What does it take to build the talent base for research and innovation? Drawing on the ReICO evidence base, this chapter compares how countries are developing the competencies their research and innovation systems need most. It analyses the supply and mix of advanced qualifications, assesses countries’ capacity to sustain and expand the talent pipeline, and profiles the foundational skills and attitudes that underpin successful careers in research and innovation.
Research and Innovation Careers Observatory 2026
Key Indicators on Talent, Mobility and Working Conditions
Abstract
Progress on R&I talent development but shortages and imbalances persist across countries
The talent base underpinning research and innovation systems is expanding in most OECD countries. More than 1% of the working-age population in OECD countries now holds a doctorate, while the number of new doctoral graduates is increasing in most countries and exceeded 300 000 in the OECD area alone in 2023. Yet this positive trend masks emerging vulnerabilities. In some countries, the age profile of highly qualified workers combined with declining population may make it increasingly difficult to replenish the future talent pipeline.
Talent is not only about quantity but also about composition. Persistent shortages in specific fields can limit research and innovation performance even where the overall supply of qualified workers is strong. Across the OECD, the disciplinary profile of advanced graduates has changed little over time: STEM fields continue to account for nearly half of all doctoral graduates (46% in 2023), while social science and business fields dominate master’s degrees (41% in 2023). Disparities between men and women also remain pronounced. Women continue to be more highly represented in non-STEM than STEM disciplines at both master’s and doctoral levels, although many countries are seeing growing shares of women’s participation in fields such as engineering and ICT.
Sustaining research and innovation capacity will also depend on the aspirations and capabilities of future generations. Countries differ markedly in the share of 15-year-old students who expect to pursue careers in science and technology, and in some cases these ambitions are weakly aligned with countries’ capacities for science-related careers. Closing this gap will be critical for strengthening tomorrow’s research and innovation workforce.
Countries have two main routes available to build their research and innovation (R&I) workforce: 1) cultivating the necessary knowledge and skills within their own populations; and 2) attracting talented people from abroad. This chapter uses the ReICO evidence base to examine the first of these routes: how R&I talent is formed within countries through acquisition of the qualifications, skills and dispositions that R&I careers demand. The second route - the capacity of countries to attract talented individuals from abroad - is addressed more fully in Chapter 4 as part of the analysis of R&I talent mobility.
As the previous chapter outlined, the potential R&I workforce comprises individuals with advanced knowledge, strong foundational and R&I-specific skills, and favourable attitudes and dispositions to engage in R&I activities. These competencies can be developed through several channels, beginning with school-level education and continuing through higher education, the workplace, continuous professional development and informal learning (Figure 2.1).
While evidence on the extent and depth of R&I competencies developed in the workplace and through informal learning remains scarce, ReICO indicators provide a signal of what formal schooling and tertiary education contribute to R&I talent, in terms of knowledge, skills and attitudes. Using these data, this chapter examines talent development across countries through several lenses: the current stock of highly qualified people, the inflow of newly qualified graduates to the labour market and composition of their fields of study, and young people's foundational skills and their motivation to pursue R&I careers. Taken together, these perspectives provide an indication of how well-placed countries are to renew their potential R&I workforce over the longer term.
Source: Authors’ elaboration.
Working at the frontier of research and innovation often requires advanced disciplinary expertise and ability to generate and apply new knowledge - competencies cultivated and certified through advanced degree programmes. As Chapter 1 noted, graduates from master’s and doctoral programmes provide a key source of talent for R&I systems and therefore warrant close monitoring by policymakers. Public interest is reinforced by the scale of public investment involved. Governments in most OECD countries finance a large share of the cost of advanced degrees, whether through block grants to higher education institutions or through research funding (OECD, 2025[1]). These considerations highlight the importance of understanding the current availability of advanced degree holders for R&I systems and the outlook for renewal of this segment of the workforce in the years to come.
Doctorate holders have grown to over 1% of the OECD working age population
Across OECD countries, in 2024, just over 1% (11 out of every 1 000) of the working age population held a doctorate degree, and 14% (144 out of 1 000) held a master’s degree, although shares vary sharply across countries (Figure 2.2). European countries have the highest shares of doctorates: in Switzerland, the share of 25-64 year-olds with a doctorate reached 33 per 1000, followed by Luxembourg (28), Slovenia (27), Sweden (22) and Germany (19). At the lower end of the scale, Mexico and Costa Rica have one or fewer doctorate holders per 1 000 of their working age population.
Per thousand of the working age population aged 25-64
Notes: Data for more countries are available on the database. 1. Reference year differs: 2021 for Australia, Belgium, Canada, Croatia, Lithuania, Malta (doctorate) and Romania, 2022 for Austria, 2023 for EU27 average (doctorate), Finland, Iceland, New Zealand, Portugal (doctorate), Slovenia, Spain and the United States.
Source: OECD, Research and Innovation Career Observatory (ReICO) Database based on multiple data sources, June 2026. Further information on the specific sources used to compute this indicator can be found in the metadata of the ReICO database.
Relative shares of master's degree holders in the working age population follow a considerably different pattern than those of doctorate holders. For example, several European Union countries with relatively low shares of doctorate holders, such as Bulgaria, Latvia, Lithuania, Poland and the Slovak Republic, have among the highest shares of master’s degree holders. This may reflect, to some extent, a traditional approach in some national education systems of integrating undergraduate and postgraduate education stages into a single “long first degree” programme which awards a master’s degree upon completion rather than a bachelor’s degree (OECD, 2025[2]). In other cases, countries with average shares of doctorate holders, such as Australia, Canada and Korea, have shares of master's degree holders below the OECD average. In these countries, short-cycle and bachelor's degrees are more prevalent as the standard entry qualification for the labour market (OECD, 2025[2]) which may limit demand among learners for master's-level credentials.
Comparing shares of doctorate holders in the working age population over time is challenging, due to gaps in time series data for individual countries and, therefore, the OECD as a whole. Countries with available data, however, tend to show an increasing trend since 2015. For instance, in Denmark the number of doctorate holders per 1 000 of the working-age population has increased by 54% since 2015, reaching 15 per 1 000 in 2024. In the United Kingdom the number reached almost 17 per 1 000 in 2024 – an increase of 45% since 2015 (Figure 2.3).
Per thousand of the working age population aged 25-64
Notes: 1. Break in series. 2. Latest year is not 2024: 2022 for Austria, 2023 for Finland, Iceland, Portugal, Slovenia and the United States.
Source: OECD, Research and Innovation Career Observatory (ReICO) Database based on multiple data sources, June 2026. Further information on the specific sources used to compute this indicator can be found in the metadata of the ReICO database.
In several countries, advanced qualifications are concentrated in older age groups
The share of the working age population holding a doctorate or master's degree indicates how large a country's research and innovation talent base is at a point in time but not how it is likely to evolve in the coming years. Two countries having identical shares of workforce with advanced qualifications may be on distinctly different pathways in terms of turnover and renewal of that workforce, as can be seen when comparing age profiles of doctorate and master’s degree holders.
Unsurprisingly, given differences in programme duration and age of entry, master’s holders on average have a markedly younger age profile than doctorate holders (Figure 2.4). Across the OECD, on average, 26% of all master’s degree holders are aged 25-34, compared to 14% of doctorate holders. On the other end of the age scale, an average of 23% of doctorate holders in OECD countries are aged 55-64, compared to 18% of master’s holders. Some countries, including several with comparatively high shares of doctorate holders among the working age population today, display still older age profiles. The share of doctorate holders aged 55-64 reaches 29% in Finland, Korea and Latvia, and exceeds 25% in a further eight countries including Germany and Sweden.
Percentage of doctorate or master’s degree holders aged 25-64
Note: 1. Reference year differs: 2021 for Australia, Austria, Belgium, Canada, Chile (doctorate), Croatia, Malta (doctorate) and Romania, and 2023 for EU27 average (doctorate), Finland, New Zealand, OECD average (doctorate), Portugal (doctorate), the Slovak Republic (doctorate), Spain, and the United States.
Source: OECD, Research and Innovation Career Observatory (ReICO) Database based on multiple data sources, June 2026. Further information on the specific sources used to compute this indicator can be found in the metadata of the ReICO database.
At master’s level, the picture is quite different. In most countries, more than 50% of master’s graduates are aged under 45, while shares exceed 60% in fifteen countries including the Republic of Türkiye (here after Türkiye) (77%) and France (68%). This much younger profile can reflect recent expansion of higher education provision (as in Türkiye) or, in some countries, an earlier loss of older qualified workers through emigration.
These age profiles of doctorate and master’s holders provide an indication of R&I workforces where renewal pressures are building. Countries with ageing populations of advanced degree holders may need a multi-faceted response, part of which includes increasing the flow of new graduates at master’s and doctoral level entering the workforce.
Over 300 000 new doctorates were awarded across OECD countries in 2023
A key aspect of talent development is the extent to which countries can sustain and expand current talent supplies through new cohorts of graduates. In 2023, OECD countries alone awarded over 300 000 new doctoral degrees and over 3.9 million master's degrees. The broader group of OECD and EU Member countries, OECD accession candidates and Key Partner countries together awarded almost half a million (about 480 000) new doctoral degrees and more than 7 million master's degrees. As Figure 2.5 shows, graduates with advanced qualifications are heavily concentrated in a few regions, in particular the European Union, the People’s Republic of China (hereafter, China), India, the United Kingdom, and the United States. Together, these five areas accounted for 70% of doctoral graduates and over 80% of master’s graduates across OECD and EU member, accession and partner countries in 2023.
Note: A total of 51 countries are included in the analysis, comprising all OECD and European Union member countries, along with OECD key partners and accession candidate countries (OECD, n.d.[3]). For each panel, countries/aggregates with the ten largest values are shown, with the remaining countries aggregated in the “Other” category. Please refer to the StatLink for data and data source by individual country.
Source: OECD, Research and Innovation Career Observatory (ReICO) Database based on OECD Education database, June 2026, complemented by data from various national sources.
On average, advanced graduate numbers are on the rise, but there are considerable differences across countries…
Across the OECD, on average, both entrants to and graduates from master's programmes rose steadily between 2015 and 2023. Overall, over this period, doctoral graduations across the OECD increased by 23% and master's graduations by 33% (Figure 2.6). New entrants to advanced programmes also increased over the same period by 26% at doctoral level and 27% at master's level. As entrant numbers are a leading indicator of future graduate output, these trends point to continued growth in graduate numbers beyond 2023, provided that entrants complete their programmes at similar rates to previous cohorts.
Both master’s entrants and master’s graduates show a steady upwards trend between 2015 and 2023. Doctoral entrants follow a less consistent pattern, remaining broadly stable between 2016 and 2020 before increasing steadily thereafter. Doctoral graduations are the only trend showing a decline over the period, with numbers decreasing or plateauing after 2016 and reaching their lowest point in 2020, coinciding with the postponement of many thesis defences during the COVID-19 pandemic. As with doctoral entrants, however, doctoral graduations resumed an upward trend from 2020 onwards.
OECD average (2015 =100)
Source: OECD, Research and Innovation Career Observatory (ReICO) Database based on OECD Education database, June 2026.
The OECD average masks considerable differences in the situation of talent pipelines across individual countries. While most countries have increased flows of graduates from master’s programmes in recent years, for doctoral education, the situation is more mixed (Figure 2.7). In general, graduate output from master’s programmes has been increasing at a faster rate than doctoral output, although there are some examples of countries that have substantially increased numbers of both doctoral and master’s graduates. Mexico shows the highest relative increases over the period, with doctoral graduates more than tripling and master’s graduates more than doubling, followed closely by Türkiye, with respective increases of 262% and 228%, albeit from a small base.
On the other end of the scale, many European countries recorded overall declines in the numbers of doctoral graduates between 2015 and 2023. Indeed, on average across the EU-27 countries, the number of doctoral graduates fell by approximately 2% over the period, compared to an average increase of 16% in the number of master’s graduates. Several Central and Eastern Europe (Bulgaria, Croatia, Czechia, Latvia, Lithuania, Romania, the Slovak Republic and Slovenia), show important declines in both doctoral and master’s graduates over this period. Another group of European countries (Denmark, Finland, Germany, Italy, Ireland and Sweden) combine decreasing numbers of doctoral graduates with increasing numbers of master’s graduates. Among this group, Ireland shows the largest difference in changes of graduates between doctoral and master’s graduates, driven by a substantial increase in master’s graduates between 2015 and 2023 (216%), reflecting major public investment in recent years to increase postgraduate throughput through dedicated upskilling, reskilling and skills conversion programmes (Higher Education Authority, n.d.[4]).
Index of change (2015=100)
Source: OECD, Research and Innovation Career Observatory (ReICO) Database based on OECD Education database, June 2026.
…..with implications for future supplies of R&I talent
Concerningly, several countries showing shrinking pools of graduates from doctoral programmes between 2015 and 2023 are already starting from a low base of population with qualifications at this level. Eight countries fall into this category, all located in Europe (Figure 2.8). Much of the contraction reflects underlying demographic decline, particularly acute across Central and Eastern Europe, which constrains countries’ capacity to provide talent for, and expand, local research and innovation ecosystems. Countries in this situation are pursuing multipronged policy responses to mitigate the risks associated with the loss of talent, including stepping up efforts to attract and retain new and returning talent from abroad, raising participation in education and training within their own populations, and addressing underrepresentation where it exists (Bruegel, 2025[5]).
In other regions, countries starting from similar low shares of population with doctoral level qualifications have been more successful in expanding numbers of doctoral graduates in recent years (Figure 2.8). Chile and Mexico fall into this category, which may reflect the impact of policy initiatives seeking to expand highly-qualified human capital in the region, for example, national initiatives to increase the numbers of higher education staff holding doctorates in Mexico (EC-OECD, 2025[6]; EC-OECD, 2025[7]) and scholarships supporting study abroad to obtain advanced qualifications in both Mexico and Chile (EC-OECD, 2026[8]; Secretaría de Ciencia, Humanidades, Tecnología e Innovación, n.d.[9]).
Few countries combine above-average shares of doctorate holders in the population with substantial increases in doctoral graduates over the period 2015-2023 (Figure 2.8). Seven countries fall into this category, while a further seven countries with higher-than-average shares of doctorate holders have seen a decline in graduates over this period. Switzerland is an outlier, combining the highest share of doctoral holders (3.3% of the working-age population in 2023) with an increase of 25% in graduate numbers between 2015 and 2023. Such changes may be driven by underlying demographics, evolving attitudes towards undertaking a doctoral degree or changing policy framework conditions underpinning doctoral education.
Notes: Data for more countries are available on the database. 1. Reference year differs for share of doctorate holders: 2021 for Australia, Belgium, Canada, Croatia, Lithuania and Romania, 2022 for Austria, 2023 for EU average, Finland, Iceland, New Zealand, Portugal, Slovenia, Spain, and the United States.
Source: OECD, Research and Innovation Career Observatory (ReICO) Database based on multiple data sources, June 2026. Further information on the specific sources used to compute this indicator can be found in the metadata of the ReICO database.
An ageing highly qualified population coinciding with a shrinking inflow of new graduates to the labour market risks creating a structural problem that is difficult to correct quickly. Efforts to grow the domestic pipeline take effect only with a long lag, especially for doctoral training which requires a comprehensive financial and time investment from governments, training organisations and the doctoral candidates themselves. For example, in Portugal, where the share of doctorate holders aged 55-64 reaches close to 30%, doctoral trainees on average reach graduation only at age 39. On the other hand, some other countries - such as Belgium and Denmark - combine a relatively young age profile of doctoral candidates with a typical age of 33 at graduation (Figure 2.9). Fewer doctoral graduates today imply a weaker future academic and research workforce, which in turn reduces the institutional capacity to supervise and train the next generation of doctoral students - a contraction cycle that is difficult to reverse without deliberate and sustained policy intervention.
Notes: Data for more countries are available on the database. 1. Reference year differs: 2021 for the Netherlands (doctorate only), 2022 for Bulgaria and Romania.
Source: OECD, Research and Innovation Career Observatory (ReICO) Database based on OECD Education database, June 2026.
Timely and granular indicators relating to talent for R&I within countries are particularly important in contexts of demographic decline, not least for assessing whether expansion, retention and attraction policies are delivering results. Traditionally, few countries or higher education institutions systematically measure and monitor stocks, flows and outcomes of individuals with doctoral level qualifications. While some countries have rolled out new initiatives in recent years (see Chapter 5), there is still considerable room for progress. In Europe, for example, a recent study showed that graduate tracking (overall, for both vocational and higher education) is an established national practice in just 18 of the 31 reviewed countries, while doctoral graduates remain much less likely to be included in any form of graduate tracking initiatives conducted by higher education institutions (European Commission, 2020[10]).
Aggregate numbers of advanced qualification holders cannot on their own indicate whether R&I systems can access the specific knowledge and skills they require. National research and innovation priorities, formulated for example in terms of seizing on the opportunities of key emerging technologies or the aim to address defined societal goals, draw on particular fields of knowledge and skillsets. Shortages in these fields can constrain research and innovation capacity even where the overall supply of highly qualified workforce is adequate (OECD, 2025[11]).
In most countries, social sciences and business account for the largest shares of master's graduates, while STEM dominates doctoral education
Field-of-study patterns differ markedly between master's and doctoral education. Social sciences and business account for the largest share of master's graduates in most OECD countries, standing at 41% of all graduates on average. At doctoral level, however, STEM fields, as defined in ReICO (Box 2.1) account for the largest share of graduates, making up 46% of graduates on average in the OECD area and reaching or surpassing 60% in France, Italy and Luxembourg (Figure 2.10).
STEM (Science, Technology, Engineering and Mathematics) is an umbrella term used by many governments and researchers to group knowledge and skills considered especially relevant for technological innovation, economic productivity and solving complex scientific and societal problems. Given their importance governments often prioritise the improvement of STEM skills in education and workforce policy (EC-OECD, 2026[12]).
Generally, STEM refers to an aggregation of fields of education and training, commonly including natural sciences, mathematics and statistics; information and communication technologies and engineering, manufacturing and construction. It should be noted that this is a broad aggregate that groups fields with quite different knowledge bases and labour markets together and can lead to differing conceptualisations across countries and organisations. Despite these complexities, however, STEM remains a valuable and widely applied aggregate for comparing countries and providing a focus for education and skills policy.
In ReICO, the categorisation of STEM comprises ISCED-F fields 05 (Natural sciences, mathematics and statistics), 06 (Information and Communication Technologies), and 07 (Engineering, manufacturing and construction) and 08 (Agricultural sciences), are used to denote the STEM category throughout. This definition can vary from some national definitions. For example, it excludes ISCED-F field 09 (health sciences) which some national definitions include. Countries with large medical doctoral programmes (notably several Southern European systems) may appear to have lower STEM shares than a broader definition would produce. Comparison across countries should be read with this boundary condition in mind.
Several countries, including France, Finland, Sweden and Estonia, show relatively high percentages of STEM graduates from both master’s and doctoral education, indicating a strong orientation of tertiary education towards technical fields (Figure 2.10). At the other end of the distribution, STEM accounts for less than one-quarter of doctoral graduates and less than 10% of master’s graduates in Costa Rica and Mexico. While these are important differences, indicators of STEM output should be interpreted with caution - high shares of doctorates graduating from STEM fields does not in itself guarantee alignment to the needs of R&I systems. Demand for individual fields within the STEM categorisation varies considerably, across R&I-intensive sectors and enterprises and in the labour market as a whole (Stephan, 2012[13]).
Nevertheless, graduates in certain STEM fields, such as engineering and ICT, are in particularly strong demand across the labour market, including in R&D-intensive sectors and enterprises. Some evidence indicates that countries expanding public investment in education in these fields can reap dividends in terms of innovation capacity (Biasi, Deming and Moser, 2021[14]). For this reason, increasing the supply of STEM skills is a frequent policy target for national governments, with several implementing new initiatives to increase supply of STEM skills in recent years (EC-OECD, 2026[15]).
Differences across countries in shares of arts, humanities and education graduates are equally substantial. At doctoral level, this field group accounts for roughly half of graduates in Costa Rica and Mexico (largely driven by high throughput of doctorates in education), compared to less than 10% in Denmark, Germany, the Netherlands and Sweden. Other countries with relatively high shares of arts, humanities and education graduates at both doctoral and master’s level include Türkiye (with 26% of doctoral graduates and 24% of master’s graduates from these fields) and Korea (21% of both doctoral graduates and 29% of master’s graduates).
Percentage of doctoral or master’s graduates in all fields of study
Source: OECD, Research and Innovation Career Observatory (ReICO) Database based on OECD Education database, June 2026.
These differences across fields partly reflect the structure of national economies, education, research and innovation systems, but they also may carry policy implications: countries whose graduate output is concentrated in fields misaligned with their research and innovation priorities may face greater difficulty in providing highly qualified staff with attractive career opportunities. Chapter 3 examines indicators of employment opportunities and outcomes for highly qualified graduates in more depth.
Field distribution of graduates has overall changed little over time….
Overall, in the OECD area, the field distribution of graduates has remained remarkably stable. Between 2015 and 2023, the shares of graduates from each field group changed by at most 3 percentage points, and in most cases by much less. The greatest changes can be observed in the shares of doctoral graduates in health, which increased from 14.4% in 2015 to 17.4% in 2023. Over the same period, doctoral graduates from STEM fields reduced from 48.1% to 46.1%, while master's STEM graduates increased by less than a percentage point, from 23.2% to 23.9% (Figure 2.11). In arts, humanities and education, smaller changes over time can be observed, by 1.4 percentage points for doctorates (from 19% to 17.6%), while the master’s share moved in the opposite direction, rising by 1.4 percentage points (14.1% to 15.5%). Shares of graduates in social sciences and businesses have changed the least over the period, from 18.6% in 2015 to 19.0% in 2023 for graduates from master’s programmes and 18.6% to 19.0% for graduates from doctoral programmes.
OECD average, percentage of doctoral or master's graduates in all fields of study
Source: OECD, Research and Innovation Career Observatory (ReICO) Database based on OECD Education database, June 2026.
….though most countries have increased graduate numbers in absolute terms
The stability observed in distribution of graduates over time at the level of the OECD area masks pronounced changes in some countries in terms of supply of advanced degree holders across fields. Changes in the flow of STEM-qualified graduates over time exemplify this. Stability or even slight reduction in field distribution over time, as shown in Figure 2.11, does not imply stagnation in supply of STEM talent for research and innovation systems. As highlighted above, most countries have expanded annual outflows of master’s and doctoral graduates in recent years, which in turn leads to increases in STEM graduates in absolute terms. Examining these changes across countries and situating them with respect to shares of graduates from STEM fields, provides an indication of how absolute supply of STEM talent feeding into research and innovation systems is changing, and how this relates to the STEM orientation of each country's graduate output (Figure 2.12).
At doctoral level (Panel A of Figure 2.12), the upper-right quadrant indicates countries with relatively high shares of graduates from STEM fields combined with higher-than-average growth in numbers of STEM graduates. Few countries (Australia, Chile, Luxembourg and Poland) fall into this category. A more common pattern among countries with high STEM shares is for lower-than-average growth or even net decline in the numbers of STEM graduates between 2015 and 2023, as is the case in Germany, Israel, Latvia and Sweden. Colombia and Türkiye are prominent examples of countries with lower-than-average shares of STEM graduates at the doctoral level and among the highest growth rates in absolute terms between 2015 and 2023.
Panel B of Figure 2.12 shows even more pronounced volume growth of graduates with STEM degrees at master’s level between 2015 and 2023, exceeding 200% in Chile, Ireland, Luxembourg and Türkiye. The upper-right quadrant (above-average share and above-average growth in absolute numbers) includes Denmark, Ireland, Greece, the Netherlands, New Zealand and Norway, while a cluster of European countries, including Czechia, Estonia, Germany, Sweden and Switzerland, also have high shares but show more subdued or declining growth.
Source: OECD, Research and Innovation Career Observatory (ReICO) Database based on OECD Education database, June 2026.
Conversely, several countries with below-average shares of STEM master’s graduates, such as Chile Colombia, Türkiye, the United Kingdom and the United States, show strong expansion in absolute numbers, showing that STEM talent pools can expand in absolute terms even without major changes in field composition. Whether these increases are sufficient to meet the needs of R&I systems depends on the scale and growth of each country's system, the share of graduates choosing R&I careers research, and demand for advanced skills driven by technological change, industrial policy, and competition from other sectors.
Women graduates continue to be highly concentrated in non-STEM fields at both master’s and doctoral level
Across OECD countries, women now account for the majority of graduates with advanced degrees. Among countries with available data, the average share of women graduates at doctoral level was 49% in 2023 and 59% at master’s level. However, as Figure 2.13 shows, shares of women graduates continue to diverge sharply across fields. In all countries with available data, the vast majority of arts, humanities and education graduates are women, with shares surpassing 80% of doctoral graduates in Latvia and master’s graduates in three countries (Bulgaria, Estonia and Latvia).
Percentage of doctoral or master’s graduates in the same field
Note: Data for more countries are available on the database.
Source: OECD, Research and Innovation Career Observatory (ReICO) Database based on OECD Education database, June 2026.
Similarly high shares of women graduates are evident across health fields at doctorate and master’s level in most countries. Social sciences and business fields show a more mixed profile - women make up the majority share of graduates at master’s level in all countries except Türkiye but make up less than half of doctoral graduates in ten countries.
Conversely, women graduates continue to be less prevalent in STEM fields at both doctorate and master’s level. At master’s level, only two countries (Greece and Iceland) have shares of women exceeding 50% in STEM, with a particularly high share of women graduates in Iceland (58%). At doctoral level a larger group of countries are closer to having equal shares of women and men STEM graduates, particularly in countries where the overall share of women among graduates is highest, such as Croatia, where women comprise 51% of STEM graduates, Romania (49%) and Lithuania (51%).
Thus, persistent gender gaps remain across broad field groupings, reflecting cultural norms, patterns of prior achievement, and the choices made during earlier phases of education (OECD/Eurostat/UNESCO Institute for Statistics, 2015[16]). These broad categories, however, mask substantial variation within them. In STEM, for instance, women's share of enrolment varies considerably from one subfield to another - often greater than the share of men in areas such as the biological and life sciences, while remaining markedly lower in fields like computing and engineering.
Disaggregating STEM into its constituents illustrates the differences in women’s share of graduates across different fields (Figure 2.14). In 2023, in natural sciences, mathematics and statistics, shares of women graduates sit closest to, and in several cases exceeded, 50%. Women account for at least 40% of graduates in these fields in most countries and make up more than 60% of graduates in Bulgaria, Croatia, Poland and the Slovak Republic. Engineering and ICT, by contrast, have considerably higher shares of male graduates. In engineering, no country had a share of women graduates exceeding 50%, although six countries (Brazil, Estonia, Greece, Lithuania, Portugal and Romania) had shares exceeding 40%. However, most countries show substantially lower shares, with the lowest values in Korea (18%) and Germany (20%). ICT presents a broadly similar profile to engineering, with women’s shares of graduates meeting or exceeding 40% in just four countries (Bulgaria, Latvia, Lithuania and Romania).
However, Figure 2.14 also highlights some evolution happening over time in several countries in the share of women graduates in ICT and Engineering fields, although changes are, for the most part, relatively modest. Between 2015 and 2023, several countries recorded an increase of more than 5 percentage points in the share of women engineering graduates, of which four countries (Greece, Iceland, Luxembourg and Portugal) increased the share by 10 percentage points or more. Increases were even more pronounced in ICT fields, where 21 countries increased their share of women graduates by at least 5 percentage points, and Belgium, Iceland and Latvia increased shares by 20 percentage points or more.
Percentage of doctoral graduates in the same STEM subfield
Note: Data for more countries are available on the database. 1. Latest year is 2022.
Source: OECD, Research and Innovation Career Observatory (ReICO) Database based on OECD Education database, June 2026.
Many countries have progressively improved their ability to identify and profile individuals with advanced qualifications, with labour force surveys increasingly able to distinguish holders of master's and doctoral degrees and support internationally comparable indicators on outcomes such as employment conditions and quality of life (as shown in Chapter 3). By contrast, statistical information on the population with other dimensions of competency for R&I careers, such as skills most relevant to R&I and positive attitudes towards R&I work, remains limited and largely relies on international skills surveys such as the OECD’s Programme for International Student Assessment (PISA) and the OECD Programme for International Assessment of Adult Skills (PIAAC). This section summarises some key insights from PISA and PIAAC relevant for R&I talent development. Considerable information gaps remain in measurement of skills and viewpoints on R&I careers on both the demand and supply sides, that will require complementary approaches, including individual-level surveys and alternative data sources (see Chapter 5).
Anticipation of successful careers in science and technology starts during youth
Longitudinal analyses of national and international data consistently show that early attention and exposure to career guidance and preparation activities is associated with better labour market outcomes for individuals (OECD, 2025[17]). From the perspective of policymakers, the perceived attractiveness and accessibility of R&I-related careers also influence future workforce supply, which motivates policy attention on teenage career readiness.
International comparisons on the career aspirations of young people are limited. PISA, the OECD’s assessment of the skills of 15-16 year-olds that has taken place at three-year intervals since 2001, provides one viewpoint. The PISA questionnaire asks students undergoing the assessment to specify the type of occupation they expect to have at around the age of 30. While the occupation classification does not allow the specific identification of R&I careers, it allows for the assessment of career aspirations in closely related and overlapping roles, notably science, engineering and ICT professionals. PISA data shows considerable gaps across countries in expectations to work as professionals in these fields, ranging from below 10% in the Netherlands to more than 25% in Peru (Figure 2.15).
In 2022, on average across OECD countries, 10.7% of students indicated that they expected to work as science and engineering professionals, a share that has remained largely unchanged in recent years (11.1% in PISA 2018 and 10.8% in PISA 2015). By contrast, average shares of students expecting to work in ICT professions, though lower in general than those expecting to work in science and engineering, are increasing over time, going from 3.4% in PISA 2015 to 5.5% in PISA 2022 (OECD, 2026[18]).
In most countries, students expecting to work in science and engineering professions outnumber those expecting to work in ICT, with career expectations in ICT typically accounting for well under half of the combined share. However, expectations for ICT careers are notably stronger in several European countries, such as Bulgaria, Estonia, Lithuania and the Slovak Republic. Signals from the labour market on employment and wage premia for these roles may influence career decisions in these countries – many of which have relatively high-wage premia for jobs in ICT (Eurostat, 2026[19]). Policy actions may also play a role. For example, in Estonia, where there has been prominent and sustained policy commitment to digital-first public and business services and where successive governments have heavily invested in ICT skills in the population, almost 15% of students expect to work in ICT professional jobs, the highest among all countries with available data. While no direct link can be drawn, the available evidence suggests that aspirations to work in science and technology may be shaped not only by career guidance programmes within schools, but by wider policy interventions and the broader economic and social context within individual countries.
Share of students expecting to be employed in occupations by age 30
Notes: Data for more countries are available on the database. 1. Data did not meet the PISA technical standards but were accepted as largely comparable.
Source: OECD, Research and Innovation Career Observatory (ReICO) Database based on PISA 2022, June 2026.
Aspirations and capacity for science careers may not always align
PISA data also enables comparison across countries of the shares of students with aspirations for careers as science and engineering professionals and shares of top performing students in the PISA science assessment. As Figure 2.16 shows, there is no relationship across countries between these respective shares. Indeed, some countries where the largest shares of students anticipate careers as science and engineering professionals also have some of the smallest shares of top performers in science, Conversely, several countries with high shares of top-performing students appear not to have especially high shares of students with career aspirations in science and engineering, indicating a potential disconnect between careers aspirations and the skills needed to realise them.
These findings, based on aggregate country-level shares, provide only a rough indication of trends and do not reveal how these variables relate at individual student level. Future ReICO work could explore this relationship more directly through analysis of PISA microdata.
Source: OECD, Research and Innovation Career Observatory (ReICO) Database based on PISA 2022, June 2026.
Adults' foundational skills for many R&I roles differ markedly across OECD countries
Quantitative and adaptive problem solving are core skills for several R&I occupations. Data from PIAAC, the OECD’s adult skills survey, shows that across countries with available data, average scores for these two skills are highly correlated and that countries have diverse levels of access to these skills within their respective populations (Figure 2.17).
Source: OECD, Research and Innovation Career Observatory (ReICO) Database based on the results of the OECD Survey of Adult Skills 2023, June 2026.
Finland, Japan and Sweden lead on both measures, while Chile records the lowest scores. From a policy perspective, these variations indicate that measures to expand development of R&I talent within domestic populations may need to take account of the differing skill foundations on which they can build.
The evidence presented in this chapter highlights both progress and emerging challenges in the development of talent for research and innovation. Expanding stocks of highly qualified individuals provide an important foundation for research and innovation activities. At the same time, significant differences remain across countries in the size, age structure and field composition of highly qualified talent pools, with some systems facing growing challenges in renewing their research and innovation workforce. In this context, ensuring that education, skills and research systems can continue to develop, attract and retain the talent needed for research and innovation will remain an important policy priority. Policy responses will need not only to ensure adequate supply of relevant advanced qualifications, but also to foster the skills and aspirations that encourage younger generations to pursue research and innovation careers.
References
[14] Biasi, B., D. Deming and P. Moser (2021), Education and Innovation, National Bureau of Economic Research, Cambridge, MA, https://doi.org/10.3386/w28544.
[5] Bruegel (2025), The demographic divide: inequalities in ageing across the European Union, https://www.bruegel.org/policy-brief/demographic-divide-inequalities-ageing-across-european-union.
[12] EC-OECD (2026), STEM skills policy theme dashboard, https://stip.oecd.org/stip/interactive-dashboards/themes/TH51.
[8] EC-OECD (2026), STIP Compass: Chile Scholarships, https://stip.oecd.org/stip/interactive-dashboards/policy-initiatives/2025%2Fdata%2FpolicyInitiatives%2F2694.
[15] EC-OECD (2026), STIP Compass: STEM Skills, https://stip.oecd.org/stip/interactive-dashboards/themes/TH51.
[7] EC-OECD (2025), STIP Compass: National Postgraduate System, https://stip.oecd.org/stip/interactive-dashboards/policy-initiatives/2025%2Fdata%2FpolicyInitiatives%2F99997519.
[6] EC-OECD (2025), STIP Compass: Researchers for Mexico, https://stip.oecd.org/stip/interactive-dashboards/policy-initiatives/2025%2Fdata%2FpolicyInitiatives%2F4563.
[10] European Commission (2020), Mapping the State of Graduate Tracking Policies and Practices in the EU Member States and EEA Countries: Final Report, Publications Office of the European Union, https://data.europa.eu/doi/10.2766/62505.
[19] Eurostat (2026), Structure of earnings survey, https://ec.europa.eu/eurostat/databrowser/view/earn_ses_monthly/default/table?lang=en.
[4] Higher Education Authority (n.d.), Human Capital Initiative (HCI), https://hea.ie/skills-engagement/what-is-human-capital-initiative-hci/.
[18] OECD (2026), ReICO database, https://data-explorer.oecd.org/s/51p.
[2] OECD (2025), Education at a Glance 2025: OECD Indicators, OECD Publishing, https://doi.org/10.1787/1c0d9c79-en.
[11] OECD (2025), OECD Science, Technology and Innovation Outlook 2025: Driving Change in a Shifting Landscape, OECD Publishing, https://doi.org/10.1787/5fe57b90-en.
[1] OECD (2025), The Financial Sustainability of Higher Education - Insights from Policy in OECD Countries, https://www.oecd.org/en/publications/the-financial-sustainability-of-higher-education_f544ccfe-en.html.
[17] OECD (2025), The State of Global Teenage Career Preparation, OECD Publishing, Paris, https://doi.org/10.1787/d5f8e3f2-en.
[3] OECD (n.d.), Members and partners, https://www.oecd.org/en/about/members-partners.html.
[16] OECD/Eurostat/UNESCO Institute for Statistics (2015), ISCED 2011 Operational Manual: Guidelines for Classifying National Education Programmes and Related Qualifications, OECD Publishing, https://doi.org/10.1787/9789264228368-en.
[9] Secretaría de Ciencia, Humanidades, Tecnología e Innovación (n.d.), Becas al Extranjero, https://www.secihti.mx/becas-al-extranjero/.
[13] Stephan, P. (2012), How Economics Shapes Science., Harvard University Press, https://doi.org/10.4159/harvard.9780674062757.