The fewest vacancies outside the pandemic since 2014
UK vacancies fell by 21,000 (2.9%) to 691,000 in July to September 2026, compared with April to June. On the year they fell by 37,000 (5.1%). The Office for National Statistics published the estimates on 20 October 2026.
Outside the coronavirus pandemic, the last time there were 691,000 vacancies or fewer was July to September 2014, with 686,000. In July to September 2020 there were 499,000, and in July to September 2022, 1,235,000.
The quarter’s fall is the largest since January to March 2026, which was also a fall of 21,000.
| July to September | Vacancies |
|---|---|
| 2014 | 686,000 |
| 2015 | 742,000 |
| 2016 | 756,000 |
| 2017 | 806,000 |
| 2018 | 853,000 |
| 2019 | 827,000 |
| 2020 | 499,000 |
| 2021 | 1,132,000 |
| 2022 | 1,235,000 |
| 2023 | 984,000 |
| 2024 | 840,000 |
| 2025 | 728,000 |
| 2026 | 691,000 |
Fewer at small firms, more at large ones
At businesses with 1 to 9 employees, vacancies fell by 21,000 (21.1%) on the quarter, to 77,000. Over the year they fell by 38,000 (32.8%), about a third. Outside the pandemic, that is their lowest level since June to August 2013.
The largest employers went the other way: over the year, vacancies rose by 2.6% at businesses with 250 to 2,499 employees and by 3.9% at those with 2,500 or more. They fell by 7.1% at businesses with 10 to 49 employees and by 5.0% at those with 50 to 249.
The ONS reports that feedback to its Vacancy Survey continues to suggest smaller firms may not be recruiting because of increases in labour costs. That is what employers tell the survey, not something the figures measure.
| Employees | On the quarter | On the year |
|---|---|---|
| 1 to 9 | -21.1% | -32.8% |
| 10 to 49 | -0.7% | -7.1% |
| 50 to 249 | -0.6% | -5.0% |
| 250 to 2,499 | +1.2% | +2.6% |
| 2,500 or more | -0.3% | +3.9% |
| All vacancies | -2.9% | -5.1% |
What a vacancy is
A vacancy is a position for which an employer is actively seeking recruits from outside the business. The figure counts positions, not job adverts, and it says nothing about applications or hires.
The estimates come from the ONS Vacancy Survey of employers. It covers the economy except agriculture, forestry and fishing, and households as employers, and its latest quarter is an early estimate that will be revised.
Each period is three months, seasonally adjusted, and compared with the three months before it and the same months a year earlier. The ONS computes changes on unrounded figures.
- A VACANCY: a position for which an employer is actively seeking recruits from outside the business or organisation. It is counted once, however many times it is advertised, and it says nothing about how many people apply.
- THE SURVEY: the ONS Vacancy Survey asks employers how many vacancies they have. It covers the whole economy except agriculture, forestry and fishing, and households as employers. July to September 2026 is an early estimate.
- THE PERIODS: vacancies are for July to September 2026; the employment and unemployment rates come from the Labour Force Survey for June to August 2026.
So 691,000 does not say that jobs were cut: it says that fewer positions were open to recruits from outside, at a time when more people were in work than a year earlier.
One survey of employers, one definition and two periods, and a reading that forgets any of them says what the source does not.
Fewer openings, more people in work
Over June to August 2026, the employment rate for people aged 16 to 64 rose by 0.2 percentage points on the quarter and on the year, to 75.3%. The unemployment rate for people aged 16 and over fell by 0.1 points on both, to 4.7%, and the economic inactivity rate was 20.9%.
HMRC payroll records count differently and point the other way: 92,000 fewer payrolled employees in June to August 2026 than a year earlier, a fall of 0.3%. The survey counts people, employees and self-employed; the payroll counts employees paid through PAYE.
There were 2.4 unemployed people per vacancy in June to August 2026, down from 2.5 in March to May.
What these figures do not say
They do not say why. The labour-cost explanation is what smaller firms tell the survey; no tax, wage rule, interest rate or technology is measured here.
They do not say that jobs were lost, how many people apply for each vacancy, or which occupations and regions are affected.
The latest quarter is an early estimate that will be revised, and none of this is a forecast.
