Do Adults Have the Skills They Need to Thrive in a Changing World?

Since the publication of results from the first Survey of Adult Skills in 2013, there has been growing interest in the importance of skills, as opposed to traditional educational qualifications, to individual and societal well-being. The OECD has repeatedly highlighted the importance of skills in today's rapidly evolving labour markets (OECD, 2023[1]; 2019[2]; 2019[3]; 2012[4]). It has also advocated policies that promote the development and use of skills, such as skills assessment and anticipation, upskilling and reskilling, and the recognition of skills acquired in non-formal and informal contexts (OECD, 2021[5]; 2021[6]; 2019[7]).

Similarly, governments around the world increasingly recognise the importance of aligning education and training with the changing needs of labour markets and societies, with an emphasis on the acquisition of skills for economic (e.g. employability or wages) and societal participation (e.g. well-being, health and political engagement). “Skills-based” or “skills-first” approaches1 are gaining prominence in human resource management decisions in organisations worldwide (OECD, 2024[8]). This shift in emphasis reflects a wider recognition that, in the 21st century, it is skills, rather than formal qualifications alone, that determine individual employability and economic success, as well as societal well-being.

Policy makers are also interested in the efficient use of skills in the economy, and therefore in reducing mismatches between skill demand and supply. While some degree of mismatch is inevitable in dynamic labour markets, persistent mismatches impose costs on economies, firms and individuals – usually in the form of lower wages, productivity and job satisfaction (Adalet McGowan and Andrews, 2015[9]). In an era of widespread labour shortages, ensuring the efficient matching of workers to jobs is vital, and prioritising skills can be part of the solution (Causa et al., 2022[10]; OECD, 2024[8]; 2022[11]).

The 2023 Survey of Adult Skills offers valuable insights into these issues and can guide strategies to close skill gaps, increase productivity, promote equitable long-term economic growth, and foster social cohesion and societal well-being (Hanushek and Woessmann, 2021[12]). This chapter explores the importance of information-processing skills for the well-being of individuals, economies and societies.2 It has three main objectives:

• examine how skills and education relate to labour-market outcomes and wages

• explore how skills affect social outcomes such as life satisfaction, health and political efficacy

• assess the extent of mismatches and their impact on wages and life satisfaction.3

Skills enable adults to perform tasks more efficiently, leading to improved employability and higher wages. The relationship between skills, productivity and earnings is well established in economic theory and supported by empirical evidence. According to standard microeconomic theory, wages generally reflect workers' productivity; therefore, individuals with higher skills, who are more productive, are expected to earn higher wages. Moreover, as skilled individuals have more to gain from employment, they are more likely to participate in the labour market to realise these benefits. The results of the first cycle of the Survey of Adult Skills confirmed that literacy and numeracy skills play a key role in labour-market outcomes, over and above the influence of formal educational attainment.

Prior to the Survey of Adult Skills, few studies had examined the labour-market rewards of skills independently of formal qualifications. Instead, formal educational attainment had been used as a proxy for skill level, thus blurring the distinctions between the two (Barro and Lee, 2013[13]; Hanushek and Woessmann, 2011[14]). This is not an unfounded assumption in theory, as there are reasonable grounds to believe that individuals with greater skills will tend to pursue education, and that education itself is a means of developing skills – meaning that skills and education will be correlated. Indeed, this idea has roots in human capital theory (Becker, 1964[15]; Mincer, 1970[16]).

However, formal education is an imperfect proxy for skills. For example, typical measures of education are too coarse to accurately capture educational quality, cannot account for variation in individuals' own skills within educational levels and can be complicated to compare across countries. There are reasons to believe that an important purpose of formal education is teaching social and emotional skills and in developing attitudes and motivations that, though crucial, are not directly captured in achievement tests (Durlak et al., 2011[17]; Heckman and Kautz, 2012[18]). Moreover, further theoretical work has raised the possibility that skill development is not the only reason why families and individuals invest in education. Rather, knowing that true ability is difficult for an employer to assess directly, they may use education to send a signal about their ability or to pass an employer’s screening process (Arrow, 1973[19]; Spence, 1973[20]; Stiglitz, 1975[21]), secure a favourable position in a job queue (Thurow, 1975[22]), or to maintain access to educational and occupational opportunities that are restricted to other segments of society (Collins, 2019[23]; Murphy, 1988[24]). There are, therefore, good reasons to examine skills independently of education.

The Survey of Adult Skills provides an opportunity to understand the relationship between skills and labour-market and social outcomes over and above their association with educational attainment. Results from the first cycle of the survey highlighted the importance of both educational attainment and skills in determining labour-market and social outcomes. The analysis also showed that the relationship between skills, employment and wages differs across countries, reflecting differences in labour-market institutions and employers in hiring, promotion and wage-setting practices (2019[27]; OECD, 2016[28]; OECD, 2013[29]). Further analysis using data from the first cycle has demonstrated that the returns to formal education tend to decline in the context of educational expansion, while this is not the case for skills (Araki, 2020[30]). Moreover, the magnitude of the association between formal education and labour-market outcomes is greatest early on in individuals' careers, with skills becoming more important as adults gain work experience and enter their prime working years (Hanushek et al., 2015[31]). This may be because employers initially rely on easily observable indicators of worker quality, such as formal education, but learn about their employees’ skills once they have been hired – a phenomenon known as “employer learning” – and reward them accordingly (OECD, 2014[32]).

The 2023 Survey of Adult Skills confirms that both skills and education are positively associated with the likelihood of being employed. This positive association reflects the fact that individuals with higher skills are more likely to be employed, as well as the fact that employment provides further opportunities for individuals to improve their skills. In addition, the analysis suggests that the strength of this relationship varies across countries and economies.

One possible explanation for differences in the extent to which formal qualifications, as opposed to skills, are associated with employment status is the degree of “skills transparency” in a given country: how informative formal qualifications are about the actual skills of individuals. If formal qualifications accurately reflect true skills (that are otherwise difficult to observe), employers may prefer to hire workers with higher qualifications (Heisig, Gesthuizen and Solga, 2019[33]). Countries with a larger gap in information-processing skills between low-educated workers and those with intermediate educational qualifications have been found to also have a larger gap between both groups in the likelihood of being employed, providing further evidence that employers treat higher qualifications as a proxy for ability (Abrassart, 2013[34]). Another possible explanation is that educational attainment captures a wider range of skills, including social and emotional skills such as perseverance or conscientiousness, that employers in some labour markets may value more than information-processing skills alone.

Average proficiency in literacy, numeracy and adaptive problem solving is higher among the employed population than the unemployed or inactive population (Table 4.1). This is true for both full-time and part-time employees. For literacy and adaptive problem solving, there are no significant differences in proficiency between full-time and part-time workers, on average across participating OECD countries and economies. In numeracy, the gap between full- and part-time workers is 13 points (272 points compared to 259). For all domains, there is a sizeable gap between the employed and the unemployed population, which in turn shows substantially greater average proficiency than the inactive population. For instance, the average numeracy proficiency for the employed population (270 points) is 23 points higher than that of the unemployed population and 36 points higher than that of the inactive population.

Analysing labour-market outcomes across different skill levels provides an alternative perspective on the relationship between skills and outcomes, focusing on comparisons between the highest and lowest skill levels. The largest differences in labour-market outcomes are in labour-force participation (Figure 4.1 Panel A). Across the OECD, 94% of high-skilled adults (numeracy proficiency of Level 4 and above) are in the labour force, compared with only 70% of low-skilled adults (at or below Level 1). The gap is 32 percentage points or greater in Austria, Finland, Germany and Italy, while Japan, Korea and Singapore have the smallest gaps at less than 15 percentage points. Even in Croatia and Korea, where only 87% of high-skilled people are active (the lowest rate among all countries and economies), this is still higher than Singapore's 79% activity rate for low-skilled adults, which is the highest rate for low-skilled workers among all countries.

Among all participating countries, high-skilled individuals are less likely to experience unemployment than low-skilled individuals (Figure 4.1, Panel B). The gap between the two groups is, on average, 5 percentage points (7% compared to 2%). In Spain and Italy, the gap is highest at 10 percentage points or more, while in Denmark, Israel, and Poland the gap is less than 1 percentage point.

High-skilled individuals are more likely to be employed full-time than their lower-skilled counterparts (Figure 4.1, Panel C). On average, 91% of high-skilled workers across OECD countries are employed full-time, compared to 82% of low-skilled workers. Czechia and Poland are exceptions, where a higher share of low-skilled individuals work full-time. The Netherlands has the lowest full-time employment rate overall, largely due to 34% of its low-skilled workers being employed part-time – the highest among participating countries.

Differences in employment status associated with numeracy proficiency persist when accounting for individuals’ social and demographic characteristics. Figure 4.2 shows the estimated change in the probability of being active in the labour market (Panel A) and being employed (Panel B) that is associated with a one-standard-deviation increase in numeracy proficiency or a one-standard-deviation increase in the number of years of education. Results for numeracy proficiency control for years of education, and vice versa, which isolates each factor’s relationship with labour market outcomes while holding the other factor constant.4

On average across participating OECD countries and economies, a one-standard-deviation increase in numeracy proficiency – 58 points on the numeracy scale – is associated with a 4-percentage-point increase in the likelihood of being active in the labour market. A similar increase in years of education results in a 5-percentage-point increase (Figure 4.2, Panel A). In this context, being “active” means that a person is either employed or actively looking for work, while being “inactive” refers to not participating in the labour market, such as those who are retired, full-time students or those not seeking employment for health or personal reasons. For education, the highest coefficients are found in Croatia, Ireland and Israel (8 percentage points or above), while for numeracy, the highest coefficients are found in Czechia, England (United Kingdom), Italy and the United States, at 7 percentage points.

Along the same lines, a one-standard-deviation increase in numeracy proficiency is associated with a 0.9-percentage-point increase in the likelihood of being employed as opposed to being unemployed (Figure 4.2, Panel B). This is similar in magnitude to the increase of 1.1 percentage points for a one-standard-deviation increase in years of education – approximately 3 years of education. The results for both numeracy proficiency and years of education are only statistically significant for some countries and economies. The largest coefficients for years of education are found for Lithuania and Spain at over 3 percentage points, while the largest coefficients for numeracy are in England (United Kingdom), Italy, Spain and Sweden, where the estimated change in probability is 2 percentage points or more.

The magnitude of the effect of numeracy on labour-market participation is greater than the effect on employment. This suggests that the decision to participate in the labour force is more closely linked to skills and education than the likelihood of employment.

The link between skills, education and employment is weaker in 2023 than it was ten years ago, when data from the first cycle of the survey were collected in most of the countries and economies taking part. One possible explanation for this is the tighter conditions prevailing in most labour markets in 2022-23 compared to 2011-12. In a tight labour market, people find work more easily, meaning that even individuals with low skills may succeed in getting hired. Figure 4.3 explores this hypothesis, by plotting the change in the unemployment rate between the two cycles (on the x-axis) against the change in the relationship between numeracy skills and the probability of being employed (on the y-axis). Overall, it shows that a fall in the unemployment rate is accompanied by a corresponding fall in the estimated effect of numeracy skills on the likelihood of being employed (it should be noted, however, that not all estimates for this effect in Cycle 1 or Cycle 2 are statistically significant at the 5% level). Tighter labour markets may thus have tended to bring workers into employment regardless of their skill level, thereby weakening the association between skills and employment.

Educational attainment and information-processing skills are thought to be positively associated with workers’ productivity. While early research into the statistical effects of human capital tended to conflate skills with educational attainment, an important contribution of the Survey of Adult Skills is that it allows them to be examined independently. Studies using this distinction have confirmed the independent effects of information-processing skills on wages (Araki, 2020[30]; Hanushek et al., 2015[31]). Further evidence from the 2023 Survey of Adult Skills suggests that the effects of educational attainment are greater than those of information-processing skills, although both remain positively associated with wages. This is possibly because educational attainment captures a wider range of skills, including social and emotional skills such as perseverance or conscientiousness. A future thematic report will explore this question by analysing data on social and emotional skills collected in the 2023 Survey of Adult Skills (OECD, 2024[25]).

While in the aggregate the returns to qualifications outweigh the returns to skills, there are important caveats. For example, it has been found that as individuals age and progress in their careers, skills become relatively more important and educational qualifications relatively less so. This suggests that employers rely on qualifications as the best indicator of potential productivity in the early years of a worker's career, while in later years the importance of educational qualifications diminishes relative to skills acquired through work experience, enabling individuals with a skill advantage to differentiate themselves (Hanushek et al., 2015[31]). In addition, the economic returns to qualifications have been found to decline in a context of overall educational expansion, whereas skills retain their premium even as the overall level of skills in a population increases. Notably, a higher overall level of skills proficiency in a population is associated with diminishing returns to qualifications (Araki, 2020[30]).

Indeed, results from the first cycle of the Survey of Adult Skills showed that skills proficiency and educational attainment have positive, significant, yet distinct effects on wages (OECD, 2019[27]; 2016[28]; 2013[29]). This section revisits this analysis, focusing on the effects of skills and years of education on wages using data from the second cycle.

Individuals with higher levels of proficiency are much more likely to be high earners (Figure 4.4). The median earner with high levels of skills earns USD 31 per hour on average, 75% more than the median earner with low levels of skills (USD 18 per hour).5 Median earnings for high-skilled workers are highest in Denmark, Switzerland and the United States, while median earnings for low-skilled workers are highest in Denmark, Norway and Switzerland. The largest absolute gaps between high- and low-skilled workers can be observed in Singapore and the United States, where high-skilled workers earn over USD 20 more per hour than low-skilled workers. Relative gaps are highest in Chile, Israel and Singapore; in Chile, the median high-skilled worker earns over three times the hourly wage of the median low-skilled worker (USD 22 per hour compared to USD 7).

Conversely, the smallest absolute gaps in median wages (USD 7 or less) are in Croatia, Poland and the Slovak Republic, which are also the countries with the lowest overall wages for high-skilled workers. Relative wage gaps in these countries are also small – Poland and the Slovak Republic are the only countries where median earnings for high-skilled adults are no more than 40% greater than median earnings for those with low skills.

Income differences become even more pronounced at the upper end of the wage distribution. Among workers with skills at Level 4 or above, high earners – i.e. at the 75th percentile for that skill level – earn USD 43 per hour, on average, compared to USD 23 per hour for the high earners among those at or below Level 1 (see Table A.4.7 (N) in Annex A). The high-skilled high earners thus earn 88% more than their low-skilled counterparts.

Numeracy skills and years of education continue to have statistically significant associations with wages independently of one another, and the magnitude of these effects varies considerably across participating countries and economies (Figure 4.5). The results presented here are adjusted for socio-demographic characteristics. On average, an increase of one standard deviation in an individual’s years of education is associated with a 16% increase in wages. This is broadly in line with existing international evidence (Clark and Abildgaard Nielsen, 2024[38]; Patrinos, 2023[39]). An increase of one standard deviation in numeracy proficiency is associated with a 9% increase in wages.6

It is important to note that skills are associated with workers’ wages even after accounting for years of education. For 27 out of 31 participating countries and economies, the relationship between numeracy and wages is statistically significant, and in France the estimated effect for numeracy proficiency in fact exceeds that of years of education. The greatest differences between the effects of numeracy and years of education are found in Singapore (22 percentage points). Numeracy proficiency has the largest estimated relationship with wages in Chile, England (United Kingdom) and Germany (over 14%), and the smallest (statistically significant) relationship in the Flemish Region (Belgium), Italy and Poland (under 6%). For education, the largest estimated relationships are in Chile and Singapore (30% or more) and the smallest are in France and Japan (10% or less).

This section considers how much of the observed variation in wages is accounted for (or “explained”) by differences in observable characteristics such as education, numeracy and literacy proficiency, work experience and socio-demographic factors. It decomposes the variance in earnings to understand how important different factors are in explaining the observed differences in wages between individuals. Wage decomposition has been used to explain, for example, gender, race or geographical wage differences (Akee, Jones and Porter, 2019[40]; Blau and Kahn, 2017[41]; He and Jiang, 2023[42]).

On average, 23% of wage variation can be accounted for by the observable variables of skills proficiency, educational attainment, field of study, job tenure, and individual characteristics – age, gender, immigrant background, and whether one lives with a partner or has children (Figure 4.6). In Croatia, Finland, Germany, Japan, Latvia and Singapore, they account for over 30% of the variation in wages, compared to less than 15% in Ireland, Israel, Korea, Poland and the Slovak Republic. The fact that observable characteristics explain only 23% of the variation in earnings on average means that 77% of the variation is determined by other factors.

Together, years of education and numeracy and literacy proficiency account for around two-thirds of the total explained variation on average. In Chile and Hungary this share exceeds 80%, whereas for Japan it is less than half. Education consistently accounts for the greatest share of wage variation. It accounts for 9% of the total variation on average and is the greatest single factor in 27 out of 31 participating countries and economies. Proficiency – across both numeracy and literacy domains – accounts for 6% of the variation in wages on average across OECD countries, and at least 10% in England (United Kingdom), France and Germany. In Korea and the Slovak Republic, conversely, skills account for less than 1% of overall wage variation.

Observable individual characteristics do not typically account for a large proportion of variation in wages, at just 4% on average across participating OECD countries and economies. However, the share is 6% in Finland and is at least 8% in Estonia, Japan and Latvia. Japan is also an outlier in the explanatory power of years of job tenure, which accounts for 9% of its wage variation, compared to the OECD average of 2%. For 11 countries the share attributable to job tenure is less than 0.5% (Figure 4.6).

The explanatory power of these observable characteristics differs by age. The total share of wage variation explained increases from 22% on average among 25-34 year-olds to 27% for 45-54 year-olds. This increase is driven in part by changes in the explanatory power of skills, which rises from 5% among 25-34 year-olds to 7% among 45-54 year-olds. This suggests that numeracy and literacy proficiency matter more for the wages of older workers than for younger ones (see Table A.4.9b (L, N) in Annex A). This increase in the salience of skills proficiency is consistent with the concept of “employer learning”, noted above (OECD, 2014[32]). Even where workers change employers, as their careers progress their formal education recedes further into the past, whereas the concrete skills they develop and demonstrate regularly during their work will remain relevant.

There are small differences in the patterns of explained wage variation by gender (Figure 4.7). On average, years of education explain a larger share of variation in wages for women than for men (Panel A), whereas numeracy and literacy proficiency accounts for a greater share of variation for men than for women (Panel B). This is consistent with research that finds that women’s wages are more closely linked to formal education, as they are often concentrated in sectors where qualifications are crucial for advancement (notably the public sector), and they may face greater barriers to being rewarded for skills acquisition alone.

However, these gender differences are small across countries. The OECD averages for men and women differ by 2 percentage points in favour of men for skills proficiency and 4 percentage points in favour of women for years of education. Differences are much more pronounced in some countries. In Canada, Croatia, Finland and Ireland, the gender gap in the share of variation explained by years of education is greater than 8 percentage points, while in Czechia, Germany and Portugal, there is a gap of 4 percentage points and more in the variance explained by numeracy and literacy proficiency.

The average total variation in wages accounted for by numeracy and literacy skills and education increased slightly between the two cycles of the Survey of Adult Skills, from 14% to 15%, although this relatively stable picture conceals considerable changes in many countries (Figure 4.8). Looking at the variation in wages accounted for by skill proficiency reveals an increase of 1.3 percentage points, from 4.5% to 5.8%, between cycles. This increase was most pronounced in Czechia (8 percentage points) and France (7 percentage points). On the other hand, there were marked decreases in Singapore (7 percentage points), the United States (6 percentage points) and Israel (6 percentage points). The change is particularly striking in Singapore, where proficiency in literacy and numeracy accounted for the greatest share of variation in wages of any country in the first cycle.

As this analysis shows, skills and education continue to be rewarded in the labour market, particularly in terms of wages, which have a much clearer relationship with numeracy and literacy skills than with employment. As discussed above, one explanation for the smaller relationship with employment is the comparatively tighter labour markets that prevailed during the second cycle (2022-23) compared to the first cycle (2011-17) of the survey; high demand for labour may have led to high employment rates regardless of skill proficiency. Although much of the variation in wages remains unexplained, information-processing skills and, particularly, education do explain a substantial proportion. It is perhaps unsurprising that education remains more important for wages, as it captures a much wider range of the knowledge, skills and attitudes that are rewarded in the labour market (Durlak et al., 2011[17]; Heckman and Kautz, 2012[18]). However, it is notable that the relative explanatory contribution of skills has increased slightly between cycles of the Survey of Adult Skills. Examining this result may provide insights into the importance that employers attach to skills and qualifications when making decisions about hiring and developing their workforce.

Large cross-country differences remain in the extent to which the labour market rewards skills and qualifications. These reflect differences in labour-market institutions, such as wage-setting mechanisms, but also employer preferences and hiring practices. In the future, policy makers could work with employers and social partners to promote wage-setting practices that reward skills rather than qualifications, thereby encouraging a wider uptake of adult learning and the continuing development of skills during people’s lives (OECD, 2024[8]). More generally, policy makers need to maintain their focus on developing the skills and qualifications of the workforce; this must be accompanied by demand-side measures, such as industry development and innovation policies, to achieve more equitable outcomes across the skills spectrum.

While employability and wages are essential contributors to individual well-being, non-economic factors contribute significantly not just to individual well-being but to social cohesion and civic engagement. Results from the first cycle of the Survey of Adults Skills showed that information-processing skills are positively associated with important aspects of well-being such as health, political efficacy, trust and volunteering, even after accounting for a range of socio-demographic variables (OECD, 2019[27]; 2016[28]; 2013[29]). However, the strength of these associations varies across countries. Generally, adults with low skills and education are least likely to report positive social outcomes, while those with high skills and education are most likely to report positive outcomes.

The first cycle of the Survey of Adult Skills also found that trust is positively associated with educational attainment and literacy proficiency. There is further evidence that education helps to develop the capacities and skills needed to build and maintain trust (Borgonovi and Burns, 2015[44]; Borgonovi and Pokropek, 2022[45]). Research has found a positive association between political efficacy and information-processing skills, with information-processing skills less associated with efficacy in societies with stronger respect for the rule of law and lower perceived corruption (Borgonovi and Pokropek, 2017[46]). Further, individuals with greater educational attainment are more likely to engage in health-promoting behaviour, benefit from better diagnosis and management of illnesses, and have longer life expectancy overall (Kakarmath et al., 2018[47]). These findings are consistent with research using other data sources finding that information-processing skills are associated with important life outcomes, such as self-reported health, teenage pregnancy or criminality (Carneiro, Crawford and Goodman, 2007[48]; Deming, 2009[49]).

This section sheds light on how literacy, numeracy and adaptive problem solving are linked to a range of social outcomes: trust in others, political efficacy, volunteering, self-reported health and general life satisfaction. Life satisfaction is a new addition to the 2023 Survey of Adult Skills (Box 4.2). The interpretation of these results is nuanced and has a normative dimension. The most favourable outcome for a country would be to have a high proportion of adults reporting positive outcomes, and for this to be irrespective of skill or educational background. Such a scenario would suggest that even individuals who are at a disadvantage in terms of skills or qualifications enjoy positive outcomes along dimensions of well-being and civic engagement at rates equal to those who are more advantaged.

Findings from the analysis of social outcomes can inform the development of education, training and workforce policies that not only improve labour-market outcomes, but also promote social cohesion and individual well-being. This broader perspective on the benefits of skills underlines their importance in promoting both personal and societal well-being.

Life satisfaction and health are integral components of individual well-being. According to the 2023 Survey of Adult Skills, they are both correlated to skills, though the relationship is more consistently statistically significant for health than for life satisfaction (Figure 4.10). It is important to note that data on life satisfaction and health are based on self-reports in the survey. For cultural reasons, citizens of one country may be more likely overall to respond positively to such subjective questions, so cross-country comparisons should be made with caution. However, the data presented in this section can still shed light on the significant differences in the proportions of respondents reporting positive outcomes at different levels of ability within societies.

Across the OECD, on average, three-quarters of individuals rank their life satisfaction as high Figure 4.9, Panel A). In Denmark, Finland, the Flemish Region (Belgium), the Netherlands and Switzerland over 85% of respondents report high life satisfaction, compared to below 55% in Japan and Korea. Across all participating countries and economies, the share of adults reporting high levels of life satisfaction is positively correlated with literacy and numeracy proficiency. On average, 84% of individuals with high numeracy skills (Level 4 and above) report high levels of life satisfaction, but only 65% of individuals with low numeracy skills (at or below Level 1) do. These averages mask considerable cross-country differences: while 92% of adults in Finland and Switzerland with high numeracy skills report high life satisfaction, only 63% of adults in Japan with such high numeracy skills do so, and the proportion falls to just 36% for Japanese with low numeracy skills compared to 81% of Finns.

When it comes to self-reported health, 41% of individuals reported having “very good” or “excellent” health on average across participating countries and economies (Figure 4.9, Panel B). Adults in Croatia, Ireland and Israel were most likely to report such positive health (at least 55%), while those in Chile, Japan, Korea and Latvia were least likely (23% or less). The differences between adults with high and low numeracy skills are pronounced: on average, 55% of adults with high numeracy skills report positive health outcomes, compared to only 28% of adults with low numeracy. Again, the OECD average masks significant heterogeneity. At the bottom end of the scale, the share of adults with low levels of numeracy reporting positive health outcomes is only 9% in Latvia and 12% in Estonia and Japan.

This positive relationship between life satisfaction, health and numeracy holds even after accounting for a number of personal characteristics. Figure 4.10 shows the percentage point difference in reporting high life satisfaction (Panel A) and positive health outcomes (Panel B) between adults with low and high numeracy skills. The regression model underlying these estimates controls for age, gender, years of education, immigrant background, parental educational attainment and whether an individual lives with a partner or has children. As the sample includes people who are not employed, no occupational controls are included.

On average, across OECD countries and economies, individuals with high numeracy skills are 11 percentage points more likely to report positive health outcomes – and 10 percentage points more likely to report high life satisfaction – than individuals with low numeracy skills. Nevertheless, some cases stand out: in Croatia, for instance, individuals with high numeracy skills are 24 percentage points more likely to report high life satisfaction, a larger effect than in any other country, while the greatest gap in health occurs in Italy (27 percentage points). In Chile, Finland, Israel and Spain, the relationship between numeracy proficiency and both health and life satisfaction outcomes is not statistically significant at the 5% level, while in 15 of the participating countries and economies there are sizeable and statistically significant associations for both outcomes. Future research into differences in the social policy environment and economic opportunities is important to contextualise differences in the relationship between skill levels and social outcomes.

The 2023 Survey of Adult Skills offers an insight into the prevalence of measures of civic engagement – political efficacy (i.e. an individual’s confidence in their ability to understand and influence political affairs), trust in others and a propensity to volunteer – and how these measures relate to adult skills (see Box 4.2 for a definition of these measures). For most countries, statistically significant associations exist between numeracy proficiency and measures of civic engagement.

On average across OECD countries and economies, 19% of individuals reported high political efficacy (Figure 4.11). Switzerland has the highest share (51%), with Czechia (33%), Finland (31%), Japan (30%) and Sweden (30%) the only other participating countries where the share was over 30%. On the other hand, in France, Italy and Croatia, less than 8% of adults report high political efficacy.

Political efficacy is generally lower among low-skilled adults. On average across participating OECD countries and economies, only 16% of individuals with numeracy proficiency at or below Level 1 report high political efficacy, compared with 26% at Level 4 and above (Figure 4.11, Panel A). However, this relationship varies considerably between countries. In 11 countries and economies, the likelihood of reporting high political efficacy increases or stays constant with each proficiency level, while in 13, the pattern follows a “U” shape, where political efficacy is higher for those at or below Level 1 than for those at Level 2 or Level 3, while adults at Level 4 and above are then most likely to report high political efficacy (see Table A.4.10 (N) in Annex A). Conversely, in Hungary, Poland, Portugal and Spain, the highest-skilled individuals report the lowest levels of political efficacy.

A high degree of trust in others is more prevalent across OECD countries and economies, with 36% of individuals reporting high trust on average. Denmark (71%), Finland (68%) and Norway (66%) have by far the highest levels of trust, whereas Chile (8%), Czechia (18%) and France (19%) are the only countries where less than one-fifth of individuals report high trust.

In all participating countries and economies, individuals with high numeracy skills are more likely than lower-skilled individuals to report high levels of trust (Figure 4.11, Panel B). On average, 25% of individuals at or below Level 1 report high levels of trust, compared to 51% of individuals at Level 4 and above. The greatest difference between high- and low-skilled individuals can be observed in the Netherlands (42 percentage points), Denmark (40 percentage points) and Sweden (40 percentage points).

Finally, 32% of adults reported at least some volunteering activity in the past year on average, with adults in Norway (52%), the United States (48%), Finland (45%), New Zealand (45%) and Denmark (44%) most likely to volunteer. The propensity to volunteer is lowest in Lithuania (15%), Korea (16%), Spain (19%) and Croatia (20%). In all participating OECD countries and economies, high-skilled individuals are more likely to volunteer than the lowest skilled: 41% compared to 22% on average (Figure 4.11, Panel C). The biggest gaps are in Germany and the United States, with differences of 32 percentage points between the two groups.

These relationships hold even when accounting for a number of socio-demographic characteristics (Figure 4.12). Roughly a quarter of participating countries and economies display some statistically significant associations across all three dimensions of civic engagement and skill levels; only Portugal and the Slovak Republic show no statistically significant associations for any of them.

For trust and the propensity to volunteer, there is a clear positive association with skills proficiency. On average across OECD countries and economies, high-skilled individuals are 10 percentage points more likely to volunteer and 18 percentage points more likely to report high trust in others than their low-skilled counterparts. The association with trust is most pronounced in Denmark, Germany and the Netherlands (all between 28 and 30 percentage points), while the propensity to volunteer has the strongest associations in Austria, France and Germany (18 to 20 percentage points). In contrast, for political efficacy, the most notable finding is that several countries show sizeable, statistically significant negative associations with proficiency level. This is the case for Poland (‑9 percentage points), Hungary (-8 percentage points) and Spain (-8 percentage points) (Figure 4.12, Panel A).

The overall picture that emerges from the 2023 Survey of Adult Skills is that, on average, skills proficiency matters significantly for both well-being and civic engagement, but the strength and nature of this relationship varies across countries. This underlines the critical importance of investing in skills development, not simply to improve labour-market outcomes, but also to enhance individual and societal well-being. Countries experiencing significant disparities in the well-being and civic outcomes of adults with different skill levels should increase their efforts to reduce these gaps through initiatives to develop skills. It will also be crucial to address the disproportionate impact of skills on social outcomes such as political efficacy. The combination of many low-skilled adults feeling unable to influence political decisions and lacking the skills to navigate complex digital information landscapes should be a concern for modern democracies. Policy makers must work to empower people of all skill levels to participate in civic life and to manage their health and well-being.

Good matches between workers’ skills and qualifications and those required by their jobs are essential to a well-functioning, productive economy. At the individual level, workers who are not well-matched to their jobs may experience lower job satisfaction and wages. At the aggregate level, mismatches may mean economies experience lower productivity and lose out on valuable human capital investment (Adalet McGowan and Andrews, 2015[9]; Allen, Levels and van der Velden, 2013[50]). Investment in human capital can be wasted when workers are not allocated efficiently to jobs. This is especially the case for workers who are over-qualified or over-skilled for their jobs.

Some natural degree of mismatch in the economy is inevitable as workers sort themselves into suitable jobs. This is particularly true for younger workers or those (re-)entering the labour force. Some people may choose to work in jobs that do not match their skills, qualifications or field of study due to factors such as personal preferences, location or family responsibilities. Nonetheless, policy makers should try to maximise the return on investment in education by promoting an efficient use of skills in the economy and therefore reducing the different types of mismatches that can occur, be they qualification mismatches, skills mismatches or field-of-study mismatches.

Structural changes in the labour market, driven by an ageing population, digitalisation and the green transition, are leading to rapidly changing demand for skills and qualifications (OECD, 2023[51]). If the skills and qualifications of the workforce do not evolve at the same pace, this can lead to mismatches. The 2023 Survey of Adult Skills provides an opportunity to assess how mismatches have evolved over the past decade as these megatrends have gathered pace.

The 2023 Survey of Adult Skills introduces some innovations in the measurement of skills mismatches, including more detailed information on self-reported skills mismatches and more detailed information on fields of study (Box 4.3). This section describes the extent of different types of mismatches – in qualifications, skills and field of study – across countries and economies (see Table 4.2 for the full set of mismatch indicators used in this chapter). It also analyses which socio-demographic groups are most likely to experience mismatches. 

On average across the OECD, about 23% of workers are over-qualified for their current job, while 9% are under-qualified, leaving just over 67% well-matched (Figure 4.13). England (United Kingdom) (37%), Japan (35%) and Israel (34%) have the highest rates of over-qualification, while the Flemish Region (Belgium) (14%), Singapore (14%) and Poland (14%) have the lowest.

The extent of over-skilling across countries and economies is similar to over-qualification. About 26% of workers consider themselves to have higher skills than their job requires, while 10% report that some of their skills are lower than what is required by their job and need to be further developed. This leaves about 64% of workers who report that their skills are well-matched to their jobs on average. Everywhere except Estonia, Finland, Japan and Norway, workers are more likely to report that they are over-skilled than that they are under-skilled.

Although over-skilling is correlated to over-qualification, there are noticeable deviations for some countries and economies (Figure 4.14). Over-skilling is highest in Israel (45%), the United States (39%) and Canada (36%) and lowest in Japan (9%), Finland (18%) and Lithuania (17%). Several factors may explain divergence between over-qualification and over-skilling rates. For example, differences in how respondents from various countries approach self-assessment questions could lead to over- or under-estimation of skills. Differences in labour-market institutions are also likely to contribute to these cross-country disparities. Even among highly educated workers, training programmes might not always be well aligned with the specific skills needed in the workplace, underscoring the importance of targeted on-the-job training and continuous adult learning.

Finally, with regards to field of study, for 38% of workers on average the field of study of their highest qualification does not align with the typical requirements in their job. Korea (49%), Japan (46%) and New Zealand (43%) have the highest field-of-study mismatch rates while Finland (29%), Croatia (31%) and the Netherlands (31%) have the lowest. Mismatches in field of study tend to be more common than over-qualification and skill mismatches. This is less of a concern if workers are choosing jobs that match their preferences and the labour market, or if they are matched to their jobs based on transferable skills. Some mismatches may include adults who have retrained in a field that is better suited to their current job, but at a lower qualification level than their highest qualification. This will be explored further in future reports.

Although it is not possible to compare skills mismatch measures between the two cycles of the Survey of Adult Skills (as explained in Box 4.3), comparisons can be made for mismatches in qualifications and field of study, due to the minimal changes in question wording and measurement. In general, mismatch rates for both have remained relatively stable over the last decade. Qualification mismatches have fallen only slightly, from 34% to 33%. This appears to be mainly due to a drop in under-qualification (from 13% to 9%), while over-qualification increased (from 21% to 23%). Meanwhile, field-of-study mismatches have fallen from 40% of workers in the first cycle to 38% in the current study (OECD, 2019[27]). Despite these small improvements, the magnitude of mismatch on both measures remains high with more work needed across the OECD to improve the matching of workers to their jobs. Moreover, more work is needed to assess whether skills-based approaches are effective in reducing skills mismatches (OECD, 2024[8]).

The 2023 Survey of Adult Skills included a new question on the types of skills that under-skilled workers feel they would need to improve to be well-matched to their jobs (Box 4.3). On average across participating OECD countries and economies, 42% of under-skilled respondents feel they need to improve their computer and software skills (see Table A.4.13 in Annex A). This is significantly more than for any other type of skill – perhaps an indication of the rapid changes to labour markets being brought about by artificial intelligence and digitalisation, and the urgency with which respondents feel they need to keep abreast with these trends. Other skills such as project management and organisational skills (26%), communication and presentation skills (26%), and teamwork or leadership skills (24%) are also cited by respondents as areas they most need to develop (see Table A.4.13 in Annex A).

However, after taking into account the overall share of under-skilled respondents, the share of all workers stating that they lack sufficient computer and software skills is only 4% across OECD countries and economies. Japan (12%), Estonia (12%), Finland (11%) and Norway (10%) have the highest overall proportions of workers who lack the digital skills required for their current jobs (Figure 4.16). This is largely because these countries have higher shares of self-reported under-skilling. Cross-country differences reflect both the supply of skills (i.e. individuals already have strong digital skills) and the demand for such skills in the labour market (i.e. their jobs may not yet require advanced digital skills). Therefore, high levels of self-reported under-skilling may indicate not only a skill gap but also a greater awareness and desire to improve, highlighting the continued need for countries to invest in digital upskilling and training initiatives to keep pace with technological progress.

Some groups of people may be more likely to be over-qualified for their jobs, which can have negative social and economic consequences for them. To understand this better, a regression analysis was carried out to estimate the probability of different groups being over-qualified, accounting for individual and job characteristics. This ensures that comparisons are only made for similarly skilled individuals working in similar occupations. Figure 4.17displays the likelihood of over-qualification by various individual characteristics, while Figure 4.18 does the same for job characteristics. See Table A.4.15 (N) in Annex A for a similar analysis on over-skilling.

The analysis of the relationship between socio-demographic characteristics and over-qualification shows that (Figure 4.17):

• Age: Theory suggests that it may take time for workers to sort themselves into well-matched jobs, with younger workers, who are new in the labour market, more likely to be mismatched (Quintini, 2011[58]). This is especially the case for younger workers who enter the labour market during a downturn, with recessions disrupting the quality of initial firm-worker matches and reducing earnings (Andrews et al., 2020[59]). The empirical analysis confirms this pattern, with older workers (45-65 year-olds) approximately 2 percentage points less likely to be over-qualified than younger workers (25-44 year-olds) on average, though this pattern is not statistically significant in many countries and economies.7

• Gender and partnership status: Workers with family responsibilities – which are still disproportionately borne by women – may choose to take a job for which they are over-qualified if it is better suited to their caring responsibilities (Goldin, 2014[60]). Previous analyses do not show a clear relationship between gender and over-qualification, with some studies finding that women are less likely to be over-qualified than men after accounting for other factors (OECD, 2016[28]; Quintini, 2011[58]). Other studies indicate that women tend to be more over-qualified than men, and in fact over-display their qualifications and education in order to attain leadership positions (Niessen-Ruenzi and Zimmerer, 2023[61]). The 2023 Survey of Adult Skills finds that on average, after accounting for occupation, single women, partnered women and partnered men are less likely to be over-qualified than single men on average (see Table A.4.14 (N) in Annex A). It should be noted that this analysis would not capture any over-qualification due to differential sorting into occupations by gender, as it only compares individuals in the same occupation (see further discussion on occupation below).

• Immigrant background: Foreign-born – or, more specifically, foreign-educated – workers may be more likely to experience mismatch due to lack of recognition of their qualifications and skills in the host labour market or because of language issues (Altorjai, 2013[62]; Larsen, Rogne and Birkelund, 2018[63]; Pivovarova and Powers, 2022[64]). Indeed, the first cycle of the Survey of Adult Skills found that foreign-born workers were more likely to be over-qualified. This finding is repeated in the current cycle, with foreign-born workers being around 5 percentage points more likely to be over-qualified than their native-born counterparts on average. Immigrant workers in Chile, Korea, Latvia, and Spain have particularly strong likelihoods of being over-qualified. This highlights the importance of immigrant integration policies, robust systems for the recognition of prior learning and the need to tackle labour-market discrimination. There are exceptions, however: in one-third of participating countries and economies, native-born adults are more likely to be over-qualified, although differences are often not significant.

• Educational attainment and skills: On average across the OECD, having more years of education is associated with a greater likelihood of over-qualification. This is partly by construction, because workers with more years of education tend to have higher qualification levels and are therefore more likely to be classified as over-qualified if they don’t work in jobs corresponding to their qualification level. Moreover, higher numeracy proficiency reduces the chance of a worker being over-qualified, although not significantly so in most countries.

The analysis of the relationship between job characteristics and over-qualification shows that (Figure 4.18):

• Firm size: The economic literature suggests that larger firms tend to pay higher wages, although this relationship has been declining over the decades (Bloom et al., 2018[65]). This may be because larger firms attract more highly skilled workers – who are in turn better at using their skills and may therefore be less likely to be over-qualified. This is supported by the analysis of the 2023 Survey of Adult Skills, which finds that on average workers in small firms are significantly more likely to be over-qualified for their job than those in large firms (1 000+ employees). The contrast is particularly stark when compared to workers in micro-enterprises (1‑10 employees), who are 10 percentage points more likely to be over-qualified than workers in very large enterprises. The relationship is even stronger – over 20 percentage points – in England (United Kingdom), Japan, Korea, New Zealand and the United States. Larger firms may be better at matching workers to jobs or may have greater capacity to provide training to ensure that workers’ skills match the needs of the job. Larger firms may also have more established internal labour markets through which workers can be moved into teams, projects and tasks that better match their skills and qualifications.

• Employment contract: Mismatch levels may also be related to the type of contract an individual holds – be that permanent or temporary (OECD, 2016[28]). Firms offer, and workers choose, different types of contracts for different reasons. For example, workers who require greater work-life flexibility may prefer fixed-term or temporary contracts. These workers may also be more likely to be over-qualified if firms set lower skill or qualification requirements for such contracts than for permanent or open-ended ones. The 2023 Survey of Adult Skills finds that workers on fixed-term contracts are slightly more likely to be over-qualified than those on permanent contracts on average. This may reflect temporary mismatches, where skilled workers have not yet found permanent work. Among employers, it may indicate a lack of appropriate matching mechanisms, such as difficulties in the recruitment and hiring process, or being unable to adequately screen and test the skills of candidates for temporary roles (OECD, 2024[8]). However, differences are mostly not statistically significant and, in some countries the relationship is in the opposite direction, highlighting the importance of context and local labour-market conditions.

• Part-time work: Workers employed on a part-time basis are 6 percentage points more likely to be over-qualified than full-time workers on average. In Czechia, Denmark, Finland, Norway and the United States, the relationship is even stronger, at 14-16 percentage points. Over-qualified workers may prefer to choose part-time jobs for reasons such as family commitments or caring responsibilities. However, this result highlights the significant aggregate loss of human capital when highly skilled individuals do not work full-time.

• Occupation: Workers employed in elementary occupations are 40 percentage points more likely to be over-qualified than workers employed in skilled occupations on average.8 This pattern holds across all countries and economies. This estimate includes any mismatch that occurs when workers switch occupations – which may be the case for some groups of workers, such as women with children, who have been found to move into different, perhaps less demanding, jobs after the arrival of children (Goldin, Kerr and Olivetti, 2022[66]; Goldin, 2014[60]).

The results for over-skilled workers are similar but the effects are much smaller (see Table A.4.15 (N) in Annex A). Moreover, the results presented in this section are in line with those of the first cycle. Taken together, these findings suggest that some socio-demographic groups are more likely to be mismatched, although the precise reasons are not yet well understood. A combination of sorting, preferences and inefficient matching mechanisms are likely to underpin the results for different groups. Different groups are also more likely to experience mismatches in different countries, highlighting the importance of context, policy and local labour-market conditions.

It is well established that mismatches have individual costs for workers – such as lower job satisfaction and wages – and wider economic costs – such as lower social welfare, lower productivity and lost investment in human capital (Adalet McGowan and Andrews, 2015[9]; Allen, Levels and van der Velden, 2013[50]). Over-qualification and over-skilling imply unused human capital and therefore a loss of potential productivity, as mismatched workers tend to under-utilise their skills. Over-qualification and over-skilling can also reduce job satisfaction and overall well-being when workers feel they are not using their skills and education to the best of their abilities. Under-qualification and under-skilling also have costs, especially when workers feel inadequately prepared to do their jobs and do not receive sufficient training to develop their skills. However, the literature has generally found that over-qualification and over-skilling result in greater economic and social costs than under-qualification and under-skilling (OECD, 2016[28]; Quintini, 2011[58]). This section investigates the economic (i.e. wages) and social (i.e. life satisfaction) costs associated with mismatches.

Analysis from the first cycle of the Survey of Adult Skills found that over-qualified workers earn about 14% less than well-matched workers with the same skills proficiency (OECD, 2016[28]). This effect is more pronounced than the effect of mismatches in skills or field of study. This section revisits the wage penalty analysis for different kinds of mismatches. It uses regression analysis to identify the association between wages and over-qualification, over-skilling and field-of-study mismatches, accounting for a range of relevant individual and job-specific characteristics. Importantly, each regression compares mismatched individuals to their equally skilled and equally educated well-matched counterparts working in the same industry and occupation. Figure 4.19 details the estimated wage penalties of mismatches.

The 2023 Survey of Adult Skills confirms the findings from the first cycle. On average across the OECD, over-qualified workers face a 12% wage penalty compared to equally skilled, well-matched workers in the same occupation and industry (Figure 4.19, Panel A). This suggests that over-qualified workers would benefit financially if they moved to a job that made better use of their highest qualification. The size of the wage penalty associated with over-qualification varies across countries and economies, with the highest penalties in Singapore (20%), Chile (20%) and the United States (19%). The results are significant for all countries and economies except for Italy and Korea.

Over-skilling is associated with a smaller wage penalty than over-qualification in most participating OECD countries and economies and the relationship is not statistically significant everywhere. On average, the over-skilled earn 2% less than their well-matched counterparts in the same occupation and industry (Figure 4.19, Panel B). This is after accounting for information-processing skills, which suggests that it is the mismatch itself and not the level of skills which drive the wage penalty. England (United Kingdom) (10%), Ireland (8%) and Spain (7%) have the highest wage penalties associated with over-skilling.

Unlike over-qualification and over-skilling, in theory, field-of-study mismatches may not necessarily lead to wage penalties where skills are transferable across jobs. If their skills are not transferable, they can translate into lower wages if this results in workers becoming over-qualified or over-skilled (Montt, 2015[54]). The 2023 Survey of Adult Skills finds that on average, field-of-study mismatches attract a 5% wage penalty, with the highest penalties occurring in Israel (11%), the United States (10%) and England (United Kingdom) (10%) (Figure 4.19, Panel C).

Taken together, these results suggest that being over-qualified remains the most important driver of lower wages among mismatched workers. This may reflect differences in underlying wage-setting structures in many OECD countries which typically tie wages to qualifications, rather than acquired skills. Cross-country differences will thus reflect the importance of these wage-setting mechanism in local labour markets as well as other differences such as bargaining models, employer preferences and hiring practices. This again highlights that policy makers should work with employers and social partners to encourage wage-setting practices that financially reward skills rather than qualifications (OECD, 2024[8]).

Mismatches can also have wider social costs for individuals and societies in terms of reduced well-being and life satisfaction. These social costs may themselves lead to poorer labour-market outcomes, in the form of reduced productivity. To complement the above analysis of economic costs, this section uses the new measure of life satisfaction included in the 2023 Survey of Adult Skills to estimate the relationship between mismatches and life satisfaction. It uses regression analysis to examine the relationship between life satisfaction and mismatches in qualifications, skills and field of study, accounting for a range of individual characteristics. Importantly, it includes wages in the regression to accounts for the fact that wages are strongly associated with happiness.

The analysis finds a negative association between over-qualification and life satisfaction, with over-qualified individuals almost 4 percentage points less likely to report being highly satisfied with their life compared to their well-matched peers, on average (Figure 4.20). This negative relationship between over-qualification and life satisfaction is strongest in the United States (12 percentage points), followed by the Flemish Region (Belgium), Ireland and Switzerland (all about 8 percentage points). Under-qualification has a similar negative relationship with life satisfaction, with those who are under-qualified being 2 percentage points less likely to report high life satisfaction. This effect is most pronounced in Italy (9 percentage points) and France (7 percentage points). Under-skilling is associated with a 5-percentage-point reduction in life satisfaction, and over-skilling with a 1 percentage point reduction (see Table A.4.17 in Annex A), although the result for over-skilling is not statistically significant. This suggests that workers’ own perceptions of their skills and how well-matched they are to their jobs is an important factor in their overall happiness, with under-skilling being a more important factor.

Furthermore, after accounting for wages, field-of-study mismatches are not significantly associated with life satisfaction suggesting that workers can still be happy working in a different field if they earn enough money. Taken together, these results suggest there may be some link between mismatch and life satisfaction, although more research is needed to understand the underlying drivers of this relationship.

[34] Abrassart, A. (2013), “Cognitive skills matter: The employment disadvantage of low-educated workers in comparative perspective”, European Sociological Review, Vol. 29/4, pp. 707-719, https://doi.org/10.1093/esr/jcs049.

[9] 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.

[40] Akee, R., M. Jones and S. Porter (2019), “Race matters: Income shares, income inequality, and income mobility for all U.S. races”, Demography, Vol. 56/3, pp. 999-1021, https://doi.org/10.1007/S13524-019-00773-7.

[50] Allen, J., M. Levels and R. van der Velden (2013), “Skill mismatch and skill use in developed countries: Evidence from the PIAAC study”, ROA Research Memorandum, No. 017, Research Centre for Education and the Labour Market, Maastricht University, https://doi.org/10.26481/umaror.2013017.

[62] Altorjai, S. (2013), “Over-qualification of immigrants in the UK”, ISER Working Paper Series, No. 2013-11, Institute for Social and Economic Research, https://www.iser.essex.ac.uk/wp-content/uploads/files/working-papers/iser/2013-11.pdf.

[59] Andrews, D. et al. (2020), “The career effects of labour market conditions at entry”, Treasury Working Paper, Australian Department of the Treasury, https://treasury.gov.au/sites/default/files/2020-06/p2020-85098-202006.pdf.

[30] Araki, S. (2020), “Educational expansion, skills diffusion, and the economic value of credentials and skills”, American Sociological Review, Vol. 85/1, pp. 128-175, https://doi.org/10.1177/0003122419897873.

[19] Arrow, K. (1973), “Higher education as a filter”, Journal of Public Economics, Vol. 2/3, pp. 193-216, https://doi.org/10.1016/0047-2727(73)90013-3.

[13] 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.

[15] Becker, G. (1964), Human Capital: A Theoretical and Empirical Analysis with Special Reference to Education, National Bureau of Economic Research, New York.

[41] Blau, F. and L. Kahn (2017), “The gender wage gap: Extent, trends, & explanations”, Journal of Economic Literature, Vol. 55/3, pp. 789-865, https://doi.org/10.1257/JEL.20160995.

[65] Bloom, N. et al. (2018), “The disappearing large-firm wage premium”, AEA Papers and Proceedings, Vol. 108, pp. 317-322, https://doi.org/10.1257/pandp.20181066.

[44] Borgonovi, F. and T. Burns (2015), “The educational roots of trust”, OECD Education Working Papers, No. 119, OECD Publishing, Paris, https://doi.org/10.1787/5js1kv85dfvd-en.

[45] Borgonovi, F. and A. Pokropek (2022), “The role of birthplace diversity in shaping education gradients in trust: Country and regional level mediation-moderation analyses”, Social Indicators Research, Vol. 164/1, pp. 239-261, https://doi.org/10.1007/S11205-022-02948-Z.

[46] Borgonovi, F. and A. Pokropek (2017), “Mind that gap: The mediating role of intelligence and individuals’ socio-economic status in explaining disparities in external political efficacy in 28 countries”, Intelligence, Vol. 62, pp. 125-137, https://doi.org/10.1016/J.INTELL.2017.03.006.

[48] Carneiro, P., C. Crawford and A. Goodman (2007), “The impact of early cognitive and non-cognitive skills on later outcomes”, CEE Discussion Paper, No. 92, London School of Economics: Centre for the Economics of Education, https://eprints.lse.ac.uk/19375/1/The_Impact_of_Early_Cognitive_and_Non-Cognitive_Skills_on_Later_Outcomes.pdf (accessed on 21 November 2024).

[10] Causa, O. et al. (2022), “The post-COVID-19 rise in labour shortages”, OECD Economics Department Working Papers, No. 1721, OECD Publishing, Paris, https://doi.org/10.1787/e60c2d1c-en.

[38] Clark, G. and C. Abildgaard Nielsen (2024), “Returns to education: A meta-study”, EHES Working Paper, No. 249, European Historical Economics Society, https://www.ehes.org/wp/EHES_249.pdf (accessed on 5 July 2024).

[23] Collins, R. (2019), The Credential Society: An Historical Sociology of Education and Stratification, Columbia University Press, New York, NY, https://doi.org/10.7312/coll19234.

[49] Deming, D. (2009), “Early childhood intervention and life-cycle skill developoment: Evidence from Head Start”, American Economic Journal: Applied Economics, Vol. 1/3, pp. 111-134, https://doi.org/10.1257/app.1.3.111.

[17] Durlak, J. et al. (2011), “The impact of enhancing students’ social and emotional learning: A meta-analysis of school-base universal interventions”, Child Development, Vol. 82/1, pp. 405-432, https://doi.org/10.1111/j.1467-8624.2010.01564.x.

[43] Fields, G. (2003), “Accounting for income inequality and its change: A new method, with application to the distribution of earnings in the United States”, in Worker Well-Being and Public Policy, Emerald Group Publishing, https://doi.org/10.1016/S0147-9121(03)22001-X.

[60] Goldin, C. (2014), “A grand gender convergence: Its last chapter”, American Economic Review, Vol. 104/4, pp. 1091-1119, https://doi.org/10.1257/aer.104.4.1091.

[66] Goldin, C., S. Kerr and C. Olivetti (2022), “When the kids grow up: Women’s employment and earnings across the family cycle”, NBER Working Paper, No. 30323, National Bureau of Economic Research, https://doi.org/10.3386/w30323.

[31] Hanushek, E. et al. (2015), “Returns to skills around the world: Evidence from PIAAC”, European Economic Review, Vol. 73, pp. 103-130, https://doi.org/10.1016/j.euroecorev.2014.10.006.

[12] Hanushek, E. and L. Woessmann (2021), “Education and economic growth”, in Oxford Research Encyclopedia of Economics and Finance, Oxford University Press, https://doi.org/10.1093/acrefore/9780190625979.013.651.

[14] Hanushek, E. and L. Woessmann (2011), “The economics of international differences in educational achievement”, in Hanushek, E., S. Machin and L. Woessmann (eds.), Handbook of the Economics of Education, Elsevier, https://doi.org/10.1016/B978-0-444-53429-3.00002-8.

[18] Heckman, J. and T. Kautz (2012), “Hard evidence on soft skills”, Labour Economics, Vol. 19/4, pp. 451-464, https://doi.org/10.1016/j.labeco.2012.05.014.

[33] Heisig, J., M. Gesthuizen and H. Solga (2019), “Lack of skills or formal qualifications? New evidence on cross-country differences in the labor market disadvantage of less-educated adults”, Social Science Research, Vol. 83, https://doi.org/10.1016/j.ssresearch.2019.06.005.

[42] He, Z. and Y. Jiang (2023), “Decomposing income inequality in the United States: 1968–2018”, Empirical Economics, Vol. 65/6, pp. 2751-2778, https://doi.org/10.1007/S00181-023-02434-6.

[47] Kakarmath, S. et al. (2018), “Association between literacy and self-rated poor health in 33 high- and upper-middle-income countries”, OECD Education Working Papers, Vol. 165, https://doi.org/10.1787/7aaeac27-en.

[63] Larsen, E., A. Rogne and G. Birkelund (2018), “Perfect for the job? Overqualification of immigrants and their descendants in the Norwegian labor market”, Social Inclusion, Vol. 6/3, pp. 78-103, https://doi.org/10.17645/si.v6i3.1451.

[55] Marcolin, L. and G. Quintini (2023), “Measuring skill gaps in firms: the PIAAC Employer Module”, OECD Social, Employment and Migration Working Papers, No. 292, OECD Publishing, Paris, https://doi.org/10.1787/903c19c9-en.

[16] Mincer, J. (1970), “The distribution of labor incomes: A survey with special reference to the human capital approach”, Journal of Economic Literature, Vol. 8/1, pp. 1-26.

[54] Montt, G. (2015), “The causes and consequences of field-of-study mismatch: An analysis using PIAAC”, OECD Social, Employment and Migration Working Papers, Vol. 2015, https://doi.org/10.1787/5jrxm4dhv9r2-en.

[24] Murphy, R. (1988), Social Closure: The Theory of Monopolization and Exclusion, Clarendon, Oxford.

[61] Niessen-Ruenzi, A. and L. Zimmerer (2023), “The Importance of Signaling for Women’s Careers”, European Corporate Governance Institute - Finance Working Paper No. 888/2023, pp. 1-83, https://doi.org/10.2139/ssrn.3987238.

[8] OECD (2024), Skills-first approaches for inclusive and efficient labour markets, OECD Publishing, Paris, https://doi.org/10.1787/4c6346da-en.

[25] OECD (2024), Survey of Adult Skills 2023: Reader’s Companion, OECD Skills Studies, OECD Publishing, Paris, https://doi.org/10.1787/3639d1e2-en.

[57] OECD (2024), Understanding Skill Gaps in Firms: Results of the PIAAC Employer Module, OECD Skills Studies, OECD Publishing, Paris, https://doi.org/10.1787/b388d1da-en.

[1] OECD (2023), OECD Employment Outlook 2023: Artificial Intelligence and the Labour Market, OECD Publishing, Paris, https://doi.org/10.1787/08785bba-en.

[51] OECD (2023), OECD Skills Outlook 2023: Skills for a Resilient Green and Digital Transition, OECD Publishing, Paris, https://doi.org/10.1787/27452f29-en.

[11] OECD (2022), OECD Employment Outlook 2022: Building Back More Inclusive Labour Markets, OECD Publishing, Paris, https://doi.org/10.1787/1bb305a6-en.

[5] OECD (2021), Getting Skills Right: Career Guidance for Adults in a Changing World of Work, OECD Publishing, Paris, https://doi.org/10.1787/9a94bfad-en.

[6] OECD (2021), OECD Skills Outlook 2021: Learning for Life, OECD Publishing, Paris, https://doi.org/10.1787/0ae365b4-en.

[56] OECD (2021), Training in Enterprises: New Evidence from 100 Case Studies, OECD Publishing, Paris, https://doi.org/10.1787/7d63d210-en.

[7] OECD (2019), Getting Skills Right: Future-Ready Adult Learning Systems, OECD Publishing, Paris, https://doi.org/10.1787/9789264311756-en.

[2] OECD (2019), OECD Employment Outlook 2019: The Future of Work, OECD Publishing, Paris, https://doi.org/10.1787/9ee00155-en.

[3] OECD (2019), OECD Skills Strategy 2019: Skills to Shape a Better Future, OECD Publishing, Paris, https://doi.org/10.1787/9789264313835-en.

[27] OECD (2019), Skills Matter: Additional Results from the Survey of Adult Skills, OECD Skills Studies, OECD Publishing, Paris, https://doi.org/10.1787/1f029d8f-en.

[35] OECD (2018), Survey of Adult Skills (PIAAC) database, http://www.oecd.org/skills/piaac/publicdataandanalysis/ (accessed on 23 August 2024).

[28] OECD (2016), Skills Matter: Further Results from the Survey of Adult Skills, OECD Publishing, Paris, https://doi.org/10.1787/9789264258051-en.

[36] OECD (2015), Survey of Adult Skills (PIAAC) database, http://www.oecd.org/skills/piaac/publicdataandanalysis/ (accessed on 23 August 2024).

[32] OECD (2014), OECD Employment Outlook 2014, OECD Publishing, Paris, https://doi.org/10.1787/empl_outlook-2014-en.

[29] OECD (2013), OECD Skills Outlook 2013: First Results from the Survey of Adult Skills, OECD Publishing, Paris, https://doi.org/10.1787/9789264204256-en.

[4] OECD (2012), Better Skills, Better Jobs, Better Lives: A Strategic Approach to Skills Policies, OECD Publishing, Paris, https://doi.org/10.1787/9789264177338-en.

[37] OECD (2012), Survey of Adult Skills (PIAAC) database, http://www.oecd.org/skills/piaac/publicdataandanalysis/ (accessed on 23 August 2024).

[26] OECD (forthcoming), Survey of Adult Skills 2023 Technical Report, OECD Publishing, Paris.

[39] Patrinos, H. (2023), 50 years after landmark study, returns to education remain strong, WorldBank Blog, https://blogs.worldbank.org/en/education/50-years-after-landmark-study-returns-education-remain-strong (accessed on 5 July 2024).

[64] Pivovarova, M. and J. Powers (2022), “Do immigrants experience labor market mismatch? New evidence from the US PIAAC”, Large-scale Assessments in Education, Vol. 10/1, https://doi.org/10.1186/s40536-022-00127-7.

[58] Quintini, G. (2011), “Over-qualified or under-skilled: A review of existing literature”, OECD Social, Employment and Migration Working Papers, Vol. 121, https://doi.org/10.1787/5kg58j9d7b6d-en.

[53] Quintini, G. (2011), “Right for the job: Over-qualified or under-skilled?”, OECD Social, Employment and Migration Working Papers, Vol. 120, https://doi.org/10.1787/5kg59fcz3tkd-en.

[20] Spence, M. (1973), “Job market signaling”, Quarterly Journal of Economics, Vol. 87/3, pp. 355-374, https://doi.org/10.2307/1882010.

[21] Stiglitz, J. (1975), “The theory of ’screening’, education, and the distribution of income”, American Economic Review, Vol. 65/3, pp. 283-300.

[22] Thurow, L. (1975), Generating Inequality: Mechanisms of Distribution in the U.S. Economy, Basic Books, New York.

[52] UNESCO Institute for Statistics (2012), International Standard Classification of Education: ISCED 2011, UNESCO Institute for Statistics, https://uis.unesco.org/sites/default/files/documents/international-standard-classification-of-education-isced-2011-en.pdf.

[67] Wolbers, M. (2003), “Job mismatches and their labour-market effects among school-leavers in Europe”, European Sociological Review, Vol. 19/3, pp. 249-266, https://doi.org/10.1093/esr/19.3.249.