Skip to content
Sign inGet started free
Photo: elliott / Unsplash · United Kingdom

Market watch · 24

Fewer people leave UK payrolls, and fewer still join them

In the year to July 2019, 274,000 more people joined UK payrolls than left them; in the year to July 2026, 101,000 more left than joined.

When the number of people on UK payrolls falls, it can be because more people are leaving paid work or because fewer are getting into it. The tax records every employer files count both flows every month, and they give a clear answer: fewer people are leaving payrolled work than before the pandemic, and fewer still are coming in. Here is how the two flows compare with 2019, what a flow counts and what it does not, and what it changes if you hire in the UK or want to get onto a payroll there.

1.91%joined a UK payroll in an average month of the year to July 2026, against 2.33% in the year to July 2019
1.94%left payrolled work in an average month, against 2.25% in the year to July 2019
-101,000is the change in the payroll count over the year to July 2026, after it grew by 274,000 in the year to July 2019

Joining a UK payroll slowed more than leaving one

Joined 20192.33%
Left 20192.25%
Joined 20261.91%
Left 20261.94%
People joining and leaving payrolled employment in an average month, as a share of the previous month’s payrolled employees, twelve months to July, UK, seasonally adjusted. Derived by the desk from the counts of the ONS and HMRC release of 15 September 2026, read 28 September 2026.

Fewer people are leaving

In the twelve months to July 2026, 7.05 million people stopped being paid through a UK payroll, against 7.77 million in the twelve months to July 2019. The count comes from the real-time records every UK employer files with HM Revenue and Customs each time it pays its staff, published by the Office for National Statistics on 15 September 2026.

The number of people on payrolls grew between the two years, from 28.9 million in July 2019 to 30.3 million in July 2026, so the rate fell further than the number: in an average month of the year to July 2026, 1.94% of payrolled employees left payrolled work, against 2.25% in the year to July 2019.

Leaving here means no longer being paid through any UK payroll: retiring, studying, caring, moving abroad, becoming self-employed and losing a job all count, together. The tables do not separate them, so a lower figure says that fewer people left paid employment, not why.

Fewer still are joining

6.95 million people started being paid through a UK payroll in the twelve months to July 2026, 1.09 million fewer than the 8.04 million of the twelve months to July 2019. In an average month, 1.91% of payrolled employees were joiners, against 2.33%.

Both flows slowed, and joining slowed more. Joiners fell by 1.09 million a year and leavers by 714,000. In the year to July 2019, 274,000 more people joined payrolls than left them, and the count grew by that much; in the year to July 2026, 101,000 more left than joined, and the count fell by 101,000.

It is not one bad month. In 18 of the 24 months from August 2024 to July 2026, more people left UK payrolls than joined them, and the year to July 2025 looked the same: 7.13 million joined and 7.22 million left, a fall of 95,000.

Provenance: Source
People who started (joined) and stopped (left) being paid through a UK payroll, summed over the twelve months to July of each year, and the change in the payroll count over the same months, which is joiners minus leavers. Seasonally adjusted counts from the ONS and HMRC; the flows begin in January 2017, and July 2026 is an early estimate.
Twelve months to JulyJoinedLeftChange in the count
20188,017,3457,696,726+320,619
20198,042,1157,767,786+274,329
20207,096,6937,674,803-578,110
20217,268,8796,794,163+474,716
20228,413,2917,559,545+853,746
20238,146,1657,610,263+535,902
20247,656,5187,386,349+270,169
20257,128,3547,223,043-94,689
20266,952,4127,053,710-101,298

What a payroll flow counts

It counts people, not jobs. A person joins in a month in which they are paid through a UK payroll and were not in the month before, and leaves in the reverse case. Someone who moves straight from one employer to another stays in the count and appears in neither flow, so these figures say nothing about how often people change jobs.

A joiner is not a hire and a leaver is not a lay-off. A joiner may be starting a first job, returning from study, caring or illness, arriving from abroad or moving from self-employment; a leaver may be retiring, studying, emigrating, becoming self-employed or losing a job.

The latest months are early estimates. The ONS notes that early estimates of both flows have historically been revised down in the following month, by broadly equal proportions, which moves the two together and leaves the gap between them largely where it is.

Provenance: SourceWhat this counts
  1. JOINING and LEAVING are movements of PEOPLE, not jobs. Someone joins in a month in which they are paid through a UK payroll and were not in the month before, and leaves in the reverse case. The count is of people, not of jobs, so a person who moves straight from one employer to another stays in it and is in neither flow.
  2. A JOINER IS NOT A HIRE and a LEAVER IS NOT A LAY-OFF. A joiner may be starting a first job, returning from study, caring or illness, arriving from abroad or moving from self-employment; a leaver may be retiring, studying, emigrating, becoming self-employed or losing a job. The published tables do not say which.
  3. The count of payrolled employees is a STOCK and the two flows move it: over any twelve months, joiners minus leavers equals the change in the count. The ONS publishes counts; every RATE here is the desk’s arithmetic, the month’s flow divided by the previous month’s count and averaged over the twelve months to July.

So 1.91% against 1.94% does not say that employers are laying people off or hiring fewer staff: it says that in an average month of the year to July 2026 slightly fewer people started being paid through a UK payroll than stopped, where in the year to July 2019 it was 2.33% against 2.25%.

Two flows and one count, from the tax records every UK employer files each time it pays its staff. A reading that forgets what they count says what the source does not.

What these counts do not say

They name no cause. The rise in employers’ National Insurance contributions in April 2025, increases in the minimum wage, artificial intelligence and interest rates have all been offered to explain a weaker UK market, and none of them is measured here.

They say nothing about vacancies, about how many people apply for each post, or about pay. And the published flows are not split by age, industry or region, so they cannot say who joined or who left.

And they forecast nothing. The next release, on 20 October 2026, will revise the latest months; nothing here says where either flow goes next.

What this means for you

If you hire in the UK

Do not read the falling payroll count as a wave of people let go and looking for work. 7.05 million people left payrolled work in the year to July 2026, 714,000 fewer than in the year to July 2019, and that count takes in retirements, study and moves abroad as well as lost jobs. The count fell because fewer people joined, not because more left: if you expected people released by other employers to fill your openings, the tax records show fewer of them than in 2019, not more.

If you want to get onto a UK payroll

6.95 million people started being paid through a UK payroll in the year to July 2026, 1.09 million fewer than in the year to July 2019, while the number of people on payrolls was 1.4 million higher. The source does not say how many were trying, so it gives no odds; it shows that the way in has narrowed more than the way out. If you are coming from study, a career break, self-employment or abroad, you are in the flow that shrank: start earlier, and apply more widely than a 2019 search would have needed.

How this was counted

Method

  1. One release. Office for National Statistics and HM Revenue and Customs, “Earnings and employment from Pay As You Earn Real Time Information, UK: September 2026”, published on 15 September 2026, seasonally adjusted tables, read on 28 September 2026 (the date is the timestamp of the downloaded file, 21:31). Table 6 gives the monthly flows from January 2017 to July 2026, and table 1 the count of payrolled employees. The next release is due on 20 October 2026.
  2. TWELVE MONTHS TO JULY. Each year is the sum of the published monthly flows from August to July, so every year covers the same months, and July 2026 is the latest month with flows. Over each twelve months, joiners minus leavers is the change in the count: 6,952,412 joiners and 7,053,710 leavers make a fall of 101,298, and the count went from 30,354,638 in July 2025 to 30,253,340 in July 2026.
  3. RATES, DERIVED. The ONS publishes counts, not rates. Each rate here is the desk’s: the month’s flow divided by the previous month’s payrolled employees, averaged over the twelve months. That gives 2.33% joining and 2.25% leaving a month in the year to July 2019, and 1.91% and 1.94% in the year to July 2026.
  4. EARLY ESTIMATES. The count for August 2026, 30,227,243, is an early estimate with no flows yet, and it is not used. The ONS notes that early estimates of the flows have historically been revised down in the following month, by broadly equal proportions.
  5. WHAT IS COUNTED. People paid through PAYE in the month, including people on paid leave, counted as people and not as jobs; pension income and self-employment are outside it. The flows begin in January 2017, so the first complete twelve months to July end in July 2018. The release also counts payrolled employees by age, industry and region; those counts are stocks, not flows, and this edition does not use them.
What the data establishes, and what it does not

What the data establishes

  • In the twelve months to July 2026, 6.95 million people started being paid through a UK payroll and 7.05 million stopped; in the twelve months to July 2019, 8.04 million and 7.77 million.
  • In an average month of the year to July 2026, 1.91% of payrolled employees joined and 1.94% left, against 2.33% and 2.25% in the year to July 2019: both flows slowed, and joining slowed more.
  • The payroll count fell by 101,000 over the year to July 2026, after growing by 274,000 over the year to July 2019; in 18 of the 24 months from August 2024 to July 2026, more people left than joined.
  • Since the year to July 2019, joiners have fallen by 1.09 million a year and leavers by 714,000.

What it does not

  • Any cause. Employers’ National Insurance, the minimum wage, artificial intelligence and interest rates have all been offered, and none of them is measured here.
  • Lay-offs or redundancies. Leaving payrolled work also means retiring, studying, emigrating or becoming self-employed, and the tables do not separate them.
  • How often people change jobs. A move straight from one employer to another is in neither flow.
  • Anything about vacancies, applicants or pay, or about who joined or left: the published flows are not split by age, industry or region.
  • Any forecast. The next release, on 20 October 2026, will revise the latest months.
Primary sources

BabZituna scores every job against your profile on six real dimensions, and publishes what it can and cannot measure. See the live market.

The markets in this piece, live

Earlier editions