20%
of 15–29-year-olds were not in employment, education or training (NEET) in 2024
OECD 2026Romania / labour market / a reading for 2026
The shortage has a shape.
Romania is counting empty jobs while leaving the route between a person and a job mostly unmeasured.
The OECD's diagnosis is a strange pairing: skilled occupations are short while substantial labour remains unused among less-skilled groups. That points to a matching failure before it points to a national absence of people.
Open the mismatchA vacancy becomes real only when the worker can reach it, qualify for it, afford the surrounding life and keep the job.
This diagram shows one vacancy passing through five gates: skill, care, place, formal work and migration. The final state is a job a worker can actually take.
55%
women's employment rate in 2024
OECD 202622%
estimated private-sector labour input that was undeclared in 2019
OECD / ELA100k
annual quota for non-EU work authorisations, 2022 to 2025
WorkinRomaniaRead the number sideways
A labour shortage can coexist with unused labour.
The word "shortage" sounds national. The evidence is more local and more specific. It describes different people, occupations, places and constraints at the same time.
The OECD reports that businesses have trouble filling vacancies, yet its diagnosis does not describe a shortage of humans in the abstract. Shortages appear mainly in skilled occupations. Unused labour is concentrated among less-skilled people, young people outside work and education, women, rural residents and people outside formal employment.
Those categories overlap with the route to a job. A person can be close to employable and still fail the match because the credential is missing, care is unaffordable, the workplace is too far away, the work is undeclared or recruitment is pointed at another occupation.
The claim is an interpretation of the OECD's evidence, not a new estimate of Romania's total labour supply.
7.2%
The labour force increase the OECD estimates from closing half of the gender employment gap.
That is a policy-sized number. It is also a reminder that care, work incentives and flexible jobs belong inside an industrial strategy.
Read the OECD calculation ↗The diagnostic
Open the same vacancy five ways.
Choose a gate. The vacancy stays the same. What changes is the question policy asks before it reaches for a quota.
Skill gate / the first obstruction
The credential can be one inch too high.
A vacancy can be open even when a person is close. One missing qualification can turn a capable worker into an invisible pool.
- Signal
- Only 48% of 25 to 34-year-olds without an upper-secondary qualification are employed, compared with 79% among those with one.
- What the count hides
- A certificate is not the same as competence. It can still decide whether an employer is allowed to see the worker.
Policy implication
Short, employer-linked training, recognition of prior learning and faster certification belong beside investment in the workplace.
The migration channel
A quota is a count, not a map.
Romania has opened the gate quickly. The harder question is what the gate is measuring, where the worker is needed, whether the person arrives and whether the job lasts.
Annual quota for non-EU work authorisations
201822,000authorisations
2022100,000ceiling reached
2025100,000ceiling renewed
The OECD warns that quotas count authorisations, not visas, entries or retained jobs.
1 in 3
work permits led to actual entry in 2023.
OECD comparison of administrative stages.1%
of third-country employment authorisations went to highly qualified migrants in 2023 and 2024.
OECD, citing the migration system's composition.3.9m
Romanian emigrants aged 15+ lived in OECD countries in 2020–2021.
OECD, using 2020–2021 data; the count is a population measure.31,913
approvals came from the 2026 quota in the first quarter alone.
IGI, 1 January to 31 March 2026.01authorisationemployer asks
→
02visaperson applies
→
03entryperson arrives
→
04matchskills fit
→
05retentionjob lasts
The occupation signal
The forecast points in two directions.
Cedefop's 2025 forecast sees demand for medium and high qualifications, while today's unused labour is concentrated among people with fewer formal credentials. A single shortage number cannot hold both facts.
Share of forecast job openings by qualification, 2022 to 2035
Cedefop baseline forecast
Cedefop projects no low-level qualification share in this measure. That is a forecast about job openings, not a claim that low-skilled workers disappear.
High-skilled / forecast shortage
ICT, science and engineering
Growth, replacement demand and supply-demand imbalance all push the shortage index upward for professional occupations.
Skilled manual / forecast shortage
Assemblers and craft workers
Some production roles combine employment growth with strong replacement needs and a thin supply of matching skills.
Skilled non-manual / forecast shortage
Personal service workers
Ageing and expanding services create demand, while the care system also removes time from the people who might fill it.
The Cedefop shortage index combines employment growth, replacement demand and supply-demand imbalance. It is a forward-looking signal, not a vacancy register.
The missing instrument
Build a map that can tell "nobody" from "not reachable".
Romania needs a labour map at the level where a match can fail: occupation, municipality, travel time, wage, credential, care and outcome.
Demand
Occupation code, employer, start date and vacancy duration. A vacancy must say what work is actually needed and how long it stays open.
Potential
Skills, experience, language, work history and willingness to move. NEET, inactive, returnee and informal workers should not vanish from the denominator.
Friction
Travel time, housing cost, childcare, eldercare, recognition and contract quality. This is where an apparent national shortage becomes a local problem.
Outcome
Filled, arrived, retained at 3 and 12 months, formal and paid. The map should learn from the match, not stop at the authorisation.
The country does not need one more number called "shortage". It needs a map of the missing link.
The intervention changes
Childcare becomes industrial policy.
Once the route is visible, the policy list stops looking like a collection of social add-ons. Each item removes a gate between an available person and an available job.
Childcare and long-term care
Freeing time is a labour supply measure. Less than 10% of children under three were enrolled in childcare in 2023.
Commuting and worker housing
A job outside the reachable daily radius is not a local match. Measure the route, not only the postcode.
Rapid retraining
Move training closer to the employer's real occupation, then measure certification and retention.
Recognition of experience
Returnees and migrants can arrive with skills that the paperwork does not yet make visible.
Formalisation
A formal job has a wage, social protection and a record. The answer to a hidden job cannot be a larger hidden labour pool.
Targeted migration
Immigration can fill real gaps. Quotas should follow occupation, place, qualification, entry and retention, with worker mobility and rights protected.
Industrial policy is partly the work of building the job's missing surroundings.
Sources and limits
The evidence is useful because it has edges.
The figures below do different jobs. Read the date, denominator and unit before putting them on the same line.
- 01
OECD Economic Surveys: Romania 2026
Main synthesis for NEET, women, care, informality, migration, local mismatch and policy recommendations.
Open chapter 2 ↗ - 02
Cedefop, 2025 skills forecast: Romania
Forecast of job openings, qualifications and the shortage index through 2035.
Open forecast PDF ↗ - 03
European Labour Authority, Romania factsheet
2019 estimate of undeclared labour input and its composition by employment relationship.
Open factsheet ↗ - 04
Eurostat, regional labour-market statistics
Regional employment data and the NUTS geography needed for a more granular map.
Open regional statistics ↗ - 05
Inspectorate General for Immigration, Q1 2026
36,925 work approvals for foreign workers, including 31,913 from the 2026 quota.
Open IGI release ↗ - 06
WorkinRomania, statistics and transparency
Official series for approvals, permits and annual worker quotas.
Open statistics page ↗
Method note. The 22% figure is a 2019 estimate of private-sector labour input, not a current headcount. The 20% figure is a NEET rate, not an unemployment rate. The 20%, 55% and 7.2% figures come from the OECD's 2026 report. Cedefop's figures are a forecast. IGI's numbers are approvals, not proof that a worker arrived, found a suitable job or stayed. The map above is a proposed measurement design, not a live vacancy map.