by Claude Opus 5.5

As of the latest data available by October 2026, what do UK labour-market indicators (employment, unemployment, inactivity, vacancies, payrolled employees, wage growth) suggest about the overall jobs picture—how has it changed since January 2026, and what are the key caveats (including Labour Force Survey reliability)?

The UK labour market in autumn 2026 is weak but no longer deteriorating quickly. Unemployment is 4.9%, below its winter peak of 5.2%. Vacancies are at their lowest outside the pandemic since 2014, payrolls are about 145,000 lower than a year ago, and pay growth has cooled by a full percentage point since January. This is a “low-hire, low-fire” market: few people are being laid off, but few are being hired, so newcomers and the long-term unemployed bear the cost. Nothing in the aggregate figures needs AI to explain it.

The headline numbers

The latest ONS release (15 September 2026) covers May–July 2026 for survey-based measures and August 2026 for payrolls. The list below compares it with the release published in January 2026.

  • Employment rate (16–64). January 2026 release: 75.1% (Sep–Nov 2025). September 2026 release: 75.1% (May–Jul 2026).

  • Unemployment rate (16+). January 2026 release: 5.1%. September 2026 release: 4.9%.

  • Unemployment level. January 2026 release: 1.84m. September 2026 release: 1.78m.

  • Inactivity rate (16–64). January 2026 release: 20.8%. September 2026 release: 20.9%.

  • Payrolled employees (flash). January 2026 release: 30.2m (Dec 2025). September 2026 release: 30.2m (Aug 2026), −145,000 on the year.

  • Vacancies. January 2026 release: 734,000 (Oct–Dec 2025). September 2026 release: 702,000 (Jun–Aug 2026).

  • Regular / total pay growth. January 2026 release: 4.5% / 4.7%. September 2026 release: 3.5% / 3.9%.

  • Real regular pay (CPIH). January 2026 release: 0.6%. September 2026 release: 0.6%.

  • Claimant Count. January 2026 release: 1.677m (Dec 2025). September 2026 release: 1.692m (Aug 2026).

  • Youth (16–24) unemployment. January 2026 release: 15.9%. September 2026 release: 16.2% (Apr–Jun 2026).

Reading the indicators together

Low hiring, low firing. The redundancy rate was 3.6 per 1,000 employees in April–June 2026, “largely unchanged over the year”. The redundancy level fell to about 106,000 in that quarter, from 145,000 in September–November 2025. Yet the unemployment level is still 83,000 higher than a year earlier, and the rise was “driven by a rise in those unemployed for more than six months” (as reported). That combination tells you where the pain is. People in jobs are mostly keeping them, but people looking for work are taking longer to find it. In March the OBR described the same pattern: “Labour market weakness still appears to be driven by entrants into the labour force struggling to find work.”

Demand is very soft. At 702,000, vacancies are at a level last seen outside the Covid period in August–October 2014. There are about 2.5 unemployed people per vacancy. CBI/Pertemps says that ratio has been stable since September 2025, so the market is slack but not getting rapidly slacker. The Bank of England’s Decision Maker Panel (September 2026) found firms’ realised employment growth over the past year was −0.2%, and they expect only +0.2% for the coming year. The Bank’s Agents expect headcount to stay “broadly flat” over the next 12 months.

There are signs of a floor. In September, most MPC members saw “signs of stabilisation in the margin of slack that had been opening up”. The KPMG/REC Report on Jobs for August recorded the first rise in permanent placements since 2022, though only a marginal one, and a fifth month of growth in temporary billings. Unemployment has now been 4.9% for four consecutive rolling quarters.

Pay is cooling, and inflation is turning against it. Regular pay growth fell from 4.5% to 3.5%. In the private sector it is only 2.9%, against 6.3% in the public sector. Real regular pay is still growing (0.6% deflated by CPIH), but CPI rose to 3.1% in August. In September the MPC said it expects CPI to reach “around 3¾% in 2026 Q4” and “slightly above 4% in 2027 Q1”. If nominal pay growth stays around 3.5%, real pay will probably turn negative over the winter. That is an inference from the Bank’s projection, not a measured outcome.

The young and London are weakest. Unemployment among 16–24-year-olds was 16.2% in April–June 2026, up 97,000 on the year. London has the highest regional unemployment rate at 6.8%, against 2.4% in Northern Ireland. There was one surprise in the payroll data: in the year to August 2026, payrolled employees under 25 fell by only 15,000, while those aged 65 and over rose by 62,000. By the ONS age bands, that implies a fall of roughly 190,000 among 25–64-year-olds (computed from the published figures, which are provisional).

What has changed since January

In January the story was “unemployment rising, vacancies flat, pay still strong”. Since then the picture has shifted in three ways:

  • Unemployment peaked earlier and lower than forecast. It reached 5.2% in November–January, then eased to 4.9% from February–April onwards. In March the OBR forecast a 2026 peak of 5⅓%. In July the Bank still projected 5.0% in Q3 and 5.1% in Q4. So far, outturns have come in below both.

  • Hiring demand kept weakening, but more slowly. Vacancies fell by about 32,000 over the period, and payrolls are still falling on the year. This is a slow grind, not a slump.

  • The macro backdrop changed sharply. In February the MPC held Bank Rate at 3.75% by 5–4, with four votes for a cut, and said rates were “likely to be reduced further”. In both July and September it held by 6–3, with three votes for a rise. An energy price shock linked to conflict in the Middle East, and to Ukraine and Russia, pushed inflation back up. Since July there is also a new Prime Minister (Andy Burnham) and Chancellor (John Healey), and the Autumn Budget is expected on 28 October.

The caveats that matter

LFS reliability

The unemployment, employment and inactivity rates all come from the Labour Force Survey. Response rates collapsed in 2023, and the LFS is still recovering. ONS says responses are now “close to their pre-coronavirus pandemic level”, with an achieved sample of 81,193 people in April–June 2026, up from 80,078 in January–March. But ONS still warns that “some volatility remains”, especially in estimates for mid-2023 and 2024. Because the survey has been enhanced since January 2024, movements since then “reflect both underlying developments in the economy and improved survey quality”.

Two further problems arose this year. First, an operational issue in May 2026 left telephone interviewing under-resourced. The MPC minutes note that this led to “increased use of imputation” for May and June. ONS judges the effect on headline estimates “minimal”. Second, a full reweighting of the LFS to updated population estimates is under way, and the first reweighted data are due towards the end of 2026. Reweighting can revise historical levels, particularly where migration estimates have changed a lot.

In practice, a 0.1–0.2 percentage point move in the unemployment rate is within the noise. ONS itself advises users to “focus on long term movements”. The LFS outputs are moving from “official statistics in development” to “official statistics”, but ONS “will not be seeking reaccreditation” for them.

The transformed LFS (TLFS) has slipped again

The original plan was to switch headline statistics to the online-first TLFS in November 2026. The April 2026 update pushed this into 2027. The August update says July 2027 “will be an important decision point” and that a November 2027 transition is now the most likely date. ONS’s reason is that “only one calendar quarter of data reflecting the latest set of design changes has been collected”. Comparisons so far show “high levels of agreement” between the two surveys on aggregate economic activity. The largest differences are in how people are classified as inactive.

So the UK will rely on the old LFS for at least another year. When the switch comes, users should expect a discontinuity, especially in inactivity.

RTI versus LFS

The payroll figures come from HMRC’s PAYE Real Time Information (RTI). These are administrative records, not a survey, so they are large, timely and unaffected by response rates. But they have three limitations:

  • They count employees on payrolls only, so they miss the self-employed.

  • The latest month is a flash estimate that is “likely to be revised”.

  • They say nothing about unemployment or inactivity.

The two sources currently disagree about direction:

  • LFS employees. Change on the year: +111,000.

  • RTI payrolled employees (August flash). Change on the year: −145,000.

  • RTI payrolled employees (July). Change on the year: −101,000.

  • Workforce Jobs (June). Change on the year: +60,000.

People and payroll jobs are different concepts. Population weighting in the LFS may also be out of date, and RTI revisions can be large. A sensible rule is to trust RTI for the direction of employee numbers, because it is less affected by survey problems. For unemployment and inactivity, use the LFS, because nothing else measures them directly, but read it alongside the Claimant Count (1.692m in August, up on the year). Vacancies come from a separate ONS business survey and are a three-month average, so they lag turning points.

Bottom line

The aggregates describe a cyclical squeeze. Demand is weak after two years of restrictive interest rates, and employers face higher labour costs (employer National Insurance and the National Living Wage) and an uncertain outlook. They show no AI shock. The Bank’s Agents still report “limited evidence of broad AI-driven reductions in employment”. Any AI effect so far is visible in particular occupations and in entry-level hiring (see 2.3), not in the headline rates.

What to watch: the 20 October release, covering June–August; whether RTI revisions narrow the gap with the LFS; long-term unemployment and youth figures; the reweighted LFS due around year-end; and whether the inflation rebound reaches pay settlements. The Agents put the average 2026 settlement at 3.6%.

Sources

From AI and Jobs: UK, October 2026