Boris Johnson, the UK’s new Prime Minister, says that the UK will leave the EU on 31 October “do or die”. With negotiations between the UK and the EU apparently at an impasse, the likelihood of a no-deal Brexit now appears to be somewhat greater than the “million-to-one against” chance to which Johnson referred in late June. Since there is a clear majority in Parliament against a no-deal Brexit, the question arises whether it can prevent such an outcome. The further question arises — given recent pronouncements by Johnson’s senior adviser, Dominic Cummings — whether the Government could obstruct Parliament if it were to seek to stop the UK from exiting the EU without a withdrawal agreement. Against this background, there are doubtless many scenarios that may need to be contemplated over the coming weeks. In this post, I consider three possible sets of circumstances and examine the legal implications of each.

Scenario 1: No confidence

It is very likely that the Opposition will move a motion of no confidence in the Government when Parliament returns in early September. If the House of Commons were to pass a motion to the effect that it has no confidence in Her Majesty’s Government, two critical constitutional provisions would thereby be engaged. The first is a cardinal constitutional convention upon which the whole system of UK Government hangs. The second is the Fixed-term Parliaments Act 2011 (FTPA). The relationship between these two constitutional provisions has, in recent days, been the source of considerable confusion.

The convention in question is that the UK Government’s authority to govern stands and falls by its capacity to command the confidence of the House of Commons. The position is set out clearly in paragraph 2.7 of the Cabinet Manual. When the Government loses a vote of confidence, convention requires that the Prime Minister should resign at an appropriate time. That does not necessarily mean that the Prime Minister should resign immediately. As the Cabinet Manual notes, circumstances may supply occasion for the Prime Minister to exercise judgement, for example by tendering his resignation only once he is in a position to advise the Queen as to whom she should appoint as his successor. There is thus no invariable constitutional requirement that Johnson should resign as soon as his Government loses a vote of confidence. However, the notion that a Prime Minister can ignore a vote of no confidence and carry on regardless is wholly without constitutional foundation. Losing a vote of confidence is a constitutional game-changer for any Government.

At this point, it is necessary to bring the FTPA into play. The effect of section 2(3) of that Act is that a motion of no confidence (in the appropriate form) triggers a 14-day period during which any of a number of things might happen. First, nothing might happen, in which case Parliament’s expression of no confidence stands and a general election must take place. Second, the incumbent Government might persuade sufficient MPs to change their mind and support it; if, in such circumstances, the Government wins a second (or further) confidence vote within the 14-day period, the Government remains in office and no early election occurs. Third, an alternative Government may be formed — a possibility that is considered in the next section of this post.

For the time being, however, a crucial question arises. What if the Government loses a vote of confidence and fails to recover the Commons’ confidence within 14 days of the first vote? The Act is clear that in such circumstances an early general election must be called. But section 2(7) provides that the timing of such an election is to be determined by the Queen “on the recommendation of the Prime Minister”. Dominic Cummings has reportedly said that the Government would simply advise the Queen that any election should take place after 31 October, thus ensuring that an early election would not prevent a no-deal Brexit. If the Government were to adopt this view, could anything be done in response? Two possibilities are worth considering.

First, Parliament could, in theory, legislate, whether by amending the FTPA as necessary or by simply enacting a one section statute instructing the Prime Minister to advise the Queen that an election be held by a given date (prior to 31 October). But while this would be legally possible, it would be extremely difficult in political terms. It would not, however, be politically impossible: Parliament has already shown itself capable of seizing control of parliamentary business in order to enact the Cooper-Letwin Bill, which sought to avert a no-deal Brexit in March 2019 by legally requiring the Prime Minister to seek an extension.

Second, it is entirely conceivable that the Prime Minister’s advice to the Queen would be challenged by way of judicial review. However, it is far from clear that any such challenge would succeed. It is certainly the case that in advising the Queen about the timing of an early election the Prime Minister is discharging a statutory duty that is impliedly imposed on him by section 2(7) of the FTPA. Within that duty is discretion that enables the Prime Minister to exercise judgement when advising the Queen, but all legal duties are legally limited, meaning that they are open to judicial review if the boundaries of the power are exceeded. However, having said all of this, the question of when an election should be held is par excellence a question of policy, thus rendering it unlikely that a court would regard the question as a justiciable one. The upshot is that a vote of no confidence could well end up triggering a general election that takes place after 31 October, meaning that a vote of no confidence on its own provides no guarantee against a no-deal Brexit.

Scenario 2: Alternative Government emerges

But what if the Johnson Government were to lose a vote of confidence and if, within the 14-day window established by the FTPA, another viable Government were to emerge? What if, for instance, MPs from a variety of parties (including some Conservative MPs) were to be in agreement that there should be a Government of national unity — a cross-party Government — led by a given senior MP? Would Johnson have to make way for such a new Government, or (as has been suggested) could he cling on until the end of the 14-day period and then call a general election? The answer to that question, in my view, is that it would be neither constitutional nor lawful for Johnson to cling on in such circumstances. I will deal each of those points in turn.

First, we have already seen that, as a matter of convention, the Government’s authority to govern hinges upon its capacity to command the confidence of the House of Commons. In turn, a Prime Minister can legitimately occupy that office only by virtue of leading a Government that commands the Commons’ confidence. We have also seen that there is, as a result, a constitutional duty upon the Prime Minister to resign should his Government lose the Commons’ confidence, albeit that that duty may not be triggered immediately. It is, however, indisputable that the duty crystallises once the Prime Minister is in a position to advise the Queen that someone else is in a position to form a Government that has the confidence of the House of Commons. It follows that if, within the 14-day period, it had become clear that a cross-party Government led by a given MP could command the confidence of the House, it would be incumbent upon Johnson to tender his resignation to the Queen and to advise her to appoint the relevant MP as the new Prime Minister.

If Johnson were to refuse to do this, there is no constitutional reason why the Queen should not dismiss Johnson and invite the relevant person to form a Government. Indeed, it is possible to go further by arguing that it would be the Queen’s constitutional duty to do so. Although the Queen normally acts only on the advice of her Ministers, the convention that she does so applies only in the absence of a more specific convention that requires the Queen to act otherwise. It is for this reason that the Queen would have to ignore ministerial advice not to grant royal assent to a Bill passed by both Houses, since a more specific convention — that the Queen grants assent to such Bills — would take precedence. Similarly, the convention that the Queen acts on an outgoing Prime Minister’s advice as to the appointment of a new Prime Minister cannot be sacrosanct. If, for instance, a Prime Minister who lost an election advised the Queen to reappoint him, she would have no option but to disregard such advice. Equally, if a Prime Minister who had lost a vote of confidence refused to resign and/or refused to advise the Queen to appoint a person who clearly could command the confidence of the House, the Queen would have no option but to dismiss the incumbent Prime Minister and ask the appropriately placed person to form a new administration. In doing so, the Queen would not be playing politics: she would be acting as the ultimate guarantor of fundamental constitutional principle in a manner fully consistent with the nature of constitutional monarchy in the UK.

Second, if the Prime Minister “squatted” in Downing Street following a vote of no confidence and following the emergence of an alternative viable government, with a view to running down the 14-day clock and triggering an election, the possibility of a legal challenge would arise (a point to which Tom Hickman has drawn attention). The argument would be a straightforward one: namely, that by acting in such a way, the Prime Minister would be frustrating one of the purposes of the FTPA and thus acting unlawfully. The Explanatory Notes to the FTPA state that the purpose of section 2(3), which establishes the 14-day window, “is to provide an opportunity for an alternative Government to be formed without an election”. Of course, Explanatory Notes are not legally binding, but in this instance they merely serve to point out that which is in any event clearly implicit in the statute itself. It is self-evident that the relevant provision in the legislation would be rendered a dead-letter if Prime Ministers were able to circumvent it by refusing the budge. As a result, it is clearly arguable that the FTPA imposes upon the Prime Minister a legal obligation to resign so as to make way for a new administration if, within the 14-day period, such an administration emerges. It also important to note that whereas a legal challenge to Prime Ministerial advice regarding the timing of an election would face an steep uphill struggle in the face of arguments about justiciability, the same is not true here: a refusal to resign in relevant circumstances would amount to a straightforward breach of a clear (implied) provision in the legislation, and would not engage any non-justiciable questions.

Scenario 3: Legislation

We have seen so far that by passing a motion of no confidence in the Government, Parliament may — subject, of course, to political considerations — be able to prevent a no-deal Brexit by producing a change of Government, either directly or following an early general election. The other main option open to Parliament is enacting legislation. As noted above, the political difficulties that would be involved in legislating in the face of Government opposition should not be underestimated, albeit that the Cooper-Letwin Bill shows that it can be done.

It is important to be clear about what the content of any legislation would need to be if it were to avert (either temporarily or permanently) the UK’s exit from the EU without a withdrawal agreement. There are certain steps that simply would not work, including some that were recently suggested by Vernon Bogdanor in the Guardian. One of Bogdanor’s suggestions was that “the Commons could legislate for a referendum before Brexit”. But this would not prevent the UK from leaving the EU by operation of EU law on 31 October unless, in addition, the UK Government requested and the EU agreed to an extension of the Article 50 period in order to create the time needed for a referendum to occur. Bogdanor also suggests that Parliament “could legislate to repeal the European Union (Notification of Withdrawal) Act and reinstate the European Communities Act, in which case the UK would remain in the EU”. But that is incorrect. The enactment of the Notification of Withdrawal Act was necessary in order to empower the Government to trigger Article 50, but repealing that legislation would not revoke Article 50: it would remain the case that the UK had lawfully initiated the withdrawal process, and under EU law the repeal of the domestic legislation under which the UK had begun the process would make not a blind bit of difference. The same goes for Bogdanor’s point concerning the European Communities Act 1972. Repealing the Notification of Withdrawal Act would not, in any event, “reinstate” the 1972 Act because the former did not repeal the latter. But even putting that point to one side, whether the 1972 Act is or is not in force has no bearing whatever on whether the UK leaves the EU on 31 October. Exit, deal or no deal, is the default position in EU law, and “reinstating” (or not repealing) the 1972 Act makes no difference at all to that.

What, then, would work? There are only really three possibilities. First, Parliament could legislate to revoke Article 50. Of course, politically many MPs would balk at such a prospect, but such legislation would be the best means by which Parliament could absolutely guarantee against a no-deal Brexit. Such legislation could be rendered politically less unpalatable from a Leave perspective by, for instance, recording within the preamble an intention that the legislation is intended to be a precursor to a confirmatory referendum. (I recognise that the CJEU’s Wightman judgment requires revocation to be unconditional, but it not clear to me that an aspirational statement regarding a possible future referendum in the preamble to a statute would breach that requirement of unconditionality.) Legislation along these lines, if enacted, would provide a cast-iron guarantee against a no-deal Brexit because it would, presumably, impose an unqualified and immediate duty upon the Prime Minister to revoke the UK’s notification under Article 50.

Second, Parliament could legislate along the lines of the Cooper-Letwin Bill by requiring the Prime Minister to seek an extension of the Article 50 period. However, this would not provide any guarantees, since it would be for the European Council to decide whether to accede to such a request. Legislation that went no further than requiring the Government to seek an extension would thus reduce the likelihood of, but would not rule out, a no-deal Brexit. That leads on to a third possibility: namely, a hybrid of the first two. Such legislation might require, in the first instance, the Prime Minister to seek an extension. However, it might go on to provide that if no extension had been granted by a given date (e.g. 30 October), the Prime Minister would be legally obliged immediately to revoke the UK’s notification under Article 50. This sort of approach, with extension as the preferred option and revocation as a last-resort failsafe, would presumably be politically less unpalatable to some MPs than legislation that required revocation without more.  


For reasons that I have explained elsewhere, the Supreme Court’s Miller judgment handed Parliament a golden opportunity to take control of the Brexit process. That opportunity was immediately squandered by parliamentarians who — for fear of being castigated as “enemies of the people” — fell over themselves to write the Government a blank cheque when they enacted the Notification of Withdrawal Act. That Act handed the Government complete discretion over when Article 50 should be triggered and provided Parliament with absolutely no instruments of control over the ensuing process. Ever since, Parliament has been playing catch-up. As the autumn unfolds, and as the cliff-edge beckons, we will see whether the majority of parliamentarians who are opposed to a no-deal Brexit can recover the situation. As I have sought to show in this post, Parliament does have options open to it. But those options are limited in legal terms — and decidedly so in political terms.