Leveraging administrative data, this chapter examines the profiles of MIS beneficiaries, their registration patterns with Public Employment Services (PES), their transitions into the labour market, as well as the jobs they enter. In Greece, around two in ten MIS beneficiaries are already employed when registering with the scheme, but their income remains below the eligibility threshold. They usually face difficulties transitioning into employment, often ending up in jobs different from those initially sought, and are more likely to be part-time or shift based. In Belgium, more than half remain on benefits one year after registration, while only a small share transition into regular employment. Among MIS beneficiaries who find work, employment is concentrated in low-skilled service and manual occupations. Despite the differences between the two countries, digital matching tools should flexibly connect MIS beneficiaries to a wide range of job with nonetheless offer promising employment prospects.
AI and Digitalisation for Employment Support in Belgium and Greece
Abstract
Creating pathways into employment for MIS beneficiaries is crucial to ensure sustainable economic and social integration. The European Pillar of Social Rights calls on Member States to combine adequate income support with inclusive labour market measures to guarantee the active inclusion of people excluded from the labour market. In both Greece and Belgium, activation measures play an important role in the overall support provided to MIS beneficiaries. However, compared to other jobseekers, the labour market integration of MIS beneficiaries is particularly challenging due to the multiple and complex barriers they face.
This chapter leverages data from the registers described in Chapter 2 for the two countries to investigate the employability of MIS beneficiaries, their profiles and the barriers they face, the extent to which they register with the PES for job-search support, and, if they enter the labour market, the types of jobs and employers they access. Together with the data analysis in Chapter 4, which takes a more comprehensive view of career pathways, including participation in ALMPs, the findings discussed in this chapter have implications for the development of the concept for digital tools to improve job matching and career progression of MIS beneficiaries in both countries (described in detail in Chapter 5).
This chapter is structured as follows: Sections 3.2 and 3.3 present the analysis results based on the Greek and Belgian data respectively. Section 3.4 concludes by summarising the key assessments for each country.
The objective of this section is to empirically investigate the profile of MIS beneficiaries and their pathways into employment. Section 3.2.1 introduces the data used for the analysis and the steps involved in data preparation. Section 3.2.2 presents the characteristics of MIS households and individual beneficiaries, focussing on factors related to their employability, their registration patterns with DYPA, and additional information collected by DYPA upon registration, distinguishing GMI jobseekers from other jobseekers, including other vulnerable groups. Section 3.2.3 analyses the types of employment contracts held by GMI beneficiaries when they enter the labour market, considering the characteristics of both the jobs they take up and the employers who hire them.
3.2.1. The analysis draws on administrative data from three registers
The analysis in this report draws on administrative data from three sources: the GMI platform, the DYPA register of jobseekers, and the ERGANI employment register. The GMI platform provides detailed data on household composition and demographics, DYPA data contain information on employability, and ERGANI records employment relationships. Linking these registers through a pseudonymised personal identifier, which protect individual privacy, enables a comprehensive analysis of the labour market situation of GMI beneficiaries. While the three registers are described in detail in Chapter 2, this section outlines the data preparation process and the samples used for the analysis.
Data from the GMI platform include all successful applications1 from the introduction of the GMI system in February 2017 until December 2024. Generally, GMI applications need to be renewed every six months and are submitted at the household level, meaning that each application contains information on the main applicant as well as on all other household members. For household members without a tax identification number (AFM ID) – mostly children – only their number is reported in each application, with no additional information.
The preparation and cleaning of the GMI data involved two main steps. First, a small number of duplicate observations (defined as cases with the same pseudonymised ID and application date) were removed (0.01% of the full sample). Many of these cases involved minors who were registered with the same pseudonymised ID as the main applicant. In such cases, only the record of the main applicant was retained. In cases where two registrations for the same individual had the same start date but different end dates, the record with the latest end date was kept. Second, adjacent GMI applications – that is, applications starting less than one month after the end of the previous application – were merged into a single continuous GMI spell. In addition, each application requires applicants to provide a range of information. Because some of this information may change over time (e.g. employment status, earnings), only the information from the first application in the spell was retained when constructing GMI spells. The resulting dataset uses the GMI spell (rather than an individual application) as the unit of observation and contains approximately 2.2 million spells and 1 million distinct individuals.
Data from DYPA cover all jobseekers who registered with DYPA for at least one day between February 2017 and January 2025, regardless of whether they received the GMI. The registration data contained a very small share of duplicates with identical pseudonymised ID and registration date (0.15%). In such cases, the record containing the most complete information was retained or, if the amount of information was identical, the record with the earliest start date was kept. In the relatively rare cases (1.14%) where a registration ended due to either the merging of the registration card or failure to renew the card, and a new registration began within one day, these consecutive registrations were merged into a single continuous unemployment spell. The resulting dataset contains detailed information on approximately 3 million unique individuals and more than 15 million unemployment spells.
Finally, data from the ERGANI employment register cover all individuals who appear in either the GMI or DYPA datasets (or both). They include records of hirings, dismissals, and contract changes for individuals in salaried employment in Greece between March 2013 (the introduction of ERGANI) and December 2024. In addition, ERGANI contains information from the annual census of employment in firms. This information was used to construct employment spells for each worker, with a spell defined as a unique combination of pseudonymised ID, employer ID, and start date. The final dataset comprises 26.4 million employment spells of approximately 2.6 million individuals. While ERGANI captures employment in some active labour market policies (ALMPs), such as wage subsidies and public works, it does not include self-employment or public sector employment. The absence of data on the self-employed – estimated to account for more than 30% of total employment in Greece, well above the OECD average (OECD, 2024[1]) – is an important limitation. An additional limitation of the employment data is that they do not capture employment in the informal sector.
In this section, different samples are used depending on the specific question analysed:
Section 3.2.2 uses two samples. The first is a snapshot of all households and individuals aged 18‑64 with an active GMI spell in December 2024, providing a profile of recent GMI beneficiaries in Greece. The second consists of all DYPA unemployment spells starting between the national roll-out of the GMI on 1 February 2017 and 31 December 2024. Linking these unemployment spells with data from the GMI platform and ERGANI allows comparison of the characteristics of these spells across three main groups of jobseekers: GMI beneficiaries registered with DYPA, vulnerable jobseekers who are not GMI beneficiaries, and all other jobseekers. A jobseeker in the DYPA data is identified as a GMI beneficiary if the registered unemployment spell overlaps, at least in part, with a GMI spell.
Sections 3.2.3 and 3.2.4 use a sample of all employment spells in the ERGANI data starting between the national rollout of the GMI on 1 February 2017 and 31 December 2024. These data are used to examine the transitions of GMI beneficiaries and other jobseekers into the labour market and to assess the characteristics of these employment spells, both in terms of the employers hiring these jobseekers and the type of jobs.
3.2.2. Households and individual beneficiaries of the GMI are a heterogeneous group
GMI beneficiaries often live in single‑adult households
As the GMI is granted to households rather than individuals, each application contains information on all household members, not only the main applicant. Since the introduction of the GMI scheme, the number of beneficiaries – both individuals and households – peaked in the second half of 2018 and has declined steadily thereafter (Figure 3.1). Between February and April 2021, the number of applications dropped sharply due to the protection measures against the spread of coronavirus. These measures extended the validity of all GMI applications due for renewal automatically by three months.2 In the most recent month available in the data (December 2024), approximately 172 000 households and 240 000 individuals received GMI.
Number of individuals and households included in applications for the GMI, Feb. 2017 to Dec. 2024
Note: All applications expiring between February and April 2021 were automatically extended by three months as part of the protective measures against the spread of the coronavirus.
Source: OECD calculations based on data from the Greek Guaranteed Minimum Income (GMI) platform.
Both household composition and characteristics, as well as the level of benefit receipt, can significantly influence the labour market participation decisions of GMI beneficiaries and thus provide insights into their employability. For example, households with children may have reduced capacity to take up employment due to caring responsibilities, while high levels of benefit receipt could reduce the incentive to seek employment if not accompanied by activation requirements. In the sample of households with an active GMI application in December 2024, only around 16% were either a single adult with at least one child or an adult with a spouse and one or more children, and just 6.4% of all households were single‑parent households (Table 3.1). In contrast, most GMI beneficiaries live in single‑adult households and therefore do not appear to face care‑related barriers to employment. The particularly high share of single‑adult households was already noted in an earlier evaluation of the GMI scheme (Marini et al., 2019[2]) and may reflect the design of the scheme, which assesses eligibility based on household rather than individual income and provides higher combined allowances to two single adults than to a two‑person household.
With respect to benefit levels, the average monthly allowance is EUR 231.90, ranging from a minimum of EUR 10 to a maximum of EUR 972, with the amount varying as a function of household composition. This allowance is well below the statutory minimum wage of EUR 830 (as of December 2024) and is insufficient to secure subsistence on its own. Indeed, most households (83.2%) also received support from the Fund for European Aid to the Most Deprived (FEAD). This suggests that employment remains necessary to complement the GMI and ensure an adequate level of income.
Characteristics of households with an active GMI application in December 2024
Note: The sample includes households with an active GMI spell in December 2024.
Source: OECD calculations based on data from the Greek Guaranteed Minimum Income (GMI) platform.
In terms of geographical distribution of GMI, roughly half of the households are in the Attica and Central Macedonia regions, mirroring population density in these regions. However, the highest density of GMI beneficiaries is found in Western Greece (Figure 3.2).
Number of GMI households per thousand inhabitants, 2024
Note: Number of GMI households with an active GMI application in December 2024 per thousand inhabitants. Population data from Eurostat on total population by NUTS 2 region on 1 January 2024.
Source: OECD calculations based on data from the Greek Guaranteed Minimum Income (GMI) platform and Eurostat.
Women, Greek nationals and individuals aged over 50 are overrepresented among GMI beneficiaries
While analysing household characteristics provides initial information on the potential barriers to employment that GMI beneficiaries might face, more granular data on individual characteristics are needed to gain deeper insights into their labour supply.
Overall, GMI beneficiaries are more often women, Greek nationals, and aged over 50 (Figure 3.3). The average duration of a GMI spell (with consecutive applications counted as a single spell) is 26 months. At the time of their first GMI application, most beneficiaries (73.9%) declare themselves unemployed, while only 20% report being employed. The remaining beneficiaries are students, individuals not able to work, children, or elderly individuals. As employment status must be reported each time the GMI application is renewed, it is informative to examine the share of applicants who ever reported being employed within a spell. This share, at 25.1%, is slightly higher than the share reporting employment in their first application, indicating that at least some beneficiaries who were not employed at registration entered employment during the GMI spell. Importantly, this shows that while most GMI beneficiaries are unemployed, a non-negligible share is actually employed or become employed during their GMI spell. One possible reason is that, despite being in employment, the household income in the six months preceding the application was below the GMI eligibility threshold.
Analysing the profile of GMI beneficiaries who were employed at least once during their GMI spell (defined as consecutive applications to the scheme) and comparing it with that of beneficiaries who never reported being employed in any application within the spell reveals notable differences between the two groups (Annex Table 3.A.1. in Annex 3.A). Unsurprisingly, beneficiaries with a disability (i.e. reporting a disability degree between 1% and 100%) are far less likely to be among those employed at least once during the spell. In contrast, men, beneficiaries aged 30 to 50, those with only compulsory education, and non-Greek nationals are more likely employed.
Multiple factors can affect the employability of GMI beneficiaries. Their educational attainment is skewed towards the lower end of the distribution, with 49.1% having completed less than upper secondary education, compared to 19.7% of the Greek population aged 25‑64 (OECD, 2023[3]). This indicates that, on average, GMI beneficiaries have lower educational attainment than the general population. Disability is another factor influencing employability. At registration, individuals with disabilities are required to submit a certificate specifying their degree of disability. While the share of GMI beneficiaries with disabilities is relatively low (4.2%), their average degree of disability is high, at 71.8%. However, this is not surprising given that a certified disability degree of 67% is required to qualify for disability benefits (European Commission, 2025[4]). Indeed, data from the GMI platform show a clear spike in certified disability degrees at exactly 67%. Finally, the receipt of other pensions or benefits may also affect employability and willingness to enter the labour market. GMI data show that only a small fraction of beneficiaries receive additional unemployment or pension allowances in addition to the GMI allowance.
Since August 2018, as part of the third pillar of the GMI (activation), beneficiaries who are able to work are required to register with DYPA in order to continue receiving the GMI. This requirement applies to individuals aged 18‑65 who are not in employment and excludes certain groups, such as people with disabilities, full-time students, apprentices in vocational schools, and participants in ALMPs without an employment relationship. Registration with DYPA is a key measure to promote the labour market integration of GMI beneficiaries, as it gives them access to a range of employment services – from counselling to participation in ALMPs such as training, public works, and wage subsidies – designed to support their transition into employment.
In the sample considered (active spells in December 2024 of individuals aged 18‑64), about two in three GMI beneficiaries are registered with DYPA and therefore have access to the activation measures of the public employment services. While some beneficiaries not registered with DYPA may be recent applicants who have not yet registered, others may belong to one of the categories mentioned above that are exempt from the registration requirement. Figure 3.3 shows that people with disabilities and those who declare being employed at the time of their first GMI application are underrepresented among those registered with DYPA compared to their share among all GMI beneficiaries, reflecting the fact that they are generally not required to register. Importantly, this indicates that activation measures are more strongly targeted towards unemployed GMI beneficiaries and less so towards those who are employed, even though the latter could also benefit from support to move into better jobs. Figure 3.3 also shows that some groups of GMI beneficiaries who tend to be more employable, such as those with upper secondary education and Greek nationals with a likely better Greek language skills and country-specific human capital, are also more likely to register with DYPA. However, it is also important to note that GMI beneficiaries registered with DYPA tend to have longer GMI spells than those who are not registered (28 months compared to 21 months).
Distribution of selected characteristics among GMI beneficiaries, by registration status with DYPA, December 2024
Note: The sample includes individuals aged 18‑64 with an active GMI spell in December 2024. Red squares show the share of people with the referred characteristics in the whole population of GMI beneficiaries. The dark blue bars show the share of people with the referred characteristics among GMI beneficiaries that are registered with DYPA. The light blue bars show the share of people with the referred characteristics among GMI beneficiaries that are not registered with DYPA. Shares are calculated within each of the eight broad categories shown in the figure. If a category of GMI beneficiaries registered with DYPA were represented in proportion to their share among all GMI beneficiaries, the length of the dark blue bars would coincide with the red squares.
Source: OECD calculations based on data from the Greek Guaranteed Minimum Income (GMI) platform and the Greek Public Employment Services (DYPA).
Many jobseekers may turn to the GMI after unsuccessful job search and potentially after exhausting their entitlement to unemployment benefits. Approximately 85% of all GMI beneficiaries registering with DYPA, register already before applying for the GMI, and more than 55% had registered over a year earlier (Figure 3.4). This group of jobseekers likely includes many long-term unemployed who require intensive support to re‑enter employment. In contrast, only 15% of GMI beneficiaries register with DYPA after applying for the GMI.
Time between registration with DYPA and application for the GMI, December 2024
Note: The sample includes individuals with an active GMI spell in December 2024 who have registered with DYPA.
Source: OECD calculations based on data from the Greek Guaranteed Minimum Income (GMI) platform and the Greek Public Employment Services (DYPA).
GMI beneficiaries are more likely than other jobseekers to belong to a vulnerable group and to seek employment in low-skilled occupations
Considering all registrations of individuals aged 18‑64 with DYPA between 1 February 2017 and 31 December 2024, approximately 21% were GMI beneficiaries.3 When registering with DYPA, extensive individual information on jobseekers is collected, such as the occupation they are seeking and whether they belong to a vulnerable or special social group.4 Vulnerable groups include jobseekers whose access to employment is limited by physical, mental, or behavioural factors, specifically persons with disabilities, former addicts, young offenders, and former prisoners. Special social groups include jobseekers who face labour market difficulties due to personal or social circumstances, such as immigrants, refugees, asylum seekers, single parents, victims of domestic violence, homeless individuals, and people from marginalised cultural groups (such as Roma or Pomaks). They also include other specific groups, such as victims of human trafficking, transgender individuals, and young adults exiting child protection facilities.
About 12% of GMI beneficiaries registered with DYPA report belonging to a special category of vulnerable jobseekers, compared to only 5% of all jobseekers registering with DYPA. These groups may face particular barriers to employment – for example, due to limited language skills (migrants and asylum seekers), cultural or social stigma, or specific constraints such as caring responsibilities for single parents. The most common vulnerability categories among GMI beneficiaries are special social groups such as Roma or Pomaks, migrants, persons with disabilities, and asylum seekers or refugees (Figure 3.5). Less common are categories such as former prisoners and young offenders, homeless individuals (although their share is higher among GMI beneficiaries than among other vulnerable jobseekers), single parents, former substance users and individuals undergoing rehabilitation, and female victims of gender-based violence.
Share of GMI beneficiaries and other jobseekers in different vulnerable categories, 1 Feb. 2017 and 31 Dec. 2024
Note: The sample includes registrations with DYPA between 1 Feb. 2017 and 31 Dec. 2024 in which jobseekers reported being in at least one vulnerability group. Of these registrations, 21.1% are overlapping with a GMI spell. Jobseekers may report multiple special categories.
Source: OECD calculations based on data from the Greek Guaranteed Minimum Income (GMI) platform and the Greek public employment service (DYPA).
GMI beneficiaries looking for a job with the help of DYPA are more likely to look for low-skilled jobs than other jobseekers registered with DYPA (Figure 3.6).5 In particular, looking for jobs on elementary occupations and as craft and related trades workers is more common among GMI beneficiaries compared to other jobseekers, while aiming to be a professional or a services and sales worker much less common. Nevertheless, vulnerable jobseekers who are not GMI beneficiaries focus on low-skilled occupations even more than GMI beneficiaries.
GMI beneficiaries exit unemployment to employment at lower rates than other jobseekers, including other vulnerable jobseekers. Twelve months after registering with DYPA, roughly 60% of all jobseekers move into employment, compared to just under 20% of GMI beneficiaries and 35% of other vulnerable jobseekers. This gap may be partly explained by the fact that some individuals register for the GMI after having already been unemployed for some time, and the longer jobseekers remain unemployed, the more difficult it becomes to re‑enter employment. Moreover, Annex Table 3.A.2. in Annex 3.A shows that certain groups of jobseekers such as older individuals and those with lower levels of education who typically face higher barriers to employment are overrepresented among GMI beneficiaries.
Vulnerable jobseekers, GMI beneficiaries and all registered jobseekers, by occupation sought
Note: The sample includes registrations with DYPA between 1 Feb. 2017 and 31 Dec. 2024. One‑digit ISCO‑08 occupations refer to the occupations sought by jobseekers.
Source: OECD calculations based on data from the Greek Guaranteed Minimum Income (GMI) platform and the Greek public employment service (DYPA).
Rates of exit to employment after registration with DYPA by jobseeker type, jobseekers registered with DYPA between Feb. 2017 to Dec. 2022
Note: Data includes registrations with DYPA between 1 February 2017 and 31 December 2022.
Source: OECD calculations based on data from the Greek Guaranteed Minimum Income (GMI) platform, the Greek public employment service (DYPA) and the employment register ERGANI.
To improve the matching of GMI beneficiaries with labour demand, it is informative to analyse how demand is expected to evolve across occupations. Skills forecasts for Greece suggest that labour demand will be strongest in the two ISCO‑08 major groups “Professionals” and “Service and Sales Workers” (CEDEFOP, 2025[5]). These occupations, along with “Clerical Support Workers” and “Plant and Machine Operators and Assemblers”, are projected to experience a net increase in the number of jobs until 2035. Other occupations are expected to see a small to moderate net decline in the number of jobs, with “Skilled Agricultural, Forestry and Fishery Workers” being particularly negatively affected. In “Elementary Occupations” – where both GMI beneficiaries and other vulnerable jobseekers are more likely to seek employment – labour demand is modest, but the overall reduction in jobs is also expected to be small. From a job matching perspective, this implies that while GMI beneficiaries are somewhat more likely than other vulnerable jobseekers to seek employment in occupations with sustained labour demand, they remain underrepresented among jobseekers targeting occupations in high demand. Therefore, for some GMI beneficiaries, re‑skilling or upskilling may be required to better align with future labour market needs.
3.2.3. GMI beneficiaries are more often employed by small employers and in the accommodation and food service sector
This and the next section focus on the employment relationships of GMI beneficiaries when they enter the labour market. Specifically, this section analyses the profile of employers hiring individuals who are either current or former beneficiaries of the GMI. The aim is to identify categories of employers that are more likely to hire GMI beneficiaries. From a job matching perspective, this information is useful as it highlights the types of employers whose vacancies could be prioritised for matching. It is important to note, however, that the analysis in this and the following section is based on realised employment matches, that is, instances where a GMI beneficiary was actually hired by an employer. Potential matches that could in theory occur but did not materialise (e.g. because the GMI beneficiary and the firm did not connect) are not captured. In addition, it is important to stress that these matches do not necessarily reflect high-quality or stable jobs, but they may still serve as a stepping stone into the labour market. Digital tools should therefore provide GMI beneficiaries with opportunities to progress towards better-quality jobs.
Figure 3.8 shows the distribution of employer characteristics across employment contracts started by GMI beneficiaries and other jobseekers. GMI beneficiaries are more often employed in micro and small firms and less frequently in medium-sized and large firms compared to the average jobseeker registered with DYPA. By sector of activity, both GMI and other vulnerable jobseekers are disproportionately employed in accommodation and food service activities – nearly 37% and 38% respectively, compared to around 33% for all jobseekers.
More generally, accommodation and food service activities account for a large fraction of all employment relationships entered by GMI beneficiaries and jobseekers registered with DYPA. Given that many jobseekers, including GMI beneficiaries, find employment in this industry, job-matching tools should ensure adequate coverage of vacancies in this sector, while also taking into account the quality of the employment opportunities offered. GMI beneficiaries are also slightly overrepresented in administrative and support service activities and in construction. In addition, they are more likely to be employed in public administration and defence; compulsory social security, possibly reflecting greater participation in public works compared to other jobseekers (Chapter 6 will provide a more detailed analysis of the participation of GMI beneficiaries in ALMPs to study this question). GMI beneficiaries are also more likely to be employed in sole proprietorships – businesses owned by a natural person – compared to other jobseekers, including vulnerable jobseekers. They are also more likely to be employed by local government authorities, such as municipalities, again possibly due to participation in public works programmes.
Distribution of employer characteristics for employment contracts by jobseeker type, Feb. 2017 to Dec. 2024
Note: Unit of analysis is the employment spell. Data include employment spells starting between 1 February 2017 and 31 December 2024. Employer size refers to the number of employees in the year the jobseeker was hired (i.e. at the start of each employment spell). Firm size categories follow the OECD/Eurostat standard: Micro (1‑9 employees), Small (10‑49), Medium (50‑249), and Large (250 or more). Employer sector is classified at the NACE Level 1 (Section). Only sectors accounting for at least 3% of all employment spells are shown separately; all other sectors are grouped under “Other sectors”. Legal type of employer is reported only for categories with at least 3% of employment spells; all other types are grouped under “Other legal forms.”
Source: OECD calculations based on data from the Greek Guaranteed Minimum Income (GMI) platform, the Greek public employment service (DYPA) and the employment register ERGANI.
3.2.4. GMI beneficiaries are more likely to work in part-time and shift jobs with lower hourly wages
This section discusses the characteristics of the jobs that GMI beneficiaries and registered jobseekers hold when entering employment. This analysis uses the employment contracts recorded in ERGANI between 1 February 2017 and 31 December 2024, distinguishing between those of GMI beneficiaries (i.e. contracts starting during or after a GMI spell) and those of other jobseekers registered with DYPA. When entering employment, GMI beneficiaries tend to earn less than other jobseekers, including vulnerable jobseekers (Table 3.2). Part of this earnings gap can be explained by differences in working hours: GMI beneficiaries are less likely to work full-time (47% compared to 55% of all jobseekers) and work fewer hours per week on average (26.1 compared to 28.4). The distribution of contractual weekly hours across jobseeker types shows that these differences are largely driven by GMI beneficiaries working more often with contracts of fewer than 20 hours per week (Annex Figure 3.A.1 in Annex 3.A).
Shares and averages of job characteristics, by jobseeker type, Feb. 2017 to Dec. 2024
Note: The unit of analysis is the employment spell. Data includes employment spells starting between 1 February 2017 and 31 December 2024.
Source: OECD calculations based on data from the Greek Guaranteed Minimum Income (GMI) platform, the Greek public employment service (DYPA) and the employment register ERGANI.
Although part-time work may be an intentional choice, making employment more accessible to individuals who cannot work full-time – for example, single parents with caring responsibilities or jobseekers with certain types of disability – it is important to note that these employment contracts are typically associated with lower hourly wages. Similarly, GMI beneficiaries also appear to be more likely than other jobseekers to enter shift work. While shift work can also be an informed and deliberate choice, it is often associated with higher health and safety risks (Saint-Martin, Inanc and Prinz, 2018[6]).
Using linear regression can help identify which characteristics of employers and jobs are associated with better job matches. For this purpose, hourly wages are used as a measure of match quality.6 The results, reported in Annex Figure 3.A.2. in Annex 3.A, show that after controlling for industry and occupation, employment contracts in medium and large firms, as well as full-time contracts, are associated with higher hourly wages. This information could be incorporated into tools designed to improve the matching of GMI beneficiaries with labour demand, for example, by prioritising job matches with medium and large firms and with contracts offering full-time employment.
Additional characteristics of the employment contracts of GMI beneficiaries that can be examined using ERGANI data include the occupation. While the final part of Section 3.2.2 presented the occupations in which GMI beneficiaries registered with DYPA seek employment, ERGANI data make it possible to explore the occupations in which they actually find employment. Figure 3.9 shows these ISCO‑08 two‑digit occupations in which more than 1% of GMI beneficiaries are employed and compares the share of GMI beneficiaries finding employment in these occupations with the share seeking employment in these occupations when registering with DYPA.
Shares of total employment and unemployment spells, by ISCO‑08 occupation, 2‑digit level
Note: Occupations are classified according to the ISCO‑08 at the 2‑digit level. Data includes only occupations where GMI applicants account for more than 1% of employment spells. “Hired” refers to the distribution of occupations for which GMI applicants obtained a contract; the sample includes employment spells starting between 1 February 2017 and 31 December 2024. “Seeking” refers to the distribution of occupations sought by GMI applicants while registered as unemployed; the sample includes all unemployment spells recorded during the same period.
Source: OECD calculations based on data from the Greek Guaranteed Minimum Income (GMI) platform, the Greek public employment service (DYPA) and the employment register ERGANI.
Overall, there is a degree of mismatch between occupations that GMI beneficiaries target and obtain. Some occupations are sought by GMI applicants far more often than they result in hires, and vice versa. For example, more than one in four GMI beneficiaries find employment as personal service workers (that is, workers providing services related to, e.g. housekeeping, catering, and hospitality) while only one in ten actively seek employment in these occupations when registering with DYPA. In contrast, the shares of GMI beneficiaries seeking employment in building and related trades, general and keyboard clerks, and labourers in mining, construction, manufacturing, and transport are greater than the shares finding employment in these occupations.
The reasons for this mismatch can be multifaceted. For example, GMI beneficiaries formerly employed in construction, building, or mining may indicate these occupations when registering with DYPA. However, as shown in Annex Table 3.A.2. in Annex 3.A, GMI beneficiaries include a relatively high share of individuals aged over 50, who generally face greater barriers to employment, particularly in physically demanding occupations. Another reason may be that some GMI beneficiaries search for jobs in occupations (e.g. office clerks) for which they lack the necessary skills and competences.
These findings have two main implications for job matching. First, they highlight the need for flexible algorithms that match jobseekers not only to jobs within the same occupation but also across occupations that, while different, have overlapping skill requirements. Second, they suggest that some GMI beneficiaries may need to change occupations to find employment. Tools designed to match GMI beneficiaries with labour demand should therefore account for potential jobs that could be attained with appropriate reskilling and upskilling through ALMPs.
This section examines the profile of Belgian minimum income scheme (Revenu d’intégration /leefloon) beneficiaries and the trajectories they follow into employment This discussion on the Belgian system is organised into three main parts. Section 3.3.1 presents the profile of MIS beneficiaries, documenting trends over time and highlighting differences across regions, socio-demographic groups, and household types. Section 3.3.2 shifts the focus to employers, describing the industries and sectors in which beneficiaries who exit into work are most often employed, and how these patterns vary across regions and educational attainment. Section 3.3.3 then turns to the jobs themselves, examining the occupational composition of employment among MIS beneficiaries, and comparing their outcomes to those of unemployment benefit recipients. Together, these sections provide a comprehensive picture of who MIS beneficiaries are, where they find work, and the types of jobs they enter.
3.3.1. Data used and analytical approach
This analysis draws on aggregate tables produced from administrative data compiled by the Crossroads Bank for Social Security (CBSS). The focus is on individuals receiving the minimum income support leefloon/revenu d’intégration (RIS) and their subsequent transitions into employment or other benefit schemes (henceforth: MIS). To capture these trajectories, the CBSS generated detailed cross-tabulations that map the profiles and pathways of recipients by key characteristics such as region, education, age, household composition, migration background and gender.7 For those who moved into employment, the analysis also examines the industries in which they found jobs and compares their profiles to those who did not enter employment. These cross-tabulations were constructed using microdata from a range of sources, including local welfare offices (CPAS), regional Public Employment Services, and national-level registers such as the National Social Security Database (ONSS).8 The dataset covers all new entrants to minimum income support across all quarters from 2011 to 2020, ensuring sufficient sample sizes and protecting confidentiality. Outcomes are observed at two points following programme entry, after four quarters and after 12 quarters, allowing for an assessment of how recipients’ situations evolve over time.
Complementary insights are drawn from secondary sources, including statistical reports from the Federal Public Service for Social Integration (SPP IS), regional monitoring reports, and previous analyses by the OECD and other research institutions.
3.3.2. The number of MIS beneficiaries has been rising, with stark regional differences
Number of beneficiaries in Belgium has been rising steadily
Belgium’s minimum income scheme is a last-resort, means-tested benefit providing a financial safety net to individuals lacking sufficient resources. In contrast to Greece’s GMI (granted at the household level), the Belgian MIS is legally an individual entitlement, embedded in the “right to social integration” framework (Law of 26 May 2002). In practice, a household’s composition and income determine the benefit: cohabiting adults receive a lower amount per person than single adults, and a higher rate applies to single adults with dependent children (OECD, 2024[7]). The minimum age for MIS is 18 (younger in special cases such as emancipated minors with children), and claimants must first exhaust other social protection rights (unemployment insurance, pensions, disability benefits, etc.).
Belgium’s minimum income scheme has expanded over time, both in absolute and relative terms. The number of beneficiaries more than doubled between 2003 and 20239 – from around 74 000 to 158 700 on a monthly average in 2023 – with particularly sharp increases after 2015. Figure 3.10 shows that the share of the population receiving RSA/leefloon has also steadily increased, with 1.3% of the population receiving the benefit in 2023, the highest in recent history. Preliminary estimates from 2025 suggest the number of beneficiaries has increased slightly to 167 800 in January (POD Maatschappelijke Integratie, 2025[8]). Reforms in 2015 that shortened the duration of youth unemployment benefits led many young people to apply for RIS, contributing to a surge in cases. At the start of 2024, approximately 1.6% of Belgium’s population depended on a RIS or equivalent social assistance benefit.
Share of the population receiving social assistance, 2003-2023
Source: Federaal Planbureau (2025[9]), Leefloners (i05) [Indicator: Percentage of the Belgian population receiving social integration income], https://indicators.be/nl/i/G01_GMI/Leefloners.
There are strong regional variations in MIS receipt
The regional distribution of beneficiaries is very uneven and differences have increase over time: as of 2023, only 0.6% of residents in Flanders received MIS, compared to 2% in Wallonia and 3.7% in the Brussels-Capital Region (Figure 3.11). In absolute terms, nearly half of all beneficiaries live in Wallonia, about 30% in Brussels and 25% in Flanders. This reflects the overrepresentation of French-speaking regions among recipients, despite Flanders accounting for more than half of the national population. There are also strong local concentrations within regions. Larger cities and former industrial centres have the highest recipiency rates of MIS. For example, in 2025, in the port city of Oostende (Flanders) about 2.7% of its inhabitants received the MIS and other urban centres like Gent and Charleroi had similarly high incidences. By contrast, some affluent and largely rural municipalities (e.g. in parts of Flemish Brabant or the South of Luxembourg province) record negligible numbers of MIS beneficiaries. The spatial pattern mirrors socio‑economic conditions: areas with a legacy of industrial decline or higher poverty see greater reliance on minimum income schemes.
Share of the population receiving minimum income scheme (MIS), by region, 2003-2023
Source: Federaal Planbureau (2025[9]), Leefloners (i05) [Indicator: Percentage of the Belgian population receiving social integration income], https://indicators.be/nl/i/G01_GMI/Leefloners.
Trajectories of MIS beneficiaries differ one year after registration
Just over half of new entrants to minimum income support are still receiving the benefit one year after registration, highlighting the persistence of reliance on the scheme for many recipients. Figure 3.12 presents the activity status of individuals 12 months after entering MIS between 2011 and 2020. Among those who exit MIS, about 13% secured employment in the regular labour market, while a further 6% participated in Article 60 public works schemes,10 together representing nearly one in five recipients making some form of transition into work. By contrast, 9% move onto unemployment benefits, suggesting that for many beneficiaries, MIS functions as a stepping stone within the broader social protection system rather than a direct bridge into sustainable employment. The remaining share either shift into other types of support or leave the register altogether, underscoring the diversity of pathways following an MIS spell.
Activity status of new minimum income scheme (MIS) spells, four quarters after registering, 2011-2020
Source: OECD calculations based on administrative data received from the Crossroads Bank for Social Security (CBSS).
Profiles of those who remain in MIS
Persistence in MIS after four quarters varies substantially across demographic groups. By region, recipients in Brussels (61.5%) and Wallonia (52.6%) remain on benefits at higher rates than the national average (51.8%), while those in Flanders (44.8%) exhibit lower persistence (Figure 3.13). Gender differences are limited, with women (53.7%) slightly more likely than men (50.0%) to continue receiving support one year after they started receiving the MIS. Educational attainment shows a clearer gradient with individuals with lower levels of education (ISCED 0‑2, 52.1%) being more likely to remain on benefits than those with mid-level (ISCED 3‑4, 43.3%) or higher qualifications (ISCED 5‑8, 44.9%). Household composition is another key differentiator: recipients living with others display the highest persistence (55.5%), followed by those living alone (50.4%), while single parents with dependent children show lower persistence (46.2%). Migration background reveals notable contrasts: first-generation foreign-born recipients have the highest persistence (61.6%), well above the average, whereas native‑born Belgians whose parents were also born with the Belgian citizenship (referred to as third-generation or later) are less likely to remain in the scheme (43.7%). Age differences are also evident: young adults aged 18‑29 show the highest persistence (55.5%), while rates decline among those aged 30‑39 (47.9%) and 40‑49 (46.1%), before rising modestly again for recipients aged 50‑64 (48.9%).
Share of people who remain in MIS one year after registering, all new spells between 2011-2020
MIS: Minimum income scheme.
Note: The horizontal line indicates the overall average. Migration background categories are defined using administrative data on country of birth and nationality of the individual and their parents. First-generation foreign-born refers to individuals born abroad with foreign nationality at birth. First-generation naturalised refers to those born abroad who later acquired Belgian nationality. Second generation refers to individuals born in Belgium with at least one foreign-born parent. Third generation or more refers to those born in Belgium to Belgian-born parents, often considered the “native” population in the statistics. Educational attainment is based on International Standard Classification for Education (ISCED) codes (0‑2, 3‑4 and 5‑8 respectively).
Source: OECD calculations based on administrative data received from the Crossroads Bank for Social Security (CBSS).
Profiles of those who transition into employment
Education, region, and household context are key determinants of successful labour market integration, while gender plays little role in the liklihood of finding employment (Figure 3.14).
Transitions from MIS into regular employment differ markedly across groups. On average, 13.2% of recipients find work in the regular labour market within a year of registering, but some groups perform well above or below this benchmark. Across regions, Flanders stands out with the highest integration into work (19.1%), while Brussels and Wallonia show considerably lower integration into the labour market (10.3% and 10.7% respectively). Educational attainment shows the expected effect on employment: almost one in four beneficiaries (23.7%) with higher qualifications (ISCED 5‑8) secure work one year after registering, compared with 18.4% among those with mid-level education (ISCED 3‑4) and only 11.8% among those with the lowest attainment (ISCED 1‑2).
Shares of former MIS spells (pooled 2011-2020) that are in formal employment 4 quarters after registering
MIS: Minimum income scheme.
Note: The horizontal line indicates the overall average. Migration background categories are defined using administrative data on country of birth and nationality of the individual and their parents. First-generation foreign-born refers to individuals born abroad with foreign nationality at birth. First-generation naturalised refers to those born abroad who later acquired Belgian nationality. Second generation refers to individuals born in Belgium with at least one foreign-born parent. Third generation or more refers to those born in Belgium to Belgian-born parents, often considered the “native” population in the statistics. Educational attainment is based on International Standard Classification for Education (ISCED) codes (0‑2, 3‑4 and 5‑8 respectively).
Source: OECD calculations based on administrative data received from the Crossroads Bank for Social Security (CBSS).
Differences by household composition also emerge. Recipients with dependent children are more likely to transition into work (15.8%) than those living alone (13.7%) or with other adults (11.1%). Migration background shows a mixed pattern: Belgians whose parents were also born with the Belgian citizenship (referred to as third-generation or later) exceed the average (reaching 15.7%), whereas first- and second-generation migrants record lower rates of finding employment (11.6% and 12.0% respectively). Gender differences are negligible, with men and women virtually identical. Age follows an inverted U-shape: transitions are weakest among the youngest (aged 18‑29 years, at 12.4%) and oldest groups (aged 50‑64, at 10.9%), while recipients aged 30‑39 and 40‑49 achieve higher rates of entry into work (14.9% and 15.2% respectively).
3.3.3. Employers who hire MIS beneficiaries are predominantly in the service sector
Service sectors as the main gateway to employment for MIS beneficiaries
MIS beneficiaries who find employment one year after exit were dispersed across a wide range of service‑oriented sectors, with limited participation in traditional goods-producing activities (Figure 3.15). The largest shares were observed in “other services” (explained in detail below) (35.8%) and administrative and support services (31.2%), together accounting for more than two‑thirds of all placements. Employment in trade, transport and hospitality also employed a sizeable group of beneficiaries (24.6%), while only a small minority entered construction (4.0%) or industry and production of goods (4.4%).
Compared to national employment patterns, this represents a very distinct profile. In Belgium overall, “other services” dominate employment (50.1%), while administrative and support services make up just 10.3% (Figure 3.15). MIS beneficiaries are more concentrated in administrative and support services, which is also true, to a lesser extent, for former unemployment benefit recipients (28.5%). Similarly, trade, transport and hospitality play a slightly larger role for MIS beneficiaries (24.6%) than both the national average (21.2%) and former UB recipients (21.8%), while traditional sectors such as industry and construction absorb much lower shares of MIS beneficiaries (together just 8.4%) than in the national workforce (18.4%).
A particularly striking finding is the prominence of administrative and support service activities (NACE N), which alone accounts for nearly one‑third of all jobs taken up by MIS beneficiaries. Unlike the broad “other services” category (including 8 NACE categories), NACE N represents a single, clearly defined set of activities, highlighting its central role in the labour market integration of this group. This sector covers a diverse range of functions, from temporary work agencies and employment placement services to cleaning and facility management, security, and office support services. The concentration of beneficiaries in activities such as temporary agency work, general cleaning of buildings and services to buildings and landscapes illustrates how re‑employment often occurs in roles that are highly accessible, characterised by relatively low formal entry requirements and where employers frequently rely on flexible or outsourced labour. The weight of NACE N underscores both the importance of these intermediary and support functions in absorbing jobseekers and the reliance of MIS beneficiaries on entry-level opportunities in sectors marked by high turnover and externalisation of services.
The other services category (NACE codes included are J‑U, excluding N) is broad, covering activities ranging from information and communication, finance, real estate and professional services to education, health and social work, arts, recreation and public administration. The high representation of this group indicates that MIS beneficiaries are not only entering elementary services jobs but also positions in publicly supported or community-oriented activities such as health and social care, which are sectors with steady demand.
By contrast, only limited numbers of beneficiaries transitioned into construction (NACE F) and industry and goods production (NACE A‑E). These sectors are often characterised by stricter qualification requirements, higher physical demands and slower job turnover. While they represent important avenues for sustainable employment, the low shares point to potential barriers in connecting MIS beneficiaries to vacancies in these areas.
Sectors of employment 12 months after minimum income scheme (MIS) or unemployment benefit (UB) entry (2011‑2020), benchmarked against 2024 national averages
Note: Data for former MIS and former UB include spells between 2011-2020 and tracks their status one year after registration.
Source: OECD calculations based on administrative data received from the Crossroads Bank for Social Security (CBSS) and StatBel Statistics on establishment units https://statbel.fgov.be/en/themes/datalab/statistics-establishment-units for the national average.
Industries of employment of former MIS beneficiaries differ by region and education
While MIS beneficiaries are generally concentrated in services, the relative weight of sectors differs across regions. In Brussels, placements are strongly oriented towards other services (41.6%), well above the national average, while administrative and support services (24.5%) employ less MIS benefiacires (Figure 3.16). These patterns may partly reflect the structure of local labour demand and vacancies, which are not captured in this analysis. This may be explained by the capital’s labour market structure, where public services, education and healthcare account for a substantial share of employment (OECD, 2023[10]). In Wallonia, the distribution is more balanced: beneficiaries enter administrative and support services (26.6%) and other services (38.9%) at levels broadly in line with the overall average, with a somewhat higher-than-average share in construction (5.2%). By contrast, Flanders stands out for its stronger reliance on administrative and support services (36.8%), significantly above the national benchmark, while other services (31.8%) are less dominant. Flanders also registers the highest integration into trade, transport and hospitality (22.9%), suggesting that retail and logistics play a more prominent role in absorbing beneficiaries there.
Share of former MIS beneficiaries employed in each sector/industry 4 quarters after registering, pooled 2011-2020
MIS: Minimum Income Scheme.
Note: Bars show the share of former minimum income scheme (MIS) beneficiaries employed in each industry within the subgroup, totalling to the entire population in employment that formerly receiving MIS. Dots indicate the overall average share employed in that industry across all groups as a benchmark.
The NACE groupings used in this analysis are based on the following classification: Industry and goods production (Sections A–E: Agriculture, forestry and fishing; Mining and quarrying; Manufacturing; Electricity, gas, steam and air conditioning supply; Water supply, sewerage, waste management and remediation activities); Construction (Section F); Trade, transport and hospitality (Sections G–I: Wholesale and retail trade, including repair of motor vehicles; Transportation and storage; Accommodation and food service activities); Administrative and support services (Section N); and Other services (Sections J–U, excluding N: including information and communication, financial and insurance activities, real estate, professional and scientific activities, public administration, education, health and social work, arts, entertainment, and other service activities). Educational attainment is based on International Standard Classification for Education (ISCED) codes (0‑2, 3‑4 and 5‑8 respectively).
Source: OECD calculations based on administrative data received from the Crossroads Bank for Social Security (CBSS).
Differences are also marked across education levels, and these reinforce patterns of labour market segmentation. Among those with the lowest qualifications (ISCED 0‑2), employment is concentrated in administrative and support services (34.1%) and other services (30.3%), with both shares closely aligned with national averages. Beneficiaries with mid-level education (ISCED 3‑4) show a similar pattern but with slightly higher integration into other services (39.0%), above the overall average. The most distinctive pattern emerges among the highly educated (ISCED 5‑8), where nearly 60% of beneficiaries find work in other services, far exceeding the national benchmark. This suggests that higher education opens access to more knowledge‑intensive and professional segments of the service economy (e.g. information, finance, education, health), while reliance on administrative and support services (20.3%) and trade, transport and hospitality (15.7%) is comparatively lower.
Taken together, these results highlight that although service activities dominate across the board, regional economic structures and educational attainment shape where beneficiaries are most likely to find jobs. Flanders channels a relatively large share of beneficiaries into administrative support and logistics, Wallonia into construction, and Brussels into broader public and social services. Similarly, higher qualifications are associated with stronger integration into knowledge‑intensive service activities, while lower qualifications link more directly to entry-level administrative support roles. These patterns point to both opportunities and vulnerabilities: while services are key drivers of employment, the concentration of less-educated beneficiaries in high-turnover segments such as cleaning or agency work may limit long-term employment stability.
3.3.4. Former MIS beneficiaries predominantly find employment in manual occupations
One year after the beginning of their MIS spell, beneficiaries who transition into employment are predominantly found in manual occupations (65%), while around a third take up positions as non-manual employees (34%). Only a very small share enter jobs as civil servants (less than 1%), underscoring the strong concentration of re‑employment in private sector manual and non-manual work. By comparison, the national workforce is much more evenly distributed across occupation types, with 33% in manual jobs, 54% in non-manual jobs, and nearly 13% in civil service positions. This contrast highlights how MIS beneficiaries’ re‑employment pathways are skewed towards manual work and under-represent public sector opportunities More detailed information on the types of jobs taken up by former MIS beneficiaries, information that would be highly relevant for developing a job-matching tool, is available in Belgian administrative registers, but could not be used in this analysis due to the lack of direct access to the underlying micro-data.
Employment pathways of former MIS beneficiaries are heavily weighted towards manual service roles
Across all sectors, the majority of MIS beneficiaries find jobs in manual (blue‑collar) occupations, but important differences emerge across industries. The administrative and support services sector stands out as the single largest employer of blue‑collar workers, absorbing more than a quarter (27%) of all beneficiaries. Trade, transport and hospitality also play a substantial role, with nearly 17% in manual jobs and a further 11% in non-manual roles. By contrast, in other services, the balance tilts more towards white‑collar employment (15%) than manual work (13%), reflecting the presence of jobs in education, health, and public services alongside lower-entry service occupations. Traditional sectors such as industry and construction account for only modest shares of overall employment, and within them the bulk of jobs are manual. Civil service positions represent only a very small fraction overall, concentrated in other services.
Compared with former unemployment benefit recipients, MIS beneficiaries are considerably more concentrated in blue‑collar occupations, particularly in support services and lower-entry service sectors (Figure 3.17). Former unemployment benefit recipients, by contrast, have a much higher representation in white‑collar roles, highlighting a clear occupational divide between the two groups.
Types of occupations 12 months after MIS or unemployment benefit entry (2011‑2020), benchmarked against 2024 national averages
MIS: Minimum income scheme, UB: Unemployment benefit.
Source: OECD calculations based on administrative data received from the Crossroads Bank for Social Security (CBSS) and StatBel Statistics on establishment units https://statbel.fgov.be/en/themes/datalab/statistics-establishment-units for the national average.
Employment outcomes differ between former MIS and unemployment benefit recipients
On average, only 13% of former MIS beneficiaries are in employment 12 months after exit, compared with 39% of former unemployment benefit recipients. Despite this difference in overall levels, the trends across groups are broadly similar for both populations (Figure 3.18): outcomes are systematically stronger among those with higher education, weaker among older age groups, and more favourable in Flanders than in Brussels or Wallonia. Looking at deviations from the mean, some differences stand out. Among former MIS beneficiaries, employment rates are relatively better for the prime‑age population (aged 30‑49) and for those with tertiary education, while both the youngest and oldest perform below average. For former unemployment benefit recipients, the gradient by education is even steeper and younger beneficiaries are relatively more successful, whereas older workers are substantially under-represented in employment.
Note: Educational attainment is based on International Standard Classification for Education (ISCED) codes: Low: 0‑2, Mid: 3‑4 and High 5‑8.
Source: OECD calculations based on administrative data received from the Crossroads Bank for Social Security (CBSS).
Occupational composition of former MIS beneficiaries differs across regions and education levels
The occupational composition of jobs taken up by former MIS beneficiaries differs markedly across both regions and education levels (Figure 3.19). In Brussels, beneficiaries in services are more often employed as non-manual employees, particularly in “other services,” where administrative, educational, and health-related roles dominate compared to other regions.11 By contrast, in Wallonia and Flanders, employment within the same sectors relies more strongly on manual work, notably in administrative and support activities as well as in trade, transport and hospitality.
Differences are even more pronounced when considering educational attainment, though in the expected direction. Among the low- and mid-qualified (ISCED 0‑2 and ISCED 3‑4), employment is largely concentrated in manual occupations, particularly in administrative support, cleaning, and trade‑related jobs. However, beneficiaries with higher education (ISCED 5‑8) exhibit a sharply distinct profile. Nearly 60% of this group finds employment in other services, with a very high share of these roles classified as non-manual employee jobs.
Note: Educational attainment is based on International Standard Classification for Education (ISCED) codes: Low: 0‑2, Mid: 3‑4 and High 5‑8.
Source: OECD calculations based on administrative data received from the Crossroads Bank for Social Security (CBSS).
Building on administrative data from Greece and Belgium, this chapter analyses the profile of MIS beneficiaries, their registration patterns with the PES, their transition rates into the labour market and, once employed, the types of jobs they take up and the employers who hire them.
Based on the analysis, the key findings for Greece are:
GMI beneficiaries are more often women, Greek nationals, and aged over 50. A significant share of GMI beneficiaries are already employed when applying for the GMI. Despite being employed, their household income remains below the eligibility threshold. Digital tools could help these individuals identify additional job opportunities to complement their existing employment and explore career pathways towards better-quality jobs.
About two in three GMI beneficiaries are registered with DYPA, and most of them were already registered before applying for the GMI. Compared to other jobseekers registered with DYPA, GMI beneficiaries more often declare belonging to a vulnerable group, and they are disproportionately older and with only compulsory or secondary education, groups whose labour market entry is typically more challenging. This is also reflected in their transition rates into employment: while roughly 70% of all jobseekers registered with DYPA find work within two years of registration, this share is only 30% for GMI beneficiaries.
GMI beneficiaries often find employment in occupations different from those they originally targeted. While many seek employment in building and related trades, general clerical work, or labouring in mining, construction, manufacturing and transport, they more often find employment as cleaners, helpers, and personal service workers. This mismatch suggests that job-matching tools should enable flexible matching, not only by occupation but also by the broader set of skills workers possess, allowing to identify jobs in other occupations that require similar skill sets. At the same time, GMI beneficiaries may require reskilling to transition into new occupations or upskilling to access jobs in their targeted occupation. Job-matching tools should therefore account for situations where matches can be facilitated by participation in upskilling or reskilling measures (e.g. ALMPs). This is particularly relevant for older workers coming from physically demanding occupations who may wish to transition into other types of work.
Certain groups, such as older individuals and those with low education, are overrepresented among GMI beneficiaries registering with DYPA compared to other jobseekers. These groups typically have low digital skills and may face barriers in using digital tools. Digital tools targeted to jobseekers should be designed in a way that is user-friendly for individuals with limited digital literacy and adequate support from employment counsellors and social workers should be provided in using the tools.
When entering employment, GMI beneficiaries are more likely than other jobseekers, including other vulnerable groups, to be employed part-time or in shift work. Although this may reflect a deliberate choice to balance employment with personal circumstances, part-time contracts are on average associated with lower hourly wages. They also more often find jobs in the hospitality sector and in micro and small enterprises. Job-matching tools should therefore ensure comprehensive coverage of vacancies in these sectors and with these employers, which often rely on non-traditional recruitment channels, while also ensuring that these vacancies meet minimum quality standards. Although such jobs can serve as stepping stones into the labour market, it is essential to provide GMI beneficiaries with opportunities to progress towards good-quality employment. Career-progression tools should therefore illustrate possible pathways towards more stable and better-quality jobs.
Based on the analysis of cross-tabulations received from the CBSS, key findings for Belgium are:
MIS beneficiaries are a diverse group, with marked differences across regions and educational backgrounds. More than half of new entrants are still on the benefit one year after registering, underlining the persistence of reliance on the scheme and the challenges of rapid labour market integration. Digital job matching and career recommendation tools could be designed to provide timely information on vacancies, training opportunities and career pathways before long-term reliance on the scheme sets in, supporting local welfare offices in the development of their social activation plans.
Around 13% of beneficiaries move into regular jobs within a year compared with nearly 40% of unemployment benefit recipients, with strong differences by region and education: outcomes are better in Flanders and among those with higher qualifications, while younger and older beneficiaries fare less well. Digital job matching and career recommendation tools would allow for customised recommendations that take into account regional labour demand and individual skill profiles. For lower-qualified beneficiaries, such digital tools could embed connections to upskilling offers and show entry-level jobs with clearer pathways to better employment.
MIS beneficiaries who find work are overwhelmingly employed in service sectors, especially administrative and support services, trade/transport/hospitality, and other services (education, health, social care). Together, these account for over two‑thirds of placements. Employment in goods-producing industries is marginal, reflecting both skill requirements and labour market barriers. Digital job matching tools should ensure broad coverage of vacancies in these sectors, including micro and small enterprises that often rely on informal recruitment. At the same time, these tools could highlight opportunities in potentially underutilised sectors (e.g. construction, manufacturing).
While the group profiles are similar in terms of trends by age, education, and region, unemployment benefit recipients have higher overall (re‑)employment rates (39% vs. 13% for former MIS beneficiaries) and are more likely to enter white‑collar roles, highlighting a divide in both the quantity and quality of employment outcomes. Digital tools could help not only to match people with immediate job openings but also map career progression routes towards more stable and higher-quality employment. Integration with training catalogues and certification recognition might be particularly valuable for beneficiaries seeking to transition into non-manual jobs.
References
[13] Belot, M., X. Liu and V. Triantafyllou (2024), “Measuring the quality of a match”, Labour Economics, Vol. 89, p. 102568, https://doi.org/10.1016/j.labeco.2024.102568.
[5] CEDEFOP (2025), 2025 skills forecast Greece, https://www.cedefop.europa.eu/files/skills_forecast_-_greece_2025.pdf.
[4] European Commission (2025), Your Social Security Rights in Greece, Publications Office of the European Union.
[9] Federaal Planbureau (2025), Leefloners (i05) [Indicator: Percentage of the Belgian population receiving social integration income]., https://indicators.be/nl/i/G01_GMI/Leefloners.
[12] Lalive, R. (2007), “Unemployment Benefits, Unemployment Duration, and Post-Unemployment Jobs: A Regression Discontinuity Approach”, American Economic Review, Vol. 97/2, pp. 108-112, https://doi.org/10.1257/aer.97.2.108.
[2] Marini, A. et al. (2019), A Quantitative Evaluation of the Greek Social Solidarity Income, World Bank Group, http://documents.worldbank.org/curated/en/882751548273358885.
[1] OECD (2024), Job Creation and Local Economic Development 2024 - Country Note: Greece, https://doi.org/10.1787/83325127-en.
[7] OECD (2024), Tax and benefit policy descriptions: Belgium 2024. OECD Tax-Benefit Policies database, https://www.oecd.org/social/benefits-and-wages/tax-benefit-country-profiles.htm.
[3] OECD (2023), Education at a Glance 2023: OECD Indicators, OECD Publishing, Paris, https://doi.org/10.1787/e13bef63-en.
[10] OECD (2023), Unleashing Talent in Brussels, Belgium, OECD Reviews on Local Job Creation, OECD Publishing, Paris, https://doi.org/10.1787/7a495020-en.
[8] POD Maatschappelijke Integratie (2025), Monitoring van het leefloon en equivalent leefloon – januari 2025 [Monitoring of the basic income and equivalent basic income – January 2025], https://socialsecurity.belgium.be/sites/default/files/content/docs/nl/sociaal-beleid-vorm-geven/WGSIC/t1_monitoring_20250403_clean_-_nl.pdf.
[6] Saint-Martin, A., H. Inanc and C. Prinz (2018), “Job Quality, Health and Productivity: An evidence-based framework for analysis”, OECD Social, Employment and Migration Working Papers, No. 221, OECD Publishing, Paris, https://doi.org/10.1787/a8c84d91-en.
[11] STATBEL (n.d.), Belgium in figures: Statistics on establishment units, https://statbel.fgov.be/en/themes/datalab/statistics-establishment-units.
Characteristics of GMI beneficiaries, by employment status, December 2024
Note: The sample includes individuals with an active GMI spell in December 2024. “Never employed during GMI spell” refers to beneficiaries who did not report being employed in any of their applications within the spell. “Employed at least once during GMI spell” refers to beneficiaries who reported being employed in at least one of their applications within the spell. The table presents the distribution of characteristics separately for GMIs by employment status, i.e. in both columns, the values for the characteristics under each bold title equal 100%.
Source: OECD calculations based on data from the Greek Guaranteed Minimum Income (GMI) platform and the Greek Public Employment Services (DYPA).
Characteristics of jobseekers registered with DYPA, by jobseeker type, 1 February 2017 and 31 December 2024
Note: The sample includes employment spells starting between 1 February 2017 and 31 December 2024 of jobseekers aged 18‑64. The table presents the distribution of characteristics separately by jobseeker type, i.e. in all columns, the values for the characteristics under each bold title equal 100%.
Source: OECD calculations based on data from the GMI platform and the Greek public employment service (DYPA).
Share of employment spells in each weekly hours category, by jobseeker type, 1 Feb. 2017 to 31 Dec. 2024
Note: The sample includes employment spells starting between 1 February 2017 and 31 December 2024. Weekly hours refer to the contracted hours at the start of the employment spell. Categories are: less than 10 hours; more than 10 to less than 20 hours; more than 20 to less than 40 hours; and 40 hours or more. The unit of analysis is the employment spell.
Source: OECD calculations based on data from the Greek Guaranteed Minimum Income (GMI) platform, the Greek public employment service (DYPA) and the employment register ERGANI.
Estimated coefficients and 95% confident intervals from an OLS regression of hourly wages
Note: The unit of analysis is the employment spell. The sample includes GMI applicants hired between 1 February 2017 and 31 December 2024. The regression controls for occupation (ISCO‑08, 2‑digit level) and firm sector of activity (NACE Rev. 2, 1‑digit level). Coefficients are reported in Euros. “Small firm”, “Medium firm”, and “Large firm” coefficients are expressed relative to the omitted category “Micro firm” (1‑9 employees).
Source: OECD calculations based on data from the Greek Guaranteed Minimum Income (GMI) platform, the Greek public employment service (DYPA) and the employment register ERGANI.
Note: At national level, blue‑collar workers, white‑collar workers and civil servants account for 33%, 54% and 13% of workers, respectively. The breakdown of the type of paid employment per NACE section is shown here below. We can see that civil servants are mainly found in public administration and defence; compulsory social security (O) and education (P). Blue‑collar workers are strongly represented in manufacturing (C), administrative and support service activities (N), construction (F), transportation and storage (H) and accommodation and food service activities (I). White‑collar workers are represented in most sectors, specifically in professional, scientific and technical activities (M), information and communication (J) and financial and insurance activities (K), in which they account for more than 90% of the staff. On the other hand, they are under-represented in accommodation and food service activities (I) and in construction (F). Source: STATBEL (n.d.[11]), https://statbel.fgov.be/en/themes/datalab/statistics-establishment-units.
← 1. Households/people that were granted GMI. Applications that were rejected (e.g. because eligibility requirements were not fulfilled) are not included in the sample.
← 2. No. D13oik.3 107/68/21‑1‑2021 (Government Gazette B’238).
← 3. Defined as cases where a DYPA unemployment spell overlaps at least partly with a GMI spell.
← 4. Annex Table 3.A.2. in Annex 3.A shows additional characteristics of GMI jobseekers and compares them with those of other vulnerable jobseekers (non-GMI) and all jobseekers.
← 5. The ISCO‑08 classification provides an indication of the skill requirements of occupations. Occupations in ISCO‑08 major group 9 (elementary occupations) generally require low levels of skills; occupations in major groups 4‑8 (Clerical Support Workers; Service and Sales Workers; Skilled Agricultural, Forestry and Fishery Workers; Craft and Related Trades Workers; Plant and Machine Operators and Assemblers) tend to require medium levels of skills; and occupations in major groups 1‑3 (Managers; Professionals; Technicians and Associate Professionals) typically require high levels of skills.
← 6. Wages are a commonly used measure of match quality (Lalive, 2007[12]; Belot, Liu and Triantafyllou, 2024[13]).
← 7. This analysis is based on cross-tabulations from the Crossroads Bank for Social Security (CBSS), which provide only part of the information available in the system. The CBSS holds much more detailed microdata (see Chapter 4), which would be essential for developing a job-matching tool for MIS beneficiaries. Full access to these data would allow richer analysis and ensure that all relevant indicators can be used to support sustainable employment transitions.
← 8. Detailed information on the variables collected by each agency is available in the corresponding Output that is available on the project’s webpage (Output 2).
← 9. Following the Russian war of aggression against Ukraine, there was a sharp increase in the number of benefit recipients in early 2022. This figure has largely stabilised since the summer of 2022. Some of the increase can therefore be explained by people who fled to Ukraine and were granted temporary protection status (POD Maatschappelijke Integratie, 2025[8])).
← 10. Article 60 allows local welfare centres (CPAS/OCMW) to hire individuals receiving minimum income support, enabling them to gain work experience and, critically, to qualify for unemployment benefits.