Keir Starmer’s decision to announce a major Defence Investment Plan while preparing to leave office exposed an awkward gap in the UK constitution. The question is not simply whether an outgoing Prime Minister has the legal power to make major decisions. Plainly, they do. The harder question is whether a Prime Minister whose political authority is visibly draining away should make decisions that bind, constrain, or politically pre-empt their successor.
The UK constitution is much better at recognising some moments of weakened authority than others. It has conventions for general elections, unclear election results, and lost confidence votes. It has far less to say about the increasingly familiar scenario of a Prime Minister leaving office mid-Parliament because their own party has decided, formally or informally, that the time has come for someone else.
The immediate controversy concerns defence spending. The Government announced a major Defence Investment Plan, with the Prime Minister presenting it in his speech announcing the plan as a response to a more dangerous international environment. The controversy lies in the timing and the fiscal consequences. Reuters has reported that Andy Burnham (expected to succeed Starmer later in July) may inherit a £4.7 billion funding gap. The Guardian has reported similar concerns, describing the plan as a £298 billion four-year package with nearly £5 billion still to be found.
Defence is, of course, not an ordinary policy field. National security decisions may be urgent, reactive and with long-term, far-reaching consequences. No serious constitutional argument can be made that a government approaching a leadership transition must stop governing. But the scale and timing of this decision sharpen a broader constitutional problem: when does normal government become legacy government?
What the rules currently say
The orthodox legal position is clear enough. A Prime Minister remains Prime Minister unless and until they resign. Where a Prime Minister resigns while their party retains a Commons majority, the usual practice is for the governing party to identify the successor, with the outgoing Prime Minister formally resigning only once that process is complete. The House of Commons Library summarises the modern position in these terms.
There is no halfway constitutional status between being Prime Minister and not being Prime Minister. The UK constitution does not recognise a semi-detached Prime Minister, an acting successor in waiting, or an incumbent whose legal powers are automatically reduced because their political authority is fading. Catherine Haddon puts the point crisply for the Institute for Government: the UK constitution has no concept of an acting Prime Minister. You are either Prime Minister or you are not.
That legal clarity sits alongside a more complicated conventional position. The UK already accepts that there are periods when ministers retain lawful authority but should exercise restraint. The caretaker convention is not a formal legal status. It is a constitutional expectation that, in certain circumstances, ministers should avoid major decisions of continuing or long-term consequence unless delay would damage the public interest. Anne Twomey’s comparative account of caretaker conventions is useful here because it connects these rules to the broader Westminster principle of responsible government.
The recognised triggers are narrow. The Cabinet Manual refers to restrictions on government activity during election periods, after a loss of confidence, and where there is uncertainty over who can command confidence after an election. The Institute for Government identifies the same basic situations. In each case, the government may continue in legal terms, but its political authority to take major, binding decisions is constrained.
Robert Hazell’s analysis of Boris Johnson’s final weeks in office remains especially useful. Hazell notes that the rules around a “lame duck” Prime Minister were fuzzy. The Johnson example raised a closely related problem: what, if anything, should limit a Prime Minister who remains legally in office after losing the authority of their own party?
Why this scenario does not fit
The Starmer-Burnham transition does not fall neatly within the existing caretaker framework. Labour still commands a Commons majority. There has been no general election, no hung Parliament, and no successful confidence vote against the government. The transition is internal to the governing party, not imposed by Parliament or the electorate.
That is why the strict constitutional answer is relatively straightforward. Starmer remains Prime Minister until he resigns. His ministers remain ministers. Cabinet government continues. The Defence Investment Plan is not constitutionally void simply because it was announced late in his premiership.
Yet that answer does not fully resolve the question of propriety. Mid-term leadership changes have become a regular feature of UK government. Whether it was Tony Blair handing over to Gordon Brown, David Cameron giving way to Theresa May, Theresa May giving way to Boris Johnson, Boris Johnson handing over to Liz Truss, or Truss giving way to Rishi Sunak, it seems almost commonplace now. As Rodney Brazier has observed in his work on choosing a Prime Minister, the rules governing how Prime Ministers enter and leave office depend on a mixture of law, convention, precedent and political practice.
These transitions were not all alike, though. Blair’s departure was a managed, if tense, transfer of power. May, Johnson and Truss left after their authority had been destroyed by their own parties. But the distinctions accentuate the problem rather than dissolving it. Whether the transition is orderly or hostile, gradual or sudden, the constitution gives very little guidance on the period between political departure and formal resignation.
The key point is therefore not that Starmer lacks power. It is that the existing caretaker rules are triggered by election periods, uncertainty over government formation, or loss of parliamentary confidence. They are not triggered by the political certainty that the Prime Minister is on the way out.
The gap between legality and legitimacy
This gap matters because Prime Ministers can still do a great deal in their final weeks. They may make appointments, approve contracts, announce spending commitments, make international undertakings or reshape departmental priorities. Some of these decisions may be reversible in practice, but others may be politically or financially difficult for a successor to undo.
Equally, we must be careful not to suggest that government should simply stop whenever a leadership contest begins. National security, in particular, may require continuity. On the issue of defence alone, in some circumstances, delay could be more irresponsible than action. But this does not resolve the constitutional tension. It simply shows why the existing rules are underdeveloped.
The civil service problem
The gap is not only political. It also has consequences for the civil service. During election periods, officials have a clearer framework for assessing whether proposed government action is urgent, unavoidable, or constitutionally sensitive. The Institute for Government notes that the enforcement of restrictions on government activity is primarily the responsibility of the Prime Minister, permanent secretaries and the Cabinet Secretary.
In an intra-party transition, the position is much less settled. A Cabinet Secretary faced with a departing Prime Minister seeking to announce a major policy has no clear textual rule to point to. The absence of a rule does not mean the decision is constitutionally unproblematic. It means the matter is left to judgment, advice and political restraint.
That is a fragile position. It places considerable weight on informal understandings inside government. It also leaves room for disagreement over whether a particular decision is necessary government business or an attempt by an outgoing Prime Minister to define the terms on which their successor takes office.
One possible alternative would be for the governing party to withdraw confidence in its own leader. But this is an imperfect failsafe. An internal party vote of no confidence may make a Prime Minister’s position politically untenable, but it does not itself remove them from office, trigger caretaker rules, or require a general election. A formal Commons vote of no confidence is different: if the Government loses the confidence of the House, the Prime Minister is expected either to resign or seek a dissolution. For a party with a working majority, however, using a Commons confidence vote to manage an internal leadership transition would be a blunt and risky tool. It could imperil the Government as a whole, rather than simply constrain the Prime Minister.
The caretaker blindspot
The controversy over defence spending therefore reveals something larger than a dispute over military budgets. It exposes a caretaker blindspot in the UK constitution. The current rules are built around confidence, elections and government formation. They are much less comfortable with the modern reality of party-managed changes of Prime Minister within a Parliament.
The formal rule remains necessary: a Prime Minister governs until they resign. The country cannot be left without responsible ministers, and the Crown must always have advisers who can act. But that formal rule does not answer every question of constitutional propriety.
The Starmer-Burnham transition shows the problem clearly. The issue is not whether an outgoing Prime Minister can govern. They can. The issue is whether the constitution gives sufficient guidance on how they should govern when everyone knows they are already in the exit corridor.
Darryn Nyatanga.
Dr Darryn Nyatanga is a constitutional law scholar whose research focuses on executive power, devolution, and constitutional change in the United Kingdom. He holds a PhD from the University of Liverpool.
The Constitution Society is committed to the promotion of informed debate and is politically impartial. Any views expressed in this article are the personal views of the author and not those of The Constitution Society.