Britain's Defence Budget Has a Chain of Command. Voters Aren't On It.

In July 2024, Rachel Reeves published a Public Spending Audit and found £22 billion of unfunded resource pressures sitting across government, with no minister having voted for any of it in the ordinary sense. The Ministry of Defence's response was to be told, along with everyone else, to cut consultancy and administration spending.

No submarine was cancelled. No general was consulted. A spreadsheet in the Treasury moved a number, and the department that runs the armed forces of a nuclear power found out afterward.

Seven months later the Prime Minister announced Britain would spend 2.5% of GDP on defence by 2027, with an "ambition" (his word, carrying no legal weight and no date attached) to reach 3% sometime in the next parliament, "subject to economic and fiscal conditions." That qualifying clause is doing more work than the pledge itself. It means the number is real until the Treasury decides it isn't, which is the position the Treasury occupies on almost everything, and which nobody in this country elected it to occupy.

The case for all this complexity is not stupid, and it deserves saying plainly before it gets dismantled. Defence procurement runs on decade-long timelines that outlast governments; a minister facing re-election in four years has every incentive to raid the equipment budget for a headline now and let the next lot deal with the gap. Boards, committees and NATO targets exist partly to stop exactly that, to bind future governments to commitments a five-year electoral cycle would otherwise shred. It's a reasonable argument. It is also, on inspection, an argument for insulating spending from politics that somehow always insulates it in one direction only: toward less, not more.

6. Public opinion

Polling on defence spending moves with headlines. A Russian incursion, a NATO summit, a warship photographed somewhere ominous, and the numbers tick up; government spending does not move with polling at all. The 2.5% figure was set by a Prime Minister's speech in February 2025, not by a white paper responding to public consultation, and the process by which "ambition" becomes "commitment" happens entirely inside government, later, if fiscal conditions allow. Public opinion is the input everyone cites and nobody actually uses.

5. Parliament

The Defence Select Committee can question ministers and publish reports that get politely filed. Parliament's actual statutory role is to approve the budget law, once a year, after the numbers are already fixed. It can, in principle, transfer funds between lines or assess whether an allocation covers a stated project, but it does not build the budget it's approving. That's the constitutional design working as intended, and the intention was never for MPs to run the numbers.

4. Defence contractors

Procurement decisions running into billions get shaped years before a minister signs anything, in relationships between the MoD and the handful of firms capable of building a frigate or a fighter jet, where switching supplier is barely a live option and the commercial leverage sits almost entirely with the seller. A minister choosing between two credible bidders is exercising choice. A minister choosing the only bidder who can deliver is exercising a formality.

3. International commitments

NATO's 2.5%-of-GDP target functions as an external anchor for domestic budgeting, which is precisely its appeal to governments that want cover for spending increases they couldn't sell on the merits alone, and precisely the problem for anyone who thinks spending priorities should originate at home. Britain's target was announced before the Strategic Defence Review that was meant to justify it had reported. The number came first. The strategy is still being assembled to fit around it.

2. The Ministry of Defence

The department that actually runs the armed forces controls day-to-day operations, procurement and contingency planning, governed by a Defence Board stacked with the Secretary of State, service chiefs and three non-executive members nobody votes for, feeding into an Investment Approvals Committee that decides what gets built. This looks like where the power lives. It isn't, quite, because every one of those bodies is deciding how to spend a total that was set somewhere else, by someone else, before the Board ever meets.

1. The Treasury

Which brings the chain back to where it always ends. The Treasury does not merely fund defence; it decides, department by department, what counts as an "unfunded pressure" worth cutting and what counts as protected. In 2010 it went further, redrawing the entire accounting framework for how defence spending is measured and reported, the "Clear Line of Sight" project, on the Treasury's own terms, for the Treasury's own convenience.

A department that can redefine what the number means before it decides what the number is has a form of power no elected defence minister has ever held. Reeves's £22 billion audit is the same instinct in daylight: defence treated as a cost centre competing with everything else for headroom that one building, staffed by officials nobody stands for election against, controls absolutely.

Line those six up and the pattern isn't complexity distributing power sensibly across expert hands. It is power concentrating in the one body structurally incapable of losing an argument, because it holds the pen on every other body's budget. The Defence Board can want what it likes. The Select Committee can ask what it likes. NATO can set whatever target flatters a Prime Minister's conference speech. None of it survives contact with a Treasury official deciding, in a room no voter has ever entered, that this year the number is smaller than promised.

Ask a nurse whether she voted for the discount rate on her hospital's capital budget and she'll look at you as if you've named a disease. Ask her whether she voted for the 2.5% defence target, or the £22 billion adjustment that shrank it before it started, and you'll get the same look, for the same reason: she didn't, and neither did you, and the ballot paper that might have let her try was never printed.