Andrea Bassanini
Emily Farchy
Sebastian Königs
António Melo
Javier Terrero Dávila
OECD Employment Outlook 2025
Can We Get Through the Demographic Crunch?
Abstract
Opposing trends in longevity and fertility imply that the OECD population is becoming older. Without further policy action, the retirement of large cohorts will shrink the pool of productive workers, while the dependent population will expand. This chapter assesses the projected impact of these trends on GDP per capita growth to 2060 and puts them, as well as their implications for public finances, into a context of intergenerational disparities – documenting the diverging income and wealth trajectories across the generations and the implications for poverty among different cohorts. Possible complementary avenues for offsetting the negative growth effect of demographic change are considered and their potential quantified based on alternative simulation scenarios.
Key findings
Opposite trends in fertility and life expectancy, and the progressive exit of baby boomers from the labour force, have already increased OECD countries old-age dependency ratio – defined as the ratio of the seniors (aged 65 years and above) to the working-age population – and will increasingly do so in the coming decades. In turn, without further policy action and changes in behaviours (e.g. people living longer may be able and willing to stay in the labour force for longer), this will drag down significantly economic growth and the capacity of OECD countries to improve their living standards.
This chapter sets the scene for the thematic part of this OECD Employment Outlook 2025, which is devoted to ageing and the labour market. It presents projection scenarios quantifying the impact of demographic change on GDP per capita growth. It also discusses current trends in intergenerational inequalities in incomes and wealth and their implications in terms of fairness of alternative policy solutions.
The key findings are as follows:
In the coming years, the size of the working-age population (aged 20‑64 years) will decline in a large number of OECD countries, while the old-age dependency ratio will continue to soar. By 2060, the working-age population will have declined by 8% in the OECD area, and by more than 30% in more than a quarter of OECD countries. The OECD old-age dependency ratio increased from 19% in 1980 to 31% in 2023 and is projected to rise further to 52% by 2060.
Because of population ageing, the employment-to-population ratio – that is, the percentage share of employed persons in the total population – is projected to decrease by 1.9 percentage points by 2060 in the OECD area. In Spain and the Slovak Republic, the slump is expected to be of 10 percentage points.
Assuming constant growth rate of labour productivity (GDP per person employed) and labour market entry and exit rate for the various groups, the projected contraction of the share of employed persons in total population implies that GDP per capita growth in the OECD area will be reduced by about 40%, falling from 1.0% per year in the 2010s to 0.6% per year on average over the period 2024‑60. This corresponds to 14% of foregone GDP per capita by 2060. In the absence of policy action, almost all OECD countries will see their GDP per capita growth declining.
While there are large differences in fertility rates across OECD countries and policies can contribute to contain their decline if not reverse it, changes in fertility trends can do little to counteract population ageing especially in the short term. Providing family policies that help the reconciliation of work and family life but also reduce the costs of raising children, especially housing costs, are key factors. However, because of changes in preferences for children, it is unlikely that such policies will enable countries to revert significantly trend decline in fertility rates. And in any event the growth dividend of higher fertility would materialise only in several decades.
Labour productivity growth has been on a declining trend for many decades. Developments in artificial intelligence (AI) and automation have made some scholars optimistic about the possibility of significantly reviving it but it is unclear whether even such a boost will be enough alone to bring productivity growth to the levels needed to compensate for the decline of labour input. While appropriate structural reforms to improve productivity can unambiguously be part of the solution, mobilising untapped labour resources will be key to maintain GDP per capita growth.
Migration can contribute to lessen the challenge demographic ageing poses to economic growth, and is already doing so. However, its potential does not appear a game changer unless net migration rates increase well above historical values. By increasing net migration rates to the 75th percentile of the cross-country distribution in 2021‑24, the median OECD country could improve its GDP per capita growth by 0.13 percentage points with respect to a hypothetical no migration scenario in which net migration flows are set to zero.
Closing the employment gap between men and women of all ages would deliver significant growth dividends: keeping all other assumptions as in the baseline scenario, it could increase annual OECD GDP per capita growth by 0.2 percentage points with respect to the baseline scenario, and by up to 0.6 percentage points in countries currently characterised by low female labour force participation. More than one‑third of these potential gains, however, would come from closing the gender gap for older workers (aged 55 years or more). Additional gains may be achieved by closing the gender gap in hours worked. Yet, care must be taken to ensure that all this goes hand in hand with closing the gender gap in unpaid work.
Mobilising further labour market participation and employment of older people in good health has a significant growth potential. By reducing the employment exit rate of older people to that of the best 10% of OECD countries, about half of the OECD countries could gain at least 0.2 percentage points of annual GDP per capita growth with respect to the baseline scenario. Adding the effect of closing the gender employment gap at older ages, the gain in annual GDP per capita growth for the whole OECD could reach 0.26 percentage points, twice as much as the gain from closing the gender employment gap for young and prime‑age people.
Overall, all these avenues to mobilise untapped labour resources should be considered as part of the solution to the growth challenges posed by population ageing. Yet, significant investments and costly policy actions are required to make them work. Nonetheless, even mobilising these untapped resources to reach two‑thirds of the above‑mentioned, very ambitious potentials in all these dimensions would allow cushioning 70% of the annual loss in GDP per capita growth due to demographic change as projected in the baseline scenario for the period 2024‑60, reaching a projected annual GDP per capita growth of 0.9%. This rate could be further increased with appropriate policies to revive productivity growth.
Failing to mobilise under-represented groups and in particular older workers in good health will not only lead to a significant reduction in GDP per capita growth but would also imply shifting the burden onto younger cohorts, as a smaller working-age labour pool will have to produce more just to maintain living standards of a larger dependent population. This would raise serious fairness issues as intergenerational inequalities have already progressed in favour of older generations in past decades:
Older generations (both 55‑64 year‑olds and 65+ year-olds) have benefited from higher income growth than the young (aged 25 to 34) in most OECD countries since the mid‑1990s, and the poverty risk has shifted away from older people towards children.
Intergenerational disparities in household wealth are large, as older generations have benefited from booming asset markets, and in particular rising house prices. Meanwhile, younger cohorts face barriers to wealth accumulation, and homeownership is increasingly out of reach for many.
A large share of public social spending is devoted to seniors in the form of pensions and health expenditures, and this part will continue to grow as populations continue to age and live longer.
The next two thematic chapters delve deeper into the challenges of, and perspectives for, mobilising labour resources at older age while making careers longer and more successful (Chapters 3 and 4). Chapter 4 also discusses the link between skills and individual productivity. Finally, Chapter 5 analyses the effect of job mobility and labour reallocation on productivity growth and discusses the potential implications that an ageing workforce, associated with lower job-to-job mobility, could have on productivity growth trends.
The world is changing fast. Several megatrends, such as digitalisation, climate change and population ageing, are transforming the world we live in and are having a deep impact on our lives, cultures, societies and living standards. The world of work will be at the forefront of these transformations and their impacts. It will have to adapt to these trends, seizing the opportunities they bring and addressing the challenges. While digitalisation (including its latest developments with generative artificial intelligence – AI hereafter) and mitigation of, and adaptation to, the effects of climate change have been widely analysed in previous editions of the OECD Employment Outlook – see e.g. OECD (2019[1]; 2023[2]; 2024[3]), the demographic transition and its consequences for the labour market have generally received less attention in recent years in this publication series. This edition of the OECD Employment Outlook aims to fill this gap.
In all OECD countries, declining fertility and increasing life expectancy have thrust population ageing to the forefront of the agenda – see e.g. OECD (2023[4]; 2024[5]). On the one hand, demographic change, and the progressive exit of baby boomers from the labour market, is turning the contribution of labour input to economic growth from positive to negative in almost all OECD countries. Widespread labour shortages in all sectors and occupations – see OECD (2023[2]; 2024[3]; 2024[6]) and Chapter 1 – are, partially, the salient symptom of this structural change. And this occurs after decades of declining productivity growth in the OECD area. On the other hand, ageing populations tend to increase demand for support and spending on health, long-term care, and pensions, and will continue to do so in the future. Absent policy action, increasing demands from a large, and in some cases expanding, consumer pool1 will have to be satisfied by the production of a shrinking labour pool, challenging living standards and even societal cohesion.
The remainder of this Outlook will focus on the consequences of demographic change for the economy and the labour market, and how the labour market can be part of the solution. This chapter aims at setting the scene by discussing the impact of demographic change on economic growth and intergenerational inequality. It is divided as follows. Section 2.1 describes key demographic trends. Section 2.2 presents a simple baseline projection scenario to describe the likely consequences of demographic change on economic growth without further policy action. Section 2.3. considers various simulation scenarios enabling a quantification of the potential of different labour market channels that could help mitigate or even fully offset these consequences. Section 2.4 focuses on trends in intergenerational inequalities and shows that income and wealth trajectories of older and younger generations have followed divergent patterns. Section 2.5 concludes and presents a roadmap of the remainder of the book.
The overarching conclusion of the chapter is that maintaining current growth rates cannot be achieved without a significant mobilisation of the employment potential of all under-represented groups. Any other strategy aiming at coping with the reduction in gross domestic product (GDP) per capita growth brought about by population ageing risks falling short of the target and bringing about significant implications for fairness, as older cohorts have been increasingly better off in recent years as compared to younger generations. The chapter is followed by three thematic chapters delving deeper into challenges of, and perspectives for, mobilising labour resources at older age while making longer careers more successful and ensuring that older workers in good health can thrive in the labour market.
People in OECD countries live longer and in better health than before. For the OECD area, longevity has steadily increased since the end of World War II (WW II), with the temporary exception of the years of the COVID‑19 pandemic (Figure 2.1). Since 1950, life expectancy at birth has gained 20 years, most of which in good health (OECD, 2017[7]; 2020[8]), even though the gap between life expectancy and health-adjusted life expectancy has increased in a number of OECD countries in recent years (Garmany and Terzic, 2024[9]; OECD/European Commission, 2024[10]). Even in recent years, however, the increase in healthy life expectancy at age 60 years represents 70% of the overall increase in life expectancy at age 60 (see Box 2.1). All these are significant achievements. At the same time, however, fertility rates, in OECD countries, which rose after the end of World War II and remained high until the 1960s, have plummeted since then. The total fertility rate (i.e. the average number of live births a woman would have by age 50 if she were subject, throughout her life, to the age-specific fertility rates observed in a given year) for the OECD area has collapsed from about 3.4 until the mid‑1960s to 1.5 in recent years, although large differences across countries remain – in 2023, total fertility rates in OECD countries ranged between 0.7 in Korea and 2.8 in Israel. According to the medium scenario of the UN population projections (United Nations, 2024[11]), fertility rates may stabilise in the future, while life expectancy will continue to increase, albeit at a slower pace (Figure 2.1).2
Average fertility rates and life expectancy at birth, OECD, 1950‑2060
Note: The figure refers to the whole OECD area. The medium scenario of the population projections is used. The vertical line indicates the first year of projections (2024).
Source: André, C., P. Gal and M. Schief (2024), “Enhancing productivity and growth in an ageing society: Key mechanisms and policy options”, https://doi.org/10.1787/605b0787-en, updated using United Nations (2024), World Population Prospects 2024, https://population.un.org/wpp/.
OECD citizens are not only living longer but also enjoying more years in good health. On average, in 2021, a 60‑year‑old individual can expect to live 23 years more, with 17.3 of those years being spent in full health – see also OECD/European Commission (2024[10]). This represents an increase in life expectancy at age 60 of 1.7 years, relatively to 2000, with 70% of these gains corresponding to increases in healthy life expectancy(Figure 2.2), defined by the World Health Organization (WHO) as the number of years expected to live in full health – i.e. excluding years of life expectancy burdened by disease or injury.
Average life expectancy and healthy life expectancy at age 60 in the OECD, years
Note: Estimates are based on 38 OECD countries. Healthy life expectancy refers to the number of years expected to live in full health, while the remaining years of life expectancy in less than full health refer to the years expected to live under the burden of disease or injury.
Source: Secretariat’s calculations based on data from the WHO Global Health Observatory.
Fertility trends imply that the relatively large cohorts born in the two decades after WW II have been continuously replaced, in recent years, by increasingly smaller cohorts entering the labour market. As baby boomers exit the labour market, the working-age population (defined as those aged from 20 to 64 years) has started declining or will start declining soon in the majority of OECD countries (Figure 2.3). Overall, the working-age population in the OECD area is projected to decline by 8% between 2023 and 2060. But in one‑quarter of OECD countries, including many Eastern Asian and Southern, Central and Eastern European countries, it is projected to fall by more than 30% – and by up to 46% in Korea. Very high rates of decline of the working-age population are also projected in four “accession” countries (Bulgaria, Croatia, Romania and Thailand) and China. By contrast, in a few OECD countries, the working-age population is still projected to increase, especially in Australia and Canada, where net migration rates are projected to be high (see Section 2.3.2 below), and Israel and Mexico, where age pyramids have still a broad base as fertility is still high or has declined below replacement levels3 only recently.
Projected percentage change in the working age population (aged 20‑64 years), 2023‑60
Note: The medium scenario of the population projections is used. OECD: Weighted average of OECD countries.
Source: Secretariat’s calculations based on United Nations (2024), World Population Prospects 2024, https://population.un.org/wpp/.
Decreasing fertility and increasing longevity also brought about a fundamental change in the age structure of the population. The share of older individuals (aged 55‑64 years) within the OECD working-age population increased by about 50% between 1980 and 2023, although in most countries it is not projected to increase much further in the next quarter of century (Figure 2.4). By contrast, the old-age population has grown dramatically and will continue to do so. Consequently, the share of the seniors (aged 65 years and above) in the population has soared and is projected to increase further at a faster pace. In the OECD area, the old-age dependency ratio, defined as the ratio of the seniors to the working-age population, increased from 19% in 1980 to 31% in 2023 and is projected to increase further to 52% by 2060 (Figure 2.5).4 By then, it will be above 50% in 30 OECD countries, and above 75% in Italy, Japan, Poland, Spain and Korea. In the latter country, UN projections suggest that in 2060 the overall dependency ratio (including also children in the numerator) will surpass 100%,5 and it will reach 70% in the OECD area as a whole.
Although there is significant uncertainty about population projections (especially regarding fertility rates), UN projections still forecast, with a probability of at least 90% in all countries except Israel, a significant increase of the old-age dependency ratio by 2060 (see Annex Figure 2.A.1).6 Absent policy action or changes in individual behaviour – for example because people living longer in good health may wish to stay in the labour force and remain employed for longer, see e.g. IMF (2025[12]) – this shift in the population structure will imply that an increasingly small labour pool will have to generate income for an increasingly large pool of people consuming but not producing, weighing heavily on the capacity of countries to continue to improve their living standards. The next section will delve more deeply into this type of accounting.
Share of individuals aged 55 to 64 years in the working-age population, various years, percentages
Notes: The working-age population is the population aged 20‑64 years. 2060 is a projection based on the medium scenario. Countries are ranked by their projected percentage point increase between 2023 and 2060, in descending order. OECD: Weighted average of OECD countries.
Source: Secretariat’s calculations based on United Nations (2024), World Population Prospects 2024, https://population.un.org/wpp/.
Ratio of old-age to working-age population, various years, percentages
Notes: The old-age population is the population aged 65 years or more. The working-age population is the population aged 20‑64 years. 2060 data concern projections. The medium scenario of the population projections is used. OECD: Weighted average of OECD countries.
Source: Secretariat’s calculations based on United Nations (2024), World Population Prospects 2024, https://population.un.org/wpp/.
The decline of the working-age population and the increase of the old-age dependency ratio pose a significant challenge to maintaining OECD countries’ GDP per capita growth. While living standards are better measured through either the OECD better life index or gross national income per capita, due to measurement and forecasting issues this chapter limits the discussion to GDP per capita growth, which is a key determinant of material living standard (see Box 2.2).
The concept of a country’s living standards typically captures average well-being in that country. At the country level, living standards are typically assessed through a measure of aggregate well-being such as the OECD Better Life Index (BLI hereafter) and its sub-indicators – see OECD (2011[13]; 2024[14]). The BLI is a synthetic index that aggregates several indicators capturing multiple dimensions of material living standards (income, wealth, jobs, earnings and housing) and quality of life (health, combining work and life, education and skills, social connections, civil engagement and governance, environmental quality, personal security, and perceived well-being).
By contrast, GDP measures the output produced domestically and GDP per capita standardises GDP by dividing it with the total population. Although it is widely used for its availability, it is a relatively imprecise proxy for the gross income of domestic households. Notably, GDP per capita includes income paid to non-residents and excludes residents’ income from production in other countries. Moreover, it does not incorporate other aspects of well-being that are not directly dependent on income, and only partially integrates information on how the various types of capital that sustain well-being are changing over time (OECD, 2011[13]). Yet, the growth rate of GDP per capita strongly correlates with the growth rate of real gross disposable income per capita of households and non-profit institutions serving households, a key component of any indicator of material living standards. For example, the correlation of these two growth rates between 2007 and 2022, across the OECD countries for which data are available, is 0.62, and increases to 0.87 upon exclusion of Ireland, a clear outlier in the relationship (Annex Figure 2.A.2, Panel A). The main reason for this result is that what is produced domestically remains, in most OECD countries, the main determinant of domestic household income. GDP per capita growth also correlates with changes in the aggregate BLI. For example, the cross-country correlation of the growth rate of GDP per capita and changes in the BLI between 2010 and 2023 is mildly positive (0.22) but soars to 0.55 when Ireland and Türkiye, two clear outliers where GDP per capita growth outpaced progress in the BLI, are excluded (Annex Figure 2.A.2, Panel B). Overall, this suggests that GDP per capita growth projections can be informative, albeit imperfectly, as regards the possible future evolution of living standards, and in particular material living standards.
In the context of population ageing, however, projected GDP per capita growth will likely overestimate the potential change in living standards. Indeed, as discussed in Section 2.4, given current policies, population ageing implies that an increasingly larger share of GDP will be spent on health and pensions just to maintain living standards among the older population. A positive, but low GDP per capita growth may be insufficient to generate sufficient additional resources to improve living standards among the whole population.
Trends in the dependency ratio will have a direct impact on GDP per capita, which is the product of aggregate GDP per person employed, and the share of employed persons in the total population (employment-to-population ratio). In turn, the latter is equal to the product of the employment rate (the ratio of employment to working-age population) and the share of the working-age population in total population.7 While it can be expected that the replacement of older cohorts with younger cohorts with stronger labour market attachment8 will somewhat increase the aggregate employment rate of people of working age, the share of the working-age population in total population will shrink dramatically, as shown in the previous section. In turn, this will significantly depress the trend of the employment-to-population ratio and, therefore, of GDP per capita growth – see also André, Gal and Schief (2024[15]).
Figure 2.6 shows the projected reductions of the ratio of employment to population, disaggregated into the contributions from the two channels outlined above – namely, the changes in the share of the working-age population in total population (population structure channel) and the changes by cohorts in the age‑specific employment rate (cohort replacement channel). The effect of the latter partially offsets that of the former due to the stronger labour market attachment of younger cohorts throughout their career. Age‑ and gender-specific employment rate projections assume constant labour market entry and exit rates for each group and are drawn from Fluchtmann, Keese and Adema (2024[16])9 and then combined with the medium scenario of UN population projections, used in the previous section.
Projected changes in the employment-to-population ratio, by channel, 2023‑60, baseline scenario, percentage points
Notes: The figure reports projected changes in the share of employed persons in the overall population in the baseline scenario. The medium scenario of the UN population projections is used. Population-structure indicates the contribution of changes in the structure of the population (and notably the decline in the ratio of working-age to total population), assuming constant the age‑ and gender-specific employment rates as in 2023. Cohort-replacement indicates the change in the contribution of the projected changes in the age‑ and gender-specific employment rates as new cohorts replace older cohorts. OECD: Weighted average of OECD countries.
Source: Secretariat’s calculations based on OECD Data Explorer, “Employment and unemployment by five‑year age group and sex – levels”, http://data-explorer.oecd.org/s/253; Fluchtmann, J., M. Keese and W. Adema (2024), “Gender equality and economic growth: Past progress and future potential”, https://doi.org/10.1787/fb0a0a93-en; and United Nations (2024), World Population Prospects 2024, https://population.un.org/wpp/.
Because of population ageing, the percentage share of employed persons in the OECD total population (employment-to-population ratio) is projected to decrease from 48.1% in 2023 to 46.2% in 2060, that is by 1.9 percentage points. This share will decline in all but three OECD countries (Israel, Mexico and Türkiye) but the fall is projected to be particularly large (more than 10 percentage points) in the Slovak Republic and Spain. Yet, the replacement of older with younger (usually more active) cohorts is expected to attenuate this reduction in many countries. Projections suggest that the drop of the employment-to-population ratio would be 1.6 percentage points larger, on average, if younger cohorts did not have stronger labour market attachment than older cohorts. With few exceptions,10 the replacement of older cohorts with younger cohorts indeed implies an increase in the employment rate of people of working age. However, this is usually insufficient to fully compensate for the effect of the change in the structure of the population, and especially the expansion of jobless old-age population, which remains the most important factor driving the (negative) dynamics of the share of employed persons in the population in almost all countries, except Colombia and Türkiye.
Under the assumption that the growth rate of labour productivity (GDP per person employed) remains approximately constant,11 the contraction of the employment-to-population ratio, in percentage terms, translates into an equal percentage effect on GDP per capita. In particular, setting the productivity growth rate in the next quarter of century at the level observed in 2006‑19,12 the baseline scenario suggests that GDP per capita growth in the OECD area will be reduced by about 40%, falling from 1.0% per year in 2006‑19 to an average of 0.6% per year in the period 2024‑60 (Figure 2.7).13 In other words, by 2060, GDP per capita in the OECD will be 14% lower than what would occur if labour input continued to grow at the same rate as during the previous business cycle.
Recent and projected annual GDP per capita growth, baseline scenario, percentages
Notes: Projected real GDP per capita growth obtained assuming the same growth of GDP per worker as in 2006‑19 and the baseline projection scenario for the employment-to-population ratio. Countries ordered by the size of the average projected growth slowdown between 2006‑19 and 2024‑60. 2006‑18 for Australia, 2013‑19 for Chile, 2015‑19 for Colombia and 2007‑19 for Korea instead of 2006‑19. OECD: Weighted average of OECD countries. GDP: Gross Domestic Product.
Source: Secretariat’s calculations based on OECD Data Explorer, “Employment and unemployment by five‑year age group and sex – levels”, http://data-explorer.oecd.org/s/253 and “Productivity levels”, http://data-explorer.oecd.org/s/254, Fluchtmann, J., M. Keese and W. Adema (2024), “Gender equality and economic growth: Past progress and future potential”, https://doi.org/10.1787/fb0a0a93-en; and United Nations (2024), World Population Prospects 2024, https://population.un.org/wpp/.
Without a pick-up in employment and productivity trends beyond baseline projections, almost all OECD countries will see their annual growth declining due to population ageing. This is because historical employment trends, including in recent years, have always been such that the labour input has provided a positive contribution to average GDP per capita growth. By contrast, current trends in population ageing imply that, in per capita terms, the contribution of the labour input will turn negative in all OECD countries except Israel, Türkiye and Mexico, as seen in Figure 2.6 above. Expected annual GDP per capita growth will remain about constant in Ireland and the United States and decline by 0.25 percentage points or less in Denmark, France and Portugal. In all these countries, however, it can be argued that the comparatively small decline is, by and large, due to an already limited contribution of the labour input to growth in recent years, as shown by the small difference between the rates of growth of GDP per capita and GDP per worker in the 2010s. At the other side of the spectrum, many Central and Eastern European countries as well as Chile, Colombia, Korea and Luxembourg will see their growth rate contract by at least 1 percentage point, although declining from very high rates in all these countries, except Austria, Germany and Luxembourg. In all these countries, if the employment-to-population ratio continued to grow at the same rate as before, GDP per capita would be at least 30% higher in 2060 than what forecasted by the baseline projections. Largely because of the negative productivity growth observed in recent years and used as a benchmark, annual growth of GDP per capita is estimated to be significantly negative in Greece, Italy and Luxembourg (below ‑0.5%), A slightly negative average annual growth of GDP per capita is also projected in Austria and Norway.14
The analysis of the previous sections suggests that, in the short and medium term, most OECD countries will face an unprecedented collapse of GDP per capita growth, which can be avoided only through bold policy action. This will require operating on several levers, which will be examined in this section.
2.3.1. Is a sufficient recovery in productivity growth in sight?
A key objective in all OECD countries is to revive growth of GDP per person employed, which has declined since the last decade of the 20th century in almost all of them (Annex Figure 2.A.3). GDP per person employed, however, is the product of hours worked per worker and hourly labour productivity (GDP per hour worked). Hours worked per worker in OECD countries are on a historically declining path, even though their fall has slowed down somewhat in recent years – see e.g. OECD (2021[17]; 2022[18]). Total hours per worker have declined since 2005 in almost all OECD countries for which data are available, for an average reduction of 5.5% – see Chapter 1. Because of these historical trends, significantly increasing hours worked by the average person may therefore be out of reach and their level can at best be expected to stabilise in the near future.15
Hourly labour productivity growth has also been on a declining trend for many years – see e.g. OECD (2024[19]) and Chapter 5.16 Yet, the diffusion of generative AI as a new general-purpose technology and the deepening of automation – see OECD (2023[2]) – have made many scholars optimistic about the possibility of significantly reviving hourly productivity growth in the same way as the diffusion of digital technologies in the 1990s eventually led to a recovery of hourly productivity growth for almost a couple of decades in a few countries – see e.g. Brynjolfsson and Mitchell (2017[20]); Aghion, Jones and Jones (2017[21]); Lu (2021[22]); Baily, Brynjolfsson and Korinek (2023[23]); Filippucci et al. (2024[24]); OECD (2024[25]); and Committee on Automation and the US Workforce (2024[26]).17 Additionally, ageing-driven increases in aggregate saving can lead to greater capital deepening thereby boosting labour productivity (see Box 2.3). Yet, in many other OECD countries, especially in Asia and continental Europe, the diffusion of digital technologies of the 1990s did not halt the secular decline in hourly productivity growth, which continued to fall more or less steadily – see e.g. Fernald, Inklaar and Ruzic (2024[27]) and OECD (2024[19]). And while in the 1990s OECD countries’ hourly productivity growth was also benefitting from the large expansion of trade and global value chains – see e.g. Grossman and Rossi-Hansberg (2008[28]); Baldwin (2012[29]); Criscuolo and Timmis (2018[30]); and OECD (2019[1]), this push is unlikely to be repeatable today, especially on such a large scale (Constantinescu, Mattoo and Ruta, 2016[31]; Goldin et al., 2024[32]). Moreover, other factors, including the acceleration of climate change, could additionally exert a negative pressure on productivity growth in the future – for example, in a recent paper, Bilal and Känzig (2024[33]) estimate that world GDP per capita would already be 37% higher today had no warming occurred since 1960, which implies an average drag on world hourly productivity growth of about 0.5 percentage points per year. This drag is likely to increase in the near future, especially if insufficient policy action is taken to counter climate change – see OECD (2024[3]). There is also some concern that an ageing labour force could dampen hourly productivity growth, although the literature on the aggregate effect of ageing on productivity growth is nuanced – see Box 2.3 and Chapters 4 and 5. For all these reasons, nevertheless, it is unlikely that labour productivity growth could be revived at a level which, alone, would suffice to compensate the projected decline in GDP per capita growth.
Ageing can affect labour productivity through multiple pathways, with no clear consensus on the direction and magnitude of its impact. Understanding this complex causal relationship requires zooming in on each of those potential pathways, ranging from micro, worker and firm-level ageing effects to broader macro impacts of aged societies on aggregate demand and public and private investment (André, Gal and Schief, 2024[15]).
The impact of individual ageing on the worker’s productivity remains ambiguous. While individual hourly productivity declines at some point in people’s lives, this point is likely to depend on multiple factors, including individual characteristics (including health and education), type of occupation (see also Chapter 4), and workplace organisation (see also Chapter 3). For many workers, the tipping point could be at relatively old age. Indeed, even if physical strength and cognitive skills, including learning capacity, deteriorate during the working life – see e.g. Prskawetz and Lindh (2006[34]) and Chapter 4 – especially when such capabilities are not regularly used (Hanushek et al., 2025[35]), older workers often take on different tasks than their younger colleagues, such as managerial roles (National Research Council, 2012[36]), and likely rely on different skill sets. Thus, firms’ ability to harness skill complementarities across an age‑diverse workforce is a key determinant of older workers’ productivity (OECD, 2020[8]).
Ageing societies may experience a less dynamic business environment, both in terms of firm entry and exit rates, and a concentration of aggregate demand in less productive sectors, potentially dragging down overall productivity growth. Entrepreneurship quality may decrease as decision-making positions are increasingly occupied by older individuals, preventing younger workers from gaining sufficient experience by the mid‑40s, the typical age to start a successful business (Azoulay et al., 2020[37]). Incumbent firms may also face less pressure from new competitors and products, as ageing consumers crystalise consumption patterns (Bornstein, 2021[38]). At the same time, demographic ageing can tilt in demand towards activities characterised by lower productivity growth, particularly long-term care services (Baumol, 1993[39]; Cravino, Levchenko and Rojas, 2022[40]).
Ageing is also likely to reshape the allocation of inputs and investment decisions in both the private and public sectors, with opposite effects on productivity growth. An ageing workforce tends to be less mobile between jobs, reducing the efficiency of the matching process, thereby significantly depressing productivity growth through the reallocation channel (see Chapter 5) and potentially contributing to labour shortages. However, these shortages can also create incentives to invest in labour-saving and labour-replacing technologies (Acemoglu and Restrepo, 2017[41]; 2021[42]). Moreover, ageing-driven increases in aggregate saving can lead to greater capital deepening thereby boosting labour productivity, under the condition that real interest rates can adjust downwards to absorb excess savings (Eggertsson, Lancastre and Summers, 2019[43]). On the contrary, rising public and private social expenditures associated with ageing (see Section 2.4) may reduce saving and crowd out productive public investments.
Ultimately, these opposing forces make it difficult to predict which effects will dominate, leaving the net impact of ageing on productivity growth ambiguous.
2.3.2. Complementary scenarios for mobilising untapped labour resources
Mobilising untapped labour resources appears therefore a necessary complementary strategy. The origin of the decline in labour input can be traced back to the secular decline in fertility rates (see Section 2.1). However, as briefly discussed in Box 2.4, while fertility policies can contribute to contain the decline in fertility rates, it is unlikely that such policies will enable countries to revert significantly the trend in population ageing, because of changes in preferences for having children. Moreover, at the least in the next quarter of century, raising the fertility rate will mechanically increase the denominator of GDP per capita without significantly increasing the numerator, with a muted, or even negative, effect on growth.
Actions that reduce the opportunity cost of rearing children, such as providing affordable childcare and strengthening work-life balance policies could serve the dual purpose of closing the gender gap in labour market participation and increasing fertility. In fact, from the 1980s onwards, women’s employment rates are positively associated with higher total fertility rates across the OECD (Fluchtmann, van Veen and Adema, 2023[44]). Similarly, expenditure on parental leaves, family allowances, housing, and early childhood education and care policies positively correlates with fertility. However, labour market, housing, education and family policies can only partially explain the evolution of fertility rates (Fluchtmann, van Veen and Adema, 2023[44]), which could point to the role of other factors such as financial insecurity, societal attitudes and norms (OECD, 2024[5]).
To better capture the full picture, it is indeed necessary to search for other drivers of fertility trends. Culture, social norms, and attitudes towards childbearing are natural candidates (Fernández and Fogli, 2009[45]; Newson and Richerson, 2009[46]). The ascension of individual self-fulfilment as an alternative life goal to family formation (Sobotka, 2008[47]), coupled with increasingly higher standards of parental roles, increase the implicit cost of childrearing (OECD, 2024[5]), which add up to the higher costs of housing and, in some countries, education, both post-compulsory and to support children of school age through private tutoring. Such mounting implicit costs can in turn influence parents to delay entering parenthood (Gustafsson, 2001[48]), have fewer children (Becker, 1960[49]), or opt for having no children at all (Baudin, de la Croix and Gobbi, 2015[50]). Given the potentially slow and even intergenerational effort required to transform social norms, attitudes, or preferences (Bisin and Verdier, 2000[51]), it is unlikely that policy will swiftly boost fertility through these channels.
Summing up, public policy does play a role in explaining at least partially cross-country differences in fertility rates, despite the generalised trend decline, but it cannot be expected that even best policy practices will revert fertility rates back to replacement levels. Even if they could succeed, higher birth rates today will materialise into larger working-age cohorts no earlier than 20‑25 years from now. During that time, rising child dependency ratios would add to the already growing old-age dependency, meaning that until the mid‑2050s, new cohorts will add much more to the denominator than to the numerator of GDP per capita. Since their contribution to the size of the working-age population will remain small in subsequent years, a gradual increase in fertility rates to replacement levels by 2035 (almost certainly an overoptimistic perspective given recent trends), would at best have little impact on GDP per capita by 2060.
Three more promising sources of labour resources will be considered in this section. They include: i) increasing employment and labour force participation of the older people in good health; ii) closing the gender gap in employment; and iii) sustaining net migration, including by better attracting talent, making the most of migrants’ skills and reducing brain drain (in countries where this is relevant).
Another source of labour resources that is not modelled here but is likely to play a role in countries with high rates of youth not in employment, education or training (NEET)18 is raising youth employment. Nonetheless, to ensure that young workers are equipped with the skills required to navigate the labour market during their whole career, effectively reducing NEET rates often entails improving education services and reducing school dropouts and/or offering full-time, second-chance educational programmes, like the US Job Corps programme, without necessarily having a significant impact on aggregate employment rates in those age categories – see e.g. Schochet, Burghardt and McConnell (2006[52]); Cohen and Piquero (2015[53]); and OECD (2018[54]; 2023[55]).19
Migration contributes to lessen the challenge posed by demographic change to GDP per capita growth
Regular migrants are already contributing to sustaining working-age population, and, in the baseline scenario, are projected to continue to do so in the future, although projected net migration rates20 in the UN medium scenario are lower than those observed in the most recent available data in many countries – 2021‑24, see Annex Figure 2.A.4.21 Yet, translating the contribution of migration flows to the population in the host country into their contribution to employment-to-population ratios is not a trivial task. On the one hand, in some countries, labour force participation and employment rates of the foreign-born are lower than those of the native‑born, especially soon after their arrival in the host country (OECD, 2024[56]). Multiple factors contribute to this pattern, including, inter alia, migration reasons (work, humanitarian, family, study, etc…), socio‑economic status, and the fact that migrants often do not have the competences in demand in host countries (Spielvogel and Meghnagi, 2018[57])22 and, at least at the beginning, may lack the experience – including language skills – to effectively navigating the host country’s labour market (OECD, 2023[58]). On the other hand, foreign-born are over-represented among the self-employed, and migrant entrepreneurs significantly contribute to job creation in host countries (OECD, 2024[56]). Moreover, migrants play a key role in reducing labour shortages in certain sectors, such as long-term care – see e.g. Rapp and Sicsic (2020[59]); Grabowski, Gruber and McGarry (2023[60]); Jun and Grabowski (2024[61]) and OECD (forthcoming[62]).
Due to data limitations, however, for the purpose of the simple simulation exercise presented in this chapter, observed employment rate gaps between native‑born and foreign-born by gender are taken as starting points. The projected percentage changes of employment rates of the foreign-born are then assumed to be the same as those of natives in the same gender and age category – that is, the computed 2023 employment rate gap in each category is assumed to be constant in subsequent years. These limitations require some caution in drawing conclusions from the results.23
As shown in Figure 2.8, should migration flows be slowed down in the near future so that to set net migration rates to zero (zero-migration scenario), projected GDP per capita growth would fall additionally by 0.06 percentage points per year in the OECD with respect to the baseline scenario (presented in Section 2.2 above), in which net migration rates are set as in the UN medium scenario (see Annex Figure 2.A.4). The decline would be greater than 0.15 percentage points only in Canada and Luxembourg, where this relatively large effect is essentially due to relatively large net migration rates forecasted in the UN medium scenario in these countries. In other countries, such as Denmark, France, Portugal, Spain and the United Kingdom, even if setting net migration to zero would have only a small impact on growth in absolute terms, these additional losses would still increase the projected loss in GDP per capita growth entailed by the baseline scenario by more than 20%. On the other side of the spectrum, countries that are projected to have, on average, negative net migration rates due to significant emigration – Mexico, Türkiye and some Central and Eastern European countries – would potentially benefit from a zero-migration world, but the projected effect would remain small.
Percentage point difference in average annual GDP per capita growth: baseline vs. alternative migration scenarios, 2024‑60
Notes: The chart compares projected percentage point differences in average real GDP per capita growth between three alternative scenarios and the baseline scenario. Zero-migration sets future net migration rates equal to 0. Moderately high migration sets future net migration rates equal to the median percentile of the cross-country distribution in 2021‑24. High migration sets future net migration rates equal to the 75th percentile of the cross-country distribution in 2021‑24. Projections assume the same change in employment rates for foreign and native‑born but different starting levels derived from observations. For Israel and Türkiye starting levels are set to be the same due to lack of data. Countries are ranked in descending order by the difference between the high migration and the zero migration scenarios. p.p.: percentage points. OECD: Weighted average of OECD countries. GDP: Gross Domestic Product.
Source: Secretariat’s calculations based on OECD (2024), International Migration Outlook 2024, https://doi.org/10.1787/50b0353e-en; OECD (2024), “Employment and unemployment by five‑year age group and sex – levels”, http://data-explorer.oecd.org/s/253, “Labour market outcomes of immigrants – Employment, unemployment, and participation rates by sex”, http://data-explorer.oecd.org/s/255 and “Productivity levels”, http://data-explorer.oecd.org/s/254; Fluchtmann, J., M. Keese and W. Adema (2024), “Gender equality and economic growth: Past progress and future potential”, https://doi.org/10.1787/fb0a0a93-en; United Nations (2024), International Migrant Stock 2024, www.un.org/development/desa/pd/content/international-migrant-stock, and United Nations (2024), World Population Prospects 2024, https://population.un.org/wpp/.
As already noted, however, net migration rates included in the baseline scenario (based on the UN medium scenario) are lower than historical rates in many countries. For OECD countries, the cross-country median of annual net migration rates in recent years was 0.46% in 2021‑24 and 0.37% in 2016‑20 against an average cross-country median of UN projected rates of 0.24% for 2024‑60 in the baseline scenario. Median net migration rates have also trended up steadily since at least 1990 (United Nations, 2024[63]) and inflows of foreign-born have never been so high (OECD, 2024[56]). Figure 2.8 considers therefore two higher migration scenarios, which might better help characterise the potential of the migration channel: a moderately high scenario, in which net migration rates are set to increase to the median levels of the 2021‑24 cross-country distribution (0.46%) in all countries by 2030 (and remain constant thereafter), except if the baseline projection is higher; and a high migration scenario, in which net migration rates are set to increase to the 75th percentile of the cross-country distribution (0.61%), again in all countries by 2030 except if the baseline projection is higher.24 It must be noted, however, that, while in the baseline and the zero migration all world flows are balanced, in the high and moderately high migration scenarios this is not the case since the same net migration rate target is used for all the countries. These two scenarios, therefore, only allow simulating the potential gains from migration separately for each country.
Potential gains from higher regular migration in terms of GDP per capita growth appear to be particularly high in many Southern, Central and Eastern European countries, where they could exceed 0.1 percentage points per year in the moderately high migration scenario and 0.2 percentage points in the high migration scenario (Figure 2.8). Overall, compared with a world without migration (zero-migration scenario), by reaching a net migration rate of 0.61% by 2030, the median OECD country could improve its GDP per capita growth by 0.13 percentage points,25 and up to 0.25 percentage points in Greece, Italy, Luxembourg and Slovenia. By contrast, gains from migration appear close to zero in countries such as the Netherlands, Israel and Mexico, due to either the projected composition of immigration flows (larger for groups with no or limited labour market participation – Israel) and/or the much lower current participation of migrants to the labour market – Mexico and the Netherlands, see OECD (2024[56]). In particular, the latter factor is likely to lead to an underestimation of the growth potential of the high migration scenarios, since employment rates of migrants would improve if increasing net migration flows are accompanied by more selective migration policies tuned towards labour market needs of the host country.
Overall, while regular migration can contribute to lessen the challenge demographic ageing poses to living standards, its potential does not appear a game changer unless net migration rates increase well above the historically high rates observed in recent years. Yet, the higher the sudden increase in net migration flows, the larger the policy interventions required to integrate higher migrant inflows into the labour market – see e.g. OECD (2016[64]; 2017[65]; 2017[66]; 2021[67]; 2023[58]). Needed interventions may range across very broad areas, for example from introduction measures, language training and specialised training (including in origin countries) to providing affordable transport and housing – see also Hermans et al. (2020[68]). Ensuring adequate access to these services is an unavoidable, integral part of well-managed migration and integration policies, but these interventions are costly and difficult to provide at scale (OECD, 2024[56]). Moreover, most migrants remain only temporarily in the host country, making it even more difficult to sustain high net migration rates over time (OECD, 2024[56]). And effectively increasing net migration rates in a way matching host countries skill needs might require retaining or reattracting native‑born citizens, and more generally, competing effectively to attract talent (Spielvogel and Meghnagi, 2018[57]; d’Aiglepierre et al., 2020[69]; Beine, Peri and Raux, 2023[70]). Last but not least, host countries’ populations are often not ready to accept large increases in foreign-born inflows, even when these could bring clear economic gains, and attempts to raise them significantly may generate strong political backlash (Hainmueller and Hopkins, 2014[71]; McCann, Sienkiewicz and Zard, 2023[72]; Boeri et al., 2024[73]), especially when flows have a strong low-skilled component (Moriconi, Peri and Turati, 2022[74]; Docquier and Rapoport, 2025[75]).
Closing the gender gap in employment can deliver large growth dividends
Reducing the employment gap between men and women of working age appears to have a clear potential. In all OECD countries, women have lower employment rates than men of the same age, and even if gender gaps in employment rates have continued to narrow in recent years, the pace of convergence has slowed down. The gender gap in employment rates of people of working age fell by 7.6 percentage points between 2000 and 2021, but its fall halved in the second decade of the 21st century (OECD, 2023[76])) and OECD evidence suggests that, in recent years, female employment has progressed significantly only in countries with the largest gender gap in employment (OECD, 2024[3]), although this progress is decelerating (see Chapter 1). Overall, this suggests that considerable progress can still be made in closing the gender gap in most countries, and that it could bring significant growth dividends.
Closing the gender employment gap between men and women of all ages in all OECD countries, while keeping all other assumptions as in the baseline scenario, could increase annual OECD GDP per capita growth by 0.2 percentage points (Figure 2.9).26 In countries where female labour force participation is particularly low (Colombia, Costa Rica, Mexico, Greece, Italy and Türkiye), the simulation suggests that closing the gender gap could increase GDP per capita by 0.3 percentage points or more and up to 0.6 percentage points in Colombia. The contribution of closing the gender gap to GDP per capita growth would instead remain modest in countries where female employment rates are already close to those of men. Specifically, in countries such as Australia, France, Israel, Latvia, Lithuania, Luxembourg and the United Kingdom, closing the gender employment gap would yield a GDP per capita growth dividend of less than 0.1 percentage points annually. In Australia and France, nevertheless, closing the gender employment gap would allow recovering more than one fourth of the loss in GDP per capita growth due to ageing (cf. Figure 2.9 with Figure 2.7). Further gains can also be obtained by reducing the gender gap in hours worked per employed person in countries where these disparities are large (see Box 2.5).
Percentage point difference in average annual GDP per capita growth: baseline vs. gender equality scenarios, contributions of different age categories, 2024‑60
Notes: The chart compares projected percentage point differences in average real GDP per capita growth between a scenario where, by 2060, for both genders, employment rates in each age category are as high as that of the gender with the highest rate and the baseline scenario. “Contr. of gender equality among young and prime age” identifies the specific contribution of closing the gender gap at age 54 years or less. “Contr. of gender equality among older workers” identifies the specific contribution of closing the gender gap for people aged 55 years or more. OECD: Weighted average of OECD countries. p.p.: percentage points. GDP: Gross Domestic Product.
Source: Secretariat calculations based on OECD Data Explorer, “Employment and unemployment by five‑year age group and sex – levels”, http://data-explorer.oecd.org/s/253 and “Productivity levels”, http://data-explorer.oecd.org/s/254; Fluchtmann, J., M. Keese and W. Adema (2024), “Gender equality and economic growth: Past progress and future potential”, https://doi.org/10.1787/fb0a0a93-en; and United Nations (2024), World Population Prospects 2024, https://population.un.org/wpp/.
Across the OECD, women typically spend fewer hours in paid work than men (OECD, 2019[77]). While Figure 2.9 above presents the potential growth contribution of hypothetical scenarios that assume full closure of the gender gap in employment rates by 2060 (i.e. the extensive margin), another potential channel stems from reducing the gender gap in hours worked per worker (i.e. the intensive margin). Data on hours worked are not available for all OECD countries, and even when available they are not always comparable (see notes to Figure 2.10). For the OECD countries where data on hours worked are available, Figure 2.10 plots the potential gains of increasing hours worked to the level of the gender that works most hours, by 2060. As before, all other assumptions are kept as in the baseline scenario. At the OECD level, GDP per capita growth would be 0.16 percentage points higher than in the baseline scenario, that is without changes in the gender employment gap, which allows adding up the potential effects simulated in this box with those from the other channels in this section.
Percentage point difference in average annual GDP per capita growth: baseline vs. gender equality in hours worked scenario, 2024‑60
Note: The chart compares projected percentage point increase in average annual GDP per capita growth (with respect to the baseline scenario) in a scenario where, by 2060, for both genders, hours worked per employed person in each age category are as high as that of the gender with the highest number of hours worked, keeping all other assumptions as in baseline scenario. Usual hours worked per employee are used for reasons of cross-country comparability. Actual hours instead of usual hours for Korea and Mexico. Hours per worker instead of hours per employee in Korea. OECD: Weighted average of countries shown. p.p.: percentage points.
Source: Secretariat calculations based on OECD Data Explorer, “Employment and unemployment by five‑year age group and sex – levels”, http://data-explorer.oecd.org/s/253, “Average usual weekly hours worked on the main job”, http://data-explorer.oecd.org/s/256 and “Productivity levels”, http://data-explorer.oecd.org/s/254; Fluchtmann, J., M. Keese and W. Adema (2024), “Gender equality and economic growth: Past progress and future potential”, https://doi.org/10.1787/fb0a0a93-en; and United Nations (2024), World Population Prospects 2024, https://population.un.org/wpp/.
This average potential effect conceals some degree of heterogeneity, however. Countries with high levels of gender equality in hours worked, namely Latvia, Hungary, and Lithuania, would experience minimal gains in growth, ranging between 0.03 and 0.04 percentage points. In contrast, for 17 countries, growth dividends from closing the gender gap in the intensive margin could be higher than closing it in the extensive margin. The contribution of closing the gender gap in hours worked could be particularly important in Austria, Germany, the Netherlands, Switzerland and the United Kingdom, where women more frequently engage in part-time employment than men (OECD, 2019[77]; OECD, 2025[78]). Indeed, for these countries, GDP per capita could potentially grow by additional 0.25 percentage points or more by 2060, relatively to the baseline scenario. Effective gains could, nevertheless, be smaller, since this exercise abstracts from quantifying the costs required to increase hours worked by women – who already take up a higher share of unpaid work (see Box 2.6 below) – to the level of men.
Gender gaps in employment are, however, asymmetrically distributed across the age distribution, which implies that, across the OECD area, more than one‑third of the potential gains in GDP per capita growth would come from closing the gender gap for older workers. And in certain countries this figure can be much higher. In some cases, this is due to very high employment rates for older men – e.g. in countries such as Iceland and Estonia, where 77% and 65%, respectively, of the potential gains from closing the gap are concentrated among older women. However, in other countries, the very rapid decline of female labour force participation at older ages is responsible for this pattern. This is the case in Chile and Hungary, for example, where reducing labour market exit rates of older women represents 64% and 70%, respectively, of the potential gains.27 Conversely, in other countries such as France, potential gains from closing the gender gap are concentrated among prime‑age (and young) women, while the effect among older people is small. This is mechanically due to the fact that employment rates among older people are relatively low for both men and women, particularly for those aged 60 years or more – see OECD (2023[4]) and below – thereby limiting potential gains from closing the gender gap in those age categories without simultaneously raising employment of older men.
Closing the gender gap would, however, require a significant increase in the female employment rate, and therefore a substantive policy effort to achieve this goal. On average, within the OECD, employment rates of working-age women (aged 20 to 64 years) should increase from 67% in 2023 to 81.7% in 2060, that is by 14.7 percentage points (Annex Figure 2.A.5), a massive leap. While the fact that younger cohorts tend to work more than older cohorts should account for about one fourth of this increase (as reflected in the baseline projections), mobilising further the female labour supply to close the gender employment gap would require a genuine, additional jump in employment rates of working-age women by 11.2 percentage points. The required increase in the employment rate of working-age women will be particularly large (more than 25 percentage points) in Türkiye and all Latin American OECD countries, except Chile. Increases of more than 15 percentage points beyond cohort effects will also be necessary in some Southern European countries such as Greece and Italy.
Policy should therefore address the various barriers that are preventing full equality between women and men in the labour market not only for equity purposes but also to contribute to offsetting the effect of demographic ageing. The 2013 OECD Recommendation of the Council on Gender Equality in Education, Employment and Entrepreneurship provides a frame of policy principles to reach this goal (OECD, 2017[79]): recommended policy interventions to foster gender equality range from promoting equal pay practices and gender- and family-friendly policies within firms, facilitating women’s access to science, technology, engineering and mathematics studies, and promoting women’s success in entrepreneurship and decision-making positions to significantly upscaling good-quality and affordable childcare and elderly care – see also OECD (2017[80]; 2023[81]). Many of these are costly policy interventions, however, especially because women take up a disproportionate share unpaid work and family and caring responsibilities (see Box 2.6), which can also act as a barrier on women’s labour force participation – see OECD (2017[80]; 2023[76]).28 Closing the gender employment gap would then require a large increase in unpaid family and care work performed by men and/or significant market or government provision of affordable childcare and elderly care.
The take up of unpaid care work and family responsibilities is very different between men and women – see e.g. OECD (2017[80]). Time use surveys allow quantifying how much time men and women spend in household-related tasks. On average, in the OECD, women spend about four hours per day in unpaid care and domestic work, against about two hours in the case of men. What is more, women spend in these activities more time than men in all OECD countries, ranging from 1.2 times the number of hours spent by men in Belgium and Norway to 5.2 times in Türkiye (see Figure 2.11).
Average time spent by women on unpaid care and domestic work, factor of men’s time, 2023
Note: Female‑male ratio of average daily time spent on unpaid domestic chores and care work. OECD: Weighted average of OECD countries.
Source: Secretariat’s calculations based on OECD Data Explorer, “Gender, Institutions and Development Database (GID-DB) 2023”, http://data-explorer.oecd.org/s/257.
Time use surveys, however, underestimate the gender gap in domestic workload. The reason is that, even when they have a job, women spend more time “on call”, that is being ready to face unexpected household occurrences during the day that require action from one adult member of the household (e.g. a sick child who must be picked up from school). This is reflected in gender differences in commuting time and preference for flexible work arrangements – see e.g. Mas and Pallais (2017[82]) and Farré, Jofre‑Monseny and Torrecillas (2023[83]). Additionally, and perhaps more important, women take up an overwhelming share of cognitive domestic work, a taxing but often invisible form of work, including for example anticipating needs, identifying options for filling them, making decisions, and monitoring progress. Evidence suggests that women take up a disproportionate share of cognitive household labour in all areas but especially anticipating needs and monitoring (Daminger, 2019[84]; Reich-Stiebert, Froehlich and Voltmer, 2023[85]). Effectively increasing labour market participation of women, without deteriorating women’s well-being, requires therefore a significant reorganisation of family responsibilities, by changing gender norms and redesigning the perimeter of domestic work.
Raising employment rates of old-age people in good health could lead to large gains in GDP per capita growth
Finally, in many OECD countries, older individuals in good health represent a significant potential source of labour resources, which is still insufficiently activated. In particular, while employment of older workers has significantly increased for workers below 60 years of age, progress can be made in many countries for older age categories (OECD, 2020[8]; 2023[4]). Employment rates across OECD countries indeed vary by a factor of four for the 60‑64 age group, and almost by a factor of seven for those aged between 65 and 69 years – see Chapter 3.29 To the extent that the cross-country variation of life expectancy in good health at 65 years does not show the same degree of cross-country variation (OECD, 2024[5]),30 this suggests significant room for improvement.
For the purpose of this chapter, an alternative higher-employment-of-older-workers scenario is constructed. In this scenario, it is assumed that, by 2060, each country would reduce the employment exit rate of older people at least to the level of the best 10% of OECD countries in each age category above the 50‑54 year‑olds.31 The reference values used are those for the gender for which the employment rate is highest.32 To avoid confusing this channel of improvement with the closure of the gender gap (and therefore double counting), gender gaps in employment for each age category are assumed to remain as in the baseline scenario.33
By reducing the employment exit rate of older people to that of the best 10% of OECD countries in each age category (that is, in the higher-employment-of-older-workers scenario, as described above), about half of the OECD countries could gain at least 0.2 percentage point of annual GDP per capita growth (Figure 2.12) with respect to the baseline scenario, close to the aggregate gain for the OECD area.34 Not surprisingly, compared with the baseline scenario, employment rates could increase in this scenario by about 10 percentage points for workers aged 60 years or more, and by a mere 3.5 percentage points for those aged from 55 to 59 years (Annex Figure 2.A.8).35
In a handful of countries, which are already close to the benchmark exit rate for all, or almost all, age categories, potential gains from further activating older workers are limited. This is especially the case in Chile, Colombia, Iceland, Israel, Japan, Korea, Mexico and New Zealand – all countries where, already in 2023, more than 20% of people aged 65 years or more had a job, and which have relatively high employment rates among old-age people in other age categories. For all these countries, potential gains in employment rates are limited, and even in the best-case scenario, employment rates among older people are likely to fall (see Annex Figure 2.A.8). This reflects changes in the composition of the top age category, with a large projected increase of the share of very old people. As a result, potential GDP per capita growth gains from mobilising further labour supply of older people are limited to less than 0.05 percentage points (see Figure 2.12).
On the other side of the spectrum, in ten continental European countries, including Austria, Belgium, France, Czechia, Greece, Italy, Luxembourg, the Slovak Republic, Slovenia and Spain, the potential gain in annual GDP per capita growth from mobilising labour capacity of older workers could be significant. Reducing the exit rate of older people so as to match the 10% best-performing countries could bring growth dividends as large as 0.4 percentage points, or even more (see Figure 2.12). Specifically, in France, Italy and Spain, it could even suffice alone to fully offset the entire projected decline in GDP per capita growth due to demographic ageing (as implied by the baseline scenario). However, the required leap in the employment rates of those aged 55 years or more in these countries would be quite high, especially for those aged between 60 and 64 years. In fact, in this scenario and in all these countries, except Czechia, the Slovak Republic and Spain, employment rates in this age category would be in 2060 at least 20 percentage point higher than in 2023, and up to 50 percentage points higher in Luxembourg (see Annex Figure 2.A.8). While in many of these countries, the cohort effect account for a small fraction of this increase, in all of them more than 70% of it is accounted by the difference in employment rates between the higher-employment-of-older-workers and the baseline scenarios.36 By contrast, in Czechia, the Slovak Republic and Spain, the greatest potential gain comes from reducing exit rates from employment for people aged 65 years of more.
Percentage point difference in average annual GDP per capita growth: baseline vs. older workers scenario, 2024‑60
Notes: The chart compares projected percentage point differences in average real GDP per capita growth rates between a scenario with high employment of older workers and the baseline scenario. The higher-employment-of-older-workers scenario is constructed by assuming that, by 2060, each country would reach in each age category above that of the 50‑54 years old (for the gender with the highest employment rate) at least the employment rate which would make the rate of decline in employment rates above 55 years of age as small as the 10th percentile of the cross-country distribution, as projected in the baseline scenario. For the other gender in each country and age category, projected employment rates are assumed to be such that they maintain the same gender employment gap as in the baseline scenario. OECD: Weighted average of OECD countries. p.p.: percentage points. GDP: Gross Domestic Product.
Source: Secretariat calculations based on OECD Data Explorer, “Employment and unemployment by five‑year age group and sex – levels”, http://data-explorer.oecd.org/s/253 and “Productivity levels”, http://data-explorer.oecd.org/s/254; Fluchtmann, J., M. Keese and W. Adema (2024), “Gender equality and economic growth: Past progress and future potential”, https://doi.org/10.1787/fb0a0a93-en; and United Nations (2024), World Population Prospects 2024, https://population.un.org/wpp/.
Sustaining activity and employment of old-age people in good health requires, nevertheless, a complex and holistic policy approach (OECD, 2020[8]; 2023[4]; 2024[86]). Policy needs to simultaneously enhance incentives to remain in the labour market, sustain employability and labour demand, often by changing employers’ attitudes and enhancing job quality and work organisation within firms, while ensuring that older workers have the right skills to thrive in the labour market and avoid in-work poverty traps. Equipping older workers with skills in demand and addressing barriers to their job mobility will also be important to avoid that increasing the share of old-age employment has negative implications for productivity growth. These issues will be discussed in detail in Chapters 3, 4 and 5. In addition, health itself is not exogenous to policy, but can be influenced through prevention policies and better management of chronic diseases so that they are less of a barrier to employment, while improving people well-being – see e.g. OECD (2022[87]; 2022[88]; 2023[89]; 2025[90]). Appropriate workplace policies are also crucial to ensure that longer working life promote better health – or at very least do not deteriorate it (see Chapter 3).
2.3.3. Combining potential gains from different mobilisation strategies
The various scenarios considered in the previous subsection are complementary, rather than mutually exclusive. Countries may indeed pursue simultaneously multiple strategies to mobilise untapped labour resources. The scenarios considered here are designed in such a way that they can be, in principle, cumulated. Even if, in most countries, none of these alternatives suffices per se, they can be part of an overall strategy, which in most cases could compensate the effect of population ageing on the employment-to-population ratio.37
According to the stylised simulations presented in the previous section, in all countries but Korea, exploiting the full potential of all these channels combined would indeed prevent the employment-to-population ratio from declining by 2060, and in certain countries it would even yield a considerable increase, as shown in Figure 2.13. The largest contribution would come from mobilising further labour market participation and employment of older people. Simultaneously reducing exit rates of older workers to best practices and closing the gender gap at older ages (that is cumulating the higher-employment-of-older-workers and the gender-equality-for-older-workers scenarios) would increase the share of employed persons in the overall population by 4.6 percentage points, twice as much as the effect of closing the gender gap for younger age categories.
However, while the goal of closing the gender gap in employment is relatively consensual within society (OECD, 2017[80]; ILO/OECD, 2024[91]), significantly mobilising labour resources from other channels often faces societal and political resistance, because of their implications for migration inflows (see the previous subsection) and for the length of careers and timing of retirement (Jensen, 2012[92]; Häusermann, Kurer and Traber, 2018[93]; Guardiancich and Guidi, 2020[94]; Bello and Galasso, 2020[95]). Yet, while closing the gender employment gap among prime‑age and young people would avoid the fall in the employment-to-population ratio in the OECD area as a whole in the simulations presented in the previous section, it would not be sufficient at a country level in all but ten countries (see Figure 2.13). Avoiding the decline of the share of employed persons in the population would therefore require, at least partially, mobilising other channels. Combining gender equality for prime‑age and young people with the high migration scenario would barely avoids the drop in employment-to-population ratios in an additional five countries. Closing the gender employment gap at all ages would stabilise the share of employed persons in the population in 17 countries. By contrast, combining the gender-equality (all ages) and the higher-employment-of-older-workers scenarios would stabilise or even further increase the employment-to-population ratio in all but three countries.
Potential gains in employment rates from combining various scenarios, 2024‑60, percentage points
Notes: The chart compares the projected percentage point loss in the share of employment to total population in the baseline scenario with gains from alternative scenarios: higher-employment of older-workers, gender equality (elderly), gender equality (prime‑age and youth), high migration. The higher-employment-of-older-workers scenario is constructed by assuming that, by 2060, each country would reach in each age category above that of the 50‑54 years old (for the gender with the highest employment rate) at least the employment rate which would make the rate of decline in employment rates above 55 years of age as small as the 10th percentile of the cross-country distribution, as projected in the baseline scenario. For the other gender in each country and age category, projected employment rates are assumed to be such that they maintain the same gender employment gap as in the baseline scenario. The gender equality scenarios assume that by 2060, for both genders, employment rates in each age category are as high as that of the gender with the highest rate in the baseline scenario. The high migration scenario sets future net migration rates equal to the 75th percentile of the cross-country distribution in 2021‑24. Proj. empl. to pop. loss: p.p. loss in employment-to-population ratios as projected in the baseline scenario. Y. & P.A.: young and prime‑age. OECD: Weighted average of OECD countries. p.p.: percentage points.
Source: Secretariat’s calculations based on OECD (2024), International Migration Outlook, https://doi.org/10.1787/50b0353e-en; OECD Data Explorer, “Labour market outcomes of immigrants – Employment, unemployment, and participation rates by sex”, http://data-explorer.oecd.org/s/255 and “Employment and unemployment by five‑year age group and sex – levels”, http://data-explorer.oecd.org/s/253; Fluchtmann, J., M. Keese and W. Adema (2024), “Gender equality and economic growth: Past progress and future potential”, https://doi.org/10.1787/fb0a0a93-en; United Nations (2024), International Migrant Stock 2024, www.un.org/development/desa/pd/content/international-migrant-stock, and United Nations (2024), World Population Prospects 2024, https://population.un.org/wpp/.
Stabilising the share of employed persons in the overall population will not be enough, however, to prevent GDP per capita growth from declining. The reason is that, as mentioned in the previous section, the contribution of the labour input to growth has always been positive in most countries until now. For example, in the whole OECD area, employment growth was still contributing as much as 0.3 percentage points to GDP per capita growth in 2006‑19.38 As a consequence, even maintaining a moderate increase in the labour input would likely be insufficient to prevent a growth slowdown. In fact, while in the whole OECD exploiting the full potential of all the labour channels considered in this section could allow cushioning the annual GDP per capita growth loss due to demographic change, as projected in the baseline scenario for the period 2024‑60, this will be insufficient in two‑thirds of OECD countries (Figure 2.14, Panel A). Beyond Ireland and the United States, where GDP per capita is not projected to decline in the baseline scenario, only in Belgium, Canada, Costa Rica, Denmark, France, Greece, Italy, Portugal, Spain and the United Kingdom mobilising all these labour resources could fully prevent GDP per capita growth from declining.39 Again, the largest contribution would come from mobilising further labour market participation and employment of older people in good health. Reducing the employment exit rate of older workers to best practices and closing the gender gap at older ages would increase GDP per capita growth by 0.26 percentage points, twice as much as the gain that would be obtained from closing the gender gap for young and prime‑age workers.40
Projected annual GDP per capita growth and potential gains from combining various scenarios, 2023‑60
Notes: Panel A compares the projected percentage point loss in real GDP per capita growth in the baseline scenario with gains from alternative scenarios: higher employment of older workers, gender equality (elderly), gender equality (prime‑age and youth), high migration. Panel B compares the average GDP per capita growth attainable from a comprehensive strategy mobilising all alternative labour channels at two‑thirds of the potential considered in Panel A and the cross-country median growth rate in 2006‑19.The higher-employment-of-older-workers scenario is constructed by assuming that, by 2060, each country would reach in each age category above that of the 50‑54 years old (for the gender with the highest employment rate) at least the employment rate which would make the rate of decline in employment rates above 55 years of age as small as the 10th percentile of the cross-country distribution, as projected in the baseline scenario. For the other gender in each country and age category, projected employment rates are assumed to be such that they maintain the same gender employment gap as in the baseline scenario. The gender equality scenarios assume that by 2060, for both genders, employment rates in each age category are as high as that of the gender with the highest rate in the baseline scenario. The high migration scenario sets future net migration rates equal to the 75th percentile of the cross-country distribution in 2021‑24. Y. & P.A.: young and prime‑age. OECD: Weighted average of OECD countries. p.p.: percentage points. GDP: Gross Domestic Product.
Source: Secretariat’s calculations based on OECD (2024), International Migration Outlook 2024, https://doi.org/10.1787/50b0353e-en; OECD (2024), OECD Data Explorer: “Employment and unemployment by five‑year age group and sex – levels”, http://data-explorer.oecd.org/s/253, “Labour market outcomes of immigrants – Employment, unemployment, and participation rates by sex”, http://data-explorer.oecd.org/s/255 and “Productivity levels”, http://data-explorer.oecd.org/s/254; Fluchtmann, J., M. Keese and W. Adema (2024), “Gender equality and economic growth: Past progress and future potential”, https://doi.org/10.1787/fb0a0a93-en; United Nations (2024), International Migrant Stock 2024, www.un.org/development/desa/pd/content/international-migrant-stock, and United Nations (2024), World Population Prospects 2024, https://population.un.org/wpp/.
For a number of countries, however, own growth in 2006‑19 may not provide an appropriate benchmark. In fact, some of them were still in a catch-up phase of development characterised by high productivity dynamism and significant growth in the labour force participation of people of working age, which cannot be expected to continue indefinitely at the same pace. Similarly, the three mobilisation scenarios considered here are highly stylised simulations that are provided to give an order of magnitude of the relative potential of each channel. As a result, they do not necessarily represent jointly attainable targets, in particular because of the costly policy interventions that are required to attain each of them. For this reason, Figure 2.14, Panel B, compares projected GDP per capita growth in a scenario incorporating the mobilisation of two‑thirds of untapped labour resources with the annual growth rate of GDP per capita of the entire OECD area in 2006‑19 (1.0%, see Figure 2.7), arguably two more reachable benchmarks. For the OECD area, mobilising these untapped resources to reach just two‑thirds of their potential would allow cushioning 70% of the annual loss in GDP per capita growth due to demographic change as projected in the baseline scenario for the period 2024‑60, reaching a projected annual GDP per capita growth of 0.9%.
The projected growth rate of 13 OECD countries (Colombia, Costa Rica, Czechia, Estonia, Ireland, Israel, Korea, Latvia, Lithuania, Poland, the Slovak Republic, Türkiye and the United States) remains above the 2006‑19 cross-country average even in the baseline scenario, despite significant growth losses for many of them with respect to their own growth rate in the same period. In two other countries (Hungary and Slovenia), growth could be kept higher than the OECD average of 2006‑19 if two‑thirds untapped labour resources from all channels were activated. Moreover, mobilising two‑thirds of untapped labour resources would suffice to keep annual GDP per capita growth within 0.5 percentage points per year from the benchmark in all but nine countries (Austria, Chile, Finland, Greece, Italy, Japan, Luxembourg, Mexico, Norway and Switzerland). The remaining gap can be closed by developing the appropriate reforms to revive productivity growth – see e.g. OECD (2018[54]), Draghi (2024[96]), André and Gal (2024[97]) and Chapter 5 – up to a realistically reachable level. For all these countries, achieving a rate of growth in GDP per person employed that is no more than half of the median for the 1990s for all countries would allow reaching this benchmark growth or remaining within about 0.1 percentage points of it (see Annex Figure 2.A.9).
Overall, these findings suggest that only a comprehensive policy strategy, involving a substantial mobilisation of untapped labour resources and making the most of the employment potential of older people, could avoid that demographic change brings about a significant decline GDP per capita growth. Any other strategy to prevent this decline would not only risk falling short of the target but would also imply a significant burden on younger cohorts, as a smaller working-age labour pool will have to produce more just to maintain living standards of a larger dependent population. This would have significant fairness implications and would jeopardise societal cohesion. The next section will delve into these issues, by looking at the extent and dynamics of intergenerational inequalities.
The results in the previous section show that mobilising untapped labour resources will be key to avoid declining GDP per capita growth and, therefore, stagnating average living standards. Yet, even if OECD countries meet this challenge, demographic change will have profound consequences for the distribution of resources across people of different ages in the population.
As the youngest baby boomers approach retirement, OECD societies with declining working-age populations will have to secure adequate incomes and care for unprecedentedly large generations of retirees. In 2021, old-age and survivors’ pensions accounted for 38% of public social expenditure across OECD countries – approximately 8.5% of GDP (see Annex Figure 2.A.10, Panel A).41 Spending on older people also accounts for about half of all health expenditure, which, at 30% of total public social spending (or 6.6% of GDP), constitutes the second-largest expenditure category.42 For comparison, OECD countries dedicated only 2.1% of GDP to family benefits, and around 1.1% to out-of-work income support for working-age adults through unemployment benefits and social assistance (OECD, 2025[98]). As OECD populations continue to age and live longer, expenditure on pensions and health is projected to rise across OECD countries, by an average of 0.09 percentage points of GDP annually until 2060, or a cumulative 3 percentage points (see Annex Figure 2.A.10, Panel B).43 Alongside implications for intergenerational fairness, this will leave increasingly little fiscal space for benefits targeted at mitigating poverty, insuring against income shocks, and supporting labour market reallocation among the working-age population.
Given the design of social protection systems in most OECD countries, increased calls on publicly financed support and services will need to be funded through the taxes and social security contributions of a shrinking working-age population.44 This shift of resources, from current and future generations of working-age adults towards the dependent old-age population, has strong implications for intergenerational equity.45 Moreover, it will compound the existing inequalities in income and wealth across generations that are documented in this section.
2.4.1. Older generations have benefited from higher income growth than the young
In over two‑thirds of OECD countries where data are available, older working-age people (55‑64) have enjoyed faster income growth than younger working-age people (25‑34) over the past three decades. The gap in equivalised disposable household income between older working-age people and the young has widened in most OECD countries since the mid‑1990s, reaching over 10% in many Nordic and Western European countries in recent years (Figure 2.15, Panel A).46 Equivalised disposable household income assigns household income after tax and transfers, adjusted for household size, to each household member. While this adjustment allows for a more accurate comparison of living standards across different types of households, it also means that rising female labour force participation alone has raised household incomes among the young. Despite this, and improved educational attainment among younger generations (Barro and Lee, 2010[99]; Cipollone, Patacchini and Vallanti, 2014[100]), the divergence in income growth has persisted. Among the few countries where younger people have enjoyed faster income growth, many are in the process of catching up with the most advanced economies, having experienced either strong labour productivity growth (Czechia, Poland and the Slovak Republic) or surging labour force participation (Chile) over the past decades. These results echo earlier research showing that the incomes of older working-age people have grown faster than those of young people across high-income countries (Pancrazi and Guiatoli, 2024[101]; Resolution Foundation, 2018[102]),47 a process driven by both larger employment gains and stronger earnings growth among older people (Bianchi and Paradisi, 2024[103]).
The incomes of people of retirement age are also catching up with those of young working-age people.48 Although across OECD countries the disposable household income of people in their late 20s and early 30s still exceeds that of people aged 65 years or more by 20%, the gap has narrowed by 10 percentage points since the mid‑1990s (Figure 2.15, Panel B). In some countries – such as Luxembourg, Italy, Spain and Norway – older people now enjoy incomes comparable to, or higher than, those of younger people.
Disparities in income trajectories across different age groups shed light on the rising intergenerational inequalities observed in most OECD countries. Each generation faces unique labour market and macroeconomic environments that shape people’s income paths. Workers employed during periods of robust labour productivity growth typically experience higher earnings growth than those employed during periods of productivity slowdown. Starting one’s career during an economic downturn – as the Millennials, who were born after 1980 and graduated in the aftermath of the global financial crisis – can leave lasting “scars”, undermining long-term labour market outcomes (Schmillen and Umkehrer, 2017[104]; Genda, Kondo and Ohta, 2010[105]) and diminishing the propensity to invest in riskier assets (Malmendier and Nagel, 2011[106]). Combined, these factors contribute to disparities in lifetime incomes across generations (Freedman, 2023[107]; Guvenen et al., 2017[108]).
Reading note: Panel A. In Denmark (left-hand panel), equivalised disposable household incomes for older working-age people (55‑64) are currently 40% higher than for younger working-age people (25‑34). In the mid‑1990s, the two groups had similar incomes. Hence, older working-age people have fared better than the young over the period. In Chile (right-hand panel), equivalised disposable household incomes for older working-age people are nearly 20% lower than for younger working-age people while in the 1990s they were 10% higher. Hence, older working-age people have fared worse than the young over the period.
Note: Observation periods vary slightly across countries: mid‑1990s refers to 1995 or the closest available year; recent year refers the latest pre‑COVID year, or 2022/23 if available. Equivalised disposable household income assigns post-tax-and-transfer household income, adjusted for household size, to each household member. Average: Unweighted average of the countries shown.
Source: OECD calculation based on the Luxembourg Income Study (LIS) Database.
Today’s older generations were fortunate to enjoy substantial income growth throughout their working-age years. In all countries where data are available, baby boomers – those born in the 1950s and early 1960s – experienced steep growth in disposable household incomes from their 20s to their mid‑50s (Figure 2.16). This steep increase mirrors sustained labour productivity growth observed in most advanced economies in the 1980s and 1990s, when the baby boomers started their careers (OECD, 2024[19]). Their incomes decline closer to retirement age as their skills depreciate (see Chapter 4) and they gradually exit the labour market (see Chapter 3). However, even well into retirement, baby boomers maintain income levels exceeding those they enjoyed in their 30s, thanks to – on average – decent pensions and the return on the savings they were able to accumulate during their working years.
While today’s younger generations generally enjoy higher real incomes than their predecessors did at the same age, they have seen slower income growth over the course of their lives. Across most countries, with few exceptions such as Italy and Spain (Figure 2.16, Panel A), each successive generation has benefitted from higher incomes than previous ones at the same age. However, income trajectories have flattened for generations following the baby boomers (Figure 2.16, Panel B). This trend is particularly pronounced for Millennials who, in many countries, have experienced limited or no income growth in their 20s and 30s. If this slowdown persists, Millennials in some countries may enter middle age with real incomes similar to, or lower than, those of previous generations. In only a third of countries for which data are available do Millennials enjoy higher incomes and similar income trajectories compared to previous generations (Figure 2.16, Panel C).
Two factors explain the flatter income trajectories of Millennials across most OECD countries:
1. Slowing labour productivity growth since the 2000s: Most advanced economies have suffered from a slowdown in labour productivity growth since the early 2000s, and particularly after the 2008 global financial crisis (GFC) (Goldin et al., 2024[32]; Fernald, Inklaar and Ruzic, 2024[27]; OECD, 2019[109]; 2024[19]). This deceleration has been more pronounced in Western and Southern Europe, and less so in Central and Eastern Europe, Korea and the United States, where productivity growth continued to be robust in recent years (OECD, 2019[109]). As highlighted above in this chapter, it appears unlikely that labour productivity growth will return to the high rates of the past in advanced economies. Millennials could therefore continue to face flatter income trajectories throughout their lifetimes.
2. Scarring effects from compounded economic downturns: Millennials were disproportionately affected by the GFC. Job losses were concentrated among younger workers, who were more likely to hold temporary and precarious contracts (Carcillo et al., 2015[110]). As a result, the share of young people neither employed nor in education or training (NEET) increased sharply and took several years to return to pre‑GFC levels in several OECD countries. Moreover, young people who secured their first jobs during the GFC earned lower wages than earlier cohorts (Schwandt and von Wachter, 2019[111]; Altonji, Kahn and Speer, 2016[112]). In many European countries the subsequent Eurozone crisis extended challenging labour market conditions for the young well into the mid‑2010s, resulting in a persistent earnings gap between Millennials and previous generations (Freedman, 2023[107]; Bentolila et al., 2021[113]).49 These challenges may result in lower pension entitlements and diminished savings capacity in old age for Millennials compared to earlier generations, while their expected expenses in old age may rise due to longer life expectancies (see Figure 2.1).
Age trajectories in real median equivalised disposable household incomes (2017 USD PPP), by generation
Note: For each generation-age group, median incomes are computed by averaging across all survey years from the 1990s to the latest available year in which a generation is observed at a given age. To improve robustness, only generation-age groups that appear in at least two survey years are included; survey years where the generation-age group represents a larger share of the population receive a greater weight. For some countries, given limited LIS data for the most recent years, Millennials are observed only up to age 35. Equivalised disposable household income assigns post-tax-and-transfer household income, adjusted for household size, to each household member. Incomes are expressed in 2017 PPP-adjusted USD.
Source: OECD calculations based on the LIS database. Consumer Price Index data are already available in the LIS database and come from the International Comparison Programme (ICP) at the World Bank.
2.4.2. Intergenerational disparities in household wealth are also large, as older generations have benefited from booming asset markets
Wealth, a critical determinant of households’ living standards and long-term economic well-being, tends to be concentrated among older generations. While income represents the ongoing flow of financial resources – through earnings from work or capital rents – wealth is the cumulative stock of past savings and asset appreciation. Wealth not only has an immediate consumption value, for example when it takes the form of housing or collateral. It also has an insurance value providing a buffer against unforeseen income shocks or consumption needs. When invested, wealth can generate future income through interest payments and dividends. Across 19 OECD economies, households headed by individuals aged 55‑65 and 65+ hold, on average, net wealth six times greater than those headed by 25‑34 year‑olds, though with large cross-country differences (Figure 2.17, Panel A).50
Evidence on long-term wealth distribution trends, though limited by data availability, indicates widening intergenerational disparities. Over the past half-century, wealth‑to‑income ratios have surged across OECD countries, driven by historic gains in equity and housing markets (OECD, 2020[114]). The wealth of homeowners now accounts for 95% of national wealth on average (Balestra, Caisl and Hermida, 2025[115]), with ownership concentrated among older cohorts who purchased homes during periods of affordability (Knoll, Schularick and Steger, 2017[116]; Causa, Woloszko and Leite, 2019[117]). In five of the six OECD countries with comparable data since the early‑2000s (Spain, Italy, Australia, Germany and Canada), the wealth gap between older and younger generations has grown; only the United States shows a decline (Figure 2.17, Panel B).51
The recent tendency towards limited later-life dissaving, and delayed bequests, risk calcifying the concentration of wealth among older people (Balestra and Tonkin, 2018[118]). Standard lifecycle models (Modigliani and Brumberg, 1954[119]) posit an accumulation of wealth during working years, followed by a period of dissaving during retirement, a broad pattern that is indeed observed in household wealth distribution data (Cowell et al., 2017[120]). However, contrary to the model’s predictions of post-retirement dissaving, older people have increasingly tended to retain wealth in their later years. Such wealth retention is likely driven by three main motivations:
1. To insure against longevity risks as longer lifespans and potential health and long-term care costs necessitate precautionary savings;
2. To provide bequests as older people retain wealth for bequests rather than to use it for consumption; and
3. To retain assets as housing market frictions and incentives to capitalise on property appreciation delay decumulation (French, Jones and McGee, 2023[121]).
A growing literature on this “retirement savings puzzle” finds that while self-insurance against long-term care, or uninsured medical expenses, dominates for most households, the bequest motive drives the saving among the wealthiest (De Nardi et al., 2025[122]). For example, in the United States, while 53% of retirees save primarily for precautionary reasons, the wealthiest 20% are motivated by bequests. Since wealth is highly concentrated, the bequest motive dominates for the majority of wealth.
Among younger cohorts, those without access to bequests and inter vivos gifts increasingly face barriers to wealth accumulation. The share of people in their 30s who own their main residence has fallen in 17 out of 24 OECD countries over the last three decades, with declines exceeding 15 percentage points in Ireland, Greece, the United Kingdom, Spain and Australia (Figure 2.18, Panel A). Lower-income young people have been particularly affected (Figure 2.18, Panel B). In the United Kingdom, between 2000 and 2017, the gap in homeownership rates between those who grew up in rented accommodation and those who grew up in owner-occupied homes has doubled (Eyles, Elliot Major and Machin, 2022[123]).52
Falling homeownership among young cohorts reflects multiple pressures: rising house prices have made ownership increasingly unaffordable for young households while at the same time mortgage access has tightened; economic and labour market conditions for young people have become less secure while the increasing cost of rent and high student debt repayments in some OECD countries has reduced their capacity to save (OECD, 2020[114]). Even inheritances – traditionally a transfer of wealth from the old to the young – are expected to arrive later in life, if they arrive (Bourquin, Joyce and Sturrock, 2020[124]), too late to aid home purchases and instead further boosting wealth levels during later working age.
Note: Results refer to the distribution of financial and non-financial assets and liabilities across households, with no adjustment made for differences in household size. Public and occupational pension wealth is not considered; wealth held in voluntary private schemes is included. Observation periods vary slightly across countries: 2000s refers to 2000 or the closest available year; recent year refers the latest pre‑COVID year, or 2022/23 if available. Median net wealth expressed in 2017 USD PPP. In Panel A, median population wealth will be higher in economies with larger older (and wealthier) cohorts. As a result, normalising by the population median will underemphasise the extent to which net wealth is concentrated among older cohorts. Average: Unweighted average of the countries shown.
Source: OECD calculations based on the Luxembourg Wealth Study (LWS) database. Panel A adapted from Balestra, Caisl and Hermida (2025[115]), “Household wealth distribution in OECD countries: trends and gaps”, using LWS data.
Note: Ownership rates include those who are paying a mortgage and those who own their home outright. Observation periods vary slightly across countries: mid‑1990s refers to 1995 or the closest available year; recent year refers the latest pre‑COVID year, or 2022/23 if already available. Average: Unweighted average of the countries shown.
Source: OECD calculations based on the LIS database.
2.4.3. The poverty risk has shifted from the old to the young
While analyses of median income trajectories and wealth accumulation patterns provide critical insights into broader trends in economic well-being across generations, substantial disparities in living standards persist, warranting special focus on the least advantaged. Previous OECD research (OECD, 2017[7]) has documented a notable shift in poverty risk between age groups from the mid‑1980s to the early 2010s, with the burden moving away from older people towards younger and prime‑aged individuals – particularly those under 25. Recent data indicate that poverty rates for older people (65+) have declined since the mid‑2000s in most OECD countries, with 18 out of 32 countries now reporting lower poverty rates for this group compared to children (under 18) (Figure 2.19).53 Notably, in 14 of these countries even the “oldest old” (76+), who traditionally face higher poverty risk, experience lower poverty rates than children. However, old-age poverty remains prevalent in some countries, such as the Baltic states and Korea, often disproportionately affecting women (OECD, 2023[4]).
Note: Poverty rates give the share of people living in households with an equivalised disposable household income of less than 50% of the national median. Equivalised disposable household income assigns post-tax-and-transfer household income, adjusted for household size, to each household member. Observation periods in Panel B vary slightly across countries: early‑2000s refers to 2000 or the closest available year; recent year refers the latest pre‑COVID year, or 2022/23 if already available. For the sake of comparability, the years used for countries with information on imputed rents are the same as in Annex Figure 2.A.11. Average: Unweighted average of the countries shown.
Source: OECD calculation based on the LIS Database.
The shift in poverty risk from older to younger generations reflects interrelated demographic, labour market, and policy developments. First, the rise in female labour force participation and growing employment of older workers has enabled older generations to accumulate longer contribution histories and greater pension entitlements. Second, pension reforms implemented since the 1990s, while raising normal retirement ages to address fiscal sustainability concerns, included sometimes long transition periods as well as “grandfathering” clauses that shield older workers from benefit reductions and tighter eligibility conditions. As a result, current retirees and older generations of workers have been less affected than future generations will be (OECD, 2019[125]; 2021[126]). Third, the shift in poverty risk may reflect the preferences of an ageing electorate and align with broader public perceptions that have traditionally viewed older people among the groups most deserving of public support (Heuer and Zimmermann, 2020[127]), though recent evidence suggests these attitudes may be evolving – see Hoynes, Joyce and Waters (2024[128]) for a discussion of social attitudes in the United Kingdom.
Moreover, conventional income‑based poverty metrics may underestimate true disparities in living standards between low-income households of different ages. This is because they do not account for differences in non-discretionary spending, notably housing – the largest expenditure item, on average, for households across all income groups (OECD, 2019[129]).54 With declining homeownership rates among younger generations (Figure 2.18), a growing share of young adults rent their homes, often at substantial cost. According to the OECD Affordable Housing Database, housing costs in the rental market exceed those experienced by owner-occupiers in 34 of 37 OECD countries even when principal repayment is included in mortgage costs (OECD, 2024[130]). The concentration of young people in expensive metropolitan areas for study or work further exacerbates intergenerational cost disparities. When combining all forms of tenure, households where the oldest member is aged between 25‑34 spend on average 14 percentage points more of their disposable income on rent or mortgage costs than households where the oldest member is older than 65 (OECD, 2025[131]). Consequently, the share of young people living with parents has been on the rise in some OECD countries (OECD, 2020[114]), delaying independent living, geographic mobility, and family formation.55
Accounting for the financial benefits of homeownership further widens measured disparities in poverty risk between older people and the young. Once imputed rent – i.e. the economic value homeowners derive from living in their own properties instead of renting at market value56 – is added to homeowners’ disposable income, poverty rates among older people decline, to 10.9% across 24 OECD countries with available data (see Annex Figure 2.A.12, Panel A).57 By contrast, poverty rates modestly rise for children (to 14.8%) and working-age people (to 10.7%), as the relative poverty threshold increases due to this adjustment, and because these groups are less likely to live in households who own their homes. When accounting for imputed rent, older people are less like to be in poverty than children in 21 of 24 OECD countries with available data – only in Estonia, Lithuania and Switzerland do older people remain the most vulnerable to poverty. In nine OECD countries (Slovak Republic, Denmark, the Netherlands, Luxembourg, Greece, France, Spain, Italy and Czechia), the poverty rate among older people is less than half that observed among children. And furthermore, examining trends over time, after accounting for imputed rent, relative poverty rates for older people have declined since the turn of the century in 8 of the 14 OECD countries with time series data available (see Annex Figure 2.A.12, Panel B), further widening the intergenerational poverty gap.
The trends described in this chapter – ageing populations, soaring dependency ratios, and a declining labour force called on to finance the needs of a growing dependent population – have profound implications for the distribution of resources between different age groups. Expanding the labour force, and strengthening participation, as outlined in the previous section, will be central to maintaining living standards. Yet beyond this, if countries are to meet anticipated fiscal demands and avoid exacerbating the implications of demographic change on intergenerational fairness, they may need to re‑examine the composition of funding sources and the balance of contributions and benefits across the life cycle. A rebalancing of labour and capital income taxation could help secure the financing of social protection and support employment and wage growth (OECD, 2024[132]). Well-designed inheritance taxes and a greater role for recurrent taxes on immovable property, could stymie further house price increases and help reduce wealth disparities (Causa, Woloszko and Leite, 2019[133]; OECD, 2018[134]; 2021[135]). Meanwhile, rebalancing social expenditure toward early-life investments – including childcare and family support – could alleviate financial pressures on young families and raise living standards, while helping to close the gender employment gap and improving long-term productivity.
The secular increase in longevity and decline in fertility have raised OECD countries dependency ratios and will continue to do so in the future. OECD countries are therefore moving from a regime in which the labour input was contributing to GDP per capita growth to another in which labour input declines will represent a drag on growth in the next 35 years. Without further policy action and/or change in individual labour market attachment, this will weigh significantly on economic growth and the capacity of OECD countries to continue improving their living standards. Moreover, the evidence suggests that demographic change has strong implications for intergenerational inequality.
This chapter has explored, by means of simple simulations of highly stylised, complementary scenarios, the potential of different avenues to address the growth challenges brought about by often dramatic population ageing. It first notes that the perspectives concerning reviving productivity growth are uncertain, and therefore relying exclusively on the hope of a sustained acceleration of productivity growth would be a dangerous strategy. By contrast, mobilising untapped labour resources appears to have a significant potential. Each of the avenues considered here (closing the gender gap, raising employment of older people and increasing migration) could have a significant positive impact on GDP per capita growth.
Increasing employment of older people appears to be one of the labour resource avenues with the largest potential. Reducing their employment exit rate to that of the best 10% of OECD countries and closing the gender employment gap at old age could increase OECD GDP per capita growth by 0.26 percentage points, about 60% of the projected decline of GDP per capita growth in the baseline scenario. Mobilising the employment potential of older people in good health would not only be necessary to effectively offset the effect of ageing on growth but also any alternative strategy would be problematic for fairness reasons as intergenerational inequalities have progressed in favour of older generations.
The remainder of this book will delve deeper into the challenges of ageing for the labour market and the policy solutions. Chapter 3 focuses on policies needed to increase activity and employment rates of old-age people and at the same time ensure that they can thrive in the labour market by accessing good-quality jobs. Beyond making sure that workers have the right incentives to stay in the workforce longer and plan for longer careers, it is crucial to guarantee that labour demand does not decline with age and that older people remain employable. Employers, governments and social partners all have a role in improving lifelong learning, job quality and promoting healthy workplaces to ensure the continuing employability and well-being of older workers. But policy attention and actions should also focus on mid-career workers, since one of the key findings of Chapter 3 is that continuous employment in one’s 50s is a crucial determinant of labour market attachment and employment trajectories in one’s 60s.
Lifelong learning is crucial: Chapter 4 indeed shows that information-processing skills and physical abilities decline with age. Moreover, labour market changes tend to make obsolete some skills held by older cohorts. Training participation and learning-by-doing, however, also decrease with age. Policies need not only to promote lifelong learning and improve training opportunities for older workers but also provide career guidance services to help mid-career and older workers reflect on their professional trajectories and address any skills gaps, especially as some professions allow more easily older workers to stay in the labour market longer.
Addressing skill gaps can also partially reduce the possible negative effect of an ageing structure of employment on productivity, which is feared by many (see Box 2.3). Beyond the fact that that productivity at some point declines with age, one of the reasons of this concern is that older workers are less mobile, so that an ageing workforce could in practice slow down growth-enhancing labour reallocation. Chapter 5 investigates how external mobility, and the related cross-firm reallocation process, raises productivity growth, and the related possibility that a greater share of older workers in employment could imply slowing productivity growth. The chapter finds that there is clear evidence of a steep negative gradient between age and job-to-job mobility, which has slowed, and may continue to slow, aggregate wage and productivity growth. Addressing frictions and bottlenecks affecting this gradient will therefore be important to avoid that workforce ageing has a negative impact on productivity growth.
References
[155] Acemoglu, D., D. Autor and C. Patterson (2024), “Bottlenecks: Sectoral Imbalances and the US Productivity Slowdown”, NBER Macroeconomics Annual, Vol. 38, pp. 153-207, https://doi.org/10.1086/729196.
[145] Acemoglu, D. et al. (forthcoming), “When Big Data Enables Behavioral Manipulation”, American Economic Review: Insights, https://doi.org/10.1257/aeri.20230589.
[42] Acemoglu, D. and P. Restrepo (2021), “Demographics and Automation”, The Review of Economic Studies, Vol. 89/1, pp. 1-44, https://doi.org/10.1093/restud/rdab031.
[41] Acemoglu, D. and P. Restrepo (2017), “Secular Stagnation? The Effect of Aging on Economic Growth in the Age of Automation”, American Economic Review, Vol. 107/5, pp. 174-79, https://doi.org/10.1257/AER.P20171101.
[21] Aghion, P., B. Jones and C. Jones (2017), “Artificial Intelligence and Economic Growth”, NBER Working Paper, No. 23928, National Bureau of Economic Research, Cambridge, MA, https://doi.org/10.3386/w23928.
[112] Altonji, J., L. Kahn and J. Speer (2016), “Cashier or Consultant? Entry Labor Market Conditions, Field of Study, and Career Success”, Journal of Labor Economics, Vol. 34/S1, pp. S361-S401, https://doi.org/10.1086/682938.
[97] André, C. and P. Gal (2024), “Reviving productivity growth: A review of policies”, OECD Economics Department Working Papers, No. 1822, OECD Publishing, Paris, https://doi.org/10.1787/61244acd-en.
[15] André, C., P. Gal and M. Schief (2024), “Enhancing productivity and growth in an ageing society: Key mechanisms and policy options”, OECD Economics Department Working Papers, No. 1807, OECD Publishing, Paris, https://doi.org/10.1787/605b0787-en.
[37] Azoulay, P. et al. (2020), “Age and High-Growth Entrepreneurship”, American Economic Review: Insights, Vol. 2/1, pp. 65-82, https://doi.org/10.1257/AERI.20180582.
[23] Baily, M., E. Brynjolfsson and A. Korinek (2023), Machines of Mind: The Case for an AI-Powered Productivity Boom, Brookings Institution, https://www.brookings.edu/articles/machines-of-mind-the-case-for-an-ai-powered-productivity-boom/.
[139] Balcázar, C. et al. (2017), “Rent-Imputation for Welfare Measurement: A Review of Methodologies and Empirical Findings”, Review of Income and Wealth, Vol. 63/4, pp. 881-898, https://doi.org/10.1111/ROIW.12312.
[29] Baldwin, R. (2012), “Global supply chains: Why they emerged, why they matter, and where they are going”, CEPR Discussion Papers, No. 9103, CEPR, London, https://cepr.org/publications/dp9103.
[115] Balestra, C., J. Caisl and L. Hermida (2025), “Mapping trends and gaps in household wealth across OECD countries”, OECD Papers on Well-being and Inequalities, No. 37, OECD Publishing, Paris.
[154] Balestra, C. and F. Oehler (2023), “Measuring the joint distribution of household income, consumption and wealth at the micro level”, OECD Papers on Well-being and Inequalities, No. 11, OECD Publishing, Paris, https://doi.org/10.1787/f9d85db6-en.
[118] Balestra, C. and R. Tonkin (2018), “Inequalities in household wealth across OECD countries: Evidence from the OECD Wealth Distribution Database”, OECD Statistics Working Papers, No. 2018/01, OECD Publishing, Paris, https://doi.org/10.1787/7e1bf673-en.
[99] Barro, R. and J. Lee (2010), A New Data Set of Educational Attainment in the World, 1950-2010, National Bureau of Economic Research, Cambridge, MA, https://doi.org/10.3386/w15902.
[50] Baudin, T., D. de la Croix and P. Gobbi (2015), “Fertility and Childlessness in the United States”, American Economic Review, Vol. 105/6, pp. 1852-1882, https://doi.org/10.1257/aer.20120926.
[39] Baumol, W. (1993), “Health Care, Education and the Cost Disease: A Looming Crisis for Public Choice”, Public Choice, Vol. 77/1, pp. 17-28, http://www.jstor.org/stable/30027203.
[49] Becker, G. (1960), “An Economic Analysis of Fertility”, Demographic and Economic Change in Developed Countries, Roberts, G. B. (ed.) (New York, NY: National Bureau of Economic Research, Inc.), pp. 209-240.
[70] Beine, M., G. Peri and M. Raux (2023), “International college students’ impact on the US skilled labor supply”, Journal of Public Economics, Vol. 223, p. 104917, https://doi.org/10.1016/j.jpubeco.2023.104917.
[95] Bello, P. and V. Galasso (2020), “The politics of ageing and retirement: Evidence from Swiss referenda”, Population Studies, Vol. 75/1, pp. 3-18, https://doi.org/10.1080/00324728.2020.1841270.
[113] Bentolila, S. et al. (2021), “Lost in recessions: youth employment and earnings in Spain”, SERIEs, Vol. 13/1-2, pp. 11-49, https://doi.org/10.1007/s13209-021-00244-6.
[103] Bianchi, N. and M. Paradisi (2024), Countries for Old Men: An Analysis of the Age Pay Gap, National Bureau of Economic Research, Cambridge, MA, https://doi.org/10.3386/w32340.
[33] Bilal, A. and D. Känzig (2024), “The Macroeconomic Impact of Climate Change: Global vs. Local Temperature”, NBER Working Papers, No. 32450, National Bureau of Economic Research, Cambridge, MA, https://doi.org/10.3386/w32450.
[51] Bisin, A. and T. Verdier (2000), ““Beyond the Melting Pot”: Cultural Transmission, Marriage, and the Evolution of Ethnic and Religious Traits*”, Quarterly Journal of Economics, Vol. 115/3, pp. 955-988, https://doi.org/10.1162/003355300554953.
[73] Boeri, T. et al. (2024), “Pay-as-they-get-in: attitudes toward migrants and pension systems”, Journal of Economic Geography, Vol. 24/1, pp. 63-78, https://doi.org/10.1093/jeg/lbad036.
[38] Bornstein, G. (2021), “Entry and profits in an aging economy: The role of consumer inertia.”, Unpublished.
[124] Bourquin, P., R. Joyce and D. Sturrock (2020), Inheritances and inequality within generations, Institute for Fiscal Studies (IFS) Report R173, London, https://doi.org/10.1920/re.ifs.2020.0173.
[20] Brynjolfsson, E. and T. Mitchell (2017), “What can machine learning do? Workforce implications”, Science, Vol. 358/6370, pp. 1530-1534, https://doi.org/10.1126/science.aap8062.
[110] Carcillo, S. et al. (2015), “NEET Youth in the Aftermath of the Crisis: Challenges and Policies”, OECD Social, Employment and Migration Working Papers, No. 164, OECD Publishing, Paris, https://doi.org/10.1787/5js6363503f6-en.
[151] Carta, F. and M. De Philippis (2023), “The Forward-Looking Effect of Increasing the Full Retirement Age”, The Economic Journal, Vol. 134/657, pp. 165-192, https://doi.org/10.1093/ej/uead051.
[133] Causa, O., N. Woloszko and D. Leite (2019), “Housing, wealth accumulation and wealth distribution: Evidence and stylized facts”, OECD Economics Department Working Papers, No. 1588, OECD Publishing, Paris, https://doi.org/10.1787/86954c10-en.
[117] Causa, O., N. Woloszko and D. Leite (2019), “Housing, wealth accumulation and wealth distribution: Evidence and stylized facts”, OECD Economics Department Working Papers, No. 1588, OECD Publishing, Paris, https://doi.org/10.1787/86954c10-en.
[100] Cipollone, A., E. Patacchini and G. Vallanti (2014), “Female labour market participation in Europe: novel evidence on trends and shaping factors”, IZA Journal of European Labor Studies, Vol. 3/1, https://doi.org/10.1186/2193-9012-3-18.
[53] Cohen, M. and A. Piquero (2015), “Benefits and costs of a targeted intervention program for youthful offenders: The YouthBuild USA offender project”, Journal of Benefit-Cost Analysis, Vol. 6/03, pp. 603-627.
[26] Committee on Automation and the U.S. Workforce (2024), Artificial Intelligence and the Future of Work, National Academies Press, Washington, D.C., https://doi.org/10.17226/27644.
[31] Constantinescu, C., A. Mattoo and M. Ruta (2016), “Does the global trade slowdown matter?”, Journal of Policy Modeling, Vol. 38/4, pp. 711-722, https://doi.org/10.1016/j.jpolmod.2016.05.013.
[40] Cravino, J., A. Levchenko and M. Rojas (2022), “Population Aging and Structural Transformation”, American Economic Journal: Macroeconomics, Vol. 14/4, pp. 479-98, https://doi.org/10.1257/MAC.20200371.
[30] Criscuolo, C. and J. Timmis (2018), “GVC centrality and productivity: Are hubs key to firm performance?”, OECD Productivity Working Papers, No. 14, OECD Publishing, Paris, https://doi.org/10.1787/56453da1-en.
[69] d’Aiglepierre, R. et al. (2020), “A global profile of emigrants to OECD countries: Younger and more skilled migrants from more diverse countries”, OECD Social, Employment and Migration Working Papers, No. 239, OECD Publishing, Paris, https://doi.org/10.1787/0cb305d3-en.
[84] Daminger, A. (2019), “The Cognitive Dimension of Household Labor”, American Sociological Review, Vol. 84/4, pp. 609-633, https://doi.org/10.1177/0003122419859007.
[122] De Nardi, M. et al. (2025), “Why Do Couples and Singles Save during Retirement? Household Heterogeneity and Its Aggregate Implications”, Journal of Political Economy, Vol. 133/3, pp. 750-792, https://doi.org/10.1086/733421.
[75] Docquier, F. and H. Rapoport (2025), “The Vicious Circle of Xenophobia: Immigration and Right-Wing Populism”, IZA Discussion Paper, No. 17754, IZA, Bonn, https://docs.iza.org/dp17754.pdf.
[96] Draghi, M. (2024), The future of European competitiveness: Part B | In-depth analysis and recommendations, https://commission.europa.eu/document/download/ec1409c1-d4b4-4882-8bdd-3519f86bbb92_en?filename=The%20future%20of%20European%20competitiveness_%20In-depth%20analysis%20and%20recommendations_0.pdf.
[43] Eggertsson, G., M. Lancastre and L. Summers (2019), “Aging, Output Per Capita, and Secular Stagnation”, American Economic Review: Insights, Vol. 1/3, pp. 325-42, https://doi.org/10.1257/AERI.20180383.
[123] Eyles, A., L. Elliot Major and S. Machin (2022), Social Mobility: Past, Present and Future, https://www.suttontrust.com/our-research/social-mobility-past-present-and-future/.
[83] Farré, L., J. Jofre-Monseny and J. Torrecillas (2023), “Commuting time and the gender gap in labor market participation”, Journal of Economic Geography, Vol. 23/4, pp. 847-870, https://doi.org/10.1093/jeg/lbac037.
[27] Fernald, J., R. Inklaar and D. Ruzic (2024), “The Productivity Slowdown in Advanced Economies: Common Shocks or Common Trends?”, Review of Income and Wealth, https://doi.org/10.1111/roiw.12690.
[45] Fernández, R. and A. Fogli (2009), “Culture: An Empirical Investigation of Beliefs, Work, and Fertility”, American Economic Journal: Macroeconomics, Vol. 1/1, pp. 146-177, https://doi.org/10.1257/mac.1.1.146.
[24] Filippucci, F. et al. (2024), “The impact of Artificial Intelligence on productivity, distribution and growth: Key mechanisms, initial evidence and policy challenges”, OECD Artificial Intelligence Papers, No. 15, OECD Publishing, Paris, https://doi.org/10.1787/8d900037-en.
[16] Fluchtmann, J., M. Keese and W. Adema (2024), “Gender equality and economic growth: Past progress and future potential”, OECD Social, Employment and Migration Working Papers, No. 304, OECD Publishing, Paris, https://doi.org/10.1787/fb0a0a93-en.
[44] Fluchtmann, J., V. van Veen and W. Adema (2023), “Fertility, employment and family policy: A cross-country panel analysis”, OECD Social, Employment and Migration Working Papers, No. 299, OECD Publishing, Paris, https://doi.org/10.1787/326844f0-en.
[107] Freedman, M. (2023), “Earnings, Cohort Effects, and Inter‐Generational Inequality: Evidence From the Luxembourg Income Study”, Review of Income and Wealth, Vol. 70/2, pp. 278-290, https://doi.org/10.1111/roiw.12641.
[121] French, E., J. Jones and R. McGee (2023), “Why Do Retired Households Draw Down Their Wealth So Slowly?”, Journal of Economic Perspectives, pp. 91–114.
[149] García-Miralles, E. and J. Leganza (2024), “Joint retirement of couples: Evidence from discontinuities in Denmark”, Journal of Public Economics, Vol. 230, p. 105036, https://doi.org/10.1016/j.jpubeco.2023.105036.
[9] Garmany, A. and A. Terzic (2024), “Global Healthspan-Lifespan Gaps Among 183 World Health Organization Member States”, Journal of the American Medical Association Network Open, Vol. 7/12, p. e2450241, https://doi.org/10.1001/jamanetworkopen.2024.50241.
[105] Genda, Y., A. Kondo and S. Ohta (2010), “Long-Term Effects of a Recession at Labor Market Entry in Japan and the United States”, Journal of Human Resources, Vol. 45/1, pp. 157-196, https://doi.org/10.3368/jhr.45.1.157.
[32] Goldin, I. et al. (2024), “Why Is Productivity Slowing Down?”, Journal of Economic Literature, Vol. 62/1, pp. 196-268, https://doi.org/10.1257/jel.20221543.
[60] Grabowski, D., J. Gruber and B. McGarry (2023), Immigration, The Long-Term Care Workforce, and Elder Outcomes in the U.S., National Bureau of Economic Research, Cambridge, MA, https://doi.org/10.3386/w30960.
[28] Grossman, G. and E. Rossi-Hansberg (2008), “Trading Tasks: A Simple Theory of Offshoring”, American Economic Review, Vol. 98/5, pp. 1978-1997, https://doi.org/10.1257/aer.98.5.1978.
[160] Gruijters, R., Z. Van Winkle and A. Fasang (2023), “Life Course Trajectories and Wealth Accumulation in the United States: Comparing Late Baby Boomers and Early Millennials”, American Journal of Sociology, Vol. 129/2, pp. 530-569, https://doi.org/10.1086/726445.
[94] Guardiancich, I. and M. Guidi (2020), “The political economy of pension reforms in Europe under financial stress”, Socio-Economic Review, Vol. 20/2, pp. 817-840, https://doi.org/10.1093/ser/mwaa012.
[138] Guillemette, Y. and J. Château (2023), “Long-term scenarios: incorporating the energy transition”, OECD Economic Policy Papers, No. 33, OECD Publishing, Paris, https://doi.org/10.1787/153ab87c-en.
[141] Guillemette, Y. and D. Turner (2021), “The long game: Fiscal outlooks to 2060 underline need for structural reform”, OECD Economic Policy Papers, No. 29, OECD Publishing, Paris, https://doi.org/10.1787/a112307e-en.
[140] Guillemette, Y. and D. Turner (2018), “The Long View: Scenarios for the World Economy to 2060”, OECD Economic Policy Papers, No. 22, OECD Publishing, Paris, https://doi.org/10.1787/b4f4e03e-en.
[48] Gustafsson, S. (2001), “Optimal age at motherhood. Theoretical and empirical considerations on postponement of maternity in Europe”, Journal of Population Economics, Vol. 14/2, pp. 225-247, https://doi.org/10.1007/s001480000051.
[146] Gustman, A. and T. Steinmeier (2000), “Retirement in Dual‐Career Families: A Structural Model”, Journal of Labor Economics, Vol. 18/3, pp. 503-545, https://doi.org/10.1086/209968.
[108] Guvenen, F. et al. (2017), Lifetime Incomes in the United States over Six Decades, National Bureau of Economic Research, Cambridge, MA, https://doi.org/10.3386/w23371.
[71] Hainmueller, J. and D. Hopkins (2014), “Public Attitudes Toward Immigration”, Annual Review of Political Science, Vol. 17/1, pp. 225-249, https://doi.org/10.1146/annurev-polisci-102512-194818.
[120] Hamilton, K. and C. Hepburn (eds.) (2017), Wealth, Top Incomes and Inequality, Oxford University Press.
[35] Hanushek, E. et al. (2025), “Age and cognitive skills: Use it or lose it”, Science Advances, Vol. 11/10, https://doi.org/10.1126/sciadv.ads1560.
[93] Häusermann, S., T. Kurer and D. Traber (2018), “The Politics of Trade-Offs: Studying the Dynamics of Welfare State Reform With Conjoint Experiments”, Comparative Political Studies, Vol. 52/7, pp. 1059-1095, https://doi.org/10.1177/0010414018797943.
[68] Hermans, K. et al. (2020), “Migration and homelessness: measuring the intersections”, European Journal of Homelessness, Vol. 14/3, pp. 13-34.
[127] Heuer, J. and K. Zimmermann (2020), “Unravelling deservingness: Which criteria do people use to judge the relative deservingness of welfare target groups? A vignette-based focus group study”, Journal of European Social Policy, Vol. 30/4, pp. 389-403, https://doi.org/10.1177/0958928720905285.
[128] Hoynes, H., R. Joyce and T. Waters (2024), “Benefits and tax credits”, Oxford Open Economics, Vol. 3/Supplement_1, pp. i1142-i1181, https://doi.org/10.1093/ooec/odad022.
[91] ILO/OECD (2024), Women at Work in G20 countries: Progress and policy action in 2023, G20 Employment Working Group, https://www.ilo.org/publications/women-work-g20-countries-progress-and-policy-action-2023.
[12] IMF (2025), World Economic Outlook 2025, International Monetary Fund, Washington, D.C., https://www.imf.org/en/Publications/WEO/Issues/2025/04/22/world-economic-outlook-april-2025?cid=ca-com-homepage-SM2025-WEOEA2025001#Chapters (accessed on 22 April 2025).
[156] Immervoll, H. (2024), “Financing social protection in OECD countries: Role and uses of revenue earmarking”, OECD Social, Employment and Migration Working Papers, No. 312, OECD Publishing, Paris, https://doi.org/10.1787/0d53155c-en.
[92] Jensen, C. (2012), “Labour market- versus life course-related social policies: understanding cross-programme differences”, Journal of European Public Policy, Vol. 19/2, pp. 275-291, https://doi.org/10.1080/13501763.2011.599991.
[148] Johnsen, J., K. Vaage and A. Willén (2021), “Interactions in Public Policies: Spousal Responses and Program Spillovers of Welfare Reforms”, The Economic Journal, Vol. 132/642, pp. 834-864, https://doi.org/10.1093/ej/ueab053.
[61] Jun, H. and D. Grabowski (2024), “Mental health in nursing homes: The role of immigration in the long-term care workforce”, Social Science & Medicine, Vol. 351, p. 116978, https://doi.org/10.1016/j.socscimed.2024.116978.
[116] Knoll, K., M. Schularick and T. Steger (2017), “No Price Like Home: Global House Prices, 1870–2012”, American Economic Review, Vol. 107/2, pp. 331-353, https://doi.org/10.1257/aer.20150501.
[161] Koutsogeorgopoulou, V. and H. Morgavi (forthcoming), “Population ageing: Fiscal implications and policy responses”, OECD Economics Department Working Papers, OECD Publishing, Paris.
[147] Lalive, R. and P. Parrotta (2017), “How does pension eligibility affect labor supply in couples?”, Labour Economics, Vol. 46, pp. 177-188, https://doi.org/10.1016/j.labeco.2016.10.002.
[22] Lu, C. (2021), “The impact of artificial intelligence on economic growth and welfare”, Journal of Macroeconomics, Vol. 69, p. 103342, https://doi.org/10.1016/j.jmacro.2021.103342.
[143] Maestas, N., M. Messel and Y. Truskinovsky (2024), “Caregiving and Labor Supply: New Evidence from Administrative Data”, Journal of Labor Economics, Vol. 42/S1, pp. S183-S218, https://doi.org/10.1086/728810.
[106] Malmendier, U. and S. Nagel (2011), “Depression Babies: Do Macroeconomic Experiences Affect Risk Taking?*”, The Quarterly Journal of Economics, Vol. 126/1, pp. 373-416, https://doi.org/10.1093/qje/qjq004.
[82] Mas, A. and A. Pallais (2017), “Valuing Alternative Work Arrangements”, American Economic Review, Vol. 107/12, pp. 3722-3759, https://doi.org/10.1257/aer.20161500.
[72] McCann, K., M. Sienkiewicz and M. Zard (2023), “The role of media narratives in shaping public opinion toward refugees: A comparative analysis”, Migration Research Series, No. 72, International Organization for Migration (IOM), Geneva.
[119] Modigliani, F. and R. Brumberg (1954), Utility Analysis and the Consumption Function: An Interpretation of Cross-Section Data, MIT Press.
[136] Morgan, D. and M. Mueller (2023), “Understanding international measures of health spending: Age-adjusting expenditure on health”, OECD Health Working Papers, No. 162, OECD Publishing, Paris, https://doi.org/10.1787/043ed664-en.
[74] Moriconi, S., G. Peri and R. Turati (2022), “Skill of the immigrants and vote of the natives: Immigration and nationalism in European elections 2007–2016”, European Economic Review, Vol. 141, p. 103986, https://doi.org/10.1016/j.euroecorev.2021.103986.
[153] Mullan, K., H. Sutherland and F. Zantomio (2011), “Accounting for Housing in Poverty Analysis”, Social Policy and Society, Vol. 10/4, pp. 471-482, https://doi.org/10.1017/s1474746411000224.
[36] National Research Council (2012), Aging and the Macroeconomy: Long-Term Implications of an Older Population, Washington, DC: National Academies Press, https://doi.org/10.17226/13465.
[46] Newson, L. and P. Richerson (2009), “Why Do People Become Modern? A Darwinian Explanation”, Population and Development Review, Vol. 35/1, pp. 117-158, https://doi.org/10.1111/j.1728-4457.2009.00263.x.
[90] OECD (2025), Does Healthcare Deliver?: Results from the Patient-Reported Indicator Surveys (PaRIS), OECD Publishing, Paris, https://doi.org/10.1787/c8af05a5-en.
[131] OECD (2025), OECD Affordable Housing Database, https://www.oecd.org/en/data/datasets/oecd-affordable-housing-database.html.
[98] OECD (2025), OECD Social Expenditure Database (SOCX), https://www.oecd.org/en/data/datasets/social-expenditure-database-socx.html.
[78] OECD (2025), Promoting Better Career Mobility for Longer Working Lives in Austria, Ageing and Employment Policies, OECD Publishing, Paris, https://doi.org/10.1787/db85473f-en.
[158] OECD (2024), Affordable Housing Database - HM1.4 Living arrangements by age groups, https://webfs.oecd.org/Els-com/Affordable_Housing_Database/HM1-4-Living-arrangements-age-groups.pdf.
[14] OECD (2024), How’s Life? 2024: Well-being and Resilience in Times of Crisis, OECD Publishing, Paris, https://doi.org/10.1787/90ba854a-en.
[56] OECD (2024), International Migration Outlook 2024, OECD Publishing, Paris, https://doi.org/10.1787/50b0353e-en.
[132] OECD (2024), Megatrends and the Future of Social Protection, OECD Publishing, Paris, https://doi.org/10.1787/6c9202e8-en.
[130] OECD (2024), OECD Affordable Housing Database - HC1.2. Housing Costs over Income, https://webfs.oecd.org/Els-com/Affordable_Housing_Database/HC1-2-Housing-costs-over-income.pdf.
[19] OECD (2024), OECD Compendium of Productivity Indicators 2024, OECD Publishing, Paris, https://doi.org/10.1787/b96cd88a-en.
[6] OECD (2024), OECD Economic Outlook, Volume 2024 Issue 2, OECD Publishing, Paris, https://doi.org/10.1787/d8814e8b-en.
[3] OECD (2024), OECD Employment Outlook 2024: The Net-Zero Transition and the Labour Market, OECD Publishing, Paris, https://doi.org/10.1787/ac8b3538-en.
[86] OECD (2024), Promoting Better Career Choices for Longer Working Lives: Stepping Up Not Stepping Out, Ageing and Employment Policies, OECD Publishing, Paris, https://doi.org/10.1787/1ef9a0d0-en.
[5] OECD (2024), Society at a Glance 2024: OECD Social Indicators, OECD Publishing, Paris, https://doi.org/10.1787/918d8db3-en.
[25] OECD (2024), “Using AI in the workplace: Opportunities, risks and policy responses”, OECD Artificial Intelligence Papers, No. 11, OECD Publishing, Paris, https://doi.org/10.1787/73d417f9-en.
[157] OECD (2023), Beyond Applause? Improving Working Conditions in Long-Term Care, OECD Publishing, Paris, https://doi.org/10.1787/27d33ab3-en.
[152] OECD (2023), Health at a Glance 2023: OECD Indicators, OECD Publishing, Paris, https://doi.org/10.1787/7a7afb35-en.
[89] OECD (2023), Integrating Care to Prevent and Manage Chronic Diseases: Best Practices in Public Health, OECD Publishing, Paris, https://doi.org/10.1787/9acc1b1d-en.
[58] OECD (2023), Introduction Measures for Newly-Arrived Migrants, Making Integration Work, OECD Publishing, Paris, https://doi.org/10.1787/5aeddbfe-en.
[76] OECD (2023), Joining Forces for Gender Equality: What is Holding us Back?, OECD Publishing, Paris, https://doi.org/10.1787/67d48024-en.
[2] OECD (2023), OECD Employment Outlook 2023: Artificial Intelligence and the Labour Market, OECD Publishing, Paris, https://doi.org/10.1787/08785bba-en.
[4] OECD (2023), Pensions at a Glance 2023: OECD and G20 Indicators, OECD Publishing, Paris, https://doi.org/10.1787/678055dd-en.
[55] OECD (2023), “Proposal for an action plan to reduce early school leaving in Spain”, OECD Education Policy Perspectives, No. 71, OECD Publishing, Paris, https://doi.org/10.1787/0c249e7a-en.
[81] OECD (2023), Reporting Gender Pay Gaps in OECD Countries: Guidance for Pay Transparency Implementation, Monitoring and Reform, Gender Equality at Work, OECD Publishing, Paris, https://doi.org/10.1787/ea13aa68-en.
[88] OECD (2022), Guidebook on Best Practices in Public Health, OECD Publishing, Paris, https://doi.org/10.1787/4f4913dd-en.
[162] OECD (2022), OECD Employment Outlook 2022: Building Back More Inclusive Labour Markets, OECD Publishing, Paris, https://doi.org/10.1787/1bb305a6-en.
[18] OECD (2022), OECD Employment Outlook 2022: Building Back More Inclusive Labour Markets, OECD Publishing, Paris, https://doi.org/10.1787/1bb305a6-en.
[87] OECD (2022), Promoting Health and Well-being at Work: Policy and Practices, OECD Health Policy Studies, OECD Publishing, Paris, https://doi.org/10.1787/e179b2a5-en.
[135] OECD (2021), Inheritance Taxation in OECD Countries, OECD Tax Policy Studies, No. 28, OECD Publishing, Paris, https://doi.org/10.1787/e2879a7d-en.
[67] OECD (2021), Language Training for Adult Migrants, Making Integration Work, OECD Publishing, Paris, https://doi.org/10.1787/02199d7f-en.
[17] OECD (2021), OECD Employment Outlook 2021: Navigating the COVID-19 Crisis and Recovery, OECD Publishing, Paris, https://doi.org/10.1787/5a700c4b-en.
[126] OECD (2021), Pensions at a Glance 2021: OECD and G20 Indicators, OECD Publishing, Paris, https://doi.org/10.1787/ca401ebd-en.
[114] OECD (2020), Housing and Inclusive Growth, OECD Publishing, Paris, https://doi.org/10.1787/6ef36f4b-en.
[8] OECD (2020), Promoting an Age-Inclusive Workforce: Living, Learning and Earning Longer, OECD Publishing, Paris, https://doi.org/10.1787/59752153-en.
[109] OECD (2019), OECD Compendium of Productivity Indicators 2019, OECD Publishing, Paris, https://doi.org/10.1787/b2774f97-en.
[1] OECD (2019), OECD Employment Outlook 2019: The Future of Work, OECD Publishing, Paris, https://doi.org/10.1787/9ee00155-en.
[77] 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.
[129] OECD (2019), Under Pressure: The Squeezed Middle Class, OECD Publishing, Paris, https://doi.org/10.1787/689afed1-en.
[125] OECD (2019), “Will future pensioners work for longer and retire on less?”, OECD Publishing, Paris, https://www.oecd.org/en/publications/will-future-pensioners-work-for-longer-and-retire-on-less_0fa49b9b-en.html.
[54] OECD (2018), Good Jobs for All in a Changing World of Work: The OECD Jobs Strategy, OECD Publishing, Paris, https://doi.org/10.1787/9789264308817-en.
[134] OECD (2018), The Role and Design of Net Wealth Taxes in the OECD, OECD Tax Policy Studies, No. 26, OECD Publishing, Paris, https://doi.org/10.1787/9789264290303-en.
[79] OECD (2017), 2013 OECD Recommendation of the Council on Gender Equality in Education, Employment and Entrepreneurship, OECD Publishing, Paris, https://doi.org/10.1787/9789264279391-en.
[66] OECD (2017), Making Integration Work: Assessment and Recognition of Foreign Qualifications, Making Integration Work, OECD Publishing, Paris, https://doi.org/10.1787/9789264278271-en.
[65] OECD (2017), Making Integration Work: Family Migrants, Making Integration Work, OECD Publishing, Paris, https://doi.org/10.1787/9789264279520-en.
[142] OECD (2017), OECD Employment Outlook 2017, OECD Publishing, Paris, https://doi.org/10.1787/empl_outlook-2017-en.
[7] OECD (2017), Preventing Ageing Unequally, OECD Publishing, Paris, https://doi.org/10.1787/9789264279087-en.
[80] OECD (2017), The Pursuit of Gender Equality: An Uphill Battle, OECD Publishing, Paris, https://doi.org/10.1787/9789264281318-en.
[64] OECD (2016), Making Integration Work: Refugees and others in need of protection, Making Integration Work, OECD Publishing, Paris, https://doi.org/10.1787/9789264251236-en.
[13] OECD (2011), How’s Life?: Measuring Well-being, OECD Publishing, Paris, https://doi.org/10.1787/9789264121164-en.
[62] OECD (forthcoming), International Migration Outlook 2025, OECD Publishing, Paris.
[10] OECD/European Commission (2024), Health at a Glance: Europe 2024: State of Health in the EU Cycle, OECD Publishing, Paris, https://doi.org/10.1787/b3704e14-en.
[101] Pancrazi, R. and G. Guiatoli (2024), “Age-Income Gaps”, Warwick Economics Research Papers No. 1504.
[34] Prskawetz, A. and T. Lindh (eds.) (2006), The impact of population ageing on innovation and productivity growth in Europe, Vienna: Vienna Institute of Demography.
[59] Rapp, T. and J. Sicsic (2020), “The contribution of the immigrant population to the U.S. long-term care workforce”, Social Science & Medicine, Vol. 263, p. 113305, https://doi.org/10.1016/j.socscimed.2020.113305.
[85] Reich-Stiebert, N., L. Froehlich and J. Voltmer (2023), “Gendered Mental Labor: A Systematic Literature Review on the Cognitive Dimension of Unpaid Work Within the Household and Childcare”, Sex Roles, Vol. 88/11-12, pp. 475-494, https://doi.org/10.1007/s11199-023-01362-0.
[102] Resolution Foundation (2018), International comparisons of intergenerational trends, https://www.resolutionfoundation.org/app/uploads/2018/02/IC-international.pdf.
[104] Schmillen, A. and M. Umkehrer (2017), “The scars of youth: Effects of early‐career unemployment on future unemployment experience”, International Labour Review, Vol. 156/3-4, pp. 465-494, https://doi.org/10.1111/ilr.12079.
[52] Schochet, P., J. Burghardt and S. McConnell (2006), National Job Corps study and longer-term follow-up study: Impact and benefit-cost findings using survey and summary earnings records data, Mathematica Policy Research, Princeton, NJ, https://wdr.doleta.gov/research/FullText_Documents/National%20Job%20Corps%20Study%20and%20Longer%20Term%20Follow-Up%20Study%20-%20Final%20Report.pdf (accessed on 28 May 2018).
[111] Schwandt, H. and T. von Wachter (2019), “Unlucky Cohorts: Estimating the Long-Term Effects of Entering the Labor Market in a Recession in Large Cross-Sectional Data Sets”, Journal of Labor Economics, Vol. 37/S1, pp. S161-S198, https://doi.org/10.1086/701046.
[137] Sengupta, M. et al. (2022), Post-acute and Long-term Care Providers and Services Users in the United States, 2017–2018, National Center for Health Statistics (U.S.), https://doi.org/10.15620/cdc:115346.
[47] Sobotka, T. (2008), “Overview Chapter 6: The diverse faces of the Second Demographic Transition in Europe”, Demographic Research, Vol. 19, pp. 171-224, https://doi.org/10.4054/demres.2008.19.8.
[57] Spielvogel, G. and M. Meghnagi (2018), “Assessing the role of migration in European labour force growth by 2030”, OECD Social, Employment and Migration Working Papers, No. 204, OECD Publishing, Paris, https://doi.org/10.1787/6953a8ba-en.
[159] United Nations (2025), World Fertility Report 2024, United Nations, https://www.un.org/development/desa/pd/sites/www.un.org.development.desa.pd/files/undesa_pd_2025_wfr-2024_advance-unedited.pdf (accessed on 15 April 2025).
[63] United Nations (2024), International Migrant Stock 2024, Online Edition, https://www.un.org/development/desa/pd/content/international-migrant-stock.
[11] United Nations (2024), World Population Prospects 2024, Online edition, United Nations, https://population.un.org/wpp/.
[150] United Nations (2007), Indicators of Sustainable Development: Guidelines and Methodologies - Methodology Sheets, United Nations, Department of Economic and Social Affairs, https://www.un.org/esa/sustdev/natlinfo/indicators/methodology_sheets.pdf.
[144] Van Houtven, C., N. Coe and M. Skira (2013), “The effect of informal care on work and wages”, Journal of Health Economics, Vol. 32/1, pp. 240-252, https://doi.org/10.1016/j.jhealeco.2012.10.006.
Percentage point increase in the ratio of old-age to working-age population, probability intervals, 2023‑60
Notes: The chart compares projected percentage point differences in the ratio of old-age to working-age population between 2060 and 2023, based on UN probabilistic projections. The UN medium projection is the median projection for each country. A rate within the 80% probability interval is projected to have a 80% probability of occurring. p.p.: percentage points. OECD: Weighted average of OECD countries.
Source: Secretariat calculations based on United Nations (2024), World Population Prospects 2024, Department of Economic and Social Affairs, Population Division, Online Edition, https://population.un.org/wpp/.
GDP per capita growth vs. other indicators
Notes: Gross household income is computed in real terms and incorporates income from non-profit institutions serving households. GDP: Gross Domestic Product. BLI: OECD Better Life Index (aggregate index). p.p.: percentage points.
Source: Secretariat’s calculations based on OECD Data Explorer, “Productivity levels”, http://data-explorer.oecd.org/s/254; “Household Indicators Dashboard”, http://data-explorer.oecd.org/s/258; OECD (2024), How’s Life 2024, https://doi.org/10.1787/90ba854a-en.
Average annual growth rate of real GDP per person employed, 1991‑2000 and change between 1991‑2000 and 2011‑19
Note: Average growth in the 1990s based on 1992‑2000 for Czechia, 1993‑2000 for Costa Rica, Hungary and Mexico, 1994‑2000 for Poland, 1995‑2000 for Estonia and the Slovak Republic, 1996‑2000 for Latvia, Lithuania and Slovenia. Average growth in the 2010s based on 2011‑17 for Australia and 2013‑19 for Greece. p.p.: percentage points. GDP: Gross Domestic Product.
Source: OECD Data Explorer, “Productivity levels”, http://data-explorer.oecd.org/s/254.
2021‑24 and 2024‑60, annual averages, percentage of the population in the previous year
Notes: Net migration rates are defined as the difference in the change of the stock of foreign-born and that of domestic-born but living abroad, divided by population size in year t‑1. Countries ordered in ascending order by average net migration rate in 2021‑24. Projected rates are based on the UN medium scenario.
Source: Secretariat’s calculations from United Nations (2024), International Migrant Stock 2024, Online Edition, www.un.org/development/desa/pd/content/international-migrant-stock, and United Nations (2024), World Population Prospects 2024, https://population.un.org/wpp/.
Percentage point difference in employment rates between each scenario (baseline and gender equality scenarios) between 2060 and 2023, women aged between 20 and 64 years
Notes: The chart compares, for female workers aged between 20 and 64 years, projected percentage point differences between employment rates in 2023, the baseline scenario in 2060, and a scenario that closes the gender gap in employment rates (Gender equality) by that date. ER: Employment rate. OECD: Weighted average of OECD countries. p.p.: percentage points.
Source: Secretariat calculations based on OECD Data Explorer, “Employment and unemployment by five‑year age group and sex – levels”, http://data-explorer.oecd.org/s/253; Fluchtmann, J., M. Keese and W. Adema (2024), “Gender equality and economic growth: Past progress and future potential”, https://doi.org/10.1787/fb0a0a93-en; and United Nations (2024), World Population Prospects 2024, https://population.un.org/wpp/.
Employment rate profiles of different age categories above 50‑54 years old
Note: The chart represents the profile of employment rates by age category for a hypothetical country (Country A) whose employment rates in some of the age categories above 50‑54 years old are, in the baseline scenario, higher than the average in the countries with exit rates below the first decile of the cross-country distribution (or, equivalently, with a ratio of employment rates in the relevant age category to those in the category 50‑54 years old above the 90% percentile). Straight lines indicate employment rates profiles in the baseline scenario. The dashed line indicates the employment rate profile in the higher-employment-of-older-workers scenario.
Percentage point difference in avg. GDP per capita growth: baseline vs. older workers scenarios, 2023‑60
Notes: The chart compares projected percentage point differences in average annual GDP per capita growth rates between the baseline scenario and two scenarios with high employment of older workers: one i) accounting for the difference in hours worked by older workers relatively to the average worker (scenario including intensive margin) and another ii) not considering differences in hours worked (considered in the main text). The higher-employment-of-older-workers scenario is constructed by assuming that, by 2060, each country would reach in each age category above that of the 50‑54 years old (for the gender with the highest employment rate) at least the employment rate which would make the rate of decline in employment rates above 55 years of age as small as the 10th percentile of the cross-country distribution, as projected in the baseline scenario. For the other gender in each country and age category, projected employment rates are assumed to be such that they maintain the same gender employment gap as in the baseline scenario. OECD: Weighted average of the countries shown. p.p.: percentage points. GDP: Gross Domestic Product.
Source: Secretariat calculations based on OECD Data Explorer, “Employment and unemployment by five‑year age group and sex – levels”, http://data-explorer.oecd.org/s/253, “Average usual weekly hours worked on the main job”, http://data-explorer.oecd.org/s/259 and “Productivity levels”, http://data-explorer.oecd.org/s/254; Fluchtmann, J., M. Keese and W. Adema (2024), “Gender equality and economic growth: Past progress and future potential”, https://doi.org/10.1787/fb0a0a93-en; and United Nations (2024), World Population Prospects 2024, https://population.un.org/wpp/.
Percentage point difference in employment rates between each scenario (baseline and older workers scenarios) between 2060 and 2023, by age category
Notes: The chart compares projected percentage point differences between employment rates in 2023, the baseline scenario in 2060, and a scenario that increases the employment rate of older workers. The alternative higher-employment-of-older-workers scenario is constructed by assuming that, by 2060, each country would reach in each age category above that of the 50‑54 years old (for the gender with the highest employment rate) at least the employment rate which would make the rate of decline in employment rates above 55 years of age as small as the 10th percentile of the cross-country distribution, as projected in the baseline scenario. For the other gender in each country and age category, projected employment rates are assumed to be such that they maintain the same gender employment gap as in the baseline scenario. OECD: Weighted average of OECD countries. p.p.: percentage points.
Source: Secretariat calculations based on OECD Data Explorer: “Employment and unemployment by five‑year age group and sex – levels”, http://data-explorer.oecd.org/s/253; Fluchtmann, J., M. Keese and W. Adema (2024), “Gender equality and economic growth: Past progress and future potential”, https://doi.org/10.1787/fb0a0a93-en; and United Nations (2024), World Population Prospects 2024, https://population.un.org/wpp/.
Annual GDP per capita growth and potential gains from different strategies, 2024‑60
Notes: The chart compares, on the one hand, the average real GDP per capita growth attainable from an alternative comprehensive strategy both mobilising all alternative labour channels at two‑thirds of their potential and reviving productivity growth and, on the other hand, the cross-country mean growth rate in 2006‑19. Productivity is supposed to be revived up to the level that the country growth of GDP per person employed would reach half of the 1991‑2000 cross-country median (the potential gains by attaining this goal are indicated as “Potential gains from prod. increase”). The higher-employment-of-older-workers scenario is constructed by assuming that, by 2060, each country would reach in each age category above that of the 50‑54 years old (for the gender with the highest employment rate) at least the employment rate which would make the rate of decline in employment rates above 55 years of age as small as the 10th percentile of the cross-country distribution, as projected in the baseline scenario. For the other gender in each country and age category, projected employment rates are assumed to be such that they maintain the same gender employment gap as in the baseline scenario. The gender equality scenarios assume that by 2060, for both genders, employment rates in each age category are as high as that of the gender with the highest rate in the baseline scenario. The high migration scenario sets future net migration rates equal to the 75th percentile of the cross-country distribution in 2021‑24. Y. & P.A.: young and prime‑age. OECD: Weighted average of OECD countries. p.p.: percentage points. GDP: Gross Domestic Product.
Source: Secretariat’s calculations based on OECD (2024), International Migration Outlook 2024, https://doi.org/10.1787/50b0353e-en; OECD Data Explorer, “Employment and unemployment by five‑year age group and sex – levels”, http://data-explorer.oecd.org/s/253, “Labour market outcomes of immigrants – Employment, unemployment, and participation rates by sex”, http://data-explorer.oecd.org/s/255 and “Productivity levels”, http://data-explorer.oecd.org/s/254; Fluchtmann, J., M. Keese and W. Adema (2024), “Gender equality and economic growth: Past progress and future potential”, https://doi.org/10.1787/fb0a0a93-en; United Nations (2024), International Migrant Stock 2024, www.un.org/development/desa/pd/content/international-migrant-stock, and United Nations (2024), World Population Prospects 2024, https://population.un.org/wpp/.
Note: Panel A allocates public social expenditure on old-age and survivors’ pensions to older individuals. To obtain age‑specific health expenditure shares relative to GDP it applies the relative spending shares presented by Morgan and Mueller (2023[136]) to the public health expenditure data provided in the OECD Social Expenditure Database (2025[98]). For countries where data on age‑specific public health expenditure are not available but which provide information on public and private health expenditure concurrently, it assumes that the age distributions of public and private health expenditure are identical. Health expenditure figures include most of countries’ public expenditure on long-term care. For countries in which health expenditure data by age group excludes long-term care, the share of long-term care expenditure allocated to those aged above 65 years is assumed to match this group’s share of long-term care beneficiaries. If that information is also missing the OECD unweighted average is used instead. Panel B displays historical (2010‑29) and projected (2024‑60) annual percentage point growth of health and pension expenditure as a share of GDP. Countries ordered by the difference between historical and projected annual changes. OECD represents the unweighted average of OECD countries in Panel A, while in Panel B it represents part of the OECD area corresponding to the 33 countries with available data.
Source: Panel A uses OECD calculations based on the Social Expenditure Database, OECD Historical population data; OECD Health Statistics; US Census Bureau Population Vintage 2019 national population estimates; Morgan and Mueller (2023[136]), “Understanding international measures of health spending: Age‑adjusting expenditure on health”, https://doi.org/10.1787/043ed664-en; Sengupta et al. (2022[137]), “Post-acute and long term care providers and services users in the United States 2017‑18”, https://dx.doi.org/10.15620/cdc:115346; Panel B uses projections from the OECD Economics Department Long-Term Model, see Guillemette and Château (2023[138]), “Long-term scenarios: incorporating the energy transition”, https://doi.org/10.1787/153ab87c-en, for a summary of the approach.
Percentage of the average wage, 2023
Note: Data refer to 2023 policies. Results for a two‑earner couple with two children aged 2 and 3 respectively. The first earner earns 100% of the average wage and the partner earns 67% of the average wage. The category “Childcare Benefits” includes entitlements related to the use of childcare facilities as well as benefits that aim to provide support to those looking after children at home. This category includes also any fee discounts for childcare users applied in particular family circumstances. The group “Changes in taxes” refers to the change in taxes related to having children, i.e. comparing the taxes paid by a couple without children to those of an identical couple with children. “Changes in other benefits” includes changes in housing and family allowances related to the use of childcare. Information on net childcare costs is currently missing for Türkiye. Average: Unweighted average of the countries shown.
Source: OECD TaxBEN model, version 2.7.1, “OECD calculator of taxes and benefit”, www.oecd.org/en/data/tools/oecd-calculator-of-taxes-and-benefits.html.
Note: Poverty rates give the share of people living in households with an equivalised disposable household income of less than 50% of the national median. Equivalised disposable household income assigns post-tax-and-transfer household income, adjusted for household size, to each household member. Observation periods in Panel B vary slightly across countries: early‑2000s refers to 2000 or the closest available year; recent year refers the latest pre‑COVID year, or 2022/23 if already available. Gross imputed rents are an estimate of the value that homeowners would pay in the rental market for their home. Most countries estimate imputed rents using a regression or stratification method based on actual rents (i.e. a prediction of the price associated to a house based on the rents paid for similar dwellings), except for CHL, COL, ITA and MEX that use the self-assessment method (i.e. the self-reported price that households believe they would pay to rent their own houses), and CZE, EST and SVK that employ the user-cost method (i.e. the rate of return to investing home equity in an interest-bearing account). See Balcázar et al., (2017[139]) for a review of rent imputation methods. Net imputed rents subtract property tax and mortgage interest payments from gross values. For countries for which net imputed rents cannot be calculated, gross values are used instead. The country rankings shown in this figure should be interpreted with caution, because of differences in imputation methods and limited data availability. Average: Unweighted average of the countries shown and for which net imputed rents were available.
Source: OECD calculation based on the LIS Database.
← 1. According to the medium scenario of UN population projection, the overall population will be larger in 2060 than in 2023 in half of the OECD countries and in the OECD as a whole (United Nations, 2024[11]).
← 2. There is, however, a large uncertainty about future fertility trends (United Nations, 2025[159]). Yet, the whole 95% probability interval around medium projections remains well below the level required for population replacement.
← 3. Usually, a total fertility rate of 2.1 is considered necessary for population replacement without positive net migration rates – see United Nations (2007[150])
← 4. In fact, it will increase steeply until 2050, to stabilise afterwards in most countries.
← 5. The overall dependency ratio (not shown in the chart) is typically defined as the ratio of the sum of the population aged less than 15 years and 65 years and more to the population aged 15‑64 years – see United Nations (2007[150]).
← 6. The medium scenario corresponds to the median of UN probabilistic projections. Other population projections, such as those of Eurostat or national statistical agencies, provide quantitatively different results although pointing in the same direction (OECD, 2023[4]). Here UN projections are preferred for comparability reasons.
← 7. Formally, , where Pop is the overall population, E is total employment, H are total hours worked, WAPop is the working-age population, is hourly labour productivity, h are average hours worked, ER is the employment rate and DR is the dependency ratio.
← 8. Especially due to the increase over time of female labour market participation.
← 9. Fluchtmann, Keese and Adema (2024[16]) take cohort-specific observed labour force participation and unemployment rates as starting points and derive future rates by assuming constant future unemployment rates and that, for each age and gender category, future labour market entry and exit rates are equal to the most recent rates of the same age category (averaged over a 5‑year period and excluding the years of the COVID‑19 crisis). The labour market entry rate is defined as the percentage change in non-employment rates from age t to age t+1 for the same cohort, when negative and zero otherwise (that is the percentage decrease when it is effectively a decrease). The labour market exit rate is defined as the percentage change in employment rates from age t to age t+1 for the same cohort, when this is negative and zero otherwise (that is, again, the percentage decrease when it is effectively a decrease). By construction, therefore, the effect of future health and education improvements and possible structural reforms that might affect entry and exit rates are not taken into account. To obtain final employment rate projections, projected age‑ and gender-specific employment rates are then combined with UN projections on population size and structure (medium scenario). Then, in Figure 2.6, the direct effect of changes in the population structure is obtained by assuming constant age‑ and gender-specific employment rates as in 2023; the cohort-replacement effect is measured as the difference between the baseline projection (which incorporates age‑ and gender-specific employment rate projections) and the constant-employment rate projection.
← 10. In Colombia, Spain and Israel, the cohort-replacement channel has a significant negative contribution due to a significant drop in the labour force of the younger cohorts in the second half of the 2010s.
← 11. 0.7% per year in the case of the whole OECD area. This assumption could be considered prima facie realistic in the long run. Recent OECD work in this area, including the OECD long-term projection model, indeed assumes constant returns to scale and constant capital-labour ratios in the long run – see e.g. Guillemette and Turner (2018[140]; 2021[141])) and Fluchtmann, Keese and Adema (2024[16]).
← 12. More recent years are excluded from the calculation of benchmark productivity trends to avoid the influence of the COVID‑19 crisis. The period 2006‑19 is chosen to cover a full business cycle and to have start and end points corresponding to similar stages of the business cycle. Robustness exercises (available upon request) using different periods (2007‑19, 2006‑18 and 2007‑18) show similar results – that is a projected drop of OECD GDP per capita of about 40%, and a projected growth reduction in all OECD countries except Ireland and the United States.
← 13. IMF (2025[12]) projects a larger slowdown (0.5 percentage points for the whole world for the period 2025‑50, and more for advanced economies). However, this is by and large due to the fact that 2016‑18, which is a period of fast labour input growth during the expansionary phase of the previous business cycle, is taken as comparison. Using the same period as benchmark, the projection method used here would deliver qualitatively similar results.
← 14. From a dynamic point of view the decline in GDP per capita growth is projected to start around 2030 and peak in 2045‑50 in most countries and the OECD area. Results are available from the Secretariat upon request.
← 15. Analysing whether reducing working hours per worker implies an increase of living standards is beyond the scope of this chapter. For an extensive discussion of working hours and living standards see OECD (2021[17]; 2022[18]).
← 16. The reason for the productivity slowdown is not consensual in the literature. Deceleration in human capital accumulation, physical capital investment and trade expansion, and technological bottlenecks leading to imbalances across sectors are among the cited explanations – see e.g. Goldin et al. (2024[32]) and Acemoglu, Autor and Patterson (2024[155]).
← 17. Other researchers are, however, less optimistic about the potential of AI in terms of aggregate productivity growth – see e.g. Acemoglu et al. (forthcoming[145]).
← 18. For example, Colombia, Costa Rica, Greece, Italy, Korea, Lithuania, Mexico, Spain and Türkiye, where more than 15% of youth aged 15 to 29 years were NEET in 2023-24 (source: https://www.oecd.org/en/topics/youth-employment-and-social-policies.html).
← 19. In certain countries, some untapped labour resources can also be found among jobless prime‑age men. From a macroeconomic perspective, however, the potential contribution of this channel is minor. In fact, the cross-country variance of employment rates of prime‑age men is small, suggesting limited potential gains from boosting employment of this group – see e.g. OECD (2017[142]; 2018[54]).
← 20. Defined as the difference between the change in the stock of foreign-born in the host country between year t‑1 and t and the change in the stock of native‑born living abroad between t‑1 and t, both expressed in percentage the overall size of the population in the host country in t‑1.
← 21. Projected net migration rates are lower in two‑thirds of the OECD countries with positive net migration rates in 2021‑24 (Annex Figure 2.A.4) and two‑thirds of those with positive rates in 2016‑20 (United Nations, 2024[63]). They are also lower in 55% of the OECD countries with positive net migration rates in 2010‑15.
← 22. Or their competences are not recognised (OECD, 2017[66]).
← 23. Especially because long-term trends in employment gaps between natives and foreign-born represents one key area of monitoring and integration policy action, which is left out of this analysis.
← 24. In these simulation exercises, the increase in net migration flows is redistributed proportionally to the projected net migration flows by gender and age category entailed by the baseline scenario after shifting the minimum of the distribution to zero.
← 25. The overall projected gains for the whole OECD area are smaller (0.10%), due to limited or no gains in some large countries such as Mexico.
← 26. Constant growth of GDP per person employed is among the maintained assumptions. This implies that the simplified gender equality scenario considered here entails a lower incidence of part-time work among women entering employment, as the female part-time rate is currently higher than in the overall population in all OECD countries – see e.g. the Statistical Annex of this publication.
← 27. In particular, in these two countries in 2023, the ratio of female employment rates between those aged 60 to 64 years and those aged 50 to 54 years were more than three times larger than the same ratio for their male peers (see OECD Data Explorer, “Employment and unemployment by five‑year age group and sex – levels”, http://data-explorer.oecd.org/s/253).
← 28. Overcoming gender disparities in childcare and elderly care is also important to reduce the exit rate of women from the labour market (Van Houtven, Coe and Skira, 2013[144]; Maestas, Messel and Truskinovsky, 2024[143]) and therefore close the gender gap in employment rates at old age – see also Chapter 3. Yet, retirement policies also play a crucial role for the gender employment gap in these age categories, since married couples tend to take joint retirement decisions, with spouses typically anticipating or delaying their exit from employment to match their husbands’ retirement (Gustman and Steinmeier, 2000[146]; Lalive and Parrotta, 2017[147]; Johnsen, Vaage and Willén, 2021[148]; Carta and De Philippis, 2023[151]; García-Miralles and Leganza, 2024[149]).
← 29. By contrast, they vary only by a factor of two in the age category 55‑59 years, and by much less than that if Türkiye is excluded.
← 30. In particular, the cross-country coefficient of variation of life expectancy in good health at 65 years is almost half of that of employment rates in the 65‑69 age category.
← 31. That is, by 2060, each country would reach in each age category above the 50‑54 year‑olds at least that employment rate that would reduce the decline in employment rates above 55 years to that at the 10th percentile of the cross-country distribution, as projected in the baseline scenario. More precisely, in the higher-employment-of-older-workers scenario, for each country and age category, the employment rate of the gender with the highest employment rate by 2060 in the baseline scenario is obtained in multiple steps: i) for each country and gender, the ratios between the employment rate of each age category above 54 years (55‑59 years, 60‑64 years and 65 years and above) and the employment rate of those aged between 50 and 54 years in the same country is computed; ii) for each country and relevant age category, the gender with the highest ratio is selected and its ratio is taken; iii) the 90th percentile of the cross-country distribution of the retained ratios is then computed for each age category above 54 years – these ratios will be called “target ratios” hereafter; and finally, iv) for each country and age category, the employment rate of the gender with the highest employment rate is multiplied by the corresponding “target ratio”, thereby reducing the exit rate from employment of this group, except if the employment rate of this group is higher in the baseline scenario (in which case it is left unchanged). This is a neater benchmark than using the top percentiles of the cross-country distribution of employment rates that confounds gender, and even prime‑age employment shifts, with employment changes due exclusively to the exit rate from employment of older workers. Note that countries with the best retained ratios are not necessarily those with the highest employment rates (although often this is the case). In certain countries, the employment rate of a given age category maybe above that of the average of the countries with retained ratios above the 90th percentile, and yet the employment rate would increase because the projected exit rate would decrease in this scenario (see Annex Figure 2.A.6). Note also that net exit rates from employment are positive also for workers aged 45‑54 years – see Chapter 3. However, they are typically very small until age 55‑59 years. For this reason, and to simplify the text, age category 50‑54 years is taken as the reference category, and the term older workers refer to workers aged 55 years and more, as standard. Nevertheless, repeating the exercise using age category 40‑44 as reference category would yield similar results.
← 32. In the baseline scenario, there are some countries and age category in which women are projected to have in 2060 a marginally higher employment rate than men.
← 33. As the projection method concerns only the gender with the highest employment rate, the employment rate of the other gender is simply obtained by applying the gender gap as projected in the baseline scenario.
← 34. As data on hours worked by age and gender are not available (or not comparable) in all OECD countries (see Box 2.5), this scenario assumes that any increase in employment of older workers would not alter the growth of GDP per employed person. This is unrealistic because older workers work fewer weekly hours. However, as shown in Annex Figure 2.A.7, for the countries for which data are available, recalculating the contribution of higher employment of older workers by assuming that each new older worker would work the same number of hours as the average of currently employed peers in the same age and gender category results in only a small reduction of the simulated potential gains from this channel – about 10% for the whole OECD area.
← 35. An additional small boost can already be expected by the fact that younger cohorts tend to work more than older cohorts, especially in the case of women (cohort effect). This simple cohort effect would already result in greater employment rates in 2060 than in 2023 in the baseline scenario, particularly for those aged between 55 and 64 years (Annex Figure 2.A.8). For those above 65 years, the overall cohort effect will be compounded by the change in the age composition of this category, and employment rates are projected to fall in the baseline scenario in many countries, especially those with already relatively high employment rates.
← 36. A large increase in employment rates in this age category is also implied for Türkiye in this scenario. However, the simple cohort effect, as included in the baseline scenario, would account for more than half of it.
← 37. It must be kept in mind, however, that these scenarios are highly stylised and presented here as a metric to grasp the magnitude of the potential of each channel. They should not be viewed as a precise forecast.
← 38. Cf. the growth rates of GDP per capita and GDP per person employed in Figure 2.7 above.
← 39. In the case of Greece and Italy, however, projected GDP per capita growth rates would remain negative.
← 40. Note that, as mentioned above, given the assumptions on productivity growth, there is a one‑to‑one correspondence of the relative effects of different channels on the employment-to-population ratio and GDP per capita growth.
← 41. The OECD Social Expenditure Database (SOCX) currently provides data on public social expenditures up to 2023, but breakdowns by spending category still only reach up to 2021 for most countries (OECD, 2025[98]).
← 42. Across the 18 OECD countries for which public health expenditures can be allocated by age, 11 devote more than half to people aged above 65 years, while only in Australia less than 45% of public health expenditure go to older people (Morgan and Mueller, 2023[136]).Health expenditure figures include most of countries’ public expenditure on long-term care, which accounts for about 1.8% of GDP across OECD countries on average but reaches more than 3% of GDP in countries such as the Netherlands, Norway, Sweden and Belgium (OECD, 2023[152]).
← 43. See also Koutsogeorgopoulou and Morgavi (forthcoming[161]).
← 44. In the two decades leading up to the COVID‑19 crisis, revenues from social contributions as a share of GDP have increased in the OECD area, from 8.4% in 2000 to 8.9% in 2019, while financing a declining share of social spending, from 47.5% down to 43.5% (Immervoll, 2024[156]).
← 45. At the same time, as population ageing undermines the sustainability of pay-as-you-go social protection financing, younger generations face declining benefit generosity. Future retirees in many OECD countries are projected to receive lower pension replacement rates than current retirees due to reforms already legislated (OECD, 2019[125]). Publicly funded long-term care systems are under-resourced in many countries (OECD, 2023[157]). Younger generations therefore confront a dual burden: financing the benefits for today’s large pensioner cohort while simultaneously accumulating private savings to compensate for less generous public benefits in their own retirement.
← 46. For data on income levels and inequality across OECD countries over long time horizons, see also the OECD Income Distribution Database, www.oecd.org/en/data/datasets/income-and-wealth-distribution-database.html.
← 47. Pancrazi and Guiatoli (2024[101]) cover a large set of OECD and non-OECD countries for the period from 2004 to 2018. Their methodological approach slightly differs from the one employed in this section in that Pancrazi and Guiatoli focus on individual rather than equivalised disposable household income, and use slightly different age brackets and average incomes across several years.
← 48. In some countries, such as Australia and Switzerland, retirees are allowed to take all or part of their mandatory pension entitlements as a lump sum when they retire. Because LIS data do not follow people over time, such lump sums cannot be smoothed across retirement years in the analysis. In any given survey year, retirees who have already withdrawn will therefore appear to have lower incomes. As the results in Figure 2.16 average across all people aged 65 or older, mixing retirees withdrawing that year with those who withdrew earlier, the net impact on the mean income of retirees is likely to be modest. By contrast, the impact on poverty statistics is potentially large (see Section 2.4.3 below).
← 49. Young people were also disproportionately affected by the COVID‑19 crisis, and their employment rates recovered more slowly than those of prime‑aged and older workers (OECD, 2022[18]).
← 50. The wealth measure used here follows earlier OECD analysis in excluding any wealth held in occupational pension schemes, though wealth in voluntary private pension schemes is included (Balestra and Tonkin, 2018[118]). Wealth differences across age groups would be larger still if occupational pensions were considered.
← 51. The wealth gap between older and younger generations has grown also in the United States when considering only the period up to 2019.
← 52. There is some evidence that within-generation wealth inequalities have risen in some countries. In the United States, the poorest Millennials have less wealth than their baby boomer counterparts, but the wealthiest Millennials have more (Gruijters, Van Winkle and Fasang, 2023[160]).
← 53. In some countries, such as Australia and Switzerland, retirees are allowed to take all or part of their mandatory pension entitlements as a lump sum when they retire. Because LIS data do not follow people over time, such lump sums cannot be smoothed across retirement years in the analysis. In any given survey year, retirees who have already withdrawn will therefore appear to have lower incomes. This is likely to overestimate the poverty rates of older people in Figure 2.19 and Annex Figure 2.A.12 for these countries.
← 54. As labour market participation among women has risen, childcare costs increasingly eat into discretionary spending. On average across the OECD, the net childcare costs facing a two‑earner couple with two children under four account for 11% of the average wage, reaching as high as 40% in the United States (Annex Figure 2.A.11, see OECD calculator of taxes and benefits, https://www.oecd.org/en/data/tools/oecd-calculator-of-taxes-and-benefits.html). For a broader discussion of the implications of essential expenditures for inequalities in living standards, see Balestra and Oehler (2023[154]).
← 55. In 17 OECD countries, more than 50% of young people in their 20s live with their parents, with the shares exceeding 75% in Korea, Italy, Greece, Spain, Portugal, Ireland and the Slovak Republic (OECD, 2024[158]).
← 56. Some tenants may not rent their properties at market value because they live in subsidised or rent-free housing. However, for seven out of the 24 OECD countries covered in the analysis, LIS data on imputed rents only provide information on homeowners. For better comparability, information on imputed rent for tenants paying below market value was therefore not considered even where available.
← 57. The country rankings shown in Annex Figure 2.A.12 should be interpreted with caution given imperfect harmonisation of imputation methods and limited data availability. For a discussion of the implications of accounting for imputed rent for poverty analysis, see Mullan, Sutherland and Zantomio (2011[153]).