What the evidence shows
About 80% of the overall increase in UK payroll jobs from July 2014 to December 2025 was recorded under non-UK nationalities. This does not measure all new hires or establish that British workers were replaced.
- The source table shows about 3.67 million more payroll jobs overall, including about 2.94 million more in non-UK categories.
- Nationality is recorded at National Insurance registration and can remain non-UK after someone becomes British.
- The figures do not identify the effect of migration on job losses or wages.
The headline is about extra jobs, not every hire
The Centre for Social Justice published its State of Dependency report on 13 October 2026. Its announcement says only one in five added payroll jobs since 2014 went to British nationals. That is another way of saying about four in five were recorded under non-UK nationalities.
We traced the figures to HMRC, the UK tax authority. The report compares July 2014 with December 2025. The calculation works, but it describes the overall increase between those dates. It does not count everyone hired during those eleven years.
Sources: Centre for Social JusticeCentre for Social JusticeHM Revenue and Customs
Where the 80% comes from
The UK had about 3.67 million more payroll jobs in December 2025 than in July 2014. About 2.94 million of that increase was recorded under EU and non-EU nationalities combined. Divide 2.94 million by 3.67 million and you get about 80%.
| HMRC category | July 2014 | December 2025 | Increase |
|---|---|---|---|
| All payroll jobs | 28,393,100 | 32,063,800 | 3,670,700 |
| UK nationality | 24,869,400 | 25,602,900 | 733,500 |
| EU nationality | 1,839,000 | 2,246,800 | 407,800 |
| Non-EU nationality | 1,684,800 | 4,214,100 | 2,529,300 |
| EU plus non-EU | 3,523,800 | 6,460,900 | 2,937,100 |
Payroll jobs means jobs paid through employers. These figures are monthly averages. They do not count self-employed work, and someone with two employee jobs can count twice. They are neither a count of unique people nor a count of new arrivals.
Sources: HM Revenue and CustomsHM Revenue and CustomsHM Revenue and Customs
A workplace can hire many people without adding many jobs
Imagine a workplace with 100 jobs at the start of the year and 110 at the end. It added ten jobs overall. But it could have hired far more than ten people if some employees left and were replaced. This is an example, not UK data.
That is why the overall increase cannot tell us which group received 80% of all hires. To answer that question, we would need data that track people starting jobs, with clear definitions of who counts as a migrant.
Sources: HM Revenue and Customs
A nationality label can stay the same after someone becomes British
HMRC uses nationality recorded when a person registers for a National Insurance number, an identifier used for tax and benefits. The nationality record is not updated when someone later becomes a British citizen.
Some people counted in a non-UK category can therefore be British citizens today. We do not have a reliable count of how many this affects in these totals. We cannot subtract an invented adjustment.
The categories do not show ethnicity, religion, visa route or asylum status. They also do not tell us when someone arrived in Britain.
Sources: HM Revenue and Customs
The figures do not show that British workers were replaced
The UK category grew by about 733,500 payroll jobs over the report's full July 2014 to December 2025 period. It still accounted for about 79.85% of all payroll jobs in December 2025. That is a share of all jobs, different from the 80% share of growth.
A shorter comparison looks different. From January 2020 to December 2025, the UK category fell by 186,300 jobs, the EU category fell by 352,000, and the non-EU category grew by 2,113,300.
Those totals cannot tell us how many British workers lost a job because of migration, or what happened to wages. They do not link an individual departure to an individual hire. That does not establish that migration has no effect; it means this table cannot measure the effect.
A fairer seasonal comparison still gives about four in five
HMRC recommends comparing the same month in different years because this series has not had seasonal changes removed. Comparing December 2014 with December 2025 gives 79.13%, rather than the report's 80.01% from July 2014.
| Period | Total increase | Non-UK increase | Share |
|---|---|---|---|
| July 2014 to December 2025 | 3,670,700 | 2,937,100 | 80.01% |
| December 2014 to December 2025 | 3,586,100 | 2,837,600 | 79.13% |
This is a useful sensitivity check. It does not overturn the rough four-in-five figure. It also does not fix the separate distinction between overall growth and all hiring.
Where the more recent non-EU growth was recorded
For January 2020 to December 2025, health and social work accounted for 617,000 of the extra non-EU-category jobs, about 29.2% of that category's national increase. Administrative and support work added 336,800, accommodation and food work 255,100, and wholesale and retail work 218,200.
We also checked the regional tables: the UK category fell in nine of the twelve reported regions, while the non-EU category rose in all twelve. These are descriptions of where growth was recorded, not evidence of why it happened or which individual workers it affected.
Sources: HM Revenue and Customs
What we have checked and what we still need
Checked: the published table reproduces roughly 80% as the non-UK share of overall payroll growth. The tidy-format table gives the same endpoints.
Still unclear: the share of all new hires received by migrants, the number of later British citizens in historical non-UK categories, and the effect of migration on British workers. Those questions need different evidence.
We have prepared focused enquiries to CSJ and HMRC, but have not sent them. No response or data-refresh reminder is scheduled. We will reassess the finding if further evidence arrives; this is not a promise of a reply.
Sources: Centre for Social JusticeHM Revenue and CustomsHM Revenue and Customs
Technical notes: exact cells, rounding and the report's dates
Use worksheet 3, cells B4:E4 for July 2014 and B141:E141 for December 2025. December 2014 is B9:E9. Column B is total, C UK, D EU and E non-EU. The tidy file cross-check is B49:C51 and B6899:C6901.
The source rounds each published category separately to the nearest 100. The full-period category changes add to 100 more than the separately rounded total change. We retain that rounding difference rather than forcing categories to balance.
The report's printed page 188, PDF page 190, refers to January 2014 as the earliest point, but its footnotes 34 and 684 refer to July. The reproduced numbers match July. This date inconsistency does not overturn the 80% calculation.
The report compares jobs from January 2020 to December 2025 with benefit figures from February 2020 to November 2025. Those are different periods and populations. They are not records linking the same people moving between benefits and work.
Sources: Centre for Social JusticeHM Revenue and CustomsHM Revenue and CustomsHM Revenue and Customs
What is established
- The full-period calculation is 80.01%; comparing December with December gives 79.13%.
- The aggregate and tidy tables agree on the endpoints and definitions.
What remains unknown
- How many hires went to migrants and how many people later became British citizens.
- Whether and how migration changed jobs or wages for British workers.
- Enquiries are prepared but unsent. No update is scheduled.
The story, in order
First month in this comparison
The starting monthly average in the HMRC table
Last month in the report's comparison
The ending monthly average
Official figures published
HMRC publishes the July 2014 to December 2025 release; files were revised on 22 May
CSJ report published
The announcement describes one in five added payroll jobs as British
YOUR TURN TO CHECK
Reproduce this investigation
Repeat the 80% calculation using the same original published spreadsheet. No advanced statistics are needed.
You will need
Excel or LibreOffice, the official ODS file and a calculator
What you can and cannot repeat
You can reproduce the recorded growth and its share. These aggregate tables cannot tell you every hire, current citizenship or the causal effect of migration.
Download the original HMRC table
Open the official release and download the nationality, region and industry ODS file. Open it in Excel or LibreOffice. Find worksheet 3, which reports UK monthly counts.
Read the two months
Read B4:E4 for July 2014 and B141:E141 for December 2025. Keep total, UK, EU and non-EU categories separate.
July: 28,393,100 | 24,869,400 | 1,839,000 | 1,684,800 December: 32,063,800 | 25,602,900 | 2,246,800 | 4,214,100Keep in mind These are monthly average jobs, not a count of every hire
Subtract the start from the end
Calculate the change in each category. Add the EU and non-EU changes together. Divide that by the change in the total.
Total: 32,063,800 - 28,393,100 = 3,670,700 Non-UK: (2,246,800 - 1,839,000) + (4,214,100 - 1,684,800) = 2,937,100 2,937,100 / 3,670,700 * 100 = 80.0147%Keep in mind Do not relabel this as the share of all hires
Compare December with December
Repeat the calculation using row 9 instead of row 4. This compares the same month in both years.
2,837,600 / 3,586,100 * 100 = 79.1277%Keep in mind The series is not seasonally adjusted
Check the meaning before drawing a conclusion
Read the commentary definitions and nationality limitations. Compare the source table with the original report. The figures alone cannot measure workers being displaced.
Sources you can check
- Just one in five added payroll jobs since 2014 went to British nationalsCentre for Social Justice / 13 October 2026Original report announcement, not an Instagram post
- State of Dependency reportCentre for Social Justice / 13 October 2026Original claim and jobs/benefits comparisons
- UK payrolled employments by nationality: July 2014 to December 2025HM Revenue and Customs / 14 April 2026Official release, updated 22 May 2026
- Commentary and definitionsHM Revenue and Customs / 22 May 2026What payroll jobs and nationality categories mean
- Nationality, region and industry tablesHM Revenue and Customs / 22 May 2026Original aggregate spreadsheet for reproducing the calculation
- Monthly counts by nationalityHM Revenue and Customs / 22 May 2026Independent format cross-check of the same monthly endpoints
We downloaded the original HMRC tables and CSJ report, reproduced the arithmetic and cross-checked the same data in a second file format. We did not obtain individual hiring records or estimate a causal effect.
Inspect the evidence
Source notes include the links and verification limits. The timeline opens in Excel or any spreadsheet app.
Photo credit and licence
Three construction workers in high-visibility clothing seated with backs to the camera in Berkeley Square, London, 3 June 2019. Neil Theasby / Wikimedia Commons. CC BY-SA 2.0. Resized for display and cropped to fit. Homepage photographs also have a dark gradient and text overlay. Original and adapted photograph: CC BY-SA 2.0. The separate TALLYCAT01 logo remains brand artwork.
Carousel photograph credits and licences
- Three construction workers in high-visibility clothing seated with backs to the camera in Berkeley Square, London
Neil Theasby. Archive London work context. The photograph does not identify the workers' nationality or establish any hiring statistic.
CC BY-SA 2.0. Cropped, resized and darkened for the carousel. The adapted photograph and its carousel slide are offered under CC BY-SA 2.0. The separate logo remains brand artwork. - Anonymous hands holding a pen and reviewing printed material outdoors
Vanessa Garcia. Report-reading illustration. This is not the CSJ report or its author.
Pexels License. Cropped, resized and darkened for the carousel. - A real calculator on a desk beside file folders, notebook and stamp
Mikhail Nilov. Calculation illustration. This is not the original HMRC table or a photographed payroll record.
Pexels License. Cropped, resized and darkened for the carousel. - Office desks and chairs
Gustavo Galeano Maz. Workstation illustration. This is not a British workplace or evidence of the hypothetical 100-to-110 example.
Pexels License. Cropped, resized and darkened for the carousel. - British passport cover, 2020
Paulwidnes1; existing Commons crop by Seewhyso. British passport cover photographed in 2020. It does not identify a person or establish when anyone became British.
CC BY-SA 4.0. Cropped, resized and darkened for the carousel. The adapted photograph and its carousel slide are offered under CC BY-SA 4.0. The separate logo remains brand artwork. - A butcher working at Borough Market in London; genuine archive photograph
Adrian Snood. Archive London work context, 2020. No claim is made about this worker's nationality or employment history.
CC BY2.0. Cropped, resized and darkened for the carousel. - Magnifying glass and paperclips
Kader D. Kahraman. Evidence-checking illustration. It does not show the original HMRC or CSJ evidence.
Pexels License. Cropped, resized and darkened for the carousel. - Hands using a smartphone
cottonbro studio. Phone-reading illustration. The phone does not display the TALLYCAT01 article.
Pexels License. Cropped, resized and darkened for the carousel.
These credits cover the carousel photographs. Original stock files are not distributed here. The logo remains separate brand artwork.
Research and production assisted by AI. Read our approach
Have evidence that could change this finding? Contact TALLYCAT01