How the pension savings are funding this year's pay award

What's happened

On 16 September 2026, the government confirmed that schools will keep a saving on their pension costs, rather than having it clawed back through next year's funding. This is helping to fund this year's pay awards for teachers and support staff.

Where the saving comes from

Every three years, an independent valuation resets how much employers must pay into the Local Government Pension Scheme (LGPS), which covers support staff.
Investment returns have been stronger than expected since the last valuation, so the scheme needs less from employers to pay the same pensions. Employer contributions have fallen by 4.9 percentage points on average.
This does not affect what any support staff member will receive when they retire. It only changes how much the employer pays in.

Why “not clawing it back” matters

Each year, DfE publishes a technical note called Schools' Costs. This is the document DfE itself uses to decide how much extra funding schools need, and it directly determines how much of any staff pay award the department considers “affordable” from that funding alone, before any further government intervention.

A fall in pension costs like this one would normally reduce DfE's own assessment of the cost pressure schools face in that note and so reduce how much extra funding it judges schools need next year, and in turn how much of a pay award it judges affordable without the government adding more money.

This is not a hypothetical risk. Through NEU campaigning, the union has repeatedly forced government to provide funding, for example the Mainstream Schools' Additional Grant. The government's commitment on 16 September 2026 means the LGPS saving will not be used to reduce next year's affordability assessment in that way. The appendix sets out exactly what each year's Schools' Costs technical note has said about LGPS costs.

The numbers

£700 million: extra funding announced in July 2026 towards this year's pay awards.
£500 million: further funding announced on 16 September 2026, from the pension saving.
Together, the government says this fully funds this year's 3.5 per cent teacher pay rise.
£1.6 billion: how much better off the government says schools will be next year than they expected to be over the summer.

Independent verification

This isn't just the government's own assessment. Luke Sibieta, Research Fellow at the Institute for Fiscal Studies, said:

“Yesterday, the government announced an effective £500 million boost to school funding in England, which equates to a permanent increase of just under 1%. This is due to the government no longer clawing back savings that have resulted from lower employer pension contributions for support staff. Alongside the £700 million top-up announced for this year in July, today's £500 million boost means that the government is now fully funding the cost of this year's 3.5% rise in teacher pay. Coming on top of other funding increases, this means that from next year schools will be £1.6 billion better off than they had expected to be this summer.”

Luke Sibieta, Research Fellow, Institute for Fiscal Studies

One thing to watch

This commitment applies only to the LGPS saving for support staff. A much larger saving on teachers' own pensions is due in April 2027, and the government has already said that one will be clawed back through cuts to school funding. Activists should expect this to be a live argument over the coming year.

LGPS in every schools' costs

Technical note

Every place the Local Government Pension Scheme (LGPS) is mentioned in each edition of DfE's Schools' Costs technical note, in chronological order. Each edition is hyperlinked to its PDF on gov.uk.

LGPS is explicitly excluded from this edition. Paragraph 5: “we are not able to assess any effects of future variations to employer rates for the Local Government Pension Scheme and Teachers’ Pension Scheme at this time.”

Paragraph 30 (methodology): “The LGPS cost changes are averaged across 79 funds; contribution rates vary widely at both the fund and the employer level (local authority or academy trust).” Rates changed +1.1, -0.4 and +0.2 per cent over the three years of the 2016 valuation.

Paragraph 26 (methodology): “The reduction in the rate of 0.5 per cent of payroll in 2020-21 corresponds to a pay bill cost pressure in 2020-21 of around -0.4 per cent.”

Paragraph 55 (methodology): “The reduction in the rate of 0.3 per cent of payroll in 2021-22 corresponds to a pay bill cost pressure... of around -0.2 per cent. For 2022-23, we estimate a zero pressure. Extrapolating to 2023-24, we estimate a pressure of 0.3 per cent.”

Paragraph 50 (methodology): “Following on from the 2021-22 pay bill pressure, we estimate negligible pressure in 2022-23. Extrapolating using previous years’ patterns of rate changes, we estimate a pressure of 0.3 per cent in 2023-24.”

Paragraph 57 (methodology): “Using the 2022 valuations report, we estimate a pressure of -1.1 per cent in 2023-24 and -0.1 per cent in 2024-25.”

Paragraph 75 (methodology): “Using the 2022 valuations report, we estimate a pressure of −0.1 per cent in 2024-25 and 0.0 per cent in 2025-26.”

Paragraph 63, footnote 28: “LGPS changes are based on historical trends in the three-year pension re-valuation cycle.”

Paragraph 111 (methodology, section B.5): “From the 2022 valuation report, and using that and earlier reports to provide an average three-year cyclical trend beyond 2025-26, we estimate pressures of:” −0.1% (2024-25), 0.0% (2025-26), −0.2% (2026-27), −0.3% (2027-28), 0.1% (2028-29).

This is the edition in place when Powell's letter was written. Its LGPS assumption is an extrapolated trend, not the actual 2025 valuation outcome - which is exactly why the real change, left alone, would only have shown up in next year's edition.