
Britain’s payroll contraction is being accompanied by a marked deterioration in the balance between people entering and leaving PAYE employment, with employee outflows rising much faster than inflows. The latest ONS and HMRC PAYE Real Time Information data show that in July 2026, 576,987 people flowed into payroll employment while 596,306 flowed out, leaving a negative monthly balance of 19,319 employees.
A year earlier, the corresponding July figures produced a positive balance of 2,443, meaning the monthly flow position deteriorated by 21,762 employees in a year.
That shift matters because it reveals something less obvious from the headline employment totals. Inflows were actually slightly higher than a year earlier, increasing by 4,015, or 0.7%. Outflows, however, increased by 25,777, or 4.5%. The deterioration in July’s payroll balance therefore reflects a much larger increase in departures from PAYE payrolls than in additions to them.
The distinction is important. The data do not show why people leave payroll employment, so the outflows should not automatically be interpreted as redundancies or dismissals. What they do show is that the recent weakening in payroll employment cannot simply be characterised as a collapse in the number of people entering PAYE employment.
A comparison of July 2025 and July 2026 makes the shift particularly clear.
| Measure | July 2025 | July 2026 | Change |
|---|---|---|---|
| PAYE inflows | 572,972 | 576,987 | +4,015 |
| PAYE outflows | 570,529 | 596,306 | +25,777 |
| Net monthly flow | +2,443 | -19,319 | -21,762 |
Finance Gazette calculations from HMRC PAYE RTI data.
The calculation also reconciles precisely with the monthly payroll data. UK payrolled employment fell from 30,272,659 in June to 30,253,340 in July, a decrease of 19,319, exactly matching the difference between July's recorded inflows and outflows. The August flash estimate subsequently fell by a further 26,097 to 30,227,243.
The July flow data therefore provide a useful decomposition of that month's payroll contraction. The number of people moving onto payrolls was not materially weaker than a year earlier; the larger change was on the other side of the equation.
Nor was July an isolated negative month. Outflows exceeded inflows in nine of the 12 months from August 2025 through July 2026. September and November 2025 recorded negative flow balances, as did February, March, April, May, June and July 2026. Only August, December and January produced positive balances during that period.
The pattern became particularly persistent in 2026. January still showed inflows exceeding outflows by 9,973, but every month from February through July produced a negative balance. June recorded 574,054 inflows against 596,122 outflows, a deficit of 22,068, followed by July's 19,319 deficit.
That does not establish what happens next, but it does show that the weakening payroll total has been accompanied by several consecutive months in which outflows exceeded inflows.
The more recent August payroll estimate does not yet have equivalent flow data attached to it, but the headline employment series suggests the contraction continued.
HMRC and the Office for National Statistics estimate there were 30.227 million payrolled employees in August 2026, down approximately 145,000, or 0.5%, from August 2025. Employment was also about 26,000 lower than in July.
The longer series shows that payroll employment has moved down from 30.428 million in August 2024 to 30.372 million in August 2025 and 30.227 million in August 2026. That represents a decline of roughly 201,000 payrolled employees over two years, although the latest August observation remains provisional.
The August number should therefore not be treated as final. ONS says its early estimates are based on around 85% of the information eventually expected, with further submissions incorporated in subsequent releases. The latest observation is consequently more exposed to revision than earlier periods.
Around 15% of data for the latest reference month are initially imputed. As actual returns arrive, those imputed values are replaced, while seasonal adjustment can create additional revisions to the series.
That limitation matters for August, but it does not remove the broader pattern visible in the more mature July figures.
Another unusually revealing part of the PAYE dataset is the distribution of the employment decline by age.
Between August 2025 and August 2026, the number of payrolled employees aged 25 to 34 fell by 144,495, almost numerically matching the entire 144,849 decline in UK payroll employment over the period.
That does not mean workers aged 25 to 34 account for all of the national decline. Other groups moved sharply in both directions, producing significant offsets.
Payrolled employment among 50-to-64-year-olds fell by 90,876, while the under-18 category declined by 41,636. In contrast, employment among workers aged 65 and over increased by 61,534, the 35-to-49 category increased by 44,355 and 18-to-24-year-olds increased by 26,270.
| Age group | Change Aug 2025-Aug 2026 | Annual change |
|---|---|---|
| Under 18 | -41,636 | -9.6% |
| 18-24 | +26,270 | +0.8% |
| 25-34 | -144,495 | -2.1% |
| 35-49 | +44,355 | +0.4% |
| 50-64 | -90,876 | -1.1% |
| 65+ | +61,534 | +4.9% |
The contrast is significant because it shows that the modest-looking 0.5% national decline is the net result of much larger movements inside the workforce. The PAYE population aged 65 and over continued to expand relatively quickly even as payroll numbers among 25-to-34 and 50-to-64-year-olds contracted.
Again, the data should not be pushed further than they allow. PAYE RTI does not establish why employment among people aged 65 and above increased, nor why the 25-to-34 category fell. Demographics, retirement decisions, migration, changes in employer demand and movements into or out of self-employment would require separate evidence before being presented as explanations.
What can safely be concluded is that Britain's payroll decline has been highly uneven by age.
The same qualification applies across industries. Administrative and support services added around 72,000 payrolled employees over the year, while wholesale and retail lost approximately 76,000. Annual employment growth ranged from a 3.0% increase in administrative and support services to a 3.3% decline in accommodation and food services.
The aggregate payroll contraction is therefore not a uniform reduction occurring at the same rate throughout the economy. Expanding sectors and age groups are partially offsetting considerably weaker areas.
That makes the flow data particularly useful. Headline employment levels describe the net result after those competing movements have taken place; inflows and outflows provide another way of seeing how that balance is being generated.
There is a further tension within the PAYE figures. Although payroll employment is contracting, median monthly pay has continued to increase.
The provisional August estimate puts median monthly pay at £2,657, up 3.5% from £2,567 in August 2025. ONS nevertheless notes that the rate of median pay growth has fallen in recent months after remaining relatively stable through much of 2023 to 2025.
The detailed distribution also shows that pay growth is not uniform. In the three-month moving-average data through July, annual pay increased by 5.6% at the 10th percentile and 5.3% at the 25th percentile, compared with 4.5% at the median and 3.9% at the 90th percentile. Growth at the 95th percentile was 3.6%, although the 99th percentile increased by 4.8%.
Those figures do not change the employment story, but they reinforce the importance of avoiding a single headline description of the labour market. Payroll numbers, payroll flows and earnings distributions are moving differently.
The most important test now is whether the flow imbalance visible through July persists when August's detailed data become available.
The August flash estimate already indicates another monthly payroll reduction of around 26,000 employees, but it cannot yet establish how much of that change reflects movements in inflows versus outflows. It is also more exposed to revision because of the amount of imputation used in the newest observation.
The next PAYE RTI release is scheduled for 20 October 2026.
If the eventual August flow data again show relatively stable inflows alongside elevated outflows, that would extend the pattern visible through the first seven months of 2026. If the balance instead changes materially, it would weaken the case that the recent outflow pattern is becoming persistent.
For now, the strongest conclusion from the underlying data is narrower but more defensible: Britain's payroll decline is occurring alongside a pronounced deterioration in employee flows, with outflows rising much faster than inflows and exceeding them for six consecutive months through July.
That provides considerably more information than the headline 145,000 decline alone.
