Nordic sub-cluster (4) · trade-decoupling
Worker-protection plus trade-decoupling exposure (no UK adjacency); squeeze arrives via decoupling, not via cross-border capital flight.
DK
FI
NO
SE
Eight storylines for the next decade, with the odds shaped by each country’s economic regime.
An optimism-pessimism spectrum from full reinstatement to a parallel-cascade collapse, three economic regimes that change which future is most likely, and the country-level trajectories underneath.
“Three economic regimes, three different sets of odds.”
The same seven “routine” futures exist for every country, but the probabilities attached to them change by economic regime, because what’s plausible in a growing economy is often not plausible when output isn’t expanding. We separate growth, secular stagnation, and post-growth. In post-growth, the standard tech-led recovery path drops from about 20% probability to about 13%, while Climate Adaptation Boom rises from about 20% to about 30%. Same menu of futures; different odds.
“The optimism path narrows the further you push the regime.”
For Austria, Luxembourg, and Turkey, only one of the routine futures consistently lands them in the safe corridor, and it is not the one most policy speeches assume. A tech-led reinstatement path is effectively closed off. Their best routine outcome runs through redirecting workers into climate-adaptation demand: healthcare, trades, and the green economy. Optimism remains possible. It just runs through climate, not conventional tech.
“S8 sits outside the spectrum on purpose.”
Seven scenarios sit on an optimism↔pessimism spectrum, and their probabilities sum to 100% within each regime. The eighth, Polycrisis Drag, is intentionally different: it assumes overlapping shocks (AI displacement plus decoupling pressure, demographic decline, climate stress, and fiscal strain), not a single dominant driver. Because there is no clean historical analogue for that combination, we treat it as a conditional “parallel risk,” not simply the worst point on the routine spectrum.
1. The Spectrum Seven routine futures along an optimism-pessimism axis, plus one parallel-cascade scenario that sits outside it. The framing is ours, not a definitive taxonomy.
How the next decade plays out for European labour markets isn’t one question, it’s at least eight. Each scenario below is a self-contained what-if: a coherent story of how AI, demographics, and institutions interact over ten years. The seven routine variants are mutually exclusive and span an optimism-pessimism axis; the eighth is a parallel-cascade pattern that can be triggered in addition to any routine variant. Other plausible futures would land between these, or outside this spectrum entirely.
S8 Polycrisis Drag · outside the spectrum
The eighth scenario sits off the optimism-pessimism line on purpose.
S8 Polycrisis Drag is a “pick-your-poison” parallel-cascade pattern, not a point on the routine spectrum. War and defence spending, budget stress, climate adaptation costs, global decoupling, political turmoil, social division, and Ukraine reconstruction combine to overwhelm institutional capacity at the same time. It is carried as a conditional that can be triggered on top of any routine variant, not folded into the spectrum.
This combination has no analogue in the 580 years of historical disruptions reviewed earlier in this analysis project (link to part 3). There is no clean historical case base to ground these joint dynamics. That is why it is treated as a conditional “parallel risk,” rather than as the worst point on the routine spectrum.
2. Three Weather Patterns Economic regimes that change which future is most likely.
Two countries can face the same AI “story,” but if one economy is growing and the other is stuck or shrinking, the same shock will land differently. That is why the project groups countries into three “weather patterns” (growth, stagnation, post-growth) to adjust which futures are more likely.
Note: “tech-led recovery” here refers to S1 New Jobs Replace Old + S3 Jobs Transform, both tech-led optimistic scenarios. Their combined probability under growth is around 20%; under post-growth, around 13%.
S1 New Jobs Replace Old
S2 Climate Adaptation Boom
S3 Jobs Transform
S4 Muddle Through
S5 Wage Cliff
S6 New Jobs Fail to Appear
S7 Bandwidth Fracture
3. Per-Country Trajectories Seven routine scenarios, baseline (2026) interpolated to scenario corridor (2035). Use selectors to pick country and emphasise one scenario.
Each line on the chart shows how one country moves from its 2026 starting “corridor” (how safe or stressed its labour market is) to where it ends up in 2035 under a specific scenario, with upward bends meaning improvement and downward bends meaning getting worse.
Scenario corridor trajectory, 2026 baseline → 2035 scenario corridor (C1 at top is the safer outcome; C3 at bottom is the worse outcome)
4. Scenario × Regime Probability Table Mid-points with 80 % CI. IPCC AR6 likelihood-scale anchors; seven routine variants sum to 1.0 per regime; S8 carried as conditional, orthogonal.
These probabilities are our best estimate of how likely each scenario is, given the country’s economic regime. The “80% confidence interval (CI)” is the range in which the probability falls in 8 out of 10 runs of our calculations.
The reskilling-capacity gap
Europe needs to retrain about 7.55 million people by 2035 due to AI effects. At first glance, the basic annual training throughput of ~3.34 M makes that seem doable. But after subtracting statistical retraining churn (~2.89 M), there is only room to retrain about 450,000 extra people per year. At that pace, the 7.55 M cohort would take roughly 15 years to clear. That is far slower than how quickly AI could change jobs, and it does not even count the training already needed for normal job turnover.
On top of that, European vocational training and university systems currently take 5–9 years to respond to a major skill shift, that is, to retool curricula, accredit new programmes, and run cohorts through to graduation. AI disrupts the affected jobs in 1–3 years, creating a several-year gap in which people are displaced before training can catch up.
This anchors the pessimism side and serves as the quantitative spine of the 15-country Class III diagnosis. Even allocating disproportionately to Class III would absorb only a marginal share of the deficit without channel expansion.
The optimism path narrows to climate Zone-C
For three countries (Austria, Luxembourg, Turkey) the only realistic “good outcome” left in the model is the “Climate Adaptation Boom” scenario where lots of new work comes from climate adaptation. The usual “tech boom brings jobs back” path does not work for them. All six other routine variants produce a corridor 2 or 3 outcome. Anchored on Cedefop 2025 country-level employment projections plus the EU Net-Zero Industry Act €100 B clean-manufacturing envelope.
AT
LU
TR
5. Capability-Floor Breach
12 countries breach the institutional adaptive-capacity floor at the 2-digit ESCO level. This result should be read as a conservative estimate at this level of aggregation. Finer-grained data would likely surface 1–2 more.
A capability-floor breach means the country’s institutions cannot routinely absorb the kind of labour-market shocks the scenarios describe.
How the 12-country count was reached
Denmark is the marginal entrant. As the job classification sharpened, the breach list grew: 3 countries at the coarsest resolution → 11 at 1-digit ISCO → 12 at 2-digit. Finer aggregation tends to surface additional breach signals, which is why the 12-country result is read as a conservative estimate.
Ranked by probability, 7 are marked high, 4 medium, and 1 low for chain-reaction (cascade) risk. The 3 Class IV countries are labeled currently failing rather than “at risk.”
6. The Squeeze Cluster Eight worker-protection economies, two distinct mechanisms. Luxembourg is explicitly NOT squeeze-flagged.
The squeeze flag is a capital-flight signal, not a labour-displacement signal. The risk is that AI investment leaves rather than that workers are displaced at home. Quantification rests on per-country counts of approximately 40 high-risk EU AI Act Annex III deployer occupations and approximately 29 Product Liability Directive post-market duty occupations. Two distinct mechanisms warrant two distinct mitigations.
Continental sub-cluster (4) · UK adjacency
Worker-protection plus UK adjacency plus Mode 1 capital-flow vulnerability: the canonical jurisdictional-buffering profile.
BE
DE
FR
NL
LU correction. An earlier draft had named only five squeeze countries (BE / DE / FR / LU / NL). The corrected reading is the eight-country pattern above, decomposed into the two sub-clusters. Luxembourg’s Class I status rests on its standalone Continental Corporatist profile plus its S2-dependent optimism path, not on any squeeze designation.
Einfache Sprache · coming soon
This page is being prepared in Einfache Sprache. The standard version is available; click Standard to read it.
Diese Seite wird in Einfacher Sprache vorbereitet. Die Standard-Version ist verfügbar.