Signal Scanner · WORKFORCE, SKILLS & ORGANISATIONAL CHANGE · 8 August 2026

The Reallocation That Isn’t: AI’s Labour Shock Lands on Places, Not on Movers

OECD evidence shows labour markets absorbing the AI and trade shock through joblessness and new entrants rather than workers switching sectors, which makes place, not occupation, the binding unit of workforce risk for employers and governments to 2030.

The reskilling consensus rests on a mechanism: as AI displaces tasks, workers move, out of exposed occupations into growing ones, out of declining places into thriving ones. The OECD's 2026 Employment Outlook found that mechanism largely absent. Adjustment, it reports, “often occurs through transitions into joblessness and job opportunities for newcomers rather than because affected workers move across sectors, leaving lasting scars for displaced workers” (OECD, 07/07/2026). Regional mobility on its own cannot close the gaps and can widen them. If people do not move, the shock is absorbed by places, and by whoever enters the labour market next.

Signal Identification

A structural shift in the incidence of displacement rather than in its scale. The signal is a missing channel: reallocation of incumbent workers across sectors and regions, the assumption underneath most reskilling and place-neutral policy, is not what the data show. Incidence falls instead on the local labour market and on the entry cohort, which changes who carries the risk and who can act on it.

Time horizon: 3-8 years (regional divergence already measurable; place-based policy decisions 2026-2028; cohort scarring visible through the 2030s)
cohort effects compound2026202720302032scarring to 2040s
Plausibility band: Medium–High
LowMediumHigh
Geographic / Jurisdictional Scope: Primary: OECD economies with measurable regional dispersion, notably the United States, the United Kingdom and the EU-27. Spillover: multi-site employers whose siting decisions concentrate exposure.
PrimaryUnited StatesUnited KingdomEU-27
SpilloverOECD widerMulti-site employers
Sectors exposed:
Corporate site selectionHR and workforce planningRegional and industrial policyFurther and higher educationData-centre host communitiesPublic employment servicesGraduate recruitment

What's Changing

Two bodies reached the missing-channel finding from different directions this cycle. Alongside the OECD result, the BIS reports firms in a wait-and-see phase consistent with low hiring and low firing, a steeper vacancy-unemployment relationship in economies further advanced in AI adoption, and unemployment up 0.75 percentage points between 2023 and 2025 in high AI-preparedness countries against flat elsewhere (BIS, 28/07/2026).

Europe shows the same shape. The Commission finds EU adjustment running through reduced recruitment rather than job cuts, with aggregate unemployment still near its historical low at around 6%, and warns that a deeper shock would convert that into rising unemployment rather than into movement (European Commission, 21/05/2026). Hiring, not firing, is where the pressure is landing, which places the cost on people trying to get in.

Where they are trying to get in matters more than it used to. Exposure is spatially sorted: in the 100 most AI-exposed US counties, between 14% and 19% of workers hold occupations where AI is already used to automate rather than augment (Brookings Metro, 03/06/2026). So is the upside. Counties receiving their first data centre saw metropolitan employment rise about 4.1% and wages 5.5%, while in less populous counties the spillovers were negligible (Fortune, 14/07/2026).

Thin exposure, thick places: the two halves of the incidence problem

Share of US workers, by AI exposure County change after a first data centre Zero measured exposure 40% Exposure of at least 0.4 under 10% Occupational exposure is thin and widely spread. It does not, on its own, sort workers by place. Metropolitan county, employment +4.1% Metropolitan county, wages +5.5% Less populous county negligible The same investment, two different labour markets. Indigo: exposure shares. Orange and red: local effects of an identical siting decision.

Source basis: Federal Reserve Bank of New York (14/05/2026); Fortune (14/07/2026).

Disruption Pathway

Stage one is visible now: divergence between local labour markets is measurable, mobility is not responding, and hiring is carrying the adjustment. Stage two runs from 2027 to 2029 as cohort effects compound: where hiring is thin, the people who never got in accumulate the scarring the OECD describes, and participation falls before employment does. Stage three settles the pattern. Either employer siting and place-based policy internalise the incidence, or divergence becomes self-reinforcing, with the local labour market rather than the worker carrying the adjustment indefinitely, which is what persistent gaps in employment, unemployment and disposable income across regions already look like (OECD, 07/07/2026).

Stresses concentrate at three points. The entry cohort is the first, because a labour market that adjusts by not hiring loads the whole cost onto people with no incumbency to protect. Single-employer communities are the second: an identical siting decision that lifts a metropolitan county's employment 4.1% does almost nothing in a thinner one (Fortune, 14/07/2026). Public finances in lagging regions are the third, since the tax base moves with the place rather than with the worker. Two adaptations follow. Employers begin treating site selection as a workforce-risk decision rather than a cost decision, and governments shift from place-neutral skills funding toward the integrated place-based instruments the OECD now calls for (OECD, 07/07/2026).

Why This Matters Now

For boards and CHROs the exposure sits in an assumption rather than a budget line. Reskilling programmes are priced on the belief that retrained people will move into adjacent work; the OECD evidence says adjustment mostly runs through exit and through new entrants instead. Three decisions change accordingly. Site selection becomes the highest-leverage workforce decision a multi-site employer makes, because it fixes which local labour market absorbs the consequences. Graduate and junior intake becomes the real adjustment margin, so cutting it is a decade-long choice rather than a cyclical saving. And workforce planning should run at local labour market level rather than by occupation group, because that is where the risk sits. For governments the implication is blunter: place-neutral skills spending is aimed at a mechanism the data do not show.

Decision-action posture for this signal: Prepare — the divergence is already measurable and the siting and intake decisions that determine incidence are being taken now, but the cohort effects that would force a Decide posture take another two to three years to surface.

Counter-Argument

The strongest objection is that neither the AI attribution nor the geography holds. The New York Fed finds less than 10 percent of US workers and vacancies in occupations with AI exposure of at least 0.4, 40 percent in jobs with zero measured exposure, and no clear trend in the share of junior postings (Federal Reserve Bank of New York, 14/05/2026). Kolko presses the geography, noting that the occupations most exposed to AI, office administrative and clerical work, are not especially concentrated in particular local labour markets (PIIE, 10/03/2026).

Both objections land on the attribution rather than the channel. The OECD finding concerns trade and technology shocks generally, and holds whichever shock is doing the work. Diffuse exposure also cuts the other way: if at-risk occupations sit everywhere and workers still do not move, the absorbing unit remains the place. Kolko's reading requires local opportunities to be genuinely available to the displaced, which is what the hiring data do not show.

Implications

This is a durable change in where workforce risk is priced rather than a transient wobble in hiring. The inflection window is 2026 to 2028, while siting decisions for the current investment cycle and the next round of regional policy are still open. Positioned to gain: employers treating local labour market depth as a siting input, regions with the surrounding density to convert investment, and education systems holding entry-level pipelines open through a low-hiring period. Positioned to lose: single-employer communities, mid-career workers in thin markets, and skills programmes funded on the assumption that trained people will move.

Early Indicators to Monitor

Disconfirming Signals

Strategic Questions

Keywords

labour market adjustment; worker reallocation; regional divergence; place-based policy; entry-level hiring; cohort scarring; AI exposure; internal migration; matching efficiency; site selection; workforce planning; OECD Employment Outlook

Bibliography

Source tiers: Tier 1, governments, regulators and intergovernmental bodies. Tier 2, think-tanks, academic institutes, major consultancies and quality data providers. Tier 3, quality journalism and specialist trade press. Tier 4, vendor, company and practitioner sources, used only as directional corroboration.

This Signal Scan is one component of a Shaping Tomorrow intelligence cycle on Workforce, Skills & Organisational Change. Every monitored topic also includes monthly Change Trackers that assess what is materially shifting in momentum, and a quarterly Decision Intelligence briefing that converts the cycle into board-level strategic implications. Cycles may also close with Action Triggers, an end-of-cycle synthesis that reads across the three formats and surfaces patterns no single deliverable carries on its own. Continuous intelligence on this topic is available as a Single-Topic or Multi-Topic programme.

Discuss a Tailored Programme

Prepared by Shaping Tomorrow: 8 August 2026