OECD Economic Surveys: Sweden 2025

Report

OECD Economic Surveys: Sweden

5 June 2025

Available in:

English
français

Download PDF

4. Strengthening skills and labour market matching

Copy link to 4. Strengthening skills and labour market matching

Jon Pareliussen
OECD

Abstract

Investments in skills raise productivity, boost the benefits from technological change, foster social cohesion and reduce vulnerabilities to disinformation. The OECD Survey of Adult Skills shows that skills are high and improving in Sweden. Older workers are more literate and have stronger labour market outcomes today than before. Younger cohorts having performed weakly in international skills assessments in the past show few signs of weak skills or labour market outcomes today. Even so, scope remains to raise efficiency and equity in compulsory school and improve the attractiveness and labour market relevance of upper secondary vocational education. The quality of university education seems to have weakened, while an increasing skill premium may sharpen the alternative cost of motherhood. Few have very weak skills in OECD comparison, but those concerned are much more likely to be unemployed. Investments in skills can increase the employability of those with no upper secondary diploma and immigrants from outside of Europe and North America, but proven and effective interventions are used too little and too late. A new programme increasing funding for the education of experienced workers should be adjusted to improve value for money and expand student numbers.

4.1. Background

Copy link to 4.1. Background

Sweden's economic model creates a range of favourable outcomes. High labour force participation, including among women and foreign-born individuals, combined with a compressed wage distribution and a comprehensive social safety net hold back income inequality, while overall wage growth tends to be compatible with price stability and cost competitiveness of exporting industries over time.

This model rests on a carefully crafted equilibrium which depends on a highly skilled population participating actively in the labour force. Sweden went from being among the top performers in the OECD’s Programme for International Student Assessment (PISA) in the early 2000s to scoring just below the OECD average in all three subjects in 2012, triggering a number of school reforms and improving results thereafter. PISA tests the skills of 15-year-olds in reading, mathematics and science and results fell in Sweden in 2022, as they did across the OECD largely due to educational disturbances during the Covid-19 pandemic (Figure 4.1). Despite these “PISA shocks”, and a rising share of foreign-born individuals in the population, Sweden’s adults are on average more literate today than a decade ago and have improved their performance compared to the OECD average considerably, as measured by the 2023 vintage of the OECD Survey of Adult Skills. The cohorts with weak results in PISA 2012 seem to have caught up and are now high-skilled on average. Older workers now have also significantly higher literacy and labour market outcomes than the previous generations.

Figure 4.1. Sweden’s PISA results bottomed out in 2012, but fell again in 2022
Copy link to Figure 4.1. Sweden’s PISA results bottomed out in 2012, but fell again in 2022

Skills are also put to good use, with the third-highest participation rate in the OECD and the tenth highest employment rate. The unemployment rate is high in international comparison, but this is largely because segments of the population who would be classified as inactive in other countries are receiving benefits that are conditional on job search in Sweden. In particular, many full-time students also seek part-time jobs, thereby artificially inflating youth unemployment numbers.

This overall strong performance leaves no room for complacency. Skills empower people and allow them to manage and embrace complexity, contributing to better-informed collective decision making and policy choices. In an era of rapid technological development, and geopolitical tensions, information processing skills are becoming increasingly important to effectively draw advantage of technologies like artificial intelligence (AI) and critically reflect on a constant stream of information to separate true from false and preserve social cohesion (OECD, 2024[1]).

The PIAAC data also reveal some areas for improvement. The skills premium associated with having a tertiary degree has fallen over the past decade while the labour market increasingly demands higher qualifications. A fairly large share of high-skilled and highly qualified workers is not used to their full potential, reducing individual earnings and holding back productivity. Gender gaps remain, although moderate compared to the OECD average.

The low-skilled, although constituting a relatively low share of the Swedish population, not only perform relatively weakly in the labour market, but also report weaker health, trust less in others and feel disaffected from politics (Pareliussen, 2025[2]). Labour shortages and long-term unemployment have increased in tandem, as a sizeable share of the working-age population has too weak skills or qualifications to be attractive to employers. There are also structural shortages of for example health care personnel and teachers, notably in rural areas experiencing severe ageing and depopulation (OECD, 2023[3]). Some immigrant groups are significantly less likely to be employed and are less paid than domestic-born on average, largely reflecting their weaker literacy skills.

Against this background, this chapter sheds light on skills and labour market outcomes in Sweden supported by new evidence from the 2023 vintage of the OECD Survey of Adult skills (OECD, 2024[1]; Pareliussen, 2025[2]). The next section discusses the distribution of skills among the adult population, how skills translate into employment and wages. The third section discusses some directions of reform in the educational system. The last section focusses on labour market matching, the incidence of weak skills and the societal cost of different types of mismatches and draws policy implications to secure value for money in employment services and adult education and training.

4.2. The OECD Survey of Adult Skills shows that skills are high overall in Sweden

Copy link to 4.2. The OECD Survey of Adult Skills shows that skills are high overall in Sweden

New data from the second vintage of the OECD Survey of Adult Skills (Box 4.1) can shed light on key questions including the skills of the “PISA shock” generations, skills and labour market outcomes of foreign-born people and older generations, as well as questions relating to labour market matching, education- and skills premiums and the incidence of discrimination based on gender and country of birth.

Box 4.1. The OECD Survey of Adult Skills

Copy link to Box 4.1. The OECD Survey of Adult Skills

The second cycle of the OECD Survey of Adult Skills, which is a product of the Programme for the International Assessment of Adult Competencies (PIAAC), took place in 2022/23, 11 years after the first cycle. The Survey of Adult Skills is the outcome of a collaboration among participating countries and economies, the OECD Directorate for Education and Skills and the OECD Directorate for Employment, Labour and Social Affairs, and an international consortium led by the Educational Testing Service (ETS). From the OECD, 29 countries, England and the Flemish region of Belgium participated, with national sample size ranging from 3 130 to 11 697.

The Survey of Adult Skills assesses literacy, numeracy and adaptive problem solving based on conceptual frameworks that define what these skills are and describe how to design assessment items to measure them. The skills assessment modules are combined with a rich background questionnaire covering variables such as age, gender, family, immigrant background, educational attainment, labour market status and health- and social outcomes.

The assessment tasks reflect how individuals apply skills across content, contexts and the cognitive processes involved. Responding to an item requires accessing and understanding the information provided, evaluating it and reasoning critically. Within each module, skills are measured on a 500-point scale, with a higher score corresponding to higher proficiency. The scale is divided into functional skill levels, each straddling 50 PIAAC score points. Individuals at the lowest levels (level one and below, scoring 175 points or less) can at best respond to simple tasks with straightforward processing and little irrelevant information. Individuals scoring at the highest levels (levels four and five, scoring 326 points or higher) can access, analyse, reason, and critically reflect on complex and abstract tasks and separate relevant from irrelevant information to arrive at the correct response.

The 2023 vintage of the OECD Survey of Adult Skills is designed to be able to statistically compare scores in literacy and numeracy to scores within the same module in the 2012 vintage. Results cannot be compared between the different modules (literacy, numeracy, adaptive problem solving), but the correlation of scores between the different items is in general very high, especially so between literacy and numeracy.

Source: OECD (2024[1])

Overall, Sweden compares well internationally. Despite demographic drags such as ageing and an increasing share of foreign-born people in the population, residents of Sweden aged 16-64 scored third highest among OECD countries in literacy, just behind Finland and Japan and just ahead of Norway. Sweden has marginally improved its performance since 2012 in a context where most countries have backtracked (Figure 4.2).

Figure 4.2. Adult skills are high in Sweden
Copy link to Figure 4.2. Adult skills are high in Sweden

Sweden also ranks third highest of all OECD countries in the other two test items, numeracy and advanced problem solving. Test scores are highly correlated across these three items, reflecting that they measure broad cognitive ability rather than just reading and mathematics. In the literacy test for example, the test-taker has to read a text, understand it and reflect on its more or less complex content to respond to questions. The higher difficulty questions require considerable logic reasoning. For simplicity, figures and analyses in this chapter therefore focus on literacy skills. Comparable figures and results for numeracy and adaptive problem solving are for the most part available in OECD (2024[1]) and Pareliussen (2025[2]), and portray an almost identical picture.

4.2.1. Education improves literacy and labour market outcomes

As is and should be the case, education and training are important drivers of adult skills in Sweden and across the OECD (Figure 4.3, Panel A). The skills premium for having a tertiary degree compared to an upper secondary degree has fallen by seven points since 2012, and the difference is statistically significant. The skills premium for individuals participating in adult education and training has increased, although this may to an extent reflect reverse causality. At the same time, the share of the population with tertiary education has increased drastically over time to well above the OECD average (OECD, 2024[1]).

Figure 4.3. Literacy, education and training are reflected in earnings
Copy link to Figure 4.3. Literacy, education and training are reflected in earnings

2023

The likelihood of being in work is strongly correlated to literacy skills and educational attainment in Sweden (Figure 4.3, Panel B). This is a sign of matching efficiency, as these measurable indicators of skills are also likely good proxies of workers’ productivity (Pareliussen, 2022[4]). Compared to 2012, literacy is somewhat less strongly correlated with employment, while having a tertiary degree has increased significantly in importance (Pareliussen, 2025[2]).

Literacy skills, education and training are also strong predictors of earnings, conditional on employment. When controlling for a range of personal characteristics, type of contract and characteristics of the employer, a 50 points higher literacy score, which corresponds to one whole functional literacy level and is roughly equal to the standard deviation of literacy scores, is associated with 6% higher log hourly earnings in Sweden. People with upper secondary education earn approximately 7% more than those who have not completed upper secondary education, while tertiary graduates earn an additional 16% per hour on average. Returns to skills and education are lower than the OECD average, where an increase of 50 PIAAC score points is associated with an 8% wage premium, and tertiary graduates earn approximately 23% more per hour than upper secondary graduates (Figure 4.3, Panel C). The same analysis shows that returns to tertiary education have increased significantly from 2012 (Pareliussen, 2025[2]). Overall, the most striking changes over time relate to tertiary education, for which a weakening correlation to literacy may indicate a falling quality of tertiary degrees, a hypothesis which is discussed in more detail below, while returns to tertiary education in terms of employment and earnings have risen.

4.2.2. Older cohorts sharpened skills, while scarring from weak PISA scores is limited

Across the OECD, average skills are lower for older age cohorts. In Sweden in 2023 this was much less the case, implying that the life-long learning in a broad sense works better here than elsewhere. Skills peak when people are in their 30s in Sweden, and only fall gently by age thereafter. For comparison, skills peak when people are in their 20s in the OECD on average and fall much faster. Swedish 16-24 year-olds for example score significantly lower than the same age groups in Finland, Estonia, Japan and Norway (Larsson, 2023[5]). Comparing the two cycles of the OECD Survey of Adult Skills shows a marked improvement for Sweden’s older working-age cohorts. People aged 45 and above consistently scored higher on average in 2023 than people of the same age did in 2012, with statistically significant differences in the 50+ age groups (Figure 4.4, Panel A). This is a cohort effect reflecting that higher-skilled cohorts have become eleven years older while the less-skilled older cohorts of PIAAC 2012 are no longer in the sample. There is no evidence of improved skills over time within these cohorts, who score at the same level as they did in 2012 (Panel B). This is nonetheless a considerably better performance than in the OECD average, where literacy scores have fallen considerably across all age groups from 2012 to 2023.

Sweden experienced its first “PISA shock” in 2012. From a position well above the OECD average in all subjects (reading, mathematics and science) in 2000 and 2003, the performance of its 15-year-olds dipped below average in all subjects in 2012 (OECD, 2019[6]). This triggered a range of school reforms (see below), and performance improved considerably among test-takers in 2015 and 2018. However, almost all the gains were reversed in a second PISA shock in 2022, a fate Sweden shared with a number of OECD countries due to the disturbances brought to schooling by the Covid-19 pandemic (OECD, 2023[7]). Sweden did not close its compulsory schools, but experienced considerable disturbances due to absent pupils and staff and was unprepared for hybrid classes. Other factors such as the increased use of screens and social media as well as falling equity and an increasing number of low performers may also have contributed. Seen in this light, it is encouraging that the youngest age group in 2023 has higher average skills than people had at the same age in 2012, despite their schooling being disrupted by the pandemic (Figure 4.4, Panel A), although the difference in coefficient estimate is not statistically significant. Comparing the same cohorts over time, the cohort aged 25-29 corresponds roughly with the PISA 2012 cohort. It has increased its average PIAAC score by 16 points since 2012 and is thereby almost on par with 35-39-year-olds, the highest-scoring age group (Panel B).

Figure 4.4. Sweden’s life-long skill outcomes are strong compared to the OECD average
Copy link to Figure 4.4. Sweden’s life-long skill outcomes are strong compared to the OECD average

Changes in performance across age groups seem to be well-explained by the other background variables. In other words, changes in literacy scores from 2012 to 2023 by age group mostly reflect differences in the level of qualifications, incidence of adult education and training and personal characteristics like immigrant background and parent education between cohorts rather than the efficiency of life-long learning. The only exception is the oldest age group, indicating that life-long learning had improved for these cohorts compared to the cohorts of that age in 2012 (Panel C). Possible explanations include quality improvements to the educational system in the past and improved employment outcomes of older workers, as those working tend to build and maintain their skills more than the inactive.

The generational shift towards higher skills and educational attainment in the older age groups is also reflected in employment. The oldest age group has improved the likelihood of employment by almost 12 percentage points since then 2012 (Pareliussen, 2025[2]) and is no longer less likely to be employed than the reference group aged 35-44 when controlling for skills and backgrounds (Figure 4.5, Panel A). Earnings also typically rise by age, which is confirmed by PIAAC data. Younger age groups earn considerably less than older ones, with hourly wages peaking when people are in their 40s or 50s both in Sweden and the OECD. Controlling for differences in education, literacy and other characteristics, wages continue to rise throughout people’s lifetimes (Panel B).

Figure 4.5. Employment and earnings vary less by age in Sweden
Copy link to Figure 4.5. Employment and earnings vary less by age in Sweden

2023

4.2.3. Immigrant employment and wage gaps are largely explained by skills

The foreign-born generally score considerably lower on the literacy test than those born in Sweden. The gap is higher than the OECD average (Figure 4.6), reflecting that the average linguistic distance between Swedish and the mother tongue of the foreign-born is high in Sweden, and that the foreign-born population on average has less education and likely poorer-quality education than natives (as further discussed below). However, the magnitude of the difference varies considerably across backgrounds, with those born in Western Europe and North America on average scoring 23 points below the native-born, those from Central and Eastern Europe scoring 61 points below, and immigrants from the rest of the world scoring 83 points below. This corresponds to more than one and a half functional literacy levels below the native-born reference group. The foreign-born with the highest educational attainment from Sweden score 21 points higher than those with qualifications from abroad on average, partly compensating for this disadvantage (Figure 4.7, Panel A).

As discussed in more detail below, immigrants from outside of North America and Europe are less likely to be employed than people born in Sweden, but these differences are statistically insignificant when controlling for literacy, education and other background variables. In addition to their impact on literacy, qualifications from Sweden directly increase the employment likelihood for immigrants (Figure 4.7, Panel B).

Figure 4.6. First- and second-generation immigrants have lower literacy proficiency
Copy link to Figure 4.6. First- and second-generation immigrants have lower literacy proficiency

Literacy proficiency, 2023

There are also wage differences between the foreign-born and domestic-born. A first group of immigrants, born in North America and Western Europe, earn 9% more on average than people born in Sweden. This is a higher premium than the same group earned in 2012, but considerably less than the 33% premium this group earns on average across the OECD. The wage premium is well-explained by this group having favourable literacy skills, education and other background variables. A second group, born in Central and Eastern Europe, earn 12% less on average than people born in Sweden. This is about the same as in 2012, and similar to the OECD average. Contrary to 2012, the wage penalty is well-explained by the literacy skills and backgrounds of this group, which largely consists of free movement work migrants from the EU (Figure 4.7, Panel C).

The third group of immigrants encompasses any immigrant coming from outside of Europe or North America. While the other two groups largely consist of different tiers of work immigrants, this group has a high concentration of humanitarian and family reunion immigrants. The reason for immigration may itself put them at a disadvantage, as work immigrants often respond to concrete job opportunities while humanitarian immigrants respond to the necessity to leave their countries of origin. Another reason for disadvantage is that their skills may be lower than those of natives, either in absolute terms due to lower educational attainment or degrees obtained from lower-quality schools and universities, or their skills may be high in their origin-country setting but mismatched to skill demand in Sweden. The latter category includes for example weak fluency in Swedish, a lack of formal recognition of degrees obtained abroad or imperfect understanding of the Swedish-specific cultural and labour-market context (Pareliussen, 2017[8]; Bussi and Pareliussen, 2017[9]). They may lack proof of qualifications because they left in a hurry, and they may have had a long period outside the labour market if the journey was long. This group of immigrants includes many who lack any type of formal education on arrival even though differences based on country of origin and gender can be large. Among immigrants in the PES Establishment programme who arrived in 2016-17, around 40% of those from Afghanistan and 50% of those from Somalia lacked formal education compared to 10% of those from Syria (Arbetsförmedlingen, 2024[10]).

Figure 4.7. Immigrants’ earnings and employment gaps reflect weaker literacy
Copy link to Figure 4.7. Immigrants’ earnings and employment gaps reflect weaker literacy

2023

On average, immigrants coming from outside of Europe or North America earn 17% less than people born in Sweden. This is about the same as in 2012, but very different from the OECD average where this group does not display a significant earnings penalty, likely due to a large extent to different group compositions and different country-specific contexts (Pareliussen, 2025[2]). Contrary to what was the case in 2012, the wage penalty is insignificant when controlling for the literacy skills and backgrounds of this group (Figure 4.7, Panel C). In other words, the composition of immigrants coming from outside of Europe or North America who are in employment has become less favourable, while their earnings are on par with the earnings of natives with similar backgrounds and literacy. These are encouraging signs that more marginal workers in this group are increasingly becoming employed and receive fair wages.

4.2.4. Gender gaps in employment and earnings remain

Gender and family characteristics are reflected in skills. Single women tend to have lower literacy skills than the reference group, which is a man in a couple. The largest difference to the reference group is found among mothers of young children, who score 14 points below on average (Figure 4.8, Panel A). The coefficient estimate is significant, although the considerable increase from 2012 is not.

Figure 4.8. Gender gaps in employment and earnings remain
Copy link to Figure 4.8. Gender gaps in employment and earnings remain

2023

This finding is not related to immigration, as the pattern is the same among the native-born as the foreign-born population. Even though absences from work in connection with childbirth can deteriorate skills, it does not seem plausible that the effect would be so large and that it would increase from 2012 to 2023. A more plausible explanation, consistent with the increasing returns to education and falling fertility rates observed over the past few years, is that growing returns to skills have led high-skilled women to increasingly prioritise career over family. Further supporting this hypothesis, falling fertility seems to be driven by women with tertiary education. In Sweden, average fertility is higher among women with tertiary education than among women with less education, but fertility among tertiary-educated women has fallen consistently since 2010, contrasting close to unchanged fertility for women with less education. The tension between career and family is also illustrated by the finding that working mothers of young children are likely to be overqualified for their jobs. Even in Sweden, if one parent has to sacrifice career progression for family reasons, this tends to be the mother, who can either put her career on hold while children are young or choose a less demanding career path which is more compatible with family (OECD, 2025[11]).

Men are more likely to be employed than women, and employed men earn more than employed women on average. Women in couples earned approximately 14% less per hour than men in 2023, controlling for literacy and a range of individual and employer characteristics. Single women earned approximately 7% less than single men. Gender gaps are relatively narrow in Sweden compared to the OECD, where women in couples earned approximately 21% less than men in 2023, and single women earned approximately 13% less than single men (Figure 4.7, Panel B and C). Even so, these adjusted gender gaps have not budged since 2012, indicating that progress towards gender equality has been slow over the past decade (Pareliussen, 2025[2]). Women remain under-represented on private company boards, in senior management positions, in many well-paid and influential professions and among entrepreneurs (Entreprenörskapsforum, 2023[12]; Ekonomifakta, 2025[13]).

No significant direct employment- or earnings penalties of having children below the age of three are found in Sweden (Figure 4.7, Panels B and C). Even though not directly attributable to having children aged below three, the gender gaps in employment and earnings under consideration are closely related to motherhood. For example, Åslund, Karimi and Sundberg (2025[14]) find that income from employment falls around 40% at the time of childbirth, and only recovers gradually and partially over a ten-year period. This persistence long after the child has grown past infancy may explain why motherhood penalties as defined here are not statistically significant, since women with older children are part of the reference group.

Gender gaps in earnings and employment are not very sensitive to immigrant background. When controlling for literacy skills, education and the other covariates, these gaps are not significantly higher for the immigrant groups (Pareliussen, 2025[2]). This is in line with findings by Åslund, Karimi and Sundberg (2025[14]). Even though they do find statistically significant differences in motherhood penalties between women originating from the least gender-equal countries and their native counterparts in detailed registry data, differences are small in magnitude and overshadowed by between-group similarities. Besides, Sweden is a rare case of a country where there are also fatherhood penalties, particularly pronounced among men in more gender-egalitarian households (proxied by the father’s share of parental leave) and among fathers who have sons relative to daughters (Sundberg, 2024[15]). This fatherhood penalty is similar among native and foreign-born fathers, regardless of gender equality in their origin countries (Åslund, Karimi and Sundberg, 2025[14]).

4.3. Policy directions to sharpen skill supply in Sweden

Copy link to 4.3. Policy directions to sharpen skill supply in Sweden

The combination of high adult skills, peaking later in life than in the OECD on average and deteriorating less by age, is likely related to a combination of a school system where people learn curiosity and how to access new knowledge (“learning to learn” as opposed to “learning by heart”). Generally good employment outcomes for all age groups also help, as high skills increase employment prospects, but employment also builds and maintains skills. A horizontal working culture encouraging independent thinking and facilitating learning by doing, and not least continued investments in skills are additional factors. Notably, the incidence of job-related adult education and training peaks around the age of 50 in contrast to the OECD average of around 30 (Figure 4.9). There is nonetheless scope for policy reform to further sharpen the skills and productivity of the majority of Swedes who are already doing well by improving the regular initial education system and by ensuring access to relevant adult education and training.

Figure 4.9. The incidence of adult education and training is high in Sweden
Copy link to Figure 4.9. The incidence of adult education and training is high in Sweden

2023

4.3.1. Rationalising coordination within the skill supply system

The supply of skills is an inherently complex affair in any OECD country, where a range of actors, public and private, are involved in formal education, informal learning activities and on-the-job-training. Sweden has a decentralised skills system, with ministries responsible for policy making and overall direction and governmental agencies and sub-national governments (regions and municipalities) responsible for implementation and delivery. In Sweden's governance model, agencies operate under collective government decisions, and they have operational independence in how to fulfil their mandates. This means that individual actors are closer to the issues and can be more responsive and adapt to local needs. On the other hand, it can lead to redundancies due to overlapping mandates and it increases the risk that individual actors optimise their operations based on their own narrow mandate and interests, resulting in a sub-optimal overall outcome (OECD, 2019[6]; OECD, 2024[16]).

In order to better structure coordination within the wider skills supply system and thereby make the most of decentralisation while minimising redundancies, the Swedish government in 2022 tasked seven national agencies, all with some responsibilities for various aspects of skills policy, to form the Inter-agency Co-operation Structure for Skills and Lifelong Learning (Myndighetssamverkan för kompetensförsörjning och livslångt lärande, MSV). MSV could be further developed into a hub for the coordination of Sweden’s skills system. A first step would be if participating agencies formalised MSV as the main forum for promoting inter-agency co-ordination on skills policy by means of a joint statement of mandate, roles, responsibilities and functions of the MSV. The existing organisational structure could also become more nimble by replacing a majority of current working groups by outcome-orientated, time-limited task forces (OECD, 2024[16]).

4.3.2. Reforming grades and national tests can shore up compulsory education

The Swedish system for initial education starts with preschool (förskola), which is voluntary and available for children aged one to five years. Compulsory education begins with reception class (förskoleklass) at age 6 followed by compulsory primary and lower secondary school (grundskola) from ages 7 to 15 provided by municipalities or approved private providers. Municipalities organise after-school activities (fritidshem), up to sixth grade (OECD, 2024[16]).

Following the “PISA shock” in 2012, substantial analytical work was carried out, identifying reform priorities which largely fit into three categories: i) funding and governance: better coordinating or even partially recentralising funding and governance to ensure that municipalities across the country can provide pupils with the resources and skilled personnel necessary to secure learning regardless of individual backgrounds; ii) school competition and choice: steering school competition and school choice to reduce school segregation and deliver for the public good and ensuring that grades fairly represent pupils’ skills and knowledge, and; iii) teaching profession: making it more attractive to be a teacher by strengthening the research base and teaching practice during initial education, clearer career paths and more cooperation, feedback and support between colleagues, coupled with clear accountability for key outcomes (OECD, 2019[6]).

A number of reforms were implemented in line with these recommendations, notably with the opening of six regional offices by the National Agency for Education as part of a new “quality dialogue” to improve how schools are run. The “Boost for Mathematics” and “Boost for Reading” programmes carried out from 2013 and 2015, respectively, were capacity-building programmes aimed at improving teaching in the respective subjects based on collegial lesson preparations and evaluations, supervised by an experienced coach. Both programmes contributed to raising results in their respective subjects (Grönqvist, Öckert and Rosenqvist, 2021[17]; Holmlund, Häggblom and Lindahl, 2024[18]). Grants have been put in place to target funding towards schools with an overweight of pupils from disadvantaged groups, and to provide career pathways through the “First teacher” reform (Statskontoret, 2021[19]). The reform seems to have been successful in promoting high-wage, highly skilled teachers, lowering teacher separations and improving student performance (Grönqvist, Hensvik and Thoresson, 2022[20]).

Sweden has so far lacked an objective benchmark of pupil performance as the national tests have been graded by their own subject teachers. In a system of school choice where funding follows the pupil, lenient grading can be one strategy to attract pupils. This lowers the average quality of schooling and allocates students unfairly and sub-optimally to higher education. External scoring of national tests by trained evaluators is set to commence in 2026. This will create an objective benchmark of school results enhancing the scope for effective governance and fair competition between school providers. It should be noted that an objective measure of results is not the same as a measure of value added at the school level, but is a necessary starting point for creating accurate value-added measures as well. A government investigation has proposed changes to the grading system, including weighing high and low grades symmetrically, suppressing the requirement to pass in certain subjects to enter upper secondary education, and adjusting grades at the school level based on the school’s average national test result (Regeringskansliet, 2025[21]). The proposed changes to the grading scale and suppression of the requirement to pass certain subjects will make the grading system fairer and more efficient and should be implemented without delay. Adjusting school grades based on national test grades also seems warranted to remove lenient grades as a way to attract students and funding, but there may be practical issues to resolve in the implementation of this proposal. For example, safeguards need to be put in place to ensure against schools asking weak performers to stay at home on the test day, and national tests can end up having outsized effects on final grades in schools with few pupils.

More could be done to compensate for socio-economic disadvantage. Compulsory and upper secondary education is provided both by municipalities and by private for-profit or non-profit providers (“independent schools”). All schools are fully funded by the municipality. When establishing new independent schools or expanding capacity in existing ones, expected impacts on the diversity of the student mix within the schools in the vicinity should be taken into account. School intake to private schools remains at the discretion of the school provider, most of which allocate school places on a first-come-first-serve basis. The resulting queuing system is unfair because it favours pupils from strong socio-economic backgrounds, and it discriminates against internal and external immigrants. Assigning school places in over-subscribed schools by lottery as in charter schools in the United States (Box 4.2), or with quotas reserved for pupils with unfavourable socioeconomic backgrounds would help (OECD, 2019[6]).

Box 4.2. School choice and public funding of private schools in other OECD countries

Copy link to Box 4.2. School choice and public funding of private schools in other OECD countries

The Swedish system is unique in the OECD context. Many countries have a large share of private schools dependent on public funding, but most other OECD countries limit dividends and/or limit entry of for-profit school providers. Only Sweden has a large publicly funded for-profit sector after Chile ended public funding of for-profit schools on equity grounds with the 2015 Inclusion and Equity Law. The private school sector is growing strongly in Sweden, and growth is almost exclusively in the for-profit segment.

The Czech school system opened up for private provision in the early 1990s as in Sweden, and a high share of private schools in Czechia depend on public funding as a result. However, entry criteria for schools receiving public funding were tightened in a 1995 law. They are not allowed to be run for profit, and public funding does not cover investment expenditure.

Denmark has a relatively large private school sector, funded with a voucher system resembling the Swedish one. However, publicly-funded Danish private schools are non-profit, they receive only slightly above 70% of the average cost per pupil in public schools, and they are allowed to charge tuition fees.

The Netherlands is another country with long traditions of private school provision. Entry of private providers is liberal, and private schools are funded equivalently to public schools, but in contrast to Sweden, only non-profit providers receive public funding.

Charter schools at the sub-national level of the United States bear strong resemblances to the Swedish system. 41 states and the District of Columbia permit charter schools to operate. Charter laws vary across states, but defining characteristics common to the Swedish system are that that they cannot charge tuition fees, and they are not permitted to impose discriminatory admission requirements. A key difference to Sweden is that if charter schools are over-subscribed, they must select pupils by lottery. Even though charter schools do not outperform public schools on average when controlling for pupil backgrounds, many charter schools, and charter school districts (such as New York City) appear to outperform traditional public schools according to various measures, and individual schools and the sector as a whole seem to improve over time. Several states have school voucher programmes or tax credits to offset the cost of attending private schools.

Source: OECD (2019[6]).

The sum of reforms seems to have made a difference, even though some important reforms remain in the pipeline. Sweden’s PISA results improved until 2018 but dropped more than the OECD average in 2022, largely reversing the gains made since 2012. How much of this was caused by Covid-19 remains an open question. Sweden did not close its compulsory schools, but there were considerable disruptions due to absences of teachers and pupils, and hybrid teaching was not well-developed. The fact that the cohorts around PISA test-taking age in 2012 have caught up with other cohorts both in terms of skills and labour market outcomes should provide some solace, and the share of youth not in education, employment and training (NEETs) remains low. However, there are indications that those close to graduation from compulsory school and upper secondary education around the time of the Covid-19 pandemic struggle to fulfil the requirements to enter university and vocational education and training. This could lead to permanently lower skills and labour market scarring of those concerned. Systematically mapping individuals concerned and following them up either in the regular school system or in municipal adult education might help. However, there is a risk that those who do not seek out existing opportunities to improve their results on their own accord may be unreceptive to such efforts.

4.3.3. Upper secondary vocational education can learn from higher vocational education

Students are in upper secondary school (gymnasieskola) from ages 16 to 19. There are 18 national programmes, 12 vocational programmes and 6 preparatory programmes for higher education. Upper secondary school is, like compulsory school, organised by municipalities and private providers. A fundamental problem with upper secondary VET that Sweden shares with many OECD countries is that it tends not to be students’ preferred choice. Only one third of pupils are enrolled in the VET track (OECD, 2024[22]). This is well below the EU average, while strong employment outcomes of VET graduates indicate that their skills are in demand. The employment rate of VET graduates (aged 20-49) is 88.3% (2021), 3 percentage points above the EU average, 5.2 percentage points higher than for graduates from general upper secondary education and 19.4 percentage points higher than for graduates with lower-level qualifications (2023) (Eurostat, 2023[23]).

A 2011 reform reduced the academic content in VET programmes as the default option making their students ineligible for tertiary education unless they made an active choice to include optional extra Swedish and English classes. It thereby narrowed the career options of VET students and reinforced a pre-existing inclination of students, especially those with relatively strong grades, to turn to the higher education preparatory programmes. As of 2023, academic content giving general eligibility to enter tertiary education was reinstated as the default in all programmes with the possibility to opt out. This has not increased interest in VET-programmes significantly so far. It is uncertain if this policy reversal will be able to reduce the stigma attached to VET programmes and thereby increase enrolment going forward.

Despite extensive voluntary cooperation between municipalities, the funding model of VET has created problems. The municipality of residence is as a rule obliged to pay the municipality or private provider where their resident pupils are enrolled an amount calculated on the basis of actual costs or the cost of organising the same programme in public schools in the municipality of residence. These rules incentivise municipalities to provide programmes that are less costly per student, and they leave considerable room for interpretation with relatively frequent litigation over funding as a result (OECD, 2019[6]; Gustafsson, Sörlin and Vlachos, 2016[24]).

Labour markets are regional, and fostering a good match between the supply of skills and what is demanded by employers calls for national and regional coordination. Students entering upper secondary school take an active part in the school choice decision while still young and with imperfect knowledge of labour market needs, increasing the likelihood of making choices they will come to regret. The PIAAC data shows that field-of-study mismatch is closely negatively correlated to the age at which a qualification was obtained (Pareliussen, 2025[2]). Contributing to mismatches, there have been signs of over-establishment of some programmes in demand by students but not by employers, and matching from upper secondary VET to jobs is very variable between programmes (SCB, 2018[25]). This calls for better coordination of the VET programmes on offer at the regional and national levels, including more active steering of the establishment of private schools, as well as strengthening career guidance (SOU, 2017[26]; OECD, 2019[6]). The strengthened regional presence of the National Agency for Education may facilitate regional cooperation, and the Inter-Agency Cooperation Structure for Skills and Lifelong Learning may also contribute to better coordination between different actors in the skills supply system on a more general basis, as discussed above. Recent legal amendments clarifying the responsibilities of municipalities to cooperate in planning and dimensioning of upper secondary education and clarifying that the programmes they decide to offer should be based on regional job market needs are further steps in the right direction. However, privately-run schools, which are financed by municipalities based on student numbers and follow the same curricula, are not required to participate in this cooperation. The risk that municipal investments in a better-matched VET offer is undercut by private providers establishing cheaper programmes which are superficially attractive to young students but with less demand from employers, thereby undermining incentives to invest in the more expensive VET programmes.

Sweden, like many other OECD countries, including Denmark, Finland and the Netherlands, offers post-secondary vocational education. In Sweden’s successful system, a dedicated agency funds vocational programmes with a clear labour market need for a time-limited period upon application from municipalities and private providers. This model is currently being piloted in adult education at upper secondary level. Sweden should pilot this system also for upper secondary vocational education. Following close evaluation of these pilot programmes and the impact of recent reforms on the quality and popularity of upper secondary VET programmes, Sweden should consider transitioning its entire upper secondary VET system into this model. This could reduce coordination problems and strengthen the connection to labour market demand. The latter should not come at the expense of general skills and the syllabus should continue to give eligibility for tertiary education as the default option.

In Sweden, higher vocational education is offered by Yrkeshögskolan (YH), which replaced qualified vocational education (kvalificerad yrkesutbildning) around the early 2010s. Students need an upper secondary diploma or can be accepted based on an assessment of their existing competencies. Programmes are mostly offered for one or two years, leading to formal qualifications at the post-secondary (ISCED 4) and tertiary (ISCED 5) levels. Shorter courses of up to six months aimed at meeting labour market needs are also provided (Eurydice, 2024[27]).

Programmes are required to correspond to real labour market needs and are delivered in close cooperation with employers and industry, meaning that the range of programmes and specialisations changes over time, with new programmes starting and old ones discontinued within a maximum of seven years and five first-year student intakes as the labour market evolves. Private education providers deliver a bit more than half of the programmes but enrol approximately 80% of the students. Municipalities deliver almost half of the programmes but enrol only about 20% of the students. A limited number of higher education institutions also offer higher vocational education. Around 80% of student places are within the five fields of economy, administration and sales, health care and social work, construction and spatial planning, technology and manufacturing, and information and communication technologies (MYH, 2024[28]; Eurydice, 2024[27]).

The number of study places has roughly doubled since 2013, and 35 000 spots were available to new students in 2023. Following the rapid expansion, outcomes have dipped slightly but remain strong. There are approximately two qualified applicants for each place. Among accepted applicants, 13% lack formal qualifications but are accepted based on an assessment of their existing competencies. The average age of students has been trending up to reach 31 years in 2023. The share of foreign-born students has increased, largely matching the share of foreign-born in the total population. The total student mass is gender balanced. Traditional gender roles still colour how students are distributed in the different programmes, but much less so than in upper secondary vocational education, with for example 33% women studying construction and spatial planning in Yrkeshögskolan in 2023 compared to 12% attending the construction programme in upper secondary VET (Figure 4.10).

The share of women has been increasing in construction and spatial planning as well as in information and communication technologies over the past five years (MYH, 2024[28]). Yrkeshögskolan is widely seen as a very successful model. Almost 67% of students graduate, 8% fail and the remaining 25% do not complete their studies. Approximately 90% of graduates get a job during the year after graduation (MYH, 2024[28]).

Figure 4.10. Educational choices are less gender-biased in higher vocational education
Copy link to Figure 4.10. Educational choices are less gender-biased in higher vocational education

2023

4.3.4. The quality of tertiary education has likely declined over the past few decades

Higher education is available to those who have completed upper secondary school. It is provided through higher education institutions including universities (universitet) and university colleges (högskola). Sweden has both public and private higher education institutions. Higher education institutions in Sweden enjoy significant autonomy within the framework of national legislation. They independently decide on aspects such as the content of study programmes, most admission criteria for students and areas of research focus. However, they must also be responsive to student demand and the requirements of the labour market (OECD, 2024[16]). All higher education institutions are to have systems in place for quality assurance. On a national level, higher education quality is monitored by the Swedish Higher Education Authority. The national quality assurance system for higher education and research consists of four components: institutional review of education and research, programme evaluation, appraisal of degree-awarding powers and thematic evaluations (Higher Education Authority, 2024[29]).

The number of students in higher education in Sweden has increased significantly since the mid-1990s, as higher education has become available to wider segments of the population (Figure 4.11). A concern which is raised from time to time in Sweden is that quality may be suffering when an increasing share of youth pursues tertiary education. Increased enrolment may reduce the teaching resources per student, and the average ability at the time of entry may have become weaker when entry has become less selective, a concern accentuated by falling school results as measured by PISA. Another concern raised is that even though the funding formula is partially based on society’s skill demands, payments per student enrolled and per student graduating also incentivize universities to provide less labour market relevant degrees if they are in high demand among students, and to ensure as many students as possible graduate, which could come at the expense of quality.

Figure 4.11. Educational attainment has increased over time
Copy link to Figure 4.11. Educational attainment has increased over time

There may be some truth to these concerns. As discussed above, the literacy premium associated with having a tertiary degree has fallen somewhat from 2012 to 2023. Investigating this further, literacy skills among tertiary graduates in 2020-23 are on average 20 score points lower than among those who graduated a decade earlier, contrasting with the OECD average, which has remained broadly unchanged (Figure 4.12, Panel A). Controlling for potential composition effects leads to a similar conclusion, with the average skills of tertiary graduates having seemingly peaked in the 1990s in Sweden, while the literacy of tertiary degree holders remained roughly constant in the OECD on average over the period (Panel B). It should be noted that due to large standard deviations and a limited sample size, the differences between cohorts are not statistically significant, and should only be seen as indicative evidence of when the value added of a tertiary degree peaked. However, the fall in literacy associated with a tertiary degree from 2012 to 2023 is statistically significant. Qualitative evidence supports this finding. Even so, Sweden invests considerable resources per tertiary student compared to the OECD average and the high overall level of skills among Sweden’s adults indicates that its education system overall delivers high quality.

Supporting the hypothesis that the value added from a tertiary degree has decreased, newer universities, which account for a considerable share of the expansion of tertiary education in Sweden, generally have students with less prior knowledge and weaker social backgrounds while older universities generally have a higher proportion of students with high admission scores and who also have parents with higher qualifications. Institutions with more students with higher needs per teaching staff member tend to lower standards by revising the content of the education and increasing the use of teaching activities and examinations that require fewer resources. In these conditions, teachers feel inadequate as they know what is required for high-quality education but lack the ability to act accordingly (Brommesson, Nordmark and Ödalen, 2024[30]). Only half of engineering students graduate, while resources spent are only about 60% of what they were in the 1990s. This leads to reduced teaching hours compared to before and compared to other countries like France and Switzerland (Sveriges Ingenjörer, 2023[31]). It should be noted that engineering is not a regulated profession in Sweden. A share of those who do not graduate may therefore still work as engineers, and in times of high demand for skilled engineers, students may be offered employment before they complete their degree.

Figure 4.12. The quality of tertiary education may have declined
Copy link to Figure 4.12. The quality of tertiary education may have declined

Despite potentially lower educational quality, the employment and wage premiums for tertiary graduates have increased over the same period in Sweden. Employment and wages reflect productivity, which depends not only human capital, but also how efficiently human capital is employed and shifting demand for the products and services produced across occupations. Such market factors may have increased the demand for skills, raising the employment and wage premiums associated with a degree, even as average skills conditional on a degree have edged down. Hensvik and Skans (2019[32]) found for example that occupations requiring a tertiary degree were on the rise in Sweden from 2001 to 2013. Research has also found that non-cognitive skills such as the ability to interact with and lead others is a strong predictor of labour market outcomes in Sweden (Edin et al., 2022[33]; Lindqvist and Vestman, 2011[34]).

Properly mapping and understanding the causes of the possible decline in the quality of tertiary education is important to continue to grow productivity by innovating and adopting new technologies and best practices. A likely partial cause is the falling ratio of teaching personnel per student, which also creates a time conflict between research and teaching (Brommesson, Nordmark and Ödalen, 2024[30]). This could in addition to reducing the quality of tertiary degrees hold back Sweden’s ability to innovate and swiftly internalise innovations at the global scientific frontier.

Easing this time conflict would require additional funding to universities, either by increasing appropriations or by allowing universities to charge tuition fees. As demonstrated above, the private gains associated with a tertiary degree are substantial, while studying in Sweden is free of charge for citizens of the EU/EEA and those with permanent residence. The principle that education should be free is deeply rooted in Sweden, and it can also be argued that progressive income taxation is a tax on the private benefits of higher education. Encouraging EU/EEA students to stay in Sweden after graduation would retain a larger share of the investment cost of their education. For others, annual tuition fees typically range between SEK 80 000 and 140 000 (USD 9 000 - 17 000). Allowing universities and university colleges to charge tuition fees is an alternative way to boost university funding. To preserve equal access to education, fees should be capped and fully compensated by increasing student loans. Additional needs-tested grants or scholarships could also be considered, although such systems are typically more complex to implement than the current universal system in Sweden. Student support could also be used more actively to address skills shortages by giving more support to students enrolled in priority fields like teaching, engineering and nursing (OECD, 2022[35]).

Across OECD countries, including the Nordics, universities play a major role in regional growth and labour supply. They stimulate the creation of knowledge networks, often support entrepreneurship and young firms, and help build and retain local skills. This generates business dynamism, capacity to absorb knowledge, and helps attract and retain skilled workers and companies, thereby raising productivity. Korean technoparks and regional clusters in the Netherlands offer examples of successful policy-led cooperation between a wide range of stakeholders to foster regional innovation and growth and increasing the resilience of the economy to industry-specific external shocks (OECD, 2021[36]).

Universities play a key role in providing a skilled workforce adapted to local labour market needs. The importance of this function is underscored by regional and sectoral labour shortages of for example teachers, engineers and care personnel discussed below. Research and education fields matching local labour market needs are particularly relevant to foster innovation and retain graduates, who are increasingly leaving their place of education after completing their studies, especially if it lies outside the larger metropolitan regions. For example, more than 70% of Umeå university students left after graduating in 2013. Further efforts to translate research into innovation especially by raising lower-performing institutions to the levels of their higher-performing peers would require deepening the involvement of tertiary education institutions in local innovation networks, in particular through enhanced incentives and support mechanisms. Further facilitating the commercialisation of research and synergies with EU research programmes could also help (OECD, 2021[36]).

4.4. Matching, skill deficiencies and adult education and training

Copy link to 4.4. Matching, skill deficiencies and adult education and training

As mentioned above, the likelihood of being in work is strongly correlated to literacy skills and educational attainment, which is a sign of matching efficiency. However, room remains to improve the matching of people in employment. For example, a share of people with high capabilities are working in mundane jobs. Furthermore, severe labour shortages exist side-by-side with a pool of long-term unemployed or inactive people potentially willing to work. The 2007-08 global financial crisis was a negative shock to matching efficiency, which kept falling from 2011 to 2017 due to high immigration from outside of Europe of people with weak competitiveness in the Swedish labour market. This latter group improved their competitiveness with their time of stay in Sweden, which improved matching efficiency from 2017 to 2022 (Böhlmark and Waisman, 2024[37]; Arbetsförmedlingen, 2024[38]). The diverse offer of adult education and training options in Sweden can both reduce mismatches and reduce the negative consequences of mismatches (Box 4.3). Most of these options are publicly financed, and various income supports are available to those embarking on studies or training. There is still room to better align the resources invested in each individual to the expected societal value created.

Box 4.3. Adult education and training in Sweden

Copy link to Box 4.3. Adult education and training in Sweden

A wide range of adult education and training options are available in Sweden to cover diverse needs.

Universities and university colleges are important providers of adult education and training. The Swedish higher education system is relatively flexible, with largely course-based education ensuring modularity and flexibility and many higher education institutions offering distance education options that allow students to complete their studies entirely online. They also provide courses designed for continuous professional development or personal enrichment that may not necessarily lead to formal higher education qualifications.

Higher vocational education (Yrkeshögskolan) is also an integral part of adult education and training and an important arena for up-skilling and re-skilling. Students are often people with work experience. A large share of students (74%) were already working before entering in 2023. The two most common reasons for these to enter Yrkeshögskolan were to change career, and to improve their career prospects within their current field of study.

Municipalities are obliged to offer adult education. Municipal adult education (kommunal vuxenutbildning, komvux) provides education for adults at compulsory and upper secondary levels. In 2020, the prioritisation rule was changed from giving priority to individuals with lower levels of education to prioritising people with the highest needs, including those who are unemployed or in need of re-skilling. Municipal adult education includes upper secondary vocational adult education (yrkesvux). It also comprises adult education for persons with intellectual disabilities (komvux som anpassad utbildning) and language classes for newly arrived immigrants (Svenska för invandrare).

Folk high schools (folkhögskola) serve two main functions in the adult education system. They provide an alternative to municipal adult education, as they are required to offer those who have not completed compulsory or upper secondary education courses designed to provide basic eligibility to higher education. Folk high schools are also an important pillar of Sweden’s liberal adult education (folkbildning), which offers non-formal adult education, not tied to academic qualifications. Courses range across several fields such as arts, crafts, music, theatre, social sciences, and vocational training. Since 2010, folk high schools have also offered motivational study courses targeting jobseekers who neither have a high school diploma nor basic eligibility for higher education. Folk high schools attract a relatively large number of students with disabilities and employ flexible study methods to accommodate their needs.

Nine independent study associations (studieförbund) make up the other main provider of liberal adult education. Study associations offer three types of courses: study circles, cultural events and “other folkbildning”, which are open to everyone. The Swedish National Council of Adult Education (Folkbildningsrådet), a non-governmental organisation, is responsible for distributing government grants to folk high schools and study associations.

Labour market training (arbetsmarknadsutbildning, AUB) is overseen by the Swedish Public Employment Service (Arbetsförmedlingen, PES) and is intended primarily for unemployed individuals in need of reskilling or up-skilling. PES focuses on providing training to those who are far from the labour market and the long-term unemployed, and procures the training programmes from external suppliers. It is also responsible for developing and implementing individual integration plans for immigrants, which includes coordinating with other skill providers, such as municipal adult education who are responsible for Swedish for immigrants.

Source: Regeringskansliet (2020_[39]); OECD (2015_[40]); OECD (2024_[16]); MYH (2024_[28]).

4.4.1. Mismatches among people in employment are around the OECD average

Good matches between workers’ skills and qualifications and those required by their jobs are needed to make full use of Sweden’s high skills to the benefit of individuals and society. Workers who are well-matched to their jobs tend to earn higher wages and be more satisfied. At the aggregate level, mismatches lower productivity as countries do not fully reap the returns from their investments in human capital (OECD, 2024[1]; Adalet McGowan and Andrews, 2017[39]). Some degree of mismatch in the economy is inevitable as workers sort themselves into suitable jobs. This is particularly true for younger workers or those (re-)entering the labour force. Some people may choose to work in jobs that do not match their skills, qualifications or field of study due to factors such as personal preferences, location or family responsibilities.

The consequences of workers being mismatched vary by type of mismatch. The most harmful mismatches from society’s point of view are those preventing people from utilising their full potential, either by keeping them out of employment, or by keeping people in jobs where they do not use their skills to their full potential. Cross-country evidence from the OECD Survey of Adult Skills indicates that the shares of workers who are underutilised by either being over-qualified or over-skilled in Sweden are around the OECD average (Figure 4.13, Panels A and B).

Figure 4.13. Skills and qualification mismatches are around the OECD average
Copy link to Figure 4.13. Skills and qualification mismatches are around the OECD average

When workers are over-qualified or over-skilled for their jobs it will not only have a macroeconomic cost, it will also come at a personal cost, as they will on average have lower wages and be less satisfied with life than well-matched ones (OECD, 2024[1]). Across the OECD, the biggest earnings penalties are associated with overqualification mismatch. In Sweden, such mismatch is associated with 11% lower earnings than well-matched workers, compared to 21% for the OECD average. People who are over-skilled earn 8% less on average than well-matched workers in Sweden. This is a slightly larger wage penalty than the OECD average of 6%, although the difference is insignificant (Figure 4.14).

An over-skilled or over-qualified worker is well-matched to the employer’s needs, so when employers in Sweden and elsewhere regularly report difficulties in recruiting people with the right skills for the job they are referring to a different type of mismatch. When people who are under-qualified or under-skilled are hired this is only a societal cost if they crowd out better-matched individuals. Companies compete for capital and talent, and the less productive employers will be less capable of offering an attractive package to skilled workers. This is driving productivity-enhancing structural change. The workers concerned and society may benefit when people who do not at the outset possess the asked-for qualifications and skills are hired and build their human capital through experience and on-the-job learning. The share of under-qualified workers in Sweden is 18%, the second highest in the OECD. On average, they earn 6% more than well-matched workers, compared to 12% in the OECD average (Figure 4.14). Underqualified and under-skilled workers display slightly lower life satisfaction in Sweden and on average across OECD countries. This is likely related to feeling inadequately prepared to do their jobs in cases where they are not receiving sufficient training to develop their skills (OECD, 2024[1]).

Figure 4.14. The consequences of mismatches vary depending on their nature
Copy link to Figure 4.14. The consequences of mismatches vary depending on their nature

Earnings per hour by type of mismatch, difference to well-matched workers, 2023

Finding the right talent for the job is not only about the level of education and skills. It is also about finding persons with specific skills that are well-matched to the job. Such skills can be acquired through education and training and through experience. Among the population in employment, the share of adults with their highest qualification in a domain which is mismatched to their current job (field-of-study mismatch) is among the lowest in the OECD (Figure 4.15). This may seem at odds with reported shortages of workers in professions in high demand such as nurses, teachers and engineers (OECD, 2023[3]), but these macro-level shortages reflect ageing, rural depopulation and too weak overall attractiveness of working in some public services, likely related to monopsony power among employers as further discussed below.

Figure 4.15. Field-of-study mismatches are prevalent across the OECD
Copy link to Figure 4.15. Field-of-study mismatches are prevalent across the OECD

Around 30-50% of workers are mismatched in OECD countries (Figure 4.15), but these field-of-study mismatches do not automatically have negative effects. In the best of worlds, such mismatches can reflect that skills are transferable, that people can pursue their talents and interests even if they are not fully aligned with their highest qualification, and that employers are able to attract the best persons to do the job, irrespective of formal qualifications. However, they can also result from people educating themselves into saturated occupations or from structural change making their qualifications obsolete. For example, 30-35% of people made redundant in Sweden find a new job in a different occupation. This typically leads to a loss of income at the individual level, but contributes to shortening unemployment spells or eliminating them altogether (Fredriksson et al., 2023[40]). In its appropriation letter for 2024 and 2025 the Government has tasked the PES to foster increased occupational and geographical movement. As a consequence, the PES increasingly requires unemployed jobseekers to apply for jobs outside of their current occupation and area of residence to maintain unemployment benefit eligibility. If successful, this adjustment would dampen unemployment and geographical mismatch, but would also increase field-of-study mismatch.

When people are forced to accept job offers in occupations for which they have less qualifications, skills and interest they become over-qualified, over-skilled or both, with the negative effects for individuals and society mentioned above (Montt, 2015[41]). In Sweden, as in the OECD on average, those with a mismatched field of study are much more likely to also be over-qualified (Pareliussen, 2025[2]). The earnings penalty to field-of-study mismatch is weak and statistically insignificant in Sweden, contrary to the OECD average (Figure 4.14) and people who are field-of-study mismatched report slightly higher levels of life satisfaction that well-matched individuals (OECD, 2024[1]), indicating that the average harm resulting from this kind of mismatch is low.

4.4.2. People with weak education or skills are poorly matched to available jobs

The overall distribution of skills is relatively egalitarian in Sweden. Only about 14% of the adult population do not have an upper secondary education diploma, which is lower than the 19% OECD average. There is a strong link between formal qualifications and general adult skills such as those measured in the OECD Survey of Adult Skills, and the country has the second-lowest share of individuals with very low functional skills across all three domains after Japan and just above Finland and Norway (Figure 4.16).

Figure 4.16. Few adults have very weak skills in Sweden
Copy link to Figure 4.16. Few adults have very weak skills in Sweden

Population shares by functional skill level in literacy, 2023

Sweden also makes good use of its human capital. The share of Sweden’s working-age population in the labour force is the third highest in the OECD and the employment rate is the tenth highest (Figure 4.17). The unemployment rate is high in international comparison, but this is largely because many full-time students also seek part-time jobs, and because segments of the population who would be classified as inactive in other countries are receiving benefits that are conditional on job search in Sweden (Fiscal Policy Council, 2024[42]).

Macro-level mismatches are strongly related to the incidence of weak skills. Approximately 500 000 people or 7.5% of the working-age population (aged 16-64) who are presumably able to work remain long-term unemployed or outside of the labour force, excluding full-time students and people on sick leave (Fiscal Policy Council, 2024[42]). At the same time, current skill shortages span from engineers to teachers, nurses and transport workers. There are regional differences, with for example considerable shortages in northern Sweden where a rapid green transition is set to increase demand for industrial workers as well as workers providing complementary public services (OECD, 2023[3]). In general, young people tend to migrate towards cities for studies and work. Combined with ageing this creates labour shortages in the countryside, notably in education, health and long-term care (OECD, 2021[36]).

Figure 4.17. Labour force participation is the third highest in the OECD
Copy link to Figure 4.17. Labour force participation is the third highest in the OECD

2024Q4

For the most part, this potential labour reserve faces decent work incentives, with likely less than 12 000 full-year equivalent workers on social assistance facing participation tax rates of above 80% after years of slow uprating of working-age benefits and strengthening of the earned income tax credit (Fiscal Policy Council, 2024[42]). Some may still face relatively weak work incentives. For example, low-income second earners receiving social assistance can lose up to 85% of earnings to taxes and benefit tapering if they go from inactivity to employment. This may contribute to the relatively low employment rate of female immigrants. Even so, the latter improved by an impressive six percentage points from 60% in 2021 to 66% in 2023, which is a considerably faster increase than for native-born women (one percentage point) and a higher employment rate than that of native-born women in some OECD countries (OECD, 2023[3]; OECD, 2025[43]).

Most of the people in this potential labour reserve are not working because they do not possess the skills demanded by employers or because information asymmetries prevent them from signalling to employers that they are sufficiently skilled. Over the past 15 years, unemployment in Sweden has become increasingly concentrated in groups with weak labour market competitiveness, defined by the Swedish Public Employment Service (PES) as people with no upper secondary education, people born outside Europe, unemployed people over the age of 55 and people with a disability that impairs their ability to work. In 2010, people from these groups accounted for approximately half of the registered unemployed. By 2024 their share had risen to just over 70%, and the number of registered unemployed in this group had increased slightly while the number of unemployed with none of these risk factors had approximately halved in the context of a strong labour market. People with one or more of these risk factors often experience long unemployment spells and are also clearly overrepresented among those outside the labour force (Fiscal Policy Council, 2024[42]).

Skill mismatches also have a sectoral dimension, with few job applicants relative to the number of job vacancies in some high-skill sectors notably healthcare, IT and education, and more applicants than available jobs in low-skill-sectors such as restaurants and customer service. Within sectors, there is also a mismatch between the skill profile of available applicants and the skills demanded in job openings. In healthcare, for example, as the number of job vacancies has increased, the number of unemployed with at least upper secondary healthcare qualifications has fallen while the number of low-qualified jobseekers to the sector has increased jointly with the number of job vacancies. This pattern is also reflected in the economy as a whole, indicating that the skills that are in demand are not sufficiently available among jobseekers (Fiscal Policy Council, 2024[42]).

Monopsony power among employers, notably in public services, may contribute to this imbalance. Over the past decade, the government has tried to address shortages in professions requiring tertiary education by tasking universities to increase study places in programmes leading to occupations in high demand, including teachers, nurses and engineers. This approach has only been partially successful (National Audit Office, 2021[44]). As argued in the previous OECD Economic Survey of Sweden (2023[3]), a fundamental issue with this policy is that it does not address the lack of attractiveness that holds back demand from students. The public sector is a major employer in healthcare, social care and education. The monopsony power of public sector employers, notably outside of the main agglomerations, has been found to lower the wages of skilled health care workers such as nurses relatively to low-skilled care personnel, likely intermediated by collective bargaining (Söderqvist and Eklund, 2024[45]). Increasing the attractiveness of occupations with shortages, notably teaching, social care and nursing, in terms of not only pay, but also career opportunities and working conditions should therefore be a priority. For example, 15-17% of educated teachers and preschool teachers do not work in the educational sector, while a considerable share of these would consider teaching again if the employment offer was more attractive (Lindqvist, Boström and Gidlund, 2022[46]). Around half of private-sector workers and 30% of public sector workers in social services and care sectors work part-time, while a third of part-time workers would like to work more hours (Larsson, 2023[47]).

Some people may have such low skills that their productivity is not sufficiently high to justify employment even at the sectoral minimum wage in the lowest-paid sectors such as restaurants and customer services. Alternatively, their productivity may be sufficient, but their documented qualifications, work experience and backgrounds are not sufficient to convince employers of their potential. In these sectors too, the number of vacancies and unemployed have increased in tandem, and employers have avoided recruiting people from the risk groups mentioned above, even though many of these jobs do not require specific qualifications. It thus appears that there is no demand at all for the labour of certain groups at current wage levels (Fiscal Policy Council, 2024[42]).

PIAAC data shows that all the risk groups have lower literacy skills than the population average, indicating, unsurprisingly, that people within the respective groups also have lower productivity on average. Residents of Sweden outside of the risk categories score approximately 300 points on average in the PIAAC literacy test. This places skills in the middle of functional literacy level three, which corresponds to middle-level functional skills. People within the risk groups score below 250 points on average, which is more than one whole functional literacy level below and corresponds to the very low end of decent functional skills (Figure 4.18, Panel A).

Employers who are not able to accurately assess individual skill levels within one of these risk groups may be reluctant to recruit from the group as such. Alternatively, if employers are better able to assess individual skills within the group, the higher-skilled individuals within each group would be more likely to become employed, while the lowest-skilled individuals would remain unemployed (Pareliussen, 2025[2]). All the risk groups show signs of sorting individuals with higher literacy skills into employment, as those not in employment have lower literacy on average. There are particularly strong signs of sorting into employment based on literacy skills for people born outside of Europe and North America (Figure 4.18, Panel A). This fits well with the finding above that the employment- and earnings- penalties this group faces relative to the native-born are largely reflecting their skills profiles.

Almost half (47%) of the population aged 16-64 which is neither in employment nor in full-time studies belong in one or more of these risk groups. Of these, 85% are either low-educated or immigrants from outside of Europe and North America, or they display a combination of risk factors including one of these (Figure 4.18, Panel B). In terms of policy priorities, this indicates that efforts towards these two groups have the highest potential to expand employment.

Figure 4.18. Groups at risk of long-term unemployment and inactivity have weak literacy skills on average
Copy link to Figure 4.18. Groups at risk of long-term unemployment and inactivity have weak literacy skills on average

2023

For people born outside of Europe and North America, PIAAC data indicates that weak labour market outcomes are closely related to individual differences in literacy, which is likely reflecting a combination of weak Swedish skills due to their foreign background and a weaker level of literacy and processing skills in absolute terms. The priority remains to build literacy skills through an individually adapted combination of Swedish language classes, education leading to formal qualifications and work-based learning, for which subsidised jobs may be a useful stepping-stone into employment. While it is encouraging that labour market matching of this group has improved since 2012, the findings above show that literacy skills have not improved significantly, pointing to continued weaknesses in Swedish Tuition for Immigrants (SFI), despite previous reform efforts (OECD, 2015[48]; OECD, 2017[49]). Important weaknesses are related to the fact that immigrants entitled to SFI are a diverse group. Classes are often not sufficiently adapted to people’s backgrounds and not sufficiently flexible to fit their other commitments. Remote learning may also not be sufficiently conducive to learning oral Swedish. 24 providers out of 30 subjected to a quality inspection in 2023 failed to meet the quality requirements set by the School Inspectorate (Skolinspektionen, 2023[50]). Intensifying quality inspections of Swedish Tuition for Immigrants and using these to map and disseminate good practices in teaching and organisation could help address these problems.

It may also be the case that more adapted classes will require increased resources. The OECD Economic Survey of Sweden (2021[36]) pointed out that municipalities tend to face difficulties in meeting standards set by the central government in some areas, especially personal assistance and integration of immigrants. The central government funds immigrant integration fully for the first two years after arrival when they are enrolled in the Introduction Programme. Funding follows the immigrant after this time as well, for example through the cost-and income equalisation system and through grants for specific services. Given the low employment rate of immigrants after the two-year integration period and that municipalities are financially responsible for persons claiming social assistance, the financing for integration may need to be re-examined.

For people with less than upper secondary education, the signalling effect seems relatively more important, as the sorting on literacy skills is weaker. Equipping people with upper secondary diplomas should be effective in improving their labour market prospects, but this would require individual ability and motivation to study, and may therefore not be effective or realistic for everyone. Shorter up-skilling activities with clear labour market relevance (Arbetsmarknadsutbildning, AUB) combined with counselling and job search assistance from the public employment service has been shown to be an effective combination for this group, as discussed below (Fredriksson et al., 2023[40]; Fiscal Policy Council, 2024[42]).

4.4.3. Vulnerable groups with weak skills need intensive and efficient interventions

People in the risk groups described above need intensive and efficient interventions to strengthen their human capital. As discussed above, work incentives are for the most part adequate. Further benefit cuts will therefore increase financial vulnerabilities, including for families with children, without substantial employment gains. A first step to address weaknesses in human capital is to have a system in place to regularly follow up with the people concerned. Sweden is well-placed in this respect, since its benefits are largely conditional on registering as unemployed and participating in active labour market policies. Of the approximately 500 000 people in the latent labour force (in 2022), 210 000 are registered as long-term unemployed. The rest are early retired (80 000), home-makers (40 000), job-seeking without meeting the LFS definition (50 000) or inactive (100 000) (Fiscal Policy Council, 2024[42]). There may be some overlaps between these groups as data is collected from different sources, and employment has been on the rise in the intervening years. Applying similar definitions on data from the Labour force survey for the age group 20-64 nonetheless implies that the latent labour force defined in this way remained above 350 000 people in 2024. Apart from the long-term unemployed, these groups are not subject to benefit conditionality, and are therefore harder to reach. There may nonetheless be possibilities to identify them and follow them up through their contact with other public services, such as health and social services. The “Equal Establishment” pilot (see below) also provides a potential model. This programme targeted intensive mapping and matching services not only to the refugee immigrants receiving benefits, but also to their relatives.

Evidence suggests that active labour market policies in Sweden work. For example, responsibility for immigrant integration was transferred from municipalities to the public employment service (PES) in 2011. As discussed above, PIAAC data does not show a significant improvement in the literacy of immigrants from 2012 to 2023, but matching seems to have improved, as gaps in earnings and employment are now fully reflecting education, skills and other characteristics which are not related to immigrant status. Furthermore, the time it takes from the date of immigration to finding employment has been shortened on average (NIER, 2023[51]). Contributing factors may be that systems have become better able to cope as the number of new immigrants has decreased, and the composition of immigrants may have become better matched to the Swedish labour market. Immigrants arriving from Syria around 2014-16 are for example faring well, as they have higher average qualifications compared to some other groups which were more numerous in the past.

Ideally, active labour market policies should work as a chain where interventions targeted to the labour market and individual needs are given at the right time to the right unemployed person. This should be possible to do in a better way than what is currently the case in Sweden. The Fiscal Policy Council (2024[42]) pointed out that bureaucratic hurdles currently prevent measures that are well-matched to individual circumstances from being used in a timely manner.

Early diagnosis, understanding people’s individual circumstances as they register as unemployed or even as they receive notice of redundancy is the first step to assign people to active labour market policies timely and correctly. The risk factors discussed above are well-known and should, together with supplementary information available to the PES, make it relatively straightforward to correctly identify a large share of those at risk of long-term unemployment early on, and identify the specific risk factors in question. This kind of diagnostic is already routine in the PES, who assign the unemployed into three risk groups (A, B and C) based on their vulnerability to long-term unemployment, assessed based on their registered data as well as caseworker assessments. However, this diagnosis tool is only applied to those who are assigned to private matching services providers, which only happens after a spell of unemployment (OECD, 2023[52]; Arbetsförmedlingens, 2024[53]). Mechanically screening all unemployed at the time of registration could facilitate early detection of people in need of intensive measures. Improving information sharing with Job security councils, which are established by collective agreement to support workers affected by structural change and redundancies (Box 4.4), may also be a low-cost way to identify early on individuals at risk of long-term unemployment.

Box 4.4. Sweden’s Job security councils

Copy link to Box 4.4. Sweden’s Job security councils

Job security councils are non-profit foundations established by collective agreement to administer support outlined in job security agreements between labour and employers unions, notably to provide support to employees who have received notice of (collective) redundancy. They were first developed in the 1970s against the backdrop of the deteriorating economic conditions in Sweden in the late 1960s and the massive job losses among white-collar workers in the wake of the oil crisis in 1973. In this context, the public employment service was not regarded by employers as providing sufficient support for white-collar workers to find new jobs. Under the 2022 amendment to the Saltsjöbaden Agreement, Job security councils became important gatekeepers the new Student Finance for Transition and Retraining Programme. In this context, a new public Job security council was opened by the Legal, Financial and Administrative Services Agency for workers not covered by a collective agreement.

The council is made up of a board of representatives from the different partners involved in the collective bargaining agreement, with the seats split equally between employer and employee representatives. The board has the task of deciding the scope and content of the support that is to be granted. The system works like an insurance mechanism and the premiums that are paid by each company vary between sectors and occupational groups (as defined by the trade unions). About 0.3% of the wage bill is paid by each affiliated company.

The advisers and consultants working in the job security council have a high degree of freedom in preparing individualised support for each affected employee. The measures provided are flexible and the support activities are tailored to the needs of each individual, taking into account their qualifications, professional interests, and personal preferences and concerns. In most cases, support activities are initiated by some form of counselling, guidance meetings, or advisory seminars, to determine the characteristics of and possibilities for the person as well as what opportunities and challenges there are. These initial activities are usually followed by further measures in the form of training or education, personal development activities, study, or support in starting a new business.

Employees facing redundancy may also receive financial compensation, in addition to general unemployment benefits. Participants in the new Student Finance for Transition and Retraining Programme can also receive top-ups to the grants and loans provided by the Swedish Board of Student Finance from their council, in accordance with the collective agreement.

In some agreements, the support activities last for a maximum of five years, or until the employee has found a new job or chosen to discontinue their relationship with the council. However, support usually is provided for a period of six to eight months.

Source: OECD Employment Outlook (2018[54]); Eurofound (2022_[55]); Author’s updates.

The next challenge is to correctly make use of this information in the activation chain. Shorter AUB “Labour market training” courses, on average around six months in duration, have proven to be effective in times of labour shortages. Courses in transport, machinery, care and ICT are among the most cost-effective, followed by courses in arts and crafts, construction and administration. Courses in retail trade and restaurants, culture and media and agriculture have low cost-efficiency. The value added of the courses is found to be particularly high among people with weak labour market positions such as those without upper secondary education, the long-term unemployed and the foreign born (Arbetsförmedlingen, 2025[56]). Such training should be prescribed more frequently in the current situation with labour shortages and long-term unemployment rising in tandem. However, the number of unemployed participating in this programme remains low and the number participating in any vocational training, including AUB and Municipal vocational adult education and training (Yrkesvux) has remained largely constant (Sundén, 2022[57]).

The main reasons why AUB is not scaled up seem to be largely of bureaucratic nature, including a lack of funding flexibility at the PES, an explicit strategy to avoid duplicating training programmes that exist in Yrkesvux, and appeals to procurement of AUB providers. Candidates placed in AUB receive the activity support benefit, for which the budget is capped by annual appropriations. This replaces the unemployment benefit, which is rules-based and thereby not capped. This could incentivise the PES to wait until unemployment benefits rights have expired (Fiscal Policy Council, 2024[42]). However, funding for AUB seems to have been sufficient in the past few years, and a more likely explanation of why AUB is not scaled up in line with political intentions is that the PES has not had enough internal capacity (under the administration appropriation) to identify candidates that can benefit from AUB. Also, approximately 30% of AUB participants quit after an introductory testing and mapping part. In the 2025 Budget the Government increased the PES administrative appropriation with SEK 79 million, and signalled further increases in 2026 and 2027.

Similarly, job counselling and placement services provided by the PES are reported under the administration appropriation, which is usually fully subscribed, while the purchase of private placement services is counted under a larger separate appropriation which is not. This contributes to steering active labour market policies towards private placement services, even though they are not proven to be better. A reorganisation of the PES to a purchaser-provider model was implemented around 2020, further accelerating the use of private services.

Evaluations show that private and public provision of similar interventions reach similar employment outcomes in Sweden, while private provision increases user satisfaction (Lundin, Söderström and Sibbmark, 2024[58]). Egebark et al. (2024[59]) found that the external providers reach similar results to the PES in terms of employment rates, sustainable employment and wages, but their costs are higher than the same services provided by the public system by 46% to support jobseekers closer to the labour market and by 71% to support jobseekers further from the labour market. In view of these evaluation results, the Swedish Government aims to further fine-tune the system in 2025, decreasing the share of contracted-out employment services again and increasing the share provided in-house. A further problem is that the training elements provided by private providers of employment services are relatively limited, and the unemployed may therefore end up with a chain of interventions that is less adapted to their actual needs when the PES relies too heavily on external providers (Sundén, 2022[57]).

As an example, a pilot project, Equal Establishment, targeted refugee immigrants and their relatives between 2018 and 2021 with a tailored intensive matching service (“Matching from day one”) whereby their preferences and skills, including informally acquired skills, were mapped by a job broker and then matched against a register of employers' recruitment needs. This pilot increased the probability of being in work or studies by approximately eight percentage points compared with the Employment Service's regular support (OECD, 2023[3]). Despite this, the programme was discontinued, partly because of the ongoing shift to a purchaser-provider model at the time. In addition the programme made intensive use of PES staff, reported under the administration appropriation, and was therefore perceived as too expensive even though the ex-post evaluation showed that the average cost was slightly lower than for participants referred to private providers (Fiscal Policy Council, 2024[42]).

There have been indications that remote and sparsely populated regions are not sufficiently covered by private sector providers after the PES reform (OECD, 2023[60]). Furthermore, municipalities have so far not been allowed to provide matching services as they are confined by law to operating within their proper geographical boundaries, while the tendered service areas often straddle several municipalities. Simplifying application requirements for providers that are already active in other regions and easing requirements on physical presence in rural areas could enhance competition in these areas. Pragmatic solutions should also be found to allow municipalities to provide matching services in fair competition with private providers (OECD, 2023[3]). Yrkeshögskolan is one example where municipalities and private providers compete on equal terms to provide publicly financed services whose users straddle municipal borders.

Subsidised job placements have also been shown to be effective, but employers are not eager to participate due to a lack of awareness, stigma towards the skills of the unemployed and a perception that participation is marred by red tape. These perceptions seem to be unjustified, as those employers who actually have participated report being satisfied with the people they employed and not being overly burdened by red tape (Fiscal Policy Council, 2024[42]). These wage subsidies are capped, and the cap has not been uprated in line with wage growth, which also gradually undermines their attractiveness. Introduction Jobs (Etableringsjobben), which is a programme combining subsidised work with training for those furthest from the labour market for a maximum duration of two years, is a new scheme introduced in 2024 that is operated by social partners with subsidies paid by the Government. This new model could potentially overcome some of the issues above, but the number of participants has so far been very low and the challenge remains to scale up the programme.

The government should put more resources into education and training as part of active labour market policies. Sweden’s active labour market policy (ALMP) spending is currently 2.2% of GDP, around the OECD average, and spending per unemployed as a share of GDP per capita places Sweden around OECD median (Figure 4.19, Panel A). Over the past decade, spending on active labour market policies has fallen as a share of GDP even as long-term unemployment has increased, and the number of programme participants has fallen (OECD, 2024[61]). The spending is heavily tilted towards employment incentives and job search assistance through private providers rather than training (Panel B). Based on the findings above, a rebalancing of resources seems warranted towards training measures, preferably by increasing the overall ALMP envelope, or alternatively by shifting ALMP spending from externally provided matching services and employment incentives.

Figure 4.19. Active labour market policies could be better targeted
Copy link to Figure 4.19. Active labour market policies could be better targeted

2022

The reluctance to duplicate vocational education already provided in municipal adult education (Komvux) seems to result in too little education and training being provided to the unemployed overall and results in gender inequalities. 80% of VET study places offered by the PES-run labour market training (AUB) programme are in male-dominated professions, while adult VET in female-dominated professions is mainly provided by Komvux. Participants in the former receive activity support benefits while studying, while participants in the latter support themselves through student loans. The PES can also direct long-term unemployed to apply for municipal adult education as a labour market measure, in which case they will receive the same activity support benefit as participants in AUB. The long-term unemployed can in principle receive the so-called education entry grant for studies at Komvux, but this support is not frequently utilised, and the government has signalled that it will be abolished from 2026. For now, women who attend upper secondary vocational training generally have less favourable financial conditions than men (Fiscal Policy Council, 2024[42]), but PES could reduce this gender inequality in study financing by directing long-term unemployed individuals to municipal adult education.

4.4.4. A significant expansion of adult education finance needs some recalibration

The 2022 amendment to the Saltsjöbaden Agreement revolutionised adult education and training in Sweden by significantly increasing workers’ rights to go on paid education leave in the new Student Finance for Transition and Retraining Programme, largely financed by the government. In return, social partners agreed to streamline employment protection legislation, notably to limit the use of non-standard work contracts and make it easier to deviate from the first-in-first-out principle in the case of redundancies (OECD, 2023[3]). These reforms were negotiated and adopted as a package. Changes to improve its individual components may therefore risk reopening the debate on what is overall a very sensible set of measures. With this caveat, there is scope to optimise the Student Finance for Transition and Retraining Programme so that it becomes available to more people and better targeted towards workers at risk of unemployment.

The new programme offers grants of up to 80% of the wage for up to 44 weeks for full-time studies, or longer for part-time studies. The subsidy is capped, but student loans are available to top up financing up to 100% of the previous wage (also subject to a cap). Participation is conditional on at least eight years of work experience. The support is administered by CSN, the Swedish Board of Student Finance. This scheme has quickly become popular across a wide range of professions, especially among middle-aged workers. Some teething issues on the administrative side are being addressed.

Job security councils established by collective agreement to support workers affected by structural change and redundancies are important gatekeepers to the programme. These councils provide help and support to develop skills, for example by means of free study and career guidance, and can in some cases provide supplementary study grants to programme participants. One innovation resulting from the new programme is that a public Job security council was opened by the Legal, Financial and Administrative Services Agency for workers not covered by a collective agreement.

The programme has become very popular. There are two application rounds per year, and out of more than 21 000 applicants in the fall of 2023, only 14% were accepted. Two-thirds of applicants were women, with acceptance rates similar across genders. 50% applied to higher vocational education, 40% to tertiary education and the rest to other studies including municipal adult education and folk high school. The respective shares of accepted applications were approximately at the same magnitudes. The median age was 40-44 years among both applicants and accepted students (SCB, MYH and UKÄ, 2024[62]).

As Swedes increasingly continue working until their late 60s and 70s in a context of rapid technological change, strengthening investments in adult education and training to make it widely accessible also to experienced workers is warranted. Participation in adult education and training can mitigate field-of-study mismatch, strengthen career prospects within one’s current field of study, or enable people to change to a field that provides better career prospects or is simply more in line with personal interests.

Even so, the support provided should be proportional to societal benefits and targeted to maximise value for money. Up-skilling and re-skilling come with an investment cost consisting of the cost of provision of the education activity and the alternative cost of lost labour. If this cost is almost fully covered by public subsidies and transfers there is a risk of over-investment in adult skills for groups who have high preferences to invest in skills and who are already high-skilled. People in employment with higher levels of education and literacy are in general more able to identify, evaluate and access relevant opportunities to invest in skills by themselves. The gains to society from such investments are equal to incremental increases in pre-tax profits and earnings, and much of these gains befall the individual and possibly their employer, who therefore have incentives to invest in their own skills on their own accord.

The situation for people with weak labour market positions is different. They will typically be less capable of correctly assessing their skill needs and of evaluating and accessing educational options. They will be more likely to require personalised interventions to fill their specific needs, be it weak language skills, low education or health issues. They may underestimate the returns to investments in skills and they may also suffer from cash constraints preventing them from making such investments on their own. The societal gain of for example moving a person from long-term unemployment to full-time employment far outstrips private gains. In money terms, the individual will only gain the difference between post-tax earnings and benefits, while the gain to society equals the full value of pre-tax full-time earnings and pre-tax increased profits of the employer, and possibly co-benefits in terms of for example reduced social exclusion and improved health. Targeting interventions to people with weak labour market positions and those who have received notice of redundancy can therefore increase value for money even if the investment cost is higher and the success rate lower than for higher-skilled individuals who are well-established in the labour market.

Experience from the two first rounds of support indicates that the level of the grant should be reduced to make it available to more people and better aligned with other supports available to invest in skills, and that programme administration should be adjusted to better target individuals at risk of unemployment. First, the programme is costly compared to other supports available. SEK 2.84 billion was allocated to the programme in the 2024 state budget, which is approximately 15% of overall student finance provided by CSN. The maximum grant available to participants is more than five times higher than the grant available in the regular student aid system. Unemployed individuals in comparable activation programmes typically receive compensation of roughly the same magnitude, but typically also have a greater need for retraining. Second, the programme is largely benefitting higher-educated individuals with low unemployment risk and low exposure to disruption from technological change and AI. The likelihood of applying increases with unemployment risk, while the likelihood of being accepted into the programme declines. This is because higher-educated individuals with low unemployment risk are likely to apply early and present a letter from their Job security council certifying that the planned training will strengthen their position in the labour market. CSN allocates funding on a first-come-first-serve basis, with priority to applicants who can present such a statement. Third, income evolves similarly post-application, both for individuals who are granted and denied support, indicating limited value added under the current design (Fredriksson and Seim, 2024[63]). These findings suggest that the system should be recalibrated by lowering the grant part considerably, especially for high-skill-low-risk applicants, possibly compensated by an increased student loan part. The system should also prioritise individuals with weaker educational backgrounds and higher risk of unemployment in the application process.

FINDINGS (main ones in bold)	RECOMMENDATIONS (key ones in bold)
Further sharpening the quality of the skills supply system
National tests in core subjects are graded locally, which introduces biases within and between schools.	Grade national tests externally to create an objective benchmark for school performance as planned and remove differences in grading leniency.
The current grading scale in lower secondary education gives and outsized weight to fail grades, and may arbitrarily prevent pupils from entering upper secondary education.	Weigh high and low grades symmetrically and suppress the requirement to pass in certain subjects to enter upper secondary education.
The socio-economic gradient of PISA scores has continued to increase.	Assign pupils to over-subscribed private schools by lottery, or with quotas reserved for pupils with unfavourable socio-economic backgrounds.
Upper secondary vocational education and training is unpopular, and there are issues relating to coordination and over-supply of programmes with weak labour market relevance. It is less responsive to labour market needs than the higher vocational education model.	Run a pilot to coordinate upper-secondary vocational education centrally as is done in higher vocational education, with the aim to transition the whole system to the new model if successful.
The literacy premium associated with a tertiary degree seems to have fallen since the 1990s.	Investigate the extent and causes of falling quality of tertiary education.
Universities are key to Sweden’s ability to innovate and swiftly internalise innovations at the global scientific frontier.	Consider funding options to facilitate more research and more teaching hours per student.
Universities contribute to a varying degree to regional development by fostering innovation and helping retain local talent.	Strengthen incentives and support to raise the contribution of universities to regional labour supply, knowledge and innovation.
Sweden’s decentralized skills supply system can lead to redundancies and weak coordination, resulting in sub-optimal overall outcomes.	Strengthen horizontal coordination of agencies and stakeholders involved in skills supply through the Inter-agency Co-operation Structure for Skills and Lifelong Learning.
Targeting resources to reduce the incidence of weak skills and increase value for money
Funding of active labour market policies has fallen while the number of long-term unemployed has risen. The share of funding allocated to training is very low.	Strengthen activation policies and prioritise up-skilling and re-skilling activities.
An increasing share of the unemployed have clear up-skilling needs, but red tape holds back the use of training and employment subsidies with proven effectiveness.	Organise activation policies according to effectiveness and the needs of the unemployed in risk groups rather than according to budgetary and organisational silos.
Providers of Swedish Tuition for Immigrants often do not adapt courses sufficiently to people’s diverse backgrounds. Remote learning adds flexibility but cannot fully substitute for in-person classes.	Intensify quality inspections of Swedish Tuition for Immigrants and use these to map and disseminate good practices in teaching and organisation, including the use of remote learning.
Vocational training by the Public Employment Service (PES) is male-dominated and come with unemployment benefits, while women largely finance training by student loans in municipal adult education.	Increase the use of municipal adult education as an active labour market policy.
The new Transition and Retraining Programme offers experienced workers five times more generous grants than what is available in ordinary student finance.	Reduce the grant part of the new Transition and Retraining Programme to better align it with normal student finance and make it available to more participants.
The new Transition and Retraining Programme is largely benefitting higher-educated individuals with low unemployment risk.	Prioritise individuals with high unemployment risk in the new Transition and Retraining Programme.

References

[39] Adalet McGowan, M. and D. Andrews (2017), “Skills mismatch, productivity and policies: Evidence from the second wave of PIAAC”, OECD Economics Department Working Papers, No. 1403, OECD Publishing, Paris.

[56] Arbetsförmedlingen (2025), Arbetsmarknadsutbildningars kostnadseffektivitet och bidrag till kompetensförsörjningen, Swedish Public Employment Service.

[38] Arbetsförmedlingen (2024), Hur effektiv är matchningen av arbetslösa till jobb? - -Trender och mekanismer 1992 - 2022, Swedish Public Employment Service.

[10] Arbetsförmedlingen (2024), Statistik-och analysuppdrag - Återrapport enligt regleringsbrev 2024, Swedish Public Employment Service.

[53] Arbetsförmedlingens (2024), Arbetsförmedlingens nya bedömningsmodell, Swedish Public Employment Service.

[14] Åslund, O., A. Karimi and A. Sundberg (2025), “Origin, norms, and the motherhood penalty”, IFAU Working Paper, No 2025:1.

[37] Böhlmark, A. and G. Waisman (2024), “Hur effektiv är matchningen av arbetslösa till jobb? – Nationalekonomiska Föreningen”, Ekonomisk debatt, Vol. 52/5, (accessed on 7 April 2025).

[30] Brommesson, D., S. Nordmark and J. Ödalen (2024), Swedish Higher Education: Caught between Offering High-Quality Education or Education for the Masses?, SNS, Stockholm.

[9] Bussi, M. and J. Pareliussen (2017), “Back to Basics – Literacy Proficiency, Immigration and Labour Market Outcomes in Sweden”, Social Policy and Administration, Vol. 51/4,.

[33] Edin, P. et al. (2022), “The Rising Return to Noncognitive Skill”, American Economic Journal: Applied Economics, Vol. 14/2, pp. 78-100, https://doi.org/10.1257/APP.20190199.

[59] Egebark, J. et al. (2024), “En effektutvärdering av arbetsförmedling med fristående leverantörer“, IFAU Report 2024:17.

[13] Ekonomifakta (2025), Kvinnor i styrelser - internationellt (accessed on 6 March 2025).

[12] Entreprenörskapsforum (2023), Kvinnor i entreprenörskap i Sverige.

[23] Eurostat (2023), Labour Force Survey, (accessed on 18 March 2025).

[27] Eurydice (2024), Organisation of post-secondary non-tertiary education, (accessed on 9 February 2025).

[42] Fiscal Policy Council (2024), Swedish Fiscal Policy - The Swedish Fiscal Policy Council’s report 2024.

[40] Fredriksson, P. et al. (2023), Konjunkturrådets rapport 2023. Strukturomvandling på svensk arbetsmarknad: konsekvenser och policyåtgärder, SNS, Stockholm.

[63] Fredriksson, P. and D. Seim (2024), Who Receives Student Finance for Transition and Retraining? An Empirical Study of the First Two Application Periods, SNS Research Brief, No 104.

[20] Grönqvist, E., L. Hensvik and A. Thoresson (2022), “Teacher career opportunities and school quality”, Labour Economics, Vol. 77.

[17] Grönqvist, E., B. Öckert and O. Rosenqvist (2021), “Does the ‘Boost for Mathematics’ Boost Mathematics?”, IFAU Working Paper, No 2021:2.

[24] Gustafsson, J., S. Sörlin and J. Vlachos (2016), Policyidéer för svensk skola, SNS, Stockholm.

[32] Hensvik, L. and O. Skans (2019), Jobbpolarisering och efterfrågan på olika förmågor på framtidens arbetsmarknad, IFAU.

[29] Higher Education Authority (2024), Improving quality – a developed methodology for UKÄ’s reviews, Report.

[18] Holmlund, H., J. Häggblom and E. Lindahl (2024), “The Boost for reading”, IFAU Working Paper, No 2024:6.

[5] Larsson, A. (2023), PIAAC Den internationella undersökningen av vuxnas färdigheter, Statistics Sweden.

[47] Larsson, M. (2023), Arbetstider år 2023, The Swedish Trade Union Confederation.

[34] Lindqvist, E. and R. Vestman (2011), “The Labor Market Returns to Cognitive and Noncognitive Ability: Evidence from the Swedish Enlistment”, American Economic Journal: Applied Economics, Vol. 3/1.

[46] Lindqvist, M., L. Boström and U. Gidlund (2022), “Teacher shortage in Sweden: time to take action?”, Education in the North, Vol. 29/2.

[58] Lundin, M., M. Söderström and K. Sibbmark (2024), Rusta och matcha - Vad händer hos privata leverantörer av arbetsmarknadsinsatser?, IFAU.

[41] Montt, G. (2015), “The causes and consequences of field-of-study mismatch: An analysis using PIAAC”, OECD Social, Employment and Migration Working Papers, No. 167, OECD Publishing, Paris.

[28] MYH (2024), Statistisk årsrapport 2024 - Myndigheten för yrkeshögskolan, Agency for higher vocational education.

[44] National Audit Office (2021), Riktade utbyggnadsuppdrag till universitet och högskolor – regeringens styrning genom utformning och uppföljning, Riksrevisionen.

[51] NIER (2023), Utrikes föddas uppåtgående sysselsättning: Hur mycket förklaras av demografi? - Konjunkturinstitutet, Specialstudie, Konjunkturinstitutet, Stockholm.

[11] OECD (2025), Korea’s Unborn Future, Understanding Low-Fertility Trends, OECD Publishing.

[43] OECD (2025), Labour market outcomes of immigrants - Employment, unemployment, and participation rates by sex, OECD Data Explorer, (accessed on 23 February 2025).

[1] OECD (2024), Do Adults Have the Skills They Need to Thrive in a Changing World?: Survey of Adult Skills 2023, OECD Skills Studies, OECD Publishing, Paris.

[22] OECD (2024), Education at a Glance 2024: OECD Indicators, OECD Publishing, Paris.

[61] OECD (2024), Labour Market Programmes, OECD Data Explorer, (accessed on 14 February 2025).

[16] OECD (2024), Strengthening the Governance of the Swedish Skills System: Final Report, OECD Skills Studies, OECD Publishing, Paris.

[3] OECD (2023), OECD Economic Surveys: Sweden 2023, OECD Publishing, Paris.

[52] OECD (2023), “Organisation of public employment services at the local level in Sweden”, OECD Local Economic and Employment Development (LEED) Papers, No. 2023/03, OECD Publishing, Paris.

[60] OECD (2023), “Participation of the social economy in the provision of Sweden’s public employment services”, OECD Local Economic and Employment Development (LEED) Papers, No. 2023/04, OECD Publishing, Paris.

[7] OECD (2023), PISA 2022 Results (Volume I): The State of Learning and Equity in Education, PISA, OECD Publishing, Paris.

[35] OECD (2022), Education at a Glance 2022: OECD Indicators, OECD Publishing, Paris.

[36] OECD (2021), OECD Economic Surveys: Sweden 2021, OECD Publishing, Paris.

[6] OECD (2019), OECD Economic Surveys: Sweden 2019, OECD Publishing, Paris.

[54] OECD (2018), OECD Employment Outlook 2018, OECD Publishing, Paris.

[49] OECD (2017), OECD Economic Surveys: Sweden 2017, OECD Publishing, Paris.

[48] OECD (2015), OECD Economic Surveys: Sweden 2015, OECD Publishing, Paris.

[2] Pareliussen, J. (2025), “Adult skills, labour market outcomes and social cohesion in Sweden”, OECD Economics Department Working Papers (forthcoming).

[4] Pareliussen, J. (2022), “How labour market outcomes reflect age, gender and skills in Korea”, OECD Economics Department Working Papers, No. 1739, OECD Publishing, Paris.

[8] Pareliussen, J. (2017), “How self-sorting affects migrants’ labour market outcomes”, OECD Economics Department Working Papers, No. 1435, OECD Publishing, Paris.

[21] Regeringskansliet, R. (2025), “Ett likvärdigt betygssystem”, (accessed on 7 April 2025).

[64] Regeringskansliet, R. (2020), “Komvux för stärkt kompetensförsörjning”, (accessed on 6 April 2025).

[25] SCB (2018), Matchningen på arbetsmarknaden - sambandet mellan utbildning och yrke, Statistics Sweden.

[62] SCB, MYH and UKÄ (2024), Studerande med omställningsstudiestöd vid yrkeshögskolan och högskolan - Myndigheten för yrkeshögskolan, Statistics Sweden, Swedish National Agency for Higher Vocational Education and Swedish Higher Education Authority.

[50] Skolinspektionen (2023), Undervisning inom kommunal vuxenutbildning i svenska för invandrare, Swedish School Inspectorate.

[45] Söderqvist, C. and J. Eklund (2024), “Wage effects from employer concentration and collective bargaining in Swedish labor markets”, In thesis: Essays on exit, voice, and technology: Industrial relations in modern Swedish labor markets, Blekinge Institute of Technology.

[26] SOU (2017), Samling för skolan - Nationell strategi för kunskap och likvärdighet, Swedish Government Official Reports.

[19] Statskontoret (2021), Olika vägar till likvärdig skola – Utvärdering av likvärdighetsbidraget till skolan. Slutrapport, The Swedish Agency for Public Management.

[15] Sundberg, A. (2024), The child penalty in Sweden: evidence, trends, and child gender, IFAU Working Paper 2024:12.

[57] Sundén, A. (2022), Yrkesutbildning löser matchningsproblemen, Swedish Trade Union Confederation.

[31] Sveriges Ingenjörer (2023), Minskade resurser til utbildning hotar Sveriges ställning som ledande ingenjörsland.

Featured topics

Agriculture and fisheries

Climate change

Development

Digital

Economy

Education and skills

Employment

Environment

Finance and investment

Governance

Health

Industry, business and entrepreneurship

Regional, rural and urban development

Science, technology and innovation

Society

Taxation

Trade

Energy

Nuclear energy

Transport

Featured topics

Agriculture and fisheries

Climate change

Development

Digital

Economy

Education and skills

Employment

Environment

Finance and investment

Governance

Health

Industry, business and entrepreneurship

Regional, rural and urban development

Science, technology and innovation

Society

Taxation

Trade

Energy

Nuclear energy

Transport

Countries A - C

Countries D - I

Countries J - M

Countries N - R

Countries S - T

Countries U - Z

Regional and global engagement

Countries

Countries A - C

Countries D - I

Countries J - M

Countries N - R

Countries S - T

Countries U - Z

Regional and global engagement

Publications

Publications

Featured publications

Data

Data

Featured data

News & events

News & events

Featured events

About OECD

About

Engage with us

Work with us

Featured topics

Agriculture and fisheries

Climate change

Development

Digital

Economy

Education and skills

Employment

Environment

Finance and investment