South West · Baseline population 94,782
Official ONS figures for the region Bath sits in, published June to August 2026. These are measured statistics, separate from the structural exposure score above — they are reported here as context and are not an input to it. Region-level data is the finest geography ONS publishes monthly; it is not a Bath-only figure.
UK average for the same period: unemployment 4.7%, employment 75.3%, inactivity 20.9%. South West ranks 10 of 12 UK regions and countries by unemployment rate.
97,965 people in South West were claiming unemployment-related benefits in September 2026, 3.3% of economically active residents, up 3,815 on September 2025. Among 16–24 year olds the count was 20,285, up 2,240 on the year. Not seasonally adjusted; this is an administrative count, not a survey estimate.
2,506,727 people were on a payroll in South West in September 2026, down 3,907 (0.16%) on the year, against 0.22% down across the UK as a whole. This is an administrative count of actual payroll records, not a survey estimate. The most recent month is a provisional flash estimate and is routinely revised.
The two UK-wide measures of employees disagree. On ONS's own like-for-like
comparison, HMRC payroll (RTI) records show payrolled employees
down 92,000 over the
year and down
37,000 over the quarter in
June to August 2026. Over the same period the Labour Force Survey shows employees
up 314,000 to
29.9 million. The provisional September
flash estimate puts payrolled employees at
30.3 million, down
66,000 on the year.
ONS attributes part of the LFS employee increase to improvements in the survey since January 2024, which have obtained more responses from employed people. Levels have therefore moved closer to RTI, but ONS cautions that LFS measures of change 'may be less aligned with RTI because they capture both real labour market shifts and changes arising from the survey improvements'. The divergence is concentrated in the 25 to 34 year old age group, which ONS is still investigating.
ONS's view remains that RTI 'currently provides the most reliable measure of employees'. As of the October 2026 release this statement appears in the LFS quality update rather than in the labour market overview bulletin itself.
We surface both rather than quietly picking one.
The August 2026 flash estimate reported a fall of 145,000 on the year; the confirmed figure is 111,000. Flash estimates are provisional and are routinely revised when more data arrive, so the September flash of -66,000 should be read as the least settled number here.
ONS research linking RTI
to the LFS, published 28 September 2026:
Linked HMRC PAYE data showed a stronger recovery in employment after the pandemic and a smaller, less prolonged rise in economic inactivity than the published LFS estimates alone suggested. It also suggested the LFS recovery during 2024 and 2025 may have reduced non-response bias, since using the linked payroll data made only a small difference to the level of employment and inactivity estimates in 2025. ONS describes the work as exploratory and is continuing the analysis.
The survey side is UK-wide; ONS publishes no regional split of this comparison. The payroll side is parsed from the ONS reference table (rtisaoct2026.xlsx). Source: ONS Labour market overview, UK: October 2026, released 20 October 2026. Next release 17 November 2026.
Sources. ONS Labour market overview, UK (time series API); ONS Labour market in the regions of the UK (Table 2, annual changes); ONS Annual Population Survey, via NOMIS (variables, percentages); Claimant count by sex and age (Universal Credit + JSA), ONS/DWP via NOMIS; Earnings and employment from PAYE Real Time Information, seasonally adjusted (Table 7, payrolled employees by NUTS1 region). Retrieved 2026-10-22. LFS and APS outputs moved from “official statistics in development” to “official statistics” on 11 August 2026 and ONS is not seeking reaccreditation; survey-based detail should be read with that in mind. Contains public sector information licensed under the Open Government Licence v3.0.
Reported UK redundancy events from the CopeCheck Layoffs tracker. These are UK-wide, not Bath events — the tracker records the country a layoff was reported in, not the town or city, so nothing here is attributed to Bath.
| Reported | Company | Roles | Sector | Source |
|---|---|---|---|---|
| 2026-08 | University of Essex | 400 | Education | www.bbc.co.uk |
| 2026-08 | Nothing | 100 | Technology | Google Alerts RSS |
| 2026-08 | Travelport | 57 | Technology | Google Alerts RSS |
| 2026-08 | TSB | 130 | Banking | bmmagazine.co.uk |
| 2026-08 | Diageo | 38 | Food & Beverage | drinks-intel.com |
| 2026-08 | WPP | 1,267 | Advertising | gbnews.com |
| 2026-08 | Supermassive Games | 75 | Technology | Google Alerts RSS |
| 2026-08 | RNLI | 75 | Charity/Non-profit | Google Alerts RSS |
74 UK events recorded for 2026, of which 11 are attributed by their source at least partly to AI or automation. Counts are individual reported events, not a national total. The upstream tracker withholds aggregates because source claims may be syndicated, aggregate, planned, aliased or overlapping. These are UK-wide and are not attributed to this city. Compiled from layoffs.fyi, TrueUp, Crunchbase, TechCrunch, and news reports. Snapshot taken 2026-10-22.
Model-generated scenario commentary, not reported fact. Specific claims about future employer or council behaviour are hypothetical; use the sector table, not the prose, as the reproducible input to the score.
Bath has assembled an economic profile that reads like a protection spell against automation: education at 13.4%, health at 13.2%, professional services at 8.4%, and a tourism economy that depends on exactly the kind of human-crafted, place-specific experience that AI can't replicate. You can't automate the Roman Baths or the Royal Crescent, and you can't replace a university lecturer with a chatbot (not yet, and not without the students noticing). The accommodation and food sector at 8.4% reflects the tourism economy, which is itself a form of resilience — people visit Bath for the physical experience of being there. Admin at 8.2% and retail at 8.6% are present but below average, making Bath one of the best-insulated cities in the dataset. The risk isn't displacement — it's affordability: Bath's desirability makes it expensive, which limits who can afford to live and work there.
Bath's economic moat is essentially 'we're too pretty and too educated to automate,' and annoyingly, that's a perfectly valid strategy. Education at 13.4%, health at 13.2%, professional services at 8.4% — over a third of the workforce does things that require physical presence in beautiful buildings, human judgement about complex matters, or both. The Roman Baths have been a tourist trap for two thousand years and show no sign of disruption. The university employs a small army of academics who are, by definition, in the business of human thought transmission. The Thermae Bath Spa charges £40 to sit in warm water and look at the abbey, and people queue for it, which tells you everything about Bath's economic resilience. The risk here isn't AI — it's that Bath becomes so expensive that only retirees and tourists can afford it, at which point the university struggles to recruit staff who can't afford the housing, and the service economy runs on workers who commute from Midsomer Norton. The Pump Room serves tea at prices that would make a Londoner blink. The Assembly Rooms host corporate events for companies based elsewhere. Bath is the economic equivalent of a trust fund — it doesn't need to work, it just needs to exist attractively. Which it does. Infuriatingly.
Bath's challenge isn't automation resilience — it's maintaining the affordability and workforce diversity that makes a functioning city rather than a heritage theme park. Protect the university's ability to recruit by addressing housing costs for academic and support staff. Develop the professional services cluster toward AI-adjacent consultancy — ethics, regulation, compliance, governance — leveraging the university's humanities and social sciences strengths. The cultural economy should be maintained and expanded, but with attention to year-round employment rather than seasonal tourism peaks. Bath's proximity to Bristol creates opportunities for knowledge-economy spillover that should be actively cultivated rather than left to happen.
Bath will host a 'Future of Work' conference at the Assembly Rooms, serve prosecco, charge £200 a ticket, and attract people who are already insulated from the problem. The council will produce a strategy that addresses affordability by building 500 'affordable' homes at £350,000 each. The university will launch an AI ethics programme that's intellectually excellent and commercially marginal. The Thermae Spa will raise its prices. The Roman Baths will raise their prices. The independent shops on Milsom Street will close one by one, replaced by chains that can afford the rents. Someone will propose a 'Bath Innovation Quarter' on the Lower Bristol Road that attracts a tech company, a yoga studio, and a restaurant. The actual challenge — keeping Bath a functioning city rather than an open-air museum with a university attached — will be addressed by planning committees that approve developments nobody who works in Bath can afford, and celebrate 'growth' that consists mainly of wealth transfer from tourists to landlords. Bath doesn't need an AI strategy. It needs a housing strategy, a wage strategy, and the humility to acknowledge that being beautiful isn't the same as being equitable.
Historical employment share by broad sector, with CopeCheck's heuristic exposure weight and contribution to the relative score. Higher weights contribute more; they are not probabilities.
| Sector | Employment % | Exposure Weight | Exposure Tier | Contribution |
|---|---|---|---|---|
| Education | 13.4% | 0.15 | low | 2.01 |
| Human Health & Social Work | 13.2% | 0.18 | low | 2.38 |
| Retail | 8.6% | 0.8 | high | 6.88 |
| Accommodation & Food Services | 8.4% | 0.48 | medium | 4.03 |
| Professional, Scientific & Technical | 8.4% | 0.3 | low | 2.52 |
| Administrative & Support Services | 8.2% | 0.85 | high | 6.97 |
| Public Administration & Defence | 6.8% | 0.22 | low | 1.5 |
| Financial & Insurance Services | 6.4% | 0.75 | high | 4.8 |
| Information & Communication | 5.8% | 0.5 | medium | 2.9 |
| Arts, Entertainment & Recreation | 4.8% | 0.2 | low | 0.96 |
| Construction | 4.2% | 0.28 | low | 1.18 |
| Manufacturing | 3.8% | 0.82 | high | 3.12 |
| Transport & Storage | 3.2% | 0.78 | high | 2.5 |
| Wholesale | 2.8% | 0.55 | medium | 1.54 |
| Real Estate | 1.6% | 0.4 | medium | 0.64 |
| Agriculture, Forestry & Fishing | 0.4% | 0.25 | low | 0.1 |
The vulnerability score is a weighted average of Bath's sector employment shares. Each broad sector carries a CopeCheck heuristic exposure weight (0.0–1.0). Frey & Osborne's study estimated occupation-level susceptibility; it did not publish these sector weights. The bundled city files also lack record-level source dates and ledgers, so this is a historical structural comparison rather than an official current statistic. The weighted average is normalised to a 0–100 scale. A score of 55.0 means Bath's workforce is moderately concentrated in higher-weighted sectors compared with other United Kingdom cities in this dataset; it does not state a 55.0% probability of displacement.