The Shape of UK Public Spending
A companion to the Tax Policy Associates map of the 85 taxes that fund the British state, built from the same OBR data. This is the other side of the ledger: every pound of public spending in 2024-25, arranged by what it actually buys.
Late last month, Dan Neidle and Tax Policy Associates published a single chart showing all 85 UK taxes on one page, each sized by how much it raises. It is the clearest picture of the revenue side of the public finances I have seen in a long time. It also left an obvious question hanging in the air. If that is where the money comes from, where does it go?
This is the other side of that ledger. It is built from the same vintage of OBR data, with the spending broken down by function: not which department holds the budget, but what the money actually buys. Click any segment to drill in. Click the centre, or Reset, to come back out.
The shape.
Two facts sit on top of everything else. The first is that just over 43p in every pound of GDP now passes through the public sector. Total Managed Expenditure was £1,285bn in 2024-25, against a GDP of £2,934bn. The second is how concentrated that spending is. Social protection, which is pensions and benefits, is £383.9bn on its own, close to 30p in every pound spent. Health is £241.8bn, another 19p.
Underneath those totals are two numbers worth saying out loud. The single largest line anywhere in government is the state pension, at £146.5bn. And the third largest thing the state buys, after the NHS and pensions, is interest on its own debt: £124.7bn, more than the schools and universities budgets combined.
Two charts that do not balance.
It is tempting to lay this chart over the tax chart and read the gap between them as the deficit. It does not work like that, and the reason is worth understanding.
The tax chart's total is National Accounts taxes, £1,013bn, or 34.5 percent of GDP. That counts taxes only. The spending here is all outlays, £1,285bn. The gap between the two is wider than the deficit, because the spending side includes everything while the tax side leaves out around £126bn of non-tax receipts, things like interest, dividends, and the surpluses of public corporations, as well as the deficit itself, which was about £153bn in 2024-25. Two honest charts of the same year do not balance, because they are measuring two different things.
The grey wedge.
One slice is deliberately dull and grey, and it is large: £127.4bn, just under a tenth of the total. It is not a service. It is mostly depreciation, the capital consumption of the public sector's buildings and equipment, plus the National Accounts adjustments needed to move from expenditure on services, which is what the functional breakdown measures, up to Total Managed Expenditure.
I have left it visible rather than quietly dropping it. The alternative is a chart whose wedges sum to less than the headline figure and never admits it, and that is the sort of small dishonesty that makes a picture untrustworthy.
Where the numbers come from.
The functional split is HM Treasury's PESA 2025, Table 5.2: public sector expenditure on services by function and sub-function, 2024-25 outturn. It is the same table HMRC uses to build the annual tax summary, so the categories are the official ones rather than mine. The headline total and the GDP figure come from the OBR public finances databank, and reconcile to PESA's total to within £0.4bn.
Where a sub-function fully decomposes, the "of which" lines are shown as the outer ring, which is how the state pension surfaces as its own slice inside old age. The spending data is published under the Open Government Licence. The companion tax chart, and the idea of drawing the system this way, are Tax Policy Associates'.
Seeing it whole.
I rebuilt this because the original made a complicated thing legible in about ten seconds, and that is most of the job. A good deal of the advisory work I do is the same instinct applied to AI systems: taking something with dozens of moving parts that no single person holds in their head, and arranging it so the shape is visible and the trade-offs can be argued. A budget and a model estate are not so different in that respect. Both are easier to govern once you can see them whole.
The data is for 2024-25. The chart will be out of date the moment the next outturn lands, which is the nature of the thing.