Using the latest available government accounts rather than treating all UK DB pensions as a direct Treasury cost.

The key figure is £60.4 billion of benefits paid in 2024-25 by unfunded public-sector pension schemes. These include NHS, teachers, civil service, armed forces, police and other schemes. The Government's accounts explicitly distinguish these from funded schemes such as local government pensions.
1. Gross saving

So, on the simplest calculation, a 40% reduction would reduce annual government expenditure by about £24.2bn.

But that's not the actual Treasury saving.

2. Allow for lost income tax

Pension income is taxable. If pensions are cut, the Treasury loses some income-tax receipts.

Because we don't have the individual income distribution of the £60.4bn in the data I'm using, I'll use a 15-20% effective tax rate as an illustration, rather than pretend it is an official estimate.

3. Then comes Pension Credit

This is where things get more complicated.

Some pensioners whose occupational pension was cut could become entitled to additional means-tested support. Pension Credit expenditure was about £6.1bn in 2025-26, so it isn't insignificant.

However, we shouldn't simply subtract £6.1bn from the saving. Most existing Pension Credit recipients would not suddenly receive another 20%, 30% or 40%, and many public-sector DB pensioners would remain well above the relevant income thresholds.

For an illustrative policy calculation, I'd model additional means-tested expenditure at perhaps 5-15% of the gross pension reduction.

That produces:

*The 5-15% assumption is illustrative, not an OBR/DWP forecast.

So, best broad estimate would be

If the Government somehow imposed an immediate cut on all currently paid unfunded public-sector DB pensions:

• 20% cut -> roughly £8-10bn annual net Treasury saving
• 30% cut -> roughly £12-14.5bn
• 40% cut -> roughly £16-19bn

Those are substantially more realistic numbers than saying a 40% cut saves £24bn.

But there's an important twist

The Government's latest accounts show £1.401 trillion of unfunded public-sector pension liabilities at 31 March 2025.

So a reduction in the promised pension isn't merely a saving on this year's expenditure.

If the reduction applied to future pension rights as well as pensions currently being paid, it would also reduce the government's enormous long-term pension liability.

But the £1.401tn figure cannot simply be multiplied by 20%, 30% or 40% to calculate a saving. It is a discounted valuation of future payments and depends heavily on assumptions about wages, inflation, interest rates and life expectancy. The Government itself warns that small changes in the discount rate can materially alter the liability.

And private-sector DB is different

This is particularly important to your original idea.

Britain has many millions of private-sector DB pension members, but cutting their pensions wouldn't give the Treasury anything like the £60.4bn saving. Those pensions are generally paid from pension funds and employer arrangements rather than directly from government expenditure.

The £60.4bn is therefore the appropriate starting point if your objective is reducing government expenditure, not if the objective is making every British pensioner financially equal.

One further calculation would be very interesting

We could take this a step further and calculate what happens to the £1.4 trillion public-sector pension liability under 20%, 30% and 40% reductions, including the effect of applying the reduction only to future accrual rather than cutting pensions already earned.

That would give us two quite different answers:

A. Cut existing pensions: large immediate cash saving, but major legal/distributional consequences.

B. Reduce future DB accrual: little immediate saving, but potentially a very large long-term reduction in government pension liabilities.

The second calculation is probably the more economically meaningful one if you're considering how the UK could make its DB/DC pension system more consistent.