Ada Zakrzewska
OECD Employment Outlook 2025
Can We Get Through the Demographic Crunch?
Abstract
As the workforce in OECD countries gets older, many fear that productivity and economic growth may slow down. This concern is often related to the view that the relationship between individual productivity and age is hump-shaped: increasing with experience at younger ages and then declining for older individuals as physical abilities and skills deteriorate. This chapter first examines three mechanisms that may affect how workers’ productivity changes with age: (i) the evolution of information-processing skills with age, (ii) the potential skills obsolescence of older workers because of labour market changes, and (iii) shifts in the occupational structure that may make it easier for workers to remain productive for longer. The chapter then investigates to what extent workers acquire knowledge through learning-by-doing and training during their careers and provides examples of government policies that could help workers stay productive as they age.
Key findings
As OECD countries are ageing, there is growing concern about the employability and well-being of older workers, who are expected to stay in the labour market for longer and account for an increasingly larger share of the workforce. Many also worry that an ageing workforce could lead to slower productivity growth and reduced economic dynamism. This is because the relationship between individual productivity and age is commonly viewed as hump-shaped: increasing with experience at younger ages and then declining for older individuals as physical abilities and skills deteriorate. Yet, there is considerable uncertainty regarding the age at which productivity starts declining. Moreover, the relationship between age and productivity is not fixed and may change as a result of greater investments in skills during the working lives, or through shifts in the occupational structure towards jobs that allow workers to stay productive for longer. This chapter investigates how the skills and jobs of older workers are related to their productivity and proposes policy solutions that could help individuals keep pace with changing labour market demands as they age.
The key takeaways from the chapter are as follows:
While the differences in skills of older and younger individuals are partly due to generational differences, some patterns, such as the decline in information-processing skills within cohorts over the past decade and slower adjustment to labour market changes among older workers, suggest that age gaps in skills may persist in the future.
Older people (55‑65 year‑olds) have weaker information-processing skills than younger individuals. Literacy and adaptive problem-solving skills are 9% lower among 55‑59 year‑olds and 12% lower among 60‑65 year‑olds compared to 25‑29 year‑olds. The gap in numeracy skills is slightly narrower, with 55‑59 year‑olds scoring 7% lower and 60‑65 year‑olds scoring 10% lower than 25‑29 year‑olds. These age gaps are observed in all OECD countries, although the size of these gaps differs. While the differences in skills across the age groups are partly due to generational differences, such as educational attainment, over the past decade, information processing skills declined within birth cohorts, and this decline was faster for older cohorts. Both findings suggest that age‑related decline in skills also plays a role. For example, over the last ten years, literacy skills declined by 5% among those who were 25‑29 in 2012 and by 9% among those who were 50‑54 in the same year. A similar trend was observed for numeracy skills.
Moreover, workforce adjustment to changing demand in the labour market, so far, has happened predominantly through younger workers. Downsizing occupations in the last decade have experienced larger increases in the average age of workers, which suggests that adjustment to changing economic structure occurs primarily through younger workers. Older workers were also less likely to use new productivity-enhancing technologies, as exemplified by lower use of ICT skills compared to younger individuals in the same occupation. For instance, the age gap in using specialised software ranged from 5% in Norway to 37% in Korea, and the use of programming language was more than 40% lower among older workers in some OECD countries. Slower adaptation to labour market trends and adoption of new productivity-enhancing technologies among older workers, especially as the share of older workers in the workforce grows, may hinder their employability and productivity, potentially weighing on the broader economy.
However, the shift in the occupational structure, away from jobs that require physical work and towards those that value experience, may make it easier for workers to remain productive for longer. While in some occupations, experience is valued and workers’ productivity can increase as they age, other occupations offer little wage growth opportunities. For example, 55‑65 year‑old managers earn over 50% more than 25‑34 year‑old managers, while the difference in wages is only 5% between the two age groups in elementary occupations. Moreover, some occupations require more physical effort than others, making it more difficult for older workers to perform those jobs, leading to lower productivity or even labour market exit. The occupational composition differs significantly across OECD countries, implying that in some countries, there are more jobs that potentially allow workers to remain productive for longer. In general, however, the share of jobs that require daily physical work decreased in most OECD countries over the past decade, while the share of high-skilled occupations, which tend to offer higher returns to experience, increased.
Lifelong learning is crucial for workers to adapt to the changing labour market, but training participation and learning-by-doing decrease with age. In 2023, only a third of 60‑65 year‑olds participated in training, compared to over half of 25‑44 year‑olds. Learning-by-doing, another important but less often studied channel through which individuals develop new skills, also decreases with age, from 62% among 25‑29 year‑olds to 45% for 60‑65 year‑olds. One reason why training participation decreases with age is lower expected returns on such investment for individuals and employers due to shorter remaining working lives. However, recent evidence showing that adults underinvest in training even when returns significantly exceed the costs suggests that individuals may also underestimate the benefits of investment in training or face barriers to training participation. Moreover, if training postpones the decision to retire, the socially optimal level of training participation of older workers may be higher than the one chosen by the individual, given that keeping the individual in the workforce increases tax revenue and reduces pension benefit payments. Both social returns to training in excess of individual returns and underinvestment in training by individuals and firms would justify government intervention to lower the cost of training for older workers and address the barriers to training that they face.
The provision of career guidance and policies to promote participation in adult learning can help workers reflect on their professional journeys and address any skills gaps. Career guidance can assist individuals in assessing whether their current roles and skills align with their ambitions and labour market needs, and in identifying training needs. To increase the use of career advice services among adults, some OECD countries subsidise career advice sessions or even provide them free of charge. This is likely to be even more effective if done as a preventive measure and accompanied by targeted outreach campaigns. Governments can also offer financial incentives for mid-career (40‑54 year‑olds) and older people to participate in adult learning. One strategy is to offer financial support for participation in adult learning, regardless of age, for example, through individual learning accounts. Additional support specifically targeted at mid-career workers may also be beneficial, given that recent research suggests that training rates are particularly sub-optimal for middle‑aged workers. Beyond financial incentives, it is important that training delivery and support services respond to the needs of mid-career and older workers, recognising the skills workers gained through informal on-the‑job learning, offering shorter modular courses, and focusing the content on practical work problems rather than abstract technical concepts.
OECD countries are ageing, and there are concerns about the employability and well-being of older workers who are increasingly expected and encouraged to stay in the labour market for longer. Many also fear that as the workforce gets older, the economy’s capacity to innovate will diminish and productivity growth will slow down (IMF, 2016[1]; André, Gal and Schief, 2024[2]; ESM, 2024[3]). One source of this concern is the commonly held view that the relationship between productivity and age is hump-shaped: increasing with experience at younger ages and then declining for older individuals as physical abilities and skills deteriorate (André, Gal and Schief, 2024[2]). In line with this view, the “demographic dividend” experienced in the past that boosted productivity when the workforce’s average age increased from mid‑30s to mid‑40s may turn into a “demographic deficit” as populations continue to age (Henriksen and Cooley, 2018[4]) – see also Chapters 2 and 5.
The first reason why workers’ productivity could decrease with age is the decline in certain cognitive skills and physical abilities (Kenny et al., 2008[5]; Desjardins and Warnke, 2012[6]). Prior research showed, for example, that older workers had lower numeracy, literacy and problem-solving1 skills than younger individuals (Paccagnella, 2016[7]). Rapid labour market changes such as those due to technological transformation or greening may further reduce the productivity of older workers if the skills they acquired early in life grow obsolete (OECD, 2019[8]; 2024[9]). On the other hand, technological change may also lead to shifts in the occupational structure, away from jobs that require physical work and towards those that value experience, making it easier for workers to remain productive for longer. Section 4.1 of this chapter examines these three mechanisms and their potential impact on the future productivity of older workers.
While the relationship between productivity and age is influenced by broad economic forces – resulting in changing labour market demand – individual choices and government policies do matter in shaping this relationship. Preventing skills decline and ensuring that older workers are equipped with the skills to thrive in a fast-changing labour market can help them stay productive for longer. This requires that the traditional three‑stage (school, work, retirement) life model gives way to a more flexible one where learning and work are intertwined and take place throughout life. Section 4.2 of this chapter explores how workers acquire knowledge both on the job and through training throughout their careers, and Section 4.3 provides examples of government policies that could help boost the skills and productivity of workers as they age.
4.1.1. Information-processing skills decline as individuals grow older
Human capital, reflected in the level of education and skills, is critical for labour market success. It is linked to both individual productivity and labour market outcomes, such as employment and wages, and to aggregate economic performance. Data from the first cycle of the Survey of Adult Skills conducted in 2012 showed that older workers had lower numeracy, literacy and problem-solving skills than younger individuals (Paccagnella, 2016[7]). This section first investigates whether these patterns persist using new data from the second cycle of the Survey of Adult Skills collected in 2023. It then attempts to disentangle the generational differences in skills and the ageing effect to shed light on whether similar skills gaps should be expected in the future.
The 2023 Survey of Adult Skills data confirm that older people (defined in this chapter as those aged 55‑65) have weaker information-processing skills, such as literacy, numeracy and adaptive problem-solving,2 compared to younger individuals.3 In 2023, literacy and adaptive problem-solving skills were 9% lower among 55‑59 year‑olds and 12% lower among 60‑65 year‑olds compared to 25‑29 year‑olds (Figure 4.1, Panels A and C). The gap in numeracy skills was slightly narrower, with 55‑59 year‑olds scoring 7% lower and 60‑65 year‑olds scoring 10% lower than 25‑29 year‑olds (Figure 4.1, Panel B).
Panels A-C show percentage differences in literacy, numeracy and adaptive problem-solving skills, compared to 25‑29 year‑olds, baseline (bar) and controlling for level and field of education (dot). Panel D shows unadjusted literacy scores by age group and country.
Note: “Baseline” estimates are adjusted for differences in gender and place of birth, while “adjusted” estimates are additionally adjusted for level of education and field of education. The sample includes Austria, Flemish Region (Belgium), Canada, Czechia, Denmark, England (United Kingdom), Estonia, Finland, France, Germany, Ireland, Italy, Japan, Korea, the Netherlands, Norway, Poland, the Slovak Republic, Spain, Sweden and the United States. Countries presented in this graph are those that participated in the Survey of Adult Skills in 2012 and 2023 to ensure comparability with the rest of this section. Literacy scores for all countries that participated in the Survey of Adult Skills in 2023 can be found in Annex Figure 4.A.1. *Caution is required in interpreting results due to the high share of respondents with unusual response patterns – see the Note for Poland in OECD (2024[10]), Do Adults Have the Skills They Need to Thrive in a Changing World?: Survey of Adult Skills 2023, https://doi.org/10.1787/b263dc5d-en. Average is the weighted average of the countries shown.
Source: 2023 Survey of Adult Skills.
The gaps in information-processing skills between older and younger workers were recorded in all participating OECD countries, but the level of skills and the size of the age gap differed. The largest gap was observed in Estonia (18%), where 25‑44 year‑olds had one of the highest literacy scores in the OECD, while older individuals lagged behind. In contrast, Sweden had the highest literacy score among older workers and a small age gap in literacy as a result (Figure 4.1, Panel D). Age gaps in numeracy and adaptive problem-solving across countries are similar to those observed for literacy skills.
Lower information-processing skills of older workers should be a source of concern, as they may be a sign of lower productivity. Workers with higher information-processing skills are, on average, more likely to be employed and earn higher wages (Box 4.1) which are, in turn, usually closely related to individual productivity – see e.g. Hellerstein, Neumark and Troske (1999[11]) and Lazear et al. (2023[12]).4 This remains true even when comparing individuals with the same level of education, suggesting that literacy skills capture additional information about the individual’s characteristics, which are rewarded in the labour market and go beyond a simple reflection of their educational attainment.
Individuals with higher literacy skills have, on average, higher employment rates and wages and the relationship becomes stronger with age. While higher employment rates and wages among those with greater skills may be influenced by factors like education, regression analysis suggests that literacy skills capture additional information about the individuals that are rewarded in the labour market and go beyond a simple reflection of their educational attainment.
Workers who have one standard deviation higher literacy skills earn, on average, 7% more at the age of 25‑34 and 14% more at the age of 55‑65, than other individuals with the same level of education, field of study and demographic characteristics. Workers with higher skills are also more likely to be employed than other individuals with the same educational attainment. The difference in the likelihood of being in employment (rather than unemployed or inactive) associated with an increase in literacy skills by one standard deviation is 4 percentage points for 25‑34 year‑olds and 8 percentage points for 55‑65 year‑olds.1
1. These results are from a regression controlling for the interaction of age dummies with years of education and gender and controlling for field of study, place of birth and country fixed effects. However, employment rates, wages and skills may be correlated with unobservable characteristics such as ability or motivation. Therefore, these results should not be interpreted as causal.
The currently observed differences in information-processing skills across the age groups are driven by a combination of cohort (generational) and ageing effects. The cohort (generational) effect refers to the impact of being born in a particular time period on the skills level of the generation due to variations in technology exposure, labour market conditions and government policies during the formative years. For example, younger workers may have higher information-processing skills because they completed more years of education or were exposed to more advanced information and communication technologies during their educational pathways. The ageing effect, on the other hand, describes the decline in certain cognitive abilities, such as memory and information-processing speed, as individuals grow older. If the cohort effect is the only driver of differences across the age groups described above, then workers who are young today should not be expected to have lower information-processing skills as they grow older.5 However, if the ageing effect also plays a role, then the skills of present-day young workers may decline as they age, with potentially negative implications for their productivity.
Comparing individuals with the same level of education and field of study reveals that the gap in information-processing skills between younger and older workers is smaller than when comparing the overall population. This suggests that the cohort (generational) effect partly explains the lower skill levels observed among older workers. Yet, differences in education and field of study alone do not fully account for the skill gap (Figure 4.1, Panels A-C) – indicating that other factors also contribute to older workers’ lower proficiency. One potential explanation is that information-processing skills capture some differences in initial education across the age groups that are not captured by years and field of education alone. For example, the quality of initial education of older people could have been lower, leading older people to have lower skills than younger people with the same number of years and field of education. Exposure to certain technologies might have also facilitated the skill acquisition of present-day younger workers compared to older workers with the same education. Another potential explanation is that lower skill levels among older people could be a sign that skills decrease with age, the ageing effect.
While disentangling these two effects further by comparing age groups in the same year is difficult, the two cycles of the Survey of Adult Skills can be used to follow cohorts over time to track how the skills of people born in the same years have changed over time. For example, the skills of individuals aged 25‑29 in 2012 can be compared to those of 35‑39 year‑olds in 2023. This analysis reveals that skills decreased as cohorts aged, and the effect was stronger amongst older individuals. For example, over the last ten years, numeracy skills declined by 3% among those who were 25‑29 in 2012 and by 6% among those who were 50‑54 in 2012. Literacy skills declined by 5% among those who were 25‑29 in 2012, and by 9% among those who were 50‑54 in 2012. The analysis follows individuals with the same level of education and place of birth, which means that the decline in skills is not due to changes in the educational composition or migration background of the cohort.6
However, it is important to note that both the changes in skills within younger and older cohorts over the last decade were not the same in all countries. For example, in Sweden, no statistically significant decrease in literacy skills was observed in the younger or older cohort (Figure 4.2). In Finland, the literacy skills of those who were 25‑29 years old in 2012 increased in the last decade, while the skills of those who were 50‑54 years old in 2012 decreased. In France and Austria, literacy skills declined over the last decade among younger and older cohorts, but the decline was stronger for older individuals.
While the above‑described changes in skills proficiency will reflect a combination of cohort and time effects, the decline in literacy and numeracy skills was more prominent for older workers than for younger ones, suggesting that the decline in skills with age may be, at least partially, due to ageing. A study using scores of a sample of adults who participated in the Survey of Adult Skills in Germany in 2012 and were retested in 2015 suggests that, when following individuals over time, literacy skills (numeracy skills) peak at the age of 41 (46) but decline afterwards (Hanushek et al., 2025[13]).
The difference in literacy skills over 10 years following cohorts
Note: The estimates are adjusted for differences in level of education and place of birth. White circles/triangles indicated results that are not statistically significant at the 5% level.*Caution is required in interpreting results due to the high share of respondents with unusual response patterns – see the Note for Poland in OECD (2024[10]), Do Adults Have the Skills They Need to Thrive in a Changing World?: Survey of Adult Skills 2023, https://doi.org/10.1787/b263dc5d-en. Average is the weighted average of the countries shown.
Source: 2012 and 2023 Survey of Adult Skills.
The decrease in literacy and numeracy skills over time was more pronounced for those without tertiary education. The literacy and numeracy skills of those who were 50‑54 years old in 2012 and had higher education decreased by 7% and 5%, respectively, over the last 10 years, while for the same age group without tertiary education, the decreases were 10% and 7%. These patterns are consistent with the cumulative (dis-) advantage hypothesis (Matthew effect), which, in the context of skills, predicts that initial skill differences become magnified over the life course. For example, those with higher educational attainment and skills obtain more complex jobs that foster further skill development (Lechner et al., 2021[14]). Indeed, significant differences in how much skills decline over time were observed across occupations (Annex Figure 4.A.2). While the literacy skills of cleaners, machine operators and building workers declined by 9% or more over the last 10 years, in various professional and managerial occupations, the decline was 4% or less. This is in line with the findings using data from repeated surveys of the same individuals in Germany (Hanushek et al., 2025[13]) and consistent with higher skills dispersion in the oldest age groups (i.e. among those who were 45‑49 and 50‑54 in 2012), which means that not all older individuals experienced the same decline in skill levels.
Further research using longitudinal data is needed to better understand the reasons why skills decline with age. One possibility is that adults and their employers increasingly underinvest in training as individuals age, resulting in lower skill levels among older individuals and potentially reducing their productivity. If that is the case, interventions to boost the skills of adults at an older age would be an appropriate policy response.7
However, an alternative explanation is that individuals who achieve higher skill levels early in life enter better-paid occupations, which require stronger information-processing skills and provide adults with opportunities to utilise these skills as they age, preventing skill decline. Adults who follow lower-paid career paths may utilise their information-processing skills less at work, so they are more likely to decline with age. In that case, the decline in skills with age would not necessarily be a result of underinvestment in skills, but rather a consequence of entering certain occupations that require fewer information-processing skills earlier in their careers and would be better addressed through early policy interventions, such as career guidance for youth.
Finally, it is also possible that as workers age, the tasks they perform shift from those that require information-processing skills to those requiring other skills, such as decision-making and managing others. In that case, information-processing skills could decline without necessarily leading to lower productivity. While the Survey of Adult Skills only collects data on information-processing skills, the analysis of tasks, such as “managing” and “influencing”, shows that while the share of workers performing these tasks increased with age in younger age cohorts, potentially compensating for the decline in information-processing skills, this effect was much weaker for older age cohorts (Box 4.2).
When following cohorts over the last decade, the share of workers managing others increased by 8 and 3 percentage points among individuals who were 25‑29 and 30‑34 in 2012, respectively (Figure 2.3). These age groups also experienced the largest increase in the extent to which they influenced others at work, for example, by giving presentations, negotiating and teaching. In contrast, among older age groups, the share of workers managing others slightly decreased over the same time period, and the increase in influencing others was less marked than among younger age groups.
The difference in the share managing others and influencing index over 10 years following cohorts
Note: Results of a regression, “overall” shows the results when controlling for gender, education and country fixed-effects, while “within occupations” includes an additional ISCO 1‑digit control. Influencing others includes teaching, giving presentations, influencing others and negotiating. The change in “managing” is not statistically significant for those who were 40‑44 in 2012 (overall estimate) and the change within occupation is not statistically significant for 35‑39 year‑olds and 40‑44 year‑olds in 2012. The change in “influencing” is not statistically significant for 50‑54 year‑olds in 2012 (both the overall and within occupation estimates). Weighted average of Austria, Flemish Region (Belgium), Canada, Czechia, Denmark, England (United Kingdom), Estonia, Finland, France, Germany, Ireland, Italy, Japan, Korea, the Netherlands, Norway, Poland, the Slovak Republic, Spain, Sweden and the United States.
Source: 2012 and 2023 Survey of Adult Skills.
4.1.2. The adjustment to the changing demand in the labour market happens predominantly through younger workers
Another mechanism that may lead to older workers lacking the skills needed in the labour market and their lower productivity, compared to younger individuals, are changes in the labour market resulting from megatrends such as digitalisation, automation and the green transition. This sub-section investigates how the adjustment to labour market trends and adoption of new productivity-enhancing technologies differ across age groups, highlighting the implications for economic dynamism and productivity as populations age.
Economic changes mean that some jobs disappear while other jobs, requiring a different skillset, are created. For example, in the past, automation displaced bank tellers but created new roles in banking, such as IT specialists to maintain digital banking systems. Moreover, many of the jobs that remain change significantly in terms of tasks and skills they require. In the past, the job of auto mechanics primarily involved using hand tools to diagnose vehicles through physical inspections, while today, this is often done using computerised diagnostic tools, which require workers to have digital skills.
While older workers are more likely to be in a job that matches their educational attainment than those who are new in the labour market, as sorting into well-matched jobs takes time (OECD, 2024[10]),8 they may be more likely to be mismatched in terms of occupations and skills currently needed in the labour market than younger people, who grew up with the new technologies, had a better chance of choosing educational pathways aligned with the current labour market demand and followed school curricula that already reflected the new skills needs.
The first piece of evidence supporting this hypothesis is the increase in the average age of workers in occupations that declined in size over the last decade, which is in line with younger workers having incentives not to enter declining occupations, or to leave them, and older workers being discouraged from exiting them due to the costs associated with occupational mobility. For example, handicraft workers, occupations for which demand decreased by 20%, experienced an increase in the average age by 2.4 years over the last decade (Figure 4.4). In contrast, in ICT professionals and business & administration professionals, occupations that grew the fastest over the last decade, the average age decreased by 0.2 and increased by 0.4 years, respectively.
These findings are consistent with prior research, which showed that in the United States, occupations that shrank became “older” – 1 percentage point contraction between 1980 and 2005 was associated with an increase in average age in the occupation by an additional 0.78 years relative to the mean (Autor and Dorn, 2009[15]).9 Similar patterns were found for routine and middle‑skill jobs (Lewandowski et al., 2020[16]; Green, 2019[17]). The change in average age was shown to be driven by a falling employment share of young workers and rising employment shares of prime‑age and older workers (Autor and Dorn, 2009[15]).
Note: The size of the bubble indicates the ratio of the average wage in the occupation and the median wage across occupations in 2023. Weighted average of Austria, Belgium, Bulgaria, Croatia, Czechia, Denmark, Estonia, Finland, France, Greece, Germany, Hungary, Ireland, Italy, Lithuania, Latvia, Malta, the Netherlands, Norway, Poland, Romania, the Slovak Republic, Slovenia, Spain and Sweden.
Source: European Union Statistics on Income and Living Conditions survey.
The second piece of evidence suggesting a more significant skills mismatch among older workers is the lower use of ICT skills by older workers compared to younger individuals in the same occupation. While it is a well-established fact that older workers have lower digital skills than younger individuals (Eurostat, 2024[18]), this alone is not necessarily a sign of a mismatch. Instead, older workers could be employed in different occupations in which they are less likely to need digital skills compared to younger individuals. The analysis of Survey of Adult Skills data suggests, however, that, even within the same occupation, older workers are less likely to use advanced ICT skills (Figure 4.5). Significant age gaps exist in the share of workers creating electronic documents, using specialised software and using a programming language.10 For example, the age gap in using specialised software ranged from 5% in Norway to 37% in Korea, and the use of programming language was more than 40% lower among older workers in some OECD countries. In contrast, older workers were similar to younger workers in terms of their use of basic ICT skills, such as using a computer at work and using the internet to communicate with others or to access information.
Percentage difference in the share of 55‑65 year‑olds and 25‑54 year‑olds using ICT at work, 2023
Note: Differences in the use of ICT within 2‑digit ISCO occupations. The difference between the age groups in using ICT to communicate with others is not statistically significant in Czechia and Spain. *Caution is required in interpreting results due to the high share of respondents with unusual response patterns – see the Note for Poland in OECD (2024[10]), Do Adults Have the Skills They Need to Thrive in a Changing World?: Survey of Adult Skills 2023, https://doi.org/10.1787/b263dc5d-en. Average is the weighted average of the countries shown.
Source: 2023 Survey of Adult Skills.
Older workers were also shown to be less likely than younger ones to take up new job opportunities, such as those created by the transition away from carbon-intensive production technologies (OECD, 2024[9]). In the United States, workers aged 55‑64 and those aged 65 and above were, respectively, 38% and 60% less likely than individuals aged 25‑34 to transition from “dirty” to “green” jobs. Moreover, among workers aged 55‑64 who left “dirty” jobs, the probability of moving to another “dirty” job was 25% higher compared to workers aged 18‑34 (Curtis, O’Kane and Park, 2023[19]).
Lower incentives for older workers to move away from declining occupations and develop new skills may be the reasons why most of the observed adjustment to labour market trends happens among younger workers. Older workers tend to have more occupation-specific expertise and if their bundle of skills cannot be fully transferred to the other occupations wage losses would have to be incurred (Kogan et al., 2022[20]; Hudomiet and Willis, 2022[21]). This effect, combined with shorter remaining working lives to reap the benefits of retraining (see Section 4.2), means that older workers may have lower incentives to exit the declining occupation. Older workers may also be more protected from displacement, as longer job tenure – often correlated with age – is a key factor in the strength of employment protection legislation (OECD, 2019[22]; Saez, Schoefer and Seim, 2023[23]). Finally, older workers may also be barred from some new occupations due to discrimination and age stereotypes – see Chapters 3 and 5.
Slower adaptation to labour market trends and adoption of new productivity-enhancing technologies among older workers, especially as the share of older workers in the workforce grows, may hinder economic dynamism, potentially weighing on the broader economy.
4.1.3. The changing occupational composition of the economy may help workers stay productive for longer
As discussed in the previous subsection, labour market changes may lead to a skills mismatch among older workers, but they also offer opportunities by shifting demand towards more age‑friendly jobs. While in some occupations, the return on experience is higher, and workers’ productivity and wages increase as they age, other occupations offer fewer development opportunities. Moreover, jobs differ in the degree to which they require physical strength that declines with age (Kenny et al., 2008[5]) and other features preferred by older workers, such as flexible working and autonomy (Maestas et al., 2023[24]) (see Chapter 3). This section investigates how the changes in the occupational composition may affect workers’ ability to remain productive for longer.
Research in the United States that tracked individual wage trajectories over time found that workers who became employed in occupations with a higher starting wage also experienced higher average wage growth in the subsequent eight years. For example, the wages of software engineers, police officers, and accountants are, on average, 100% higher eight years after they start work in their occupation, while the wages of cosmetologists, maids, or childcare workers barely change at all (Abraham et al., 2024[25]).
The difference in salaries between 55‑65 and 24‑34 year‑olds varies considerably across occupations, suggesting that wage growth trajectories vary across occupations as well over a longer time horizon. 55‑65 year‑old managers earn over 50% more than 25‑34 year‑old managers, controlling for gender and country fixed effects, while the difference in wages is only 4% between 25‑34 and 55‑65 year‑olds in elementary occupations (Figure 4.6). This is consistent with the higher dispersion in hourly earnings among older workers compared to younger age groups observed in the Survey of Adult Skills data.11
Difference in earnings of 55-65 and 25-34 year-olds by occupation, 2023
Note: Results of a regression of log earnings on age dummy, controlling for gender and country fixed effects. Earnings include bonuses and earnings by self-employed individuals. Data includes Austria, Flemish Region (Belgium), Canada, Chile, Czechia, Denmark, England (United Kingdom), Estonia, Finland, France, Germany, Hungary, Ireland, Israel, Italy, Japan, Korea, Latvia, Lithuania, the Netherlands, New Zealand, Norway, Poland, Portugal, the Slovak Republic, Spain, Sweden, Switzerland and the United States.
Source: 2023 Survey of Adult Skills.
Different wage trajectories may be related to the variation in returns to experience both within the job and due to differences in career opportunities. Returns to experience have been shown to be positive among skilled workers, insignificant among low-skilled workers (Dustmann and Meghir, 2005[26]), and higher for non-routine tasks than for routine ones (Gonzaga and Guanziroli, 2019[27]).
Moreover, some occupations require more physical effort than others, making it more difficult for older workers to perform those jobs, leading to lower productivity or even labour market exit.12 For example, 81% of building and related trades workers and 73% of cleaners and helpers work physically every day, while this is the case for only 6% of business and administration professionals and 3% of ICT professionals (Annex Figure 4.A.3).
The occupational composition differs significantly across OECD countries and may mean that in some countries, more jobs allow workers to remain productive for longer. In Chile, for example, 46% of workers are employed in jobs that require working physically for a long time every day, the highest share among the countries included in the survey (Figure 4.7). At the same time, only 33% of workers in Chile hold high-skilled occupations, where the increase in wages is the highest with age. In contrast, in the Flemish Region (Belgium), far fewer jobs require physical effort (32%) and almost 60% of workers are employed in high-skilled occupations where wages increase significantly with age.
Share of the employed working physically every day and share employed in high-skilled occupations by country
Note: High-skilled occupations include Managers, Professionals and Technicians and Associate Professionals. *Caution is required in interpreting results due to the high share of respondents with unusual response patterns – see the Note for Poland in OECD (2024[10]), Do Adults Have the Skills They Need to Thrive in a Changing World?: Survey of Adult Skills 2023, https://doi.org/10.1787/b263dc5d-en.
Source: 2023 Survey of Adult Skills.
In the future, changes to the occupational composition, away from jobs that require physical work and towards those that value experience, may make it easier for workers to remain productive for longer.13 Over the last decade, the share of jobs that require working physically every day decreased in most OECD countries (Annex Figure 4.A.4), while the share of high-skilled occupations that tend to offer higher returns to experience increased.
The changes in the structure of economic activity within countries may not be exogenous and, instead, may be affected by demographic changes. For example, ageing countries may specialise in industries with high returns to experience and shift away from industries in which skills and abilities that depreciate with age, such as physical strength, are more important (Cai and Stoyanov, 2016[28]). On the other hand, ageing may also lead to changes in the structure of consumption, for example, increasing demand for long-term care (Causa et al., 2025[29]; Oliveira Martins et al., 2005[30]). Finally, robots and AI will also affect the tasks performed by humans, potentially leading to a decrease in demand for physical tasks (ITF, 2023[31]).
Compared to younger workers, older workers tend to be at a disadvantage in terms of the information-processing skills they possess and the jobs they perform. Whether these patterns will be replicated in the future will depend on economic changes affecting occupational composition, but it will also depend on how well workers are equipped to adapt to the changing labour market and whether they are supported to develop their skills as they age.
In this context, lifelong learning has a critical role to play. It can help workers prevent skills atrophy, adapt to the changing labour market, and move into higher-productivity and better-paying jobs. Recent evidence from Denmark demonstrates that the returns on investment in human capital can be high for adults, with earnings found to be 25% higher among those with a vocational qualification who returned to education to complete a bachelor’s degree (Humlum, Munch and Plato, 2023[32]).14 Randomised evaluations of sector-focused training programmes also found substantial and persistent earnings gains following training, ranging from 12% to 34%, depending on the scheme (Katz et al., 2022[33]). This is consistent with positive returns to training found in the meta‑analysis of active labour market policies (Card, Kluve and Weber, 2018[34]). Work-related training was also found to be correlated with direct measures of productivity, with 1 percentage point increase in training associated with an increase in value added per hour of 0.6% and 0.3% in hourly wages (Dearden, Reed and van Reenen, 2006[35]).
This section investigates how training participation and on-the‑job learning evolve as workers age, highlighting differences across countries.
4.2.1. Training participation and learning-by-doing decrease with age
Participation in formal and non-formal adult learning decreases with age. In 2023, only a third of 60‑65 year‑olds participated in adult learning in the 12 months preceding the survey, compared to over half of the 25‑44 year‑olds (Figure 4.8, Panel A). The most notable drop in adult learning participation occurs between the ages of 55‑59 and 60‑65. Non-formal training was significantly more common than formal learning (i.e. training that led to a qualification) across the age groups, but this was particularly the case among older individuals, with only 1% of 60‑65 year‑olds participating in formal learning.
Learning-by-doing, another important but less often studied channel through which individuals develop new skills, also decreases with age. Among 25‑29 year‑olds, 62% report learning by doing at least every week compared to 45% among 60‑65 year‑olds (Figure 4.8, Panel B). Learning by doing is only observed for those who are employed, and employed individuals tend to have higher rates of training, which likely explains why learning-by-doing decreases less with age than training participation in the wider population.15
Workers aged 55‑65 are 7 percentage points less likely to participate in training and 14 percentage points less likely to be learning-by-doing compared to 25‑34 year‑olds with the same level of education and occupation. This suggests that the lower degree of learning is not just due to generational differences, such as different educational attainment leading to differences in occupational composition, but instead, it is related to age.
Participation in non-formal training and learning-by-doing differ significantly across socio‑economic groups, both in terms of levels and trends as workers age. For example, participation in non-formal learning for those with tertiary education fluctuates around 60% between the ages of 25 and 54 and only starts decreasing afterwards, i.e. 10 years later than in the general population (Annex Figure 4.A.5). Among tertiary educated 60‑65 year‑olds, non-formal training participation remains at a relatively high level of 49%. In contrast, the training participation rate of those without upper secondary education is lower and starts decreasing earlier in life, dropping to 9% among 60‑65 year‑olds. Learning-by-doing is also significantly higher among the tertiary educated, but it decreases gradually throughout life, regardless of the level of education.
Note: Learning by doing in Panel B refers to individuals who reported that their job involves learning-by-doing from the tasks they perform at least once a week. Graph presents a weighted average of Austria, Flemish Region (Belgium), Canada, Chile, Czechia, Denmark, England (United Kingdom), Estonia, Finland, France, Germany, Hungary, Ireland, Israel, Italy, Japan, Korea, Latvia, Lithuania, the Netherlands, New Zealand, Norway, Poland, Portugal, the Slovak Republic, Spain, Sweden, Switzerland and the United States.
Source: 2023 Survey of Adult Skills.
The lower participation in non-formal learning of older individuals may stem from either a lower willingness to train (defined as the likelihood that workers identified learning opportunities that they wanted to take up) or greater barriers to accessing training, such as time constraints or costs of training courses. Evidence indicates that a lower willingness to train is likely a key factor. The share of the population that wanted to participate in training (regardless of whether they did)16 decreases with age, from around 60% among 25‑44 year‑olds to 37% among 60‑65 year‑olds (Annex Figure 4.A.6). A similar pattern is observed in the share of the population that participated in training less than they wanted to, which also decreases with age, from 28% among 25‑34 year‑olds to 17% among 55‑65 year‑olds (see Box 4.3 for reasons why people participate in training less than they want to).
Time constraints are less of a barrier to training participation for older people than for other age groups. Among 55‑65 year‑olds, 7% reported participating in training less than they wanted to due to time constraints – 5% due to work-related reasons and 2% due to family reasons (Figure 4.9). In contrast, 15% of 35‑44 year‑olds indicated time constraints as a barrier to training participation, with 8% pointing to work-related time constraints and 7% to family responsibilities.
“Other” barriers to training participation, such as an unexpected event (e.g. health problems) or other reasons that the survey is not capturing, play a more important role among older people. A similar share of the 55‑65 year‑old population participated in training less than they wanted to due to “other” reasons, as in other age groups, while the overall share of the 55‑65 population that participated in training less than they wanted was lower (Figure 4.9). This means that “other” reasons were a proportionally bigger barrier to training participation for older people compared to younger individuals.
Share of the population that participated in training less than they wanted, by reason
Note: Barriers related to training availability include no suitable training available, training at an inconvenient time or location, training cancelled or postponed, not meeting training prerequisites and lack of employer support. Other reasons include something unexpected came up and unspecified reasons. Graph presents a weighted average of Austria, Flemish Region (Belgium), Canada, Chile, Czechia, Denmark, England (United Kingdom), Estonia, Finland, France, Germany, Hungary, Ireland, Israel, Italy, Japan, Korea, Latvia, Lithuania, the Netherlands, New Zealand, Norway, Poland, Portugal, the Slovak Republic, Spain, Sweden, Switzerland and the United States.
Source: 2023 Survey of Adult Skills.
Moreover, employers who are a key source of funding for training are less likely to invest in the training of older workers. Over half of the non-formal training in which individuals participated in the year preceding the survey was at least partially funded by employers, making firms the leading source of financing for workforce development. The share of the population that received non-formal training funded by employers in the year preceding the survey decreases from about 30% among 30‑44 year‑olds to 22% among 55‑59 year‑olds and 14% among 60‑65 year‑olds. While this is partly due to lower employment rates at older ages, the share of employed individuals who participated in non-formal training funded by employers also decreases with age, from about 35% among 30‑44 year‑olds to 29% and 25% among the 55‑59 and 60‑65 year‑olds, respectively.
4.2.2. Large differences in training participation of older workers are observed across countries in the OECD
While participation in non-formal learning of older individuals is lower than among the prime‑age population (25‑54 year‑olds) in all OECD countries, both the level of participation among older people and the age gap differ significantly. The highest level of non-formal training participation among 55‑65 year‑olds is observed in Norway, Finland and Denmark (around 50%) (Figure 4.10). In contrast, the lowest non-formal training participation in this age group was recorded in Poland, Korea and the Slovak Republic (less than 20%). Non-formal training participation of older people is highly correlated with the training participation of prime‑age individuals, but the age gap ranges from 4 to 26 percentage points. For example, Canada has the fourth-highest non-formal training participation of prime‑age people but a significant age gap (26 percentage points), ranking only twelfth for non-formal training participation of older people.
Participation in non-formal learning in the year preceding the survey, 25‑54 year‑olds and 55‑65 year‑olds
Note: *Caution is required in interpreting results due to the high share of respondents with unusual response patterns – see the Note for Poland in OECD (2024[10]), Do Adults Have the Skills They Need to Thrive in a Changing World?: Survey of Adult Skills 2023, https://doi.org/10.1787/b263dc5d-en. Average is the weighted average of the countries shown.
Source: Survey of Adult Skills 2023.
Learning-by-doing is also less common among older workers than it is among prime‑age workers in all OECD countries, but the gap is narrower than for training. Learning-by-doing at least every week was most common among older workers in Portugal, Ireland and Spain (above 55%), and it was the lowest in Poland, Lithuania, Korea and Hungary (25% or less) (Annex Figure 4.A.7). While the intensity of learning-by-doing among older workers is correlated with the level of learning among prime‑age workers, here, too, the overall level is not all that matters. For example, Italy and the Slovak Republic have a similar share of prime‑age workers learning by doing, but while in the Slovak Republic the age gap was narrow, in Italy, a large age gap was observed, meaning that older workers were significantly less likely to learn on the job in Italy than in the Slovak Republic.
4.2.3. Older workers train less due to lower expected returns, but they may also be underestimating the benefits of training
One reason why older people are less likely to engage in training, and employers are less likely to fund their training, are lower expected returns on such investments due to shorter remaining working lives. Economic theory suggests that individuals make investment decisions by weighing expected returns against associated costs. Thus, a person is likely to pursue training only if its discounted returns outweigh both the financial and psychological costs involved. Since older workers generally have fewer years left in the workforce compared to their younger counterparts, the return on investment in training is lower for them. Consequently, lower participation in training among older workers is expected (OECD, 2019[8]; Picchio, 2021[36]). Similar arguments may apply to the decision to change jobs, which requires effort to learn new skills on the job to adapt to the new work environment (see Box 4.4) and to employers’ decision to invest in the training of older workers.
Moreover, if acquiring new skills is more challenging for older workers than for younger ones, the increased difficulty raises the cost of learning, further lowering the expected returns on training. While such effects would be consistent with changes in fluid intelligence – the ability to solve problems in novel situations independently of acquired knowledge – which is found to decrease with age (Desjardins and Warnke, 2012[6]), evidence to suggest that older participants perform worse than their younger counterparts on tests after completing training is limited to software training (Charness and Czaja, 2006[38]). Moreover, the effectiveness of training for older workers may also depend on training methods, which are often designed to meet the needs of younger workers (Picchio, 2021[36]).
The age gap in non-formal training participation is larger in countries with lower effective labour market exit age,17 especially for women (Figure 4.12).18 This is consistent with individuals and firms making decisions about participation in training based on the return on investment. For example, in Korea, where the effective labour market exit age for women was the highest (67) in 2022, the gap in non-formal training participation of older and prime‑age women was relatively small (11%). In contrast, in Poland, the Slovak Republic and Austria, where the effective labour market exit age of women was 61‑62 in 2022, the participation in non-formal training of women aged 55‑65 was less than half of the non-formal training participation rate of prime‑age individuals.
Extending working lives – for example, through pension reforms – could lead to higher training participation if workers and firms make training investment decisions based on expected returns. Indeed, evidence suggests that pension reforms19 in the Netherlands and Italy had a positive impact on the training participation of workers in their 50s (Montizaan, Cörvers and De Grip, 2010[39]; Brunello and Comi, 2015[40]). This is consistent with the findings of other studies, which show that a drop in mortality rates, which extended the potential period during which investment benefits can be reaped, increased investment in general human capital (Kalemli-Ozcan, Ryder and Weil, 2000[41]; Jayachandran and Lleras-Muney, 2009[42]). If the expected rate of return fully explained the lower training participation of older workers, this outcome could be efficient from the perspective of the individual, the employer and the economy.
Workers who stay in the same job for longer are more likely to develop job-specific skills and may eventually feel less need to learn new skills to meet the demands of the job they hold. This is consistent with evidence showing that workers with longer tenure at the company are less likely to be learning-by-doing, even after controlling for internal mobility within the company (Figure 4.11, Panel A). Given that job mobility declines with age – only 6% of 45‑64 year‑olds change jobs annually compared to 17% under the age of 30 (OECD, 2024[37]) – and tenure increases (Figure 4.11, Panel B), lower mobility could be responsible for lower learning-by-doing among older workers.
Note: Results from a regression controlling for 2‑digit occupation, education, having moved to a different position in the company, having experienced a change in tasks and responsibilities and having moved to a different unit or department and country fixed effects. The 95% confidence interval are shown. Data for Austria, Flemish Region (Belgium), Canada, Chile, Czechia, Denmark, England (United Kingdom), Estonia, Finland, France, Germany, Hungary, Ireland, Israel, Italy, Japan, Korea, Latvia, Lithuania, the Netherlands, New Zealand, Norway, Poland, Portugal, the Slovak Republic, Spain, Sweden, Switzerland and the United States.
Source: 2023 Survey of Adult Skills.
Difference between the participation rate in non-formal learning of 55‑65 year‑olds and 25‑54 year‑olds
Note: *Caution is required in interpreting results due to the high share of respondents with unusual response patterns – see the Note for Poland in OECD (2024), Do Adults Have the Skills They Need to Thrive in a Changing World?: Survey of Adult Skills 2023, https://doi.org/10.1787/b263dc5d-en.
Source: Survey of Adult Skills 2023 for training participation; effective labour market exit age based on OECD (2022), Pensions at a Glance 2022.
However, there is a risk that individuals and employers may underestimate the returns on investment in training older workers and, therefore, choose an inefficiently low level of training participation, for example, if they believe that the ability of older workers to learn new skills is lower than it actually is. When hiring, employers expect mid-career and older applicants (aged 45 and above) to be a worse fit in terms of skills than younger applicants and to be less willing to learn new skills and try new technologies (OECD/Generation: You Employed, Inc., 2023[43]). These negative attitudes towards older workers may be partly driven by stereotypes. Evidence from correspondence studies shows that 50‑year‑old workers with the same qualifications and experience were less likely to be called back for interviews than younger workers due to discrimination (Baert, 2017[44]; Neumark, Burn and Button, 2019[45]; Carlsson and Eriksson, 2019[46]), and similar mechanisms may apply to firms’ decisions to provide training. Moreover, older workers themselves may have lower confidence in their own ability to learn new skills, as they are less used to learning (Posthuma and Campion, 2009[47]). Individuals may also choose a level of training participation that is below the socially optimal level if the returns on investment are primarily captured by employers (see also Section 4.3.2).
Finally, if training postpones the decision to retire, the socially optimal level of training participation of older workers may be higher than the one chosen by the individual or their employer, given that keeping older individuals in the workforce increases output and tax revenues and reduces pension benefit payments – see also Chapter 2. Research is, however, inconclusive about the effect of training on the employment of older workers and retirement decisions, with the results ranging from positive (Herrbach et al., 2009[48]; Kristensen, 2012[49]; Piccio and van Ours, 2013[50]; Berg et al., 2015[51]) to no effect (Stenberg, de Luna and Westerlund, 2012[52]; Boockmann, Fries and Göbel, 2018[53]).
Both social returns to training in excess of individual returns and the underestimation of the return on training investment by individuals and firms would justify government intervention to lower the cost of training for older workers.
To avoid mismatches at older ages, workers should regularly reassess their skills and career objectives and participate in training to bridge any skills gaps. Governments can aid these efforts by offering low-cost career guidance services and promoting lifelong learning opportunities that reflect the learning needs of workers at different ages. This section presents examples of government interventions that facilitate access to career guidance and adult learning opportunities for older people and discusses preventive measures targeting mid-career individuals that can help workers age better – see Chapter 3 for policies that support better use of skills within firms.
4.3.1. Career guidance for adults
Career guidance can support adults in reflecting on their professional journeys and identifying steps to achieve a satisfying career as they age. First, advisors can assist individuals in assessing whether their current roles and skills align with their ambitions and labour market demand, while also considering their ability to sustain their jobs over time. Based on such assessments, advisors can suggest pathways for career progression or occupational shifts and propose training courses that will help the individual achieve their new career objectives. While the literature on the impact of career guidance for employed individuals is limited, offering career guidance to job seekers at risk of long-term unemployment in France was shown to increase the probability of finding a job (J-PAL, 2015[54]) and in the United States, individualised career guidance offered to job seekers increased the probability of finding a job as well as the salary and fringe benefits of those jobs (Mathematica, 2019[55]) – see also Chapter 3.
Yet, the use of career guidance remains low in many OECD countries, especially among mid-career and older individuals. In 2020/21, 38% of mid-career adults (40‑54 years olds) and 27% of older adults (55‑64 year‑olds) in countries that participated in the OECD’s Survey of Career Guidance for Adults had spoken to a career advisor in the preceding five years. That was significantly lower than the use of career guidance among younger people (25‑39 year‑olds), at 49% (OECD, 2021[56]; 2022[57]).20
While internationally comparable data for a broader number of countries is limited, the Adult Education Survey shows that similar patterns are observed for how often workers looked for information on learning opportunities. Mid-career workers (45‑54 year‑olds) and older workers (55‑64 year‑olds) are less likely to look for information on learning opportunities than younger workers (25‑34 year‑olds) (Figure 4.13). The highest share of mid-career workers who looked for information was recorded in the Netherlands (40%) and the lowest in Türkiye and Bulgaria (6% and 5%, respectively).
Share of individuals who searched for information on learning possibilities by age, 2022
Source: Eurostat (2024), “Search for information on learning possibilities by type of learning and age”, Adult Education Survey, https://doi.org/10.2908/TRNG_AES_183 (accessed on 28 May 2025).
To increase the use of career advice services among adults and prevent skills mismatches at older ages, some OECD countries offer free or subsidised career advice services. While such services are commonly provided to unemployed individuals, schemes available to all adults, regardless of their employment status, also exist. For example, in Switzerland, individuals aged 40 or older can take advantage of free professional assessments, provided they are not already eligible for similar services due to unemployment status or disability. In Flanders and Brussels, adults who have at least seven years of work experience are eligible for “career vouchers” to participate in subsidised career guidance sessions delivered by accredited providers (see Box 4.5). Skills Development Scotland centres offer free advice to people regardless of age (Skills Development Scotland, n.d.[58]). In France, funding available to adults under individual learning accounts can be used to cover the cost of career advice services. In a number of countries, such as Austria, Germany and Estonia, Public Employment Services provide career guidance to all adults regardless of their employment status (OECD, 2021[56]). Finally, in Singapore, an initiative called Volunteer Career Advisors matches individuals, especially mature workers, with experienced professionals in the selected industry who provide peer-level support and career guidance to help the individual advance their career or transition to other jobs (see Box 4.5).
Career Counselling for those aged 40 and over, Switzerland
Under the Viamia scheme in Switzerland, individuals aged 40 and over are offered a free professional assessment, provided they are not already eligible for similar services due to unemployment or disability. The process begins with participants completing a self-assessment questionnaire that focuses on their skills and motivation, along with submitting their CV. This is followed by a one‑on-one interview with a counsellor to analyse their personal and professional situation in light of current labour market trends and demands. If one session is not sufficient, additional interviews are arranged. The process concludes with the participants, supported by the counsellor, defining concrete actions to maintain or improve their employability.
Viamia is a joint initiative of the Swiss federal government and the cantons. It was launched in 2021 as a pilot in 11 cantons and expanded nationwide in 2022. Between 2021 and 2024, a lump sum of CHF 1 200 was allocated per participants, which was estimated to cover six hours of counselling. In the period 2022‑23, 13 169 individuals benefited from the scheme, out of whom the vast majority were women (70%), employed individuals (82%) and 40‑50 year‑olds (71%).
Participant feedback has been overwhelmingly positive, with an average satisfaction rating of 5.5 out of 6, and 98% of participants indicating they would recommend the programme to others.
Career vouchers in Flanders and Brussels, Belgium
In Flanders and Brussels, workers can use career vouchers (loopbaancheques) to participate in subsidised career guidance sessions. Individuals can benefit from two vouchers every six years. The first voucher allows them to participate in 4 hours of guidance, and the second one in 3 hours of guidance provided at a recognised career centre. The co-payment from the individual to use the voucher is EUR 45, while the estimated cost of career guidance is EUR 182 per hour (i.e. EUR 1 274 for seven hours). To be eligible, individuals need to live in Flanders or Brussels, be employed and have at least seven years of work experience in paid employment or self-employment. Individuals can apply for the voucher online on the VDAB portal, where the list of career centres can also be found.
Between 2013 (when the scheme was set-up) and 2021, 155 000 citizens, or 5.2% of the employed population, used the vouchers. In 2024, 41% of participants were 30‑40 years old and 33% were 40‑50 years old, which was significantly above the share of the employed that the respective age groups constitute (25% and 25%). Workers above the age of 50 were underrepresented (22% of participants and 32% of the employed). Career guidance centres participating in the scheme are required to demonstrate that at least 30% of the individuals they served belonged to disadvantaged groups, which includes those above the age of 50.
Volunteer Career Advisors initiative, Singapore
Under the Volunteer Career Advisors initiative, Workforce Singapore is building a pool of advisors from professional communities. These advisors are then matched with individuals, especially mature workers, who are employed in the same sector or who are looking to join the sector and seek peer support and career guidance. The initiative has been launched for the Accountancy, Electronics & Semiconductors, Healthcare, Information & Communication Technology, Sustainability and Retail sectors, and will be progressively rolled out in more sectors. The initiative is part of a broader SkillsFuture Mid-Career Support Package (see more details in Box 4.6).
Note: Employment data refers to Belgium.
Source: Scheme in Switzerland: The Federal Department of Economic Affairs, Education and Research (n.d.[59]), Viamia, https://viamia.ch/fr/#kontakt; Ecoplan (2024[60]), Évaluation de viamia: Mise en œuvre et effets de l’offre, https://formationprofessionnelle2030.ch/images/projekte/viamia/Rapport_final_Evaluation_de_viamia_2023_Ecoplan.pdf; Scheme in Belgium: VDB (2021[61]), Monitoringsrapport Loopbaanbegeleiding, https://extranet.vdab.be/system/files/media/bestanden/2022-03/Monitoring%20Loopbaanbegeleiding%202021_2.pdf; VDAB (n.d.[62]), Alles over loopbaancheques, www.vdab.be/orienteren/loopbaanbegeleiding/alles-over-loopbaancheques; VDAB (2024[63]), Cijfergegevens Volledig, https://extranet.vdab.be/system/files/media/bestanden/2025-01/Cijfergegevens%202024.pdf; Eurostat (2025[64]), “ Employment by sex, age and citizenship”, https://doi.org/10.2908/LFSA_EGAN; Scheme in Singapore: Workforce Singapore (2025[65]), Volunteer Career Advisors Initiative, www.wsg.gov.sg/home/individuals/attachment-placement-programmes/volunteer-career-advisors-initiative.
Making career advice available may not be sufficient, and additional outreach efforts may be needed to boost the use of career guidance services. In the OECD’s Survey of Career Guidance for Adults, the second most common reason for not using career advice services, indicated by about 20% of mid-career individuals who did not participate in career guidance, was the lack of knowledge that such services existed. Similarly, a survey in Australia found that 42% of 50‑69 year‑olds did not know where to access career advice (National Careers Institute, 2022[66]). Moreover, many individuals do not feel the need to use career advice services, which was the most common reason cited for non-participation in the OECD’s Survey of Career Guidance for Adults. Half of mid-career women and 60% of mid-career men did not use career advice services because they did not think they needed them (OECD, 2021[56]). While for many of those individuals, that may be a rational choice, for others, it may be a result of the underestimation of opportunities and benefits that career advancement or change can offer.
The field of healthcare could serve as a model for proactive early engagement with individuals to help prevent adverse outcomes later in life. Many countries contact individuals who reach a certain age and invite them to participate in health screenings to identify health problems early, such as mammography for women who turn 45. Similar outreach and tests, in the form of career guidance sessions, could be conducted to test the “health” of one’s career and prevent, rather than cure, any skill gaps. Statistical profiling could be used to predict an individual’s likelihood of leaving the labour force early and reach out only to those at high risk to reduce the cost of the intervention.
Employers can also play an important role in supporting employee career development through regular career conversations and/or mid-career reviews (OECD, 2024[67]). For example, in France, employers are required to regularly provide career guidance to their employees. In 2008, France adopted a law that puts an obligation on companies to prepare “older workers” plans either in-house or in collaboration with social partners, which was subsequently replaced in 2013 by a requirement to conduct an appraisal review with all employees, regardless of age, every other year. Schneider Electric is an example of a French multinational company that put in place a Senior Talent Program to support workers in the late stages of their careers (aged 51 and above), which includes workshops that prepare both experienced workers and their managers to engage in career conversations at least once a year (OECD, 2024[68]). However, this approach should not be thought of as a replacement for career advice provided by an independent career guidance organisation, given that it is likely to focus on an individual’s future career within the existing organisation without considering the opportunities in different sectors or areas of activity (Eurofound, 2017[69]).
4.3.2. Support to participate in training for mid-career and older individuals
Training is another key lever for enhancing older workers’ skills and productivity, and policy makers play an important role in encouraging investment in training by both individuals and employers.
In this context, a key question faced by policy makers is whether to support lifelong learning for all or to introduce additional measures specifically targeting mid-career and older workers. As discussed in the preceding section, the returns to training decrease with age due to a shorter remaining working life, which reduces the incentives for both employers and individuals to pay for training at older ages. If social returns to training are higher than private returns, for example, because training participation delays retirement, thus reducing the government’s pension spending, providing additional financial incentives to older individuals and their employers may be socially desirable. Similarly, additional support for mid-career and older workers may be appropriate if they are sufficiently distinct from younger individuals in terms of educational attainment or behavioural aspects such as risk aversion, making them more likely to choose an inefficiently low level of training than younger individuals.
Evidence from Denmark suggests that current rates of training are sub-optimal for workers between the ages of 40 and 50 in particular, both from the government and individual perspectives. A study which analysed the returns to a reskilling subsidy after a workplace injury found that only 6% of middle‑aged workers reskilled after injury, while the costs to the government (covering tuition and benefits during training) pay for themselves through higher taxes and lower future benefit payments for 36% of these workers. A similar take‑up of reskilling was found to be optimal from individual and social perspectives. By contrast, the reskilling rates among the youngest and oldest workers (age 20‑30 and 60‑65, respectively) were found to be close to the social optimum (Humlum, Munch and Plato, 2023[32]). While it is uncertain whether these results generalise to other contexts and countries, they suggest that underinvestment in training may differ by age, potentially providing justification for additional support for specific age groups.
Several OECD countries provide more substantial financial incentives for firms to invest in the skills of mid-career and older workers, or prioritise older workers when distributing funding for training. For example, in Germany, firms can receive subsidies (Förderung Beschäftigtenqualifizierung nach §82 SGB III) that cover between 25% to 100% of expenses associated with staff training depending on the company size and the characteristics of workers, with training costs of workers over the age of 45 being fully covered for companies with less than 500 employees (Bundesagentur für Arbeit, 2025[70]). In Austria, employers can obtain funding for the training costs of certain groups of workers (Qualifizierungsförderung für Beschäftigte), one of them being individuals above the age of 50. The subsidy covers 50% of the training fees, up to the limit of EUR 10 000 per person (Bundesministerium für Arbeit und Wirtschaft, 2024[71]). In Poland, employers can apply for funding to cover the training costs of their employees from the National Training Fund (Krajowy Fundusz Szkoleniowy), with 80% of costs covered for SMEs and 100% for microenterprises. While the overall funds available each year are capped, priority is given to the training for workers who are 45 years old or older (Wojewódzki Urząd Pracy w Krakowie, 2024[72]).
In Singapore, additional funding for training is offered directly to mid-career individuals without the need for employer support, which boosts individual choice and responsibility with regard to training. Higher subsidies to participate in training, in some cases above 90%, are offered to individuals who are 40 or older. The remaining costs can be covered through SkillsFuture Credit, which operates like an individual learning account. In addition, Singapore is planning to introduce an income‑replacing allowance for mid-career individuals participating in selected full-time training programmes (see Box 4.6). In Ireland, subsidised upskilling and reskilling courses (Skills to Advance) target workers aged 50 or older, among other vulnerable groups (Solas, n.d.[73]).
Singapore offers additional funding for training to mid-career and older individuals in addition to its already generous funding for adult learning. First, the Mid-Career Enhanced Subsidy, available to Singaporean citizens aged 40 and above, covers at least 90% of fees for courses funded by the Ministry of Education and up to 90% of courses funded by SkillsFuture Singapore. These courses range from technical and vocational education programmes to postgraduate level.
In addition, Singapore offers SkillsFuture Credit, which operates like an individual learning account. All citizens aged 25 and above receive SGD 500 (EUR 350), which can be used to participate in approved training. The funding can be used at a time convenient for the individual and does not expire. In 2024, those aged 40 and above received a top-up of SGD 4 000 (EUR 2 800) to upgrade their skills. While the selection of courses for individuals 40+ is more restrictive than for the general population, it still includes approximately 7 000 courses, which are listed on MySkillsFuture website.
The support measures can be used jointly. For example, a course worth SGD 17 200, costs only SGD 1 720 after the Mid-Career Enhanced Subsidy and the remaining costs can be covered through the SkillsFuture Credit, meaning that there are no out-of-pocket expenses for individuals aged 40+.
Finally, in early 2025, Singapore is planning to introduce a new SkillsFuture Mid-Career Training Allowance, which will offer income replacement to those participating in selected full-time, long-form programmes. The allowance will be 50% of the average income of the individual over the last 12‑month period, capped at SGD 3 000 (EUR 2 100) a month. The individuals will be able to use the allowance for up to 24 months over their lifetime.
Source: Skills Future Singapore (2025[74]), SkillsFuture Mid-Career Enhanced Subsidy, www.skillsfuture.gov.sg/initiatives/individuals/enhancedsubsidy; SkillsFuture Singapore (2025[75]), SkillsFuture Level-Up Programme, www.myskillsfuture.gov.sg/content/portal/en/career-resources/career-resources/education-career-personal-development/SkillsFuture_Level-Up_Programme.html?_gl=1*ptwrgi*_gcl_au*MzY4Nzc4OTI3LjE3MzY3ODA1MT.
Broader financial support for adult learning, regardless of age, that enables individuals to access training opportunities at any stage of their career can also help workers age better. This can be implemented through individual learning accounts (ILA), which provide individuals with a budget to spend on training at any point in their lives.21 Examples of countries that either have an ILA scheme or are piloting one include France, Singapore, Lithuania and Czechia. Greece has established a legal framework for ILAs, offering more generous funding depending on characteristics, such as age, and is preparing to pilot the scheme. A similar scheme, but focused on loans, where individuals are offered subsidised loan entitlement that can be used for education and training throughout life, is being piloted in the United Kingdom (Lifelong Loan Entitlement). Instilling the lifelong learning mindset from an early age and providing individuals with opportunities to upgrade their skills throughout life will not only improve their labour market position but is also likely to encourage them to seek training at an older age.22
Another important question is how the cost of training should be shared between the labour market participants. Employers are well-positioned to ensure workers are equipped for a changing labour market, given that research suggests that they often reap substantial returns from training. For example, in France and Sweden, firms were found to capture 70% and 65% of the returns to firm-sponsored training, respectively (Ballot, Fakhfakh and Taymaz, 2006[76]) and, in the United Kingdom, work-related training was associated with a higher increase in productivity than in wages (Dearden, Reed and van Reenen, 2006[35]). However, social returns to training in excess of private returns may justify the government also sharing the cost of the training investment.
To ensure fair funding responsibilities, collective bargaining can be leveraged, as was done in Sweden to introduce financial support to participate in training based on Education Support for Transition (Omställningsstudiestöd) agreement. Collective bargaining can also help ensure that upskilling efforts translate into higher wages and opportunities for career progression, thereby increasing the incentives for individuals to participate in training. Where funding for training is offered via government schemes, cost-sharing between the government and employers can be facilitated through the introduction of a firm levy scheme, used to partially finance a dedicated training fund. For example, in France, employers are obliged to contribute to professional training (Contribution à la Formation Professionnelle, CFP), which finances various skills development initiatives, including the ILA.
Yet, financial support alone may not be enough to boost the training participation of older workers, especially those with lower qualifications and skills. In Lithuania, France, Czechia and the Netherlands, 30‑49 year‑olds were the main beneficiaries of ILAs and were overrepresented among its users compared to the general population, while those above the age of 50 and below the age of 30 were underrepresented (OECD, 2025[77]).
Outreach strategies involving social partners, Public Employment Services, and local community organisations, followed by career guidance, can help individuals of all ages identify skill gaps and motivate them to engage in training. For example, Unionlearn, part of the British trade union umbrella organisation, supports skills development through a network of Union Learning Representatives. These workplace‑based union members assess employees’ learning needs and co‑ordinate training activities in collaboration with employers (Unionlearn, 2024[78]). Public Employment Services and community-based organisations can further promote training participation, including among those outside the workforce.
In addition, it is important that adult learning systems offer mechanisms for recognition of prior learning and provide flexible learning options that respond to the constraints that adults are facing, such as lack of time due to work and family responsibilities. Experienced workers often possess skills gained through informal on-the‑job learning, which are difficult to showcase through formal qualifications. The availability of recognition of prior learning systems allows adults to formally validate the knowledge they have acquired in an informal setting, which can not only directly improve the job prospects of older workers by making their skills visible to employers, but also facilitate participation in adult learning (OECD, 2019[22]). This is because recognising existing competencies, combined with shorter, modular courses, helps increase the relevance of training schemes by allowing individuals to focus only on content that addresses their skill gaps, shortens training duration and lowers its cost.
Finally, adapting the content of training to the needs of older workers may increase the effectiveness of training. Research suggests that the training currently on offer may be less effective for older workers because it does not relate to their everyday experiences (Schirmer et al., 2022[79]) or focuses on abstract technical content rather than practical work problems (Zwick, 2011[80]). Here, too, collective bargaining can be leveraged to identify and articulate workers’ training needs (OECD, 2019[81]).
As the workforce in OECD countries gets older, it is increasingly important to ensure that workers stay productive as they age. The economic changes which shift occupational composition away from jobs that require physical work and towards those that value experience may make it possible for workers to remain productive for longer. However, there is a risk that these benefits will be offset by a decline in skills with age. Today, older workers tend to be at a disadvantage in terms of the information-processing skills they possess. The decline in information-processing skills within cohorts over the past decade and slower adjustment to labour market changes among older workers suggest that, if no action is taken, these age gaps in skills may persist in the future, which may weaken older worker’s capacity to remain in employment as well as hinder economic dynamism and slow down productivity growth, potentially weighing on the broader economy.
To address this risk, there is a need for an urgent shift from the traditional three‑stage (school, work, retirement) life model to a more flexible one where learning and work take place throughout life. Governments can help make that a reality by facilitating access to career guidance and adult learning, particularly focusing on mid-career and older workers, for whom the underinvestment in skills is the greatest. The field of healthcare could serve as a model for proactive early engagement with individuals to help prevent adverse outcomes later in life. Outreach and tests, in the form of career guidance sessions, could be conducted to test the “health” of one’s career and prevent, rather than cure, any skill gaps. To keep this process cost-effective, these sessions could be limited to individuals who statistical profiling suggests are most at risk of early exit from the labour force, such as those with lower educational attainment. Individual learning accounts could also be leveraged to boost training participation. Their advantage is that they not only provide individuals with funding to upgrade their skills throughout life but also allow governments to flexibly offer additional top-ups to target groups, for example, mid-career and older workers, when such needs arise.
References
[25] Abraham, L. et al. (2024), Wage Growth and Initial Occupations: Examining Differences Across Worker Demographics Using Longitudinal Data, RAND Corporation, Santa Monica, https://www.rand.org/pubs/research_briefs/RBA3492-1.html.
[2] 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.
[15] Autor, D. and D. Dorn (2009), “This Job Is “Getting Old”: Measuring Changes in Job Opportunities Using Occupational Age Structure”, American Economic Review, pp. 45-51, https://doi.org/10.1257/aer.99.2.45.
[44] Baert, S. (2017), “Hiring Discrimination: An Overview of (Almost) All Correspondence Experiments Since 2005”, https://docs.iza.org/dp10738.pdf.
[76] Ballot, G., F. Fakhfakh and E. Taymaz (2006), “Who Benefits from Training and R&D,the Firm or the Workers?”, British Journal of Industrial Relations, Vol. 44/3, pp. 473-495, https://doi.org/10.1111/j.1467-8543.2006.00509.x.
[51] Berg, P. et al. (2015), “The Relationship between Establishment Training and the Retention of Older Workers: Evidence from Germany”, NBER, https://www.nber.org/system/files/working_papers/w21746/w21746.pdf.
[53] Boockmann, B., J. Fries and C. Göbel (2018), “Specific measures for older employees and late career employment”, The Journal of Economics of Ageing, Vol. 12, pp. 159-174, https://doi.org/10.1016/j.jeoa.2018.03.005.
[40] Brunello, G. and S. Comi (2015), “The side effect of pension reforms on the training of older workers. Evidence from Italy”, The Journal of the Economics of Ageing, Vol. 6, pp. 113-122, https://doi.org/10.1016/j.jeoa.2015.02.001.
[70] Bundesagentur für Arbeit (2025), Weiterbildung für Ihr Personal – Zukunft für Ihr Unternehmen [Further training for your staff - the future for your company], https://www.arbeitsagentur.de/k/weiterbildung-qualifizierungsoffensive (accessed on 16 May 2025).
[71] Bundesministerium für Arbeit und Wirtschaft (2024), Qualifizierungsförderung für Beschäftigte [Qualification support for employees], https://www.oesterreich.gv.at/themen/arbeit_beruf_und_pension/aeltere_arbeitnehmer/1/2/Seite.2010140.html (accessed on 16 May 2025).
[28] Cai, J. and A. Stoyanov (2016), “Population aging and comparative advantage”, Journal of International Economics, Vol. 102, pp. 1-21, https://doi.org/10.1016/j.jinteco.2016.04.006.
[82] Caplin, A. et al. (2023), “How Worker Productivity and Wages Grow with Tenure and Experience: The Firm Perspective”, NBER Working Paper, https://www.nber.org/system/files/working_papers/w30342/w30342.pdf.
[34] Card, D., J. Kluve and A. Weber (2018), “What Works? A Meta Analysis of Recent Active Labor Market Program Evaluations”, Journal of the European Economic Association, Vol. 16/3, pp. 894-931, https://doi.org/10.1093/jeea/jvx028.
[46] Carlsson, M. and S. Eriksson (2019), “Age discrimination in hiring decisions: Evidence from a field experiment in the labor market”, Labour Economics, Vol. 59, pp. 173-183, https://doi.org/10.1016/j.labeco.2019.03.002.
[29] Causa, O. et al. (2025), “Labour shortages and labour market inequalities: Evidence and policy implications”, OECD Economics Department Working Papers, No. 1832, OECD Publishing, Paris, https://doi.org/10.1787/14e62ec0-en.
[38] Charness, N. and S. Czaja (2006), “Older Workers Training: What We Know and Don’t Know”, Washington, DC: AARP.
[19] Curtis, E., L. O’Kane and R. Park (2023), “Workers and the Green-Energy Transition: Evidence from 300 Million Job Transitions”, National Bureau of Economic Research, Vol. 5, https://doi.org/10.3386/w31539.
[35] Dearden, L., H. Reed and J. van Reenen (2006), “The Impact of Training on Productivity and Wages: Evidence from British Panel Data”, Oxford Bulletin of Economics and Statistics, https://cep.lse.ac.uk/textonly/people/vanreenen/papers/OBESArticle.pdf.
[6] Desjardins, R. and A. Warnke (2012), “Ageing and Skills: A Review and Analysis of Skill Gain and Skill Loss Over the Lifespan and Over Time”, OECD Education Working Papers, No. 72, OECD Publishing, Paris, https://doi.org/10.1787/5k9csvw87ckh-en.
[26] Dustmann, C. and C. Meghir (2005), “Wages, Experience and Seniority”, The Review of Economic Studies, Vol. 72/1, pp. 77-108, https://doi.org/10.1111/0034-6527.00325.
[60] Ecoplan (2024), Évaluation de viamia: Mise en œuvre et effets de l’offre, https://formationprofessionnelle2030.ch/images/projekte/viamia/Rapport_final_Evaluation_de_viamia_2023_Ecoplan.pdf (accessed on 23 May 2025).
[3] ESM (2024), Population ageing and productivity: The innovation channel, https://www.esm.europa.eu/ESM-briefs/population-ageing-and-productivity-innovation-channel.
[83] EU-OSHA, Cedefop, Eurofound and EIGE (2017), Towards age-friendly work in Europe: a life-course perspective on work and ageing from EU Agencies, Publications Office of the European Union, Luxembourg, https://www.eurofound.europa.eu/en/publications/2017/towards-age-friendly-work-europe-life-course-perspective-work-and-ageing-eu.
[69] Eurofound (2017), Changing places: Mid-career review and internal mobility, Publications Office of the European Union, Luxembourg, https://data.europa.eu/doi/10.2806/42599.
[64] Eurostat (2025), Employment by sex, age and citizenship, https://doi.org/10.2908/LFSA_EGAN (accessed on 23 May 2025).
[18] Eurostat (2024), Digital skills in 2023: impact of education and age, https://ec.europa.eu/eurostat/web/products-eurostat-news/w/ddn-20240222-1.
[27] Gonzaga, G. and T. Guanziroli (2019), “Returns to experience across tasks: evidence from Brazil”, Applied Economics Letters, Vol. 26/20, pp. 1718-1723, https://doi.org/10.1080/13504851.2019.1593927.
[17] Green, A. (2019), “What is happening to middle skill workers?”, OECD Social, Employment and Migration Working Papers, No. 230, OECD Publishing, Paris, https://doi.org/10.1787/a934f8fa-en.
[13] 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.
[11] Hellerstein, J., D. Neumark and K. Troske (1999), “Wages, Productivity, and Worker Characteristics: Evidence from Plant‐Level Production Functions and Wage Equations”, Journal of Labor Economics, Vol. 17/3, pp. 409-446, https://doi.org/10.1086/209926.
[4] Henriksen, E. and T. Cooley (2018), Demographics and long-run growth, https://cepr.org/voxeu/columns/demographics-and-long-run-growth.
[48] Herrbach, O. et al. (2009), “Perceived HRM practices, organizational commitment, and voluntary early retirement among late-career managers”, Human Resource Management, Vol. 48/6, pp. 895-915, https://doi.org/10.1002/hrm.20321.
[21] Hudomiet, P. and R. Willis (2022), “Computerization, obsolescence and the length of working life”, Labour Economics, Vol. 77, p. 102005, https://doi.org/10.1016/j.labeco.2021.102005.
[32] Humlum, A., J. Munch and P. Plato (2023), “Changing Tracks: Human Capital Investment after Loss of Ability”, https://docs.iza.org/dp15992.pdf.
[1] IMF (2016), Why Productivity Growth is Faltering in Aging Europe and Japan, https://www.imf.org/en/Blogs/Articles/2016/12/09/why-productivity-growth-is-faltering-in-aging-europe-and-japan.
[31] ITF (2023), “Adapting (to) automation: Transport workforce in transition”, International Transport Forum Policy Papers, No. 122, OECD Publishing, Paris, https://doi.org/10.1787/905fdc2c-en.
[42] Jayachandran, S. and A. Lleras-Muney (2009), “Life Expectancy and Human Capital Investments: Evidence from Maternal Mortality Declines”, The Quarterly Journal of Economics, Vol. 124/1, pp. 349-397, https://doi.org/10.1162/qjec.2009.124.1.349.
[54] J-PAL (2015), Public or Private? Job Counseling in France, https://www.povertyactionlab.org/sites/default/files/publication/public-or-private-job-counseling-in-france_0.pdf.
[41] Kalemli-Ozcan, S., H. Ryder and D. Weil (2000), “Mortality decline, human capital investment, and economic growth”, Journal of Development Economics, Vol. 62/1, pp. 1-23, https://doi.org/10.1016/S0304-3878(00)00073-0.
[33] Katz, L. et al. (2022), “Why Do Sectoral Employment Programs Work? Lessons from WorkAdvance”, Journal of Labour Economics, Vol. 40, https://doi.org/10.1086/717932.
[5] Kenny, G. et al. (2008), “Physical work capacity in older adults: implications for the aging worker”, Am J Ind Med, Vol. 51/8, pp. 610-25, https://doi.org/10.1002/ajim.20600.
[20] Kogan, L. et al. (2022), “Technology, Vintage-Specific Human Capital, and Labor Displacement: Evidence from Linking Patents with Occupations”, https://doi.org/10.3386/w29552.
[49] Kristensen, N. (2012), “Training and Retirement”, IZA, Institute for the Study of Labor (IZA), Bonn, https://nbn-resolving.org/urn:nbn:de:101:1-201204239412%0a.
[12] Lazear, E. et al. (2023), “Productivity and Wages: What Was the Productivity–Wage Link in the Digital Revolution of the Past, and What Might Occur in the AI Revolution of the Future?”, in Research in Labor Economics, 50th Celebratory Volume, Emerald Publishing Limited, https://doi.org/10.1108/s0147-912120230000050007.
[14] Lechner, C. et al. (2021), “Stability and change in adults’ literacy and numeracy skills: Evidence from two large-scale panel studies”, Personality and Individual Differences, Vol. 180, https://doi.org/10.1016/j.paid.2021.110990.
[16] Lewandowski, P. et al. (2020), “Ageing of routine jobs in Europe”, Economic Systems, Vol. 44/4, p. 100816, https://doi.org/10.1016/j.ecosys.2020.100816.
[24] Maestas, N. et al. (2023), “The Value of Working Conditions in the United States and the Implications for the Structure of Wages”, American Economic Review, Vol. 113/7, pp. 2007-47, https://doi.org/10.1257/aer.20190846.
[55] Mathematica (2019), New Study Finds Positive Impacts of Staff Assistance, Inconclusive Evidence of Training Effectiveness for Participants in Two Large Publicly Funded Employment Programs, https://www.mathematica.org/news/new-study-finds-positive-impacts-of-staff-assistance-inconclusive-evidence-of-training-effectiveness (accessed on 16 May 2025).
[39] Montizaan, R., F. Cörvers and A. De Grip (2010), “The effects of pension rights and retirement age on training participation: Evidence from a natural experiment”, Labour Economics, Vol. 17/1, pp. 240-247, https://doi.org/10.1016/j.labeco.2009.10.004.
[66] National Careers Institute (2022), Late career, https://www.yourcareer.gov.au/resources/research/late-career (accessed on 16 May 2025).
[45] Neumark, D., I. Burn and P. Button (2019), “Is It Harder for Older Workers to Find Jobs? New and Improved Evidence from a Field Experiment”, Journal of Political Economy, Vol. 127/2, pp. 922-970, https://doi.org/10.1086/701029.
[77] OECD (2025), Advancing Adult Skills through Individual Learning Accounts: A Step-by-Step Guide for Policymakers, Getting Skills Right, OECD Publishing, Paris, https://doi.org/10.1787/08e1bdaf-en.
[68] OECD (2024), “Career paths and engagement of mature workers”, OECD, Paris, https://www.oecd.org/content/dam/oecd/en/topics/policy-issues/ageing-and-employment/Career-Paths-and-Engagement-of-Mature-Workers.pdf.
[10] OECD (2024), Do Adults Have the Skills They Need to Thrive in a Changing World?: Survey of Adult Skills 2023, OECD Skills Studies, OECD Publishing, Paris, https://doi.org/10.1787/b263dc5d-en.
[9] OECD (2024), OECD Employment Outlook 2024: The Net-Zero Transition and the Labour Market, OECD Publishing, Paris, https://doi.org/10.1787/ac8b3538-en.
[37] 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.
[67] OECD (2024), Promoting Better Career Mobility for Longer Working Lives in the United Kingdom, Ageing and Employment Policies, OECD Publishing, Paris, https://doi.org/10.1787/2b41ab8e-en.
[57] OECD (2022), Strengthening Career Guidance for Mid-Career Adults in Australia, Getting Skills Right, OECD Publishing, Paris, https://doi.org/10.1787/e08803ce-en.
[56] OECD (2021), Career Guidance for Adults in a Changing World of Work, Getting Skills Right, OECD Publishing, Paris, https://doi.org/10.1787/9a94bfad-en.
[81] OECD (2019), Negotiating Our Way Up: Collective Bargaining in a Changing World of Work, https://doi.org/10.1787/1fd2da34-en.
[8] OECD (2019), OECD Employment Outlook 2019: The Future of Work, OECD Publishing, Paris, https://doi.org/10.1787/9ee00155-en.
[22] OECD (2019), Working Better with Age, Ageing and Employment Policies, OECD Publishing, Paris, https://doi.org/10.1787/c4d4f66a-en.
[84] OECD (2013), OECD Skills Outlook 2013: First Results from the Survey of Adult Skills, OECD Publishing, Paris, https://doi.org/10.1787/9789264204256-en.
[43] OECD/Generation: You Employed, Inc. (2023), The Midcareer Opportunity: Meeting the Challenges of an Ageing Workforce, OECD Publishing, Paris, https://doi.org/10.1787/ed91b0c7-en.
[30] Oliveira Martins, J. et al. (2005), “The Impact of Ageing on Demand, Factor Markets and Growth”, OECD Economics Department Working Papers, No. 420, OECD Publishing, Paris, https://doi.org/10.1787/545827207132.
[7] Paccagnella, M. (2016), “Age, Ageing and Skills: Results from the Survey of Adult Skills”, OECD Education Working Papers, No. 132, OECD Publishing, Paris, https://doi.org/10.1787/5jm0q1n38lvc-en.
[36] Picchio, M. (2021), “Is training effective for older workers?”, IZA World of Labor, https://doi.org/10.15185/izawol.121.v2.
[50] Piccio, M. and J. van Ours (2013), “Retaining through training even for older workers”, Economics of Education Review, Vol. 32, pp. 29-48, https://doi.org/10.1016/j.econedurev.2012.08.004.
[47] Posthuma, R. and M. Campion (2009), “Age Stereotypes in the Workplace: Common Stereotypes, Moderators, and Future Research Directions”, Journal of Management, Vol. 35, pp. 158-188, https://doi.org/10.1177/0149206308318617.
[23] Saez, E., B. Schoefer and D. Seim (2023), “Deadwood Labor? The Effects of Eliminating Employment Protection for Older Workers”, NBER Working Papers, https://www.nber.org/papers/w31797.
[79] Schirmer, W. et al. (2022), “Digital skills training for older people: The importance of the ‘lifeworld’”, Archives of Gerontology and Geriatrics, Vol. 101, https://doi.org/10.1016/j.archger.2022.104695.
[58] Skills Development Scotland (n.d.), Scotland’s Careers Services, https://www.skillsdevelopmentscotland.co.uk/what-we-do/scotlands-careers-services/our-centres.
[74] Skills Future Singapore (2025), SkillsFuture Mid-Career Enhanced Subsidy, https://www.skillsfuture.gov.sg/initiatives/individuals/enhancedsubsidy (accessed on 23 May 2025).
[75] SkillsFuture Singapore (2025), SkillsFuture Level-Up Programme, https://www.myskillsfuture.gov.sg/content/portal/en/career-resources/career-resources/education-career-personal-development/SkillsFuture_Level-Up_Programme.html?_gl=1*ptwrgi*_gcl_au*MzY4Nzc4OTI3LjE3MzY3ODA1MTU (accessed on 23 May 2025).
[73] Solas (n.d.), Skills to Advance, https://www.solas.ie/programmes/skills-to-advance/employees/.
[52] Stenberg, A., X. de Luna and O. Westerlund (2012), “Can adult education delay retirement from the labour market?”, Journal of Population Economics, Vol. 25, pp. 677–696, https://doi.org/10.1007/s00148-010-0350-8.
[59] The Federal Department of Economic Affairs, Education and Research (n.d.), Viamia, https://viamia.ch/fr/#kontakt (accessed on 23 May 2025).
[78] Unionlearn (2024), Union Learning Reps, https://www.unionlearn.org.uk/union-learning-reps-ulrs (accessed on 16 May 2025).
[63] VDAB (2024), Cijfergegevens Volledig, https://extranet.vdab.be/system/files/media/bestanden/2025-01/Cijfergegevens%202024.pdf (accessed on 23 May 2025).
[62] VDAB (n.d.), Alles over loopbaancheques, https://www.vdab.be/orienteren/loopbaanbegeleiding/alles-over-loopbaancheques (accessed on 23 May 2025).
[61] VDB (2021), Monitoringsrapport Loopbaanbegeleiding, https://extranet.vdab.be/system/files/media/bestanden/2022-03/Monitoring%20Loopbaanbegeleiding%202021_2.pdf (accessed on 23 May 2025).
[72] Wojewódzki Urząd Pracy w Krakowie (2024), Co to jest Krajowy Fundusz Szkoleniowy? [What is the National Training Fund?], https://wupkrakow.praca.gov.pl/kfs-wup-24.
[65] Workforce Singapore (2025), Volunteer Career Advisors Initiative, https://www.wsg.gov.sg/home/individuals/attachment-placement-programmes/volunteer-career-advisors-initiative (accessed on 23 May 2025).
[80] Zwick, T. (2011), Why Training Older Employees is Less Effective, http://ftp.zew.de/pub/zew-docs/dp/dp11046.pdf.
Unadjusted literacy scores by age group and country
Note: The difference in literacy scores between 25‑44 and 55‑65 year‑olds is not statistically significant in New Zealand. *Caution is required in interpreting results due to the high share of respondents with unusual response patterns – see the Note for Poland in OECD (2024[10]), Do Adults Have the Skills They Need to Thrive in a Changing World?: Survey of Adult Skills 2023, https://doi.org/10.1787/b263dc5d-en.
Source: 2023 Survey of Adult Skills.
Percentage difference in literacy proficiency between 2012 and 2023 for those who were 25‑54 years old in 2012, controlling for gender and place of birth, by occupation
Note: Occupations shown are those for which the estimates are statistically significant at the 5% level. Weighted average of Austria, Flemish Region (Belgium), Canada, Czechia, Denmark, England (United Kingdom), Estonia, Finland, France, Germany, Ireland, Italy, Japan, Korea, the Netherlands, Norway, Poland, the Slovak Republic, Spain, Sweden and the United States.
Source: 2012 and 2023 Survey of Adult Skills.
Share of workers working physically every day by occupation
Note: Weighted average of Austria, Flemish Region (Belgium), Canada, Chile, Czechia, Denmark, England (United Kingdom), Estonia, Finland, France, Germany, Hungary, Ireland, Israel, Italy, Japan, Korea, Latvia, Lithuania, the Netherlands, New Zealand, Norway, Poland, Portugal, the Slovak Republic, Spain, Sweden, Switzerland and the United States.
Source: 2023 Survey of Adult Skills.
Note: High-skilled occupations include Managers, Professionals and Technicians and Associate Professionals. Panel A shows the share of workers who report that their job involves working physically for a long period every day. *Caution is required in interpreting results due to the high share of respondents with unusual response patterns – see the Note for Poland in OECD (2024[10]), Do Adults Have the Skills They Need to Thrive in a Changing World?: Survey of Adult Skills 2023, https://doi.org/10.1787/b263dc5d-en.
Source: 2012 and 2023 Survey of Adult Skills.
Note: Weighted average of Austria, Flemish Region (Belgium), Canada, Chile, Czechia, Denmark, England (United Kingdom), Estonia, Finland, France, Germany, Hungary, Ireland, Israel, Italy, Japan, Korea, Latvia, Lithuania, the Netherlands, New Zealand, Norway, Poland, Portugal, the Slovak Republic, Spain, Sweden, Switzerland and the United States.
Source: 2023 Survey of Adult Skills.
Share of the population that participated in training less than they wanted to
Note: Weighted average of Austria, Flemish Region (Belgium), Canada, Chile, Czechia, Denmark, England (United Kingdom), Estonia, Finland, France, Germany, Hungary, Ireland, Israel, Italy, Japan, Korea, Latvia, Lithuania, the Netherlands, New Zealand, Norway, Poland, Portugal, the Slovak Republic, Spain, Sweden, Switzerland and the United States.
Source: 2023 Survey of Adult Skills.
Share of workers who were learning-by-doing at least weakly, 25‑54 year‑olds and 55‑65 year‑olds
Note: *Caution is required in interpreting results due to the high share of respondents with unusual response patterns – see the Note for Poland in OECD (2024[10]), Do Adults Have the Skills They Need to Thrive in a Changing World?: Survey of Adult Skills 2023, https://doi.org/10.1787/b263dc5d-en.
Source: 2023 Survey of Adult Skills.
← 1. In the first cycle of OECD Survey of Adult Skills, problems solving in technology-rich environment was measured. This concept was defined as “the ability to use digital technology, communication tools and networks to acquire and evaluate information, communicate with others and perform practical tasks” (OECD, 2013[84]).
← 2. In the 2023 Survey of Adult Skills, adaptive problem-solving was measured instead of problem-solving in a technology-rich environment. Adaptive problem solving was defined as: “the capacity to achieve one’s goals in a dynamic situation in which a method for solution is not immediately available. It requires engaging in cognitive and metacognitive processes to define the problem, search for information, and apply a solution in a variety of information environments and contexts” (OECD, 2024[10]).
← 3. On average, literacy and numeracy skills decreased compared to 2012 for a weighted sample of data from Austria, Flemish Region (Belgium), Canada, Czechia, Denmark, England (United Kingdom), Estonia, Finland, France, Germany, Ireland, Italy, Japan, Korea, the Netherlands, Norway, Poland, the Slovak Republic, Spain, Sweden and the United States
← 4. However, caution should be taken when extrapolating from the wage‑skills link to the productivity-skills relationship, as recent literature questions the equivalence between wages and productivity (Caplin et al., 2023[82]).
← 5. Even if skills do not decline with age, people who are young today may have lower information-processing skills today and when they are older compared to young people of the future, e.g. if the future generations have higher educational attainment.
← 6. The second cycle of the Survey of Adult Skills, was designed to ensure the results are comparable with those of the first cycle. However, some innovations have been introduced to improve the content, design and delivery of the assessment and, as a result, the assessments in both cycles are not identical. Please see (OECD, 2024[10]) for more detailed information about methodological changes between the cycles.
← 7. Policy response is also needed to prevent the overall decline in information-processing skills, which might be due to e.g. lower quality of education that focuses on standardised tests rather than developing critical thinking and analytical skills, a decline in attention spans due to consumption of fast digital content or other impact of technological change.
← 8. Older workers (45‑65 year‑olds) are approximately 2 percentage points less likely to be over-qualified than younger workers (25‑44 year‑olds) on average, though this pattern is not statistically significant in many countries.
← 9. The average age of the working population rose by 3.3 years during 1980 through 2005.
← 10. However, this pattern may be the result of older workers being less likely to hold jobs that require the use of ICT even within the same 2‑digit ISCO occupation and not necessarily because they lack the necessary skills.
← 11. For example, among the 25‑29 year‑olds, earnings of the top 25% and top 10% earners are 40% and 91% higher than of the median earners, respectively. In contrast, among the 60‑65 year‑olds, the top 25% earn 67% more than the median, and the top 10% earn 3 times the median.
← 12. Among workers in physically demanding occupations (elementary occupations, plant and machine operators and assemblers), more than 30% said they will not be able to continue working until the age of 60, as compared to 20% among professionals and clerical support workers (EU-OSHA, Cedefop, Eurofound and EIGE, 2017[83]).
← 13. However, it should be noted that less physically demanding jobs may still be associated with high stress levels, leading to burnout and other mental health problems affecting workers employability and productivity.
← 14. The authors use Danish administrative data to study work accidents in a setting where individuals can either receive disability insurance payments or opt for rehabilitation benefits, which are set at the same level and are available to those who participate in formal education or vocational training with firms. Authors exploit differences in eligibility driven by prior vocational training and find that reskilled workers earn 24% more than before their injuries and do not end up on antidepressants. Reskilling subsidies for injured workers pay for themselves four times over.
← 15. For comparison, participation in non-formal training among employed 60‑65 year‑olds is similar to the share of 60‑65 year‑olds who are learning by doing.
← 16. Sum of those who participated in non-formal training and those who wanted to participate in training but did not.
← 17. The average effective age of labour market exit is defined as the average age of exit from the labour force for workers aged 40 and over.
← 18. Other factors, such as different age gaps in educational attainment between countries may also contribute to the age gap in training participation. Higher training participation may also lead to higher effective labour market exit age, as discussed in the subsequent paragraphs.
← 19. In 2006, the Dutch Government abolished favourable tax treatment of early retirement plans for public sector employees born in 1950 or later. This resulted in a drop in pension benefits when workers retire early and stronger incentives to continue working (Montizaan, Cörvers and De Grip, 2010[39]). In Italy, starting from 1995, access to seniority pensions was tightened, and individuals were required not only to have 35 years of contributions (sufficient condition before the reform), but also to reach a minimum age, which was progressively increased to 57 (Brunello and Comi, 2015[40]).
← 20. Average includes Argentina, Australia, Brazil, Canada, Chile, France, Germany, Italy, Mexico, New Zealand and the United States
← 21. When designing and implementing ILA, it is essential to verify that training providers comply with minimum quality requirements to avoid fraud. It is also important to clearly define eligibility criteria to prevent beneficiaries from enrolling in leisure courses and to ensure that ILA-funded training is aligned with the scheme’s objectives.
← 22. The effectiveness of the schemes that offer financial incentives to participate in training, whether they target employers or individuals, will depend on the exact design features and the context in which they operate. Therefore, it is critical that countries take an experimental approach to implementing such schemes, which involves continuous monitoring of their implementation, rigorous evaluation to assess whether the schemes achieve the desired impact in a cost-effective way and continuously adapting the scheme based on the results.