Pensions at a Glance 2025

OECD and G20 Indicators

Report

OECD Pensions at a Glance

27 November 2025

Available in:

English
français

Download PDF

2. Gender pension gap

Copy link to 2. Gender pension gap

Abstract

This chapter starts by showing recent and projected trends in the pensions of women relative to those of men in OECD countries, and breaks down gender differences across pension components. The second section zooms into the key drivers of the gender pension gap, which results mainly from differences in lifetime earnings between men and women due to different labour market trajectories in terms of employment, hours worked and hourly wages. The chapter then raises normative questions about the role of pension policy in dealing with the gender pension gap. The next section details the pension rules that directly or indirectly affect gender disparities in pensions based on the OECD pension model. The following section focusses on gender disparities arising in asset-backed pensions. The chapter ends by discussing policy implications.

Introduction

Copy link to Introduction

The much lower pensions of women relative to men’s raise important social and policy concerns. Pension differences between men and women largely reflect and add up to gender disparities in the labour market and the disproportionate burden of unpaid care responsibilities faced by women. The higher longevity of women and the gender pension gaps (GPG) combine into higher women’s old-age poverty risks.

The instruments included in pension systems that limit the transmission of labour market disparities into retirement income also help reduce GPGs. In particular, first-tier pensions are higher for those less attached to the labour market and partially compensate for resulting low pensions due to low earnings during the whole working life. Reducing income inequality and alleviating poverty are often part of the objectives of pension systems along with limiting the fall in living standards at retirement and insuring against the uncertainty related to the length of individual lives, the so-called longevity risks.

Pension policy is influenced by normative choices regarding broader family policies, particularly reflected in the design of the following instruments. Childcare‑related pension credits specifically aim at compensating for the impact of childcare breaks on pension benefits. Pension bonuses for having children irrespective of experiencing an employment break can partially offset the indirect impact of parenthood on career development. Survivor pensions mainly benefit women due to their higher life expectancy and lower pension entitlements. Despite having been available for a few decades in some countries, pension splitting has not gained much popularity.

Key findings and policy implications

Copy link to Key findings and policy implications

Pension outcomes

Women receive monthly pensions that are about one‑quarter lower than men’s on average across OECD countries, ranging from less than 10% lower in Czechia, Estonia, Iceland, the Slovak Republic and Slovenia to more than 35% lower in Austria, Mexico, the Netherlands and the United Kingdom, and even 47% lower in Japan.
The large average gender pension gap (GPG) across OECD countries has declined from 28% in 2007 to 23% in 2024. It is projected to further decline in all countries for which such projections exist.
The GPG is the key indicator of average differences in pension levels between men and women. However, it does not measure differences in standards of living between older men and older women because living standards account for other sources of income, household compositions and income sharing within households. Also, older people without a pension are not accounted for in measuring the GPG. There is actually no correlation across OECD countries between the GPG and the gender gap in average disposable income among people 66 or older. On average in the OECD, gender disparities in household disposable income are substantially lower than the gender pension gaps, 10% among people aged 66+ in 2023, which is less than half the average GPG.
In 2024, on average across OECD countries, women are expected to live 22.8 years after having effectively left the labour market compared to 18.6 years for men, hence 4.1 years more or about one‑quarter longer. Men and women start receiving earnings-related old-age pensions at similar ages in many OECD countries, but the gender difference is large in countries that provide eligibility to pensions to women at lower ages: 4.3 years in Poland, 3.0 years in Chile, 2.8 years in Hungary and 2.0 years in Austria.
Among OECD countries, more than three‑fifths of beneficiaries of first-tier pensions are women, against half of beneficiaries of mandatory earnings-related pensions. There is a substantial underrepresentation of women in own earnings-related pensions (excluding survivor pensions) in Belgium, Costa Rica, Greece, Italy, Japan, Korea, Luxembourg, the Slovak Republic and Spain.
Gender differences in lifetime earnings are the main driver of the gender pension gap as a large part of pension benefits is earnings-related. Gender differences in the expected career duration, hours worked and hourly wages between men and women make a similar contribution to the large gender gaps in expected lifetime earnings averaging 35% across OECD countries in 2023.
At 34 years, the expected career duration for women was almost 6 years (or 15%) shorter than for men on average in the OECD in 2023. The gender gap in the expected career duration strongly declined from 18 years in 1980 to 6 years in 2023 on average across countries, mostly due to longer careers of women. After having declined by 1.5 hours since 2008, the difference in weekly working hours between employed men and women is still relatively large at 5.1 hours on average across OECD countries, or about 13%. The gender gap in hourly wages is also large at about 11% on average across OECD countries. It has declined by around 4 percentage points (p.p.) since 2008.

Pension rules

Despite a converging trend over the last 30 years, women can still retire without penalty at lower age than men in nine OECD countries. Based on current legislation, this gender difference in the normal retirement age will be eliminated in Austria, Lithuania and Switzerland, while it will persist in Colombia, Costa Rica, Hungary, Israel, Poland and Türkiye, negatively affecting women’s pension levels.
Mothers can retire between four months and four years earlier than childless women, depending on the country and the number of children, in Czechia, France, Italy, the Slovak Republic and Slovenia. On the other hand, avoiding pension penalties requires delaying retirement in the case of a five‑year childcare break in Greece and Portugal, as well as in France and Spain for a ten‑year break.
Mandatory pension systems cushion about half of the effects of a five‑year childcare‑related employment break on pensions for mothers with two children on average across OECD countries. Nine OECD countries give credits just for having had children or provide pension bonuses to parents, irrespective of whether a career break occurred.
Austria, Belgium, Denmark, Estonia, Finland, France, Germany, Hungary, Ireland, Lithuania, Norway, Spain and the United Kingdom also credit periods spent providing informal family care for adults, which is mainly done by women.
Korea and the United States provide spousal supplements, Japan credits periods towards the contribution-based basic pension when spouses are not employed and Belgium applies higher accrual rates for couples in contributory pensions. These instruments provide specific benefits for couples in which spouses do not have their own pension or only a very low pension.
Longevity differences between men and women are ignored in the calculation of pension benefits in mandatory public pensions in all OECD countries. This is consistent with the pooling of longevity risks across the whole population. Given women’s lower pension entitlements, ignoring longevity differences between men and women avoids lowering further women’s monthly pensions. In the European Union, private pension schemes cannot, by law, take into account longevity differences between men and women to calculate pension benefits, even when they are funded defined contribution.
In defined contribution schemes in countries outside the EU where benefits depend on gender-specific life expectancy, women’s pensions are negatively affected: funded defined contribution (FDC) schemes pay less every month to women than to men for the same amount of accumulated assets due to women’s longer expected duration of benefit receipt. A recent pension reform in Chile will eliminate the negative impact of higher women’s longevity on pensions from the FDC scheme by providing a compensating bonus to women as if they had men’s mortality tables.
Despite having been available for a few decades in some countries, pension splitting, i.e. the sharing of earned pension entitlements within a couple, has not gained much popularity.

Policy implications

The most efficient measures to reduce the GPG over the long term need to focus on tackling gender differences in employment, hours worked and wages. In particular, the unequal share of unpaid care between men and women as well as persistent disparities in education and labour market pathways have large implications.
While pensions cannot fully compensate for inequalities building up from education to labour market pathways, reducing income inequality in old age is often part of the objectives of pension systems. Policy instruments that reduce the impact of labour market inequalities on retirement-income differences tend also to reduce the GPG. They include progressive pension formulae, minimum contributory and basic pensions and pension credits for employment breaks.
Countries wanting to promote gender equality in the labour market and reduce the GPG should eliminate earlier access to pensions for women.
High levels of first-tier benefits, particularly when means-tested, strongly reduce pension inequalities and thereby the GPG.
Care‑related pension credits are an effective instrument to cushion the impact of relatively short employment breaks, especially at low-income levels. They mainly benefit women given the strongly unequal division of childcare tasks between men and women.
Protecting survivors’ standards of living following the partner’s death remains an important policy objective. Survivor pensions reduce the gender pension gap in mandatory earnings-related schemes by about one‑third on average. Women benefit from survivor pensions much more than men in all OECD countries where such a scheme exists, and they account for 88% of recipients on average across OECD countries. To support women’s longer careers, recipients should not be eligible to a permanent survivor pension before the retirement age.
Communication efforts should increase women’s awareness of the possibility and importance of splitting retirement entitlements upon divorce. Still, while splitting pension rights is fairly easy to implement in defined contribution and point systems or in defined benefit systems that are based on straightforward accrual rates, it is more complicated to do so in complex and fragmented pension systems as well as in schemes with loose links between contributions and pension entitlements
Reducing minimum earnings or hours worked requirements to be covered by pensions and lowering eligibility conditions related to the minimum contribution records to access pensions would also help decrease gender disparities in old-age income.

Gender disparities in pensions

Copy link to Gender disparities in pensions

Gender pension gaps are large but steadily declining

Women receive pensions that are about one‑quarter lower than men’s on average across OECD countries. In 2024, the gender pension gap (GPG), which measures, among pension recipients aged 65 and over, the difference in the average pension level between men and women relative to that of men, was 10% or less in Czechia, Estonia, Iceland, the Slovak Republic and Slovenia while it was more than 35% in Austria, Mexico, the Netherlands and the United Kingdom, and even 47% in Japan (Figure 2.1). These large GPGs result principally from diverging employment and wage trajectories between men and women, as analysed in the next section. Low pensions currently received by old-age women contribute to the lower confidence of working-age women about whether they will be able to access adequate old-age benefits (Frey, Alajääskö and Thomas, 2024[1]).

The average gender pension gap across OECD countries has declined from 28% in 2007 to 23% in 2024. The most significant decreases took place in Germany, Greece and Slovenia where the gap narrowed by more than 15 p.p. between 2007 and 2024, as well as in Luxembourg, Norway, Portugal, Türkiye by more than 10 p.p. In many OECD countries, strongly declining labour market differences between men and women (next section) are driving this reduction in the GPG, but it takes time for these changes to be fully reflected in lower pension inequalities.1

Figure 2.1. The gender pension gap has declined steadily across countries
Copy link to Figure 2.1. The gender pension gap has declined steadily across countries

Difference between the average pension of men and women relative to the average pension of men

In addition, some pension reforms have contributed to reducing the GPG. For example, Slovenia increased the retirement age for women more than for men since 1999, thereby reducing gender differences; it also introduced additional pension credits for combining part-time work and childcare in 2012 (OECD, 2022[7]). Austerity measures taken during the Global Financial Crisis in Greece were targeted at reducing the highest pensions, generally held by men, leading to a reduction in the GPG (Danchev et al., 2024[8]). In the United Kingdom, the role of the earnings-related component within public pensions has gradually decreased since 2002 and the statutory retirement age of women increased from 60 to 65 between 2010 and 2018, converging with that of men. These measures resulted in a decline of the GPG in the public scheme from 25% for those born in the 1940s to 5% for those born in the 1950s (Cribb, Karjalainen and O’Brien, 2023[9]); nevertheless the total GPG remains large in the United Kingdom at 36% because private pensions, which play a large role, amplify pension inequalities. In France, the pension credits for childcare that were introduced in the 1970s slightly contributed to the large decline in the gender gap in old-age pensions between those born in 1930 and in 1955 (DREES, 2024[10]).

According to projections for several countries, the gender pension gap will decline substantially over time. Using microsimulation models for five European countries, Barslund et al. (2021[11]) estimate that the downward trends in employment and wage differences between men and women will nearly eliminate the GPG by 2050 in Portugal and Slovenia, and lead to a strong reduction to 10% or less in Belgium and Luxembourg. This is despite significant gender gaps remaining in part-time work in all four countries. In Switzerland, the GPG is projected to decline to a lesser extent from 29% to 22% between 2018 and 2070. In the Nordic countries, the GPG is also projected to decline (Andersson, 2023[12]): in Denmark the gender gap in occupational and private pension wealth would disappear for cohorts retiring after 2050 compared to a gap of 22% in 2021; in Finland, the gender gap in public pension would decline from 26% in 2017 to 19% in 2045, and to 15% in 2085; in Norway, the GPG would diminish to 10% in 2033; and, in Sweden, the GPG would shrink to 19% in public pensions and 35% in occupational schemes around 2050. In France, the gender pension gap is projected to steadily decline to 7% by 2060 and stabilise at this level thereafter (COR, 2024[13]).

The gender pension gap does not measure gender differences in standards of living. By contrast, disposable income better captures standards of living and is calculated at the household level. It is the same for each partner within a couple by definition, while pensions that enter the GPG are specific to each individual. Moreover, while disposable income takes into account all sources of income, the GPG does not account for earnings, which make up one‑quarter of disposable income on average among people aged 65 or more (Chapter 7). Also, older people with no individual pensions are not accounted for in measuring the GPG.

Gender differences in household disposable income among older people are substantially lower than gender pension gaps. On average across OECD countries, the gender gap in disposable income among people older than 65 was 10% in 2023. It exceeds 15% in Latvia and Lithuania while it is less than 5% in Chile and the Slovak Republic (Figure 2.2). Moreover, there is no correlation across countries between the gender pension gap and the gender gap in disposable income. For example, Chile and Sweden have a similar GPG, around 20%, but the gender gap in disposable income is more than three times higher in Sweden than in Chile, maybe related to fewer older women living alone in Chile as grandparents, particularly widows, tend to live in the household of their children (Scroope, 2017[14]).

Figure 2.2. Gender gaps in disposable income and in pensions are not correlated across OECD countries
Copy link to Figure 2.2. Gender gaps in disposable income and in pensions are not correlated across OECD countries

Beyond lower average incomes, older women also face higher poverty risks than older men. Women aged older than 65 face higher rates of income poverty compared to men in all OECD countries (Figure 2.3), with the exception of Costa Rica and Iceland. The old-age income poverty rate – defined as the percentage of people living in households with equivalised disposable income less than 50% of the median in the total population – is 17% among women and 12% among men on average in OECD countries.

Figure 2.3. Older women are more likely to be in income poverty than older men
Copy link to Figure 2.3. Older women are more likely to be in income poverty than older men

Share of 66+ with income less than 50% of the median equivalised household disposable income, 2022 or latest

Gender imbalances in pension coverage are substantial

Women’s pensions are lower than men’s partially due to the gender composition of beneficiaries across different pension schemes. This is because first-tier and survivor schemes, in which women are overrepresented, provide lower benefits than earnings-related schemes, in which women are underrepresented.

Women rely more often on first-tier pensions than men. Among OECD countries, more than three‑fifths of beneficiaries of first-tier pensions – minimum contributory pensions, residence‑based or contribution-based basic pensions and old-age safety-net benefits (Chapter 3) – are women, against half of beneficiaries of mandatory earnings-related (second-tier) pensions. By comparison, women make 56% of people aged 65 or more in OECD countries on average (Figure 2.4). The share of women receiving first‑tier pensions is close to 70% or more in Austria, Finland, Latvia, Lithuania, Norway, Portugal and Sweden. In Austria, Finland, Norway and Sweden, this can be attributed to first-tier benefits topping up earnings-related entitlements, while for Latvia and Lithuania it is mainly due to the large proportion of women among older people resulting from very high mortality rates among men.

There are significantly fewer women than men among earnings-related pension recipients (excluding survivor pensions) than among older people, by about 6 p.p. on average among the 29 OECD countries for which data are available. Exceptions are Canada, Denmark, Finland, Germany, Hungary, Latvia, Lithuania, Norway, Poland, Sweden and Switzerland, where the gender balance of recipients of mandatory earnings-related pensions, almost fully mirrors the gender composition of the population aged 65 or more. The difference between the share of women among recipients of earnings-related pensions and among older people is large at around 10 p.p. in Greece, Italy and the Slovak Republic, around 15 p.p. in Belgium, Luxembourg and Spain, and 20 p.p. in Korea; it is even larger at around 25 p.p. in Costa Rica and Japan. Beyond low employment rates of women, substantial differences in some countries result from specific pensions features. For example, in Japan, workers working less than 20 hours a week (mainly women, as in other countries) do not contribute to and build entitlements from earnings-related pensions. In Belgium some women with very small pension entitlements give up their own pensions so that their partners can receive them at a higher rate (75% instead of 60% of the reference wage), thereby increasing total household income. Czechia requires 35 years of contributions to access old-age pensions, Italy 20, and Costa Rica and Spain 15 years, conditions that are less likely to be met by women.

Women are also underrepresented among voluntary (third-tier) pension recipients. In six out of seven countries for which data are available, women make up a smaller share of third-tier pension recipients than their share in the population aged 65 and over: only 40% of third-tier pension recipients are women in Norway, 41% in Belgium, 43% in Switzerland, 45% in Ireland, 46% in Costa Rica and 48% in Germany. However, in New Zealand, the share of women among third-tier pension recipients is not different from their share in the population aged 65 and over, at slightly more than 50%.2 Among the working-age population, the proportion of women participating in voluntary schemes is usually lower than that of men (OECD, 2021[16]). Women tend to work in sectors, such as education, health and social work, that are less likely to provide occupational plans than men-dominated sectors, such as manufacturing. In addition, eligibility criteria based on a minimum number of working hours or on a minimum income threshold tend to restrict women’s ability to join asset-backed pension plans more than men’s, as women are overrepresented among part-time workers and earn less than men. These criteria exist for occupational pension plans in Canada, Japan, Switzerland and the United Kingdom (minimum income thresholds), as well as in Japan and Korea (minimum number of working hours) (OECD, 2021[16]).

Women benefit from survivor pensions much more than men in all OECD countries where such a scheme exists. In all 27 OECD countries shown in Figure 2.4, women account for more than 70% of survivor pension recipients, with an average across countries of 88%. In Chile and Japan, nearly all recipients (97% and 98%) are women. By contrast, Latvia has the lowest share at 73%, which can be attributed to survivor benefits being limited to only one year and not subject to any means-testing.

Figure 2.4. Women receive first-tier and survivor pensions more often than men but are less covered by earnings-related pensions
Copy link to Figure 2.4. Women receive first-tier and survivor pensions more often than men but are less covered by earnings-related pensions

Share of women among pension beneficiaries by scheme type (%) and among the population aged 65+ (%), 2024 or latest

First-tier and survivor pensions mitigate gender differences in pension income

Differences in average pension benefits between men and women vary substantially across pension components. First, first-tier pensions generally compensate for low earnings-related pensions or provide flat-rate benefits, which are in percentage terms more beneficial to low earners. When topping up low pensions, first-tier benefits are higher for those less attached to labour markets.3 Second, mandatory earnings-related (second-tier) pensions have closer links with earnings histories, although the extent of the link varies with the design of specific schemes. Third, voluntary (third-tier) pensions are closely linked to voluntary contributions and tend to provide higher entitlements for workers with higher income, often men.

The gender gap in mandatory earnings-related pensions excluding survivor pensions was 27% on average among 28 OECD countries in 2023. It was around 12% in Czechia and Latvia and almost zero in Costa Rica where the defined benefit earnings-related scheme is highly redistributive (Figure 2.5). By contrast, the gender pension gap in earnings-related pensions exceeded 40% in Chile, the Netherlands and Portugal.

As first-tier pensions are the most redistributive among all schemes, sometimes topping up values from second-tier pensions, women often receive higher first-tier pensions than men. This is the case in Latvia, Norway, Portugal and Sweden where the gender gap among recipients of first-tier pensions was very negative, around ‑20%, in 2023 (Figure 2.5). In Austria, the gap was positive and high (18%), but still much lower than the gap in earnings-related pensions (38%). This positive gap is likely related to the fact that the old-age allowance is granted at a higher rate to couples, but the couple rate is transferred to only one person in the household, often men. Japan and Spain also have positive gaps, although much lower than for earnings-related schemes.

Figure 2.5. First-tier pensions lower the total gender pension gap
Copy link to Figure 2.5. First-tier pensions lower the total gender pension gap

Gender pension gap by scheme across selected OECD countries, excluding survivor pensions, 2024 or latest

Gender pension gaps in voluntary pensions, whether occupational or personal, are larger than in mandatory earnings‑related pensions. Indeed, voluntary pensions rarely include redistributive components and usually have higher contributions among people with high income – who are more likely to be men. The gender gap among recipients of voluntary pensions exceeds 40% in Belgium, Costa Rica, Germany and Switzerland, while it was 24% in Ireland and 15% in New Zealand.4 In Ireland, the gender pension gap is almost entirely due to voluntary pensions because the only mandatory contributory pension scheme does not depend on earnings and treats up to 20 years of care as working periods.5 In Canada, the gender gap in private pensions at 25% in 2021 was much higher than in mandatory earnings related pensions (CPP/QPP) at 16% (Pay Equity Office, 2024[20]).6 Generally, in OECD countries, among the working-age population, women tend to contribute less than men to their asset‑backed pension plans, therefore accruing less (OECD, 2021[16]). Moreover, in many funded defined contribution (FDC) pension plans, women have to rely on their accumulated assets for longer given their longer average life expectancy.

Survivor pensions reduce the gender pension gap in mandatory schemes by about one‑third on average. On average across 24 OECD countries for which data are available, the gender gap in mandatory earnings-related schemes would be 37% if survivor pensions were excluded while the GPG was actually equal to 23% in 2023 when including survivor pensions (Figure 2.6). Survivor pensions reduced the GPG by about 20 p.p. or more in Chile, Costa Rica, Germany, Greece, Israel, Japan, Luxembourg, Portugal and Spain. The impact of survivor pensions on the GPG is only marginal in Czechia, Norway, Latvia, the Slovak Republic and Sweden, due to their low amount (Czechia, Latvia, Norway and the Slovak Republic), or low coverage (Sweden).

Figure 2.6. Survivor pensions lower the gender pension gap substantially in many countries
Copy link to Figure 2.6. Survivor pensions lower the gender pension gap substantially in many countries

Gender pension gap in second-tier (mandatory earnings-related) pensions with and without survivor pensions, 2023 or latest.

Women live longer in retirement than men

Women generally live longer after leaving the labour market. On average across OECD countries, women would live 22.8 years after having left the labour market compared to 18.6 for men, hence 4.1 years more or 22% longer, based on 2024 mortality rates (Chapter 6). Expected years of life after labour market exit are higher for women in all OECD countries, with differences exceeding 6 years in Colombia, Costa Rica and Poland (Figure 2.7). New Zealand records the lowest difference of 2.0 years. Coincidentally, this 22% gap in the expected years of life after labour market exit in favour of women is almost the same numerical amount as the average gender gap in monthly pensions of 23% discussed earlier. This implies that the total amount of pensions paid to men and women over the retirement period may end up being similar on average across OECD countries.

On average, life expectancy differences between men and women explain three‑quarters of the difference in life expectancy at the average labour-market exit age. Indeed, at age 65 in 2024, women have a remaining life expectancy of 21.6 years compared to 18.5 years for men.7 The other quarter is due to women leaving the labour market earlier than men, by 1.1 years on average across OECD countries (see below).8 However, not all additional years lived by women are spent in good health, an issue discussed later in the Chapter.

Figure 2.7. Gender gaps in remaining life expectancy at the average labour market exit age
Copy link to Figure 2.7. Gender gaps in remaining life expectancy at the average labour market exit age

Contribution of differences in mortality and in labour market exit age between men and women in OECD countries, in years, 2024 or latest

Men and women start receiving earnings-related old-age pensions at similar ages in many OECD countries, 64.2 and 63.6 years, respectively, on average.9 The age difference is much larger in countries that provide pension eligibility at lower ages to women than to men: 4.3 years in Poland, 3.0 years in Chile, 2.8 years in Hungary and 2.0 years in Austria (Figure 2.8). By contrast, women start claiming pensions around one year later than men in France, Italy, Norway and Spain. France, Italy and Spain provide earlier access to full pensions to people with long careers and to those covered by special schemes for hazardous or arduous jobs, who are more often men (OECD, 2023[5]). In Norway, women more often than men receive disability pensions, which are transformed into old-age pensions only at age 67, women less often qualify for early retirement due to shorter insurance record and women less frequently combine work with pensions.

Figure 2.8. In a few OECD countries, women start receiving old-age earnings-related pensions at substantially younger ages than men
Copy link to Figure 2.8. In a few OECD countries, women start receiving old-age earnings-related pensions at substantially younger ages than men

Average age of new beneficiaries of old-age earnings-related pensions (excluding survivor pensions) by gender in OECD countries, 2023 or latest

The gender gap in the average age of labour market exit widened between 1970 and 2000 but has almost halved since then to 1.1 years in 2024 (Figure 2.9). These trends seem to be related to slightly different timing of pension reforms affecting men and women. Labour market exit ages declined for both men and women between the 1970s and the 1990s, which was concomitant with measures encouraging early retirement in the context of rising unemployment. Since the mid‑1990s, these measures have been reversed, and pension reforms have tightened early-retirement schemes (Boulhol, Lis and Queisser, 2023[21]). The labour market exit age stopped declining on average across OECD countries, initially for men around the mid‑1990s and then for women at the turn of the century. Recent reform trends toward the unification of pension eligibility conditions for men and women in many countries are likely to result in further narrowing the gender gap in retirement patterns between men and women.

Figure 2.9. The gender gap in the average age of labour market exit widened between 1970 and 2000 but has almost halved since
Copy link to Figure 2.9. The gender gap in the average age of labour market exit widened between 1970 and 2000 but has almost halved since

OECD average of labour market exit ages for men and women, 1970-2024, years

Long-term trends in gender labour market inequalities

Copy link to Long-term trends in gender labour market inequalities

Pension benefits are largely based on contributions made throughout working lives, although the extent of the link between pension and contribution levels depends on the design of the pension system. Employment, hours worked and hourly wages over the career determine lifetime contributions, with lifetime earnings being the product of three components:

Lifetime earnings = career duration in years * average hours worked per year * hourly wage

This section discusses in turn the long-term trends in each component of this breakdown, namely career duration, average hours worked and hourly wages, and then combines them together to show the full picture of gender disparities in lifetime earnings among OECD countries.

Women have shorter careers than men

Employment rates are lower among women than among men for all age groups in most OECD countries, translating into much shorter career durations. The employment rate among women aged 20‑64 was 67% against 82% for men on average across OECD countries in 2023. Based on the age structure of employment rates, the expected career duration was, at 34.3 years, almost 6 years lower for women than for men in the OECD on average in 2023 (Box 2.1). The expected career duration of women varies from less than 25 years in Costa Rica and Türkiye to more than 40 years in Iceland, the Netherlands and New Zealand, and, for men, they vary from around 35 years in France, Greece, Italy, Luxembourg and Spain to more than 45 years in Iceland, Japan, Mexico, New Zealand and the Netherlands (Figure 2.10). In the Netherlands, long careers of both men and women coexist with the large use of part-time employment. By construction, as with the standard measure of life expectancy, expected career duration is only based on current employment rates by age and gender and does not take into account any past data or projections (Box 2.1).

Box 2.1. Measuring expected career duration

Copy link to Box 2.1. Measuring expected career duration

The expected career duration is equal to the average employment rate across 5‑year age groups between 15 and 74 years multiplied by 60 years. It shows what would be the average expected duration of employment in a given year if the employment rates observed that year were applied to the whole career. This is akin to the standard measure of life expectancy that measures what life expectancy would be in a given year for a given cohort if that cohort had the same age‑specific mortality rates in the future as those observed for that year (for the whole population of different cohorts and therefore at different ages) – this means that this measure of life expectancy does not make any projection of changes in future health conditions, which translate into changes in mortality rates. Likewise, the expected career duration measure does not project changes in employment rates. Eurostat provides a similar measure of expected duration of working life (Eurostat, 2024[22]) with two important differences. First, the Eurostat measure is based on labour force participation rates, while the OECD expected career duration herein uses employment rates, because the latter are more consistent with the calculation of lifetime earnings using average hours worked and hourly wages. Second, the Eurostat measure also accounts for mortality rates until retirement while the OECD expected career duration does not, because mortality in periods before the age of claiming pensions has no direct impact on own pension entitlements.

Figure 2.10. Expected career duration differs substantially between men and women
Copy link to Figure 2.10. Expected career duration differs substantially between men and women

Expected career duration in OECD countries in 2023

The gender gap in expected career length is less than 2 years in Estonia, Finland, Latvia and Lithuania, but it exceeds 15 years in Colombia, Cost Rica, Mexico and Türkiye (Figure 2.11). Very large gender gaps in Colombia and Mexico are driven by both exceptionally long careers of men and strikingly short careers of women. The expected career duration is more than 5 years lower among women than men in ten other countries, from the United States (5.1 years) to Korea (8.4), Greece (9.0), Italy (9.4) and Chile (10.6).

The gender gap in expected career duration has declined by about 40% every 20 years since 1980 on average across countries. More precisely, it declined substantially from 17.9 years in 1980 to 10.2 years in 2000, 6.4 years in 2020 and 5.9 years in 2023 on average across OECD countries (Figure 2.11). This resulted mainly from the large increase in career duration for women from 27.9 to 33.3 years between 2000 and 2023, which was more than twice larger than the increase for men from 37.8 to 40.2 years. Beyond economic reforms, structural changes such as improvements in health and education, and shifts toward more flexible work arrangements have contributed to higher employment of both men and women (OECD, 2025[24]). Over the whole period, the gap narrowed across the board in countries with both the highest and lowest initial gaps. However, since 2000, the largest declines, of more than 8 years, were observed in countries with very large initial gaps: Chile, Costa Rica, Ireland, Mexico and Spain.

Figure 2.11. Large reduction of gender gaps in average career duration across all OECD countries
Copy link to Figure 2.11. Large reduction of gender gaps in average career duration across all OECD countries

Difference in the expected career duration between men and women in years

If recent trends continue, the gender gap in career duration will be much lower for cohorts entering the labour market now. For those born in 1950‑1954, hence having reached 70‑74 in 2020‑2024, the observed gender gap is equal to almost 11 years on average across OECD countries, much larger than the gap in the expected career duration measure of 5.9 years in 2023 based on employment rates observed in 2023 across different age groups (therefore belonging to different birth cohorts). Employment rates of women born in 1950‑1954 have a characteristic M-shape, with a decline around the age of having the first child and an increase thereafter until about age 50 (Figure 2.12). This M-pattern has disappeared among younger cohorts in many OECD countries and is no longer visible for the cohort born in 1963‑1967. Between these two cohorts, employment rates of women increased substantially from age 25‑29, while for men the increase is large for the 55‑59 age group only. If past trends are extrapolated, the cohort entering the labour market now will have a gender gap in career duration of about 3 years, or about half the 2023 measure.

Figure 2.12. Employment rates of women are lower than men’s in all age groups
Copy link to Figure 2.12. Employment rates of women are lower than men’s in all age groups

Employment rates for men and women born in 1950‑1954 and 1963‑1967, OECD average

Overall, shorter working careers of women are due to lower employment across all age groups, with large gaps (15 p.p. or more) between age 25 and 44 years on average across OECD countries. Women enter the labour market half a year later than men on average10 and they leave the labour market more than one year before men.

Although on a continued decreasing trend, gender gaps in expected career length will likely persist. Shorter working lives among women are mainly due to deeply entrenched traditional gender roles in many countries, the burden of dual work-family responsibilities for women and the lack of affordable childcare options (OECD, 2023[25]). In particular, the low employment of mothers with dependent children endures (OECD, 2024[26]). Moreover, three factors contribute to women exiting the labour market earlier than men. First, they tend to be younger than their partner in heterosexual couples and retirement decisions are interrelated within couples – although less now than in the past (Moghadam, Puhani and Tyrowicz, 2024[27]). Second, women still provide care more often than men, including for older family members, which often discourages them from having paid work at older ages. Third, ageism may affect older women more strongly than men, and, for example in Australia older women are more likely than older men to be perceived by their peers as having outdated skills, being slow to learn new things or having unsatisfactory results at work (CGEPS, 2023[28]).

Working women spend less hours in paid work than working men

The difference in weekly working hours between male and female workers is still relatively large at 5.1 hours on average across OECD countries, or about 13% in relative terms. In 2023, the gender gap in working hours ranged from around 1 hour in Hungary, Latvia and Lithuania, where part-time employment is rare, to more than 7 hours in Austria, Costa Rica, Germany, Iceland, Ireland, Japan, Mexico, New Zealand, the Netherlands and Switzerland, where part-time employment is more common especially among women (Figure 2.13). The difference in working time between men and women has its counterpart in the unequal share of unpaid work, especially care resposiblities, being borne by women. In some countries, for example Korea and Mexico, long working hours of full-time employees are sometimes incompatible with women’s disproportionate responsibility for unpaid work (OECD, 2023[25]). When mothers engage in paid work, they work fewer hours in many countries than both women without dependent children and fathers (OECD, 2019[29]).

The gender gap in working hours has significantly decreased, from 6.6 to 5.1 hours between 2008 and 2023 on average. This decline has been driven by reduced working hours among men, from 42.4 in 2005 to 40.0 in 2023 on average across OECD countries. By contrast, women’s working hours remained roughly stable over this period. The reduction in the gender gap in working hours was much lower in the previous 15 years, as it decreased by only 0.4 hours between 1993 and 2008 on average across OECD countries.11

Figure 2.13. Gender gaps in average working hours have declined
Copy link to Figure 2.13. Gender gaps in average working hours have declined

Difference in the average weekly working hours between men and women, 1993-2023, in hours

Women still earn substantially less than men per hour of work

The gender gap in hourly wages is large at 11.4% on average across OECD countries. The gender gap in hourly wages measures the difference in the average hourly wage between men and women among all employees as a percentage of men’s.12 The average gap in hourly wages tends to be lower in countries with low women’s employment, likely due to few women working in low-paying jobs. The gap varies from less than 5% in Colombia, Costa Rica, Greece, Italy, Luxembourg, Mexico and Türkiye to more than 20% in Estonia, Germany, Israel, Japan and Korea (Figure 2.14).

Figure 2.14. The gender gap in hourly wages is very large in some countries
Copy link to Figure 2.14. The gender gap in hourly wages is very large in some countries

Gender gap in average hourly wages among all employees, 2023 or latest, hours

Over the past 15 years, the gender wage gap among full-time workers declined substantially in many OECD countries, by 3.9 p.p. on average. Declines were 10 p.p. or larger in Iceland and Ireland while the gap increased by more than 5 p.p. in Chile, Hungary, Latvia and Türkiye (Figure 2.17).13

Figure 2.15. The gender wage gap has decreased in most OECD countries since 2008
Copy link to Figure 2.15. The gender wage gap has decreased in most OECD countries since 2008

Change in the gender wage gap among full-time workers between 2008 and 2022, in p.p.

The much larger role of women as primary caregivers explains a significant part of the gender gap in hourly wages. Hourly wages are lower for mothers than for childless women once other similar characteristics are accounted for (OECD, 2024[26]). Many empirical papers find a negative impact of giving birth on earnings trajectories, while no fatherhood penalty is observed (Bertrand, Goldin and Katz, 2010[31]; Ciminelli, Schwellnus and Stadler, 2021[32]). In their broad meta‑analysis, Cukrowska-Torzewska and Matysiak (2020[33]) find that mothers’ lower wages are mostly explained by the negative impact of childcare‑related employment breaks on human capital deterioration and by women’s choices of jobs and occupations that pay less to accommodate family responsibilities.14 Moreover, in the United States, Wilde, Batchelder and Ellwood (2010[34]) find that wage trajectories diverge sharply among high skilled women between non-mothers and mothers after (but not before) they had children, while there is little difference among low skilled women. In France, however, having children lower more labour income of mothers at lower end of the wage distribution (Pora and Wilner, 2019[35]). Actually, in many countries, gendered educational choices and occupational pathways often diverge substantially even before childbirth, including in the United Kingdom (Strauss and Borrett, 2025[36]). Although they have higher levels of education on average, women less often choose STEM (science, technology, engineering and mathematics) education, which is typically associated with higher wages, and more often public services, education and care‑related occupations, which often pay less but improve work/life balance, through e.g. family-friendly working hours (OECD, 2023[37]).15 More equal educational choices contributed to the decline of the gender wage gap over time in the United States (Altonji et al., 2025[38]). The gender hourly wage gap results partially from women working more often part-time and less frequently long hours compared to men. A number of studies find that the total number of hours worked seems to be positively correlated to the hourly wage level.16

It is very difficult to precisely separate and quantify the impact of preferences versus that of discrimination, on wages. The “choice” to prioritise part-time and flexible work arrangements over working long hours, to request pay increases and promotion less frequently17 and to pursue lower-paying occupations, e.g. those related to care, may actually reflect deep-seated social norms or stereotypes rather than innate preferences (Ciminelli, Schwellnus and Stadler, 2021[32]).18 Moreover, part of the gender wage gap is likely to reflect negative attitudes towards women in the workplace. The substantial impact of preferences, social norms, stereotypes, wage bargaining strategies and negative attitudes towards women on the gender wage gap may explain why a significant proportion of the gender wage gap is left unexplained by individuals’ and jobs’ characteristics, both within and between firms. For example, differences in job characteristics and in observable characteristics between men and women workers (age, education, etc.) explain only around one‑fifth of the gender wage gap in EU countries (Leythienne and Pérez-Julián, 2022[39]). Discrimination and bargaining practices are estimated to account for 10% of the gender wage gap in France and Sweden, 15% in Denmark and Portugal and 20% in Hungary (Palladino et al., 2024[40]). Finally, in the United States, Maloney and Neumar (2025[41]) find, based on a novel index of misogyny constructed from Google Trends data, that a significant part of the gender wage gap results from negative attitudes towards women.

Despite strong improvements, the gender gap in lifetime earnings is very large

Differences in the expected career duration, hours worked and hourly wages between men and women combine into large gender gaps in expected lifetime earnings averaging 35% across OECD countries (Box 2.2). This total gap varies from 14% in Lithuania, 17% in Slovenia and less than 25% in Finland, Latvia, Luxembourg, Portugal, the Slovak Republic and Sweden to about 50% or more in Costa Rica, Japan, Korea, Mexico and Türkiye (Figure 2.16). On average across OECD countries, each of the three components has a similar contribution of a about one‑third with career duration contributing slightly more (14 p.p.) to the expected lifetime earnings gap while hours worked and wages contribute 11 p.p. and 10 p.p., respectively.

Box 2.2. Gender gap of expected lifetime earnings

Copy link to Box 2.2. Gender gap of expected lifetime earnings

The gender gap of expected lifetime earnings is close to the sum of the gaps in the three dimensions, with the exact formula being:

Equation 2.1.

$r_{g a p} = \frac{{r_{m} - r}_{w}}{r_{m}} = 1 - \frac{r_{w}}{r_{m}} = 1 - \frac{l_{w}}{l_{m}} \frac{h_{w}}{h_{m}} \frac{w_{w}}{w_{m}} = 1 - (1 - l_{g a p}) (1 - h_{g a p}) (1 - w_{g a p}) ~ l_{g a p} + h_{g a p} + w_{g a p}$

with r, l, h and w denoting the gender gap in lifetime earnings, career length, hours worked and hourly wages, respectively.

Beyond averages, the main contributing factors differ across countries. In Latin America countries, Czechia, Greece, Italy, Poland and Türkiye, the high gap in expected career duration is the main factor. In these countries except for Chile, Czechia and Poland, this coincides with very low hourly wage gaps, which likely results from large obstacles for women to enter the labour market. Conversely, in Korea, the career-length gap is also large, but the main factor is the high difference in hourly wages between men and women. The above‑average hourly wage gap is the main contributing factor to the gender gap in expected lifetime earnings in Estonia, Finland, Hungary, and Latvia, while in Austria, Germany, Israel and Japan gaps in both hours worked and hourly wages make a significant contribution. The hours-worked gap boosts the gender lifetime earnings gap in Australia, Iceland, Ireland, the Netherlands, New Zealand, Switzerland and the United Kingdom.

Figure 2.16. Women earn one‑third less than men over the lifetime on average across OECD countries
Copy link to Figure 2.16. Women earn one‑third less than men over the lifetime on average across OECD countries

Contribution of expected career duration, working hours and hourly wages to the gender gap in expected lifetime earnings, in p.p., 2022

Over the last 20 years, the gender gap in lifetime earnings decreased in all OECD countries, and very strongly on average by 14 p.p. between 2002 and 2022 (from 49% to 35%). The largest reductions were recorded in countries with large initial gaps: by more than 20 p.p. in Belgium, Greece, Iceland, Luxembourg, Portugal and Spain, although countries with the largest initial gaps, Korea, Mexico and Türkiye, saw only average or low declines (Figure 2.17). All OECD countries have managed to improve women’s employment and reduce the gender gap in expected career duration over the last two decades. Overall, employment trends account for more than half of the 14‑p.p. reduction in the gender gap in expected lifetime earnings, followed by hourly wages (about one‑third) and hours worked (slightly more than one‑tenth). In most OECD countries, gaps were reduced in all three dimensions. However, there is substantial scope for further reductions; keeping the current pace requires strong policy efforts, in particular to overcome women’s underrepresentation in occupations that provide higher wages and to reduce labour income losses among mothers after childbirth (Bertrand, 2020[42]).

Figure 2.17. Changes in career length and hourly wages strongly reduced gender gaps
Copy link to Figure 2.17. Changes in career length and hourly wages strongly reduced gender gaps

Change in the expected gender gap in lifetime earnings in p.p., 2002-2022

Four normative questions about the role of pension policy in addressing the gender pension gap

Copy link to Four normative questions about the role of pension policy in addressing the gender pension gap

Pension policies are shaped by broad normative dilemmas. Some of these dilemmas influence decisions regarding the extent to which pensions should compensate for labour market outcomes. They also affect the selection of instruments for addressing these outcomes, such as targeting parents, carers, women or couples. Another area of debate is about how to deal with gender longevity differences. This section discusses such normative dilemmas.

Should pension systems mitigate the effects of gender labour market inequalities?

Reducing income inequality and alleviating poverty are often part of the objectives of pension systems on top of consumption smoothing and insurance against longevity risks. Pension systems therefore often aim at reducing the impact of labour market inequalities on retirement-income differences, which contributes to reducing the GPG. This, however, can be achieved to a limited extent only, especially given other objectives of pension systems. For example, close links between earnings and pension entitlements increase the transmission of inequality from working age into old age but are consistent with consumption smoothing, i.e. limiting income losses faced by individuals when moving into retirement. The weight countries give to the redistribution and the consumption-smoothing objectives is a political choice, which depends on societal preferences. In OECD countries, pension systems transmit on average about two‑thirds of overall lifetime earnings inequality on to pension inequality (OECD, 2017[43]).19

Policy instruments that limit pension inequality tend to reduce the gender pension gap. These instruments reduce the impact of lower wages, shorter working hours and shorter careers on pension benefits. They include progressive pension formulae, minimum contributory and basic pensions and pension credits for employment breaks, including childcare credits that mainly benefit women given the strongly unequal division of childcare tasks between men and women. Figure 2.18 provides a snapshot of pension instruments more or less directly targeting women, mothers, couples and care providers. One straightforward way to limit the transmission of income inequality into old age is to have a high level of first-tier benefits (minimum contributory pensions, contribution-based or residence‑based basic pensions or targeted benefits), which are unrelated to previous earnings.

Figure 2.18. Measures affecting women’s pensions
Copy link to Figure 2.18. Measures affecting women’s pensions

Should there be additional pension measures that specifically deal with gender differences in wages and employment? The answer is not straightforward given the difficulty to disentangle the sources of gender inequalities arising in the labour market. For example, the larger use of part-time work among women may result from individual preferences, choices within couples or gendered social norms. While these explanations all lead to the unequal division of household tasks and unpaid work within couples and to different occupational choices (see above), they may have different policy implications. Whether pension policy instruments should correct inequalities that result from how heterosexual couples divide tasks between themselves is not obvious. Moreover, such additional instruments perpetuate gendered social norms, as related debates in Nordic countries have emphasised for several decades (Andersson, 2023[12]; Schmauk and Kridahl, 2024[44]).

Offsetting the impact of gender discrimination seems to provide a clearer justification for additional pension measures, but this also raises complex questions. On the one hand, addressing discrimination can be more clearly justified than part-time employment because people do not choose to suffer from discrimination while part-time work may result from genuine choices. On the other hand, as discussed in the previous section, it is not easy to disentangle the effects of discrimination on labour market outcomes from that of other factors, which can be addressed by general redistribution instruments. An additional difficulty arises from the horizontal equity perspective: compensating women for discriminatory practices through pension measures would require similar compensation mechanisms for some other discriminatory practices affecting other population groups, such as race‑based, migration-related or disability-related discrimination.

A general normative principle is that first-best policies should tackle inequalities when they arise, rather than putting a large burden on pension systems to try to correct them. These inequalities are steadily building up during working age or even before, during the education period, even though they are declining along various dimensions, as discussed above. These first-best policies include combating gender stereotypes, fighting against discrimination and promoting an equal division of household and care tasks within couples. Yet, if these policies have not been in place in time or have not produced effective results, should, for example, women be granted a pension bonus? This would be the most direct way to reduce the GPG.

Should specific pension policy instruments target women and couples?

Earlier retirement ages for women than for men, spousal pension supplements, survivor pensions and pension splitting are, to some extent, based on the notion of the single‑breadwinner model. In its extreme form, women do not participate in the labour market, do unpaid work at home and are financially dependent on men. In that case, initial choices related to sharing care responsibilities become permanent as switching roles becomes more costly over time, which perpetuates the gendered division of tasks. This model looks as something of the past as women’s employment has increased substantially over the recent decades. Yet, pension instruments that are still based on the single breadwinner model can incentivise behaviour that perpetuates women’s financial dependence on their partner.

In some countries, women are allowed to retire earlier than men but then with a lower pension. This drives old-age inequalities and raises the gender pension gap because retiring earlier results in lower pension entitlements. Since the late 1970s countries have been making pension systems more gender neutral (Boulhol, Lis and Queisser, 2023[21]). Making pensions accessible to women at lower ages than men perpetuates gendered social norms and is difficult to justify given women’s longer expected lives. Earlier access given to women is consistent with the views that women’s primary role is to provide care, including for grandchildren and older family members, that women should not work at older ages, more generally, and that wives should be able to retire together with their husbands who are older on average.

Spousal supplements or higher accrual rates of contributory benefits for couples result in higher pensions being granted to one‑earner couples than to single earners, as in Belgium, Japan, Korea and the United States. Some European countries abolished – at least for new comers – benefits for financially dependent spouses over the past decades in response to the rise of the two‑earner household model, for example the United Kingdom in 2010, France in 2011 and the Netherlands in 2015 (Brown and Fraikin, 2022[45]).

Almost all OECD countries cover survivor risks for at least some parts of the population, with eligibility criteria for and coverage of survivor pensions differing substantially across countries (OECD, 2018[46]). Following the death of a partner, survivor pensions have pursued two main objectives. First, they protect widows or widowers from poverty risks by cushioning sharp drops in disposable income to low absolute levels. This is less relevant now than in the past, as nowadays all OECD countries have instruments directly targeted at poverty alleviation. Second, more relevant today, they contribute to insuring against the decrease in disposable income and standards of living upon the death of the partner, in the same way as old-age pensions help avoid a sharp drop in income when moving out of paid work upon retirement. As women live longer, are often the younger spouse and earn less, they tend to benefit substantially more from survivor pensions even if the rules are gender neutral.

Some pension systems introduced the option to split pension entitlements within households, but its use remains marginal. While survivor pensions provide protection to individuals less attached to the labour market in the event of the partner’s death, pension splitting provides income protection to the partner less attached to the labour market also in the event of divorce/separation. It is fairly easy to implement splitting in defined contribution and point systems or in defined benefit systems that are based on straightforward accrual rates, but it is more complicated to introduce splitting in complex and fragmented pension systems as well as in schemes with loose links between contributions and pension entitlements. Splitting some pension rights tends to provide more financial security to women, especially in the case of divorce. Due to the higher life expectancy of women, shifting pension entitlements from men to women through pension splitting boosts total pension spending, negatively affecting pension finances. Conversely, pension splitting would often lower survivor pensions expenditure and first-tier pension expenditure if means-tested. While pension splitting can efficiently reduce pension inequalities within couples, it cannot replace survivor pensions in smoothing income after the partner’s death.

Should pensions reflect longevity differences between men and women?

Longevity differences between men and women are ignored in the calculation of pension benefits in mandatory public pensions in all OECD countries. Benefits do not account for women’s longer lives in public pension schemes, be they defined benefit, points, notional or funded defined contribution. This is consistent with the pooling of longevity risks across the whole population. Moreover, given women’s lower pension entitlements, ignoring longevity differences between men and women avoids lowering further women’s monthly pensions. However, ignoring longevity differences more broadly is sometimes challenged as, within genders, it reduces the progressivity of pension systems given that high-income people tend to live longer than low-income individuals.

In the European Union, private pension schemes cannot take into account longevity differences between men and women to calculate pension benefits, even when they are funded defined contribution. The law forbids to use the information about gender-specific mortality tables for setting both annuity premiums and benefits, as it would be perceived as discriminatory against women (Council of the European Union, 2004[47]). However, the higher share of women among beneficiaries of a specific pension plan tends to put pressure on annuity providers to increase premia. Higher premia in turn discourage men from taking annuities if they are not mandated, boosting premia further. This well-known mechanism of so-called adverse selection in insurance markets leads to the underuse of annuities, among other factors. Outside the EU, women’s higher life expectancy lowers their monthly retirement income from defined contribution schemes, such as in Australia, Costa Rica and Israel, because accumulated assets need to finance pensions over a longer period. In these countries, however, annuitisation is not mandatory, and effectively large amounts of payments are made through lump sums or programmed withdrawals, which leave women to spread these payments over longer periods, de facto reducing their monthly benefits, or risking outliving the assets.

Gender differences in healthy life expectancy at older ages are smaller than differences in remaining life expectancy, or even non-existent according to some measures. Based on some subjective survey data, men and women can expect to live similar numbers of years without any health limitations or in good health at the age of 65 on average across EU countries (Di Lego, Di Giulio and Luy, 2020[48]). Some other subjective measures calculated by Di Lego, Di Giulio and Luy (2020[48]) show higher numbers of healthy life years for women than men but with smaller gender differences than in total life expectancy; indeed, in all EU countries, women can expect to spend more years without any severe health conditions than men at age 65. Gender differences in healthy life expectancy are also smaller than in life expectancy according to model-based estimates, using current rates of ill‑health and mortality: women can expect to live 2.1 more healthy years at age 60 than men on average across OECD countries, compared to the life expectancy difference of 3.4 (WHO, 2025[49]).

It is not the purpose of pension systems to deal with differences across population groups in health status during retirement and therefore in healthy life expectancy. One exception may be when these health differences result from hazardous or arduous jobs, as discussed in the 2023 edition of Pensions at a Glance. Indeed, pension systems pursue different objectives that relate in different ways to providing income from the retirement age until death; whether, and if so how, they should account for health status during retirement is not straightforward. Other policies outside the pension area are better suited to deal with health-related issues. Healthcare systems aim to prevent, postpone and eventually deal with health deterioration with age directly, while disability benefits and long-term care systems compensate for poor health outcomes.

Should pensions compensate or reward mothers and carers?

Pension systems commonly compensate for at least part of pension entitlements lost while providing childcare, including through pension credits. These childcare‑related credits can be linked to previous earnings, to maternity/paternity and parental-leave benefits (pension credits for parental leave can generally be shared between parents) or be flat-rate; they are limited in time, either for a given number of years or granted up to some child’s age; moreover, they may compensate for reduced working hours. They are generally less generous for longer breaks and for older children (Chapter 5). The parent who actually provides childcare receives pension credits, hence, while fathers can benefit from them, they actually do so much less often than mothers. Similarly, pension credits can apply to employment breaks taken to provide care to older individuals or adults with disabilities, which women predominantly do.

Pension credits reward caring for children and limit gender inequalities in retirement income. Pension credits are particularly valuable tools in countries where mothers face big obstacles to resume paid employment, whether due to the scarcity of childcare services, discriminatory labour market practices or other factors. By helping carers to qualify for old-age pension, pension credits contribute to reducing old-age poverty and enhancing retirement-income adequacy. Pension credits should partly compensate carers for pension entitlements lost during the provision of childcare without unnecessarily prolonging employment breaks and without excessively inflating fiscal costs. For example, in Estonia and Sweden, credits are given based on 100% and 75% of the nationwide average income, respectively, resulting in higher replacement rates for low earners. Likewise, pension contributions during parental leave are proportional to past earnings capped at 60% of the average wage in Poland.

Beyond compensating for breaks in employment, some pension systems provide higher benefits to mothers, and benefits typically increase with the number of children. Such instruments target mothers without generating disincentives to work or incentivising reduced working hours, and they also compensate the GPG beyond employment breaks, e.g. for the motherhood penalty in wages.

Beyond reducing old-age income inequalities, providing benefits for mothers and pension credits for childcare can serve other policy objectives. They are sometimes seen as part of a package of broader family policies aimed to compensate for the direct and indirect costs of raising children (Letablier et al., 2009[50]). Such policy packages often include public provision of childcare, child benefits and preferential tax treatment of families with children. As pension credits do not compensate for the direct and immediate costs of having children, they cannot effectively substitute these other family-policy instruments. Compared to spousal supplements or gender-specific pension rules, child-related pension credits are more aligned with modern family policies. Alternatively, providing bonuses for mothers and pension credits for childcare are sometimes justified as part of pro-natalist policies. Decreasing fertility accelerates the ageing of the population structure and undermines the finances of PAYG pension systems. However, the argument that increasing pension entitlements for having children raises fertility through financial incentives is dubious and the empirical evidence supporting this is lacking.

What countries do: pension rules and gender inequalities

Copy link to What countries do: pension rules and gender inequalities

Gender differences in retirement ages

Women still have a lower normal retirement age than men in nine OECD countries. The normal retirement age is the age at which one can retire after a full career without penalty. However, based on current legislation, this gender difference will be eliminated in Austria, Lithuania and Switzerland, while it will persist for the generation entering the labour market in 2024 in Colombia, Costa Rica, Hungary, Israel, Poland and Türkiye (Figure 2.19, Panel A).20 Moreover, in Israel, the gender gap in the statutory or normal retirement age will have narrowed from five to two years between 2022 and 2032, while there has been no gender difference in the minimum age to access occupational pensions since 2014. In Chile, men and women have the same eligibility conditions to the residence‑based basic pension, but women can retire five years earlier in mandatory defined contribution pensions. In Italy, the statutory retirement-age gap closed in 2019, but there are still some gender differences in eligibility conditions.21

Over the last 30 years, pension eligibility conditions have converged between men and women in some countries, including Belgium, Czechia, Germany, the Netherlands and the Slovak Republic.22 Moreover, Belgium gradually eliminated the higher accrual rates benefiting women between 1998 and 2009. In 2019, Slovenia decided to eliminate from 2025 women’s earlier access to pensions and their related higher accrual rate, which was in place to limit the impact of the lower retirement age on pension entitlements; Slovenia was, the last OECD country to provide a higher accrual rate to women. By contrast, in 2025, Mexico introduced an earlier access to residency-based basic pensions for women at age 63, to be expanded at age 60 in 2026, while men remain eligible from age 65 (Chapter 1). Moreover, in 2024, Colombia started to gradually reduce the period required to qualify for a full contributory defined benefit pension from 1 300 to 1 000 weeks by 2036, while maintaining the 1 300‑week requirement for men.

In countries where women can retire earlier or where benefits depend on gender-specific life expectancy, women’s pensions are negatively affected. Due to lower normal retirement ages, women will have lower pensions in Costa Rica, Colombia, Hungary, Israel, Poland and Türkiye (Figure 2.19, Panel B). A 5‑year difference in the retirement age lowers pensions of full-career women by 25% compared with those of men having the same wages in Poland, but by only 6% in Colombia. This is because, beyond lower related entitlements, pensions are automatically adjusted to the age of claiming pensions in Poland’s NDC scheme, while this is not the case in Colombia, where additionally the 80% cap to the replacement rate means that additional years of work do not accrue additional pension entitlements.23 Higher women’s life expectancy also lowers their future pensions from defined contribution schemes in Australia, Costa Rica and Israel because defined contribution schemes in these countries pay less every month to women than to men for the same amount of accumulated assets, for example as annuities are calculated with gender-specific mortality tables. A recent pension reform in Chile (Chapter 1) will eliminate the negative impact of higher women’s longevity on pensions from the FDC scheme by providing a compensating bonus to women as if they had men’s mortality tables, financed by additional pension contributions paid by everyone (Chapter 1). Before this reform, the gender gap in future theoretical pension was about 6% (for the same career and the same wages). In 2024, Colombia eliminated the option to switch contributions between the FDC scheme and pay-as-you-go DB scheme and from 2025 pension contributions for earnings up to 2.3 times the minimum wage will finance the DB scheme only (Chapter 1). The implementation of this reform is uncertain after the Constitutional Court suspended the reform in June 2025. Thereby, for earnings up to the threshold, higher women’s longevity is no longer affecting their benefits. The 2004 European Union directive mandates the use of unisex mortality tables (Chen and Vigna, 2017[51]).24

Figure 2.19. Women have lower retirement ages in some countries, reducing their future pensions
Copy link to Figure 2.19. Women have lower retirement ages in some countries, reducing their future pensions

Mothers can retire earlier than childless women in some countries. In Czechia, France, Italy, the Slovak Republic and Slovenia, mothers can retire between four months and four years earlier than childless women, depending on the county and the number of children. For example, in France, each child adds two years to the contributory record of a mother; therefore, mothers can reach the full retirement-age condition at younger ages than childless women.25 In Czechia, the possibility for mothers to retire earlier will disappear in 2037 based on current legislation. In the Slovak Republic and Slovenia, fathers can alternatively benefit from this measure.26 This earlier access to pensions for mothers reduce the future normal retirement age – the age when a full-career worker entering labour market at age 22 can retire without penalty – only in France and the Slovak Republic, and by one year for mothers of two children (Table 2.1).27 In Slovenia, in case of full careers, the normal retirement age for mothers is the same as for childless women. In this case, mothers receive a bonus equal to a one‑year accrual for each of the first three children. However, for example when having two children, mothers can retire 16 months earlier if they reach the contribution‑length requirement of 40 years before age 60 and they forgo the pension bonus for having children. In Italy, the normal retirement age for mothers and childless women is the statutory retirement age, but pensions of mothers are increased by applying a more favourable transformation coefficient in the notional defined contribution pension formula.28 Retiring three years before the statutory retirement age will be possible for women with a long contribution record of 41.8 years or high enough pensions (2.8 times the old-age social allowance, which was 55% of the average wage in 2024). Mothers with two children who do not qualify for these early retirement options can retire 8 months before the statutory retirement age, but, in that case, they have to forego the more favourable transformation coefficient.

Avoiding pension penalties requires delaying retirement in the case of a five‑year childcare break in Greece and Portugal, as well as in France and Spain for a ten‑year break. Moreover, in Slovenia, mothers with a five‑year break can only access pensions later than full-career mothers, as in Luxembourg in the case of a ten‑year break. This is because childcare‑related credits offset only part of long breaks (Table 2.1). For example, in Slovenia, a 40‑year contribution record is required to retire before the statutory retirement age of 67 in the future and pension credits cover one year of contributions per child. Hence, a mother of two children taking a five‑year break will have to retire three years later than a full-career woman.

Table 2.1. Motherhood or childcare‑related employment breaks affect normal retirement ages in seven OECD countries

Copy link to Table 2.1. Motherhood or childcare‑related employment breaks affect normal retirement ages in seven OECD countries

Future normal retirement ages for women with two children starting their career at age 22 with a full career or with a 5‑ or 10‑year employment break for childcare compared with childless women

Country	Future normal retirement age, full career childless women (A)	Retirement age adjustment for a mother of two children, compared to (A)
Country		Full career	Having a 5‑year-career break and working until retiring without penalty	Having a 10‑year-career break and working until retiring without penalty
France	65	‑1	‑1	2
Greece	66		1	5
Luxembourg	62			2
Portugal	68		1	2
Slovak Republic	69	‑1	‑1	‑1
Slovenia	62		3	3
Spain	65			0.5

Source: OECD pension model.

Small impact of pension indexation rules on the gender pension gap

A more generous indexation of pensions in payment benefits relatively more individuals with higher life expectancy, and thus tends to reduce the GPG. The effect of indexation on the GPG comes from gender differences in life expectancy, which imply that indexation affects women for longer periods on average. Therefore, moving from e.g. price to wage indexation reduces the GPG.

However, there are trade‑offs. Through the same mechanism, related to differences in life expectancy across population groups, a more generous indexation benefits more, within genders, the socio‑economic groups with longer expected lives, thereby increasing income inequality as the most disadvantaged groups tend to have shorter lives. Moreover, while a more generous indexation benefits everyone, it raises pension expenditure. It is therefore more insightful to compare the impact of indexation alternatives for a given level of pension expenditure. While price indexation is needed to sustain the purchasing power of pensions, more than price indexation for a given level of total spending reduces pensions at retirement for everyone: in that sense a more generous indexation is likely to come at the cost of lower benefits during the first part of the retirement period, negatively affecting the socio‑economic groups with lower life expectancy.

Quantitatively, pension indexation has a limited impact on the gender pension gap. To measure the impact of indexation through gender differences in life expectancy, it is assumed that the other key components, initial pensions and normal retirement ages are the same between men and women in each country. With these assumptions, the theoretical gender pension gap would be 1.3% larger on average across countries if pensions were indexed to prices in every country than if they were indexed to wages.29

Pensions mitigate the transmission of earnings inequalities into old age

By boosting old-age income at the lower end of the income distribution, first-tier benefits lower the gender pension gap. On average across OECD countries, based on current legislation, a person born in 2002 who will not have worked at all during the entire life and therefore not contributed towards pensions will receive old-age benefits equivalent to 16% of the gross average wage (Annex 2.A). Workers with a full career from age 22 in 2024 and earning 25% of the average wage (as an order of magnitude close to working half-time at the minimum wage in many countries) can expect old-age benefits totalling 24% of the average wage on average across OECD countries. This typically implies large replacement rates.

The following cases compare the future theoretical pension entitlements of women who have had a full career and earned the average wage with those who have experienced either lower pay, as implied by the current gaps in hours and hourly wages, or a shorter expected career duration, as described above. These cases do not take into account survivor pensions. The results are produced with the OECD pension model.

Even around the average wage, many mandatory pension schemes mitigate the transmission of gender wage gaps into the gender pension gap. The gender gap in pay (or total wages, made of hours worked and hourly wages) averages 23% across OECD countries and the resulting gender pension gap averages 15% among workers without career breaks, representing a reduction of almost one‑third. This reduction exceeds 20 p.p. in Australia, Estonia, Ireland, Israel, New Zealand and Korea, and is around 15 p.p. in Canada, Czechia, Denmark, Iceland, Mexico, the United Kingdom, Switzerland and Japan (Figure 2.20, Panel A). All of these countries have strong redistributive components in earnings-related schemes or substantial basic pensions. In other OECD countries, mandatory pensions transmit almost all gender wage differences around the average wage.

The average 15% gender gap in career length across countries would translate into a future average gender gap in pension entitlements of 12%. However, this exercise underestimates the extent to which pensions mitigate career gaps, as it does not include pension credits for employment breaks due to childcare or unemployment. Compared to reducing the pay gap, pensions have a substantially less pronounced mitigating impact on employment gaps in countries with substantial contributory-based basic pensions, for example in Czechia, Ireland and Korea (Figure 2.20, Panel B). Shorter careers have no impact on the future mandatory pensions in Mexico and the United States as long as the contribution periods are at least 20 and 35 years, respectively. In the United States, the full pension accrual is reached after 35 years of contribution, and in Mexico a recently introduced pension top-up guarantees a 100% replacement rate up to the 2024 average wage for those with at least 20 years of contributions. Other mechanisms, such as higher or lower accrual rates at older ages, and uprating past earnings or contributions to more or less than wages, play minor roles.30 Neither low earnings (up to the average wage) nor short careers reduce pensions in Australia for the assumed case, as the residence‑based basic pension (Age pension) fully compensates for reduced occupational pensions (Superannuation).

Figure 2.20. Pensions mitigate the transmission of gender earnings gaps into pension gaps
Copy link to Figure 2.20. Pensions mitigate the transmission of gender earnings gaps into pension gaps

Care‑related pension credits accrue pensions for care periods

Most OECD countries better protect the impact of childcare‑related employment breaks on pensions than of unemployment. Pension credits are designed to compensate for a break in “working time” so that there are no significant gaps in pension entitlements (Natali et al., 2024[52]). Credits for childcare typically cover career breaks until children reach a certain age. They are generally less generous for longer breaks and for older children.31 Some countries (Czechia, Greece, Hungary and Luxembourg) factor childcare into assessments of eligibility but disregard them when computing the earnings base, thereby limiting the negative impact. Childcare credits were introduced between the mid‑1980s and the mid‑1990s in Austria, Belgium, Germany, Ireland, Japan, Norway and Switzerland, and in the 2000s in Denmark, Finland, Korea and Portugal. In 2024, Australia decided to finance contributions for the mandatory DC scheme (Superannuation) from the public purse for the period of parental leave (up to 6 months), with payments from July 2026 (Chapter 1). Also in 2024, Colombia introduced a reduction in the career-length requirement to access pensions by up to 50 weeks of childcare for each of the first three children. The implementation of this reform is uncertain after the Constitutional Court suspended the reform in June 2025.

The design of childcare credits is largely gender neutral, and the pension credits beyond the maternity leave can be granted to either parent who actually gives up work to provide care. Still, in practice, most of the breaks are used by mothers, for example, more than nine in ten in Czechia, Germany, Hungary, Ireland and the Slovak Republic, three‑quarters in Lithuania and Italy, and two‑thirds in Finland.

Nine OECD countries give credits just for having had children or provide pension bonuses to parents, irrespective of whether a career break occurred. Extra years of credit are given in Austria, France, Germany, Korea and Slovenia, a more favourable conversion factor is applied in Italy, and a pension bonus is given in Czechia, Hungary and Spain. In Austria, Czechia, Germany and Slovenia, parents decide who receives the extra years, and they can be split between them. In France, since 2023, six extra quarters are credited to mothers only, two quarters can be split between parents, and a 10% increase of pensions for having at least three children goes to both parents. In Italy, the pension bonus for having children applies to mothers only and, in Spain, the pension bonus is granted to the parent having the lower pension. As a result, in these nine countries, mothers of two children can expect their total pension entitlements to be higher than those of childless women with the same career, from about 2.1% in Spain to 6.2% in France (Figure 2.21).

Mandatory pension systems cushion about half of the effects of a five‑year employment break on pensions for mothers with two children on average across OECD countries. On average across OECD countries, a five‑year employment break for childcare reduces pensions of mothers with two children by 5% at the average‑wage level, while this five‑year break represents a shorter career of about 11% on average.32 In countries without or with very weak compensatory mechanisms, such as Israel and Türkiye, a five‑year employment break for childcare indeed reduces pensions by 11% for people earning the average wage (Figure 2.21). In eight OECD countries the impact of such an employment break is less than 1%: Belgium, Ireland and Japan grant substantial pension credits for childcare; in Colombia, Spain and the United States maximum accruals are reached after 30, 37 and 35 years, respectively; in Mexico a top-up guarantees 100% replacement rate up to the 2024 average wage (indexed with prices – Chapter 1); and, in New Zealand, only the residency-based basic pension is mandatory. In France and Spain, taking also into account pension credits for mothers, the pension entitlements of mothers of two children with a five‑year employment break are 2‑3% higher than those of full-career childless women, while they are 4% lower on average across OECD countries. In Czechia and Hungary, the credits granted to mothers for two children exactly offset the impact of the five‑year employment break. Going beyond these typical cases, in France, childcare‑related pension credits and bonuses for mothers compensate women almost fully for the impact of having children, including through reduced wages and hours worked, on their pension entitlements (Bonnet and Rapoport, 2019[53]). Taken together, credits for having children and for childcare‑related employment breaks result in pensions of mothers of two children experiencing a five‑year employment break being only 4% lower than those of a childless full-career woman on average across OECD countries.

Low-earners are better protected against the impact of childcare‑related breaks in some OECD countries (Chapter 5). In Germany having a child gives one parent a credit of one pension point annually for three years, thereby making it equivalent for pension purposes to earning the average wage throughout the credit period. In Estonia and Sweden, credits are given based on the nationwide average income and 75% thereof, respectively, resulting in higher benefits for low earners. Austria and the Slovak Republic provide flat-rate credits during childcare breaks which are worth more to lower earners. In Australia, Colombia, Iceland and Poland safety-nets and minimum pensions compensate particularly low-earning mothers for entitlements lost during childcare.

Figure 2.21. Pensions cushion significantly the impact of a five‑year break for childcare in many OECD countries
Copy link to Figure 2.21. Pensions cushion significantly the impact of a five‑year break for childcare in many OECD countries

Effects of having two children and having a 5‑year employment break for childcare on gross total pension entitlements at the average‑wage level

In asset-based pensions, pension right accruals may continue during periods of maternity and parental leave. However, this is not the case in some countries, such as Austria, New Zealand and the United States, where employers generally stop contributing on behalf of mothers on maternity leave (OECD, 2021[16]). In Australia, 81% of employers who offer parental leave pay contributions to defined contribution pensions on that leave (WGEA, 2025[54]). In Estonia, Iceland, Latvia, Poland, the Slovak Republic and Sweden, the government or the social security institute pays contributions to the mandatory pension account of mothers on maternity or parental leave, while in Chile, Germany and Lithuania mothers receive public subsidies into their pension plan based on the number of children. Even when contributions continue during leave, the earnings base used to calculate these contributions is lower than past earnings in some countries (e.g. Estonia, Iceland and Poland), thereby reducing the level of contributions compared to a period of full activity.

Some countries also credit periods spent providing informal family care for adults. For example, Germany, Norway and the United Kingdom grant pension credits to both employed and not employed informal carers who provide at least 10, 22 and 20 hours of care per week, respectively. In Finland, provided that they register at the municipality, family caregivers are entitled to a care allowance that accrues pension rights, which amount is higher for more intensive care needs (Euro carers, 2025[55]). Similarly in Denmark, Estonia and Hungary, the care allowance accrues pension entitlements for carers. In Austria, the government has been paying pension contributions for informal carers since 2009. Furthermore, employment leave for caring accrues pension rights in Belgium, France and Spain. In Ireland, up to 20 years of providing family care counts towards the contribution-based basic pension. In Lithuania, the parent or a stepparent taking care of an adult child with disabilities is covered by pension insurance.

Survivor pensions, pension splitting and spousal benefits

Almost all OECD countries provide some protection against the death of a spouse or a partner through survivor pensions. Consumption smoothing, i.e. limiting the risk of a fall in standards of living, is currently the key objective pursued by survivor pensions, which de facto help reduce the pension gap between men and women. In the 2018 edition of the Pensions Outlook, the OECD undertook an in-depth analysis of survivor pensions in OECD countries (OECD, 2018[46]). Coverage by permanent survivor pensions is included in mandatory contributory pensions in all countries except Australia, Latvia, the Netherlands, New Zealand, Norway, Sweden and the United Kingdom. While marriage used to be required to access survivor pensions, an increasing number of countries have expanded survivor benefits to civil unions and even cohabitations. Most countries require that the partnership had lasted for some minimum period.33 Moreover, over the last decades, gender differences in eligibility for survivor pensions have been eliminated in many countries, but a few exceptions remain. In Israel and Japan, the access for men is more restricted than for women, but this gender disparity will be eliminated in Japan in 2028. Until 2022, men in Switzerland were only eligible for survivor pensions if they had a dependent child, whereas this condition has never applied to women. Except for eight OECD countries, survivor pensions are granted after divorce, treating this entitlement as a right acquired during the marriage, even though the consumption-smoothing objective is not relevant in that case when the ex-partner dies. In Finland, the survivor pension after divorce is linked to the alimony payment. In 2024, Canada eliminated survivor pensions for separated couples who had split their pension entitlements.

The impact of survivor pensions on the gender pension gap is expected to decrease as women’s labour market outcomes are improving and survivor pensions are means-tested in most OECD countries. Between 2011 and 2021, expenditures on survivor pension decreased from 1.0% to 0.8% of GDP on average across OECD countries, while old-age pension expenditures increased from 7.6% to 8.5% of GDP (OECD, 2025[56]). Survivor pensions pay around half of the deceased’s mandatory contributory pension to never‑working survivors on average across OECD countries, and more than four‑fifths in Mexico, Poland and the United States (OECD, 2018[46]). Most countries reduce the survivor benefits for spouses based on their own pension entitlements. In the case of both partners with the same full career at the average wage, the survivor pensions replace about one‑quarter of the deceased’s mandatory pension on average across OECD countries. Targeting survivor pensions towards low earners is particularly strong in Austria, Canada, Estonia, Ireland, Japan, Slovenia and the United States.

Survivor pensions available from early ages discourage women’s employment and thereby might increase the gender pension gap. No minimum age requirements apply for receiving a permanent survivor pension in Austria, Canada, Chile, Costa Rica, Colombia, Ireland, Italy, Korea, Luxembourg, Mexico, Spain and Türkiye while only widowed persons (who are neither disabled nor have dependent children) above a certain age are eligible in 17 OECD countries. The lowest minimum age is 35 years in Portugal and 40 years in Israel. Hence, while recipients should not be eligible to a permanent survivor pension before the retirement age, survivor pensions are helpful to insure against the decrease in disposable income relative to the situation prevailing before the death of the partner, in the same way as old-age pensions help avoid a sharp drop in income upon retirement.

Splitting of pension entitlements means transferring old-age pension entitlements from one partner to the other. Splitting can take place while contributing, upon separation or upon retirement. For ongoing relationships, splitting provides the partner who is less attached to the labour market with more financial independence and security. This independence becomes even more important when the couple separates or after the death of the partner.

Despite having been available for a few decades in some countries, pension splitting has not gained much popularity. In Canada, Germany, Japan and the Netherlands, pension splitting is the default option when a marriage ends. Canada introduced pension splitting in the event of divorce in 1978, and survivor pensions are no longer paid if pensions entitlements are split in the CPP scheme in 2025 or later. Germany introduced the possibility to split pensions in 1977 (West Germany back then) and, in 2002, introduced the option of trading the entitlement to a survivor pension for a 50‑50 pension splitting when the younger spouse retires (Schmauk and Kridahl, 2024[44]). For the couple as a whole, survivor pensions are generally more beneficial than the 50‑50 splitting and the take‑up rates of splitting have been very low. In Japan, pensions can be shared upon mutual agreement during divorce proceedings. Alternatively, the financially dependent spouse can submit a request for splitting. In effect, pension splitting upon divorce is relatively common, with almost 38 000 splitting arrangements in 2023, or about one‑fifth of the number of divorces. In occupational pensions in the Netherlands, pension splitting during divorce has been possible (and encouraged as the default option) since 1995, and during marriage since 2007. Without being the default option, Chile introduced pension splitting for divorced couples in 2009, and courts can split pension entitlements, at the default 50‑50 rate, even without mutual agreement. Since then, only 7 530 men and 170 women transferred their pension entitlements to their partners’ accounts. Occupational and private pensions can be divided by court order following divorce in Canada, Denmark, Germany, Ireland and Sweden.34 Occupational pensions are not automatically split in a divorce in the United Kingdom but are considered part of the marital assets. Additionally, in Austria, Denmark, Germany, Ireland, the Netherlands and Sweden, registered partnerships or other legally recognised unions may be eligible for pension splitting upon separation.

Beyond the Netherlands, pension splitting can occur for ongoing partnerships in Australia, Austria, Canada, Sweden and Switzerland. Switzerland is the only country to have made pension splitting mandatory. Since 1997, half of the joint couple’s earnings during the marriage is used to calculate individual benefits in the public scheme. In Australia, spouses can split up to 85% of contributions to the DC superannuation scheme upon request without divorcing. Austria allows the transfer of up to half of the employed parent’s public pension entitlements to the caregiving parent’s pension account within the first 14 years after childbirth. In Sweden, it is possible to transfer entitlements in the funded part of public pensions (premium pensions) between spouses (OECD, 2021[16]). Transferring pension entitlements to partners with higher life expectancy, e.g. from men to women, inflates total expenditure, and to offset this, a charge of 6% of the transferred assets is levied by the Swedish pension system.

Korea and the United States provide spousal supplements, Japan credits periods towards the contribution-based basic pension when spouses are not employed, and Belgium applies higher accrual rates for couples in contributory pensions. Spousal supplements provide specific benefits for spouses who do not have their own pension or who have a very low one. Spousal supplements benefit spouses who have relied on their partners financially for whatever reason, and married couples are treated more favourably than informal couples or single persons. In Belgium, after a full career, the replacement rate increases from 60% to 75% of the higher-earning partner if this is more beneficial for the couple than applying the 60% rate to both spouses separately. In the United States, the spousal supplement is equal to 50% of the higher individual pension within the couple, and the lower pension is deducted from the spousal benefit. In Korea, a small flat-rate supplement is paid to the partner whose spouse does not receive their own pension. In Japan, some out-of-work spells of spouses of workers are credited towards the contribution-based basic pension, even though no contributions are paid. Finland abolished spousal benefits in 2001. Many OECD countries apply different rates to singles and couples for residence‑based basic pensions and targeted benefits to account for household economies of scale related e.g. to housing costs (Chapter 5).

Specific issues affecting the gender pension gap in asset-backed pensions

Copy link to Specific issues affecting the gender pension gap in asset-backed pensions

Beyond labour market factors, behavioural and cultural factors may affect individual decisions linked to retirement and retirement saving. Women frequently demonstrate higher risk aversion than men, which can translate into a preference for lower-risk investments and therefore lower returns on average for their retirement savings. This seems to be related to differences in attitudes towards risk taking and willingness to compete, as well as in financial education levels (OECD, 2021[16]; Buser, Ranehill and van Veldhuizen, 2021[57]; Charness and Gneezy, 2012[58]). Given that women already tend to hold conservative investments, they are less likely to switch to a riskier alternative investment option if the default already matches their risk aversion level. For example, in Italy and Latvia, the default investment option in asset-backed pension plans is a conservative investment strategy. While this curbs the risks, it also reduces the expected return that women could get on their savings over the entire accumulation phase. Furthermore, financial advisors may be subject to gender stereotypes and assume a greater risk aversion for women, reinforcing the already higher risk aversion of women compared to men (OECD, 2021[16]). Attitudes towards saving also differ between men and women as women may delay or avoid saving for retirement because they feel more vulnerable to short-term financial hardship, or they are more likely to prioritise current family members needs over their own old-age security (OECD, 2021[16]).

Lower levels of financial literacy may also lead women to engage less in retirement planning. On average, men have slightly higher levels of financial literacy than women (OECD, 2023[25]). Gender differences in financial knowledge tend to be significant in Estonia, Finland, Greece, Luxembourg and Sweden (OECD, 2023[59]). Lower levels of financial knowledge imply that women have lower knowledge than men of concepts like time value of money, simple and compound interest, and risk diversification that are crucial for making informed decisions about long-term savings and pensions.

Policy discussion

Copy link to Policy discussion

Gender pension gaps are large and represent an important topic for pension policy. Women receive pensions that are about one‑quarter lower than men’s on average across OECD countries. Moreover, older women face much higher poverty risks than older men in almost all OECD countries. While pensions cannot fully compensate for inequalities that build up during the working life, limiting the impact of these inequalities on pension differences between men and women is among the priorities facing policymakers in the pension area. Mitigating the transmission of labour market disparities into the gender pension gap is also consistent with supporting families with children as part of broader family policies objectives. Redistributive pension policies differ substantially across countries as they depend on individual tastes and social preferences towards, among others, old-age inequality, the relative value of paid work and care, the role of marriage in society and the importance of having children (Barr, 2019[60]).

Gender differences in lifetime earnings are the main driver of the gender pension gap as a large part of pension benefits is earnings-related. Still, not all lifetime earnings inequalities are transmitted into pensions and, in particular, the gender gap in lifetime earnings is significantly larger than the GPG. Gender differences in employment, hours worked and hourly wages make a similar contribution to the gender gap in lifetime earnings (about one‑third each) on average across OECD countries. Gender gaps in lifetime earnings have been declining across cohorts, mostly driven by higher female employment. Yet, disparities in labour market outcomes between men and women remain large and are unlikely to disappear in the foreseeable future. As a result, the most efficient measures to reduce the GPG over the long term need to focus on tackling persistent gender differences in employment, hours worked and wages. In particular, the unequal share of unpaid care between men and women as well as gender disparities in education pathways and the labour market have large implications. Policy priorities in these areas go beyond this report and are discussed in other OECD publications (OECD, 2023[25]; OECD, 2024[26]; OECD, 2025[61]).

Countries wanting to promote gender equality in the labour market and reduce the gender pension gap should eliminate earlier access to pensions for women. Earlier access to pensions by women is a legacy of the past inherited from the single‑breadwinner model. Having the same pension eligibility conditions for men and women help reduce gender gaps in career duration. By contrast, earlier eligibility ages to pensions for women in a few OECD countries results in lower pension entitlements, raising the GPG. Based on current legislation, Colombia, Costa Rica, Hungary, Israel, Poland and Türkiye will maintain differences in the normal retirement age between men and women. In Chile, men and women have access to public pensions from age 65, but only women can claim pensions from the mandatory funded scheme five years earlier. Furthermore, providing mothers with the possibility to retire at a lower age, as is the case in Czechia, Italy, the Slovak Republic and Slovenia, is difficult to justify.

Reducing minimum eligibility conditions to access pensions as much as possible would help lower the gender income disparities in old age. Such conditions include long contribution records, minimum earnings or minimum hours worked to access pensions. For example, Czechia requires 30 years of contributions to access pensions, Japan and Korea require 20 and 15 hours of work per week, respectively, to be covered by earnings-related pensions. Some countries only cover mandatorily those earnings above a certain threshold, which amounts to around 10% of the gross average wage in Austria, Germany, and the United Kingdom, and around 20% in Italy, Japan and Switzerland. Japan will eliminate this threshold in 2028. In Canada, Japan, Switzerland and the United Kingdom, minimum income thresholds constrain access to asset-backed occupational pension plans (OECD, 2021[16]). More generally, these conditions tend to penalise workers with short careers, low earnings and frequent part-time employment; as a result, they disproportionately affect women.

High levels of first-tier benefits strongly reduce pension inequalities and thereby the gender pension gap. First-tier pensions comprise programmes offering the first layer of social protection in old age, and for which past earnings are irrelevant in the calculation of retirement income. While these benefits are generally gender neutral, they benefit women more. When first-tier benefits play a large role relative to earnings-related pensions, this limits the transmission of earnings inequalities into pensions. However, it also provides less protection against the income drop upon retirement for many workers, thereby lowering the incentives to contribute. Similar trade‑offs apply to other choices when designing first-tier pensions: residency-based basic pensions are more effective than contribution-based instruments in reducing gender inequalities, as the latter are linked to individual labour-market histories. In Denmark, Iceland, Israel and New Zealand, non-contributory first-tier benefits pay more than 30% of the average wage, which is also the case for full contribution-based basic pensions in Belgium, Colombia, Luxembourg, Slovenia, Spain and Türkiye. The normative discussion about these trade‑offs in the design and levels of first-tier benefits should take into account their gender implications.

Care‑related pension credits are an effective instrument to cushion the shock of relatively short employment breaks, especially at low-income levels. Childcare‑related credits compensate for about one‑half of pension entitlements lost during a 5‑year childcare‑related break on average across OECD countries. Such credits can be also expanded to cover reduced hours needed to reconcile care and work, as for example in Germany, Portugal and Slovenia. However, they should limit the risk of permanently trapping those who have interrupted their careers in part-time jobs. The credited entitlement can be linked to the amount of the care‑related benefits, be it maternity, paternity or parental leave. Alternatively, it can be directly linked to pre‑break earnings up to some ceiling, economy-wide average earnings, the minimum wage or any other flat-rate amount. Among these choices, pre‑break earnings provide the strongest link between earnings and benefits while earnings ceiling or flat-rate amounts provide higher entitlements to low-income workers for a given total fiscal cost. The duration of the credited periods for childcare should not be excessively long to support a faster return to employment and limit the negative impact of the break on career progression, provided that childcare services are accessible. Subsidised credits for childcare may also apply to private pensions, but these subsidies should be capped or based on flat-rate amounts to mitigate old-age inequalities.

Pension entitlements granted to mothers irrespective of interrupting their career for childcare can compensate for reduced hours worked and lower wages, the so-called motherhood penalty. They can also complement measures that support families with children more broadly. These instruments benefit mothers without disincentivising work. For example, France and Germany grant some pension credits to mothers irrespective of whether they interrupted their careers. The direct link between pension entitlements and having children, rather than taking childcare‑related employment breaks, simplifies benefit calculation and may compensate for the motherhood penalty – related, for example, to reduced working hours and slower career progression. If such entitlements are linked to past individual earnings, they better compensate for individuals’ loss of earnings while flat-rate entitlements provide better protection to low earners.

Ignoring women’s higher longevity for pension benefit calculation avoids substantially increasing the gender pension gap further. It is also consistent with evenly pooling longevity risks across the whole population. While women live longer than men, by around three years after age 65 on average across OECD countries, mandatory public-pension benefits of women are not affected by this difference in any OECD country. The principle of ignoring gender longevity differences applies also to private pensions in the European Union, as opposed to other parts of the world. Even though it decreases the GPG, using unisex mortality tables for annuity calculations in defined contribution schemes discourages men from taking annuities if annuitisation is voluntary. This contributes to longevity risks remaining largely uninsured in voluntary pensions. Outside the EU, defined contribution schemes pay less every month to women than to men for the same amount of accumulated assets due to their longer expected retirement period. Furthermore, as for all groups with higher life expectancy at older ages, women tend to benefit more from generous pension indexation.35 The trade‑off about how much to frontload pensions and how much to index them over time for a given expenditure level should obviously take into account broader implications than those related to the GPG.

Survivor pensions substantially lower the gender pension gap and decrease old-age poverty of women in most OECD countries. Women benefit more than men from survivor pensions due to both their lower own entitlements and the fact that they often outlive their partners. However, apart from reducing the GPG, survivor pensions have pursued two main objectives. First, they have protected widows or widowers from poverty risks to offset sharp drops in disposable income to low absolute levels. This is less relevant now than in the past, as nowadays all OECD countries have instruments directly targeted at poverty alleviation. Second, more relevant today, they have contributed to insuring against the decrease in disposable income relative to the situation prevailing before the partner’s death, in the same way as old-age pensions help avoid a sharp drop in income when moving out of paid work upon retirement. This second objective remains valid despite the substantial reduction in employment differences between men and women. To support women’s longer careers, recipients should not be eligible for a permanent survivor pension before the retirement age (OECD, 2018[46]). Instead, at these younger ages a temporary benefit should be accessible following the partner’s death to help adapt to the new situation.36

Communication efforts should increase women’s awareness of the possibility and importance of splitting retirement entitlements upon divorce. Still, while splitting pension rights is fairly easy to implement in defined contribution and point systems or in defined benefit systems that are based on straightforward accrual rates, it is more complicated to do so in complex and fragmented pension systems as well as in schemes with loose links between contributions and pension entitlements. Splitting pension rights, including in public schemes, should replace survivor pensions for separated couples and it may be mandated in divorce settlements, in line with how other assets are split. For separated couples, the death of the former partner does not generally affect the survivor’s income – unless alimony was granted – so survivor pensions are not needed to smooth income. For ongoing partnerships, pension splitting cannot replace survivor pensions to smooth income upon the partner’s death. Splitting pension rights within couples enhances gender equality and is consistent with sharing resources broadly within partnerships, although some countries favour the individual treatment of partners (OECD, 2018[46]).

Policymakers can take actions to reduce the gender gap in asset-backed pension arrangements. While asset-backed pension arrangements should aim to be gender neutral, reducing the gap in assets and benefits between men and women requires adjusting pension plan rules as well as additional communication effort (OECD, 2021[16]). Increasing the availability of pension plans in industries predominantly employing women and relaxing eligibility requirements to join a plan would improve women’s access to these arrangements. To increase the availability of retirement savings arrangements in industries predominantly employing women, several options exist: mandating occupational pension plans, providing incentives for employers to establish occupational arrangements for their employees, or increasing the availability of personal arrangements. Once women have access to a plan, both men and women could be encouraged to join one and contribute to it by using nudges (e.g. automatic enrolment), providing financial incentives to participate, as well as using tailored educational workshops and communication that convey the importance of having their own savings for retirement and the importance of regular contributions.

References

[62] Afonso, A. and C. Blanco Arana (2024), “The Persistence of Gender Pay and Employment Gaps in European Countries”, SSRN Electronic Journal, https://doi.org/10.2139/ssrn.4991813.

[74] Age UK (2018), For love and money: Women’s pensions, expenditure and decision-making in, https://www.ageuk.org.uk/siteassets/documents/reports-and-publications/reports-and-briefings/money-matters/rb_aug18_women_retirement_expenditure_and_pensions.pdf.

[38] Altonji, J. et al. (2025), Decomposing Trends in the Gender Gap for Highly Educated Workers, National Bureau of Economic Research, Cambridge, MA, https://doi.org/10.3386/w34133.

[12] Andersson, J. (2023), Gender-equal pensions in the Nordics, Nordic Council of Ministers, http://pub.norden.org/temanord2023-506.

[11] Barslund, M. et al. (2021), The future of Gender Pension Gaps, DG JUST, https://www.migape.eu/pubs/MIGAPE%20The%20future%20of%20Gender%20Pension%20Gaps.pdf.

[42] Bertrand, M. (2020), “Gender in the Twenty-First Century”, AEA Papers and Proceedings, Vol. 110, pp. 1-24, https://doi.org/10.1257/pandp.20201126.

[31] Bertrand, M., C. Goldin and L. Katz (2010), “Dynamics of the Gender Gap for Young Professionals in the Financial and Corporate Sectors”, American Economic Journal: Applied Economics, Vol. 2/3, pp. 228-255, https://doi.org/10.1257/app.2.3.228.

[53] Bonnet, C. and B. Rapoport (2019), “Is There a Child Penalty in Pensions? The Role of Caregiver Credits in the French Retirement System”, European Journal of Population, Vol. 36/1, pp. 27-52, https://doi.org/10.1007/s10680-019-09517-0.

[21] Boulhol, H., M. Lis and M. Queisser (2023), Trends in Pension Reforms in OECD Countries, Routledge, London, https://doi.org/10.4324/9781003150398.

[45] Brown, A. and A. Fraikin (2022), “The old-age pension household replacement rate in Belgium”, The Journal of the Economics of Ageing, Vol. 23, p. 100402, https://doi.org/10.1016/j.jeoa.2022.100402.

[57] Buser, T., E. Ranehill and R. van Veldhuizen (2021), “Gender differences in willingness to compete: The role of public observability”, Journal of Economic Psychology, Vol. 83, p. 102366, https://doi.org/10.1016/j.joep.2021.102366.

[72] Card, D. et al. (2025), The gender gap in career trajectories: do firms matter?, NBER, http://www.nber.org/papers/w33730.

[4] CASEN (2022), Encuesta de caracterización socioeconómica nacional, Ministerio de Desarrollo Social y Familia., https://observatorio.ministeriodesarrollosocial.gob.cl/encuesta-casen-2024.

[28] CGEPS (2023), Intersectionality at work: building a baseline on compounded gender inequality in the Victorian public sector, Commission for Gender Equality in the Public Sector, https://www.genderequalitycommission.vic.gov.au/sites/default/files/2023-10/Intersectionality-At-Work-Report.pdf.

[58] Charness, G. and U. Gneezy (2012), “Strong Evidence for Gender Differences in Risk Taking”, Journal of Economic Behavior & Organization, Vol. 83/1, pp. 50-58, https://doi.org/10.1016/j.jebo.2011.06.007.

[51] Chen, A. and E. Vigna (2017), “A unisex stochastic mortality model to comply with EU Gender Directive”, Insurance: Mathematics and Economics, Vol. 73, pp. 124-136, https://doi.org/10.1016/j.insmatheco.2017.01.007.

[32] Ciminelli, G., C. Schwellnus and B. Stadler (2021), “Sticky floors or glass ceilings? The role of human capital, working time flexibility and discrimination in the gender wage gap”, OECD Economics Department Working Papers, No. 1668, OECD Publishing, Paris, https://doi.org/10.1787/02ef3235-en.

[76] COR (2025), Évolutions et perspectives des retraites en France, Conseil d’orientation des retraites, https://www.cor-retraites.fr/sites/default/files/2025-06/RA_2025_def_publi.pdf.

[13] COR (2024), Évolutions et perspectives des retraites en France, https://www.cor-retraites.fr/sites/default/files/2024-12/RA_2024_finale_impression.pdf.

[47] Council of the European Union (2004), Directive 2004/113/EC implementing the principle of equal treatment between men and women in the access to and supply of goods and services, https://eur-lex.europa.eu/eli/dir/2004/113/oj/eng.

[9] Cribb, J., H. Karjalainen and L. O’Brien (2023), The gender gap in pension saving, lnstitute for Fiscal Studies, https://ifs.org.uk/publications/gender-gap-pension-saving.

[64] Cubas, G., C. Juhn and P. Silos (2019), Coordinated Work Schedules and the Gender Wage Gap, National Bureau of Economic Research, Cambridge, MA, https://doi.org/10.3386/w26548.

[33] Cukrowska-Torzewska, E. and A. Matysiak (2020), “The motherhood wage penalty: A meta-analysis”, Social Science Research, Vol. 88-89, p. 102416, https://doi.org/10.1016/j.ssresearch.2020.102416.

[8] Danchev, S. et al. (2024), “Equally poorer: inequality and the Greek debt crisis”, Fiscal Studies, Vol. 45/3, pp. 359-375, https://doi.org/10.1111/1475-5890.12384.

[48] Di Lego, V., P. Di Giulio and M. Luy (2020), “Gender Differences in Healthy and Unhealthy Life Expectancy”, in International Handbooks of Population, International Handbook of Health Expectancies, Springer International Publishing, Cham, https://doi.org/10.1007/978-3-030-37668-0_11.

[10] DREES (2024), Les retraités et les retraites, https://drees.solidarites-sante.gouv.fr/sites/default/files/2024-10/RR24.pdf.

[55] Euro carers (2025), Towards carere-friendly societies, https://eurocarers.org/country-profiles.

[2] Eurostat (2025), Gender pension gap by age group, https://doi.org/10.2908/ILC_PNP13.

[22] Eurostat (2024), Duration of working life averaged 36.9 years in 2023, https://ec.europa.eu/eurostat/web/products-eurostat-news/w/ddn-20240725-1.

[75] Finsider (2025), Veľký prehľad: Kedy pôjdete po novom do dôchodku? Na jeho získanie treba odpracovať viac, https://www.finsider.sk/servis/novy-odchod-do-dochodku/.

[1] Frey, V., L. Alajääskö and J. Thomas (2024), “Gendered perceptions of social protection across OECD countries”, OECD Social, Employment and Migration Working Papers, No. 311, OECD Publishing, Paris, https://doi.org/10.1787/f3e002c2-en.

[63] Goldin, C. (2014), “A Grand Gender Convergence: Its Last Chapter”, American Economic Review, Vol. 104/4, pp. 1091-1119, https://doi.org/10.1257/aer.104.4.1091.

[60] Holzmann, R. et al. (eds.) (2019), Chapter 17. Gender and Family: Conceptual Overview, Washington, DC: World Bank, https://doi.org/10.1596/978-1-4648-1455-6.

[67] Kiessling, L. et al. (2024), “Gender differences in wage expectations and negotiation”, Labour Economics, Vol. 87, p. 102505, https://doi.org/10.1016/j.labeco.2024.102505.

[71] Konle-Seidl, R. (2021), “Precarious but popular? The German mini-job scheme in comparative research on work and welfare”, Journal of International and Comparative Social Policy, Vol. 37/3, pp. 293-306, https://doi.org/10.1017/ics.2021.11.

[69] Langner, L. (2015), “Within-couple specialisation in paid work: A long-term pattern? A dual trajectory approach to linking lives”, Advances in Life Course Research, Vol. 24, pp. 47-65, https://doi.org/10.1016/j.alcr.2015.02.002.

[50] Letablier, M. et al. (2009), The costs of raising children and the effectiveness of policies to support parenthood in European countries: a Literature Review, Eurpean Commission, https://www.ined.fr/fichier/s_rubrique/19548/158bis.fr.pdf.

[39] Leythienne, D. and M. Pérez-Julián (2022), Gender pay gaps in the European Union, European Union, https://ec.europa.eu/eurostat/en/web/products-statistical-working-papers/-/ks-tc-22-002.

[6] LIS (2025), Luxembourg Income Study Database, https://www.lisdatacenter.org/our-data/lis-database/.

[41] Maloney, M. and D. Neumark (2025), Does the Gender Wage Gap Actually Reflect Taste Discrimination Against Women?, National Bureau of Economic Research, Cambridge, MA, https://doi.org/10.3386/w33405.

[27] Moghadam, H., P. Puhani and J. Tyrowicz (2024), “Pension reforms and couples’ labour supply decisions”, Labour Economics, Vol. 91, p. 102627, https://doi.org/10.1016/j.labeco.2024.102627.

[52] Natali, D. et al. (2024), Study supporting the monitoring of care credits in occupational pension schemes., Publications Office of the European Union.

[23] OECD (2025), Employment and unemployment by five-year age group and sex - levels, http://data-explorer.oecd.org/s/1a3.

[77] OECD (2025), Employment Outlook, OECD, https://doi.org/10.1787/194a947b-en.

[61] OECD (2025), Gender Equality in a Changing World: Taking Stock and Moving Forward, Gender Equality at Work, OECD Publishing, Paris, https://doi.org/10.1787/e808086f-en.

[30] OECD (2025), Gender wage gap, OECD, https://data-explorer.oecd.org/s/31i.

[15] OECD (2025), Income Distribution Database, https://www.oecd.org/en/data/datasets/income-and-wealth-distribution-database.html.

[24] OECD (2025), OECD Employment Outlook 2025: Can We Get Through the Demographic Crunch?, OECD Publishing, Paris, https://doi.org/10.1787/194a947b-en.

[56] OECD (2025), Social expenditure aggregates, https://data-viewer.oecd.org?chartId=da4a8a64-f57e-4ac5-b249-e48dbff6113f.

[68] OECD (2024), Enrolment rate by age, https://data-explorer.oecd.org/vis?df[ds]=dsDisseminateFinalDMZ&df[id]=DSD_EAG_UOE_NON_FIN_STUD%40DF_UOE_NF_ENRL_RATE&df[ag]=OECD.EDU.IMEP&df[vs]=1.0&dq=._T......A......F%2BM.Y_LT3%2BY3T5%2BY6T14%2BY15T19%2BY20T24%2BY25T29&pd=2022%2C2022&to[TIME_PERIOD]=f.

[26] OECD (2024), Megatrends and the Future of Social Protection, OECD Publishing, Paris, https://doi.org/10.1787/6c9202e8-en.

[37] OECD (2023), Beyond Applause? Improving Working Conditions in Long-Term Care, OECD Publishing, Paris, https://doi.org/10.1787/27d33ab3-en.

[25] OECD (2023), Joining Forces for Gender Equality: What is Holding us Back?, OECD Publishing, Paris, https://doi.org/10.1787/67d48024-en.

[59] OECD (2023), “OECD/INFE 2023 International Survey of Adult Financial Literacy”, OECD Business and Finance Policy Papers, No. 39, OECD Publishing, Paris, https://doi.org/10.1787/56003a32-en.

[5] OECD (2023), Pensions at a Glance 2023: OECD and G20 Indicators, OECD Publishing, Paris, https://doi.org/10.1787/678055dd-en.

[7] OECD (2022), OECD Reviews of Pension Systems: Slovenia, OECD Reviews of Pension Systems, OECD Publishing, Paris, https://doi.org/10.1787/f629a09a-en.

[16] OECD (2021), Towards Improved Retirement Savings Outcomes for Women, OECD Publishing, Paris, https://doi.org/10.1787/f7b48808-en.

[29] OECD (2019), Part-time and Partly Equal: Gender and Work in the Netherlands, Gender Equality at Work, OECD Publishing, Paris, https://doi.org/10.1787/204235cf-en.

[46] OECD (2018), Chapter 7. Are survivor pensions still needed?, OECD Publishing, Paris, https://doi.org/10.1787/pens_outlook-2018-en.

[43] OECD (2017), Preventing Ageing Unequally, OECD Publishing, Paris, https://doi.org/10.1787/9789264279087-en.

[40] Palladino, M. et al. (2024), “The role of bargaining and discrimination in the gender wage gap in France: A cross-country perspective”, OECD Social, Employment and Migration Working Papers, No. 315, OECD Publishing, Paris, https://doi.org/10.1787/1fd68687-en.

[20] Pay Equity Office (2024), Understanding the Gender Pension Gap in Canada, https://payequity.gov.on.ca/wp-content/uploads/2024/06/PEO_Understanding-the-Gender-Pension-Gap-in-Canada-EN-1.pdf.

[66] Pelley, E. and M. Carnes (2020), “When a Specialty Becomes “Women’s Work”: Trends in and Implications of Specialty Gender Segregation in Medicine”, Academic Medicine, Vol. 95/10, pp. 1499-1506, https://doi.org/10.1097/acm.0000000000003555.

[35] Pora, P. and L. Wilner (2019), Les trajectoires professionnelles des femmes les moins bien rémunérées sont les plus affectées par l’arrivée d’un enfant, https://www.insee.fr/fr/statistiques/4226475.

[44] Schmauk, S. and L. Kridahl (2024), “Who receives most? Gendered consequences of divorce on public pension income in West Germany and Sweden”, Ageing and Society, pp. 1-24, https://doi.org/10.1017/s0144686x23000703.

[14] Scroope, C. (2017), Chilean culture, Cultural Atlas, https://culturalatlas.sbs.com.au/chilean-culture/chilean-culture-family.

[70] Sin, I., S. Stillman and R. Fabling (2022), “What Drives the Gender Wage Gap? Examining the Roles of Sorting, Productivity Differences, Bargaining, and Discrimination”, The Review of Economics and Statistics, Vol. 104/4, pp. 636-651, https://doi.org/10.1162/rest_a_01000.

[18] SSA (2024), Annual statistical supplement, 2024.

[3] Statistics Canada (2024), Income of individuals by age group, sex and income source, Canada, provinces and selected census metropolitan areas, https://doi.org/10.25318/1110023901-eng.

[36] Strauss, A. and D. Borrett (2025), UK gender pay gap begins at graduation as women are quickly out-earned, Financial Times, https://www.ft.com/content/cb0e209c-0b08-4749-9d49-0ecfabe5a96e.

[17] UN (2024), World Population Prospects 2024: Dataset., https://population.un.org/wpp/.

[73] Weinkopf, C. (2014), Women’s Employment in Germany. Robust in Crisis but Vulnerable in Job Quality, Revue de l’OFCE, Presses de Sciences-Po.

[54] WGEA (2025), Australian employers paying up for mums and dads on parental leave, https://www.wgea.gov.au/newsroom/parental-leave-scorecard.

[49] WHO (2025), Healthy life expectancy (HALE), WHO, https://www.who.int/data/gho/data/indicators/indicator-details/GHO/gho-ghe-hale-healthy-life-expectancy-at-birth.

[34] Wilde, E., L. Batchelder and D. Ellwood (2010), The Mommy Track Divides: The Impact of Childbearing on Wages of Women of Differing Skill Levels, National Bureau of Economic Research, Cambridge, MA, https://doi.org/10.3386/w16582.

[65] Wiswall, M. and B. Zafar (2017), “Preference for the Workplace, Investment in Human Capital, and Gender*”, The Quarterly Journal of Economics, Vol. 133/1, pp. 457-507, https://doi.org/10.1093/qje/qjx035.

[19] ZUS (2024), Struktura wysokości świadczeń wypłacanych przez ZUS po waloryzacji w marcu 2024 roku.

Annex 2.A. Benefits of older people with no or little contributory pension entitlements

Copy link to Annex 2.A. Benefits of older people with no or little contributory pension entitlements

A person born in 2002 who will not have worked at all for his or her entire life will receive old-age benefits equivalent to 16% of the gross average wage on average across OECD countries, ranging from around 5% in Czechia, Hungary, Korea, Lithuania and Türkiye to over 30% in Denmark and New Zealand (Annex Figure 2.A.1). Workers with a full career from age 22 in 2024 and earning 25% of the average wage (as an order of magnitude this would be close to working part-time at the minimum wage in many countries) can expect old-age benefits totalling 24% of the average wage on average across OECD countries. This is half more than the 16% of the average wage provided to individuals who have never worked. Full-career workers with such very low earnings can expect benefits that are 10 p.p. higher than those of individuals who have never worked in several countries including Chile, Czechia, Costa Rica, Estonia, Hungary, Greece, Latvia, Korea, the Slovak Republic and Türkiye. In Colombia and Mexico, full minimum pensions are projected to equal 52% and 35% of the average wage, respectively, while safety-net benefits would remain low. No mandatory contributory pensions exist in New Zealand, while in Austria, France, Germany, Ireland, the Netherlands, Poland, Slovenia, Spain, and Switzerland mandatory contributory pensions for full-career workers earning 25% of the average wage will be no more than three points higher than non-contributory benefits.

Working for 20 years at 50% of the average wage provides benefits that are similar to those of working a full career at 25% of the average wage, although there are some exceptions (Annex Figure 2.A.1). In Denmark, Germany, Latvia, Luxembourg, the Slovak Republic and the United Kingdom some components of pensions are prorated for career length and thereby longer careers result in higher pensions than shorter careers with similar earnings. Conversely, in Mexico, given the recent reform, the low earner with only 20 years of contributions will receive a benefit equivalent to 100% of their last earnings, twice that of the full career at 25% of average earnings. In the Netherlands, only earnings above a certain threshold accrue occupational pension rights, while low earnings accrue no additional pension entitlements beyond the basic pension. In Hungary, workers earning less than the full-time monthly minimum wage have their accrual rates prorated relative to the minimum wage. This results in a double penalty as the pension is reduced by both the lower reference wage and the lower accrual rate. Additionally, the accrual rate is at a substantially higher rate for the first 15 years of career than for following years, benefiting more workers with shorter careers.

Annex Figure 2.A.1. Pensions mitigate old-age inequalities for low earners
Copy link to Annex Figure 2.A.1. Pensions mitigate old-age inequalities for low earners

Future pensions as percentage of the average wage for: 1) an individual who never works, 2) a part-time worker earning 25% of the full-time average wage throughout the whole career, and 3) a worker earning 50% of the average wage and working 20 years before retiring

Notes

Copy link to Notes

← 1. In Colombia and Mexico, however, the GPG increased by 14 and 7 percentage points along with a strong increase of pension coverage among women.

← 2. Based on data provided by countries for Belgium, Costa Rica, Germany, New Zealand, Norway and Switzerland, and on data included OECD (2021[16]) for Ireland.

← 3. Benefits from contribution-based basic pensions are non-earnings-related but with some link to past employment as they are contributory.

← 4. Based on data provided by countries for Belgium, Costa Rica, Germany, New Zealand and Switzerland, and on data included OECD (2021[16]) for Ireland. The data on gender gaps in voluntary pensions come from administrative sources and might not be fully consistent with the gender pension gaps reported from surveys.

← 5. From 2026, Ireland aims to expand the coverage of voluntary pensions through auto‑enrolment, which would improve pension prospects of many men and women, with an undetermined impact on the gender pension gap.

← 6. The share of private pensions in total pensions receipt increased from 48% to 58% between 1976 and 2021. During this period, the gender pension gap increased in the CPP/QPP from 8% to 16% and the total gender pension gap increased slightly, from 15% to 17%. However, this surge in the CPP/QPP happened before 1990s, and since mid‑1990s, the gender pension gap in both voluntary and mandatory earnings-related schemes has been steadily declining, by one‑third in total. The expansion of private pensions by itself has increased the GGP by 3 p.p., but it was offset by a decline in the gender pension gap in private pensions from 30% to 25%.

← 7. The gender difference in life expectancy at 65 varies from around 2 years in Iceland and the United Kingdom to about 5 years in Japan, Korea, and Lithuania.

← 8. This gender difference in the average labour market exit age is very high, at 6 years in Colombia and 5 years in Costa Rica, while in Estonia, France, Japan and Spain women leave the labour market at an older age than men on average, by around half a year, as well as in Korea by 2 years. The residual factor results from measuring life expectancy at different ages for men and for women.

← 9. Across countries, the average labour market exit age is closely related to but differs from the average age at which pensions start to be received. Starting to receive old-age pensions is only one way of exiting the labour market, as, on the one hand, workers may stop working and live on their savings, partner’s income, safety-net benefits or disability pensions, while, on the other hand, workers may combine receiving old-age pensions and working.

← 10. Men enter the labour market at 21.1 years, about half a year earlier than women on average across OECD countries. Average labour market entry ages are calculated using a similar methodology as the one used to calculate average labour market exit ages, which are reported in Chapter 7. Men enter the labour market by more than 1.5 years earlier than women in Czechia, Estonia, Hungary, Mexico, Poland and the Slovak Republic, while they enter around half a year later than women in Ireland, Israel, and Switzerland, and even 1.7 years later in Korea. In Israel, Korea and Switzerland, military conscription delays labour market entry particularly for men, while in Czechia, Estonia, Poland and the Slovak Republic, the enrolment rates of women in tertiary education relative to men are exceptionally high (OECD, 2024[68]): the mandatory military conscription for men lasts 20 months in Korea, 6 months in Switzerland and 32 months in Israel. In Israel, military conscription is also mandatory for women and lasts 24 months, i.e. 8 months less than for men.

← 11. The gender gap in working hours in 2023 is not smaller than 30 years ago in Germany, Greece, Korea and Spain. In Germany, between 1993 and 2023, the gender gap in working hours first increased and then decreased, reflecting broader changes in the labour market. Between 1993 and 2008, average hours worked declined more for women than for men, as the increase in women’s employment was primarily driven by part-time work and mini-jobs (Weinkopf, 2014[73]). During this period, in 2003, the government introduced so-called mini-job contracts, which have been exempted from mandatory social security contributions for monthly earnings up to a ceiling, which is equal to EUR 556 in 2025. Between 2008 and 2023, the men’s working hours decreased more strongly than women’s. The introduction of the statutory hourly minimum wage in 2015 might have contributed to the reduction of hours worked by men (Konle-Seidl, 2021[71]).

← 12. The gender gap in hourly wages is one component of the gender gap in lifetime earning needed for the method used in this chapter (see next sub-section for the exact formula for the break-down). It differs from the gender wage gap measured on a monthly basis for full-time workers at median wage, which is often reported, e.g. in OECD (2023[25]).

← 13. Although general trends across countries are clear, country-specific trends should be interpreted with caution as the sectors covered for the measurement of wages can vary over time.

← 14. They estimate the motherhood penalty, defined as the difference between wages of mothers compared to childless women with similar characteristics, to be around 3.7% on average across all available studies. This is also consistent with the gender wage gap widening with age and reducing the financial incentive for women to stay in employment (OECD, 2023[25]; OECD, 2025[77]).

← 15. The strong gender segregation of women into lower-paying jobs is also observed within occupations, e.g. within medical professions (Pelley and Carnes, 2020[66]). In addition, the sorting of women into slow-growth firms is found to account for one‑fifth of the gender wage growth gap in Italy, and women who have a child within 5 years of entering work experience particularly slow wage growth (Card et al., 2025[72]).

← 16. First, part-time work and other flexible work arrangements may slow human capital accumulation, contributing to the gender gap in hourly wages (Wiswall and Zafar, 2017[65]). Part-time employment also limits promotion opportunities (OECD, 2023[25]). Afonso and Blanco Aran (2024[62]) estimate that higher part-time employment by women significantly increases the gender gap in hourly wages based on a quantitative analysis covering a number of European countries. Second, firms tend to provide higher hourly wages to individuals who work long hours and work during unusual time schedules, who are more often men (Goldin, 2014[63]; Cubas, Juhn and Silos, 2019[64]).

← 17. A significant part of the gender wage gap in New Zealand is related to women being less willing to bargain or less successful at bargaining to capture firm-specific rents (Sin, Stillman and Fabling, 2022[70]). Furthermore, substantial gender gaps in wage expectations exist even before entering the labour market indicating the significant influence of differences in perceptions about own abilities and in bargaining approaches between men and women (Kiessling et al., 2024[67]).

← 18. Ciminelli, Schwellnus and Stadler (2021[32]) find that, on average, “sticky floors” – i.e. persistent disadvantages over women’s working lives from labour market entry to retirement – related to individual preferences for some occupations, social norms, gender stereotyping and discrimination account for 40% of the gender wage gap, while the “glass ceiling” – i.e. limited career progression – related to e.g. the motherhood penalty and preferences for working less hours in more flexible environment accounts for around 60%. The importance of the “glass ceiling” is especially large in most Northern and Western European countries, while “sticky floors” explain the major part of the gap in most Central and Eastern European countries.

← 19. Recent OECD estimates confirm this order of magnitude.

← 20. Hungary offers women only an option to retire at any age after a 40‑year career, while other conditions have been equalised between men and women following measures taken in 1997. In Türkiye, it will be eliminated for those starting their careers in 2028. In Austria, the initial five‑year gender gap in retirement ages is being eliminated between 2024 and 2033, following legislation introduced in the 1990s. In Lithuania, retirement ages are converging for men and women between 1995 and 2026. In Switzerland, the three‑year gender gap in statutory retirement ages was reduced in 2001 and will be eliminated in 2028.

← 21. In Italy, women can access early retirement after a one‑year shorter career than in the case of men: women with disabilities, providing care or being dismissed can retire from age 61 with 35 years of contributions as of 2025, subject to an age-specific benefit reduction. Before 2024, all women could use this pathway.

← 22. The pension eligibility conditions were equalised between men and women in Belgium between 1997 and 2009, in Czechia between 1995 and 2011, in Portugal, between 1994 and 2000 and in the Slovak Republic between 2004 and 2014 (Finsider, 2025[75]). In Germany, women’s retirement age was lower than men’s between 1957 and 2009. In the Netherlands, the equal treatment of men and women both in terms or retirement ages and benefit calculation were set in 1990.

← 23. Hence the 6% gender difference in pensions in Colombia is only due to pensions in payment being projected to increase less than wages as the initial replacement rate for men and women is the same at the point of retirement. For higher earners, the gender different is larger as part of the pension in Colombia will come from the mandatory FDC scheme, which adjusts benefits for both the lower retirement age of women and their higher longevity.

← 24. In 2011, the European Court of Justice also ruled that pension contributions and fees must not differ between men and women.

← 25. In France, motherhood adds additional years to the contributory record but does not reduce the minimum retirement age of 64. As the full pension will require 43 years of contribution record, with labour market entry at age 22, a mother with a full career will be able to retire at 64 without penalty, while a childless woman will not access a full pension before age 65.

← 26. In Czechia, a woman can retire one year earlier when having one child, two years earlier with two children, three years earlier with three or four children, and four years earlier with five or more children. In the Slovak Republic, the retirement age for women who raised children is lowered by 6 months for the first three children. If the mother cannot benefit from this early-retirement possibility, the right is transferred to the father. In Hungary, only women are eligible to retire without any age condition after 40 years of contributions. In Italy, the early retirement for women can be reduced by one year for each of their first two children, and they can also reduce the statutory retirement age by four months for each child, up to 12 months. In Slovenia, the retirement age can be reduced by up to four years for mothers, depending on the number of children. Alternatively, mothers can choose to increase their benefits. Fathers can also benefit, with the retirement age being reduced by up to two years.

← 27. Given the assumed entry age of 22 years, Table 2.1 does not includes neither Hungary and Italy nor the early-retirement option for mothers in Slovenia, which conditions are given in the text.

← 28. The transformation coefficient for mothers is more favourable because it is based on the actual retirement age plus one year for a mother of one or two children, or plus two years for a mother of three of more children.

← 29. These figures are based on rough simulations based on the OECD pension model assumptions and OECD-average mortality rates for men and women. For a stronger real-wage growth of 3% instead of 1.25% as assumed in the OECD pension model, the gender pension gap would be 2.8% higher with price indexation compared with wage indexation.

← 30. Furthermore, higher accrual rates in the early years of a career reduce the impact of shorter careers in Hungary, Slovenia and Spain. The opposite is true in Greece and Luxembourg, where the accrual rate increases with tenure. Additionally, career breaks at the beginning of a career have lower impact on pension benefits than those occurring at older ages when past earnings are uprated with less than the average wage growth. This occurs in Belgium, France, Portugal and Spain in defined benefit schemes as well as in Italy and Poland in NDC schemes. The opposite is true for funded DC schemes that are assumed to provide higher rates of return than wage-growth rates. In Spain, multiple mechanisms affect the transmission of employment breaks: i) working beyond 37 years does not lead to higher accruals; ii) the accrual rate is higher for the first years of work; iii) the reference wage will be based on only the best 27 out of last 29 years; iv) and, conversely, missing periods are imputed using the minimum pension base when calculating the reference wage.

← 31. Many OECD countries credit time spent caring for very young children (usually up to 3 or 4 years-old) as insured periods and consider it as paid employment. However, once children are aged 6 years or older any credit given for this extended period is usually only to determine eligibility for early retirement and the minimum pension, and not to raise benefits.

← 32. Assuming labour market entry at age 22, given the average future normal retirement ages of 66 years across countries, the average length of a full career will be 44 years. A five‑year break thus shortens it by 11%.

← 33. In Spain for example, five years of cohabitation are required. Around half of OECD countries provide survivor benefits to civil unions, and Canada, Hungary, Japan, Korea, Mexico, Portugal, Slovenia and Spain grant survivor pensions to cohabitating couples that meet additional conditions. Some OECD countries require a minimum marriage length to grant survivor benefits, ranging from 6 months to 10 years. In Estonia, the divorced spouse can receive the benefit upon reaching the statutory retirement age within three years of the divorce, provided that the marriage lasted for at least 25 years.

← 34. In Denmark, pension splitting of occupational pensions following divorce is only possible if specified in a prenuptial agreement.

← 35. In a budget-neutral way, generous pension indexation is offset by lower initial pensions when retiring, which penalises people with low life expectancy.

← 36. Consistent with the view that survivor pensions perpetuate stereotypical secondary role of women in the labour market, Norway and Sweden have eliminated survivor pensions and thereby they do not provide benefits to address the drop of income following partners’ death.

Featured topics

Agriculture and fisheries

Climate change

Development

Digital

Economy

Education and skills

Employment

Environment

Finance and investment

Governance

Health

Industry, business and entrepreneurship

Regional, rural and urban development

Science, technology and innovation

Society

Taxation

Trade

Energy

Nuclear energy

Transport

Featured topics

Agriculture and fisheries

Climate change

Development

Digital

Economy

Education and skills

Employment

Environment

Finance and investment

Governance

Health

Industry, business and entrepreneurship

Regional, rural and urban development

Science, technology and innovation

Society

Taxation

Trade

Energy

Nuclear energy

Transport

Countries A - C

Countries D - I

Countries J - M

Countries N - R

Countries S - T

Countries U - Z

Regional and global engagement

Countries

Countries A - C

Countries D - I

Countries J - M

Countries N - R

Countries S - T

Countries U - Z

Regional and global engagement

Publications

Publications

Featured publications

Data

Data

Featured data

News & events

News & events

Featured events

About OECD

About

Engage with us

Work with us

Featured topics

Agriculture and fisheries

Climate change

Development

Digital

Economy

Education and skills

Employment

Environment

Finance and investment