Main findings and policy pointers

The world is changing and this is having a profound impact on labour markets: on the number and nature of jobs, as well as on the distribution of work and income. No country is left unaffected, including Belgium.

Important demographic changes are underway and populations are ageing rapidly. While this trend is slightly less marked in Belgium than in its neighbouring countries, the old-age dependency ratio is nonetheless projected to increase in Belgium from around 30% today to 43% in the next 30 years meaning that, in 2050, there will be just over two people of working age for every person aged 65+, compared with over three today (Figure 1, Panel A).

The world is more closely integrated along global supply chains and, being a small, open economy, a large share of jobs in Belgium are dependent on demand from abroad. To give just one example, the share of business sector jobs in Belgium that depend on demand from abroad has increased from 46% in 2005 to 54% in 2015. In larger countries like Germany and France, this share is much lower – 34% and 29%, respectively (Figure 1, Panel B).

Last but not least, new technologies are finding their way into the workplace and are changing the tasks that workers focus on, how they carry out their work, as well as the skills that are required for the job. Firms in Belgium are among the most digitalised in Europe, with 17.5% scoring “high” or “very high” on the Digital Intensity Index of the European Commission, compared to just 9.3% and 7.0% in France and Germany, respectively. Belgium also ranks third in the OECD in terms of the share of employment that is in highly digital-intensive sectors (OECD, 2020[1]) (Figure 1, Panels C and D).

These mega-trends are associated with, on the one hand, opportunities for the creation of new, digitally-intensive jobs, but also, on the other hand, the risk of obsolete jobs disappearing and significant changes in the tasks done and skills used at work. Across the OECD on average, 14% of jobs are at high risk of automation, while another third are at risk of significant change over the next 10 to 20 years. These risks are similar in Belgium (Figure 2, Panel A). 16% of workers in Belgium say that the introduction of ICT has already led to a change in the tasks that they performed at work (Figure 2, Panel B).

The risk of automation is significantly higher for the low-educated. In Belgium, 26.4% of the low-educated are at high risk of automation, compared to just 4.5% of the high-educated (Nedelkoska and Quintini, 2018[2]). In recent years, automation has affected “routine” occupations in particular – i.e. occupations where many of the tasks performed by workers are accomplished by following explicit rules. As a result, the share of middle-skill occupations in overall employment has declined, while the share of low- and high-skilled jobs has increased. This phenomenon is also sometimes referred to as “job polarisation”. In Belgium, as in neighbouring countries, there has been strong growth in the share of high-skilled jobs, but relatively low growth in the share of low-skilled jobs (Figure 3). In other words: fewer job opportunities have been created for the low-educated in Belgium than for the more educated.

The low-educated in Belgium are in an unfavourable starting position to face these new challenges. The employment rate of the low-educated aged 20 to 64 in Belgium is low (46.6% in 2018) and it has been falling since the beginning of the millennium (from 51.2% in 2000) (Figure 4). Among neighbouring countries, only France has also experienced a drop in the employment rate of the low-educated over the same period, but it remains above that of Belgium (51.6% in 2018). In the Netherlands, the employment rate of the low-educated has remained more or less constant (62.6% in 2018), while it has increased significantly in Germany (60.8% in 2018).1

Job quality for low-educated workers in Belgium is good relative to the situation in neighbouring countries. Average earnings of the low-educated in Belgium (EUR 2 7602) for the period 2016-18 were considerably higher than they were in Germany (EUR 1 670) and in France (EUR 2050), but marginally lower than in the Netherlands (EUR 2 880). The earnings distribution in Belgium is more compressed than in neighbouring countries. In both France and Germany, there are far more low-educated workers bunched in the bottom of the earnings distribution (Figure 5). While the Netherlands and Belgium are more comparable in terms of the median earnings of low-educated workers, the Netherlands also has more workers than Belgium in the bottom tail of the earnings distribution.3

Another sign of higher job quality among the low-educated in Belgium is the relatively low prevalence of non-standard and flexible work contracts. The share of fixed-term contracts among the low-educated in Belgium is only 11.3%, compared to 18.8% in the Netherlands, 19.4% in France and 20.6% in Germany (Figure 6). France (5.0%), the Netherlands (7.1%) and Germany (5.5%) all have significantly higher shares of temporary work agency (TWA) work among the low-educated than Belgium (3.7%). Part-time work is more common among the low-educated in the Netherlands (47.6%) and Germany (32.8%) than it is in Belgium (30.3%). And, although own-account work is relatively common in Belgium (7.4%), it is still less prevalent than in the Netherlands (11.2%). That being said, the share of non-standard work among the low-educated in Belgium has been growing faster than among the more educated.

Based on current trends, the employment rate of low-educated workers in Belgium is expected to continue to fall. While the share of the labour force without an upper-secondary qualification is expected to fall, so is the number of jobs held by low-educated workers. Between now and 2030, the number of jobs held by low-educated workers is projected to fall from 770 000 to 610 000 – i.e. by 21% (Cedefop, 2019[3]). By contrast, the number of jobs held by medium- and high-educated individuals is forecast to grow by 100 000 (+5%) and 430 000 (+20%), respectively. As a result, the employment rate for the low-educated could fall by 7 percentage points by 2030, while in the Netherlands and Germany, it could be 2 and 3 percentage points higher, respectively.

Future job opportunities for low-educated workers are more likely to be in lower-wage industries. In addition to the fall in the overall number of job opportunities for the low-educated, the latter will continue to shift away from sectors that used to offer them relatively high-quality jobs (e.g. manufacturing) and into sectors with lower-quality jobs (e.g. services). These structural shifts are expected to lead to a 1.4% decline in wages for the low-educated, as well as a small increase (1.2%) in the share of low-educated workers in non-standard work.

The low-educated represent a shrinking share of the labour force in Belgium, but this share is larger than in neighbouring countries. In Belgium, the share of low-educated in the population has dropped from 41% in 1998 to 22% in 2018. However, the share of low-educated in Belgium remains higher than in neighbouring countries: 20% in both the Netherlands and France, and 14% in Germany. In addition, important regional differences remain. The share of low-educated in the workforce is considerably higher in the Brussels-Capital Region (28%) and Wallonia (25%) than in Flanders (19%).

The composition of the low-educated has changed significantly over time. In 1998, 50% of the low-educated were women, 15% were migrants and 28% were aged 55-64. In 2018, 47% were women, 33% were migrants, and 36% were aged 55-64. Overall, the low-educated in Belgium are now more likely to be male and, in particular, older and from an immigrant background. These changes mean that policies aimed at promoting better labour market outcomes among the low-educated need to take account of their changing demographic characteristics. Different groups face different challenges.

Migrants represent a particularly vulnerable group who will need special attention. Belgium’s share of migrants (20% of the population aged 20-64) is similar to Germany’s (21%) but considerably higher than that of France and the Netherlands (14-15%). Moreover, migrants in Belgium tend to be less qualified than in neighbouring countries: 43% of migrants in Belgium from outside the EU have no (formally recognised) secondary qualification, compared to only 32% in the Netherlands, 37% in France and 38% in Germany.

There is significant skills mismatch in Belgium, and most new job opportunities will require at least an upper secondary qualification. Before the COVID-19 crisis hit, the Belgian labour market showed signs of tightness, despite the low (and falling) level of employment among the low-educated. This reflects an important imbalance in skills demand and supply. In Belgium as a whole, there are 30 employed persons with higher education for every unemployed person with higher education. Among the low-educated, there are just under seven employed persons for every unemployed person. An analysis of the existing shortage occupation lists indicates that less than 1% require low-educated workers. The vast majority (nearly two thirds) require an upper secondary qualification. Upskilling would therefore offer significantly better employment opportunities to many low-educated workers in Belgium.

Boosting investments in initial education will need to be accompanied by measures to encourage lifelong learning. Ensuring that all young people leave school with the right skills to quickly find a good job is essential. Too many young people still struggle to find a foothold in the labour market due to poor educational attainment. However, efforts will be needed to ensure that people continue learning even after they have left school. This is an area where Belgium underperforms by international standards, particularly among the low-educated. Participation in adult learning in Belgium (70%) is similar to the EU-28 average, and below participation in France (79%) and the Netherlands (87%). Among low-educated workers, participation in adult learning in Belgium (54%) is below the EU-28 average (60%).

Activation measures need to be better targeted and give more weight to training. Belgium spends a relatively large share of its GDP on active labour market policies to get the unemployed back into work. In 2017, Belgium dedicated 0.88% of GDP to active labour market policies (i.e. excluding spending on passive measures such as out-of-work income maintenance and support, as well as early retirement). This is higher than in France (0.87%), Germany (0.65%) and the Netherlands (0.64%). However, Belgium’s spending on active policies is heavily skewed towards measures that are more likely to suffer from large deadweight losses (e.g. employment incentives). Another difference between Belgium and other countries is that it spends comparatively less on training: just 0.14% of GDP, compared to double that in France (0.28%). Yet training can improve employability and could also help address the important skills mismatch that exists in Belgium today.

Labour costs in Belgium are amongst the highest in the OECD. The challenge for raising the employment rate of low-educated people in Belgium does not lie only in a lack of skills, but also in the cost of hiring to employers. Annual gross labour costs in Belgium for single workers (no children) earning 67% of the average wage4 were USD 54 408 in 2019 (with equal purchasing power). At that level of wages, only Germany and Switzerland had higher labour costs (USD 56 483 and 56 252, respectively). Labour costs for this group in Belgium were 7% higher than in the Netherlands5, and 32% higher than in France. They were also higher than in the Nordic countries. While Germany may have higher labour costs than Belgium at 67% of the average wage, more low-educated workers earn below that threshold than in Belgium.

One of the reasons for these high labour costs in Belgium is high employer social security contributions. For a single worker without children earning 67% of the average wage, employer social security contributions in Belgium represented 26% of gross earnings in 2019. While employer SSCs are even higher in France (27% of gross earnings), they are significantly lower in Germany (20%). In the Netherlands, employer social security contributions are also much lower than in Belgium, however once compulsory non-tax payments are included this gap is considerably narrowed (though a gap still remains).

Belgium operates a number of reductions in social security contributions for employers, but these could be better targeted. Recent reforms reduced the social security contributions payable by employers. The reductions were greatest for workers with the lowest earnings, however there were also some reductions at the top of the wage distribution. In addition to these “structural” reductions, there are a number of more targeted reductions in employer social security contributions for specific groups – some of which are federal (e.g. reductions for new hires) while others are regional (e.g. reductions for older and younger workers). While these reductions in employer social security contributions may benefit low-educated workers more than other types of workers, there is scope to further improve the targeting of these reductions and minimise deadweight loss.

Collectively negotiated wages in Belgium are substantially above the national minimum wage. Labour costs at the national minimum wage in Belgium are in line with those of neighbouring countries. However, in practice, only around 1% of workers earn the national minimum wage and most low-educated workers earn considerably above that. This is because wages for most workers are set in sector collective agreements and, on average, these are an estimated 20% above the national minimum wage. This set-up contrasts with France, for example, where collective agreements seldom raise the minimum wage above the national minimum. Moreover, these sector minimum wages in Belgium rise automatically each year in line with inflation (France and the Netherlands abandoned automatic wage indexation in the 1980s). Finally, high collective bargaining coverage in Belgium (96% – compared just 58% in Germany and 86% in the Netherlands) means that these minimum wages apply to almost the entire workforce.

If wages do not reflect productivity, this could result in lower employment. As long as higher wages reflect higher productivity, the former are justified. However, there is some evidence in the case of Belgium that productivity and wages do not always align, particularly across regions. For example, there is virtually no difference between the earnings distribution of Flanders and that of Wallonia (Chapter 4), despite productivity in Flanders exceeding that of Wallonia by about 14% – a gap which has grown wider over the past 10 years. Where such gaps in productivity and wages arise, the employment prospects of the low-educated are likely to be harmed, particularly in a context of limited inter-regional mobility.

Improving the labour market outcomes of the low-educated will also require measures on the supply-side that go beyond investments in skills. Low demand from employers and a lack of skills are only part of the explanation behind poor labour market outcomes for the low-educated in Belgium. Some groups face poor work incentives. For example, if a single parent with two children starts work in Belgium at the minimum wage, 87% of his/her additional earnings will be lost to either higher taxes or lower benefit entitlements. These effective tax rates on participation are considerably lower in France (61%) and in the Netherlands (56%). Work incentives are better in France than in Belgium for all household types (Table 1).

Work incentives can be improved through lower taxes when moving into work, as well as higher in-work benefits. Compared to neighbouring countries France and the Netherlands, the taxes and social security contributions that an individual in Belgium pays when he or she moves into work are considerably higher: 24.5% of gross earnings for a single person with two children earning 67% of the average wage, compared to 14.0% in France and -1.0% in the Netherlands (Figure 7).6 Belgium introduced a refundable tax credit (the fiscal work bonus) in 2011. For someone earning just under EUR 20 000 per year, the bonus was equivalent to around EUR 800 (or 4% of gross earnings). However, the median low-wage worker would not be eligible to the work bonus because his/her earnings are too high. In France, the Prime d’activité is an in-work benefit comprising a lump-sum amount that depends on family composition and a work bonus based on individual earnings. In 2019, an average person earning the minimum wage in France would receive EUR 241 per month in Prime d’activité, and a couple with two children EUR 407 per month (DF, 2020[8]). Even someone earning at 67% of the average wage would still be eligible to the Prime d’activité (Figure 7). Moreover, in Belgium, second earners face strong tax disincentives to work because income can be transferred between spouses for tax purposes.

The Belgian unemployment benefit system also generates disincentives to work. It is relatively difficult to access unemployment benefits in Belgium. However, once individuals are in receipt, there are fewer conditions imposed on individuals to keep their benefits. In particular, availability requirements are not particularly stringent and individuals have more leeway than in other countries for refusing job offers. In addition, the generosity of unemployment benefits is on the high side in Belgium. The average net replacement rate of income in unemployment of low-paid workers in Belgium is 10% higher than in France and 8% higher than in Germany. In neighbouring countries, the starting replacement rate tends to be lower, and there is also a sharp drop when individuals move on to unemployment assistance. The withdrawal of unemployment benefits when people move into work makes up the main component of the participation tax rate (Figure 7).

In addition to poor work incentives, many low-educated workers in Belgium face barriers to work which they will need help to overcome. Compared to neighbouring countries, a far larger share of non-employment among the low-educated in Belgium is due to disability and caregiving duties. Similarly, even though prime-age low-educated workers in Flanders have very good employment outcomes, this is not the case for older low-educated workers. Efforts aiming to extend working lives will continue to be needed, not just in Flanders but in Belgium as a whole.

The regulation of individual dismissals of regular workers in Belgium is among the strictest in the OECD. Such regulation is also strict the Netherlands, but less so in France and especially in Germany. Severance pay in particular is high in Belgium.

High employment protection on permanent contracts is unlikely to affect the employment rate of low-educated workers. However, in combination with lax rules on hiring on fixed-term contracts, it is likely to trap such workers in more precarious work. There are relatively few restrictions on hiring on fixed-term contracts in Belgium – particularly compared to France. In France, the valid use of fixed-term contracts is limited to specific cases (there are no such restrictions in Belgium) and also their maximum duration is limited to 18 months (unlimited for the first contract in Belgium). In countries that combine strict employment protection on permanent contracts with less strict regulation on the hiring on temporary contracts, the labour market tends to exhibit duality and workers can get trapped in less stable careers.

The COVID-19 crisis has hit low-wage, low-educated workers in particular. While it is still very early to assess the full impact of COVID-19 on different labour market groups, the emerging international evidence suggests that vulnerable workers have borne the brunt of the costs (OECD, 2020[10]). Low-paid, often low-educated workers have been particularly affected. On the one hand, many of them ensured the continuation of essential services during the lockdowns, often at a substantial risk of exposing themselves to the virus while working. On the other hand, low earners were: less likely to be able to work from home; more likely to have been working in sectors affected by shutdowns; and more likely to have suffered job or earnings losses. Evidence from Belgium’s temporary lay-off scheme (tijdelijke werkloosheid / chômage temporaire) confirms that the greatest number of benefits were awarded in sectors that employ many low-educated workers (e.g. Services to buildings and landscape activities; Food and beverage service activities; and wholesale trade) (FOD Economie, 2020[11]).

COVID-19 adds more urgency to the measures proposed in the present report, but their sequencing may need to be carefully considered. COVID-19 is likely to lead to a widening of the gap in labour market outcomes between high- and low-educated individuals. To help the low-educated back into work, measures will be required on the demand-side (lower labour costs to encourage hiring) as well as on the supply side (training, job search incentives, and access to re-employment services) – as outlined above. But the sequencing of reforms will be important. For example, it may not be appropriate in the midst of an economic crisis generated by COVID-19 to engage in reforms of employment protection legislation or to reduce the generosity of unemployment benefits – both of which can wait until the crisis subsides. On the other hand, reforms that strengthen work incentives (e.g. reasons for rejecting job offers) should be initiated straightaway, as should training measures that help workers find jobs in areas of high demand. Efforts to reduce labour costs are also high priority, which could take the shape of (targeted) reductions in employer social security contributions. The social partners should make efforts to ensure that low-educated workers are not priced out of the labour market during this difficult time. The judicious use of opt-out clauses in collective agreements and/or reductions in working time could help firms to survive that have become temporarily unprofitable and, hence, to preserve employment.

In addition to the recommendations outlined above, some temporary measures will be needed to support low-educated workers in the short-run. In the initial stages of the crisis, Governments across the OECD have put in place measures to preserve employment. Belgium has used both temporary lay-offs (tijdelijke werkloosheid / chômage temporaire) and short-time work (Corona-tijdskrediet / crédit-temps corona). Some of these measures will continue to be needed in the months ahead but, given their heavy burden on public finances, will need to be targeted on those sectors/occupations that continue to face the greatest difficulties. Indeed, the recovery has been uneven, with activity recovering more in some parts of the economy than in others. There is also a need to ensure that there are strong enough incentives for firms to move off such support and for workers to move on to more viable jobs when activity does pick up. In addition, workers will need help getting back into work, and doing so safely. Good occupational safety and health practices that limit the spread of contagion are a top priority, particularly for low-income workers who more frequently take up jobs that expose them to physical contact and a higher risk of infection. Low-educated workers are often less likely to be able to tele-work, and so targeted measures that provide family care support, part-time leave support and extent the right to flexible working arrangements may also be needed, particularly in the case of further lockdowns. While continuing to provide adequate income support will be important, care will need to be taken that resources are used wisely and targeted through the use of income and asset tests where possible. Finally, Belgium will need to start thinking about how to avoid that the crisis leaves long-term scars on the employment prospects of certain groups hit hardest by the crisis. Specific support packages that include guidance, training, hiring subsidies, etc. may need to be put in place for low-skilled youth in particular.


[3] Cedefop (2019), Cedefop, 2018 Skills Forecasts, (accessed on 27 August 2020).

[8] DF (2020), Prime d’activité : à quel montant avez-vous droit?, Dossier Familial, (accessed on 8 June 2020).

[11] FOD Economie (2020), Economische impact van het coronavirus, (accessed on 25 August 2020).

[2] Nedelkoska, L. and G. Quintini (2018), “Automation, skills use and training”, OECD Social, Employment and Migration Working Papers, No. 202, OECD Publishing, Paris,

[6] OECD (2020), OECD Economic Surveys: Belgium 2020, OECD Publishing, Paris,

[10] OECD (2020), OECD Employment Outlook 2020: Worker Security and the COVID-19 Crisis, OECD Publishing, Paris,

[1] OECD (2020), OECD Going Digital Toolkit, (accessed on 5 June 2020).

[12] OECD (2020), Taxing Wages 2018-2019: Non-Tax Compulsory Payments as an Additional Burden on Labour Income in 2019, OECD Publishing, Paris,

[4] OECD (2019), In-Depth Productivity Review of Belgium, OECD Publishing, Paris,

[7] OECD (2019), OECD Skills Strategy Flanders: Assessment and Recommendations, OECD Skills Studies, OECD Publishing, Paris,

[9] OECD (2018), Key policies to promote longer working lives: Belgium Country note 2007 to 2017, OECD, Paris, (accessed on 7 June 2020).

[5] OECD (2015), OECD Economic Surveys: Belgium 2015, OECD Publishing, Paris,


← 1. At the time of going to press, micro-data from the EU-LFS for 2019 were not yet available so, for consistency purposes, all estimates presented in this report are based on the 2018 EU-LFS. For 2019, aggregate EU-LFS data indicated that, for the age group 15-64, 36.0% of the low-educated were in employment, compared to 38.7% in France, 49.4% in Germany and 61.3% in the Netherlands.

← 2. This estimate obtained with the EU-SILC is very similar to the estimate obtained using the Belgian labour Force Survey.

← 3. Qualitatively similar results obtain if focusing only on full-time employees. Differences in the distribution of hours worked, and by extension hourly wages, across countries does not, therefore, appear to be a first-order concern when examining differences in the earnings distribution.

← 4. The focus is on 67% of the average wage rather than the minimum wage because, in Belgium, the average earnings of low-educated workers are considerably higher than the minimum wage and much closer to 67% of the average wage (in fact, they are slightly above 67% of the average wage).

← 5. The labour cost calculations presented in this report do not include non-tax compulsory payments. In both Belgium and Germany, this includes compulsory private insurance against work-related accidents and occupational diseases. In the Netherlands, these labour cost calculations omit compulsory (employee and employer) pension payments as well some contributions to health insurance. If these compulsory payments were included, the gap in labour costs between Belgium and the Netherlands would be greatly reduced, although labour costs would remain higher in Belgium, particularly for some family types (OECD, 2020[12]).

← 6. Again, in the case of the Netherlands, these calculations do not take into account compulsory contributions to private pension and health schemes, so the actual differences with Belgium are likely to be smaller.

Metadata, Legal and Rights

This document, as well as any data and map included herein, are without prejudice to the status of or sovereignty over any territory, to the delimitation of international frontiers and boundaries and to the name of any territory, city or area. Extracts from publications may be subject to additional disclaimers, which are set out in the complete version of the publication, available at the link provided.

© OECD 2020

The use of this work, whether digital or print, is governed by the Terms and Conditions to be found at