The number, and what it counts
Cedefop's skills forecast expects the loss of 32% of the jobs held by people with low qualifications by 2035 (Cedefop research paper 5620, April 2026).
The ten occupational groups with the worst outlook — clerical support, protective services, skilled agricultural and forestry work, handicraft and printing trades, food processing and garment trades, stationary plant operators, drivers and mobile plant operators — hold almost 44 million jobs in the EU between them. More than 10 million of those are held by low-qualified workers; more than 30 million are jobs with low skill requirements.
Note carefully what is being counted. This is a projection about jobs held by people who did not complete upper secondary education. It is not a forecast of unemployment, and it is not a statement about any one country.
The comparison that matters more
Same publication, different measurement. The share of workers in jobs with low skill requirements, broken down by education level (Cedefop European skills and jobs survey, 2021):
- 64% of those with lower secondary education or less (ISCED 0–2)
- 59% with upper secondary or post-secondary non-tertiary (ISCED 3–4)
- 36% with a higher education degree (ISCED 5–8)
Twenty-eight percentage points between the bottom row and the top one. Cedefop reports the association as statistically significant (χ², p < 0.001).
This is not a promise about anyone's salary. It is the measured difference in where people end up working.
What the publication does not say
There is no country breakdown of the 2035 projection. The −32% is an EU aggregate. Anyone quoting it as "Germany is about to lose a third of its jobs" has added something the source does not contain.
The publication also carries one finding that is easy to get backwards, and that Cedefop's own news page summarised in the opposite direction. Among adults holding a vocational qualification, those in engineering, manufacturing and construction, healthcare and services are more likely to be working in a low-skilled job. Fields with lower proportions are ICT, business administration, arts and humanities, social sciences, natural sciences, mathematics and statistics, and generic programmes.
Two cautions before anyone uses that. It describes holders of vocational qualifications, not university graduates — the study did not measure the second group this way. And it describes where people currently work, not what a degree in that field is worth.
The one country table there is
Cedefop measured what share of jobs in each country carry low skill requirements, defined across literacy, numeracy, interpersonal skills and problem-solving (ESJS microdata, 2021):
| Country | Jobs with low skill requirements |
|---|---|
| Lithuania | 60.1% |
| Latvia | 59.2% |
| Estonia | 54.3% |
| Germany | 52.1% |
| Netherlands | 48.6% |
| Austria | 45.7% |
Across the EU the range runs from 40.3% in Ireland to 60.9% in Czechia.
Read this for what it is: the structure of an entire labour market, every worker included. It is not a graduate outcome, and the publication never presents it as one. A country can hold a large share of low-requirement jobs and a healthy market for degree holders at the same time — the two numbers describe different parts of the same economy. The table is here because it is measured and public, not because it ranks destinations.
What follows for a decision
Not a country. If you were hoping this data would tell you that Austria is a better bet than Lithuania, it does not, and reading it that way is the error the publication guards against in its own text.
What it does support is narrower and more useful. The employment structure that is shrinking is the one entered without a completed qualification, and the gap between finishing a degree and not finishing one is 28 percentage points wide in where people work. That is an argument about finishing, not about where.
For the "where" question the relevant data is occupation-level rather than country-level: vacancy rates and pay for your specific occupational group, per country, which Eurostat publishes and which we show on the route result next to each field. Those numbers differ between our six destinations. This one does not.
One more thing worth keeping. Adults with low qualifications take part in training at less than half the rate of the population as a whole, and only 8% of workers doing only basic digital tasks strongly felt a need to improve their digital skills. The gap compounds itself — which is a reason to treat the first qualification as the thing that is hard to add later.