Labour leadership candidate Owen Smith yesterday became the highest profile politician to date to endorse a second referendum on Brexit. But how feasible is this? Alan Renwick suggests that a referendum of the type Smith proposes, on whether or not to accept the terms of Brexit agreed with other EU members, is possible. However, much will depend on how public opinion evolves over the coming years. It is far too early to say whether opinion is likely to shift away from Brexit or not.

In launching his bid for the Labour leadership yesterday, Owen Smith said there should be a second referendum on Brexit once the terms of the deal on future relations between the UK and the EU have been negotiated. In doing so, he became the most high-profile politician to endorse a response to last month’s vote that is attractive to many of those who would like us to remain in the EU. But is a second referendum actually feasible?

There is no doubt that it is possible: parliament can legislate for a referendum on any topic any time it wants. But whether such a vote could deliver the outcome that its advocates intend requires careful consideration. Four key questions need to be answered.

What sort of referendum are we talking about?

To begin with, we need to ask what sort of second referendum we have in mind. Three sorts have been suggested in the course of recent discussions of Brexit:

1. The first is simply a rerun of the referendum that we have already had. Over four million people have signed a petition saying that – because the result of the referendum was tight and, given turnout, only 37 per cent of those eligible to vote backed Brexit – a second vote should be held before confirming the decision. It is clear anecdotally that many of those taking this view are Remain supporters who are angry that Leave won last month on the basis of what they see as a deeply mendacious campaign. They hope that, now the stakes are somewhat clearer, a second vote would yield a different outcome.
2. The second option is that last month’s result is taken as showing general dissatisfaction with our current EU membership rather than a specific desire to leave the EU altogether. Rather than triggering the withdrawal process, the government could seek a deeper renegotiation of our membership terms, then go to the country arguing for continued EU membership on those revised terms. This approach was apparently advocated by Boris Johnson last year, and he seemed to toy with it again after announcing in February that he would campaign for Leave.
3. The third option is that we go ahead with triggering Article 50 but hold a second referendum once the negotiations have been completed, on whether to accept the deal that has been struck. This is the sort of referendum that is now advocated by Owen Smith.

It is clear that the second of these options is off the table. Indeed, as I wrote in February, it was never a serious runner. The country has voted for Brexit. No politician – particularly no politician who leads the contemporary Conservative Party – could take the referendum result to mean anything else. Indeed, merely by suggesting that the vote might not have been motivated in large part by concerns about immigration, Boris Johnson appears to have fried his chances of winning the Conservative leadership.

The first option has much to recommend it in principle. There are good reasons for saying that such a momentous change as departure from the EU should not be made on the basis of a slender majority measured at just one moment in time.  A more common procedure for major constitutional change in democratic countries is that the legislature adopts the change and a referendum then ratifies it. This means that opinion is effectively tested three times: first, the public votes in a government that advocates such change; then the legislature debates and adopts the change; finally, the public get to consider the matter specifically and decide whether they really want to go ahead.  That is the process that led, for example, to the Scottish independence referendum of 2014, building on the SNP’s majority victory in the Scottish Parliament elections of 2011. Here, by contrast, the current government was not elected on a manifesto of leaving the EU, and parliament has never voted for leaving the EU: opinion has been tested only once.

Nevertheless, while the case for saying we need to tighten the procedures around referendums to make them more robust is overwhelming, we are where we are. The rules for this referendum were defined in law last year and cannot now be changed. The future role and conduct of referendums ought to be investigated and reformed – as the Constitution Unit strongly urges – but cannot be retrofitted after the fact.

The case for a referendum of the third type – the type now proposed by Owen Smith – is, at least in principle, much stronger. Referendums ought to present voters with a choice between clear options.  The Venice Commission’s Code of Good Practice on Referendums says (at paragraph 3.1c), ‘The question put to the vote must be clear’ and ‘electors must be informed of the effects of the referendum’.  But last month’s referendum could not present voters with a clear choice, because what Brexit means could not, prior to the negotiations, be known. The Constitution Unit’s Robert Hazell has long argued that Scottish independence ought not to happen without a process involving two referendums – the first on whether to start negotiations, the second on whether to accept the agreed terms. The same might reasonably be said to apply to Brexit.

Still, there are three further questions that need still to be answered.

What would the question be?

Let us suppose that the second referendum is, as Owen Smith has suggested, a vote on whether to accept the terms of the deal that is negotiated between the UK and the EU over the next few years.  We might think that voters can just be asked a simple yes/no question on whether they accept the deal or not.

But that won’t do, because rejecting the deal on offer could mean completely different things to different people. In fact, rejecting the deal could mean any of three things:

1. that we should not go ahead with Brexit after all, but rather stay within the EU;
2. that we should reject the Brexit terms on offer and go back for a better deal;
3. that we should give up on trying to negotiate the terms of Brexit and leave without any kind of deal.

It is the first of these options that Owen Smith is proposing. In essence, we would be saying to voters, ‘Now you can see what Brexit will actually look like, do you still want it or not?’ It is possible also to imagine circumstances in which a referendum might be held on the second option: in order to strengthen its hand with the EU, a government might ask voters to endorse a demand for a better deal. The third option, by contrast, is distinctly unattractive: there is general agreement that a ‘disorderly departure’ from the EU would be disastrous: it would leave us facing tariffs on our exports to Europe, imperil the position of UK nationals living in the EU and EU nationals based in the UK, and damage our reputation around the world.

Does Article 50 permit such a referendum?

But whether any of these referendums are feasible depends crucially on how we interpret Article 50 of the Lisbon Treaty – the article, as is now well known, that sets out the procedure through which a country can leave the EU. Article 50 is both vague and untested, which is leading to much debate over its meaning. Most attention has focused so far (as a glance through the last few weeks of posts on the UK Constitutional Law Association’s blog will attest) on whether the government can trigger Article 50 without parliament’s consent. Parliament probably will endorse the trigger, so, before too long, this issue is likely to fade away. The next question is whether a country, having declared its intention to leave the EU under the terms of Article 50, can withdraw that declaration.

This question is crucial for whether a referendum as proposed by Owen Smith can happen. Article 50 says that a country that triggers the exit process does automatically leave after two years, unless an extension is unanimously agreed by all the member states. It says nothing about whether a declaration of intent to leave can be withdrawn.

If we cannot withdraw that declaration, a vote in a second referendum to remain after all will have no immediate effect: we will have to leave the EU and then apply to rejoin, probably under less favourable terms than we have at present.

So is Article 50 in fact so constraining? The answer is, probably not. But just how constraining it is remains a matter of interpretation. There appear in fact to be three ways in which an Article 50 declaration might not be irrevocable:

1. First, the member states could simply allow indefinite extensions to the two-year negotiating period – a possibility that was raised by Professor Steve Peers in a prescient post in 2014. This might get round any tricky legalities. But it would clearly leave the UK at the mercy of every other member state: extensions require unanimous agreement; some states could readily use this fact to extract concessions from the UK on any number of issues. So this route should be seen as a last resort.
2. Second, the general view among lawyers is that, if all the member states agree, an Article 50 declaration can be withdrawn: as Professor Mark Elliott puts it, ‘It is plain that it [the Article 50 process] can be aborted by agreement’. This provides for a stable and permanent solution. But it again requires unanimity. So if this is how Article 50 is interpreted and the UK votes in a second referendum to stay in the EU after all, we might have to make concessions before we are allowed to remain.
3. Third, it might be – but this is much less certain – that there is an implied right in Article 50 for a state to withdraw an Article 50 trigger unilaterally. It is to this crucial question that our constitutional lawyers could usefully now turn their concerted attention. In a very helpful post last week, Charles Streeten summarized the key contributions on this point so far and set out his view, that a unilateral right to rescind does exist. The Lords EU Select Committee, drawing on evidence from two senior lawyers, reached the same conclusion in May. But uncertainty remains; early resolution would be desirable.

In short, if the UK votes to remain in the EU after all in a second referendum held after Article 50 has been triggered, we will be able to stay. But whether we will have to accept concessions in order to get this remains unclear.

Is such a referendum politically feasible?

I have suggested above that there are two strong reasons in constitutional and democratic theory for suggesting that a second referendum – on whether to remain in the EU or leave on the terms by then negotiated – would be not only possible, but also desirable. First, leaving the EU is a momentous decision with huge constitution, political, economic, social, and cultural ramifications.  Such a decision ought to be taken only after careful and prolonged consideration.  Second, voters last month could not know what Brexit would mean, because the terms of withdrawal had not been defined. To say that voters should not be allowed to change their minds when they see what the withdrawal terms actually are would be deeply undemocratic.

Nevertheless, the majority of voters supported Brexit last month.  Many of them were clearly people who feel they have been ignored by the political elite for decades and who are glad finally to have been heard in some way. Chris Hanretty estimates that 421 of the 574 Westminster constituencies in England and Wales voted to leave. In these circumstances, is it really plausible that MPs might risk incurring the wrath of their voters and further enraging those parts of society that have long felt marginalised by seeking a second referendum that might reverse the result of the first?

The answer, simply, is that it depends on how public opinion evolves over the coming years. If public opinion remains broadly as it is, then a second referendum looks unlikely. If, by contrast, opinion shifts clearly and lastingly away from Brexit, then a second referendum looks distinctly possible. That is clearly much more likely if an election comes before Brexit and a government committed to holding a second referendum enters office.

It is much too early to say whether that is likely. John Curtice finds that slightly more Leave than Remain voters now regret their choice, and that there is also some remorse among those who did not vote. But what matters is what happens from now on. If the economy turns down and the negotiations fail to yield the benefits that Leave campaigners foretold, pressure to reverse last month’s vote will likely rise. Conversely, if favourable trade deals are rapidly signed and the short-term economic costs prove milder than many forecast, support for Brexit could consolidate. What will look politically tenable in two, three, or four years’ time is today impossible to predict.

About the author

Dr Alan Renwick is the Deputy Director of the Constitution Unit.