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

Technological change and the need for a green transition are increasing the demand for advanced information-processing skills and other high-level cognitive skills. In parallel, social and demographic developments, including population ageing, educational expansion and migration, continue to alter the supply of skills. Over the past decade, many OECD countries have seen an increase in the share of older adults in their populations – a group that typically has lower skills. This trend is projected to continue. Immigration has increased in most countries, but significant shares of migrants lack sound foundation skills. At the same time, increasing shares of the population are attaining high levels of education.

How well countries manage to upgrade and adapt the skills of their adult population to meet changing demands depends, among other factors, on whether more education translates into better skills, how well the skills of older adults are maintained and developed, and how well people with immigrant backgrounds are supported in overcoming potential cultural and language barriers to develop strong skills. Furthermore, providing everyone with equal opportunities to develop their skills is increasingly important for addressing skills shortages and demands – and even more crucial to ensuring social mobility and fairness in society.

The Survey of Adult Skills tracks the development of key information-processing skills among adults over time by administering comparable assessments in repeated cycles. This chapter describes changes in adults’ literacy and numeracy proficiency in the countries and economies that participated in both the first (between 2012 and 2017) and the second cycle (2023) of the survey. It examines shifts in the skills distribution for the entire adult population as well as for specific socio-demographic groups defined by age, education, immigrant background, gender and parental education. The results provide evidence to help answer important policy questions related to the development of the skill supply and the inclusiveness of skills provision: How has adults’ skills proficiency changed over time and across countries? Has the relationship between education and skills changed with expanding educational attainment? How has proficiency among older age groups evolved? Have the skill levels of immigrants approached those of non-immigrants? Do socio-economic disparities in skills persist?

The results show that the skills proficiency of the adult population has declined in many countries and economies between the two cycles of the Survey of Adult Skills. Such declines have been most pronounced among the most vulnerable groups in society – the lowest-performing adults in a country’s population, adults with below upper secondary education, adults with low-educated parents, and often immigrants or older adults. This is exacerbating inequalities along various dimensions, including socio-economic background, immigrant background and age. Gender gaps in literacy have essentially disappeared, but this was due mainly to the fact that men saw larger declines in their proficiency than women. In almost all countries, women continue to be significantly less proficient than men in numeracy. Increases in educational attainment did not compensate for the skill losses among the adult population.

The Survey of Adult Skills provides comparable measures of proficiency over time in the domains of literacy and numeracy. The instruments and methodology used to assess these skills in the 2023 Survey of Adults Skills are largely similar to those used in the first cycle but have been updated to better reflect the way individuals encounter and use information in contemporary society and to enhance the quality of the assessment. Box 3.2 discusses the implications of these methodological changes for comparing results across cycles. In the domain of problem solving, the results from the second cycle for adaptive problem solving cannot be compared to those from the first cycle for problem solving in technology-rich environments due to differences in the underlying constructs.

Twenty-seven countries and economies participated in both cycles of the survey. While only one round of the second cycle has been conducted so far (in 2022/23), the first cycle was carried out in three rounds: Round 1 in 2011/12, Round 2 in 2014/15 and Round 3 in 2017. As different countries participated in the three rounds, the amount of time that has elapsed between the two data collections is not the same for all countries and economies. The majority of them (21 out of 27) participated in Round 1 of the first cycle, 11 years before the second cycle. For this reason, this chapter often refers to changes that occurred “over the past decade”, for ease of exposition. Five countries participated in Round 2 of Cycle 1, eight years before the second cycle. Hungary participated in Round 3 of Cycle 1, only six years before the second cycle (the United States also participated in Round 3, see Box 3.1). Because of these differences, the size of the change in proficiency between the cycles is not comparable across the participants in the different rounds of the first cycle. All the figures in this chapter, therefore, group countries and economies according to when they participated in the first cycle, and no results are given for the average across OECD countries.

Figure 3.1 shows the change in average literacy and numeracy proficiency between the first and second cycle of the Survey of Adult Skills. Literacy proficiency increased significantly in Finland (by 15 points) and Denmark (by 9 points) over 11 years, and remained stable in 14 countries and economies. Eleven countries experienced a significant decline in literacy proficiency. In four of these, the decrease was particularly large: 31 points in Poland, 28 points in Lithuania, 23 points in Korea and 21 points in New Zealand.

The trends in numeracy proficiency were more favourable. Eight countries and economies recorded significant increases in numeracy proficiency, with the largest gains observed in Finland (17 points), Singapore (17 points) and Estonia (9 points). Proficiency remained stable in 12 countries and economies and significantly decreased in 7. Sizeable falls in numeracy proficiency were observed in Lithuania (22 points) and Poland (21 points).

Overall, literacy and numeracy proficiency have tended to evolve in similar directions to each other across the participating countries and economies. The most notable exception to this pattern is Singapore, where a considerable increase in numeracy proficiency was accompanied by stable proficiency in literacy. Similar trends are seen in Canada, England (United Kingdom), Estonia, the Netherlands and Norway, albeit with relatively smaller changes in numeracy. In contrast, in Austria, Czechia, France and Israel, numeracy remained stable, while literacy declined.

In most countries and economies participating in the Survey of Adult Skills, the demographic composition of the adult population has changed in the years between the cycles. Larger migration flows and ageing populations may have contributed to declining proficiency within countries, as immigrants and older adults often display lower levels of skills. In order to isolate the impact of these demographic changes, data from the 2023 Survey of Adult Skills were reweighted to match the demographic distribution of the population (in terms of age, gender and immigrant background) observed in the first cycle.1 This exercise provides an estimation of the proficiency that would have been observed in 2023, had the composition of the population remained the same as in the previous cycle.

Figure 3.1 shows that accounting for demographic changes does have an impact on the estimated evolution of skills. However, this impact is generally modest, and is not seen in all countries and economies. The analysis suggests that, in Norway and Sweden, average scores in literacy would have increased rather than remained stable, had the demographic profile of the adult population in the second cycle remained the same as in the first cycle. Were it not for demographic changes, literacy scores would have declined by 5.6 points in Austria and by 5 points in France (compared to the observed declines of 11.9 and 6.7 points, respectively), but these changes would not have been statistically significant. In other words, changes in the demographic composition of the population are one factor that accounts for the lack of improvement or the decline in literacy proficiency in these four countries. Similarly, average numeracy proficiency would have improved in Chile, the Flemish Region (Belgium), Germany, Spain and Sweden, were it not for population changes, while it would not have decreased significantly in Korea and the United States.

Literacy and numeracy proficiency in the Survey of Adult Skills is described in terms of six proficiency levels, ranging from Below Level 1 to Level 5 (see Chapter 2). This report denotes as low performers adults that scored at the two lowest levels (at or below Level 1), and as high performers those who scored at Levels 4 or 5. In nine countries, the share of low-performing adults in literacy has increased, while the share of high-performing adults has remained stable or (in the case of Estonia) increased (Figure 3.2). In Denmark, Finland, Germany, the Flemish Region (Belgium) and Norway, the share of adults scoring at or below Level 1 in literacy remained stable, while the share of adults performing at Levels 4 or 5 increased. In all these countries, it can be said that skills polarisation has increased, as more adults now perform at either the lowest or the highest levels.

Conversely, in Korea, Lithuania, New Zealand, Poland and the Slovak Republic, there was a shift towards lower levels of literacy proficiency. In these countries, the increase in the share of low-performing adults coincided with a fall in the share of those scoring at Levels 4 or 5.

A similar pattern of polarisation in proficiency also emerges for numeracy, albeit with greater increases in the share of high-performing adults (Figure 3.3). In 11 of the participating countries and economies, the proportion of adults scoring at Levels 4 and 5 in numeracy has increased, while the proportion of those scoring at or below Level 1 has remained unchanged. In four other countries – Austria, Czechia, Korea and the United States – the share of low-performing adults in numeracy has increased, while the share of high performers has remained stable. In all these countries, relatively fewer adults than before score at intermediate proficiency levels.

In contrast, Finland and Singapore recorded improvements at the lowest and the highest proficiency levels in numeracy, with relatively fewer low-performing adults and relatively more high performers. Hungary, Lithuania, New Zealand, Poland and the Slovak Republic have seen a shift towards lower levels of numeracy proficiency, with relatively more low performers in numeracy than in the previous assessment, and relatively fewer adults scoring at the highest proficiency levels.

Changes in average skills proficiency do not necessarily correspond to changes at different points of the proficiency distribution within the country. Figure 3.4 illustrates changes in proficiency scores between the two cycles at different points of the literacy and numeracy proficiency distributions – the 10th, 25th, 75th and 90th percentiles. When adults are ranked according to their proficiency score, the 10th percentile marks the point below which the lowest-performing 10% of adults score, and the 25th percentile is the point below which the lowest-performing 25% of adults score. Changes in the scores at these percentiles, therefore, indicate how proficiency has changed among the lowest-performing adults in the country between cycles. Likewise, the 75th and the 90th percentile are the points above which the highest-performing 25% and 10% of adults score (or, conversely, below which 75% and 90% of adults fall). Changes in the scores at these percentiles reflect changes in proficiency among adults with the highest skills in a country or economy.

In 16 countries and economies, literacy proficiency declined significantly at the bottom end of the distribution, that is at the 10th and often the 25th percentile (Panel A, Figure 3.4). In 13 of these countries and economies, the proficiency distribution widened – either because proficiency at the upper ends increased (Estonia, the Flemish Region [Belgium] and Germany), or because it declined by less than at the bottom of the distribution (Hungary, Korea and New Zealand), or because it remained unchanged (Austria, Czechia, France, Italy, Israel, Japan and the United States). In Lithuania, Poland and the Slovak Republic, the decline in literacy proficiency was similar at both the bottom and the top of the distribution.

In contrast, in Chile, Canada, England (United Kingdom), Ireland, the Netherlands, Spain and Sweden, the distribution did not change significantly over time. In Finland and Denmark, proficiency increased at the 25th, 75th and 90th percentiles, while remaining unchanged at the 10th percentile. In Norway, proficiency increased significantly at the 75th and 90th percentile, with no change at the lower points of the distribution. Singapore saw a decline in proficiency at the 25th percentile.

The shifts in the numeracy distribution over time follow a similar pattern, in that proficiency at the bottom of the distribution often changed less favourably than proficiency at the top (Panel B, Figure 3.4). In Austria, Czechia, Hungary, Italy, Korea, New Zealand and the United States, numeracy proficiency at the 10th and/or the 25th percentiles significantly decreased, while proficiency at the 75th and 90th percentiles changed less or remained stable. Declines in average numeracy proficiency in these countries are thus largely due to declining proficiency among the lowest-performing adults. In comparison, numeracy proficiency did not change at the bottom of the distribution but increased at the top in Estonia, the Flemish Region (Belgium), France, Germany, Ireland, Japan and the Netherlands. For Estonia and the Netherlands, this means that improvements among the highest-performing adults in the country were the main drivers of the overall increase in numeracy proficiency.

However, five countries and economies (Denmark, England [United Kingdom], Finland, Norway and Singapore) recorded improvements in numeracy at both the lower and the higher ends of the distribution, while Canada saw an increase at the 25th percentile. In contrast, Lithuania, Poland and the Slovak Republic saw sizeable declines in proficiency at all four points. Israel recorded a decrease at the 75th percentile. In Chile, Spain and Sweden numeracy, like literacy, remained unchanged across the distribution.

These findings reveal important insights that are not evident when looking solely at average proficiency scores. First, the declines in average literacy proficiency noted above are largely the result of declines in proficiency among adults at the bottom of the national proficiency distribution. Second, the different trajectories at the bottom and top of the proficiency distribution have widened the gap between the lowest- and the highest-scoring adults in society in both assessment domains, even in countries where average scores have remained constant. In 17 participating countries and economies, the proficiency gap in literacy proficiency between the top- and the bottom-performing 10% of adults, known as the inter-decile range, has widened significantly (see Table A.3.3 (L) in Annex A). In numeracy, the inter-decile range widened in 13 countries and economies (see Table A.3.3 (N) in Annex A). The majority of participating countries and economies thus saw increases in skills-related inequalities between the two cycles of the Survey of Adult Skills.

The Survey of Adult Skills was designed to provide reliable comparisons with the results of the International Adult Literacy Survey (IALS), administered in 21 countries between 1994 and 1998, and the Adult Literacy and Life Skills Survey (ALL), conducted in 13 countries between 2003 and 2007. There are 17 countries and economies which participated in both the Survey of Adult Skills and IALS, with results between 13 and 18 years apart, depending on the country. Comparisons between ALL and the Survey of Adult Skills are available for seven countries and are presented in Table A.3.23 (N) in Annex A. A brief overview of the relationship between the three surveys is provided in Box 3.3.

Figure 3.5 presents how average literacy proficiency has evolved between IALS and the first and second cycles of the Survey of Adult Skills. It groups countries and economies according to the pattern of their long-term trends. Of all the participants, only Finland exhibits a pattern of overall improvement, with mean proficiency in 2023 significantly exceeding that found by IALS by 15 points. In Denmark, Germany, Norway and Sweden, proficiency improved or remained stable over the two cycles of the Survey of Adult Skills, following a drop between IALS and the first cycle of the survey. Conversely, in Czechia and the United States, average proficiency remained stable between IALS and the first cycle of the Survey of Adult Skills but has since decreased. In Hungary, Poland and New Zealand, sizeable increases in literacy proficiency were followed by substantial declines. In Italy and England (United Kingdom), a moderate increase between IALS and the first cycle of the Survey of Adult Skills was followed by no significant change in the next cycle. Finally, in Canada, Chile, the Flemish Region (Belgium), Ireland and the Netherlands, proficiency remained unchanged throughout the entire period.

The data presented in Figure 3.5 span up to three decades and cover adults born between the 1930s and 2000s who started their schooling between the 1940s and early 2010s. They thus reflect influences that go as far back as the Second World War. Against the background of the significant social, political and economic changes of this period, the development of adults’ literacy proficiency appears relatively stable. The differences in average scores between IALS and the Survey of Adult Skills were below 20 points in all the participating countries and economies except Poland and Sweden.

As discussed in the previous section, relative increases in the size of groups with generally lower skill levels, such as immigrants or older adults, can contribute to declines in skills proficiency. On the other hand, an increase in the share of highly educated adults can contribute to improvements in the average skills levels of the population. Along with these compositional changes in the adult population, country-level changes in proficiency depend on how skills proficiency within socio-demographic groups evolves. For example, the impact of a growing share of older adults in the population on overall skills levels depends on how well the skills of older adults are maintained and developed. Similarly, the extent to which educational expansion improves the proficiency of the population depends on whether the highly educated maintain the same high level of skills.

This section focuses on how the relative size and the skills proficiency of socio-demographic groups defined by immigrant background, age and educational attainment have changed across the cycles of the Survey of Adult Skills, and how much these changes have contributed to the proficiency changes in the overall population. The focus is on the literacy domain, while results in numeracy generally follow a similar pattern. The latter results are only briefly outlined and presented in detail in Annex A.

When interpreting the results of this analysis, it is important to keep in mind that the changes in the socio-demographic composition of the adult population are estimated using data collected in the two cycles of the Survey of Adult Skills. The surveys were designed to be representative of the adult population, but differences in the propensity of adults to participate in the surveys can result in adults with certain characteristics being over- or under-represented in the data. While the data are weighted to bring the composition of the sample more in line with the known distribution of characteristics in the population, some discrepancies with other data sources may remain. These discrepancies can sometimes be explained by differences in coverage, timing or definitions between the Survey of Adult Skills and other

sources (see Reader’s Guide). Another reason for differences with other surveys relates to the exclusion of doorstep respondents in the second cycle of the survey and literacy-related non-respondents from the first cycle, which mechanically excludes adults with insufficient language proficiency. A more detailed discussion of these issues is presented in OECD (2024[9]) and OECD (forthcoming[7]).

Over the past few decades, the education systems of OECD countries have been expanding. Completion of upper secondary education has become commonplace, and an increasing number of individuals are attaining tertiary education (OECD, 2023[13]). The extent to which this development improves the key information-processing skills of the population depends on how effectively education translates into strong skills.

The relationship between education and skills is complex. The formation of skills through formal education takes place in a variety of institutions – from early childhood education and care programmes, through schools and universities, to providers of non-formal education and training for adults. How effectively these institutions contribute to the development of relevant skills depends on the quality of teaching and learning, which is a function of multiple factors – including teachers’ training, the curriculum and the learning environment, to name a few (OECD, 2018[14]). As educational systems expand, maintaining high-quality teaching and learning for a growing number of students spending more years in education becomes a major challenge for countries (OECD, 2020[15]).

However, care should be taken not to attribute all changes in skills proficiency related to educational attainment to changes in the quality of education. Individuals select and are selected for advanced education levels based on their abilities and skills. This means that those who already have strong skills are more likely to attain higher levels of education. As access to higher education levels expands, students at these levels become more diverse in terms of academic ability and skills proficiency. Simultaneously, the average ability of adults with lower levels of educational attainment is likely to decline as the relative size of this group shrinks, other things being equal. Consequently, a change in skills proficiency among adults with the same level of education over time may reflect changes in the composition of students rather than changes in education quality.

In addition, skills formation also happens outside the education system, for example in the family and the workplace or through more or less structured adult training and learning opportunities. Adults with different educational attainment may have systematically different levels of access to such opportunities. This is another reason why not all changes in skills should be attributed to changes in formal education.

This section presents changes in literacy proficiency between cycles among adults with below upper secondary, upper secondary and tertiary education as their highest educational qualification. The analysis is restricted to adults over the age of 25, as they are most likely to have completed their formal education.

During the period between the two cycles of the Survey of Adult Skills, the education level of the population aged 25-65 has changed in all participating countries and economies (Figure 3.6).2 In all countries except Israel, New Zealand and Poland, the proportion of adults with tertiary education has increased. This is due to two factors: older cohorts with less education are no longer part of the target population and enrolment rates among younger cohorts have increased. The increase in the share of tertiary-educated adults between 2012 and 2023 was most pronounced in the Flemish Region (Belgium), Ireland and Korea, where it rose by more than 14 percentage points. Meanwhile, the share of adults with below upper secondary education has declined in almost all countries and economies.

Across the countries and economies participating in both cycles of the Survey of Adult Skills, declines in literacy proficiency were most widespread among low-educated adults (Figure 3.7). The average proficiency of adults with below upper secondary education significantly decreased between the two cycles in all countries and economies except in Canada, Denmark, Finland, Israel, Spain and Sweden. These decreases ranged from 12 points in the Flemish Region (Belgium) to 46 points in Lithuania. The proficiency of adults with upper secondary education fell in 19 countries and economies, remained unchanged in 7, and increased in Finland. Among tertiary-educated adults, average proficiency fell in 13 countries and economies, and only increased in Finland.

Changes in numeracy followed a similar pattern, with declines in numeracy proficiency being more common among low-educated adults (see Table A.3.11 (N) in Annex A). Thirteen countries and economies showed a significant decrease in the average numeracy score of this group, while only Singapore recorded an increase. The average numeracy proficiency of tertiary-educated adults increased only in Canada, Estonia and Finland, and declined in Hungary, Israel, Italy, Korea, Lithuania, New Zealand, Poland and the Slovak Republic. Particularly large declines in the numeracy proficiency (over 20 points) of tertiary-educated adults were found in Lithuania, Poland and the Slovak Republic.

The widespread decline in proficiency among low-educated adults has widened disparities in proficiency by educational background in most of the participating countries and economies. Austria, France, Japan, New Zealand and Norway saw large increases in the proficiency gap between adults with below upper secondary and tertiary education (Figure 3.8) In contrast, in Singapore, which has the largest inequality in literacy by educational attainment, the gap changed less because proficiency declined among both low-educated and highly educated adults. The gap narrowed in Poland due to declining literacy proficiency among tertiary-educated adults (Figure 3.7).

The numeracy proficiency gap between low- and highly educated adults widened significantly in 11 countries and economies, with large increases (over 20 points) in Austria, Japan and New Zealand. In Singapore, the gap in numeracy narrowed due to improvements among the low-educated, while in Israel and the Slovak Republic it narrowed because of proficiency declines among tertiary-educated adults (see Tables A.3.11 (N) and A.3.12 (N) in Annex A).

When interpreting proficiency trends for different education groups, it is important to consider that the proficiency of younger and older adults with the same educational attainment may have evolved differently due to their different educational experiences. Older adults may have received a very different education from that received by younger people today. The relationship between older adults’ formal education and their proficiency is also presumably weaker due to the length of time they have been out of education. In contrast, trends in the proficiency of younger adults with different attainment levels are more influenced by the state of education and training in recent years. These trends are also of greater interest to governments, as they devote considerable public resources to education and are relying on the skills of the younger population to meet future skills demands in the labour market.

Figure 3.9 shows that changes in literacy among 25-44 year-olds are similar to those among 45-65 year-olds across the different levels of educational attainment. Exceptions to this pattern are observed among those with lower educational attainment. In particular, in Austria, England (United Kingdom), Estonia, Finland, the Flemish Region (Belgium), France, Germany, Japan, Korea and Norway, declining proficiency among adults with below upper secondary education are primarily seen among 45-65 year-olds, while proficiency among low-educated 25-44 year-olds either improved or changed less since the previous cycle. This suggests that, in these countries and economies, factors other than recent changes in education policy are driving declines in proficiency among low-educated adults.

Numeracy proficiency developed largely similarly among adults in different age groups but with the same level of educational attainment (see Table A.3.11 (N) in Annex A). As with literacy, age differences are mainly observed among low-educated adults. In particular, Austria, England (United Kingdom), Estonia, Finland and France recorded declines in numeracy proficiency among low-educated adults aged 45-65, but not among younger adults with the same educational attainment, after accounting for changes in the population composition by age, gender and immigrant background. In contrast, in Italy, Lithuania, New Zealand and Poland, declines in numeracy proficiency among the low-educated were more pronounced among younger adults, after accounting for demographic changes. In Israel, Singapore and Spain, numeracy proficiency did not change significantly among low-educated young adults and improved among low-educated older adults.

The populations and workforce of OECD countries are ageing, as a result of declining fertility rates, increasing longevity and labour-market policies that delay retirement and encourage longer careers (André, Gal and Schief, 2024[16]). To understand the capacity of an ageing society and workforce to adapt to the demands of the modern, skills-driven economy, means understanding how age relates to skills. The Survey of Adult Skills provides important insights into this relationship.

The 2023 Survey of Adult Skills makes it easier to disentangle some of the mechanisms that link age to skills proficiency. The addition of a second cycle allows the proficiency of the same birth cohort to be compared at two points in time, in other words at different ages. This sheds light on the so-called ageing effect on skills: how skills change as a consequence of growing older. This may be due to the gradual maturation and subsequent decline of cognitive abilities that result from biological ageing. It also depends on the opportunities available to individuals to develop and maintain their key information-processing skills. Since the Survey of Adult Skills does not track the same individuals over time, the focus is on how birth cohorts, rather than individual adults, have gained, maintained or lost skills, on average, over time.

The second cycle also enables different birth cohorts of the same age to be compared. This provides insights into the so-called cohort effect: differences in skills proficiency that are due to the specific experiences of different generations. Cohort effects provide information about how skills change as a consequence of important shifts in society, such as educational expansion, changes in education quality, migration flows, or changes in employment, occupations and work. Cohort effects are likely to vary widely across countries and economies due to differences in these factors, especially due to the different paces at which education expands.

Over recent decades, all OECD countries have seen a steady increase in the share of older adults in the population. This trend is projected to continue, with the proportion of people aged 65 and over expected to increase from 18% in 2023 to 25% by 2050, and that of people aged 50 and over from 37% to 44% (OECD, 2024[17]). In the years between the two cycles of the Survey of Adult Skills, most participating countries and economies have experienced a change in the age structure of their population (Figure 3.10). This shift was most pronounced in Austria, Germany, Italy and Korea, where the share of 55-65 year-olds increased by 5 percentage points or more between 2012 and 2023. Among countries taking part in Round 2 of the first cycle, this share has grown by 5 percentage points in Lithuania and 4 percentage points in Singapore between 2015 and 2023.

Figure 3.11 presents differences in literacy proficiency between adults of the same age at the time of the first and second cycle (i.e. cohort effects). It shows how proficiency has changed between assessments within five different age groups across the participating countries and economies. The evolution of proficiency in the youngest age group will reflect, to a larger extent, current and recent changes in education, such as expanding participation or improving education quality due to recent reforms. In contrast, changes in proficiency among the older age groups may be the result of influences that occurred long before the period covered by the assessment – reaching as far back as the 1950s and 1960s, when the cohort aged 55-65 in the first Survey of Adult Skills began compulsory schooling. In addition, differences among older adults will also reflect more recent changes in the world of work that would have provided or limited their opportunities to maintain and develop skills.

The average literacy proficiency of young adults aged 16-24 improved significantly in England (United Kingdom), Finland and Norway between the two cycles. In contrast, in eight countries, 16-24 year-olds demonstrated lower proficiency in the second cycle than their peers in the first cycle, with particularly pronounced declines in Lithuania, New Zealand, Poland and the Slovak Republic. Among 25-34 year-olds, three countries saw improvements in proficiency and eight countries significant decreases. The trend for older adults was more widely negative: in 12 countries and economies, 55-65 year-olds had lower literacy proficiency in the second cycle than the same age group in the previous cycle. Only Denmark, Finland, Spain and Sweden have seen improved proficiency among older adults.

In numeracy, improvements among young adults aged 16-24 were recorded in Chile, Denmark, England (United Kingdom), Estonia, Finland, Ireland, Japan, the Netherlands, Norway and Singapore (see Table A.3.7 (N) in Annex A). In most of these countries, with the exception of Ireland and Japan, as well as in Canada, numeracy skills also improved among 25-34 year-olds. Numeracy proficiency among adults aged 55-65 increased in Finland, the Flemish Region (Belgium), Germany, the Netherlands, Norway, Singapore, Spain and Sweden. In contrast, Hungary, Lithuania, New Zealand, Poland and the Slovak Republic saw declines in numeracy proficiency among all age groups (except among 55-65 year-olds in New Zealand). In Israel, Italy, Korea and the United States, proficiency declined among 25-34 year-olds. The United States recorded declines among 55-65 year-olds, and Korea among 35-44 year-olds.

Ageing effects offer additional insights into how skills proficiency relates to age. Figure 3.12 compares the average literacy proficiency of cohorts in the second cycle to their proficiency in the previous cycle, when they were between 5 and 11 years younger, depending on the assessment round. For example, adults aged 55-65 in the second cycle are compared to (different) adults who were aged 44-54 in Round 1 of the first cycle, to show how the literacy proficiency of adults born in 1958-68 has evolved on average between these ages. The analysis excludes migrants who have lived in the country/economy for less than ten years in order to isolate the impact of migration on the composition of cohorts.

The figure shows that age-related skills declines are widespread. In almost all countries, the literacy proficiency of the older cohorts has decreased since the last assessment. The exceptions are Denmark, Germany and Sweden, where the average literacy proficiency of adults who were aged 55-65 in the second cycle did not change significantly with age. In the remaining countries and economies, the magnitude of the age effect varies widely. For example, in the Flemish Region (Belgium), the mean literacy proficiency of the oldest cohort in the second cycle only declined by 11 points between the ages 44-54 and 55-65, compared to declines of around 40 points in Korea and Poland. These cross-country differences suggest that investments in skills have a role to play in delaying the decline in skills in older age.

Countries and economies also differ in the onset of age-related skills loss. In six countries and economies, age-related skills loss seems to begin at a relatively young age: Hungary, Korea, Lithuania, Poland, Singapore and the Slovak Republic all recorded age-related declines in literacy among those aged 27-34 in the second cycle. In 12 other countries/economies, age-related declines were first observable among those aged 35-44 in the second cycle. The remaining countries – Canada, Denmark, England (United Kingdom), Estonia, Finland, the Flemish Region (Belgium), Germany, Norway and Sweden – saw clear skill gains across cycles for the cohort aged 16-23 in 2012 and 27-34 in 2023 while declines in literacy were only apparent among those in the older two cohorts.

In comparison, age-related declines of numeracy skills were less pronounced (see Table A.3.10 (N) in Annex A). Two-thirds of the participating countries and economies recorded declines in numeracy proficiency among the cohort aged 55-65 in the second cycle. These declines varied from 9 points in Canada to 32 points in Poland. Only Hungary, Lithuania, Poland and the Slovak Republic recorded age-related declines in numeracy among the cohort aged 27-34 in 2023. These countries, together with Austria, Czechia, Israel, Italy, Korea and New Zealand, saw declines in numeracy proficiency between the cycles among adults aged 35-44 in 2023.

Figure 3.13 shows how the proficiency gap between younger and older adults has changed between the cycles across participating countries and economies. Three scenarios can be distinguished according to whether changes in the gap result from an overall improvement in proficiency among both age groups, with one group improving more than the other; an overall decline, with one group declining more; or a divergence between the groups, with the proficiency of one group increasing and that of the other decreasing.

The first scenario, of net improvement in proficiency, is found in Spain and Sweden, where the literacy proficiency of 55-65 year-olds improved and that of 25-34 year-olds did not change significantly between cycles (Figure 3.11). As a consequence, the difference in proficiency between the age groups has narrowed. This change suggests adults have more opportunities to develop and maintain skills, even though these opportunities do not seem to benefit younger and older adults in the same way.

The second scenario, of overall proficiency decline, is found in several other countries. In Austria, Chile, Czechia and France, the age-related gap in literacy widened as proficiency declined among older adults but remained stable among younger adults (Figure 3.11). In Korea and Lithuania, the proficiency gap widened as a result of larger proficiency declines among older adults than among younger adults. In contrast, the Slovak Republic saw narrowing age-related proficiency gaps because of relatively larger declines among the younger age group. All these changes suggest less favourable conditions for developing skills, either for younger or older people in society or for both.

The third pattern, of divergence, is also observed in fewer countries. Notably, Estonia experienced improvements in proficiency among younger adults alongside declines among older adults that led to a sizable increase in the age-related skills gap (Figure 3.11). A similar pattern, albeit with non-significant changes in proficiency among the age groups, is found in England (United Kingdom).

In the numeracy domain, proficiency improved unevenly among younger and older adults in eight countries/economies (see Tables A.3.7 (N) and A.3.8 (N) in Annex A). In particular, Canada, Chile, Denmark, Estonia and England (United Kingdom) saw improvements among 25-34 year-olds and no change among 55-65 year-olds which widened age-related differences in numeracy proficiency. Singapore, Spain and Sweden saw relatively stronger increases in numeracy proficiency among the older age group, narrowing the gap. Israel, Italy and the Slovak Republic also saw a narrowing of the numeracy gap, as numeracy declined among older adults and remained stable or declined less among younger adults. In Austria, the gap widened due to a decline in the older age group. France exhibited a pattern of divergence, with (non-significant) increases in numeracy proficiency among younger adults and (non-significant) decreases among older adults leading to a widening of the age-related gap in numeracy.

Since the first Survey of Adult Skills, most OECD countries have seen unprecedented numbers of migrants arriving in their territories, many of whom are asylum seekers (OECD, 2024[18]).3 This has posed challenges to many countries, while also presenting opportunities. Migrants increase and diversify the supply of labour, which can contribute to economic growth and help mitigate the negative impacts of population ageing (OECD, 2016[19]). A precondition for this is that they have the skills required to succeed in daily life and work in the host country (OECD, 2019[20]). While some migrants will already possess these skills, others need to overcome substantial language and cultural barriers to acquire them.

How immigrants’ skills develop depends on two main factors: the profile of migrants coming in and the institutions and policies that support their integration. Countries often admit migrants with specific qualifications to address labour shortages or support innovative sectors. In contrast, during the refugee crises of the 2010s, many countries, particularly in Europe, admitted large numbers of low-skilled migrants who were unfamiliar with the host country’s language. This can alter the average skills proficiency of the immigrant population and – depending on the size of the group – the adult population as a whole. Once migrants have settled in the host country, there are usually mechanisms in place to support their social and labour-market integration, such as language training and assistance in searching for and applying for work. These policies also impact migrants’ skills levels. Their effectiveness may vary across countries and economies due to differences in quality but also differences in the linguistic proximity between migrants’ countries of origin and destination.

The Survey of Adult Skills provides important insights into changes in immigrants’ skills. It shows how the share and profile of migrants change over time and how skills proficiency evolves within different migrant groups and between migrants and native-born adults. However, these analyses are necessarily limited to respondents with sufficient knowledge of the language of the assessment. Only such respondents were surveyed in the first cycle of the survey. Although the doorstep interview provides better coverage of immigrants with poor knowledge of the national language(s) in the second cycle (see Box 3.2), these cases were excluded from the analysis of proficiency changes to maximise the comparability of samples across cycles. This means the results presented here are not representative of the entire immigrant population and should be interpreted accordingly.

The changes in the relative size of the immigrant population observed in the Survey of Adult Skills largely mirror those observed from other sources (OECD, 2024[18]). Most countries and economies that participated in both cycles of the survey saw an increase in the share of their foreign-born populations over the last decade (Figure 3.14). According to the data collected in the Survey of Adult Skill, between 2012 and 2023 the proportion of foreign-born adults grew by more than 5 percentage points in Austria, Canada, England (United Kingdom), Germany, Ireland, Italy, Norway, New Zealand, Spain and Sweden. In contrast, this share fell in Estonia and Israel.4 In interpreting these results, it is important to keep in mind that adults with insufficient proficiency of the language in which the survey was administered (literacy-related non-respondents in the first cycle and doorstep interview respondents in the second cycle) are excluded from the analysis, and these are often foreign-born adults.5

The composition of the migrant population represented by the Survey of Adult Skills has also changed over time. Most countries saw improvements in the educational level of foreign-born adults, which probably reflects common trends of educational expansion in both host countries and countries of origin (see Table B.3.12 (Trend) in Annex B). The strongest educational gains are observed in Estonia and the Netherlands. In contrast, the share of low-educated immigrants increased in the Flemish Region (Belgium) and remained stable in Germany, Israel, Italy and New Zealand. Some countries and economies have also seen the linguistic diversity of the foreign-born population covered by the survey increase since the first cycle. For example, Austria, Estonia, the Flemish Region (Belgium), Israel and the Netherlands saw an increase in the share of foreign-born adults who do not speak the language of assessment at home (see Table B.3.11 (Trend) in Annex B). This may have implications for migrants’ literacy and numeracy proficiency, as the assessments are conducted in the countries’ official national or regional language(s).

Figure 3.15 shows the evolution of literacy proficiency for two groups: foreign-born adults of foreign-born parents and native-born adults of native-born parents. Larger fluctuations in literacy proficiency are observed among the foreign-born population, with statistically significant declines in proficiency recorded in 11 countries. These trends are likely to reflect changes in the socio-demographic and linguistic composition of the foreign-born population, driven by recent migration waves.

Overall, the patterns of change for the native-born population largely mirror those of the entire adult population, but there are exceptions in some countries. In the Flemish Region (Belgium), Germany and Norway, the literacy proficiency of adults without an immigrant background has improved since the first cycle, and it has remained stable in France. Meanwhile, the proficiency of adults with an immigrant background developed less favourably in these countries and economies.

Other countries have also experienced less favourable trends for immigrants than for native-born adults. In Chile and Singapore, literacy proficiency declined among immigrants, but not among native-born adults. In Austria and New Zealand, the decline in literacy proficiency for foreign-born adults was more pronounced than for native-born adults.

Migrants’ numeracy skills have developed more positively in many countries over the past decade. Canada, Denmark, England (United Kingdom), Estonia, Finland, Norway and Sweden saw improvements in the proficiency of foreign-born adults (see Table A.3.14 (N) in Annex A). In Finland, Norway and Sweden, these gains were particularly large, exceeding 20 points. In contrast, numeracy proficiency among immigrants only declined significantly between cycles in Hungary and New Zealand.

When interpreting changes in the literacy proficiency of immigrants it is important to consider how long they have been living in the host country. Learning a new language and integrating into a new society takes time, while policies supporting immigrants’ integration also take time to contribute effectively to skills formation. Recent migrants will have had less of both time and such resources. Their skills proficiency is typically lower and more strongly influenced by the opportunities available in their country of origin as well as their linguistic distance to the host country. Changes in proficiency among recently arrived migrants must therefore be interpreted in the context of the changing profile of incoming migrants, which is shaped by each country’s immigration patterns and policies. The skills profile of long-term migrants may also vary across time due to changing patterns of immigration, including return migration in the country of origin. In any event, the skills of long-term migrants are more strongly influenced by conditions in the host country. Changes in their skills over time can therefore offer insights into the effectiveness of the host country’s policies in fostering integration.

Some countries have seen literacy proficiency for foreign-born adults evolve differently among those who have lived in the host country for less than ten years and those who have lived there longer (Figure 3.16). In Germany, the skills of recent migrants in 2023 are significantly lower than those of recent migrants in the previous cycle and no changes are observed among long-term migrants. While the average score in 2023 of recent migrants to Germany (203 points) was among the lowest across participating countries and economies, in the first cycle the average score for this subgroup (248 points) was close to the OECD average (240 points; see Table A.3.14 (L) in Annex A). In contrast, in Austria and New Zealand, literacy proficiency declined among both recent and long-term migrants. Estonia, France, Ireland and Singapore saw significant declines among long-term migrants.

Declining literacy proficiency among immigrants leaves them at a significant disadvantage. It means that more immigrants will find it difficult to participate effectively in their host countries’ labour markets or integrate into their communities. Declines in skills among migrants also pose a challenge for society as a whole, as a widening skills gap between migrants and the native-born population risks undermining equity and social cohesion. In eight of the countries and economies with available data, the gap in literacy proficiency between migrants and non-migrants has widened (Figure 3.17). This was typically due to proficiency among foreign-born adults changing less favourably than that of native-born adults (Figure 3.15). Only Norway and the Flemish Region (Belgium) have seen the gap widening because the skills of native-born adults improved while those of foreign-born adults remained stable.

The widening of the proficiency gap between foreign- and native-born adults was particularly pronounced in Chile, Germany and Singapore (Figure 3.17). As described above, Germany experienced an increase in the relative size of its immigrant population, coupled with low skills among newly arrived migrants (Figure 3.14; see also Table A.3.14 (L) in Annex A). In Singapore, literacy proficiency among long-term migrants has significantly declined (Figure 3.16). Accounting for differences between foreign- and native-born adults in educational attainment, use of the language of the assessment and demographic characteristics accounts for only a fraction of the change in the proficiency gap in these countries. This suggests that additional factors are at play in widening the difference in literacy between these groups.

Numeracy proficiency increased more strongly among migrants than among native-born adults in Canada, England (United Kingdom), Estonia, Finland, Norway and Sweden, narrowing the gap in numeracy between the two groups (see Table A.3.15 (N) in Annex A). In contrast, the gap widened in the Flemish Region (Belgium), Germany and Singapore as native-born adults improved their proficiency more than migrants did.

How have changes in the relative size and average level of proficiency of different socio-demographic groups contributed to overall changes in proficiency of the entire adult population? This section decomposes the changes in countries’ mean literacy scores between the two cycles of the Survey of Adult Skills into two parts. The first relates to the changing composition of the population by age, education and immigrant background, while the second relates to changes in proficiency within these groups. Concretely, the first part quantifies the contribution of increasing educational attainment, ageing and immigration to the average proficiency trend in literacy of the population, assuming the relationship between literacy proficiency and socio-demographic characteristics remained constant. The second part shows how changes in the literacy proficiency of each socio-demographic group contribute to aggregate changes, net of the impacts linked to changes in the size of these groups (see Box 3.4). Table A.3.28 (N) of Annex A presents decomposition results for numeracy.

Overall, changes in educational attainment appear to have a stronger influence on countries’ average literacy proficiency than ageing or immigration (Figure 3.18). This is due to literacy proficiency being on average more strongly related to educational attainment than to age or immigrant status. In most countries and economies, changes in the composition of the population by educational attainment are predicted to contribute to higher average proficiency, as overall levels of education have increased in most countries. These predicted contributions are greatest in Singapore, England (United Kingdom) and the Netherlands. As shown in Figure 3.6, the share of tertiary-educated adults in the adult population aged 25-65 increased by 13 percentage points in England (United Kingdom) and the Netherlands between 2012 and 2023, while the share of adults with below upper secondary education fell by 13 percentage points in Singapore between 2015 and 2023. Panel A in Figure 3.18 shows that changes in educational attainment are associated with an increase of more than 6 points in the average literacy scores in these countries and economies.

At the same time, the relatively stronger declines in proficiency among low-educated adults observed in most countries and economies are associated with substantial decreases in proficiency overall (Figure 3.18, Panel A). This is especially the case in Austria, Estonia, France, Hungary and New Zealand, where the gap in literacy proficiency between adults with below upper secondary or upper secondary education and those with a tertiary education widened considerably. In Estonia, for example, the widening skills gap between adults with tertiary education and those with below upper secondary education contributed to a reduction in overall proficiency by 4 points, while the widening gap between tertiary-educated adults and those with upper secondary attainment contributed to a reduction of 7 points.

Changes in the age composition of populations are not associated with changes in average proficiency because changes to the age distribution have been limited over the years that separate the two cycles of the Survey of Adult Skills (Figure 3.18, Panel B). However, in some countries and economies, differences in how proficiency has changed across different age groups are associated with a lower overall average score. This is the case in Estonia, where proficiency improved among adults aged 25-44, remained unchanged among adults aged 45-54 and declined for those aged 55-65 (Figure 3.11). The fact that older adults’ proficiency deviates from the improvement among adults in prime working age is linked to a reduction of 10 points in the average proficiency of the overall population. In France, the declining proficiency of older adults, against a background of stable proficiency among 25-44 year-olds, also contributes negatively to the population’s average proficiency (by -4 points).

Finally, immigration seems to have a limited impact on countries’ average proficiency. Increases in the immigrant population (assuming constant proficiency in this group and net of impacts of age and education) are only associated with lower mean scores in Norway and Sweden (Figure 3.18, Panel C). As shown in Figure 3.14, these countries saw considerable increases in the share of foreign-born adults in the population between 2012 and 2023 (9 percentage points in Norway and 8 percentage points in Sweden). Conversely, declines in the mean proficiency of the immigrant population relative to the native population (assuming a constant size of the immigrant population and net of impacts of age and education) contributed negatively to the overall level of proficiency in Austria, Germany, New Zealand and Singapore.

Globalisation and technological change have made information-processing skills more important for individuals’ success in work and life. It is therefore crucial to ensure that opportunities for developing skills are accessible to all, regardless of socio-economic background, religion, ethnicity, disability or gender. This section examines how skills disparities related to gender and parental education have changed between the cycles of the Survey of Adult Skills – key to evaluating how countries and societies are faring in their efforts to promote equal opportunities, gender equality and social mobility.

The first cycle of the Survey of Adult Skills found that most OECD countries had a relatively narrow gender gap in proficiency in key information-processing skills (OECD, 2019[10]). There were no significant gender differences in literacy proficiency in most countries and economies, while almost all recorded moderate differences in numeracy in favour of men. These gender differences were typically more pronounced among older adults. Since the first cycle, gender differences in educational attainment have diminished among older adults in many countries, so gender gaps in skills proficiency might be expected to have decreased (OECD, 2024[24]).

Figure 3.19 shows how the skills proficiency of men and women changed between cycles of the survey. In one-third of the participating countries and economies, literacy proficiency has declined among men and has remained unchanged or declined to a lesser degree among women. In Norway, literacy proficiency improved significantly among women, but not among men. Meanwhile, proficiency in numeracy has developed similarly for both genders in most countries and economies. The exceptions are Israel and the United States, where numeracy proficiency declined more strongly among men, and Chile, where numeracy proficiency increased by 15 points among women and remained stable among men.

The stronger declines in literacy proficiency among men has led to a narrowing of the gender gap in literacy in 10 countries and economies (Figure 3.20, Panel A). In many countries, women now score higher than men (see Chapter 2). In numeracy, the proficiency gender gap narrowed only in Chile, Israel and the United States (Figure 3.20, Panel B). Accounting for differences in socio-demographic characteristics between men and women (such as educational attainment) has little impact on the observed changes in the gender gap.

Breaking down the analysis by age provides a more nuanced picture. Although most countries and economies have seen similar changes in the proficiency of men and women across different age groups, in some cases proficiency declined mostly among older men, while in other cases gender proficiency trends diverged among younger age groups. Figure 3.21 shows the resulting changes in the gender proficiency gap among adults aged 16-24, 25-44 and 45-65.

In literacy, differences by age emerge in several countries (Figure 3.21, Panel A). In Canada, Chile, Germany and Korea the gender gap in literacy in favour of men narrowed among older adults due to relatively larger declines in proficiency among older men. In Chile, this pattern also applies to 25-44 year-olds. In Israel, New Zealand and Singapore, a change in the literacy gender gap in favour of women was observed among those aged 25-44 due to greater proficiency declines among men in that age group.

A similar pattern is observed for numeracy (Figure 3.21, Panel B). Chile, Germany, Korea and the United States saw a narrowing gender gap in numeracy among the oldest age group. In Chile and the United States, men and women aged 45-65 now have equal numeracy scores, on average (see Table A.3.5 (N) in Annex A). Chile, Israel, New Zealand, Poland and Singapore saw gender differences in numeracy narrow among 25-44 year-olds. In Israel, New Zealand and Poland, men and women aged 25-44 now have the same level of numeracy proficiency, while in Chile and Singapore, men in this age group still have an advantage in numeracy over women (see Table A.3.5 (N) in Annex A).

These findings suggest that, despite successes in reducing gender differences in educational attainment and skills proficiency, more can be done to promote gender equality. In particular, future policies should aim to further encourage girls and women to engage with numeracy-related tasks in school, work and everyday life, as gender differences in numeracy remain unchanged in many countries. Policies should also address the declining proficiency in literacy among men. This will require a deeper investigation of the factors behind this trend – whether it relates to different study choices, different career choices, or different patterns of use of reading at work and in daily life.

A vast empirical literature identifies three major sources of social inequalities in skills proficiency (Boudon, 1974[25]; Breen and Goldthorpe, 1997[26]; Erikson et al., 2005[27]; Lucas, 2001[28]). First, individuals from disadvantaged backgrounds have a relatively weaker foundation of skills, as they have been raised in social settings that provide less support for learning. This reinforces further disadvantages by limiting their opportunities to advance in education and acquire further skills. Second, disadvantaged individuals tend to be less aware of the range of educational pathways available to them, as their parents typically lack education and familiarity with the education system. This leads to poor educational choices and reduced chances of progressing to higher levels of education. Third, individuals from disadvantaged backgrounds face greater financial constraints on their pursuit of education.

The impacts of social background on skills can be reduced through well-designed policies, including comprehensive and high-quality early childhood education and care services, education counselling services for students and parents, financial support for students, and education and training systems that offer learning opportunities to older adults. The results of the Survey of Adult Skills suggest that the effectiveness of such efforts varies widely since there are substantial cross-country differences in the association between skills proficiency and socio-economic background (see Chapter 2). This section examines how the association between key information-processing skills and social background has changed between the survey cycles across participating countries and economies. This comparison, coupled with detailed knowledge of how policies and institutions have evolved in individual countries and economies, can help identify their strengths and weaknesses in reducing social inequalities.

The analyses use parents’ educational attainment as a proxy for socio-economic background. As in Chapter 2, it groups adults into three categories: those who have low-educated parents (neither parent has attained upper secondary education), those with medium-educated parents (at least one parent has attained upper secondary education) and those with highly educated parents (at least one parent has attained tertiary education). Measuring socio-economic background in this way requires separate analyses by age, since there are generational differences in how common it was to have highly educated parents and the advantages linked to it. Detailed analyses by age are presented in Tables A.3.18 (L) and (N) in Annex A and discussed in the next section, while Figure 3.22 presents trends for the entire adult population.

In all countries and economies except Chile, Denmark, England (United Kingdom), Finland, Norway, Spain and Sweden, the literacy proficiency of adults with low-educated parents has declined. These declines were not statistically significant in the Flemish Region (Belgium), Italy, the Netherlands and Singapore after accounting for changing demographics (i.e. age, immigrant background and gender). Literacy proficiency among adults with medium-educated parents declined in 15 countries and economies, while 12 experienced declines among those with highly educated parents. The declines in the latter group were generally smaller than among adults with lower-educated parents. Only Finland recorded significant improvements in literacy among adults with highly educated parents compared to the previous cycle.

In 14 countries and economies, the proficiency gap in literacy between adults with low-educated parents and those with highly educated parents has widened due to declining proficiency in the former group (Figure 3.23). Only the Slovak Republic and Spain have seen the gap narrow, due to declining proficiency among those with highly educated parents. Accounting for differences in various characteristics, including educational attainment, reduces the widening of the gap in most countries, indicating that changes in inequality are linked to a shift in the composition of socio-economic groups.

The numeracy proficiency of adults with low-educated parents has declined in 11 of the participating countries and economies (see Table A.3.17 (N) in Annex A). Most of these countries saw declines in the numeracy proficiency of adults with medium-educated parents as well. In Singapore, numeracy proficiency among adults with low-educated parents has increased, while in Chile, Finland and Spain this increase was only significant after accounting for demographic changes. In comparison, seven countries saw a decline and four recorded an increase in the numeracy proficiency of adults with highly educated parents. As numeracy generally developed more favourably in this group than among adults with low-educated parents, inequalities in numeracy proficiency increased in many countries (see Table A.3.19 (N) in Annex A).

This section shows how the proficiency gap associated with socio-economic background (as proxied by parental education) has changed across cycles within three age groups: 16-24 year-olds, 25-44 year-olds and 45-65 year-olds. Due to insufficient sample sizes of some groups, adults with low- and medium-educated parents are considered as a single group (denoted as adults whose parents have at most an upper secondary education) and compared to adults with at least one tertiary-educated parent.

Skills-related inequalities in literacy have developed differently in different age groups (Figure 3.24). Canada, Norway and Sweden recorded an increase in the socio-economic gap in literacy among 16-24 year-olds, while socio-economic differences among older adults have remained stable or, in the case of Sweden, have decreased. In Hungary and New Zealand, the socio-economic gap in literacy has increased both among 16-24 and 25-44 year-olds. In another group of countries, including Austria, Estonia, Germany, Israel, Japan and Korea, inequalities increased among 45-65 year-olds and/or 25-44 year-olds, but not among the youngest adults. France saw an increase in social inequalities in skills among the oldest age group and a decrease among those aged 25-44. Next to Sweden, proficiency disparities among older adults narrowed in Singapore and Spain.

The socio-economic gap in numeracy increased especially among young adults (aged 16-24) in Hungary, New Zealand and Sweden, while Germany saw an increase in the gap among those aged 25-44 (see Table A.3.21 (N) in Annex A). In France, Lithuania and Spain, inequalities in numeracy proficiency declined among adults aged 25-44. Singapore and Spain saw a decline in the gap among those aged 45-65, while in Estonia this gap has widened.

To sum up, inequalities in education and skills remain a key policy challenge in many countries. Policies should aim to provide better access to educational opportunities for disadvantaged adults, particularly disadvantaged youth. Policies should also help socio-economically disadvantaged adults to make the most of these opportunities. Perhaps more important than further widening access to education is ensuring that compulsory schooling provides a solid skills foundation. This would offer everyone an equal start in developing higher-order skills and knowledge, and pursuing higher levels of education.

[16] André, C., P. Gal and M. Schief (2024), “Enhancing productivity and growth in an ageing society: Key mechanisms and policy options”, OECD Economics Department Working Papers, No. 1807, OECD Publishing, Paris, https://doi.org/10.1787/605b0787-en.

[21] Blinder, A. (1973), “Wage Discrimination: Reduced Form and Structural Estimates”, The Journal of Human Resources, Vol. 8/4, p. 436, https://doi.org/10.2307/144855.

[25] Boudon, R. (1974), Education, Opportunity, and Social Inequality: Changing Prospects in Western Society, Wiley, New York.

[26] Breen, R. and J. Goldthorpe (1997), “Explaining educational differentials”, Rationality and Society, Vol. 9/3, pp. 275-305, https://doi.org/10.1177/104346397009003002.

[27] Erikson, R. et al. (2005), “On class differentials in educational attainment”, Proceedings of the National Academy of Sciences, Vol. 102/27, pp. 9730-9733, https://doi.org/10.1073/pnas.0502433102.

[2] Hogan, J. et al. (2016), U.S. Program for the International Assessment of Adult Competencies (PIAAC) 2012/2014: Main Study and National Supplement Technical Report, National Center for Education Statistics, US Department of Education, https://nces.ed.gov/pubs2016/2016036_rev.pdf (accessed on 8 July 2024).

[23] Jann, B. (2008), “The Blinder-Oaxaca decomposition for linear regression models”, The Stata Journal, Vol. 8/4, pp. 453-479, https://doi.org/10.1177/1536867X0800800401.

[28] Lucas, S. (2001), “Effectively maintained inequality: Education transitions, track mobility, and social background effects”, American Journal of Sociology, Vol. 106/6, pp. 1642-1690, https://doi.org/10.1086/321300.

[22] Oaxaca, R. (1973), “Male-female wage differentials in urban labor markets”, International Economic Review, Vol. 14/3, p. 693, https://doi.org/10.2307/2525981.

[18] OECD (2024), “International migration database”, OECD International Migration Statistics, https://doi.org/10.1787/data-00342-en.

[24] OECD (2024), OECD Data Explorer: Adults’ educational attainment distribution, by age group and gender, https://data-explorer.oecd.org/ (accessed on 24 September 2024).

[17] OECD (2024), OECD Data Explorer: Population projections, https://data-explorer.oecd.org/ (accessed on 30 August 2024).

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

[13] OECD (2023), Education at a Glance 2023: OECD Indicators, OECD Publishing, Paris, https://doi.org/10.1787/e13bef63-en.

[29] OECD (2023), International Migration Outlook 2023, OECD Publishing, Paris, https://doi.org/10.1787/b0f40584-en.

[15] OECD (2020), Resourcing Higher Education: Challenges, Choices and Consequences, Higher Education, OECD Publishing, Paris, https://doi.org/10.1787/735e1f44-en.

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

[10] 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.

[3] OECD (2019), Technical Report of the Survey of Adult Skills (3rd Edition), OECD, Paris, https://www.oecd.org/content/dam/oecd/en/about/programmes/edu/piaac/technical-reports/cycle-1/PIAAC_Technical_Report_2019.pdf (accessed on 10 July 2024).

[14] OECD (2018), Education Policy Outlook 2018: Putting Student Learning at the Centre, OECD Publishing, Paris, https://doi.org/10.1787/9789264301528-en.

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

[19] OECD (2016), Perspectives on Global Development 2017: International Migration in a Shifting World, OECD Publishing, Paris, https://doi.org/10.1787/persp_glob_dev-2017-en.

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

[8] OECD (2013), Technical Report of the Survey of Adult Skills (PIAAC), https://www.oecd.org/content/dam/oecd/en/about/programmes/edu/piaac/technical-reports/cycle-1/PIAAC_Technical_Report_2013.pdf (accessed on 30 August 2024).

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

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

[11] Paccagnella, M. (2016), “Literacy and numeracy proficiency in IALS, ALL and PIAAC”, OECD Education Working Papers, No. 142, OECD Publishing, Paris, https://doi.org/10.1787/5jlpq7qglx5g-en.

[1] Rampey, B. et al. (2016), Skills of U.S. Unemployed, Young, and Older Adults in Sharper Focus: Results From the Program for the International Assessment of Adult Competencies (PIAAC) 2012/2014: First Look, National Center for Education Statistics, US Department of Education, https://nces.ed.gov/pubs2016/2016039.pdf (accessed on 8 July 2024).

[12] Statistics Canada (2024), International Adult Literacy Survey (IALS), Statistics Canada, https://www150.statcan.gc.ca/n1/ (accessed on 23 September 2024).