The 21st century is being reshaped by crises and megatrends, including pandemics, geopolitical instability, climate change, demographic shifts and rapid technological innovation. These forces create uncertainty and change skills demands, widening existing socio-economic and regional disparities in access to opportunities. This chapter defines the set of skills needed to thrive in 21st-century labour markets and highlights the importance of reducing disparities in skills development. Factors outside the control of individuals, including gender, immigrant status or socio-economic and residential backgrounds, have a significant impact on skills acquisition and returns to skills. These factors can exacerbate skills shortages, constraining opportunities for economic growth and potentially fuelling social tensions. At the same time, rapid changes in technologies, economies and societies compress the window for policy adaptation. This chapter details how agile lifelong learning policy responses that are evidence-based, engage stakeholders, and are timely and responsive to 21st-century challenges can ensure that what appear as constraints can instead become opportunities for individuals to thrive.
OECD Skills Outlook 2025
Abstract
The challenges of the 21st century require timely policy solutions that can be adapted to respond to rapidly evolving circumstances. A central goal is to reduce skills disparities driven by circumstances beyond individuals’ control, such as gender, parental education and occupation, immigrant background, age, and childhood residential context. Addressing them supports economic growth, innovation and the fairer distribution of benefits. Identifying the policy interventions needed to reduce skills disparities first requires mapping how skills are distributed across individuals with different socio-demographic characteristics, and which factors determine such disparities. By putting into place the initiatives detailed in this report, policymakers can make a concrete difference to the lives of their populations by reducing skills disparities and disparities in the returns to skills.
Agile governance is key. Lifelong learning policies can mitigate disparities between groups. Such policies should be flexible, data-driven and iterative, enabling rapid adjustments to address rapidly shifting circumstances based on continuous monitoring and feedback mechanisms. Effective systems rely on strategic collaborations between national and local governments, social partners, and learners.
Lifelong learning must be firmly embedded within broader economic development and innovation strategies. Effective lifelong learning frameworks integrate education at all stages, from early childhood through compulsory education, tertiary education and ongoing adult training. Collaboration across ministries aligns education and training goals with broader governance strategies, such as those related to digitalisation, health and well-being, or inclusion, resulting in greater efficiency and improving the outcomes of programmes and policies.
Investments should be made in initiatives that foster educational and occupational mobility. Collaboration with stakeholders should be reflected in investments to foster educational and occupational mobility, such as accessible education pathways that support horizontal and vertical mobility between vocational education, academic programmes, and adult upskilling and reskilling initiatives. This approach helps maintain workforce adaptability, ensuring that the supply of skills remains closely aligned with changing labour market requirements.
Reducing disparities related to socio-demographic characteristics from the earliest stages of education is essential. Skills disparities begin to take shape early through differences in access and quality of early childhood education and care (ECEC) services. The effects of these disparities on individuals’ skills only widen with time. Investment in high-quality, comprehensive and universally accessible ECEC systems, combined with targeted support measures, can reduce the emergence of skills gaps that originate from circumstances that lie beyond individuals’ control. Integrating services such as health, nutrition and parental employment support within early education provision further strengthens these interventions, particularly for disadvantaged communities.
Compulsory schooling systems must balance excellence, equity and labour market relevance. Funding formulas need to take into account skills disparities, while curricula should ensure academic proficiency in core 21st-century skills, social and emotional skills, and career readiness. Initiatives that promote inclusive and stereotype-free education, particularly those related to gender and social mobility, help reduce occupational segregation and improve long-term employment outcomes.
Adult training programmes must balance access, quality and labour market relevance. Working adults or those with caretaking responsibilities face significant time and financial constraints to accessing training, posing and worsening barriers to their uptake. Existing adult education systems frequently favour individuals already advantaged by prior educational attainment, thus exacerbating skill divides. It is essential to ensure quality and relevance through rigorous accreditation processes, outcome-oriented funding mechanisms and robust provider accountability frameworks. Targeted support for underserved populations, including older adults, migrants, and individuals with low levels of skills proficiency or low levels of formal education, should also be prioritised.
Career guidance can be an important route to better opportunities and outcomes for individuals, while also helping to fill growing skill and labour shortages. Comprehensive and bias-free career guidance, including information and counselling, should be expanded and modernised to address information gaps and tackle persistent stereotypes. Career counselling capacities, especially in disadvantaged schools and public employment services, could be enhanced through the use of digital and personalised support services that channel people towards training and facilitate labour market entry.
A skills-first approach in labour markets could help reduce skills shortages and promote inclusion. Governments can lead by example, for instance by integrating skills-based hiring methods into public sector recruitment practices. They can also encourage employers to adopt skills-based hiring models through incentives. Standardised frameworks for skills assessment, recognition and credential portability should be developed to ensure transparency and facilitate job mobility. Together with policies related to adult training, skills-first talent management practices within firms can promote lifelong learning and broader policy goals related to innovation and addressing labour market shortages.
The first quarter of the 21st century has been marked by a series of acute crises – including global pandemics, geopolitical conflicts and economic disruptions – and long-term systemic changes, often referred to as megatrends, such as climate change, demographic shifts and digital innovations. The interplay between these crises and megatrends creates complex challenges for skills development, with sudden disruptions amplifying existing disparities, and structural changes redefining skills demands and altering the labour market.
For instance, although technological progress can drive economic growth, it may also exacerbate disparities in income and labour market opportunities, as uneven access to digital tools and education opportunities will mean that some individuals will be less able to become more productive through the use of technology. Disparities could also arise from differences in the suitability of technology to enhance the productivity of human work in certain, but not all, sectors and occupations, leading to divergent patterns in economic and labour market opportunities for workers. Climate change and its mitigation policies increase demand for certain skills while reducing demand for other skills, with transformations needed to reach net-zero targets creating opportunities for innovation, job creation and social progress. However, climate change also risks deepening regional inequalities, with lower-income areas often bearing disproportionate environmental burdens and lacking the resources for effective adaptation. Demographic shifts, particularly population ageing in high-income countries, further strain labour markets, underscoring the urgent need for policies that address emerging skills gaps. Understanding these forces, how they relate to each other and how they impact skills demands is essential for designing skills policies that respond to the evolving demands of 21st-century economies. By responding to these challenges effectively, socio-economic disparities in life chances both between and within countries can be reduced.
However, the rapidity of changes compresses the window for policy adaptation, stretching conventional governance cycles. In practice, education and training systems, which set curricula, qualifications frameworks and funding agreements often several years in advance, find it difficult to redesign programmes, upskill instructors and scale delivery quickly enough to meet newly emerging skills requirements. Employment services, industrial strategies and social protection instruments face similar challenges: they must process labour market intelligence, align incentives and deploy resources quickly; however, they frequently rely on data and decision-making structures that are not well-suited to rapid decision making and shifts in service provision. These timing mismatches risk creating skills shortages, reducing the potential for economic growth and diminishing the return on public and private investments. If skills policies are to keep pace with rapid 21st-century economic and societal changes, more agile policy architectures, underpinned by timely data, iterative evaluation and close collaboration between government, employers, social partners and training providers, must be developed.
1.1.1. Increased uncertainty across the globe
Crises exert both direct and indirect effects on economies and societies by disrupting everyday activities and undermining the stability of institutional frameworks. They force economic and social actors to recognise the possibility of disruption and, by so doing, generate uncertainty. Perceptions of uncertainty appear to be higher and more volatile in the first quarter of the 21st century than in the last decades of the 20th century (Ahir, Bloom and Furceri, 2022[1]).
The rising frequency of crises (and the associated increase in uncertainty) could itself be viewed as a megatrend. Historians such as Adam Tooze argue that the global context is shaped by a “polycrisis” of overlapping and compounding challenges (Henig and Knight, 2023[2]). This perspective reflects the volatility that has marked the late 20th and early 21st centuries. However, others caution against viewing current turmoil as uniquely permanent (Frankopan, 2024[3]), and argue that from a longer historical perspective, periods of acute disruption have arisen in the past and may not necessarily imply a lasting transformation. While the short-term impacts of crises are clear, their long-term significance as a defining trend in the 21st century remains uncertain.
Against a background of interlinked crises and pervasive uncertainty, designing and implementing skills policies becomes increasingly complex. At the same time, heightened public awareness about the risks for economies and societies of systemic vulnerabilities can serve as a catalyst for improving skills policies. The 20th century saw the primacy of information-processing skills – numeracy, literacy and adaptive problem solving – established at the societal level, fuelling changes in many education systems, which adjusted their teaching to support the promotion of these skills. The 21st century could be the century that urges societies to embrace a more inclusive definition of what basic skills are and that values and recognises diverse talents and contributions, challenging policymakers, educators and employers to develop innovative approaches to broader skills development efforts. Therefore, although greater uncertainty poses risks and may increase social disparities, it also provides an opportunity to advance skills policies that benefit a wider set of adults.
1.1.2. Societal and demographic changes
Over the course of the 20th century, OECD countries experienced profound changes in their demographic profiles. A key development was the marked expansion of educational opportunities, which notably benefited women (Barro and Lee, 2013[4]; OECD, 2022[5]). This shift reflected growing awareness of education as a critical driver of economic growth and social mobility, and set the stage for higher female participation in the labour market and their greater voice in political decision making and representation in government.
Family structures also underwent major transformations during this period. For instance, dual-earner couples with higher levels of education tended to have fewer children than previous generations, a pattern that scholars have linked to changing family norms, delayed marriage and the realisation of women’s career aspirations (Lesthaeghe, 2010[6]). Fertility rates declined overall (Bongaarts, 2009[7]), while heightened standards of living, advances in medical care and better access to healthcare contributed to increased life expectancy. Together, these changes reshaped household compositions, altered population age structures, and introduced new pressures on pension systems and care provision for older adults.
Urbanisation, driven by industrialisation and the search for employment opportunities in increasingly knowledge-based economies, has been another significant demographic trend of the 20th century. The movement from rural to urban areas has led to important societal changes, including the growth of metropolitan centres and new spatial patterns of work and residence. Meanwhile, international migration flows have intensified, adding another layer of complexity to demographic evolution in OECD countries and increasing ethnic and cultural diversity.
In the 21st century, many of these demographic shifts appear set to continue, although with different impacts across countries. In particular, falling fertility rates in OECD countries, combined with the high birth rates following the Second World War, are creating tensions as a shrinking pool of young workers struggles to support rising pension costs and the costs of public healthcare systems (United Nations, 2024[8]). Coupled with rising life expectancy, these trends heighten the need for policies to equip people of all ages with relevant skills to ensure that they remain available in the labour market for longer. Encouraging older adults to remain in the workforce longer is becoming increasingly common in many OECD countries to reduce dependency ratios and labour shortages. However, this will require substantial improvements in the provision, quality and attractiveness of lifelong learning opportunities so that workers can continue to adapt to new technologies and changing work environments.
As a result of all these factors, skills policies in the 21st century will need to be far-reaching and flexible. Beyond traditional academic pathways, greater investment in vocational education and training (VET), online learning, and credential-updating programmes will be vital to accommodate a workforce that is both ageing and increasingly diverse (OECD, 2021[9]). Policies promoting continuous reskilling – particularly for mid-career and older workers – will be critical, as will tailored support for migrants and vulnerable groups – such as adults with low levels of formal educational qualifications – to ensure that they can fully participate in the labour market. In this context, broadening what is counted as a “relevant skill” – from digital proficiency to interpersonal communication – might also prove necessary for effectively promoting economic and social well-being.
1.1.3. Social mobility and inequality
During the 20th century, social mobility and economic inequality among OECD countries changed drastically as a result of industrialisation, expansions in educational opportunities, globalisation and technological development. In particular, while the mid-20th century witnessed an unprecedented increase in opportunities for upward social mobility and a narrowing of income and wealth gaps, the last quarter of the century saw the re-emergence of pronounced income disparities and the solidification of class boundaries (OECD, 2018[10]). Education emerged as both a driver of opportunity and a mechanism of stratification, with skill-biased technological change – i.e. when new technologies make workers with certain skills much more productive, and therefore more in demand, than those without such skills – amplified the returns to advanced qualifications, while systemic disparities in educational quality perpetuated intergenerational disadvantage for many (Acemoglu, 2002[11]; Berman, Bound and Machin, 1998[12]).
Social mobility can be understood in two distinct ways: absolute mobility, which reflects changes in the socio-economic status of successive generations, and relative mobility, which reflects the likelihood that individuals with lower socio-economic backgrounds can catch up with those from more advantaged families. Rising absolute mobility may coexist with low relative mobility if the benefits of economic growth primarily accrue to those already in favourable positions.
The decades after the end of Second World War marked a “golden age” of social mobility in many OECD countries. Rapid industrial expansion, strong unionisation and progressive taxation created conditions in which 60% of children born into working-class families achieved a higher occupational status than their parents (Causa and Johansson, 2009[13]). Cohorts born between 1950 and 1970 experienced 22% higher upward mobility rates than pre-war generations. In the Nordic countries in particular, parents’ level of education was closely and consistently linked to positive outcomes for their children, showing a nearly direct progression from one generation to the next (Causa and Johansson, 2009[13]). The post-war period was also characterised by a convergence in living standards, facilitated by progressive taxation, the establishment of social welfare programmes and a marked increase in labour productivity.
However, from the mid-1970s onwards, while absolute poverty continued to decline in many parts of the world, relative inequality within countries increased (Cingano, 2014[14]). Moreover, a combination of deindustrialisation and skill-biased technological change significantly reshaped labour markets (Acemoglu, 2002[11]). For example, OECD data indicate that for cohorts born after 1975, the likelihood of upward mobility stagnated (OECD, 2018[10]). Income inequality also followed a U-shaped curve across the 20th century, with the Gini coefficient1 falling from 0.55 in 1920 to 0.30 in 1970 in advanced economies (indicating reduced income inequality) before rebounding to 0.45 in 2020 (Coady and Dizioli, 2017[15]).
The first two decades of the 21st century have been defined by increasing precariousness among lower- and middle-income groups, alongside soaring incomes and wealth concentration at the very top of the distributions. New digital technologies and the transition to a knowledge-based economy have created opportunities for some, particularly those with in-demand technical and information-processing skills. However, many have been left behind, especially in regions heavily dependent on declining industries for employment opportunities (OECD, 2024[16]) or those where access to quality education and training remain constrained (UNESCO, 2024[17]). This polarisation has exacerbated social stratification, diminishing the prospects for upward mobility.
Although correlational in nature, evidence suggests that countries with higher levels of income inequality often exhibit lower rates of mobility from one generation to the next, an association that is referred to as the Great Gatsby Curve (Corak, 2013[18]). In fact, within countries, increases in income inequality have been accompanied by declines in social mobility, possibly reflecting the fact that when income or wealth disparities are wide, access to quality education, healthcare and social networks becomes highly stratified, enabling those from more privileged socio-economic backgrounds to maintain their advantage. Disparities in access to education and training, both during childhood and throughout adulthood, play a critical role in determining whether education fulfils its potential as the great equaliser. If only those in already privileged situations can benefit from high-quality learning opportunities, both absolute and relative mobility may suffer.
1.1.4. Climate change
Throughout the 20th century, economic growth in OECD countries often occurred at the expense of the environment, resulting in high CO2 emissions and loss of biodiversity. Recognising the urgent need to address climate change, the 21st century has witnessed more determined efforts to cut greenhouse gas (GHG) emissions, reduce waste and transition to a low-carbon future. However, this shift requires individuals and firms to acquire new skills to adapt to rapidly changing labour markets. Managing this shift effectively necessitates an approach that anchors public investment in comprehensive re- and up-skilling programmes, creates clear job-to-job pathways, and establishes formal frameworks for information sharing, consultation, and collective bargaining.
Although the net effect of the green transition on total employment is projected to be modest, it will entail significant shifts across industries, occupations and regions (OECD, 2024[16]). Around 20% of workers in OECD countries are employed in green-driven occupations,2 while 6% are in GHG-intensive jobs (OECD, 2024[16]). Green-driven jobs, particularly new and emerging ones, often require higher-level skills related to management, technology and professional services, whereas jobs not driven by the green transition often call for more medium- and low-skilled competencies. Demand for interpersonal and digital skills is set to rise substantially, while the need for certain traditional manufacturing skills is expected to decline (OECD, 2023[19]). Without adequate education and training, low-carbon growth could be undermined by skills shortages.
In addition to reducing emissions, green-driven occupations tend to offer higher wages overall, although this advantage mostly applies to jobs that typically require tertiary-level qualifications and high levels of information-processing skills. By contrast, individuals in green-driven roles that do not require tertiary-level qualifications and associated skills proficiency are often paid less than similar roles in GHG-intensive or neutral sectors (OECD, 2024[16]). This is problematic, as attracting workers to these occupations is essential to avoid skills becoming a severe bottleneck for delivering the net-zero transition. Displaced workers in high-emission industries are often male, older and have lower educational attainment, making it harder for them to transition to similarly well-paid roles (Barreto et al., 2024[20]; OECD, 2024[16]).
1.1.5. Technological developments
Empirical estimates of the impact of digital technologies on employment in the 20th century and early part of the 21st century are mixed: whereas some reveal that technological developments have led to a growth in employment opportunities (Dixon, Hong and Wu, 2021[21]; Koch, Manuylov and Smolka, 2021[22]), others suggest that technological developments have reduced employment possibilities for workers (Acemoglu and Restrepo, 2020[23]). Overall, empirical evidence suggests that past waves of technological developments did not lead to overall lower employment opportunities and net job destruction in the long run (OECD, 2019[24]). In fact, throughout the 20th century, the employment-to-population ratio rose and the unemployment rate did not change (Autor, 2015[25]).
In the past, job losses resulting from computer automation tended to be more pronounced for low-wage occupations, occupations in the manufacturing sector (Mann and Püttmann, 2023[26]) and generally among workers conducting routine work (Gaggl and Wright, 2017[27]). As a result of past waves of technological progress, today’s workplaces demand people who can solve non-routine problems. Few workers, whether in manual or knowledge-based occupations, use only repetitive actions to perform their job tasks. As technologies capable of performing rule-based tasks were introduced, the importance of people’s ability to solve complex problems that could not simply be solved by applying pre-specified rules grew. While computers gradually took over “the expected”, individuals increasingly had to deal with the “unexpected and the unfamiliar”, often working alongside computers (Autor, Levy and Murnane, 2003[28]). Whereas technological developments in the past led to the creation of computers and robots that could only follow narrowly specified rules, today, machine learning algorithms allow automata to perform a considerably broader set of tasks that lack rule-based solutions. As a result, the set of tasks that can be performed by technologies is radically different. The advent of artificial intelligence (AI) systems may dramatically change the demand for skills in the future as non-routine tasks fall within the scope of what automata can perform reliably. On the one hand, technology may obviate the need for humans to perform certain tasks. On the other hand, technologies may complement humans, requiring workers to learn to work effectively with new technologies as some tasks, but not all, will be affected by automation.
Estimates suggest that as many as 80% of the US workforce could have at least one in ten tasks affected by the use of large language models (LLMs), and around two in five employees might experience an impact on at least half of their tasks (Eloundou et al., 2023[29]). Recent analyses of online job advertisements in the United Kingdom and the United States indicate that between 2021 and 2024, skills requirements for the average job changed by approximately one-third, with one in four jobs experiencing shifts in up to three-quarters of required skills (Lightcast, 2025[30]; 2025[31]). Moreover, in the United States, the pace of change between 2021 and 2024 (a three-year period) matched that observed between 2016 and 2021 (a five-year period) (Lightcast, 2025[31]). The most disrupted occupations generally require extensive training or tertiary qualifications. In contrast, jobs experiencing minimal changes in skill requirements do not typically require advanced qualifications and frequently involve demanding physical tasks.
The labour market implications of LLMs are not yet fully understood and may shift over time, depending on technological developments and policy decisions. Early evidence points to mixed effects, with LLMs both substituting and complementing human labour. As the technology advances, LLMs are becoming capable of performing a wider range of tasks with greater speed and proficiency. Evidence from early 2025 shows that LLMs can substitute for human labour in tasks such as writing and translation (Demirci, Hannane and Zhu, 2025[32]; Qiao, Rui and Xiong, 2024[33]), and have also been shown to complement human skills and enhance the productivity of less experienced workers (Brynjolfsson, Li and Raymond, 2025[34]; Noy and Zhang, 2023[35]). In some cases, productivity gains arise when workers possess AI-related knowledge, particularly of machine learning systems, enabling them to command higher wages than peers with similar skills but without such expertise (Stephany and Teutloff, 2024[36]). Ultimately, policy decisions around adoption and regulation will determine their broader impact on workers and labour markets (Autor, 2024[37]).
The emerging literature on generative AI points to a complex interplay between displacement and augmentation. Some scholars warn of widespread automation of middle-skill tasks (Acemoglu and Restrepo, 2018[38]; Frey and Osborne, 2017[39]), while others highlight the potential for new job creation when AI is used to enhance, rather than replace, human capacities (Brynjolfsson and Mitchell, 2017[40]). Whether LLMs ultimately displace existing work or foster new forms of employment likely depends on how the technologies evolve and the related organisational adoption practices, regulatory frameworks and institutional incentives put in place (Autor, 2024[37]).
Evidence from consultancy markers and online freelancers, for which rapid shifts in skills demands and employment opportunities are more easily observable, indicate that demand tends to dry up for less experienced workers, whereas demand for more experienced workers remains sustained, with an increase in workers taking on complex tasks that currently lie beyond the capabilities of AI systems (Demirci, Hannane and Zhu, 2025[32]; Lysyakov and Viswanathan, 2021[41]; Teutloff et al., 2025[42]). On the supply side, workers in occupations highly exposed to generative AI substitution may suffer reductions in earnings and employment opportunities (Hui, Reshef and Zhou, 2023[43]; Liu et al., 2024[44]), whereas those in jobs complemented by AI may increase their earnings (Qiao, Rui and Xiong, 2024[33]). Current evidence, including the OECD’s (2024[45]), indicates that negotiated adoption with worker consultation and training provision (including dedicated training time) is associated with better outcomes for workers and can help steer AI toward augmentation rather than displacement.
These findings can be reconciled by distinguishing task complexity and, by proxy, required experience: in online consultancy marketplaces, decreases in demand are especially pronounced for short-term jobs and for jobs requiring novice workers to perform tasks complementary to the capabilities of existing LLMs. These results imply that the increased productivity of novice workers within firms (Brynjolfsson, Li and Raymond, 2025[34]) may reduce the need for novice freelancers. At the same time, the developing capabilities of LLMs appear to increase the need for workers with experience working on complex tasks.
At the economy level, the pace of technological change is not affecting all jobs equally, as indicated in Figure 1.1. There is a trade-off between the speed of change (and consequent need for upskilling and reskilling) and wages: occupations experiencing rapid change tend to command higher-than-average wages, whereas occupations relatively unaffected by technological change tend to command below-average wages. Figure 1.1 uses data from Lightcast (2025[46]) to map, at the occupational level, how much the set of skills demanded by employers in online job advertisements changed between 2021 and 2024; it also shows monthly earnings for the same set of occupations, based on the 2023 Survey of Adult Skills (OECD, 2024[47]). Speed of change is measured through the Skills Disruption Index, a standardised index ranging from 0 to 100, where a higher score indicates a greater degree of skill change (Lightcast, 2025[46]). Occupations highly affected by technology – such as mathematicians, actuaries and statisticians, and application programmers – show relatively high rates of skills change as well as high earnings. In contrast, occupations less affected by technology, such as domestic housekeepers, show relatively low wages as well as a low rate of skills change. Figure 1.1 suggests that a willingness to upskill and reskill is critical to be able to operate in many highly paid occupations, as the skills requirements in these occupations are changing rapidly.
Note: The Skills Disruption Index was calculated using Lightcast data and measures how employer skill requirements have evolved across different occupations. It ranges from 0 to 100, where a higher score indicates a greater degree of skill change. The size of the bubbles represents the number of workers (in million), categorised as fewer than 0.5 million, between 0.5 and 5 million, and more than 5 million. Monthly earnings and the number of workers in each occupation were calculated based on data from the OECD Survey of Adult Skills.
Source: Lightcast (2025[46]), The Speed of Skill Change, https://lightcast.io/resources/research/speed-of-skill-change and OECD (2024[48]), PIAAC data and methodology, www.oecd.org/en/about/programmes/piaac/piaac-data.html.
1.1.6. Aligning skills policies with evolving global realities
As described in this section, crises and megatrends are reshaping the nature of work, influencing both the types of skills demanded and the ways in which individuals acquire, develop and update these skills. Preparing for future challenges requires not only an assessment of current skills needs and the extent to which they are met by individuals’ existing skills, but also a fundamental re-evaluation of how skills policies must adapt to evolving social and economic realities. The uncertainty generated by megatrends and crises raises questions about whether skills policies, largely devised in the 20th century, remain fit for purpose.
Promoting high levels of skills development among as many individuals as possible and across diverse socio-demographic groups is essential in the 21st century. First, given the increasing diversity of societies and the complexity of the information landscape, a range of skills enable individuals to access, critically evaluate and effectively use information. Skills allow individuals to engage constructively in civic life, reducing social divides and helping maintain trust in institutions (OECD, 2021[49]; 2024[47]). Second, in countries experiencing low fertility rates, a shrinking workforce can lead to persistent labour shortages, threatening both economic growth and existing welfare arrangements. Expanding the workforce pool requires equipping a larger share of the population with relevant skills and facilitating their entry into the labour force. Reducing disparities in skills development and removing barriers to the effective use of skills can mitigate demographic pressures, enhance adaptability to changing labour demands and sustain productivity. Third, cultivating a broad range of skills supports the development of diverse teams. When people from various backgrounds and expertise areas collaborate, they foster greater innovation and creativity – skills that drive competitiveness in an increasingly complex and uncertain marketplace.
This OECD Skills Outlook provides an overview of the skill set of adult populations in countries that took part in the 2023 Survey of Adult Skills, a product of the OECD’s Programme for the International Assessment of Adult Competencies (PIAAC). It focuses on persistent and emerging disparities, notably those associated with gender, socio-economic background and childhood residential context. The report has two main objectives. First, to help countries identify the current stock of skills in their populations and evaluate their readiness to tackle social and economic challenges; and second, to highlight opportunities for how the 21st century can deliver on the defining narrative of the 20th century: equality of opportunity and meritocracy.
The analysis of adult skills data thus incorporates historical insights on how various countries and economies have attempted to reduce inequalities and encourage social mobility through skills development. The data in this report cover adults born between 1947 and 2006, corresponding to the oldest participants (65-year-olds) in the 2012 Survey of Adult Skills (PIAAC) cycle and the youngest respondents (16-year-olds) in the 2023 cycle. This means that the parents of the oldest cohort members were born during the First World War or even at the start of the 20th century. Because the analysis focuses on the 21st century, it also considers how education systems are equipping today’s children for tomorrow’s challenges. To that end, data from adult skills assessments are supplemented with information on school-aged populations, such as from the Programme for International Student Assessment (PISA), the Trends in Mathematics and Science Study (TIMSS), the Progress in International Reading Literacy Study (PIRLS), the International Civic and Citizenship Education Study (ICCS), and the International Computer and Information Literacy Study (ICILS).
Crises and megatrends shape skills demand by influencing job creation and destruction, and they impact skills supply by determining population movements and guiding individuals’ choices about education and training. They can also redefine existing roles: if tasks within a job evolve, then the skill set required to perform that job changes. At a basic level, each worker possesses a unique combination of skills, with varying proficiency levels, and each job consists of tasks requiring specific skills and knowledge. Where demand and supply intersect, economic growth and individual well-being are enhanced. By contrast, persistent mismatches hamper both.
The value of understanding the processes through which skills develop, both within and outside formal education and training systems, increases in an environment marked by shifts in the nature of work, the rapid diffusion of new technologies and greater dependence on information access. Examining the distribution of skills across socio-demographic groups, as well as the opportunities to acquire and practice them, reveals whether individuals are well-prepared to contribute to shape the economies of the 21st century, and whether they will have the 21st-century skills required to thrive.
The term 21st-century skills began to be used in the early 2000s, as scholars, governments and others began to realise that the traditional approach to skills, which focused on gaining knowledge, needed to evolve as technology progressed, changing the nature of work. In the past 25 years, skills assessments have been developed at the national and international level, with previous OECD work detailing the commonalities and differences in how they refer to 21st-century skills (Foster and Piacentini, 2023[50]). What these frameworks have in common is that they describe 21st-century skills as being relevant or applicable in many fields and associated with higher-order skills and behaviours that represent the ability to transfer knowledge, cope with complex problems and adapt to unpredictable situations. In contrast, this report uses the term 21st‑century skills to refer to the set of skills needed to thrive in the 21st century and for which comparative evidence on representative samples of adults exists. This list is not exhaustive, as thriving in certain jobs or contexts that are growing in importance in the 21st century depends on a wide range of specialised skill sets. However, the goal of the report is to consider skills that underpin the capacity of individuals to acquire such specialised skills and to thrive in a wide range of contexts and situations.
This report considers nine 21st-century skills, as detailed in Figure 1.2. These include information-processing skills (also referred to as core 21st-century skills in this report), as well as the ancillary skills of willingness to delay gratification and the social and emotional skills of extraversion, emotional stability, agreeableness, conscientiousness and open-mindedness – measured through the Big Five personality module3 – which are increasingly considered as critical determinants of labour market success (Almlund et al., 2011[51]; Heckman, Stixrud and Urzua, 2006[52]). By integrating information-processing skills (literacy, numeracy and adaptive problem solving), willingness to delay gratification, and social and emotional skills within a single conceptual framework, this report aims to provide policymakers, educators, training providers and employers with a comprehensive evidence base to re-evaluate or redesign their skills policies.
Selected analyses are presented to illustrate young people’s readiness to tackle the emerging challenges of the 21st century, covering the domain-specific skills of science, global competence, collaborative problem solving, creative problem solving, creative thinking, financial literacy, computer and information literacy, and computational thinking. This holistic approach ensures that learners are equipped with the skills needed not only to meet immediate challenges, but also the emerging demands of a rapidly evolving economic and social landscape. Previous international assessments of adult populations have mostly focused on information-processing skills, particularly literacy, numeracy and problem solving in technology-rich environments (OECD, 2012[53]). While these skills remain essential, they must be complemented by an expanded set of skills. However, while all of the aforementioned skills are important, this report focuses on the core 21st-century skills as more data are available, enabling more meaningful analyses. Selected analyses are provided for the ancillary skills.
1.2.1. Information-processing skills
Proficiency in literacy, numeracy and adaptive problem solving was measured in the 2023 Survey of Adult Skills through a low-stakes, untimed, standardised assessment administered in respondents’ homes by trained interviewers (OECD, 2024[47]). The low-stakes nature of the assessment meant that individuals who chose not to participate in the assessment were not penalised. Similarly, performance in the assessment had no consequences for participants, who did not receive any feedback on their score. This protocol means that the assessment design does not necessarily reveal respondents’ maximal capacity to solve tasks in the relevant domains – as defined in the relevant assessment frameworks (OECD, 2021[49]) – rather, it captures the level of achievement individuals obtain by exerting the amount of effort and engagement they would typically invest in solving everyday online cognitive tasks. In effect, the scores represent “typical performance” under routine, low-stakes conditions rather than peak or optimal performance observed in high-stakes testing environments such as job searches, applications for positions into selective educational institutions, or tasks with large economic and financial consequences.
In high-stakes environments that focus on maximal performance (typically characterised by high-pressure and unavoidable consequences), test-takers often experience anxiety and a sense of hopelessness, especially when they receive no actionable feedback (Kankaraš and Suarez-Alvarez, 2019[54]). This might be especially critical for marginalised populations, who may have lower familiarity with the test format and content, or who may process information differently than their peers. By emphasising typical performance, assessments such as the Survey of Adult Skills can provide a more accurate representation of test-takers’ skills and learning needs, thereby supporting fairer educational outcomes and more targeted support.
This distinction is important, as research in educational psychology has long indicated that test performance is significantly influenced by motivational factors (Kyllonen et al., 2024[55]; Ulitzsch et al., 2021[56]). In particular, research on school-aged populations, for which results from comparable assessments administered under low-stakes and high-stakes conditions are available, indicate that the level of effort individuals put in during a testing situation can result in large differences in outcomes, with differences having important implications for the evaluation of between-group differences (Braun, Kirsch and Yamamoto, 2011[57]; Duckworth and Yeager, 2015[58]; Sackett, Borneman and Connelly, 2008[59]). Furthermore, psychometric research indicates that results in low-stakes assessments capture not only underlying cognitive abilities but also situational factors – such as test-taking motivation and engagement – that are critical to the observed performance (Messick, 1989[60]; Wise and Kong, 2005[61]).
Of particular relevance to this report is the fact that motivational and contextual factors can contribute to shaping disparities in performance both across countries and economies and within different population groups (Borgonovi, Ferrara and Piacentini, 2023[62]). If motivation and effort truly capture how adults in different countries and economies handle literacy, numeracy and adaptive problem solving in everyday life, then the differences in performance on a low-stakes test like the Survey of Adult Skills represent valid distributions. This perspective is essential for understanding cross-national differences and socio-economic disparities in information-processing skills, as it suggests that part of the variation may be attributable to differential levels of test motivation and context-specific factors. Problems would only arise if motivation and effort reflected variations in attitudes towards testing.
The importance of motivation
In 21st-century societies and workplaces, the capacity to motivate oneself emerges as a potentially critical differentiator of labour market outcomes. As AI and other technologies handle routine information-processing tasks at high speeds, human value creation increasingly hinges on the desire to engage creatively with ambiguous, open-ended challenges. This involves not being satisfied with merely passable outcomes but rather actively seeking to engage with material beyond what is strictly necessary or mandated. Such intrinsic motivation fosters innovative thinking and creative problem solving, attributes that, to date, remain beyond the full reach of automated systems.
Information-processing skills rarely develop into a fully automated process. To reason critically with mathematical content, evaluate and reflect on written text, or solve problems in dynamically evolving situations requires substantive and continuous cognitive effort. Even when individuals have gained basic levels of proficiency in a particular skill, applying such skill to many real-life tasks across domains necessitates a deliberate, ongoing investment of effort. In practice, people often default to pre-existing habits to minimise effort, implying that high-level proficiency in these domains remains linked to motivation.
Although technological advancements increasingly enable the use of monitoring devices to evaluate output and productivity with enhanced precision (Milanez, Lemmens and Ruggiu, 2025[63]), goods and services that are high in value added are increasingly “intangible”, meaning that their production is linked to workers’ intrinsic motivation. Even in environments where routine tasks are measurable, qualities such as creativity, critical thinking and the determination to exceed minimal requirements give firms a competitive edge; however, these are not easily captured by the quantitative metrics used in algorithmic management tools (Milanez, Lemmens and Ruggiu, 2025[63]). Consequently, assessing individuals’ information-processing skills in conditions that reveal their willingness to put in effort and motivation under low-stakes conditions can help countries identify the readiness of workers to remain competitive in a rapidly evolving work environment.
Recent research suggests that generative AI technologies may help to reduce disparities in cognitive ability by automating routine intellectual tasks (Dell’Acqua et al., 2023[64]). If the gap in basic information-processing skills is narrowed by such technological innovations, then differences in outcomes are more likely to be driven by how individuals engage with tasks, as well as whether they are able to harness these tools by deciding if tasks are within the ability frontier of AI applications. In this scenario, the willingness to work beyond standard requirements becomes the decisive factor in individual productivity and innovation. This shift underscores the growing importance of motivational engagement as a primary driver of success in modern workplaces.
Against this backdrop, performance in literacy and numeracy as measured in the Survey of Adult Skills declined between 2012 and 2023 – see Chapter 3 in the Survey of Adult Skills (OECD, 2024[47]). Such a decline may indicate a broader erosion of individuals’ ability to solve tasks or a decrease in motivation and engagement with challenging cognitive tasks. Irrespective of the cause, this trend is problematic when the demands of the 21st-century economy increasingly require workers to exceed routine performance. If adults are less inclined to engage deeply with complex material under low‑stakes conditions, both individual competitiveness and overall societal advancement may be adversely affected.
Literacy
In the 2023 Survey of Adult Skills, literacy is defined as “accessing, understanding, evaluating, and reflecting on written texts in order to achieve one’s goals, to develop one’s knowledge and potential, and to participate in society” (OECD, 2021, p. 42[49]). Given the prevalence of written communication in various aspects of life, proficiency in literacy is crucial for adults across their personal, social and professional spheres. Throughout the day, adults engage in diverse reading activities, from delving into extensive pieces of continuous text to swiftly scanning pages for pertinent information. These activities encompass reading emails, leaflets, timetables and instruction manuals.
Numeracy
Numeracy encompasses “accessing, using, and reasoning critically with mathematical content, information and ideas represented in multiple ways in order to engage in and manage the mathematical demands of a range of situations in adult life” (OECD, 2021, p. 19[49]). To succeed in work, life and citizenship, the skills and knowledge needed have changed. Individuals are being presented with ever-increasing amounts of information of a quantitative or mathematical nature through internet-based or technology-based resources. This information has to be located, selected, filtered, interpreted, at times questioned and doubted, and analysed for its relevance to the responses needed.
Adaptive problem solving
Adaptive problem solving involves “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, 2021, p. 19[49]). Adaptive problem solving captures the fact that individuals need to be vigilant to changes, adaptive and willing to modify their plans in pursuit of their goals. It comprises three key features that emphasise individuals’ capacity to flexibly and dynamically adapt their problem-solving strategies to a dynamically changing environment; identify and select among a range of available physical, social and digital resources; and monitor and reflect on their progress in solving problems through metacognitive processes (i.e. the ability to calibrate one’s comprehension of the problem, evaluate potential solutions and monitor progress towards the goals).
1.2.2. Willingness to delay gratification
The ability to delay gratification has long been recognised as a crucial determinant of individual and collective outcomes (Hanushek et al., 2025[65]). Willingness to delay gratification reflects how people trade off immediate benefits against potentially greater future gains, influencing personal savings, educational investment, health choices, retirement planning and many other domains of life (Chao et al., 2009[66]; Hunter et al., 2018[67]; Borghans and Golsteyn, 2006[68]; Kang and Ikeda, 2013[69]; Becker and Mulligan, 1997[70]; Ross and Mirowsky, 1999[71]). Access to high-speed internet and smartphones provides immediate gratification in the form of social media, news, entertainment and constant communication, conditioning users towards short-term rewards (Przybylski and Weinstein, 2012[72]). At the same time, rapid shifts in consumer services such as same-day delivery, instant food delivery apps and streaming media platforms foster expectations for immediacy, leading individuals to increasingly favour immediate satisfaction over long-term planning and restraint. Constant exposure to instant gratification may decrease individuals' ability to delay gratification, making the achievement of long-term goals more challenging. The modern workplace rewards employees who demonstrate long-term strategic planning, emotional intelligence and disciplined behaviour. Workers who can delay gratification are better able to persevere through challenging tasks, commit to long-term projects and invest in their professional development. Furthermore, personal growth demands the ability to delay immediate comfort for future gains. Whether acquiring new skills, pursuing health goals or achieving financial independence, delayed gratification underpins effective decision making, discipline and perseverance (Baumeister, Vohs and Tice, 2007[73]).
In the 2023 Survey of Adult Skills, respondents were asked to report how willing they would be to delay gratification, i.e. to give up something that is beneficial to them today in order to benefit more from it in the future (Falk et al., 2018[74]). The survey uses an 11-point scale ranging from (0) “completely unwilling to do so” to (10) “very willing to do so” (Falk et al., 2018[74]). This measure is easy to translate into different languages and cultural contexts. Because it does not specify what is being given up, it avoids a major criticism of the widely used “money earlier or later” (MEL) paradigm in which participants choose between monetary rewards at different points in time (Cohen et al., 2020[75]). With such time-preference experiments, empirical findings suggest that discount rates (the rate an individual is willing to delay gratification, with higher rates meaning less likely to delay) vary across different types of rewards, with primary rewards such as food being discounted more steeply than money. In other words, when given a choice between receiving something now or later, individuals are often less willing to delay gratification (or apply a higher discount rate) for certain types of rewards; for example, they would be less willing to delay gratification for food than for money.
Individuals who are more willing to delay gratification often exhibit better academic performance, higher income levels and improved health outcomes than those less willing to delay gratification. This willingness to defer immediate rewards is closely linked to self-regulatory capacities that enable individuals to invest in education, adopt healthy lifestyles and plan effectively for retirement. In contrast, those who report lower willingness to delay gratification are more likely to engage in impulsive behaviours, potentially exacerbating long-term socio-economic inequalities and increasing health risks. Over generations, deficits in willingness to delay gratification can have a compounding effect, as children of parents with a low willingness to delay gratification may inherit or learn similar behaviours (Brenøe and Epper, 2022[76]), reinforcing existing disparities. If certain socio-demographic groups systematically exhibit a lower willingness to delay gratification, inequalities may widen in terms of wealth, employment, health, education and overall quality of life. Given that preferences and behaviours can be passed from one generation to the next, any existing disparities in willingness to delay gratification risk compounding over time.
1.2.3. Social and emotional skills
This report refers to behavioural tendencies as “social and emotional skills” (Kankaraš, 2017[77]; OECD, 2024[47]) rather than as “traits”. It does this to emphasise their malleability and teachability and to align with previous OECD work on children (OECD, 2024[78]) and adults (OECD, 2025[79]). In the context of this report, social and emotional skills refer to the Big Five personality traits of extraversion, emotional stability, agreeableness, conscientiousness and open-mindedness (Soto and John, 2017[80]). Contrary to popular belief, social and emotional skills are not fixed individual attributes but should be viewed as dispositions subject to change over time and amenable to concerted cultivation. What is common across the nine skills analysed in this report is that they are essential for adaptability, interpersonal effectiveness and long-term success in uncertain environments.
There is growing evidence that social and emotional skills have a profound influence on an individual’s academic performance, employability and capacity to adapt to rapidly shifting economic and technological landscapes. For example, conscientiousness has been linked to higher levels of achievement and dependable work habits, while openness to experience fosters innovation and receptivity to learning new skills, which are critical attributes in an age marked by the rise of AI. Meanwhile, emotional stability helps individuals cope with stress and uncertainty. Even outside of labour markets, social and emotional skills play a key role in maintaining healthy relationships and civic engagement, making them an integral component of public and social policy considerations (OECD, 2015[81]).
The social and emotional skills identified in the 2023 Survey of Adult Skills were measured using instruments originally developed in the context of the Big Five personality model, but there are differences between psychological studies that consider personality traits and social studies that adopt social and emotional skills as their preferred terminology, with the latter considering social and emotional skills in terms of functional capabilities and personality traits as behavioural tendencies (Steponavičius, Gress-Wright and Linzarini, 2023[82]). In the past, the promotion of social and emotional skills in educational contexts was driven by the aim of identifying vulnerabilities and deficits (Steponavičius, Gress-Wright and Linzarini, 2023[82]). By contrast, current studies of social and emotional skills stress their importance for obtaining positive outcomes and typically consider how these skills can enable young people to reach their ambitions (Kern et al., 2016[83]; Taylor et al., 2017[84]). There are important conceptual differences between how measurements of information-processing skills and socio-emotional skills are to be interpreted (e.g. whereas a higher literacy score is always better, the same is not true for higher extraversion “score”), which are discussed in the following section. The Reader’s Guide provides an overview of how the Big Five domains are measured.
1.2.4. Measurement and interpretative considerations
A critical conceptual difference exists between information-processing skills and social and emotional skills (a detailed explanation is provided in the Reader’s Guide). Information-processing skills are often regarded as unidimensional because an increase in literacy, numeracy and adaptive problem solving is inherently beneficial. For example, more capacity to process information typically improves labour market outcomes, civic engagement and the ability to reach one’s goals, implying that “more is always better”. By contrast, the optimal level of social and emotional skills varies depending on context and over the life course. For instance, greater perseverance can be beneficial in some settings but may hinder flexibility in others; being highly agreeable can aid collaboration and teamwork but reduce one’s ability to exert authority and achieve decisions in time sensitive, conflictual situations. Thus, the normative judgement of “more is better” does not apply to social and emotional skills. This is especially the case in the 21st century, where contexts are rapidly shifting and evolving because of uncertainty arising from crises and megatrends.
This context dependency underscores why a system aiming solely to maximise social and emotional skills may not be optimal. Instead, adaptability often emerges from a balance of different skills – both cognitive and social and emotional – in a diverse population. If uncertainty is a key feature of the 21st century, then adaptability becomes paramount. Megatrends demand a labour force skilled not just in abstract thinking – typical of information-processing skills – but also in social and emotional skills, enabling them to collaborate, innovate and modulate their responses to rapid changes.
From a measurement perspective, information-processing skills were assessed in the 2023 Survey of Adult Skills using a test specifically designed to distinguish individuals with different levels of proficiency, rather than relying on individuals’ self-perceptions. The tasks within this test were developed to capture what it means to have different levels of literacy, numeracy and adaptive problem solving skills proficiency, thereby offering a more direct and accurate evaluation of these competencies. By contrast, willingness to delay gratification and social and emotional skills were assessed using self-reporting questionnaire instruments and are thus not objective assessments but rather reflect the image individuals have of themselves and the extent to which they believe those particular behavioural tendencies are seen as desirable by the people they value and in the circumstances they are generally exposed to. This difference, alongside the role of context, has implications for the cross-country comparability of different sets of skills. The measurement framework adopted to assess information-processing skills, and their unidimensional nature, means that a cross-country comparable scale could be established. This allows the identification of increasing levels of proficiency that reflect individuals’ abilities to solve tasks of a particular level of difficulty. By contrast, social and emotional skills are context specific, and, as such, no cross-country comparable scale was established. As a result, although it is possible to establish whether an individual is more extroverted or more introverted than others within their country, it is not possible to compare levels of extraversion across countries. Similarly, because behavioural deviations from a country’s standard behaviour are likely to be country specific, the variability in each social and emotional skill is constrained to be the same within each country. This means that it is not possible to derive absolute comparisons between countries in the levels and variability of social and emotional skills.
Table 1.1 details correlations between information-processing skills and social and emotional skills across OECD countries. Literacy, numeracy and adaptive problem solving are highly correlated: the average correlation between literacy and numeracy is 0.87, between literacy and adaptive problem solving is 0.86, and between numeracy and adaptive problem solving it is 0.84. By contrast, social and emotional skills appear to be little correlated with each other except for emotional stability and extraversion (correlation 0.20), conscientiousness and extraversion (0.22), conscientiousness and emotional stability (0.27), and conscientiousness and agreeableness (0.25). These estimates are in line with meta-analytic evidence and reflects the specific measurement tool adopted (self-rating and small number of questions per skill) (Park et al., 2020[85]). While Table 1.1 presents averages of country-specific correlation coefficients across OECD countries, patterns of relations can differ within countries. These results align with the literature on information-processing skills and social and emotional skills, and guided the analytical approach used in this report to derive skills profiles based on the nine skills dimensions.
OECD average, 2023
Note: All coefficients are statistically significant at the 1% level except for those between literacy and agreeableness, and extraversion and agreeableness.
r is greater than 0.8; 
r is between 0.2 and 0.8; 
r is between 0.1 and 0.2; 
r is below 0.1
Source: Calculations based on OECD (2024[86]), Survey of Adult Skills (PIAAC) 2nd cycle database, www.oecd.org/en/data/datasets/piaac-2nd-cycle-database.html.
In today’s rapidly changing economic landscape, skills disparities pose significant challenges to both labour market efficiency and broader economic growth. Market failures often mean that some individuals are not afforded the opportunities necessary to develop the skills increasingly needed for success in the 21st century. Disparities in skills development opportunities, stemming from unequal access to economic and cultural resources inherited from previous generations, have large social and economic costs. The economic costs of skills disparities are manifested through skills shortages, reduced productivity and ultimately lowered growth.
By examining observable gaps in skills across various population groups, this report considers how structural barriers and policy limitations currently hinder many individuals from realising their full potential. The inability to develop essential skills for the 21st century restricts innovation and economic dynamism, and exacerbates labour market imbalances. Addressing these failures can improve outcomes for individuals and serve the broader economic objectives of enhancing productivity and stimulating growth. Societies that constrain opportunities for social advancement for individuals with unfavourable circumstances do not make the most of the potential that exists in the population.
Disparities in the skills individuals possess and in the economic and social returns associated with these skills can arise because of factors that individuals can control – such as the choices they make – and those they cannot control – such as the characteristics of their parents. Identifying the role played by circumstances that lie beyond an individual’s control, also referred to as the “accidents of birth”, is critical if skills policies are to effectively expand opportunities for all while respecting individuals’ personal preferences and agency.
These issues have been explored by economists and political scientists. According to Roemer, success in life is broadly determined by two elements: “circumstances”, over which individuals have no control and for which they should not be held responsible, and “effort”, which represents factors within an individual’s control (Roemer, 2012[87]). In an ideal meritocratic system, rewards should reflect effort rather than the advantages conferred by birth. Rawls (1971[88]) and Dworkin (1977[89]) have further contended that evaluating individual welfare requires looking at both outcomes and the processes leading to them, ensuring that personal responsibility is maintained while mitigating the undue influence of structural disadvantages. Amartya Sen’s capability approach similarly emphasises that genuine equality of opportunity depends on enhancing individuals’ abilities to pursue lives they value, rather than merely equalising access to resources (Sen, 1999[90]).
Empirical evidence supports the notion that people tend to accept disparities stemming from differential effort over those arising from unearned advantages (Benabou and Ok, 2001[91]; Fong, 2001[92]; Starmans, Sheskin and Bloom, 2017[93]). This behavioural preference underscores the importance of using neutral terminology – such as “disparities” – to describe gaps in skills. Unlike “inequality” and “inequity”, which carry moral judgements of fairness, “disparities” simply denote observable differences between groups.
Nonetheless, the theoretical distinction between effort and circumstance is challenging in practice. An individual’s background invariably shapes their capacity to exert effort by imposing physiological, psychological and material constraints. As Sapolsky has questioned, the very concept of free will is interwoven with these constraints, complicating the separation between choice and circumstance (Sapolsky, 2023[94]).
Philosophical debates on equality of opportunity provide a useful framework for understanding these issues. On the one hand, meritocratic ideals suggest that outcomes should reward genuine talent and hard work. On the other, sociological research reveals that structural factors can restrict individuals’ choices and impede skill development. As such, market failures occur when structural barriers prevent individuals from realising their full potential, leading to skills shortages that have broader economic consequences in terms of lower growth and productivity.
This report distinguishes between two types of disparities. Within-group disparities refer to differences in skills among individuals sharing similar socio-demographic characteristics. Such variations can result from differences in talent, motivation or random life events. By contrast, between-group disparities arise from structural barriers and unequal access to opportunities, for example, differences in educational attainment or labour market outcomes between men and women or between individuals with socio-economic backgrounds. These disparities often stem from long-standing institutional practices and require targeted policy interventions.
The meritocratic ideal assumes that individual success is primarily a function of ability, effort and achievement. Yet, in practice, meritocratic systems may inadvertently perpetuate existing disparities if they overlook the structural constraints that limit some individuals’ opportunities. Rapid digitalisation, the transition to a green economy and shifting demographic trends have amplified the importance of continually developing relevant skills. However, unequal access to economic and cultural resources across generations means that not everyone has the same opportunity to adapt to these changes. Lack of opportunities for skills development of certain population groups can contribute to labour shortages in high-growth sectors and undermine overall economic productivity.
From an economic perspective, addressing these disparities is not simply a question of fairness – it is a strategic imperative. By improving access to high-quality education and training, policymakers can enhance the productive capacity of the workforce, stimulate innovation and support sustainable economic growth. Moreover, targeted interventions can help to bridge the gap between structural disadvantages and the demands of modern labour markets, ensuring that the benefits of technological and environmental transformations are more widely shared.
Equality of opportunity is achieved when the distribution of outcomes depends only on effort, rather than on circumstances. Identifying and enabling individuals to exert the effort they are capable of and to fully exploit the talents they have is at the core of the meritocratic ideal. According to proponents of meritocracy, the advantage of economic and social structures organised around this ideal is that it allows individuals to translate raw ability into positive outcomes through effort, without being constrained by society.
Individual responsibility should be maintained and rewarded because it is an important motivational driver at both the individual and social level, leading to innovation and growth. However, disparities that arise because of circumstances beyond individuals’ control should be reduced because they harm social cohesion rather than enhance drive and motivation.
Analyses show that people’s preferences for policy action to reduce socio-economic disparities in life outcomes reflect their beliefs about the role of effort rather than circumstances in explaining such outcomes (Alesina and Giuliano, 2011[95]), with individuals more accepting of inequality resulting from differential effort rather than unequal circumstances (Fong, 2001[92]). Even in contexts with high levels of economic disparities, when individuals believe that their children will be socially mobile and benefit from unequal resource distributions that will reward their effort, they refrain from supporting redistributive policies (Benabou and Ok, 2001[91]). From a behavioural perspective, individuals favour fair distributions – i.e. distributions that reflect the level of effort individuals put in to achieving certain outcomes – over equal distributions – i.e. distributions in which outcomes are the same irrespective of effort. In fact, when fairness and equality clash, people prefer fair inequality to unfair equality (Starmans, Sheskin and Bloom, 2017[93]). These insights underscore the importance of carefully distinguishing between the source of disparities observed in people’s life outcomes and whether they arise from “choice factors” or from “accidents of birth”.
The term “accidents of birth” highlights that differences in skills are observed across population groups – whether by gender, socio-economic background, migration history or other characteristics – and that these differences sometimes translate into varying returns, such as higher or lower pay and better or worse working conditions. The intention is not to dismiss concerns about fairness or the importance of equality of opportunity; rather, it is simply an acknowledgement that skills and their benefits vary for many reasons. Some of these reasons are likely linked to broader social structures or institutional practices, while others reflect the varied ways people exercise their own agency.
The extent to which individuals' skills vary across populations reflects the degree of skills inequality within societies, raising crucial questions about fairness and efficiency. To better understand the underlying dynamics of these disparities, it is essential to quantify not only the overall variation in a population’s core 21st-century skills, but also to identify how much of this variance can be attributed to socio-demographic characteristics beyond individuals' control. In this report, these socio-demographic characteristics constitute gender, parental education and occupation, immigrant background, and childhood residential context.4 These characteristics shape individuals' developmental trajectories, potentially forming opportunities for talent to flourish irrespective of effort.
To illustrate the role that “accidents of birth” play in shaping skills development and the returns to skills, this report examines key dimensions that reflect structural factors influencing individuals’ opportunities to develop their skills. These dimensions include: parental education, parental occupation, gender, childhood residential context, immigrant background and age (which also reflects birth cohort effects). Each of these indicators provides insights into the socio-economic and cultural resources that are inherited at birth, and the extent to which market failures can restrict skill development.
1.3.1. Emergence of disparities in skills development and labour market outcomes
Disparities in skills development and how they translate into labour market outcomes can emerge through multiple pathways: educational access and quality, socio-economic barriers, discrimination and bias, and policy and institutional frameworks. Differential access to high-quality education and training programmes can result in significant skill gaps between different groups. Factors such as funding disparities, geographic location and availability of advanced coursework contribute to these disparities. Socio-economically disadvantaged individuals may face additional challenges in acquiring skills, including limited access to resources and financial constraints, which may lead them to combine the pursuit of education with part-time work or engagement in caregiving duties. Individuals living in poverty also suffer from higher stress levels (Brisson et al., 2020[96]) and poor nutrition (Vilar-Compte et al., 2021[97]), which can impede their cognitive, psychological and physical development (Black et al., 2008[98]; Bryan et al., 2004[99]; Lupien et al., 2009[100]). In severe cases, material deprivation and psychological trauma can limit individuals’ ability to acquire skills. Systemic discrimination based on race, ethnicity, gender or other socio-demographic characteristics can hinder individuals’ ability to attain and use their skills effectively in the labour market, resulting in unequal employment opportunities and outcomes. The design and implementation of skills policies play a critical role in either mitigating or exacerbating inequalities. Policies that fail to consider the diverse needs of different demographic groups may inadvertently reinforce existing disparities. The rapidly changing nature of skills demands in the 21st century can further entrench disparities if policies do not adapt to support continuous skills development among different groups.
Parental educational attainment and occupational status
Parental education and occupation are widely recognised as central determinants of the socio-economic environment into which an individual is born and develops. Higher parental education often correlates with greater access to information, networks and cultural capital, all of which can enhance a child’s learning environment. Similarly, parental occupation reflects not only economic resources but also social status and the opportunities for parents to provide for their children’s cognitive and social development. By identifying if the distribution of skills in the population varies systematically across individuals with parents who have different levels of educational attainment and occupational status, this report traces the transmission of advantages and disadvantages from one generation to the next, detailing some of the structural barriers that constrain skills development.
In this report, adults are categorised into two groups depending on the occupation their parents or guardians held when they were 14. The first group comprises adults with at least one parent/guardian who worked as a manager, professional, technician or associate professional (these are high-status occupations, classified as groups 1, 2 and 3 according to the occupational categories defined in the International Standard Classification of Occupations [ISCO]). The second group comprises adults with parents/guardians who worked as clerical support workers; service and sales workers; skilled agricultural, forestry and fishery workers; craft and related traded; plant and machine operators, and assemblers; and elementary occupations (these are low-status occupations, classified as ISCO occupation groups 4-9).5
Adults are also categorised into two groups depending on the educational attainment of their parents or guardians at the time respondents were age 14, as defined by the International Standard Classification of Education (ISCED) 2011. The first group comprises adults who have at least one parent/guardian with a tertiary-level qualification (ISCED levels 5, 6, 7 and 8). The second group comprises adults whose parents/guardians did not complete a tertiary-level degree. Educational attainment was assessed based on qualifications currently or formerly available in the country concerned, followed by the conversion from the national educational attainment levels to ISCED 2011.
Childhood residential context
The environment in which a person grows up – whether rural, small town or large city – plays a significant role in shaping access to quality education, opportunities to engage in extra-curricular activities, exposure to labour market opportunities and local economic resources. In some countries, non-urban settings may offer less in terms of resources and access to advanced educational programmes than urban centres, curtailing opportunities for skills development for individuals from these areas. In other countries, inner-city neighbourhoods may expose young people to economic deprivation. For some, this could be a motivational driver to change their circumstances, but for others, it may foster a lack of aspirations. By incorporating residential context, the analysis in this report acknowledges how geographic factors, often beyond individual control, contribute to long-term disparities in skills development and in the returns to skills. Residential context in childhood is captured by an indicator showing whether the respondent lived in a city, the suburbs/outskirts of a big city, a town, or a village/farm in the countryside at age 14.
Gender
Men and women continue to experience different educational and labour market trajectories. By analysing disparities in the skills men and women have, as well as differences in the labour market returns to skills, it is possible to consider some of the persistent barriers related to gender stereotypes, societal expectations and potential biases in access to education and training. Through this analysis, this report highlights the economic implications of these disparities.
Immigrant background
In terms of immigrant background, the focus of this report is on the children of immigrants, i.e. individuals who were born in the country in which they reside but whose parents were not, and those who were born in a different country and moved to their current country of residence as children (before the age of 18), but not those who migrated to their country of residence as adults. This captures another critical dimension of the “accidents of birth”, as for these individuals, migration was not a choice they made but a circumstance that shaped their life. Migration often comes with both challenges and opportunities, ranging from linguistic and cultural barriers to enhanced diversity and exposure to varied skill sets. This dimension is essential for understanding how inherited cultural capital and the integration process influence skills development and subsequent labour market outcomes. Because immigrant populations differ widely across countries and economies in terms of their size, composition and migration histories, average results on differences by immigrant background may, more than for other dimensions, differ greatly from results for specific countries or economies.
Age and birth cohort
Age serves as an important proxy not only for an individual’s life-course stage but also for the historical and economic context in which they were raised. Because the data used in the report are mostly cross-sectional, age generally reflects lifecycle effects, cohort-specific influences and period effects (namely the specific time at which the data were collected). Age differences reflect differences related to the ageing process but also differences across generations in education policy, technological advancements and economic conditions, which all contribute significantly to shaping the opportunities available to different generations to develop their skills and use them effectively in the labour market. This dual lens is crucial for identifying trends that may be masked if only one of these aspects is considered. In this report, adults are categorised into three groups depending on their age: 16-29, 30-49 and 50-65. The first category (16‑to-29 year-olds) groups individuals who are in the “learning” phase. According to prior evidence, young adults generally improve their proficiency in information-processing skills up until their late 20s (Borgonovi et al., 2017[101]; Paccagnella, 2016[102]). The second category (30-49 year-olds) groups prime-age workers, and the third category (50-65 year-olds) groups mature adults.
1.3.2. The local impact of crises and megatrends
Countries differ in how crises and megatrends have shaped their populations in the past and in how they are expected to continue shaping them in the future. For example, over the course of the 20th century, some countries have shifted from being primarily places of emigration to predominantly destinations for newcomers. Likewise, distinct patterns of urbanisation, educational participation and economic development have led to marked differences in population structures across countries and over time.
Given that the 2023 Survey of Adult Skills collects representative data on adult populations aged 16 to 65, it offers a unique opportunity to compare birth cohorts of population groups that are often central to skills interventions and whose prevalence has changed over the course of the second part of the 20th century and the first quarter of the 21st century. These groups may not participate equally in skills development, may encounter barriers to labour market entry or broader social engagement, or may receive uneven rewards for their skill sets.
When comparing cohorts of individuals born between 1958 and 2007, all countries and economies participating in the 2023 Survey of Adult Skills experienced a decline in the share of individuals classified as socio-economically disadvantaged, as measured by parental education (Figure 1.3, Panel A) and occupation (Panel B). On average across OECD countries for which data are available, 79% of adults born between 1958 and 1973 (aged between 50 and 65 in 2023) did not have a tertiary-educated parent. This share stood at 49% for adults born between 1994 and 2007 (aged between 16 and 29 in 2023). Some of the most pronounced decreases in the proportion of individuals without tertiary-educated parents are observed in East Asian countries such as Japan, Korea and Singapore, while smaller decreases appear in Germany, Italy and the Slovak Republic. Similarly, on average across OECD countries, the share of individuals with parent(s) working in low-status occupations dropped from 65% for individuals born between 1958 and 1973 to 46% for those born between 1994 and 2007. This share decreased in all 31 countries and economies participating in the 2023 Survey of Adult Skills, with the largest changes in Denmark, Portugal and Singapore, and the smallest in Estonia, New Zealand and the Slovak Republic.
Urbanisation has led to internal migration from rural to urban settings. On average across OECD countries with available data, the share of individuals who spent their formative years on a farm or in a rural environment declined from 39% among adults born between 1958 and 1973 (aged between 50 and 65 in 2023) to 30% among adults born between 1994 and 2007 (aged between 16 and 29 in 2023) (Figure 1.4, Panel A). By contrast, the proportion of individuals who grew up in towns or small cities rose from 28% among the older cohorts to 33% among the younger cohort (Panel B), and the share of those raised in large cities or their suburbs increased from 33% to 37% (Panel C). However, not all countries experienced similar patterns.
These patterns of internal migration have implications for skills development. Rural to urban shifts influence educational and labour market opportunities, with larger metropolitan areas often offering broader access to advanced training and employment options. Ensuring that those in rural areas are not disadvantaged requires addressing gaps in infrastructure, digital connectivity and local training programmes. Similarly, rapid urban growth can strain existing systems in towns and cities, underscoring the need for policies that distribute resources and services proportionately, enabling both older and younger residents to acquire relevant skills for a changing labour market.
Note: Panel A: Differences between the shares of 50-65 year-olds and 16-29 year-olds are shown next to country names. All differences are statistically significant. Parental education (at respondents’ age 14) distinguishes between adults with at least one parent who attained tertiary qualifications (ISCED 2011 5, 6, 7 and 8) and those with no parent educated at the tertiary level. Panel B: Differences between the shares of 50-65 year-olds and 16-29 year-olds are shown next to country names. All differences are statistically significant. Parental occupation (at respondents’ age 14) is based on the International Classification of Occupations (ISCO) and grouped into high-status: managers, professionals, and technicians and associate professionals (ISCO 1-3); and low-status: clerical support workers; service and sales workers; skilled agricultural, forestry and fishery workers; craft and related trades workers; plant and machine operators, and assemblers; and elementary occupations (ISCO 4-9).
Countries and economies are ranked in descending order of the share of 16-29 year-olds.
Source: OECD (2024[48]), PIAAC data and methodology, www.oecd.org/en/about/programmes/piaac/piaac-data.html.
Note: Statistically significant differences between the shares of 50-65 year-olds and 16-29 year-olds are shown next to country and economy names. Childhood residential context (at respondents’ age 14) refers to whether the respondent grew up in a village, town or city.
Countries and economies are ranked in descending order of the share of 16-29 year-olds.
Source: OECD (2024[48]), PIAAC data and methodology, www.oecd.org/en/about/programmes/piaac/piaac-data.html.
The analyses presented in this report indicate that skills disparities remain pervasive across OECD countries, carrying significant consequences for individuals and economies. Core 21st-century skills – literacy, numeracy and adaptive problem solving – as well as social and emotional skills and willingness to delay gratification, remain unevenly distributed across groups, as explored in Chapter 2. There are also unequal opportunities in education and training throughout the life course, which often widen initial skills gaps, as those with advantaged socio-economic backgrounds attain more and higher-quality education and participate more in upskilling over time, as explored in Chapter 3. The analyses also show that skills disparities translate into pronounced differences in labour market outcomes, with important inefficiencies in how talent is developed and used, as explored in Chapter 4.
Together, these findings highlight a dual challenge for policymakers: first, to broaden access to skills development from early childhood into adulthood, and second, to ensure that those skills are effectively matched to productive and rewarding jobs. Meeting this challenge is essential for both equity and economic growth. Inequalities in skills and opportunity not only undermine fairness but also impede economic potential by wasting talent and hampering productivity growth. By inhibiting social mobility and the efficient allocation of talents, inequality of opportunities may trigger slower growth (Soldani et al., 2024[103]). Expanding access to quality education from childhood and upskilling throughout working lives is key to achieving more inclusive growth.
The discussion of policy actions in this chapter is structured around critical policy domains spanning the lifecycle – from early childhood education to formal schooling, lifelong adult learning systems, and career guidance and labour market practices – reflecting the need for a comprehensive, system-wide response. For each domain, targeted policies that can narrow skills gaps, address inefficiencies in skill development and matching, and promote both greater equity and stronger economic performance are discussed. Examples of best practices from across the OECD are explored to illustrate how countries can tackle these challenges in practice. The overarching message is clear: closing skills disparities and improving skills matching is not only a social imperative but also an economic necessity in the 21st century, one that requires co‑ordinated action from education systems, employers and governments to unlock the full productive capacity of communities.
1.4.1. Skills disparities from early childhood through to adulthood: Key findings
Skills are a strong predictor of success in modern knowledge-driven economies. Initial skills disparities related to socio-demographic circumstances beyond an individual’s control – such as gender, parental education and occupation, childhood residential context, and immigrant background – strongly shape 21st-century skills acquisition, as revealed in Chapters 2 and 3. Although such disparities are widespread across countries, the size of the gaps varies greatly, suggesting that policy action and interventions can reduce the depth of the disadvantage faced by specific population groups.
The uneven distribution of skills results in unequal employment opportunities and outcomes in the labour market: workers with higher proficiency in literacy, numeracy and adaptive problem solving are more likely to be employed and, once employed, tend to earn significantly more. For example, other things being equal, a one standard deviation (SD) increase in numeracy proficiency is associated with about 5% higher hourly earnings (see Chapter 4, Table 4.2, Column 5). Similarly, higher educational qualifications are associated with markedly higher earnings: adults with an upper secondary qualification with a vocational orientation earn 3% more than otherwise similar adults who have not completed upper secondary education (see Chapter 4, Table 4.2, Column 5). Groups with lower skills and who did not pursue advanced qualifications face severe disadvantages in the labour market, with most disparities in employment rates and wages across socio‑economic groups explained by differences in skills and engagement in lifelong learning. For example, adults whose parents did not attain tertiary education are less likely to be employed (78% employment rate) than those with tertiary-educated parents (86%), but there is almost no difference after controlling for individuals’ own education and skills (see Chapter 4, Figure 4.1).
Age
Across OECD countries, mature adults (50-65 year-olds) have lower core 21st-century skills than young adults (16-29 year-olds). For example, on average across OECD countries, 50-65 year-olds have a numeracy proficiency that is 0.36 SD lower than the proficiency of 16-29 year-olds. Age-related differences in 21st-century skills are especially pronounced in Chile and Singapore, where mature adults score 0.89 SD lower than their younger counterparts, whereas in New Zealand and Sweden, mature adults score at the same level as younger adults (Table 1.2).These differences reflect expansions in educational opportunities in past decades and the fact that many young adults are actively engaged in skills development because they are still students or participate in non-formal learning.
Large skills gaps between individuals with different socio-economic backgrounds and gender emerge early and often widen as individuals age. For example, children from socio-economically disadvantaged families already lag behind their more advantaged peers in core skills in primary school: by age 10, on average across OECD countries, the difference in numeracy proficiency between children with and without tertiary-educated parents is 0.64 SD, ranging from 0.34 SD in the Denmark to 1.06 SD in Hungary (Table 1.2). As students progress through compulsory schooling, schools in most countries play an equalising role, partly narrowing initial gaps in core 21st-century skills such as literacy and numeracy. For example, for the same cohort of students as above, the average gap by age 15 was 0.42 SD, standing at 0.60 SD in the Czech Republic (hereafter ‘Czechia’) but only 0.26 SD in Denmark (Table 1.2).
Note: Numbers in bold indicate that the respective indicator of a specific country is significantly different from zero at the 5% level. Parental education (at respondents’ age 14) is based on the International Standard Classification of Education (ISCED) 2011 and grouped into tertiary-educated (having at least one parent who had attained tertiary education [ISCED 5, 6, 7 and 8]) and non-tertiary-educated parents.
1. Unadjusted differences in numeracy are the differences between the two averages for each contrast category.
2. Achievement growth in mathematics at age 10 and 15 is based cohorts born 2004/2005.
A country’s performance is statistically significantly above the OECD average at the at the 5% level.
A country’s performance is statistically significantly below the OECD average at the at the 5% level.
Source: Calculations based on OECD (2024[104]), Survey of Adult Skills (PIAAC) 2nd cycle database, www.oecd.org/en/data/datasets/piaac-2nd-cycle-database.html; IEA (2011[105]), TIMSS 2011 Grade 4 Database, https://doi.org/10.58150/IEA_TIMSS_2011_G4; IEA (2015[106]), TIMSS 2015 Grade 4 Database, https://doi.org/10.58150/IEA_TIMSS_2015_G4; OECD (2018[107]), PISA 2018 Database, www.oecd.org/pisa/data/2018database/; and OECD (2022[108]), PISA 2022 Database, www.oecd.org/en/data/datasets/pisa-2022-database.html.
Socio-economic background
Once formal schooling ends, skills disparities between different socio-economic groups widen during youth and adulthood, reflecting differences in formal and informal learning opportunities. Individuals from advantaged backgrounds are more likely to complete tertiary educational qualifications, select into fields of study with high labour market returns and engage in continuous adult learning, whereas disadvantaged individuals more often end their education earlier and have fewer upskilling opportunities. This cumulative advantage means that education pathways, and who gains and maintains skills throughout life, still critically depends on family resources.
Disparities in access to adult-learning opportunities mean that highly skilled workers, who often have an advantaged upbringing, keep improving their skills while those with lower levels of skills proficiency risk falling further behind, a pattern that tends to reinforce existing disparities. Socio-economic differences in adult learning largely stem from structural factors such as educational pathways and job contexts. Workers in occupations that do not typically require advanced qualifications at entry or who work in occupations with a large share of precarious employment contracts have fewer training opportunities offered by employers and face more barriers (e.g. cost, time, lack of information) to pursuing learning themselves. Meanwhile, adults with strong foundations and resources are often more able and motivated to engage in continuous learning, and employers tend to invest more in their training. The content of training also differs, with adults from disadvantaged backgrounds over-represented in narrow, job-specific courses, whereas adults from advantaged backgrounds more often pursue courses that develop skills such as management, foreign languages or advanced information and communication technology (ICT). In short, early disparities tend to compound over time: initial gaps in schooling lead to different employment and training trajectories, which then further widen the skills divide in adulthood.
Disparities in skills development and educational qualifications are a major channel through which socio-economic advantage (or disadvantage) is passed across generations, with spillover effects onto employment prospects. Wage disparities between those from high and low socio-economic backgrounds shrink dramatically when accounting for skills differences: adults with tertiary-educated parents on average earn 9% more than those without tertiary-educated parents – a difference that is fully explained by differences between the two groups in skills and educational qualifications (Table 1.3). It is important to note, however, that the regression models for parental education and parental occupation each control for the other. Because these two dimensions of socio-economic background are closely linked (the average correlation across OECD countries is around 0.5) this approach may understate the full effect of family background. In statistical terms, this reflects multicollinearity between highly correlated predictors, which can lead to attenuation bias and smaller estimated coefficients. In practice, many individuals experience overlapping disadvantages across parental education and occupation, meaning that the cumulative impact of social origin on skills and earnings is often stronger than what each measure in isolation suggests.
Note: Numbers in bold indicate that the respective indicator of a specific country is significantly different from zero at the 5% level. Parental education (at respondents’ age 14) is based on the International Standard Classification of Education (ISCED) 2011 and grouped into tertiary-educated (having at least one parent who had attained tertiary education [ISCED 5, 6, 7 and 8]) and non-tertiary-educated parents.
1. Unadjusted differences are the differences between the two averages for each contrast category.
2. Adjusted differences are based on a regression model that takes into account differences associated with gender, age, parental occupation, childhood residential context, immigrant background and respondents’ educational attainment.
3. Basic adjusted differences in hourly earnings are adjusted for differences in gender, age, parental occupation, childhood residential context, immigrant background. Korea is not included in these estimates.
4. Fully adjusted differences in hourly earnings are adjusted for differences in gender, age, parental occupation, childhood residential context, immigrant background, respondents’ educational attainment, skills, volunteering activities, and participation in non-formal adult education and training. Korea is not included in these estimates.
A country’s performance is statistically significantly above the OECD average at the at the 5% level.
A country’s performance is statistically significantly below the OECD average at the at the 5% level.
Source: Calculations based on OECD (2024[104]), Survey of Adult Skills (PIAAC) 2nd cycle database, www.oecd.org/en/data/datasets/piaac-2nd-cycle-database.html.
Unequal participation in education and training after the end of compulsory schooling is rooted in different skills levels, as well as choices and opportunities, and is a major driver of widening skills disparities between individuals with different socio-economic backgrounds. Individuals from more advantaged socio-economic backgrounds not only tend to study longer (achieving higher formal qualifications), but they also often study differently, choosing fields and formats that yield greater skills payoff. For example, 62% of adults whose parents had high-status professions attain a bachelor’s degree or higher compared to only 14% of adults whose parents had low-status professions (see Chapter 3, Figure 3.7).
There are also striking disparities in adult learning between those from different socio-economic backgrounds. Adults whose parents had high-status professions are 14 percentage points more likely to participate in non-formal learning than adults whose parents had low-status professions, a difference that was larger than 20 percentage points in Italy and as low as 4 percentage points in Korea (Table 1.4).
Gender
Girls outperform boys in reading during secondary school, but this advantage diminishes in adulthood. For boys, their advantage in numeracy grows larger from adolescence into adulthood. These patterns point to “sticky floors” and “glass ceilings” in skills development – i.e. lower-skilled groups (often those from low socio-economic backgrounds or certain minority groups) struggle to rise from the bottom, and some high-skilled groups (e.g. women) face barriers in reaching the very top skills levels. Even though women are more likely than men to pursue advanced educational qualifications, they are less likely to pursue educational trajectories in science, technology, engineering and mathematics (STEM) fields. For example, on average across OECD countries, women are 28 percentage points less likely than men to have completed bachelor’s or more advanced programmes with a STEM orientation and 52 percentage points less likely to have completed vocational programmes with a STEM orientation. The difference between men and women in participation in bachelor programmes is the smallest in Portugal, at 19 percentage points, and the largest in Germany, where it is as high as 40 percentage points. Disparities in participation in VET courses are largest in Canada, the Flemish Region (Belgium) and New Zealand and smallest in Israel (Table 1.5).
Women are less likely than men to be employed and, when employed, earn less than men. These differences remain pronounced even when differences in skills and qualifications are accounted for. After controlling for socio-demographic characteristics, on average across OECD countries, men are 8 percentage points more likely to be employed than women (Chapter 4, Table 4.1). In addition, women earn roughly 16% less per hour than men on average – a gap that remains large even after controlling for educational attainment and skills proficiency (Table 1.5). This implies that factors beyond skills – such as occupational stereotypes, discrimination, differences in work hours or caregiving responsibilities and access to information and to opportunities – play a significant role in explaining gender labour market disparities. Women and men tend to work in different jobs and industries, with women-majority occupations often paying less than men-majority occupations, even when they have similar skills requirements. On average, around 13% of women work in men-majority occupations and 15% of men work in women-majority occupations. These patterns reinforce wage gaps, as fields with more men (such as technology or engineering) usually offer higher wages than fields with more women (such as care and education) at equivalent skills levels (Table 1.5).
Note: Numbers in bold indicate that the respective indicator of a specific country is significantly different from zero at the 5% level. Parental occupation (at respondents’ age 14) is based on the International Classification of Occupations (ISCO) and grouped into high-status: managers, professionals, and technicians and associate professionals (ISCO 1-3); and low-status: clerical support workers; service and sales workers; skilled agricultural, forestry and fishery workers; craft and related trades workers; plant and machine operators, and assemblers; and elementary occupations (ISCO 4-9).
1. Unadjusted differences are the differences between the two averages for each contrast category.
A country’s performance is statistically significantly above the OECD average at the at the 5% level.
A country’s performance is statistically significantly below the OECD average at the at the 5% level.
Source: Calculations based on OECD (2024[104]), Survey of Adult Skills (PIAAC) 2nd cycle database, www.oecd.org/en/data/datasets/piaac-2nd-cycle-database.html.
Note: Numbers in bold indicate that the respective indicator of a specific country is significantly different from zero at the 5% level.
1. Unadjusted differences in numeracy are the differences between the two averages for each contrast category.
2. Basic adjusted differences in hourly earnings are based on a regression model that takes into account differences in age, parental education, parental occupation, childhood residential context, immigrant background, and part-time and public sector employment.
3. Fully adjusted differences in hourly earnings are based on a regression model that takes into account differences in age, parental education, parental occupation, childhood residential context, immigrant background, part-time and public sector employment, respondents’ educational attainment, skills, volunteering activities, and participation in non-formal adult education.
A country’s performance is statistically significantly above the OECD average at the at the 5% level.
A country’s performance is statistically significantly below the OECD average at the at the 5% level.
Source: Calculations based on OECD (2024[104]), Survey of Adult Skills (PIAAC) 2nd cycle database, www.oecd.org/en/data/datasets/piaac-2nd-cycle-database.html.
Childhood residential context
Adults who grew up in cities on average have higher levels of proficiency in core 21st-century skills than those who grew up in towns or villages. For example, across OECD countries, adults who grew up in cities have, on average, 0.18 SD higher numeracy proficiency than those who grew up in villages. Differences related to childhood residential context are especially pronounced in Chile, where adults who grew up in cities have considerably higher proficiency in numeracy than those who grew up in villages (0.72 SD), and in the Netherlands, where adults who grew up in villages have considerably higher proficiency in numeracy compared to those who grew up in cities (0.29 SD) (Table 1.6). These differences reflect, to an extent, differences in the socio-economic background of urban and rural communities: on average, the difference decreases from 0.18 SD to 0.11 SD when controlling for parental education and occupation, and from 0.11 SD to 0.04 SD when further controlling for educational attainment (Table 1.6).
Youth from urban communities are far more likely to complete tertiary education than those from rural areas. On average across OECD countries, 40% of adults who grew up in cities completed a bachelor’s degree or higher qualification compared to only 26% of those who grew up in villages (see Chapter 3, Figure 3.7). The gap in the percentage of adults who obtained a bachelor’s degree or higher qualification between those growing up in cities and villages is largest in Korea (30 percentage points) and smallest in the Flemish Region (Belgium) (1 percentage point) (Table 1.6).
Adults who grew up in cities are also more likely to participate in adult education than those who grew up in villages. The difference is 5 percentage points on average across OECD countries, and fully reflects differences in socio-economic condition between the two groups. The largest advantage for those growing up in cities is observed in Chile (19 percentage points), whereas in the Netherlands, there is no statistically significant difference between the different groups (Table 1.6).
Differences in labour market outcomes are also pronounced. For example, on average across OECD countries, individuals who grew up in cities earn 7% more than adults with similar socio-demographic characteristics who grew up in villages. This difference is most pronounced in Chile, where it is as large as 23%, whereas there are no differences in the United States and Israel. Crucially, the wage penalty associated with growing up in rural areas is mostly the result of differences in educational attainment and skills between adults growing up in different contexts: the difference is reduced to around 3% once controlling for educational attainment and skills (Table 1.6).
Note: Numbers in bold indicate that the respective indicator of a specific country is significantly different from zero at the 5% level. Childhood residential context (at respondents’ age 14) refers to whether the respondent grew up in villages, towns or cities. Estimates for the unadjusted participation rate in adult education and training exclude Japan, Korea and the United States.
1. Unadjusted differences in numeracy are the differences between the two averages for each contrast category.
2. Basic adjusted differences in numeracy are based on a regression model that takes into account differences associated with gender, age, parental education, parental occupation and immigrant background.
3. Fully adjusted differences in numeracy are based on a regression model that takes into account differences associated with gender, age, parental education, parental occupation, immigrant background and respondents’ educational attainment.
4. Unadjusted differences in the participation rate are the differences between the two averages for each contrast category.
5. Adjusted differences in the participation rate are based on a regression model that takes into account gender, age, parental education, parental occupation, immigrant background, respondents’ education and occupation.
6. Basic adjusted differences in hourly earnings are based on a regression model that takes into account differences in gender, age, parental education, parental occupation, immigrant background, and part-time and public sector employment.
7. Fully adjusted differences in hourly earnings are based on a regression model that takes into account differences in gender, age, parental education, parental occupation, immigrant background, part-time and public sector employment, respondents’ educational attainment, skills, volunteering activities, and participation in non-formal adult education.
A country’s performance is statistically significantly above the OECD average at the at the 5% level.
A country’s performance is statistically significantly below the OECD average at the at the 5% level.
Source: Calculations based on OECD (2024[104]), Survey of Adult Skills (PIAAC) 2nd cycle database, www.oecd.org/en/data/datasets/piaac-2nd-cycle-database.html.
Immigrant background
Adults who are the children of immigrants on average have lower levels of proficiency in core 21st-century skills than those who do not have an immigrant background.6 For example, across OECD countries, adults who are the children of immigrants have 0.14 SD lower literacy proficiency than those who do not have an immigrant background. Differences related to having an immigrant background are especially pronounced in the Flemish Region (Belgium) (0.53 SD), whereas in Canada, Chile, Hungary, and Israel adults who are the children of immigrants have higher literacy proficiency than individuals without an immigrant background (Table 1.7). These differences reflect variations across countries in migrant composition as well as integration opportunities and selection mechanisms. On average across OECD countries, skills differences based on immigrant background do not reflect to a large extent differences in socio-demographic characteristics nor educational trajectories.
Across OECD countries, differences in participation in further education based on immigrant background are small: when adults with similar background characteristics are compared, the children of immigrants are 1.82 percentage points more likely than adults without an immigrant background to have completed a bachelor’s degree or higher (Table 1.7).
The gap in the percentage of adults who obtained a bachelor’s degree or higher qualification in favour of the children of immigrants is widest in Canada (18 percentage points), whereas adults without an immigrant background are more likely to have obtained a bachelor’s degree or more advanced qualifications in the Estonia, Flemish Region (Belgium), Italy and Latvia (Table 1.7).
Wages are similar, on average across OECD countries, between adults who are the children of immigrants and adults without an immigrant background. However, in Hungary, the children of immigrants earn 21% less than adults without an immigrant background but with similar socio-demographic characteristics, a difference that remains large, at 19%, when educational attainment and skills differences between the two groups are also taken into account. By contrast, in Estonia, the children of immigrants earn 18% more than adults without an immigrant background but with similar socio-demographic characteristics, a difference that remains large, at 12%, when educational attainment and skills differences between the two groups are also taken into account (Table 1.7).
Note: Numbers in bold indicate that the respective indicator of a specific country is significantly different from zero at the 5% level.
Groups by immigrant background distinguish between children of immigrants and non-immigrants. Children of immigrants were born in the country in which they reside, but their parents were not, or they were born in a different country and moved to their current country of residence before the age of 18. Non-immigrants are born in the country of residence as well as their parents. Japan is not shown due to small sample size.
1. Unadjusted differences in literacy are the differences between the two averages for each contrast category.
2. Basic adjusted differences in literacy are based on a regression model that takes into account differences associated with gender, age, parental education, parental occupation and childhood residential context.
3. Fully adjusted differences in literacy are based on a regression model that takes into account differences associated with gender, age, parental education, parental occupation, childhood residential context and respondents’ educational attainment.
4. Unadjusted differences in the participation rate are the differences between the two averages for each contrast category.
5. Adjusted differences in the participation rate are based on a regression model that takes into account gender, age, parental education, parental occupation, childhood residential context, respondents’ education and occupation, and social and emotional skills.
6. Basic adjusted differences in hourly earnings are based on a regression model that takes into account differences in gender, age, parental education, parental occupation, childhood residential context, and part-time and public sector employment.
7. Fully adjusted differences in hourly earnings are based on a regression model that takes into account differences in gender, age, parental education, parental occupation, childhood residential context, part-time and public sector employment, respondents’ educational attainment, skills, volunteering activities, and participation in non-formal adult education.
A country’s performance is statistically significantly above the OECD average at the at the 5% level.
A country’s performance is statistically significantly below the OECD average at the at the 5% level.
Source: Calculations based on OECD (2024[104]), Survey of Adult Skills (PIAAC) 2nd cycle database, www.oecd.org/en/data/datasets/piaac-2nd-cycle-database.html.
1.4.2. Promoting the agility and co‑ordination of skills policies
To navigate uncertainty and rapid change, OECD countries need to reform not just which skills are taught, but how skills policy itself is made, moving away from rigid, fragmented policy processes towards agile, co‑ordinated and lifelong approaches. By embracing governance models that support quick adaptation, cross-sector collaboration and integration across all stages of learning, countries can be better positioned to respond to emerging skills demands by supporting all population groups.
Agility means that policy frameworks should be framed in ways that can be adapted quickly as new information emerges, rather than remaining fixed until the next reform cycle. For example, the Danish Ministry of Children and Education is considering recommendations put forward by a specially convened expert group on AI in education to shortening the timelines and streamlining the process of reform to national assessments, enabling testing formats to keep pace with technological progress (Danish Government, 2024[109]). This entails moving away from one-off reforms towards an iterative approach, where policies are regularly evaluated and refined.
Some countries have treated the implementation of curriculum changes not as the end point but as a learning phase: feedback is gathered during rollout, and mid-course corrections are made in real time. Such an approach shortens the impact lag of policy reforms by addressing problems early and prevents the entrenchment of ineffective measures. This approach is particularly relevant for policies related to 21st‑century skills, which require an agile governance that embraces pilots, rapid feedback loops and experimental adjustments, allowing skills policies to recalibrate in real time as conditions evolve. This stands in contrast to the prevailing “plan then implement” approach of the 20th century and is better suited for an age when foresight is imperfect and change is constant.
Cross-ministerial co‑ordination is a key pillar of the governance model needed for responsive reform in the 21st century. Skills development and effective skills use cut across the traditionally separate domains of education, labour, local economic development, migration, social protection and industrial policies. However, government structures are often compartmentalised, with different ministries and agencies overseeing early childhood education, schooling, vocational training, higher education and workforce adult learning programmes (OECD, 2020[110]; 2024[111]). These silos can lead to fragmented actions and missed opportunities for synergy. No single ministry can tackle emerging skills challenges alone; for instance, ensuring that workers have the skills for a green economy involves not just education ministries but also those concerned with industry, energy and the environment. Whole-of-government co‑ordination, co‑operation and collaboration is therefore essential, and how to effectively promote this remains a key challenge in strengthening skills systems governance in many OECD countries (OECD, 2020[110]; 2024[111]). An agile reform process rests, for example, on the establishment of joint governance mechanisms – such as inter-ministerial skills councils or strategy units reporting to the centre of government – to align objectives and actions across departments. Such bodies can enable rapid collective responses – for example, aligning a sudden upskilling initiative with both education providers and employment services. Agility in skills policy goes hand in hand with collaboration: flexible governance must be underpinned by strong horizontal co‑ordination so that the left and right hands of government move in the same direction.
Current governance arrangements for skills are not only slow but are also often fragmented. Responsibilities for developing and using skills are widely distributed across levels of government, as well as ministries and public agencies, with little active co‑ordination. People acquire and apply skills throughout their life – in early childhood, schools, technical and tertiary education, and through adult learning and labour market programmes – but each of these stages tends to fall under different institutional mandates. The result is that the governance of skills policy does not follow the learner’s life-course journey but remains split into silos: one agency oversees school curricula, another manages vocational training incentives, another handles unemployment retraining, and so on. Often, even within adult learning, there is a divide; for example, labour ministries focus on training the unemployed while education ministries support other adult learners (OECD, 2024[112]).
This fragmentation leads to gaps, overlaps and inefficiencies. Formal mechanisms to co-ordinate adult learning policy are “rare” and frequently ad hoc, resulting in inconsistent decisions across programmes and a lack of synergy between initiatives. Without co‑ordination, well-intended policies can work at cross purposes or leave critical needs unaddressed. For instance, a labour ministry might subsidise training in isolation from education authorities, leading to duplicated efforts in some areas and blind spots in others. Such misalignment exacerbates skills mismatches in the economy: training supply does not fully correspond to labour market demand when educational planning is disconnected from employment strategy. In practical terms, this means that employers face shortages of workers in certain fields while workers in other fields cannot find jobs commensurate with their qualifications. Fragmented governance thus undermines the overall effectiveness of skills investments, wasting resources and eroding public trust in reform. Addressing this issue is not merely a bureaucratic concern but central to reducing mismatches and improving productivity.
Breaking these silos requires institutionalised co‑ordination across the policy cycle – from joint design of reforms, to synchronised implementation, to shared monitoring and evaluation (OECD, 2024[112]). Some countries have moved to create umbrella structures for skills policy, such as national skills councils or inter-agency task forces that bring together education, labour and economic actors under one strategy (OECD, 2020[110]; 2024[111]). Such bodies can help align funding streams, reconcile differing priorities and ensure information flows between stakeholders. The guiding principle is that all relevant actors should be “on the same page” about skills needs and policy responses. International experience shows that when co‑ordination is strong – whether via formal councils, cross-ministerial agreements or unified strategies – it is possible to reduce duplication and fill the gaps between programmes. Overcoming fragmentation, therefore, is a precondition for an effective and responsive skills policy system. It enables a shift from a patchwork of initiatives to a cohesive ecosystem in which education and training provision, at all levels, actively complements labour market and economic development goals.
Many OECD countries are recognising the need for a coherent national lifelong learning strategy that creates learning pathways that span from early childhood, through formal schooling and tertiary education, into adult upskilling and reskilling. The aim is to enable seamless transitions at every stage so that individuals can continuously develop new competencies and navigate career shifts. Effective roadmaps for lifelong learning transformations involve creating flexible pathways that connect different levels and types of learning to enhance learning opportunities for youth and adults and avoid dead ends. In practice, this might involve strengthening links between secondary vocational programmes and higher education, or between university degrees and on-the-job training, so that learners can move vertically or horizontally without unnecessary barriers. It also means improving the recognition and portability of skills – for example, via national qualifications frameworks that ensure credentials from different institutions or regions are comparable and valued across the labour market (OECD, 2025[113]). Likewise, international and national quality assurance frameworks that can strengthen the quality and recognition of non-formal training are essential for promoting the credibility of alternative learning pathways and skills validation mechanisms (OECD, 2024[114]).
A national lifelong learning strategy also provides a platform to better align the supply of skills with evolving demand. When all relevant ministries and stakeholders co-operate under one vision, it becomes easier to steer the education and training system in line with economic needs. For instance, co‑ordinated foresight exercises can be embedded into a lifelong learning strategy to anticipate emerging skills requirements from megatrends (such as the digital or green transitions) and to feed this intelligence into curriculum updates, career guidance and training investments across the board. Skills foresight data, as well as normative efforts (e.g. competence frameworks and guides, policies, and clear governance) can help in this endeavour. An integrated strategy can thereby act as a bridge between the education sector and employers, ensuring that what learners are taught reflects the competencies actually required in workplaces. International policy research emphasises that fostering more flexible learning pathways goes hand in hand with enhancing the labour market relevance of individuals’ skills. Flexibility – in programme design, delivery and credentialing – is associated with improved agility and customisation of learning to meet both learner needs and industry demand. More concretely, a coherent lifelong learning strategy could set common goals (e.g. increasing STEM skills or digital literacy across age groups), align funding to those priorities, and create accountability for outcomes such as reduced skills mismatches or higher adult training participation. This approach represents a shift from piecemeal policies to a system-wide plan for skills development. Several countries have already moved in this direction.
1.4.3. Early childhood education and care: Levelling the playing field from the start
Investing in high-quality, well-funded early childhood education and care (ECEC) – with qualified, multi-disciplinary staff and access to continuing professional development – is one of the most powerful strategies to reduce skill disparities and promote equity. The early years (from birth to age 5) are when critical cognitive, social and emotional capacities take shape, and when disadvantages rooted in family circumstances can take hold. Research shows that high-quality early education can mitigate the effects of socio-economic disadvantage by boosting children’s development before formal schooling begins. Conversely, gaps that exist at school entry tend to persist or widen over time, making it harder (and more costly) to remedy them later. Thus, OECD countries increasingly recognise ECEC as a cornerstone of an inclusive skills strategy. For example, the UK’s Sure Start Children’s Centres continue to be recognised as an essential, wide-ranging and cost-effective policy tool to reduce the disparities in skills and outcomes that begin in childhood and widen and accumulate over the life course. Sure Start was launched in 1999 as a network of “one-stop shops” that provide a variety of services, including health services, early learning, parenting support and parental employment help, to the communities – often in the most deprived areas of England – in which they were set-up (Carneiro et al., 2025[115]) (see Chapter 3, Box 3.5). Independent evaluations and workforce evidence identify stable funding (pre-2010) and specialist capacity as central to this impact (Carneiro, Cattan and Ridpath, 2024[116]; unison, 2024[117]).
Expanding access to affordable and high-quality ECEC, especially for disadvantaged children, should be a primary policy goal for countries. Currently, participation in early childhood programmes varies widely across countries and socio-economic groups. Children from higher-income families are more likely to be enrolled in preschool or childcare, while those with disadvantaged socio-economic backgrounds – who have the most to gain – often face barriers such as cost, limited provision in their area or non-flexible hours that conflict with parents’ work. To close these gaps, countries should work toward universal access for at least one or two years of preschool before primary school, as well as affordable childcare options for younger children. Given finite budgets, an equitable approach entails targeted subsidies and fee waivers to eliminate financial obstacles for low-income households. For example, several OECD countries cap parental fees as a share of household income or provide free ECEC hours for families below certain income thresholds. Such measures have proven effective in raising participation among disadvantaged children. In Finland, the Right to Learn Program promotes equal learning opportunities for all children by targeting those facing socio-economic, language or learning barriers. Parents may receive allowances to cover ECEC fees, while schools in deprived areas may receive additional resources, such as specialised teachers and support staff, as well as teacher training to help them identify and support students requiring additional help, enhancing learning outcomes in the long run (OECD, 2022[118]) (see Chapter 3, Box 3.5).
Expanding the supply of ECEC services in underserved areas is crucial. Rural or low-income urban communities often have fewer high-quality centres. Governments can invest in infrastructure and provide incentives for providers to operate in these areas so that no community is left behind. Innovative delivery models – for instance, mobile ECEC units, public-private partnerships or flexible provision during non-standard hours – can help reach families with irregular work schedules or those in remote locations. The overarching goal should be that all children, regardless of socio-economic background, can access early education at least from age 3 or 4, if not earlier, setting them on an equal footing as they enter compulsory schooling.
The quality of ECEC is critical to achieving positive outcomes. Disadvantaged children in particular benefit from enriching, nurturing ECEC environments that support all areas of development (language, cognitive skills, social-emotional learning, self-regulation, etc.). Policy should therefore focus on providing high-quality ECEC across all settings and for all children, while also directing additional support to those with greater needs. Key policy levers include establishing strong quality standards and curricula, investing in the ECEC workforce through training and decent working conditions, and monitoring and supporting continuous improvement in providers. Many countries are adopting curriculum frameworks for ECEC that balance play-based, child-centred approaches with intentional skill-building aligned to children’s developmental stages. Such curricula ensure that early learning is holistic – fostering social and emotional development alongside early literacy, numeracy and motor skills – and culturally responsive. To promote equity, curriculum guidelines can include strategies for supporting children with additional needs (e.g. second language learners or those with special educational needs) within mainstream settings. Equally important is the professional development of ECEC staff. Caregivers and teachers need the skills to create responsive, language-rich interactions and to address diverse needs. The OECD’s Starting Strong studies emphasise building a competent ECEC workforce, including through minimum qualification requirements, ongoing training and coaching for educators on inclusivity and bias awareness. Governments should ensure that resources and training for quality improvement are particularly targeted at centres serving high proportions of disadvantaged children so that quality gaps do not compound existing inequities (OECD, 2025[119]).
Making ECEC a vehicle for reducing skills disparities also means designing inclusive services that are linked with broader support systems for children and families. Some countries provide additional bilingual aides or cultural liaisons in ECEC centres to ensure that children of immigrant or minority backgrounds feel represented and supported. Beyond the classroom, ECEC can serve as a hub connecting families to other services. Integrating early education with health, nutritional and parental support programmes yields multiple benefits (see Sure Start Children’s Centre’s in the United Kingdom [England]; Chapter 3, Box 3.5). For example, offering parenting workshops through ECEC centres can empower low-income parents to better support their children’s development. Smoothing transitions to primary school is also important: sharing child development records with primary teachers, aligning curricula between preschool and first grade, and familiarising children with their future schools help sustain the gains from early education and maximise the positive equalising effects of schooling.
In sum, a policy mix that combines universal entitlements with targeted support – universal in the sense that every child can participate, targeted in providing extra resources to those who start from behind – can reduce early skill gaps and reduce disparities later on. Countries such as Finland and Sweden, for instance, have near-universal enrolment in ECEC and invest heavily in quality, which has been associated with smaller socio-economic disparities in academic readiness. The economic case is strong: evaluation studies suggest that investments in the early years have high returns, particularly for disadvantaged children, as money spent yields high long-term societal benefits through better education and employment and reduced social costs (Barnett, 2011[120]; Heckman, 2006[121]; Heckman et al., 2010[122]).
1.4.4. Formal education: Strengthening opportunities for all
As children move into formal education (primary, secondary and tertiary levels), the challenge for policy is to sustain early interventions or reduce disparities that emerge early on, ensuring that the education system provides equal opportunities for all learners to develop 21st-century skills. Formal schooling remains the single biggest influence on skills acquisition in youth and an engine of social mobility, or conversely, of social reproduction. The evidence from Chapter 3 indicates that compulsory schooling often narrows socio-economic skills gaps relative to early childhood, but significant disparities remain, and what happens after secondary school (in higher education or entry to work) can widen disparities yet again. By contrast, gender disparities widen as children progress in their education, with evidence from France suggesting that a gender gap in mathematics favouring boys is established as soon as children enter school (Martinot et al., 2025[123]). The gender gap extends beyond mathematics to science subjects more generally, as TIMSS 2023 data show an 8-point performance advantage for boys over girls in science in CM1, a gap that was previously non-significant (La Direction de l’évaluation, de la prospective et de la performance, 2024[124]).Therefore, policymakers must pursue a two-fold strategy in formal education: 1) improve the quality of education for all; and 2) target interventions to support disadvantaged students and reduce the effect of societal norms and stereotypes that shape the relative achievement possibilities of boys and girls. In this way, all students will be able to achieve their full potential.
To achieve this, all schools must have the resources and capacity to deliver high-quality education, regardless of the socio-economic or geographic context. Within-country differences in school quality (e.g. availability of trained teachers, class sizes, curricular breadth) often map to socio-economic divides, as wealthier communities can attract more resources. To counter this, many OECD countries use funding formulas to allocate additional resources to schools serving disadvantaged or high-needs student populations. Ensuring adequate and progressive funding – more per-pupil spending on those who need it most – has been linked to smaller learning gaps. This can be achieved either by providing more teachers to classes catering for disadvantaged populations or more qualified and trained teachers who can better manage the additional challenges arising from working with children with disadvantaged socio-economic backgrounds.
Countries with higher cumulative public investment up to secondary level have narrower adult skills disparities by socio-economic background, but the association is not strong, as highlighted in Chapter 3. This means that it is not merely about spending more but about spending effectively: directing funds to evidence-based programmes (such as early literacy interventions, tutoring for struggling students or teacher incentives for remote areas) yields the greatest impact. To break the link between socio-economic background and educational outcomes, inclusive practices and targeted support must reach into the daily practices of schools and classrooms. One key policy is early identification and support for struggling students. Diagnostic assessments and monitoring can help flag learning gaps early (for instance, difficulties in reading or mathematics in primary years), triggering additional help such as reading specialists, small-group tutoring or individualised learning plans. Targeted support can also address the non-academic barriers that often impede learning for disadvantaged youth; for example, countries may provide school meals, mental health counsellors or after-school programmes in low-income communities to mitigate the effects of poverty on learning readiness.
The rigid tracking of students into “vocational” vs. “academic” streams at an early age tends to reinforce social disparities, because disadvantaged students are disproportionately placed into less challenging/rewarding tracks, limiting their skills development. Policies that delay tracking, promote comprehensive schooling or ensure permeability between tracks (so that late bloomers can advance) are associated with more equitable skills outcomes. Moving forward, advances in the personalisation that AI affords could help education systems combine the benefits of tracking and ability grouping, which include the possibility to provide learning in ways that best align with students’ preferences, talents and interests, with the benefits of comprehensive schooling, which allows for diverse developmental trajectories and being stimulated by diversity.
Because of its personalisation potential, AI adoption may allow education systems to reap the benefits of differentiation as well as integration, allowing greater flexibility in educational offering without the stigma associated with rigid differentiation. Adaptive algorithms can tailor content, pace and feedback to each learner’s needs, supporting mastery for high-achievers and targeted remediation for struggling students. For example, the Canadian government has invested in the development and integration of AI into an existing digital language learning platform to help students learning French as a second language. The platform gathers information about each student’s unique learning profile and provides them with personalised educational content (Government of Canada, 2025[125]). These types of tools may be particularly well-suited to address the varied challenges faced by students with an immigrant background or those whose native language is different than that of the language of instruction. This is important, as across the OECD, students from these groups perform less well and report a lower sense of belonging in schools than those without an immigrant background (see Chapter 3, Box 3.1 and Box 3.6). Furthermore, schools can adopt heterogeneous grouping and individualised expectations that are tailored to all students’ individual abilities. In this way, AI may enable differentiation within the classroom (such as more challenging material for advanced students and remedial support for those behind) without sacrificing students’ outcomes, rather than separating students into different schools or programmes too early.
To supplement this, providing strong guidance and support for transitions is crucial – for example, helping students from under-represented groups prepare for and apply to tertiary education through mentorships or college/career counselling can increase their access to higher levels of study. In Spain, the Programme for Educational Guidance, Advancement, and Enrichment (Programa para la orientación, avance y enriquecimiento educativo; PROA+) aims to reduce school dropout and low performance, especially in disadvantaged areas, with a focus on digitalisation and modernisation of educational infrastructure from ECEC to higher education. The programme channels resources into schools with high levels of need, often in rural or socio-economically disadvantaged areas. Schools commit to responsive service provision through agreements with educational authorities that outline the shared objectives and define the specific actions and resources needed to achieve these objectives (See Chapter 3, Box 3.10).
Tertiary education – including university and high-level vocational programmes – has become increasingly important for 21st-century skills and labour market prospects. However, access to tertiary education remains highly unequal, with students from wealthier, urban or university-educated families far more likely to enrol and complete degrees. To narrow skill gaps, OECD countries should pursue policies that broaden participation in tertiary education and improve completion rates for under-represented groups. Measures can include financial support (needs-based scholarships, reduced tuition or free college for low-income students), outreach and bridging programmes (special admission pathways or preparatory courses for first-generation college students). Strong academic and social support should also be provided on campus (tutoring, peer mentoring, etc.) to ensure that students not only enter but also succeed in higher education. For example, the Irish Higher Education Access Route helps facilitate entry into higher education for students with disadvantaged socio-economic backgrounds, with applicants who meet certain financial, social and cultural criteria able to access reserved places with amended entry requirements at participating higher education institutions (see Chapter 3, Box 3.5). Additionally, some countries fund “second-chance” routes for adult entry into tertiary programmes, which allows individuals (often with disadvantaged socio-economic backgrounds) who did not continue education after secondary school to access higher qualifications later in life. Expanding such pathways – including through part-time and online tertiary study options – can reduce socio-economic disparities in advanced skills.
All of these policies are compatible with maintaining the high quality and labour market relevance of the knowledge and skills taught in higher education institutions. Policies tying funding to student support services and graduate outcomes can involve employers in their design and scope (especially in vocational tertiary tracks) and can help align tertiary expansion with skills needs and student success. Quality assurance mechanisms – such as school inspections or performance monitoring – should specifically attend to equity indicators, ensuring that disadvantaged students are not left behind. Some countries have successfully implemented targets to reduce socio-economic disadvantage, with goals including reducing the proportion of low-performing students or the gap between different student groups in core subjects, and are holding the system accountable to these goals. Quality assurance mechanisms can exist at the system level or beyond – for example, the European Union’s Education and Training Monitor provides regular updates on member states’ progress towards EU-level targets for education policy for 2025 and 2030. The Monitor provides updates for all 27 EU member countries, linking evidence and experiences across contexts, and proposing linkages and policy solutions (European Commission, 2024[126]).
Beyond access to tertiary education, there is a notable form of disparity regarding differences in field-of-study choices between men and women, which have long-term implications for skills and careers, as explored in Chapter 3. For instance, despite the fact that women now often have higher overall tertiary attainment than men, they remain under-represented in STEM fields and mathematics-intensive courses. Furthermore, students with disadvantaged socio-economic backgrounds may be less likely to pursue advanced science or technology studies due to lack of initial preparation or encouragement, or because, despite their talent, in purely rank systems they may have lower achievement than their advantaged peers. While procedurally meritocratic, these systems are substantively unequal because achievement disparities determining access to advanced qualifications arise because of factors that lie beyond individuals’ control.
These enrolment patterns contribute to skill gaps in areas such as numeracy and digital skills, and later to pay gaps, as STEM careers are often higher paying. Policymakers should thus implement initiatives to broaden participation in high-value fields. Examples include providing extra tutoring and orientation in mathematics/science for under-represented students; promoting role models and mentors (such as women engineers visiting schools to inspire girls); offering scholarships for low-income students in STEM; and combating stereotypes through curriculum and career counselling. Some countries have introduced targeted programmes such as coding camps for girls or partnerships between elite universities and disadvantaged secondary schools to demystify fields such as engineering and computer science. For example, the French “Girls and Maths” (Filles et Maths) plan has recently introduced measures to encourage girls to choose more mathematics-intensive study pathways throughout their formal schooling, and ultimately enrol in mathematics-related higher education pathways. In the Slovak Republic, the “You too in IT”, (Aj Ty v IT) initiative supports girls and women to pursue careers in ICT by offering coding workshops for secondary school girls, upskilling for career changers and teacher training (see Chapter 3, Box 3.3).
Setting out clear and achievable targets or quotas for the enrolment of individuals from under-represented groups can incentivise training providers and other social partners to mobilise towards a common goal, especially if the process of target-setting was transparent and participatory. The Austrian National Strategy for the Social Dimension in Higher Education, which aims for a minimum of 10% of women and men in any field of education at any higher education institution (excluding doctoral studies) by 2025 and has longer‑term goals to increase this to 30%, exemplifies such an approach (Federal Ministry of Science, Research and Economy, 2017[127]).
Within vocational education, encouraging a diversity of enrolments can both reduce labour market gender segregation and ensure that talent is developed in all sectors. Targeted campaigns can help students identify VET and apprenticeships of interest to them. For example, the Irish National Apprenticeship Office oversees the “Facts, Faces, Futures” campaign, which promotes the participation of women in apprenticeships by targeting all-girls schools with specific content, which is delivered by female role models (see Chapter 3, Box 3.3).
In addition to addressing pressures and hurdles external to the school environment (such as broader societal stereotypes and segregation in the labour market, as discussed above), policymakers should consider pressures internal to the school environment, such as the development of social and emotional skills for lifelong learning and success in the labour market. Such initiatives can be implemented throughout the formal schooling lifecycle; for example, the Spanish “In their Shoes” (En Sus Zapatos) programme models the positive impacts of a school environment that fosters skills related to conflict resolution and emotional management for children aged between 4 and 17, while France’s “Roped party of Success” (les Cordées de la réussite) programme boosts academic ambition and combats self-censorship among disadvantaged secondary school students towards higher education (see Chapter 3, Box 3.6). In Chile, the Life Skills Program (Programa Habilidades para la Vida, HPV) supports students from pre-kindergarten to 12th grade by fostering emotional, social and cognitive development throughout all stages of schooling. The programme tackles emotional regulation and adaptation, school co-existence and psychosocial well-being, and academic-related challenges (see Chapter 3, Box 3.10).
In short, achieving equity in formal education is not just about getting everyone through the door, but also guiding all students towards the full range of learning opportunities that match their interests and the economy’s needs. It should be noted that education policy alone cannot equalise outcomes without support from other social policies. Issues such as child poverty, housing and health affect students’ ability to learn. A holistic approach that co‑ordinates education with social protection (income support for poor families), healthcare (vision, dental, mental health services in schools) and community development amplifies the impact of school-based interventions. Some OECD countries have moved towards “community schools” or integrated service models that provide a wraparound support system for disadvantaged students (for example, co-locating health clinics and social workers on school campuses). These approaches recognise that to truly give every child a fair shot at developing their skills, barriers outside the classroom must also be addressed.
1.4.5. Lifelong learning and adult training systems: Promoting continuous skills development for all
Closing skills gaps and adapting to fast-changing skills demands require that learning does not stop at graduation. A strong adult learning system that encompasses on-the-job training, continuing education, reskilling and upskilling programmes, and second-chance education is critical to both economic dynamism and social inclusion. Yet, as highlighted in Chapter 3, participation in adult learning remains highly unequal and, in some countries, worryingly low or even declining. Those who could benefit most from retraining (older workers in declining industries, adults without basic qualifications, migrants needing language skills, etc.) often face the greatest obstacles to accessing adult learning. Meanwhile, employers may lack the resources to provide training that could be especially valuable for low-skilled or temporary workers. As a result, adults’ skills stagnate or become obsolete, even as job vacancies for skilled workers go unfilled and new technologies demand continuous learning. To address this, OECD countries need to fundamentally rethink and strengthen their adult learning systems. Incremental tweaks are unlikely to suffice; instead, comprehensive policy packages are required to incentivise lifelong learning, lower barriers to access and ensure training provision is aligned with future skills needs.
The first priority is to make adult learning more accessible and attractive, particularly to under-represented groups. Cost and time are two major barriers often cited. Many adults, especially in low-paid jobs, cannot afford tuition fees or income loss during training. Governments can tackle cost barriers through public financial incentives: for example, grants or learning vouchers for low-income adults, tax credits or subsidies for employers who train low-skilled workers, and individual learning accounts (ILAs) that give every adult a personal budget for training. A number of countries are experimenting with ILAs – France’s Compte Personnel de Formation provides a notable example of a nationwide scheme that credits training hours/money to individuals, with higher credits for the low-qualified. Singapore’s SkillsFuture Credit is another, providing direct training credits (e.g. SGD 500 at age 25 plus top-ups at later ages) to encourage continuous learning (OECD, 2025[113]). To be effective, these instruments must be designed with inclusion in mind, with more support targeted at those with lower qualifications or in non-standard work (who tend to receive less employer training).
The other major barrier is time, with busy working adults with care responsibilities struggling to find time for training. To counter this barrier, policies such as training leave and more flexible learning options are crucial. Many OECD countries have introduced or expanded rights to training leave – allowing employees to take paid or unpaid time off for education. To make these policies inclusive, it is important to ensure that training leave is usable by those who need it most. This may entail providing wage replacement or stipends during leave so that low-income workers can afford to take time off, raising awareness about the entitlement, and incentivising employers to approve leave requests. Training opportunities can also be provided through intra-organisational mobility. For example, the Belgian Federal Public Administration developed a “Talent Exchange” programme through which 21 public organisations and agencies can request for temporary staff from the other organisations within the network for help on a specific project. The programme has been used as a tool to address skills shortages while giving civil servants the opportunity to develop their skills and learn on-the-job (OECD, 2025[113]). Where training leave is statutory, governments should monitor uptake among different groups and adjust parameters if, for example, only high-skilled workers are found to be benefitting.
Flexible training formats are another way of helping to reconcile learning with work and life schedules. This includes offering modular courses, evening/weekend classes, and online or blended learning. The rise of micro-credentials – short courses certifying specific skills – is a promising development to allow adults to upskill in small, manageable steps. Policymakers should support the expansion of accredited micro-credential programmes and ensure that they can be “stacked” towards larger qualifications, enabling continuous progression. The recognition of prior learning (RPL) is another powerful tool, as by assessing and crediting skills that individuals have acquired informally (through work or other experiences) it can reduce the time and cost needed for adults to obtain new qualifications. Countries such as Portugal (via its Qualifica Centres) and Finland have established robust RPL systems to help adults, especially those with low formal education, get credentials for skills they already have. Scaling up RPL and making the process simpler could greatly improve uptake, as many adults are unaware of or may be intimidated by current RPL procedures. The promotion of RPL systems and their benefits can incentivise the uptake of non-formal skills development opportunities, as highlighting the potential labour market and other benefits of learning can encourage individuals to reconcile learning with competing priorities that share common goals. For example, volunteering opportunities can provide individuals with an immigrant background the language learning and networking opportunities needed to integrate into local labour markets, and incentives such as tax-breaks or priority access to jobs can encourage learning through volunteering or other community-based initiatives (see Chapter 3, Box 3.11).
Overall, adult learning must be made easy to access, with minimal bureaucratic hurdles, and be actively promoted and incentivised. Outreach and guidance are needed to engage adults who do not naturally seek training. This could involve local one-stop centres for adult education advice, online portals linking individuals to courses and subsidies, and community-based initiatives to motivate learning (such as volunteering opportunities, peer learning groups, or leveraging public libraries and community colleges as access points for lifelong learning). Some local governments are reinvesting in physical spaces that provide wraparound services for young and adult learners while also regenerating urban spaces. For example, the UTOPÍAS initiative in Iztapalapa, Mexico City, aims to transform neglected urban spaces into free-access centres offering educational, cultural, health and recreational services (see Chapter 3, Box 3.10).
A robust adult learning system depends on a partnership between the public sector, individuals and employers. Currently, responsibilities are often blurred or fragmented: some firms invest heavily in training their staff (although training quality and relevance varies), while others, especially small and medium-sized enterprises (SMEs), have limited resources to dedicate to employee upskilling or reskilling. To address this, policies can foster a culture of shared investment in skills. For instance, some countries use training levies or collective schemes whereby employers contribute to a national training fund, often with the option for firms to recoup funds when they train employees. Such schemes, when well-designed, encourage all firms to play a role and generate pooled resources to finance training for SMEs or for the workforce at large. For example, Norway’s “Skills Plus” Work agenda includes resources for employers to provide workplace-based literacy, numeracy, digital and communication skills training on-the-job for individuals with low qualifications (see Chapter 3, Box 3.5). In Poland, there are efforts encouraging employers to aid unemployed individuals to re-enter the labour market, with co-financing for salaries and certain social security contributions exemptions available for employers onboarding senior workers who may need significant upskilling or other support transitioning back into the workforce (Polish Government, 2024[128]).
Another approach to boost adult learning within firms is to highlight the business case for training: governments and industry bodies can disseminate evidence that training improves productivity and innovation, thus incentivising firms to invest. Small firms may need support to navigate training options, for example through sector-based training consortia or public programmes that provide training needs assessments for companies. Ireland and Estonia have achieved notable progress in adult learning participation in the past decade by implementing long-term reforms that combine financial incentives, training leave, employer involvement and quality assurance of training. These comprehensive approaches signal that piecemeal efforts will not solve the issue, and that the incentives of all actors must be aligned.
As countries ramp up adult learning, they must also ensure that the training provided is of high quality and relevant to evolving skill demands (OECD, 2024[114]). One risk of expanding decentralised adult training through financial incentives is the proliferation of low-quality courses or credentials that do not actually improve workers’ prospects. Strong quality assurance mechanisms are needed that include accrediting providers, setting standards for curricula and instructors, and monitoring outcomes (such as employment or wage gains for training participants). New forms of learning, such as online courses and micro-credentials, pose particular quality challenges, and several countries (e.g. New Zealand) have moved to establish specific quality criteria for micro-credentials and frameworks to integrate them into the national qualification system. Sharing information on provider performance and course outcomes can empower learners to choose effective programmes. This calls for improved labour market intelligence and mechanisms to feed this intelligence into training provision. For example, skills forecasting and sector skills councils can identify emerging skills needs (related to digitalisation, green transition, etc.). Public training funds can prioritise courses in those areas, while forecasting data and competence frameworks can help design criteria to both guide this prioritisation and ensure course quality. Some countries have created “occupation skill frameworks” in partnership with industry. For example, Singapore’s Industry Transformation Maps (Workforce Singapore, 2025[129]) involve industry leaders in defining evolving occupational standards, which are used to guide modular training development. Moreover, jobs and skills frameworks that identify the key competencies required in each job can help design training programmes and career mobility. In France, the Public Service Jobs Directory (Répertoire des métiers de la fonction publique – RMFP) offers such a framework by systematically mapping the skills associated with each public job. The RMFP is regularly updated to reflect changes in jobs and skills, and upcoming developments aim to strengthen its alignment with labour market evolutions. This tool not only helps French public employers anticipate changes and guide training provision, but also helps French citizens understand the evolving skills needs of French public institutions. The public employment service can also play a role by steering jobseekers towards training for occupations in shortage. In essence, an adaptive and forward-looking adult learning system is needed that not only responds to current skills gaps but that also anticipates future ones, thereby enabling workers to prepare for the jobs of tomorrow rather than jobs that may be in decline. Lessons from higher education institutions that have integrated some micro-credentials and massive open online courses (MOOCs) into their broader curricula can be valuable for employers looking to outsource or procure training content or tools for their employees.
Given the evidence that market forces alone leave many disadvantaged adults behind, reducing skills disparities must be a design principle of adult learning policies. This means continuously asking questions such as: are our programmes reaching those with the lowest skills? Are older workers and migrants seeing improved access? Considering the needs of vulnerable populations and offering learning in non-formal settings, such as in the community, can make upskilling more approachable for adults with negative schooling experiences or low initial skills levels.
In summary, a reinvigorated adult learning system in OECD countries should be inclusive, proactive and well-coordinated. It should treat learning as a lifelong right and necessity, just as health systems treat healthcare as an essential service with universal access, but with extra care given to those most at risk of “skills poverty”. This is not just a social policy agenda – it is an economic one, as continuous skills development is essential for productivity, innovation and resilience in the face of technological and demographic changes. Investments and reforms in adult learning will mitigate skills shortages and enable broader participation in an economy’s growth by creating a more adaptable and skilled workforce.
1.4.6. Career guidance and transitions: Supporting informed and inclusive pathways
Individuals face transitions throughout their lives: from school to further education or work, from one job to another, or back into the workforce after a break. How well people navigate these transitions can significantly affect whether skills are effectively utilised and whether labour market disparities widen or narrow. Individuals’ ability to steer through professional challenges effectively at least partially relies on the education and training support available to them, as well as complementary services that support decision making. Career guidance and career development support are key policy domains to facilitate smoother transitions and better skills matching. Particularly for young people, access to quality career guidance during schooling can help ensure they make informed decisions aligned with their talents and labour market opportunities, rather than being derailed by lack of information or social biases.
Misalignments often begin early, as highlighted in Chapter 4. For example, youth with disadvantaged socio-economic backgrounds may have limited awareness of certain careers or underestimate the level of education and training needed, leading to sub-optimal choices that perpetuate inequality. Students with a disadvantaged socio-economic background generally have lower levels of career development activities and networks than students with advantaged socio-economic backgrounds (OECD, 2024[130]), relying more on schools for career information. Girls and boys also still exhibit gendered career aspirations in line with stereotypes, contributing to later occupational segregation and wage gap (Korlat et al., 2022[131]). Strengthening career guidance systems with an equity focus is therefore a high-impact policy lever.
Ensuring universal access to career guidance from an early age is critical. Policymakers should make career-related learning and guidance a standard part of the education journey for all students, beginning well before the end of secondary school. Evidence suggests that children start forming ideas about jobs early in life and that by early adolescence, many have already narrowed their aspirations, often based on gender norms or the limits of the information and role models they have been exposed to. Introducing age-appropriate career education in primary school – for instance, activities that broaden children’s awareness of different occupations and counter stereotypes (e.g. female scientists, male nurses as role models) – can plant the seed for broader and higher aspirations. By lower secondary (around age 13-15), all students should receive guidance on the diverse educational pathways and careers available, including vocational routes, apprenticeships, academic routes, etc., so that they can make informed choices for upper secondary. Schools with a large concentration of disadvantaged or vulnerable students may require additional resources to provide high-quality career guidance in a systemic way. Governments must be cognisant of this and provide the additional support needed. For example, in Ireland, the Delivering Equality of Opportunity in Schools programme provides preferential and extra funding for eligible schools to provide career guidance activities. In the United Kingdom, programmes such as Inspiring the Future and Speakers for Schools connect schools with volunteers and public figures who work exclusively with public institutions to try to democratise access to hard-to-reach networks, information and ultimately career pathways (see Chapter 3, Box 3.9). Countries can also set policies that guarantee every student has access to a certain minimal provision – for example, a mandated number of career counselling sessions or compulsory participation in work exploration activities (such as workplace visits or job shadowing). Some education systems allocate extra counsellor positions or partner with community organisations to reach at-risk youth. The guiding principle is early and proactive engagement: waiting until the end of secondary school to offer career advice is too late for many, as they might have already made curriculum choices that limit their options. Starting guidance early, and making it an ongoing process throughout schooling, leads to better alignment between students’ capabilities and interests, and the needs of the labour market.
Enhancing the quality and inclusiveness of career guidance is as important as broadening access. Many teachers or school counsellors feel under-prepared to guide students in a rapidly changing world of work and require additional training on labour market trends, counselling techniques and bias-free guidance. In some countries, guidance is delivered by a mix of professionals, including dedicated career counsellors and teachers or social workers who integrate career development into their roles. Building this professional capacity with the proper resources is vital to ensure that students get expert support.
Beyond informing students about options, career guidance is an important way of imparting career management skills and employability skills to students. Guidance programmes should actively teach young people how to make decisions, set goals, search for opportunities and present themselves (CV writing, interview training, etc.). These competencies help level the playing field, as more advantaged students often pick up such skills from family and social capital, whereas disadvantaged students rely on school to learn how to navigate careers. An inclusive guidance approach also means tailoring support to those who need extra help: for example, special outreach to students who are disengaged or at risk of dropping out, providing alternate pathways such as vocational training or re-engagement programmes so that no young person is left without a pathway. Some best practices include “career coaching” schemes targeting disadvantaged youth or mentorship programmes that pair professionals from industry with students lacking role models.
In conjunction with guidance, all students should have the opportunities to build social capital and gain exposure to the world of work. One consistent finding is that students with lower socio-economic backgrounds have less access to informal networks and work experience that can inform their career choices. For example, they may not have family friends who are engineers or lawyers to ask for advice, and their part-time job opportunities or internships might be limited. Schools can act as connectors to the world of work to compensate for this gap. Policies encouraging partnerships between schools and employers are highly beneficial. This might involve organising career talks where diverse professionals (including women in STEM, leaders from minority communities, etc.) come to speak about their jobs; facilitating short work placements or job shadowing for students; and running programmes such as “career fairs” or “industry days” in schools. These activities help demystify professional environments and expand students’ horizons. They also allow students to develop social capital – contacts and insights that are often concentrated among those with more advantaged socio-economic backgrounds. Inviting workplace volunteers or alumni from the school (especially those with similar backgrounds) to mentor students can be powerful. Some countries have national initiatives linking schools with local businesses and community mentors to ensure that every student gets at least one meaningful work-based learning experience before graduation. For example, the State of Maryland in the United States fosters partnerships between public higher education institutions, local employers and non-profits that promote local workforce development and the uptake of relevant learning and training opportunities throughout young adults’ educational and early professional trajectories. Non-profits provide students with career guidance while local employers provide information, mentorship and employment opportunities (Robbins, 2024[132]). Such measures can democratise access to career-related networks.
Involving students’ families is also key: schools can engage parents, including those with disadvantaged socio-economic backgrounds, in the career guidance process by providing them with information and encouraging career conversations at home. Family-oriented career workshops or one-on-one meetings between counsellors and parents can help align support around the student while providing crucial information to parents, who may need to engage in upskilling or reskilling themselves.
Gender imbalances in career choices contribute significantly to persistent labour disparities. Career guidance can play a critical role in addressing gender stereotypes and broadening the aspirations of individuals belonging to groups under-represented in certain industries. Career guidance policies should explicitly aim to challenge gender stereotypes and encourage the exploration of non-traditional careers for all genders. This can be done by ensuring that career materials and activities are free of bias (e.g. showing men and women in diverse roles), providing specific programmes to boost the confidence of under-represented groups (for example, coding clubs for girls, or nursing career camps for boys), and offering experiential learning that breaks down gender norms. For instance, some education systems have introduced “Girls’ Days”, where female students visit companies in engineering/manufacturing, as well as campaigns to attract more boys into teaching or healthcare. In Germany, such events are part of a larger, annual campaign, supported by several federal ministries, that mobilises resources and political capital to support women entering men-majority fields. Adjacent initiatives and support include investments in childcare at universities and co‑ordination with employers to ensure equity in recruitment practices (see Chapter 3, Box 3.8).
Career counselling should pay attention to different needs by gender, with evaluations suggesting that work experience and positive role models can significantly impact young people’s decision making. Adolescent boys often exhibit more career uncertainty and could benefit from earlier intervention to engage them in thinking about their futures. Meanwhile, girls, despite high academic performance, may self-select out of certain careers due to confidence gaps or societal messages (OECD, 2024[130]). Just as it is important to promote the uptake of STEM subjects for girls in schools to ensure that they are able to access high-paying and in-demand skills and occupations, such as those related to STEM, there is a range of actions that governments can undertake to bring more men into historically women-majority fields, such as health and social care. Australia, Germany and the United Kingdom have all invested in targeted outreach campaigns, research, setting up new training institutions, and recruitment drives and educational materials to promote men’s enrolment in training related to nursing (see Chapter 3, Box 3.8). Similarly, the Norwegian government has used targets and positive discrimination policies to bring more men into teaching and childcare professions (see Chapter 3, Box 3.8), while some Spanish public higher education institutions have established promotion plans to ensure women can access the same STEM teaching roles as men (see Chapter 3, Box 3.8). Thus, guidance should encourage self-exploration and critical thinking about career-social relationships – essentially helping youth critically examine why they might be inclined towards certain jobs and to question limiting narratives. At the same time, structural issues, such as societal expectations and the working conditions and renumeration pertaining to these occupations, should be addressed to address the problems pertinent to all teachers and health and social care workers across many OECD countries.
While schools are a major focus, career guidance is equally important for adults at transition points – such as those unemployed, considering a career change or re-entering the workforce (e.g. parents returning after childcare break or older workers). Public employment services (PES) and adult career counselling centres should be strengthened, with the centralisation of personalised counselling, skills assessment and the navigation of training options able to greatly help adults find pathways to better employment. For example, the Australian Victoria’s Jobs Victoria Career Counsellors Service has been offering free, personalised career guidance to all adults, regardless of employment status. The services are delivered by the Career Education Association of Victoria and staffed by professionally endorsed practitioners through the Career Industry Council of Australia, including specialists for Aboriginal communities and people with disabilities (see Chapter 3, Box 3.9). This type of support can only be mobilised in collaboration with social partners and requires targeting and tailoring to particular groups. For example, Austrian labour foundations (Arbeitsstiftungen) offer tailored training programmes, career counselling and job placement services in contexts such as mass lay-off events, or help disadvantaged populations to re-enter the workforce (Cedefop, 2020[133]; OECD, 2025[134]; 2025[135]).
Such support, supplemented by targeted interventions that consider local labour market dynamics and issues, are particularly critical in the context of industrial transitions (e.g. coal miners needing to shift to new sectors due to the green transition) and technological disruptions. Investing in career guidance for adults – including digital tools such as online skills assessment and job matching platforms – helps to reduce labour market mismatches and shortens unemployment spells. For example, the Danish “Boss Ladies” project seeks to attract more women into construction, an industry where women currently represent just 9% of the workforce. The project employs ambassadors, collaborates with VET institutions, provides career guidance and apprenticeship matching, and connects women with training opportunities, providing critical entry points into the industry (see Chapter 3, Box 3.3).
PES providers also play a crucial role in guiding unemployed individuals back into the labour market, and there are a range of possible interventions for the diverse needs of these populations. Policies such as Luxembourg’s FutureSkills Initiative targets unemployed individuals with secondary education, offering three months of training followed by a six-month public sector internship. Spain’s “Vives Emplea Saludable” integrates physical, emotional and social well-being into employment support programmes among unemployed individuals (see Chapter 3, Box 3.5). In designing coherent and comprehensive strategies to increase labour force participation through career guidance and adjacent programmes, policymakers should consider the breadth of support that can be provided and different means of delivery.
Career guidance complements adult learning policies, as even if training opportunities exist, many adults will not utilise them optimally without guidance to identify which skills to develop for sustainable careers. Some countries have introduced individual case management for those who are long-term unemployed or vulnerable workers, pairing them with career counsellors or job coaches. Others have created mid-career “check-ups”, where workers at mid-life (age 40-50) are offered a free career review to encourage upskilling before their skills become obsolete. These practices recognise that career development is a continuous process – not a one-time choice made in youth – and that guidance should be available throughout a person’s working life.
By helping people make better-informed educational and job choices, career guidance improves the efficiency of skills utilisation in the economy, reducing mismatches, and contributes to fairer outcomes by ensuring that talent is not lost due to lack of information or connections. For policymakers, this means treating career guidance as an integral part of the skills development ecosystem, investing in its reach and quality, and embedding it at multiple life stages.
1.4.7. Developing robust skills monitoring for diverse adult populations
The way adult skills are assessed can significantly influence the observed gaps between different socio-demographic groups. Different assessment tools may portray proficiency levels differently across heterogeneous populations, meaning that what appears as a “gap” between groups can widen or narrow depending on how and what is measured. For example, standard literacy tests (conducted in the host-country language) may exaggerate skills gaps for immigrants by not fully accounting for language barriers. Whereas in the past language barriers were hard to dismantle, technology can now limit their impact on people’s possibility to effectively participate in certain jobs and, increasingly, certain labour markets that operate in English rather than other official languages (Marconi, Vergolini and Borgonovi, 2023[136]). Similarly, assessments not designed to cater to diverse participants can misrepresent group differences. For example, adults with disabilities and impairments often score lower on traditional literacy and numeracy tests than the general population, which partly reflects test formats that fail to accommodate their needs. If universal design features are lacking (e.g. accessible interfaces or adapted time limits), testing environments can prevent certain groups from demonstrating their true competencies, thereby skewing the measured gaps. Even gender differences in skills may fluctuate with the choice of assessment framework: research finds that men perform better on certain types of literacy items (e.g. texts that are short, non-continuous or contain numerical content), whereas women do as well as men on other item types (Borgonovi, 2022[137]).
Failing to consider the full diversity of learners can mask the reasons behind low achievement in certain segments of the population, limiting policymakers’ capacity to respond. Who and how we assess matters: robust monitoring tools must be sensitive to gender, socio-economic, cultural and age differences so that measured skills gaps reflect reality rather than artefacts of the test (Borgonovi and Suarez-Alvarez, 2025[138]).
In OECD labour markets today, populations are more diverse than ever in terms of gender and migration, socio-economic and educational backgrounds. However, many adult skills assessments were not originally designed with such heterogeneity in mind. This mismatch has tangible consequences. Standardised surveys that ignore diversity risk presenting an overly narrow picture of the skills individuals have and miss opportunities to inform policy with meaningful data. Developing more inclusive and flexible assessment tools would yield a more accurate and complete skills profile of the entire adult population.
Modernising adult skills monitoring systems to better capture the full spectrum of abilities across diverse groups of learners involves both widening coverage (ensuring under-represented groups participate) and refining measurement approaches. Reflecting diverse life experiences and skills applications can be achieved by tailoring assessments to adults with different experiences. In practice, this could mean offering test items that better mirror real-world tasks in various cultural settings, or, in the context of international assessments, allowing nationally relevant modules alongside an international core to acknowledge country-specific differences in skill sets. Advancing the flexibility and accessibility of assessments can help to better represent the strengths and vulnerabilities of different population groups. Technological innovations provide promising avenues: computer-based adaptive testing can adjust question difficulty to an individual’s skills level, producing more precise results for both low and high performers. Using large item banks that include easier items and practical scenarios can improve measurement at the lower end of proficiency, where many disadvantaged adults are found. Multimedia and interactive problem solving tasks may also engage adults with different learning styles or language profiles, thereby capturing skills that traditional text-heavy tests miss. Importantly, embedding universal design and accommodations (such as assistive technologies or extra time) can enable people with disabilities or low digital literacy to fully participate, making the assessment far more representative. These improvements would help ensure that observed skills gaps truly signal areas in need of intervention, rather than reflecting biases of the instrument.
Assessment systems that embrace heterogeneity yield more precise data for decision making and better serve broader social goals. By designing tools that consider the heterogeneity in the characteristics of target populations, policymakers can identify the needs of entire communities and thus formulate effective, inclusive skills policies (Borgonovi and Suarez-Alvarez, 2025[138]). Building more adequate adult skills monitoring tools is not an end in itself but a means to drive both personalised learning pathways and smarter public policy. High-quality skills data, disaggregated by gender, socio-economic status, age, immigrant status or region, provide invaluable insights for guiding interventions. At the individual level, improved assessments can support tailored guidance and training. When adults receive detailed feedback on their skills strengths and weaknesses, they and their instructors can pinpoint what upskilling or retraining is needed. For instance, assessments used in learning settings help educators identify each learner’s gaps in real time so that they can adjust instruction accordingly. Even large-scale surveys primarily intended for monitoring could be made more useful to participants. Incorporating feedback mechanisms into surveys such as the OECD Survey of Adult Skills – for example, by giving test-takers an indication of their results and areas to improve – would make the experience more rewarding for adults and increase their engagement. This in turn yields more reliable data for policymakers, creating a virtuous circle: adults are motivated to participate and improve, and policymakers get better information on which to act. Ultimately, an accurate skills assessment can empower individuals to make informed decisions about further training, career moves or skills certification. It becomes a diagnostic tool guiding adults to the opportunities that match their needs – whether a numeracy refresher course for someone with low basic skills or an advanced ICT workshop for an older worker adapting to technological change. For example, the Slovakian "Digital Seniors" project initiative provides digital training to over 100 000 seniors and people with disabilities by providing training and tools to incentivise ongoing learning, highlighting the need for a multi-faceted approach to supporting sustained learning (see Chapter 3, Box 3.10).
At the system level, comprehensive and granular skills data enable more effective policy design, evaluation and resource allocation. When monitoring tools capture which groups are falling behind and in what areas, governments can target investments to where they will have the greatest impact. For example, if assessments reveal that adults in certain rural regions have particularly low digital skills, training funds and programmes can be directed there. Likewise, persistent gender or socio-economic skills gaps in the data can trigger the design of specialised outreach and support (such as literacy programmes for low-educated adults or mentoring for women in STEM fields). Robust measurement also helps in tracking progress and accountability: policymakers can set baselines and objectives (e.g. closing the urban–rural gap in basic skills or improving immigrant language proficiency over time) and use periodic assessment results to evaluate whether policies are working. System-level assessments are crucial for identifying equity gaps and guiding strategic planning towards national skills goals. Moreover, linking assessment outcomes with administrative data (on employment, earnings, training uptake, etc.) allows for the rigorous evaluation of which skill investments yield returns. This evidence-based approach supports allocating public funding efficiently – bolstering programmes that demonstrably raise skills levels in target groups and redesigning or discontinuing those that do not. In addition, transparent results from skills monitoring can strengthen accountability for institutions and training providers, ensuring the responsible use of public funds and quality assurance in adult learning.
1.4.8. Aligning skills with labour market needs: Improving matching and inclusive hiring through a skills-first approach
As countries work to build skills through education and training, they must also improve the matching of skills to jobs and remove barriers that prevent individuals from securing employment. Labour market institutions and employer practices can have a profound influence on whether skills disparities translate into opportunity gaps in the labour market. Policies in this domain aim to ensure that 1) employers can find and recognise the skills they need (addressing skill shortages); 2) workers can signal and use their skills effectively (addressing skills underutilisation and mismatch); and 3) progress is made against discriminatory or inefficient hiring practices so that everyone with the requisite skills can secure employment. Better matching can promote productivity and growth as firms get the talent they need and workers are more productive in roles that match their abilities. It can also promote inclusion by tackling biased hiring practices.
PES play a crucial role in matching jobseekers to jobs and in facilitating upskilling for those who are unemployed. Active labour market policies (ALMPs) such as job placement assistance, targeted training programmes and hiring subsidies can help those with skills gaps or other disadvantages find employment. For example, jobseekers who lack certain in-demand skills can be directed to short training courses (in co‑ordination with local employers who have vacancies). Hiring subsidies or apprenticeship incentives encourage employers to not overlook candidates with disadvantaged socio-economic backgrounds or disrupted learning trajectories, such as youth without experience, the long-term unemployed or persons with disabilities, and to instead focus on training them on-the-job. It is important that these programmes are well-targeted and coupled with counselling so that participants are matched to opportunities suited to their aptitudes. Many OECD countries are modernising their PES by using digital tools to profile jobseekers and predict who might need more intensive support, while simplifying access to services online. Public employment specialists and career counsellors should be trained to address biases, take into account individuals’ needs and aspirations, and understand local labour market needs to improve outcomes. Governments may outsource this work to community-based, non-profit, private and other mission-driven organisations, which can help bridge the gap between policies and the practicalities of helping those they are intended to serve (see Chapter 3, Box 3.9) (OECD, 2025[113]).
ALMPs also need to keep pace with changing skill demands, which involves close co‑ordination with education providers and employers. Fostering collaboration between PES providers, training providers and industry representatives can help address skills shortages (for example, when a new local investment creates demand for certain skills, the local workforce system can respond with a targeted training and placement programme). The ultimate goal is to prevent unfilled vacancies while skilled workers remain unemployed or underemployed due to shortcomings in the recruitment process. By continuously upgrading the skills of jobseekers and connecting them with opportunities, ALMPs contribute to both lowering unemployment and creating a more inclusive labour market.
Traditional hiring practices often rely heavily on formal credentials, such as degrees, certifications and prior job titles as proxies for skill, which can inadvertently exclude talented candidates who have followed non-formal learning pathways. For example, a seasoned IT professional may have a breadth and depth of self-taught coding skills but no relevant academic credential, and as a result may be filtered out of the recruitment process if a vacancy requires a university diploma.
A skills-first approach (see Chapter 4, Box 4.1) emphasises individuals’ skills rather than how they were acquired. In practice, it means enhancing the transparency of skills information, making it easier for employers to identify the competencies they seek and for workers to showcase their skills, thereby improving labour market matching. A skills-first approach has implications for all stakeholders in the labour market: for employers, it means rethinking their recruitment and human resource management strategies; education providers may need to re-evaluate their curricula, teaching practices and means of certifying skills; and individuals may need to consider how best to demonstrate their existing skills and identify new skills required to succeed in the labour market.
Existing forms of hiring bias such as prejudiced decision making or disregarding non-traditional skills signals have created undue barriers for individuals striving for occupational mobility. A skills-first approach can promote occupational mobility by opening pathways to opportunities for individuals who have the skills required for a particular job but who have been historically under-represented in a given occupation or industry as a result of their inherent characteristics or occupational experiences. For example, even though more women hold higher education degrees, they are less likely to signal the full range of skills and experiences associated with the qualification. A skills-first approach to hiring and human resource management that emphasises clear and direct communication about individuals’ skills could help address these gaps in communication.
Policymakers can encourage the skills-first approach by working with employers and industry bodies to develop better skills assessment tools and frameworks. For instance, creating occupational skills profiles and encouraging the use of skills tests or portfolios in recruitment can allow candidates to demonstrate ability regardless of socio-demographic characteristics. Governments themselves, as major employers, can lead by example by removing unnecessary degree requirements in public sector hiring and implementing skills assessments. Robust credentials and skills recognition, particularly for migrants and older workers, can enable their qualifications and skills to be recognised by employers in the host country. Similarly, mid-career workers who acquire new skills through short courses or online learning need those skills to be visible to employers. Expanding digital credential platforms (for verified micro-credentials and RPL) and the use of e-portfolios and professional networking tools are other potential avenues of promoting skills-first practices in the labour market.
Policymakers should also focus on addressing occupational segregation and promoting diversity in high-demand fields. For example, women and socio-economically disadvantaged workers are often clustered in lower-paying roles or sectors. This is not purely a pipeline issue but reflects workplace environments and societal norms. Policymakers can support initiatives that improve diversity and inclusion in high-growth fields traditionally dominated by men or those with advantages socio-economic backgrounds, such as engineering and finance. For example, enforcing strong anti-discrimination laws and pay transparency can diminish gender pay gaps and hiring biases. For instance, France introduced a Gender Equality Index in 2019, requiring companies with more than 50 employees, and later public institutions, to publish annual scores based on pay gaps, promotions, and representation indicators. This policy aims to foster transparency and accountability, putting measurable pressure on organisations to reduce gender inequalities. Some countries require larger companies to report on gender pay differences or diversity statistics, creating pressure for organisations to improve. Supporting returnship programmes (short-term programmes to help professionals return to careers after a break, often used to reintegrate women after having children) can help reclaim lost skills and reduce attrition from the labour force. In tech sectors, programmes that retrain women or minority workers from other fields into coding or data analytics roles have shown success when coupled with hiring commitments from industry. Governments can fund or co-sponsor such conversion programmes as part of their skills strategy.
While the onus to upskill often rests with the individual, is it up to employers to make the best use of employees’ skills. Many employees have more or different skills than their job requires, which can lead to disengagement and lower productivity (Adalet McGowan and Andrews, 2015[139]). Governments and employer associations can promote workplace practices that fully use and further develop workers’ skills, such as the Belgian Federal Public Administration Talent Exchange (OECD, 2025[113]), or similar programmes that enable employees to gain experience across different parts of an organisation. Supporting SMEs to adopt such practices (e.g. via consulting or benchmarking initiatives) could yield productivity gains.
Dialogue between employers and education providers can ensure that skills taught (especially soft skills such as communication and collaboration) are leveraged in the workplace. New graduates being underutilised would be an inefficiency to address through job design or improved graduate recruitment practices. The increasingly sophisticated capabilities of AI and its encroachment on previously untouched research-intensive jobs has created unemployment and underemployment for many recent graduates. For example, in the United States, research and consulting roles, many of which are held by recent graduates and early-career professionals, are declining quickly and at a higher rate than overall postings (Stahle, 2025[140]).
In conclusion, aligning skills with labour market needs requires interventions on both the demand side (employer practices, job design, anti-discrimination) and the supply side (employment services, mobility and training incentives). By reducing mismatches and biases, countries can reap a “double dividend”: a more efficient allocation of human capital – which research shows could boost productivity significantly (Adalet McGowan and Andrews, 2017[141]) – and a more inclusive labour market, where an individual’s background is less determinative of their job prospects. Many disparities, especially those related to socio‑economic background, would diminish if skills were equalised; for those gaps that remain, such as the gender gap, labour market-focused solutions are needed on top of skills policies. Several countries have launched national initiatives for employers to commit to skills-based hiring and some are using “diversity charters” to encourage firms to implement inclusive recruitment and progression policies. Sharing and scaling up such best practices will be crucial. Ultimately, the effectiveness of a nation’s skills investments will be undermined if those skills are not properly used. Therefore, skills policy and labour market policy must work hand in hand, ensuring that improving the skillset of the population translates into real gains in economic prosperity and social inclusion.
References
[11] Acemoglu, D. (2002), “Technical change, inequality, and the labor market”, Journal of Economic Literature, Vol. 40/1, pp. 7-72, https://doi.org/10.1257/0022051026976.
[23] Acemoglu, D. and P. Restrepo (2020), “Robots and jobs: Evidence from US labor markets”, Journal of Political Economy, Vol. 128/6, pp. 2188-2244, https://doi.org/10.1086/705716.
[38] Acemoglu, D. and P. Restrepo (2018), Artificial Intelligence, Automation and Work, National Bureau of Economic Research, https://doi.org/10.3386/w24196.
[141] Adalet McGowan, M. and D. Andrews (2017), “Skills mismatch, productivity and policies: Evidence from the second wave of PIAAC”, OECD Economics Department Working Papers, No. 1403, OECD Publishing, Paris, https://doi.org/10.1787/65dab7c6-en.
[139] Adalet McGowan, M. and D. Andrews (2015), “Labour Market Mismatch and Labour Productivity: Evidence from PIAAC Data”, OECD Economics Department Working Papers, No. 1209, OECD Publishing, Paris, https://doi.org/10.1787/5js1pzx1r2kb-en.
[1] Ahir, H., N. Bloom and D. Furceri (2022), “The World Uncertainty Index”, NBER Working Papers 29763, National Bureau of Economic Research, Vol. 29763, https://ideas.repec.org/p/nbr/nberwo/29763.html (accessed on 18 July 2025).
[95] Alesina, A. and P. Giuliano (2011), “Preferences for Redistribution”, in Handbook of Social Economics, Elsevier, https://doi.org/10.1016/b978-0-444-53187-2.00004-8.
[51] Almlund, M. et al. (2011), “Personality Psychology and Economics”, in Handbook of the Economics of Education, Handbook of The Economics of Education, Elsevier, https://doi.org/10.1016/b978-0-444-53444-6.00001-8.
[37] Autor, D. (2024), Applying AI to Rebuild Middle Class Jobs, National Bureau of Economic Research, https://doi.org/10.3386/w32140.
[25] Autor, D. (2015), “Why are there still so many jobs? The history and future of workplace automation”, Journal of Economic Perspectives, Vol. 29/3, pp. 3-30, https://doi.org/10.1257/jep.29.3.3.
[28] Autor, D., F. Levy and R. Murnane (2003), “The skill content of recent technological change: An empirical exploration”, The Quarterly Journal of Economics, Vol. 118/4, pp. 1279-1333, https://doi.org/10.1162/003355303322552801.
[120] Barnett, W. (2011), “Effectiveness of early educational intervention”, Science, Vol. 333/6045, pp. 975-978, https://doi.org/10.1126/science.1204534.
[20] Barreto, C. et al. (2024), “The “clean energy transition” and the cost of job displacement in energy-intensive industries”, OECD Social, Employment and Migration Working Papers, No. 310, OECD Publishing, Paris, https://doi.org/10.1787/abf614d1-en.
[4] Barro, R. and J. Lee (2013), “A new data set of educational attainment in the world, 1950–2010”, Journal of Development Economics, Vol. 104, pp. 184-198, https://doi.org/10.1016/j.jdeveco.2012.10.001.
[73] Baumeister, R., K. Vohs and D. Tice (2007), “The strength model of self-control”, Current Directions in Psychological Science, Vol. 16/6, pp. 351-355, https://doi.org/10.1111/j.1467-8721.2007.00534.x.
[70] Becker, G. and C. Mulligan (1997), “The endogenous determination of time preference”, The Quarterly Journal of Economics, Vol. 112/3, pp. 729-758, https://doi.org/10.1162/003355397555334.
[91] Benabou, R. and E. Ok (2001), “Social mobility and the demand for redistribution: The Poum hypothesis”, The Quarterly Journal of Economics, Vol. 116/2, pp. 447-487, https://doi.org/10.1162/00335530151144078.
[12] Berman, E., J. Bound and S. Machin (1998), “Implications of skill-biased technological change: International evidence”, The Quarterly Journal of Economics, Vol. 113/4, pp. 1245-1279, https://doi.org/10.1162/003355398555892.
[98] Black, R. et al. (2008), “Maternal and child undernutrition: Global and regional exposures and health consequences”, The Lancet, Vol. 371/9608, pp. 243-260, https://doi.org/10.1016/s0140-6736(07)61690-0.
[7] Bongaarts, J. (2009), “Human population growth and the demographic transition”, Philosophical Transactions of the Royal Society B: Biological Sciences, Vol. 364/1532, pp. 2985-2990, https://doi.org/10.1098/rstb.2009.0137.
[68] Borghans, L. and B. Golsteyn (2006), “Time discounting and the body mass index”, Economics & Human Biology, Vol. 4/1, pp. 39-61, https://doi.org/10.1016/j.ehb.2005.10.001.
[137] Borgonovi, F. (2022), “Is the literacy achievement of teenage boys poorer than that of teenage girls, or do estimates of gender gaps depend on the test? A comparison of PISA and PIAAC.”, Journal of Educational Psychology, Vol. 114/2, pp. 239-256, https://doi.org/10.1037/edu0000659.
[62] Borgonovi, F., A. Ferrara and M. Piacentini (2023), “From asking to observing. Behavioural measures of socio-emotional and motivational skills in large-scale assessments”, Social Science Research, Vol. 112, p. 102874, https://doi.org/10.1016/j.ssresearch.2023.102874.
[101] Borgonovi, F. et al. (2017), “Youth in transition: How do some of the cohorts participating in PISA fare in PIAAC?”, OECD Education Working Papers, No. 155, OECD Publishing, Paris, https://doi.org/10.1787/51479ec2-en.
[138] Borgonovi, F. and J. Suarez-Alvarez (2025), “How can adult skills assessments best meet the demands of the 21st century?”, OECD Social, Employment and Migration Working Papers, No. 319, OECD Publishing, Paris, https://doi.org/10.1787/853db37b-en.
[57] Braun, H., I. Kirsch and K. Yamamoto (2011), “An experimental study of the effects of monetary incentives on performance on the 12th-grade NAEP reading assessment”, Teachers College Record: The Voice of Scholarship in Education, Vol. 113/11, pp. 2309-2344, https://doi.org/10.1177/016146811111301101.
[76] Brenøe, A. and T. Epper (2022), “Parenting values and the intergenerational transmission of time preferences”, European Economic Review, Vol. 148, p. 104208, https://doi.org/10.1016/j.euroecorev.2022.104208.
[96] Brisson, D. et al. (2020), “A systematic review of the association between poverty and biomarkers of toxic stress”, Journal of Evidence-Based Social Work, Vol. 17/6, pp. 696-713, https://doi.org/10.1080/26408066.2020.1769786.
[99] Bryan, J. et al. (2004), “Nutrients for cognitive development in school-aged children”, Nutrition Reviews, Vol. 62/8, pp. 295-306, https://doi.org/10.1111/j.1753-4887.2004.tb00055.x.
[34] Brynjolfsson, E., D. Li and L. Raymond (2025), “Generative AI at work”, The Quarterly Journal of Economics, Vol. 140/2, pp. 889-942, https://doi.org/10.1093/qje/qjae044.
[40] Brynjolfsson, E. and T. Mitchell (2017), “What can machine learning do? Workforce implications”, Science, Vol. 358/6370, pp. 1530-1534, https://doi.org/10.1126/science.aap8062.
[116] Carneiro, P., S. Cattan and N. Ridpath (2024), The short- and medium-term effects of Sure Start on children’s outcomes, https://ifs.org.uk/publications/short-and-medium-term-impacts-sure-start-educational-outcomes.
[13] Causa, O. and Å. Johansson (2009), “Intergenerational social mobility”, OECD Economics Department Working Papers, No. 707, OECD Publishing, Paris, https://doi.org/10.1787/223106258208.
[133] Cedefop (2020), Empowering Adults Through Upskilling and Reskilling Pathways, European Centre for the Development of Vocational Training, https://doi.org/10.2801/692917.
[66] Chao, L. et al. (2009), “Time preference and its relationship with age, health, and survival probability”, Judgment and Decision Making, Vol. 4/1, pp. 1-19, https://doi.org/10.1017/s1930297500000668.
[14] Cingano, F. (2014), “Trends in income inequality and its impact on economic growth”, OECD Social, Employment and Migration Working Papers, No. 163, OECD Publishing, Paris, https://doi.org/10.1787/5jxrjncwxv6j-en.
[15] Coady, D. and A. Dizioli (2017), “Income inequality and education revisited: Persistence, endogeneity and heterogeneity”, Applied Economics, Vol. 50/25, pp. 2747-2761, https://doi.org/10.1080/00036846.2017.1406659.
[75] Cohen, J. et al. (2020), “Measuring time preferences”, Journal of Economic Literature, Vol. 58/2, pp. 299-347, https://doi.org/10.1257/jel.20191074.
[18] Corak, M. (2013), “Income inequality, equality of opportunity, and intergenerational mobility”, Journal of Economic Perspectives, Vol. 27/3, pp. 79-102, https://doi.org/10.1257/jep.27.3.79.
[109] Danish Government (2024), Ekspertgruppen om ChatGPT og andre digitale hjælpemidler [The Expert Group on ChatGPT and other digital tools] (In Danish), Ministry of Children and Education, http://www.uvm.dk/-/media/filer/uvm/aktuelt/pdf24/april/240423-ekspertgruppen-om-chatgpt-og-andre-digitale-hjaelpemidlers-rapport.pdf (accessed on 2025 August July).
[64] Dell’Acqua, F. et al. (2023), “Navigating the Jagged Technological Frontier: Field Experimental Evidence of the Effects of AI on Knowledge Worker Productivity and Quality”, SSRN Electronic Journal, https://doi.org/10.2139/ssrn.4573321.
[32] Demirci, O., J. Hannane and X. Zhu (2025), “Who is AI replacing? The impact of generative AI on online freelancing platforms”, Management Science, https://doi.org/10.1287/mnsc.2024.05420.
[21] Dixon, J., B. Hong and L. Wu (2021), “The robot revolution: Managerial and employment consequences for firms”, Management Science, Vol. 67/9, pp. 5586-5605, https://doi.org/10.1287/mnsc.2020.3812.
[58] Duckworth, A. and D. Yeager (2015), “Measurement matters: Assessing personal qualities other than cognitive ability for educational purposes”, Educational Researcher, Vol. 44/4, pp. 237-251, https://doi.org/10.3102/0013189x15584327.
[89] Dworkin, R. (1977), Taking Rights Seriously, Harvard University Press.
[29] Eloundou, T. et al. (2023), GPTs are GPTs: An Early Look at the Labor Market Impact Potential of Large Language Models, https://doi.org/10.48550/arXiv.2303.10130.
[126] European Commission (2024), Education and Training Monitor 2024 – Comparative Report, Directorate-General for Education, Youth, Sport and Culture, https://doi.org/10.2766/815875.
[74] Falk, A. et al. (2018), “Global evidence on economic preferences”, The Quarterly Journal of Economics, Vol. 133/4, pp. 1645-1692, https://doi.org/10.1093/qje/qjy013.
[127] Federal Ministry of Science, Research and Economy (2017), “National strategy on the social dimension of higher education”, https://www.sozialerhebung.at/sozdim/strategiepapier/Strategie_2017_englisch.pdf.
[92] Fong, C. (2001), “Social preferences, self-interest, and the demand for redistribution”, Journal of Public Economics, Vol. 82/2, pp. 225-246, https://doi.org/10.1016/s0047-2727(00)00141-9.
[50] Foster, N. and M. Piacentini (eds.) (2023), Innovating Assessments to Measure and Support Complex Skills, OECD Publishing, Paris, https://doi.org/10.1787/e5f3e341-en.
[3] Frankopan, P. (2024), “The age of crisis”, Cambridge Journal of Regions, Economy and Society, Vol. 18/1, pp. 249-254, https://doi.org/10.1093/cjres/rsae043.
[39] Frey, C. and M. Osborne (2017), “The future of employment: How susceptible are jobs to computerisation?”, Technological Forecasting and Social Change, Vol. 114, pp. 254-280, https://doi.org/10.1016/j.techfore.2016.08.019.
[27] Gaggl, P. and G. Wright (2017), “A short-run view of what computers do: Evidence from a UK tax incentive”, American Economic Journal: Applied Economics, Vol. 9/3, pp. 262-294, https://doi.org/10.1257/app.20150411.
[125] Government of Canada (2025), The Government of Canada invests in artificial intelligence and teaching French as a second language from early childhood, New release, 10 May 2023, Canadian Heritage, http://www.canada.ca/en/canadian-heritage/news/2023/05/thegovernment-of-canada-invests-in-artificial-intelligence-and-teaching-french-as-a-second-language-from-early-childhood.html (accessed on 9 July 2025).
[65] Hanushek, E. et al. (2025), “Patience and subnational differences in human capital: Regional analysis with Facebook interests”, The Economic Journal, https://doi.org/10.1093/ej/ueaf045.
[121] Heckman, J. (2006), “Skill formation and the economics of investing in disadvantaged children”, Science, Vol. 312/5782, pp. 1900-1902, https://doi.org/10.1126/science.1128898.
[122] Heckman, J. et al. (2010), “The rate of return to the HighScope Perry Preschool Program”, Journal of Public Economics, Vol. 94/1-2, pp. 114-128, https://doi.org/10.1016/j.jpubeco.2009.11.001.
[52] Heckman, J., J. Stixrud and S. Urzua (2006), “The Effects of Cognitive and Noncognitive Abilities on Labor Market Outcomes and Social Behavior”, Journal of Labor Economics, Vol. 24/3, pp. 411-482, https://doi.org/10.1086/504455.
[2] Henig, D. and D. Knight (2023), “Polycrisis: Prompts for an emerging worldview”, Anthropology Today, Vol. 39/2, pp. 3-6, https://doi.org/10.1111/1467-8322.12793.
[43] Hui, X., O. Reshef and L. Zhou (2023), “The short-term effects of generative artificial intelligence on employment: Evidence from an online labor market”, SSRN Electronic Journal, https://doi.org/10.2139/ssrn.4527336.
[67] Hunter, R. et al. (2018), “Association between time preference, present-bias and physical activity: implications for designing behavior change interventions”, BMC Public Health, Vol. 18/1, https://doi.org/10.1186/s12889-018-6305-9.
[106] IEA (2015), Trends in International Mathematics and Science Study (TIMSS) 2015 Grade 4 Database, International Association for the Evaluation of Educational Achievement, https://doi.org/10.58150/IEA_TIMSS_2015_G4.
[105] IEA (2011), Trends in International Mathematics and Science Study (TIMSS) 2011 Grade 4 Database, International Association for the Evaluation of Educational Achievement, https://doi.org/10.58150/IEA_TIMSS_2011_G4.
[142] IEA (1995), Trends in International Mathematics and Science Study (TIMSS) 1995 Grade 4 Database, International Association for the Evaluation of Educational Achievement, https://doi.org/10.58150/IEA_TIMSS_1995_G4 (accessed on 8 July 2025).
[69] Kang, M. and S. Ikeda (2013), “Time discounting and smoking behavior: Evidence from a panel survey”, Health Economics, Vol. 23/12, pp. 1443-1464, https://doi.org/10.1002/hec.2998.
[77] Kankaraš, M. (2017), “Personality matters: Relevance and assessment of personality characteristics”, OECD Education Working Papers, No. 157, OECD Publishing, Paris, https://doi.org/10.1787/8a294376-en.
[54] Kankaraš, M. and J. Suarez-Alvarez (2019), “Assessment framework of the OECD Study on Social and Emotional Skills”, OECD Education Working Papers, No. 207, OECD Publishing, Paris, https://doi.org/10.1787/5007adef-en.
[83] Kern, M. et al. (2016), “The EPOCH measure of adolescent well-being”, Psychological Assessment, Vol. 28/5, pp. 586-597, https://doi.org/10.1037/pas0000201.
[22] Koch, M., I. Manuylov and M. Smolka (2021), “Robots and firms”, The Economic Journal, Vol. 131/638, pp. 2553-2584, https://doi.org/10.1093/ej/ueab009.
[131] Korlat, S. et al. (2022), “Gender Typicality and Prestige of Occupational Aspirations in Adolescents: The Relevance of Agency and Communion”, Journal of Career Development, Vol. 50/2, pp. 405-424, https://doi.org/10.1177/08948453221100744.
[55] Kyllonen, P. et al. (2024), “Charting the future of assessments”, ETS Research Report Series, Vol. 2024/1, pp. 1-62, https://doi.org/10.1002/ets2.12388.
[124] La Direction de l’évaluation, de la prospective et de la performance (2024), TIMSS 2023 en CM1, https://www.education.gouv.fr/timss-2023-en-cm1-analyse-des-questions-de-sciences-415955.
[6] Lesthaeghe, R. (2010), “The unfolding story of the Second Demographic Transition”, Population and Development Review, Vol. 36/2, pp. 211-251, https://doi.org/10.1111/j.1728-4457.2010.00328.x.
[46] Lightcast (2025), The Speed of Skill Change, Lightcast, https://lightcast.io/resources/research/speed-of-skill-change (accessed on 18 March 2025).
[31] Lightcast (2025), The Speed of Skill Change: Every Job in the US is Changing. Are you?, https://4906807.fs1.hubspotusercontent-na1.net/hubfs/4906807/Speed%20of%20Skill%20Change_US_Lightcast-final.pdf.
[30] Lightcast (2025), The UK Skills Revolution: Building a Data-driven Skills Systen in an Era of Disruption, https://www.datocms-assets.com/62658/1738249792-uk-skills-revolution-final-30-jan.pdf.
[60] Linn, R. (ed.) (1989), Validity, New York, NY: American Council on education and Macmillan.
[44] Liu, J. et al. (2024), “Generate” the Future of Work through AI: Empirical Evidence from Online Labor Markets, https://doi.org/10.48550/arXiv.2308.05201.
[100] Lupien, S. et al. (2009), “Effects of stress throughout the lifespan on the brain, behaviour and cognition”, Nature Reviews Neuroscience, Vol. 10/6, pp. 434-445, https://doi.org/10.1038/nrn2639.
[41] Lysyakov, M. and S. Viswanathan (2021), “Threatened by AI: Analyzing Users’ Responses to the Introduction of AI in a Crowd-sourcing Platform”, SSRN Electronic Journal, https://doi.org/10.2139/ssrn.3758338.
[26] Mann, K. and L. Püttmann (2023), “Benign effects of automation: New evidence from patent texts”, Review of Economics and Statistics, Vol. 105/3, pp. 562-579, https://doi.org/10.1162/rest_a_01083.
[136] Marconi, G., L. Vergolini and F. Borgonovi (2023), “The demand for language skills in the European labour market: Evidence from online job vacancies”, OECD Social, Employment and Migration Working Papers, No. 294, OECD Publishing, Paris, https://doi.org/10.1787/e1a5abe0-en.
[123] Martinot, P. et al. (2025), “Rapid emergence of a maths gender gap in first grade”, Nature, https://doi.org/10.1038/s41586-025-09126-4.
[63] Milanez, A., A. Lemmens and C. Ruggiu (2025), “Algorithmic management in the workplace: New evidence from an OECD employer survey”, OECD Artificial Intelligence Papers, No. 31, OECD Publishing, Paris, https://doi.org/10.1787/287c13c4-en.
[35] Noy, S. and W. Zhang (2023), “Experimental evidence on the productivity effects of generative artificial intelligence”, Science, Vol. 381/6654, pp. 187-192, https://doi.org/10.1126/science.adh2586.
[113] OECD (2025), Empowering the Workforce in the Context of a Skills-First Approach, OECD Skills Studies, OECD Publishing, Paris, https://doi.org/10.1787/345b6528-en.
[134] OECD (2025), OECD Employment Outlook 2025: Can We Get Through the Demographic Crunch?, OECD Publishing, Paris, https://doi.org/10.1787/194a947b-en.
[135] OECD (2025), Promoting Better Career Mobility for Longer Working Lives in Austria, Ageing and Employment Policies, OECD Publishing, Paris, https://doi.org/10.1787/db85473f-en.
[119] OECD (2025), Reducing Inequalities by Investing in Early Childhood Education and Care, Starting Strong, OECD Publishing, Paris, https://doi.org/10.1787/b78f8b25-en.
[79] OECD (2025), Skills that Matter for Success and Well-being in Adulthood: Evidence on Adults’ Social and Emotional Skills from the 2023 Survey of Adult Skills, OECD Skills Studies, OECD Publishing, Paris, https://doi.org/10.1787/6e318286-en.
[130] OECD (2024), Challenging Social Inequality Through Career Guidance: Insights from International Data and Practice, OECD Publishing, Paris, https://doi.org/10.1787/619667e2-en.
[112] OECD (2024), Co-ordinating Adult Learning Policies: Mechanisms for Inter-Institutional and Stakeholder Co-ordination, OECD, Paris, https://www.oecd.org/en/publications/co-ordinating-adult-learning-policies_725d9938-en.html (accessed on 10 July 2025).
[47] 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.
[111] OECD (2024), Insights from Skills Strategies in the European Union: Lessons Learnt for Developing and Implementing Effective Skills Policies, OECD Skills Studies, OECD Publishing, Paris, https://doi.org/10.1787/0bf9e78e-en.
[16] OECD (2024), OECD Employment Outlook 2024: The Net-Zero Transition and the Labour Market, OECD Publishing, Paris, https://doi.org/10.1787/ac8b3538-en.
[104] OECD (2024), PIAAC 2nd cycle database, Survey of Adult Skills, OECD, Paris, https://www.oecd.org/en/data/datasets/piaac-2nd-cycle-database.html (accessed on 10 December 2024).
[48] OECD (2024), PIAAC data and methodology, OECD, Paris, https://www.oecd.org/en/about/programmes/piaac/piaac-data.html (accessed on 10 December 2024).
[114] OECD (2024), Quality Matters: Strengthening the Quality Assurance of Adult Education and Training, OECD Skills Studies, OECD Publishing, Paris, https://doi.org/10.1787/f44a185b-en.
[78] OECD (2024), Social and Emotional Skills for Better Lives: Findings from the OECD Survey on Social and Emotional Skills 2023, OECD Publishing, Paris, https://doi.org/10.1787/35ca7b7c-en.
[86] OECD (2024), Survey of Adult Skills (PIAAC) 2nd cycle database, OECD, Paris, https://www.oecd.org/en/data/datasets/piaac-2nd-cycle-database.html (accessed on 10 December 2024).
[45] OECD (2024), “Using AI in the workplace: Opportunities, risks and policy responses”, OECD Artificial Intelligence Papers, No. 11, OECD Publishing, Paris, https://doi.org/10.1787/73d417f9-en.
[19] OECD (2023), OECD Skills Outlook 2023: Skills for a Resilient Green and Digital Transition, OECD Publishing, Paris, https://doi.org/10.1787/27452f29-en.
[5] OECD (2022), Education at a Glance 2022: OECD Indicators, OECD Publishing, Paris, https://doi.org/10.1787/3197152b-en.
[118] OECD (2022), “Finland’s Right to Learn Programme: Achieving equity and quality in education”, OECD Education Policy Perspectives, No. 61, OECD Publishing, Paris, https://doi.org/10.1787/65eff23e-en.
[108] OECD (2022), PISA 2022 Database, OECD, Paris, https://www.oecd.org/en/data/datasets/pisa-2022-database.html (accessed on 5 November 2024).
[9] OECD (2021), OECD Skills Outlook 2021: Learning for Life, OECD Publishing, Paris, https://doi.org/10.1787/0ae365b4-en.
[49] OECD (2021), The Assessment Frameworks for Cycle 2 of the Programme for the International Assessment of Adult Competencies, OECD Skills Studies, OECD Publishing, Paris, https://doi.org/10.1787/4bc2342d-en.
[110] OECD (2020), Strengthening the Governance of Skills Systems: Lessons from Six OECD Countries, OECD Skills Studies, OECD Publishing, Paris, https://doi.org/10.1787/3a4bb6ea-en.
[24] OECD (2019), OECD Employment Outlook 2019: The Future of Work, OECD Publishing, Paris, https://doi.org/10.1787/9ee00155-en.
[10] OECD (2018), A Broken Social Elevator? How to Promote Social Mobility, OECD Publishing, Paris, https://doi.org/10.1787/9789264301085-en.
[107] OECD (2018), PISA 2018 Database, OECD, Paris, https://www.oecd.org/pisa/data/2018database/ (accessed on 5 November 2024).
[81] OECD (2015), Skills for Social Progress: The Power of Social and Emotional Skills, OECD Skills Studies, OECD Publishing, Paris, https://doi.org/10.1787/9789264226159-en.
[53] OECD (2012), Literacy, Numeracy and Problem Solving in Technology-Rich Environments: Framework for the OECD Survey of Adult Skills, OECD Publishing, Paris, https://doi.org/10.1787/9789264128859-en.
[102] 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.
[85] Park, H. et al. (2020), “Meta-analytic five-factor model personality intercorrelations: Eeny, meeny, miney, moe, how, which, why, and where to go.”, Journal of Applied Psychology, Vol. 105/12, pp. 1490-1529, https://doi.org/10.1037/apl0000476.
[115] Payne, J. (ed.) (2025), The short- and medium-term effects of Sure Start on children’s outcomes, The Institute for Fiscal Studies, https://ifs.org.uk/sites/default/files/2025-05/IFS%20Report.%20The%20short-%20and%20medium-term%20effects%20of%20Sure%20Start%20on%20children%E2%80%99s%20outcomes.pdf (accessed on 8 July 2025).
[128] Polish Government (2024), Projekt ustawy o rynku pracy i służbach zatrudnienia [Draft Act on the labor market], https://www.gov.pl/web/premier/projekt-ustawy-o-rynku-pracy-i-sluzbach-zatrudnienia.
[72] Przybylski, A. and N. Weinstein (2012), “Can you connect with me now? How the presence of mobile communication technology influences face-to-face conversation quality”, Journal of Social and Personal Relationships, Vol. 30/3, pp. 237-246, https://doi.org/10.1177/0265407512453827.
[33] Qiao, D., H. Rui and Q. Xiong (2024), AI and Jobs: Has the Inflection Point Arrived? Evidence from an Online Labor Platform, https://doi.org/10.48550/arXiv.2312.04180.
[88] Rawls, J. (1971), A Theory of Justice, The Belknap Press of Harvard University Press.
[132] Robbins, S. (2024), Unlocking Workforce Potential with the Community College Skills-Based Hiring Playbook, ACCT Now, https://perspectives.acct.org/stories/unlocking-workforce-potential-with-the-community-college-skills-based-hiring-playbook (accessed on 9 July 2025).
[87] Roemer, J. (2012), “On several approaches to equality of opportunity”, Economics and Philosophy, Vol. 28/2, pp. 165-200, https://doi.org/10.1017/s0266267112000156.
[71] Ross, C. and J. Mirowsky (1999), “Refining the association between education and health: The effects of quantity, credential, and selectivity”, Demography, Vol. 36/4, pp. 445-460, https://doi.org/10.2307/2648083.
[59] Sackett, P., M. Borneman and B. Connelly (2008), “High stakes testing in higher education and employment: Appraising the evidence for validity and fairness”, American Psychologist, Vol. 63/4, pp. 215-227, https://doi.org/10.1037/0003-066x.63.4.215.
[94] Sapolsky, R. (2023), Determined: A Science of Life Without Free Will, Penguin Press.
[90] Sen, A. (1999), Development as Freedom, Oxford University Press, New York.
[103] Soldani, E. et al. (2024), “Policy approaches to reduce inequalities while boosting productivity growth”, OECD Economics Department Working Papers, No. 1819, OECD Publishing, Paris, https://doi.org/10.1787/af121c8f-en.
[80] Soto, C. and O. John (2017), “The next Big Five Inventory (BFI-2): Developing and assessing a hierarchical model with 15 facets to enhance bandwidth, fidelity, and predictive power”, Journal of Personality and Social Psychology, Vol. 113/1, pp. 117-143, https://doi.org/10.1037/pspp0000096.
[140] Stahle, C. (2025), May 2025 US Labor Market Update: What Indeed’s Data Reveals About the Early Impact of Changing Policies on the Job Market, Hiring Lab, http://www.hiringlab.org/2025/05/22/may-2025-us-labor-market-update/ (accessed on 10 July 2025).
[93] Starmans, C., M. Sheskin and P. Bloom (2017), “Why people prefer unequal societies”, Nature Human Behaviour, Vol. 1/4, https://doi.org/10.1038/s41562-017-0082.
[36] Stephany, F. and O. Teutloff (2024), “What is the price of a skill? The value of complementarity”, Research Policy, Vol. 53/1, p. 104898, https://doi.org/10.1016/j.respol.2023.104898.
[82] Steponavičius, M., C. Gress-Wright and A. Linzarini (2023), “Social and emotional skills: Latest evidence on teachability and impact on life outcomes”, OECD Education Working Papers, No. 304, OECD Publishing, Paris, https://doi.org/10.1787/ba34f086-en.
[84] Taylor, R. et al. (2017), “Promoting positive youth development through school‐based social and emotional learning interventions: A meta‐analysis of follow‐up effects”, Child Development, Vol. 88/4, pp. 1156-1171, https://doi.org/10.1111/cdev.12864.
[42] Teutloff, O. et al. (2025), “Winners and losers of generative AI: Early evidence of shifts in freelancer demand”, Journal of Economic Behavior & Organization, Vol. 235, p. 106845, https://doi.org/10.1016/j.jebo.2024.106845.
[56] Ulitzsch, E. et al. (2021), “Model meets reality: Validating a new behavioral measure for test-taking effort”, Educational Assessment, Vol. 26/2, pp. 104-124, https://doi.org/10.1080/10627197.2020.1858786.
[17] UNESCO (2024), Global Education Monitoring Report, 2024/5, Leadership in education: Lead for learning, United Nations Educational, Scientific and Cultural Organization, https://doi.org/10.54676/eflh5184.
[117] unison (2024), Cuts to children’s centres hitting the most disadvantaged, https://www.unison.org.uk/news/press-release/2024/06/cuts-to-childrens-centres-hitting-the-most-disadvantaged/.
[8] United Nations (2024), World Population Prospects 2024, Department of Economic and Social Affairs, https://population.un.org/wpp/ (accessed on 12 July 2025).
[97] Vilar-Compte, M. et al. (2021), “AUrban poverty and nutrition challenges associated with accessibility to a healthy diet: A global systematic literature review”, International Journal for Equity in Health, Vol. 20/1, https://doi.org/10.1186/s12939-020-01330-0.
[61] Wise, S. and X. Kong (2005), “Response time effort: A new measure of examinee motivation in computer-based tests”, Applied Measurement in Education, Vol. 18/2, pp. 163-183, https://doi.org/10.1207/s15324818ame1802_2.
[129] Workforce Singapore (2025), Jobs Transformation Maps, https://www.wsg.gov.sg/home/employers-industry-partners/jobs-transformation-maps.
← 1. The Gini coefficient measures the extent to which distribution (for example of income) among individuals or households within an economy deviates from a perfectly equal distribution. A Gini index of 0 represents perfect equality, while an index of 100 implies perfect inequality.
← 2. Green driven occupations comprise all jobs that are likely to be positively affected by the net-zero transition, even those that are not green in themselves. In particular, this includes the following categories: 1) green new and emerging occupations: new occupations (entirely novel or “spinoffs” from an existing occupation) with unique tasks and worker requirements (e.g. biomass plant engineers; carbon trading analysts; solar photovoltaic installers); 2) green-enhanced skills occupations: existing occupations whose tasks, skills, knowledge, and external elements, such as credentials, tend to be altered because of the net-zero transition (e.g. arbitrators, mediators, and conciliators; architects; automotive specialty technicians; farmers and ranchers). However, even if the net-zero transition alters the characteristics of these jobs, in non-green sectors of the economy (e.g. certain GHG-intensive industries, such as chemicals, fossil fuel power generation), these occupations may still be associated with the old (non-green) list of tasks, skills, knowledge and credentials, and their demand may therefore not necessarily grow in the short term; 3) green increased demand occupations: existing occupations in increased demand due to the net-zero transition but with no significant changes in tasks or worker requirements. Some occupations in this group can be considered as directly contributing to low emissions and clearly involve green tasks (e.g. environmental scientists and specialists; forest and conservation workers) but most are not and should rather be seen as in support of green economic activities (e.g. construction workers; drivers; chemists and materials scientists) (OECD, 2024[16]).
← 3. The social and emotional skills used in the OECD Skills Outlook 2025 are based on the BFI-2-XS Extra Short Big Five Inventory. This is the instrument administered across most of the countries that took part in the OECD Survey of Adult Skills.
← 4. This report uses parental education, parental occupation, immigrant background and childhood residential context to distinguish groups. Parental occupation (at respondents’ age 14) is based on the International Classification of Occupations (ISCO) and grouped into high-status: managers, professionals, and technicians and associate professionals (ISCO 1-3); and low-status: clerical support workers; service and sales workers; skilled agricultural, forestry and fishery workers; craft and related trades workers; plant and machine operators, and assemblers; and elementary occupations (ISCO 4-9). Parental education (at respondents’ age 14) is based on the International Standard Classification of Education (ISCED) 2011 and grouped into tertiary-educated (having at least one parent who had attained tertiary education [ISCED 5, 6, 7 and 8]) and non-tertiary-educated parents. Childhood residential context (at respondents’ age 14) refers to whether the respondent grew up in a village, town or city. Groups by immigrant background distinguish between children of immigrants, immigrants and non-immigrants. Children of immigrants were born in the country in which they reside, but their parents were not, or they were born in a different country and moved to their current country of residence before the age of 18. Immigrants are defined as those who migrated to their country of residence at age 18 or older. Non-immigrants were born in their country of residence, as were their parents.
← 5. In cases where information is available for only one parent/guardian, the occupation of that parent/guardian is used for the analysis.
← 6. For a definition of population subgroups by immigrant background please refer to section 1.3.1.