UK Unemployment Holds at 4.9 Percent as Payrolls Fall 145,000 and Pay Growth Slows to 3.9 Percent
Britain’s unemployment rate was 4.9 percent in the three months to July, up 0.2 percentage points on the year and largely unchanged on the quarter. The Office for National Statistics published the figures at 07:00 London time on Tuesday. Underneath the steady headline, its early estimate of payrolled employees for August fell 145,000 on the year to 30.2 million, vacancies dropped to their weakest level outside the pandemic since 2014, and total pay growth slowed to 3.9 percent, the lowest reading since late 2020.
The two measures are telling different stories and the Statistics Office says plainly which one it trusts more for employees. Its stated view is that the tax data currently provides the most reliable measure of employees; the survey remains the only source for unemployment and inactivity.
The survey is flat and the tax data is falling
| Measure, three months to July 2026 | Rate | Change on the year |
|---|---|---|
| Unemployment, 16 and over | 4.9% | +0.2 pts |
| Employment, 16 to 64 | 75.1% | -0.1 pts |
| Economic inactivity, 16 to 64 | 20.9% | -0.1 pts |
Labour Force Survey estimates as published. All three are described by the Statistics Office as largely unchanged on the latest quarter.
Against that, the payrolled employee count from tax records fell 101,000, or 0.3 percent, between July 2025 and July 2026, and 19,000 on the month. On the three month window that matches the survey, payrolls fell 84,000 on the year and 39,000 on the quarter. The early August estimate, which the Statistics Office marks provisional and likely to be revised, is a fall of 145,000 on the year and 26,000 on the month.
The claimant count rose on both the month and the year, to 1.692 million in August, also provisional.
Vacancies are back where they were in 2014
| Vacancies, June to August 2026 | |
|---|---|
| Level | 702,000 |
| Change on March to May | -8,000, or 1.1% |
| Change since January to March | -16,000 |
Early estimates as published.
The Statistics Office makes the comparison itself: outside the coronavirus period, the last time there were 702,000 or fewer vacancies was August to October 2014, when there were 701,000. That is twelve years, on our calculation.
It also passes on what employers are telling its Vacancy Survey, which is that smaller firms may not be recruiting because of increases in labour costs. That is the release naming a cause, and it is worth noticing that a statistical bulletin has done so.
The flatness matters as much as the level. The Statistics Office describes the estimates as broadly flat since the start of the year, with a fall of only 16,000 since the first quarter. So this is not a collapse in hiring demand inside the quarter; it is a level that has settled twelve years back and has not moved off it.
Pay growth is slowing, and the public and private sectors have split
| Average weekly earnings, three months to July 2026 | Annual growth |
|---|---|
| Regular pay, excluding bonuses | 3.5% |
| Total pay, including bonuses | 3.9% |
| Regular pay, public sector | 6.3% |
| Regular pay, private sector | 2.9% |
| Regular pay, real terms on CPIH | 0.6% |
| Total pay, real terms on CPIH | 0.9% |
As published. The Statistics Office notes that public sector growth continues to be affected by variations in the timing of pay awards this year.
Total pay growth of 3.9 percent is down from 4.2 percent in the previous three month period, and the Statistics Office puts the last lower reading at September to November 2020, when it was 3.7 percent. Regular pay at 3.5 percent has been relatively stable across five consecutive three month periods after a year of slowing growth.
The gap between the sectors is 3.4 percentage points on our calculation, with public sector workers seeing regular pay rise more than twice as fast as private sector workers. Real regular pay growth of 0.6 percent means the average worker is barely ahead of prices.
The Bank of England saw these numbers a day early
The bulletin discloses that the Bank of England was granted exceptional pre-release access to it, and to the accompanying tables, at 10:00 on Monday 14 September, so that the data were available for the Monetary Policy Committee meeting held that day. The Statistics Office publishes an exchange of letters covering the arrangement.
The Committee announces its decision on Thursday 17 September.
Why it matters: the headline rate has not moved and everything beneath it has softened. Payrolls are down 145,000 on the year on the early August estimate, vacancies are at a level last seen in 2014 outside the pandemic, and total pay growth is the weakest in almost six years. The Statistics Office’s own position is that the tax data is the more reliable measure of employees, which points at the falling series rather than the flat one. The split between 6.3 percent public sector pay growth and 2.9 percent private is the sharpest single number in the release, and it is the one that will be read hardest before Thursday.
Outlook: the Monetary Policy Committee decides on 17 September having had this release since Monday morning, which is disclosed in the bulletin itself. The next labour market release is 20 October, and the figure to watch is whether the August payroll estimate of minus 145,000 survives revision, since the Statistics Office has flagged it as provisional and first estimates on this series move.
Sources: Office for National Statistics.
