legislated over the last two years and described above have the largest positive impact on future net replacement rates in Mexico, as well as in Chile and the Slovak Republic. While Mexico’s future replacement rates for average and low earners were close to the OECD average before, they have now increased by 18 and 47 p.p., respectively. This is the consequence of the introduction of the public top-up to the mandatory private FDC scheme described above. In Chile, the increased employer contributions flowing into individual FDC accounts as well as guarantee bonds increase the future net replacement rate for average‑earning men by 17 p.p. to 61%. The replacement rate for women with the same earnings increases by 19 p.p. due to the introduction of the benefit compensating the negative impact of women’s higher life expectancy on FDC annuities, bringing their pension up to the same level as that of men in that case. The introduction of the 13th-month pension in the Slovak Republic increases net replacement rates by 6 p.p. for an average earner and 7 p.p. for the low earner, to 79% and 86%, respectively. The reform in Slovenia increases the replacement rate of an average earner by 6 p.p. to 71%, and for a low earner by 8 points to 100%. This is based on cases where workers receive the same relative wage throughout their entire careers, however, whereas many people see an increasing earnings profile over their careers. Hence, for many people the increase in pension entitlements will be lower as the higher accrual rate will to some extent be offset by the extension of the reference period, on top of lower indexation for everyone.
Reforms have generated more moderate changes in replacement rates in Korea, Latvia and Switzerland. In Korea, the increase in the pension scheme’s target replacement rate results in an increase in the net replacement rate of 3 p.p. for an average earner and 4.5 p.p. for a low earner. In Latvia, the impacts of the temporary reallocation of contributions from the FDC to the NDC scheme and of the change in taxation of pensions largely cancel each other out for the average earner. The introduction of a 13th month pension in the public earnings-related scheme in Switzerland has increased replacement rates by 2 and 3 p.p. for average‑ and low-earner cases, respectively. Given the career-length and scheme‑membership assumptions underpinning the baseline case, Colombia’s structural pension reform does not affect replacement rates here: both before and after the reform, an average and a low earner with a full career receive a pension fully in the DB scheme, the rules of which have not been changed.
Future net replacement rate from mandatory schemes after a full career from age 22 in 2024
Note: Normal retirement age between brackets. Low earners earn 50% of the average earner. Low earners in Colombia, New Zealand and Slovenia are at 64%, 63% and 56% of average earnings, respectively, to account for the minimum wage level.
Source: See Chapter 4, Table 4.4, https://stat.link/ic8ung.
Changes in withdrawal options of funded pensions
Lithuania and Türkiye have introduced the possibility to make lump-sum withdrawals from funded pensions. Lithuania has decided to allow people to take out 25% of the account balance of the FDC pension at any time once in their life. The total account balance can now also be taken up in the five years prior to reaching the statutory retirement age if the amount remains below a certain threshold.42 In case of a serious health condition impeding making further contributions in the future, the full account can be withdrawn without tax or deduction at any time. In Türkiye, withdrawals can now be made in case of marriage, purchasing a home, natural disasters or university education. Each reason for withdrawal can only be used once, and each time up to half of the account balance can be withdrawn.43 These options are not aligned with the OECD Recommendation that early access to retirement savings should be a measure of last resort and based on individual circumstances of hardship (OECD, 2022[37]). In 2021, the Netherlands legislated the option to withdraw up to 10% of the pension as a lump sum upon retirement. While initially foreseen to take effect in 2022, the law was revised and its implementation postponed. A ministerial communication now set its earliest data of implementation in July 2026.
Other small changes in contributions
Poland and Türkiye took small measures modifying contribution subsidies to business owners. Poland has de facto introduced an 8.5% reduction in social contributions (which include pension contributions) provided to small-business owners: social contributions of small-business owners are fully paid by the state for one month per year in order to reduce the cost of running a business. During the “contribution holiday”, owners of businesses employing fewer than 10 people do not pay social contributions for themselves, with no impact on their social-security entitlements. Türkiye has reduced the subsidy for social contributions paid by employers from 5 to 4 p.p., except in manufacturing where it remains at 5 p.p.
Coverage reforms
Changes in coverage
Ireland will be introducing automatic enrolment in its occupational pension scheme, whereas Lithuania has decided to abolish its automatic enrolment policy. After several delays, Ireland legislated automatic enrolment in 2024 and is now expected to start automatically enrolling new and current employees aged between 23 and 60 into FDC occupational pensions from January 2026. Employees will be enrolled if gross earnings exceed EUR 20 000 on an annual basis or 30% of average earnings unless they are already enrolled in an occupational pension scheme. Contributions are paid on the part of earnings below EUR 80 000 per year. Those who are auto‑enrolled can opt out or suspend their contributions after six months of mandatory participation. Enrolment is totally voluntary for employees not meeting the auto‑enrolment conditions. The total contribution rate is set to increase from 3.5% in 2026 to 14% in 2036. Employer contributions match employee contributions, and the state contributes one‑third of that amount: by 2036, employer and employee will each contribute 6%, and the state 2%. It is not possible to pay in more than the set rate. Lithuania, by contrast, has decided to abolish auto‑enrolment in its FDC pension scheme and move back to fully voluntary coverage. People will be able to withdraw the contributions they made and the returns on those contributions, exempt from personal income tax, in 2026‑2027. The part of the individual account financed from subsidies would flow to the social insurance fund and be converted into supplementary pension points in the points-based pension scheme.
Japan, Korea, Mexico and Switzerland have taken steps to expand coverage of existing pension schemes to new categories of workers. Japan has increased coverage of part-time workers, as well as of workers in specific sectors. Previously, part-time workers only qualified for pension build-up if: they worked at least 20 hours per week; their earnings exceeded a threshold corresponding to around 20% of gross average earnings; and, they worked in a business with more than 50 employees. While the condition of 20 hours worked remains in place, the earnings limit is abolished by 2028, and between 2027 and 2035 the company-size threshold is gradually phased out. Furthermore, while coverage is mandatory for businesses with at least five full-time employees, there were exceptions for businesses in agriculture, forestry, fishing, bars, restaurants and hotels. These exemptions are eliminated from 2029, although only for new businesses entering those sectors. Korea extended contribution subsidies in its voluntary pension scheme for low-income people who are not mandatorily covered by pension insurance (e.g. self-employed or people working in small businesses).44 Mexico, which expanded mandatory coverage to domestic workers in 2022, has extended it again to digital platform workers from June 2025. In Switzerland, contributions paid after the statutory retirement age now result in supplementary pension build-up until reaching age 70. When combining work and pensions, the pension can be recalculated once to include the extra years worked.
Coverage for periods of care, marriage and survivor’s benefits
Several OECD countries have improved pension provision for parents over the last two years. Australia has introduced childcare‑related credits, and Czechia and Korea have expanded these credits. In Colombia, the suspended pension reform includes the introduction of childcare credits. Australia introduced childcare credits in the FDC scheme for government-funded parental leave for children born or adopted as of July 2025. For the period of parental leave (up to 120 and 130 working days for children born as of July 2025 and July 2026, respectively), the government has now started paying superannuation contributions. Contributions are made at the level of the Superannuation Guarantee, at 12% of the parental-leave benefit. The impact on total pension entitlements is limited, however, as the higher FDC pension is to a large extent offset by a reduction in the targeted benefit. Colombia’s pension reform pending review by the Constitutional Court would make the DB component of its newly reformed pension system more accessible to mothers by crediting 50 weeks of contributions per child towards the career-length condition to qualify for the DB pension, for up to three children. This would effectively lower the eligibility threshold to qualify for the DB pension from 1 000 to 850 weeks for a mother of three children.
In Czechia, pension entitlements for parents will change for the first two children from 2027. One parent now receives a three‑year credit per child, irrespective of whether he or she is working, while the flat-rate childcare supplement per child remains in place for subsequent children. As the same period cannot be credited twice for having two children below age 3, an average‑earning parent builds up the same pension in case of a five‑year career break to care for two children born two years apart, as someone without children.45 In Korea, childcare credits for a period of up to 12 months per child were previously only available from the second child onward. From 2026, credits will be available from the first child.46
Czechia has introduced the option for spouses or registered partners to split pension entitlements from 2027. For periods during which both partners are in employment, couples in Czechia will be able to ask for the pension entitlements of each partner to be calculated on the average earnings of both partners. Given strongly redistributive elements in the Czech earnings-related pension, such as relatively low earnings thresholds above which earnings are only partially or not at all included in pension calculation, pension splitting through sharing earnings in pension calculation could allow couples with large income differences to increase the total pension they receive as a couple. Belgium might introduce voluntary pension splitting as well. Indeed, the Belgian Government agreement of January 2025 includes a commitment to introducing the option to voluntarily split pensions, and, if legislated, the higher replacement rate for individuals with a dependent spouse with little or no pension entitlements would be limited to some specific instances.
Canada, Japan, Poland and Slovenia have made changes to their survivor’s pensions. Canada no longer pays survivor’s pensions to the surviving spouse of a separated couple if they had requested a pension split of their Canada Pension Plan entitlements. Japan has reformed the survivor’s pension to make it gender neutral from 2028 onwards. Previously, women received a permanent survivor’s pension after widowing or a five‑year transitional benefit if they became widowed before turning 30 years old, while men could only access a transitional benefit if they became widowed from age 55 or a permanent survivor’s pension from age 60. Under the new rules, both men and women are entitled to the five‑year transitional benefit after losing their spouse before age 60, and to the permanent survivor’s pension thereafter. The reform will not reduce entitlements for people who are already receiving survivor’s benefits, for those who are over age 60, for women who are over 40 in 2028, and for people with children younger than 18 years old. In addition, Japan introduced a new component to its survivor’s pension: if the deceased spouse’s earnings were higher than those of the surviving spouse, part of the deceased spouse’s earnings-related pension record is added to that of the surviving spouse. At the same time, the income test for survivor’s benefits is abolished.47 While Poland previously did not allow combining a survivor’s pension with an old-age pension, it is now possible to some extent. Widow(er)s can now choose whether to receive their full personal pension plus 15% of the survivor’s pension, or whether to receive the full survivor’s pension plus 15% of their personal pension. From 2027, the amount of the second pension benefit will increase from 15% to 25%. In Slovenia, the eligibility age for survivor’s pensions will increase by two years in line with other age thresholds in the pension system (see above). Between 2028 and 2035, the eligibility age will increase from 58 to 60, in increments of three months per year. At the same time, survivor’s benefits are increased from 70% to 75% of the deceased spouse’s pension in 2026 and to 80% in 2027. Switzerland is also expected to reform its survivor’s pension.48
Pension reforms in progress
Austria and Norway are planning to change retirement ages and early retirement options. Austria would restrict access to the early-retirement scheme known as the “Korridorpension”, which is one of four early-retirement schemes alongside three schemes covering long-term insured and people performing physically demanding work. For “Korridorpension”, while the statutory retirement age is 65 (for men, and from 2033 also for women), the minimum retirement age would gradually be increased from 62 to 63 years and the insurance years required to take up the pension from 40 to 42 years. Following a 2024 parliamentary agreement, Norway is in the process of legislating a two‑thirds link between the retirement age and life expectancy, and is planning to simultaneously reduce the effective penalty in case of early retirement. The retirement age would automatically be adjusted annually in monthly adjustments. As part of the political agreement, the impact of the penalty in case of early retirement would be reduced by the introduction of a flat-rate supplement. The measure would provide a top-up to persons retiring between the age of 62 and 65, with these age limits increasing along with the normal retirement age. The full supplement, around 4% of gross economy-wide average earnings, would be paid to persons retiring at 62, and the amount will gradually be reduced as people retire closer to the normal retirement age. The benefit, called a “hardship scheme”, is not targeted at jobs or occupations considered arduous or hazardous, but instead it is based on an underlying assumption that if people do not work longer in response to increasing retirement ages, they probably are unable to. People choosing to be in the scheme will only be able to combine work and pensions to a very limited extent. The benefit is designed to mitigate the negative impact of the early-retirement penalty to some extent, in particular for people with low earnings, and is considered too modest to have a substantial effect on early-retirement incentives.
In Belgium, the new government’s agreement contains a wide range of pension-related measures expected to be passed in 2025. They are primarily aimed at containing the increase in pension expenditure due to population ageing. These among others includes: a cap on the amount of credited periods that can make up an individual’s insurance career; the introduction of a bonus-penalty scheme as well as a new early-retirement option from age 60 with at least 42 years worked; a further harmonisation of the pension scheme for civil servants with that of private‑sector employees; the closing of early-retirement options for certain occupations, including for military and train staff, who can currently retire at age 56 and 55, respectively; and, tighter residence requirements to receive social assistance for older people.
Since the 2023 edition of Pensions at a Glance, the replacement rates for the Netherlands are based on FDC occupational pensions. The rules that entered into force in 2023, obliging pension funds to transition from FDB to FDC schemes by 2028, still apply today. Yet, transitional measures remain a topic of political debate, which generates uncertainty. Funds are encouraged to transfer DB entitlements to the new pension system and whether to force pension funds to consult their members on transitioning already built-up entitlements remains a topic of intense debate. A proposed amendment to force such consultations was rejected by Parliament in May 2025 with a margin of one single vote.
References
[32] AIReF (2025), Opinion on the long-term sustainability of the general government: Demographic and climate change, Opinion 2/25, https://www.airef.es/wp-content/uploads/2025/03/Opini%C3%B3n_sobre_la_sostenibilidad_de_las_AAPP_largo_plazo/AIReF.-Opinion-on-the-Long-term-Sustainability-of-General-Government.-Demography-and-Climate-Change.pdf.
[14] BBVA (2022), Qué pensión cobraré si retraso mi jubilación más allá de la edad de jubilación ordinaria [What pension will I receive if I delay my retirement beyond the ordinary retirement age], https://www.jubilaciondefuturo.es/es/blog/que-pension-cobrare-si-retraso-mi-jubilacion-mas-alla-de-la-edad-de-jubilacion-ordinaria.html.
[11] Blöndal, S. and S. Scarpetta (1999), “The Retirement Decision in OECD Countries”, OECD Economics Department Working Papers, No. 202, OECD Publishing, Paris, https://doi.org/10.1787/565174210530.
[31] Board of Trustees of the Federal Old-Age and Survivors Insurance and Federal Disability Insurance Trust Funds (2025), The 2025 Annual Report of the Board of Trustees of the Federal Old-Age and Survivors Insurance and Federal Disability Insurance Trust Funds, https://www.ssa.gov/oact/TR/2025/index.html.
[35] Central Statistical Bureau of Latvia (2025), Number of pension recipients by average size of pension granted (PPP050), https://stat.gov.lv/en/statistics-themes/social/benefits-allowances-pensions/tables/ppp050-number-pension-recipients.
[23] Dewhurst, E. (2016), “Proportionality Assessments of Mandatory Retirement Measures: Uncovering Guidance for National Courts in Age Discrimination Cases”, Industrial Law Journal, Vol. 45/1, pp. 60-88, https://doi.org/10.1093/indlaw/dww004.
[16] Eurostat (2023), Persons receiving an old-age pension and continued working at the beginning of pension receipt by reason (lfso_23pens08).
[8] Human Mortality Database (2025), Human Mortality Database, supported by the Max Planck Institute for Demographic Research (Germany), the University of California, Berkeley (USA), and the French Institute for Demographic Studies (France) - data accessed on 10 June 2025, http://www.mortality.org.
[34] International Monetary Fund. European Dept. (2025), “Spain”, IMF Staff Country Reports, Vol. 2025/121, p. 1, https://doi.org/10.5089/9798229012782.002.
[24] Lazear, E. (1979), “Why Is There Mandatory Retirement?”, Journal of Political Economy, Vol. 87/6, pp. 1261-1284, https://www.jstor.org/stable/1833332.
[9] Lee, R. and Y. Zhou (2017), “Does Fertility or Mortality Drive Contemporary Population Aging? The Revisionist View Revisited”, Population and Development Review, Vol. 43/2, pp. 285-301, https://doi.org/10.1111/padr.12062.
[6] Levine, M. and E. Crimmins (2018), “Is 60 the New 50? Examining Changes in Biological Age Over the Past Two Decades”, Demography, Vol. 55/2, pp. 387-402, https://doi.org/10.1007/s13524-017-0644-5.
[13] Ministerio de Inclusión, Seguridad Social y Migraciones (2021), Memoria del análisis de impacto normativo del anteproyecto de ley de garantía del poder adquisitivo de las pensiones y de otras medidas de refuerzo de la sostenibilidad financiera y social del sistema público de pensiones, https://tinyurl.com/z8u7nfk2.
[33] OECD (2025), OECD Economic Surveys: Spain 2025, OECD Publishing, Paris.
[2] OECD (2025), OECD Employment Outlook 2025: Can We Get Through the Demographic Crunch?, OECD Publishing, Paris, https://doi.org/10.1787/194a947b-en.
[21] OECD (2024), OECD Economic Surveys: Japan 2024, OECD Publishing, Paris, https://doi.org/10.1787/41e807f9-en.
[1] OECD (2024), Society at a Glance 2024: OECD Social Indicators, OECD Publishing, Paris, https://doi.org/10.1787/918d8db3-en.
[15] OECD (2024), Strengthening the Hungarian Pension System, OECD Publishing, Paris, https://www.oecd.org/content/dam/oecd/en/topics/policy-sub-issues/structural-reforms/country-tailored-policy-reforms/Strengthening-the-Hungarian-Pension-System.pdf.
[5] OECD (2023), Beyond Applause? Improving Working Conditions in Long-Term Care, OECD Publishing, Paris, https://doi.org/10.1787/27d33ab3-en.
[30] OECD (2023), Pensions at a Glance 2023: OECD and G20 Indicators, OECD Publishing, Paris, https://doi.org/10.1787/678055dd-en.
[20] OECD (2022), OECD Reviews of Pension Systems: Korea, OECD Reviews of Pension Systems, OECD Publishing, Paris, https://doi.org/10.1787/2f1643f9-en.
[4] OECD (2022), OECD Reviews of Pension Systems: Slovenia, OECD Reviews of Pension Systems, OECD Publishing, Paris, https://doi.org/10.1787/f629a09a-en.
[37] OECD (2022), Recommendation of the Council for the Good Design of Defined Contribution Pension Plans, https://legalinstruments.oecd.org/en/instruments/OECD-LEGAL-0467.
[3] OECD (2021), Pensions at a Glance 2021: OECD and G20 Indicators, OECD Publishing, Paris, https://doi.org/10.1787/ca401ebd-en.
[25] OECD (2019), Working Better with Age, Ageing and Employment Policies, OECD Publishing, Paris, https://doi.org/10.1787/c4d4f66a-en.
[36] OECD (2018), OECD Reviews of Pension Systems: Latvia, OECD Reviews of Pension Systems, OECD Publishing, Paris, https://doi.org/10.1787/9789264289390-en.
[17] OECD (2018), Policy Brief on Ageing and Employment: Council Recommendation on Ageing and Employment, OECD Publishing, Paris, https://www.oecd.org/els/emp/Flyer_AE_Council%20Recommendation.pdf.
[19] OECD (2018), Working Better with Age: Korea, Ageing and Employment Policies, OECD Publishing, Paris, https://doi.org/10.1787/9789264208261-en.
[12] OECD (2017), Pensions at a Glance 2017: OECD and G20 Indicators, OECD Publishing, Paris, https://doi.org/10.1787/pension_glance-2017-en.
[18] OECD (2017), Preventing Ageing Unequally, OECD Publishing, Paris, https://doi.org/10.1787/9789264279087-en.
[27] OECD (2017), Preventing Ageing Unequally, OECD Publishing, Paris, https://doi.org/10.1787/9789264279087-en.
[28] OECD (2013), OECD Employment Outlook 2013, OECD Publishing, Paris, https://doi.org/10.1787/empl_outlook-2013-en.
[22] Oliveira, A. (2016), “A Freedom Under Supervision: The EU Court and Mandatory Retirement Age”, in Challenges of Active Ageing, Palgrave Macmillan UK, London, https://doi.org/10.1057/978-1-137-53251-0_2.
[26] Rabaté, S. (2019), “Can I stay or should I go? Mandatory retirement and the labor-force participation of older workers”, Journal of Public Economics, Vol. 180, p. 104078, https://doi.org/10.1016/j.jpubeco.2019.104078.
[10] Schokkaert, E. and P. Van Parijs (2003), “Debate on Social Justice and Pension Reform”, Journal of European Social Policy, Vol. 13/3, pp. 245-263, https://doi.org/10.1177/09589287030133003.
[29] Schols, J. et al. (2022), De loopbaansamenstelling van (toekomstig) gepensioneerde werknemers, https://www.plan.be/sites/default/files/documents/REP_Pension_202201_NL.pdf.
[7] Scott, T. and V. Canudas-Romo (2024), “Decomposing the Drivers of Population Aging: A Research Note”, Demography, Vol. 61/4, pp. 1011-1021, https://doi.org/10.1215/00703370-11481955.
[38] Superintendencia de Pensiones (2025), Número y monto promedio, en U.F., de las pensiones pagadas en el mes por modalidad, según tipo de pensión (Al 31 de marzo de 2025), https://www.spensiones.cl//inf_estadistica/afipen/mensual/2025/03/m00.html.
← 1. Due to COVID‑19, net migration rates were much lower in 2020 and 2021 than in the previous years. Hence, 2019 is a better reference to assess the increase in the net migration rate in 2022.
← 2. The impact of migration on the evolution of the old-age to working-age ratio can be small, even if net migration has increased in recent years. The decomposition does not just take into account the evolution of migration over the last 10 years, but rather over the full lives of the cohorts concerned since birth. Higher net migration rates would have to be sustained for some time before they really start weighing on the composition of the full population on active age relative to the population in old age.
← 3. For people born in 2002 without 40 years of contributions, early retirement will be possible from the age of 66 (which is also the normal retirement age for full-career workers). In that case, a permanent penalty of 6% per year of anticipation applies.
← 4. The bonus is calculated differently depending on whether the career is shorter or longer than 44 years and six months. Estimates are computed based on the OECD pension model.
← 5. Estimates are computed based on the OECD pension model.
← 6. After 2% per year of early uptake or deferral from 2026, 4% from 2030. Both the bonus and the penalty will depend on career length: the penalty would only apply to people with fewer than 35 years effectively worked whereas the bonus would only apply to people with at least 35 years effectively worked. Maternity and care periods would be credited in the 35‑year career to determine whether a bonus or penalty applies. Such career-length conditions undermine the effectiveness of the bonus and penalty to remove disincentives for working longer as they conflict with the principle of actuarial neutrality and exclude large groups of people.
← 7. Combining work and pensions is the dominant form of working beyond the retirement age as pension deferral is not very common. In the chapter dedicated to flexible retirement in the 2017 edition of Pensions at a Glance, it was noted that only 2% of individuals aged 65‑69 in the EU continued in employment without claiming a pension (22% among the 60‑64) (OECD, 2017[12]).
← 8. There is no bonus for the deferral of the remaining 60% or 80%, although Slovenia has a very high accrual rate for years worked after the statutory retirement age. Combining the lack of bonus on the 60% deferred pension benefit with the very high accrual rate results in a pension build-up that is close to actuarially neutral, being it in a complex way (OECD, 2022[4]).
← 9. In Germany, the employee is exempt from paying pension contributions when combining work and pension receipt after the normal retirement age, although employers still have to pay contributions while no further pension entitlements are built up. The employee can waive the exemption and pay contributions as well, in which case additional pension entitlements are accrued. Similarly, in Türkiye, a working pensioner pays reduced social contributions and build no pension entitlements, but the employer pays regular contributions. Pension entitlements can be built up further if the employee suspends pension receipt and pays regular contributions.
← 10. In Luxembourg, while employees can request to have their employee contributions reimbursed, employer contributions cannot be recuperated. As reimbursement has to be requested, this practice is effectively a tax on ignorance.
← 11. After one year of deferral, work can be combined with up to 45% of an individual’s full pension; combining a full pension with work requires that uptake is deferred by at least five years. For every year of combining work and pension, the share of the pension received further increases by an additional 5 p.p. up to a maximum of 100%. Hence, a person who defers uptake by one year and then combines work and pension, receives 45% of the pension in the first year of combining work and pension, 50% in the second year, and 55% in the third year.
← 12. The earnings limit does not apply to people with 45 years of contributions.
← 13. All countries with mandatory retirement for civil servants except Colombia, Ireland, Italy and Türkiye do foresee an option to extend civil service employment beyond the mandatory retirement age under certain conditions, such as performance requirements or if retirement would result in the loss of capabilities in the civil service. Either by extending their appointment or by rehiring them, civil service employment can be extended, typically for a period of three to five years and often in the form of renewable one‑year extensions or contracts.
← 14. Chile applies a lower minimum wage to people who are hired after the statutory retirement age of 65 to make it more attractive for employers to hire retirees. However, this is not a case of mandatory retirement as it does not apply to people who reach the statutory retirement age under contract, but only to people who are recruited after the statutory retirement age.
← 15. After the elimination of the retirement age in March 2023 for people who entered the labour market before 8 September 1999, the normal retirement age in Türkiye even dropped to 47 for men and 46 for women.
← 16. While Türkiye is an absolute outlier for people retiring now, its normal retirement age is set to increase fast as it will be 65 for men entering the labour market in 2024.
← 17. Increases in the statutory retirement age require parliamentary approval in Denmark. Under current rules, the retirement age revisions take place every five years and take effect 15 years after approval (OECD, 2021[3]).
← 18. For each cohort, the career-length condition for early retirement is equal to the statutory retirement age for the cohort minus 23, i.e. the difference between the statutory retirement age (63) and the career-length condition (40) applying to the 1960 cohort who reached the statutory retirement age in 2023.
← 19. Deductions of one year in case of a single child or two years for multiple children remain in place.
← 20. Normally, the pension is suspended if earnings exceeded the income threshold for minimally employed workers (known as the “Geringfügigkeitsgrenze”, EUR 551 per month in 2025). In 2024 and 2025, a person can earn up to 40% of that monthly threshold in excess of that limit over the full year (i.e. up to EUR 220 for the full year in 2025) before the pension is suspended.
← 21. A working-time reduction of between 25% and 75% was already possible two years prior to the normal retirement age in case of a “relief contract”, through which the retiree is gradually replaced by an unemployed person or someone previously employed on a temporary contract. Since 2025, partial retirement under this type of contract is possible from three years before the normal retirement age, although in that case, only a working-time reduction of between 20% and 33% is possible in the first year.
← 22. The compensation is lower in case of retirement before 65.
← 23. The contribution-based basic pension increases by 0.1 UF per year of contributions. UF, or Unidad de Fomento, is a unit of account used in finance in Chile. The average FDC old-age pension paid out in March 2025 was 6.73 U.F., or around CLP 264 000 (Superintendencia de Pensiones, 2025[38]).
← 24. In Chile, the targeted benefit is increased, but the minimum and maximum thresholds between which the benefit is gradually withdrawn remain unchanged, respectively at 64% and 102% of gross average earnings. As a higher benefit has to be fully withdrawn between the same two limits, the withdrawal rate is higher.
← 25. The FDC component would furthermore be strengthened by gradually eliminating the 0.8 p.p. administrative fee, so that contributions to the individual account would increase from 13.2% to 14.0%.
← 26. Individuals with between 300 and 999 weeks of contributions would instead receive their contributions as an annuity with a tax-financed top-up, capped at 80% of the minimum wage. The targeted benefit, the lump sum and the annuity would be accessible three years after the normal retirement age, from age 65 for men and 60 for women, aligned with the eligibility ages of the previous targeted benefit.
← 27. On the part of earnings exceeding 4 times the minimum wage, a contribution of between 1.5% and 3.0% (on the part of earnings exceeding 20 times the minimum wage) would be paid. Contributions from pensions would be somewhat lower, with a 1% contribution rate on the part of total pensions between 10 and 20 times the minimum wage, and a 2% contribution rate on the part of total pensions above that threshold.
← 28. A fund was created to pay the top-up. In addition to sleeper accounts of people over 70 (it remains possible for people to reclaim their pension from the fund), the fund is financed from a variety of sources including among others assets seized by the state, and profits of state‑owned enterprises.
← 29. The increase applies partially to singles born between 1954 and 1963 as well.
← 30. Indexation will resume earlier if the index exceeds the 2023 level by 10.2%.
← 31. The threshold is reduced from EUR 18 306 to EUR 15 000 for singles and from EUR 29 290 to EUR 24 000 for couples. The withdrawal rate applied to income increased somewhat as well, from 41.3% to 43.5%.
← 32. The increase for between 40 and 49 years of contributions (3.0 p.p.) and between 50 and 59 years of contributions (5.0 p.p.) remain the same, but the increase as of the 60th year of contributions was raised from 7.0 to 7.5 p.p.
← 33. Only 3 120 days of effective employment are required to access the minimum pension for part-time employment, in which case the minimum pension benefit is prorated to the number of days effectively worked relative to the number of days worked by someone with a full 45‑year career of full-time work. Some non-worked periods in particular in relation to caregiving are nonetheless included to determine whether a person has attained the required number of days of effective employment.
← 34. This only refers to periods during which pension entitlements are built up as an employee in Belgium. People may have built up entitlements as self-employed or abroad as well.
← 35. Under Slovenia’s new rules, the best consecutive 40 years are taken into account in the pension calculation, but the five years with the lowest earnings are excluded; for people with at least 28 years of contributions, one year can be excluded from the pension calculation, increasing to five years in case of a career of at least 40 years of contributions. The period will be extended by two years each year from 2028 to 2035.
← 36. For a worker with earnings increasing from 60% to 123% of average earnings over the career, the reform reduced the effective real annual rate of return – i.e. the implicit rate of return on an individual’s contributions paid to finance the individual’s pension benefits – from 2.7% to 2.3%. This does remain substantially above the real internal rate of return of 1.6%, which is the level that would sustainably finance pension promises from contributions in that case.
← 37. The age of automatic enrolment remains at 18 years, but it is now also possible for people to opt in from age 16 and receive government contributions – matched employer contributions are only mandatory from 2028 onward.
← 38. The 13th month pension is specific to the type of pension received, e.g. an old-age pensioner receives the average old-age pension, whereas survivor’s or disability pension recipients receive the average of their respective types of benefits.
← 39. The Slovak Republic also increased the earnings ceiling below which contributions are due from 7 to 11 times gross average earnings, both for employees and the self-employed.
← 40. In addition, Portugal made a one‑off payment to people with a pension below three times IAS in 2024 of between EUR 100 and EUR 200 depending on pension level.
← 41. The supplement is gradually rolled out again, starting with those who retired after its abolishment in 2012 and 2013, so that by 2029, all pensioners who have worked before 1996 will receive a supplement.
← 42. A 3% deduction applies to withdrawals before the normal retirement age, but these withdrawals are exempt of personal income tax. Upon retirement, part of the pension should be taken out as an annuity if the account balance exceeds a certain threshold.
← 43. Withdrawals for education are paid as a four‑year annuity; all others as a lump sum.
← 44. Subsidies previously covered people who are subscribed to the scheme but ceased to pay contributions, whereas the new subsidies cover people who subscribe to the scheme in general. The subsidies are to cover half of the pension contributions for up to 12 months; the maximum contribution subsidy is yet to be determined. the contribution subsidy that incentivises these workers to enrol voluntarily has been increased for the first 12 months of paying contributions.
← 45. The new credited periods are based on average earnings for people having below average earnings, and based on the individual’s previous earnings for people earning more.
← 46. In addition, Korea previously capped total childcare credits at 50 months over the full career. This cap is abolished from 2026 as well.
← 47. Japan furthermore increased the flat-rate supplement for pension recipients living with children younger than 18.
← 48. This is in the context of the 2022 ruling by the European Court of Human Rights that the current legislation in Switzerland treats men and women unequally. Widows currently receive a lifelong annuity irrespective of age in case of children or, in case there are no children, from age 45 provided they have been married for at least five years. Widowers, by contrast, can only receive a survivor’s pension in case they have children. The proposal is to instead introduce an annuity until the youngest child turns 25, irrespective of the recipient’s sex or whether or not the couple was married. For people without children under 25 at the moment their partner passes away, a two‑year transitional benefit would be paid, except in case the surviving partner is 58 or older at the time of death and the loss of their partner would result in precarity. Of people who are already receiving a survivor’s pension, those below age 55 without dependent children would be moved to the two‑year transitional benefit.