12 December 2018

If Parliament cannot agree a way forward, it may feel there is no alternative but to refer the final decision to the public, says Akash Paun. That is not the simple solution it appears.

The chaos at Westminster makes a second referendum on Brexit a serious possibility for the first time. If it does happen, the key to successfully designing the referendum process must be to ensure that the legitimacy of the outcome is respected by as many people as possible, in all parts of the country.

First is the referendum question itself. In 2016, as in most other previous referendums in the UK, the choice offered to voters was a simple one: in this case between Remain and Leave. However, as it stands, there are at least three possible ways out of the present political impasse – the Prime Minister’s deal, a no deal Brexit and Remain.

To exclude any of these options from the ballot paper risks undermining the legitimacy of the process. The question is whether all three options could be included on the ballot, to provide both legitimacy and a clear answer.

First-past-the-post would not work

This raises the question of what system should be used to determine the winner.

The simplest approach would be for voters simply to mark their preferred option of the three. The option with most votes would win, as in a general election held on a first-past-the-post basis.

The risk is that the successful option could win with the support of only 40% or so of the electorate – for instance because the pro-Brexit vote would be divided between the two Leave options. This outcome would be contested and could lead to an even more divided nation.

A second option is to use preferential voting

Voters rank the three options in order of preference. The least popular option would then be eliminated and the second preferences reallocated to the other two options. This should ensure that the winning option is either a first or second preference of over 50% of voters. However, this could leave a winning option that scored poorly on first preferences.

A different approach would have two questions

One way of addressing those issues would be to have a two-stage question. Question 1 would be a choice between Leave and Remain, as in 2016. Question 2 would be a choice between the two possible flavours of Brexit: Deal or No Deal.

If Remain won a majority on Question 1, then Brexit would be cancelled.

If Leave won a majority on Question 1, then the results of Question 2 would determine whether the Withdrawal Agreement were ratified (Deal) or whether the UK should walk away on a No Deal basis. All sides in the argument would have had the opportunity to put their option to voters. Another option would be to ask the questions the other way round: the best way to leave and then whether to leave on that basis. But that would have to be done in two separate polls, like the French Presidential election.

But there would still be other difficult issues to resolve

Even a complex two-stage question would run the risk that the four UK nations again vote in different ways.

Parliament would also need to decide if this time it wanted to make the referendum legally binding. The 2016 referendum was ‘advisory’ but a second referendum would have to be definitive – so there are arguments that the result should bind Parliament. That is relatively simple for either the Remain outcome, where the UK would simply revoke Article 50, or No Deal – though the UK may not have much time to prepare to walk away with no transition. 

But it is more problematic in the case of the Prime Minister’s deal: Parliament would have to pass the Withdrawal Agreement but the Political Declaration is, as many people pointed out, miles away from being a clear statement of the UK’s future relationship with the EU, which has to be negotiated after we have left. There would still be scope for many arguments on the shape of that relationship. And Parliament would still want to scrutinise the way in which the Government was choosing to put the Withdrawal Agreement into law, which as we have warned raises important constitutional issues.

The Electoral Commission would, in the normal way, assess the referendum question for intelligibility and bias. It would, in this case, also have to nominate at least three official lead campaign groups: for Remain, Leave/Deal and Leave/No Deal – though there might have to be more, to reflect the choice of second preferences.

A second referendum is only one way out of the present impasse, but with the Prime Minister’s deal facing rejection in the House of Commons, the prospect must be taken seriously. There is a risk that it is seen as an easy way out, and not enough thought is given to how to design it. Should the country be faced with another vote, Parliament and the Electoral Commission must ensure the vote takes place under rules that maximise the chances of the outcome being respected.


There are bigger problems than those identified here both with preferential voting and with a two-stage choice.

First, with preferential voting, it's a reasonable assumption that a large majority both of those favouring No Deal, and those favouring Remain, would see leaving on the current deal as their second preference. However, first preferences might split, say, No Deal 34%, Deal on offer 32%, Remain 34%. Result: a forced choice between extremes, with the option best placed to reunite the country excluded.

Second, with a two-stage question, voters favouring the deal on offer, but fearing No Deal as the worst outcome, might feel obliged to answer Remain to the first question to avoid the risk that No Deal would win on the second. Result, again, a polarised choice between extremes.

There are ways round this, I know, but they are complex to explain and easy to characterise as manipulation.

The only justification for a referendum now is that parliament can find no other solution. To make the result legally binding seems wrong; our entire political system is predicated on representative democracy and parliament must therefore make the final decision. If parliament cannot agree, it goes back to the people to seek further guidance, not to seek an instruction. Given that the one point of certainty currently is that parliament would never vote for “no deal”, it would be misleading to put that option on the ballot paper. This simplifies things massively as the choice would then be between remain and the PM's deal. Whilst the latter also currently has no majority in Parliament it is reasonable to suppose that if a majority of the public backed it, they would then vote it through.

The only way a referendum could happen is if Mrs May decides to let the Tory Party go hang. I am certain she would rather have no-deal than destroy her Party - and nothing can stop her in this regard, no-deal being the default legal position.

Hi Akash, if a second referendum were to happen, how could one ensure that the public had the opportunity to actually assess the role played by the US and EU in running public and social services ? I wondered if increased transparency, say, by publishing the OJEU lists for Goods, Works and Services with a note saying which trade agreement applied would work? For example, the Services Directive 2006/123/EC says that Health and Welfare services are exempt from its scope yet the directive contains backdoors through the words 'household support services' and it is largely US backed equity trusts that have made gains. Public Procurement is also covered by the broader scope of the Services Directive and impacts directly on devolution, eg the Scots are far more diligent and open about how procurement can (can not) work in service area's, say, for the elderly and their constitutional SO's are fairly well constructed to guide delegated decision making rignts in far mopre rigorous way than English SO's. Give me call if you'd like to unpack this further, alecfraher@googlemail.com, Cheers, Alec

Since Theresa May currently is opposed to a second referendum, how exactly do you think it is going to come about under the present Government, Akash?

Unlike David Cameron's 2015 commitment to the first EU Referendum, the May Government of 2017 came to power with the reverse commitment, i.e. to implement Brexit. Unless there is a change of Government and an election context where one of more parties campaign with an explicit manifest commitment to hold EUref2, and the winner who comes to power, either by outright win or coalition. has made that commitment, there won't be a second EU Referendum unless the government falls - or changes its mind.

How do you see the scenario unfolding?

Peter Curran

Hi Peter. It's hard to predict a specific path from here to there. There are still many uncertainties.

But if the PM's deal is rejected by the Commons in January, and No Deal Brexit looms, then the PM (or at least some of her supporters) may start to float the idea a second Ref as the only way to break the deadlock. Of course, if another referendum (with Remain on the ballot) starts to look at all likely, then some Brexiteers (the ERG etc) would probably rally around the PM's Deal despite their serious misgivings, to avoid the risk that Brexit is reversed entirely.

Another election is also a possibility. As it stands the Government would win a no confidence vote, but this is in the hands of the DUP...

Akash Paun
Institute for Government

The legitimacy of any referendum that included remain as an option would be questioned by those who voted leave last time whatever the procedures. The legitimacy of our democracy would be undermined too: it would mean that no significant change could be achieved by the ballot.

Hi. People would undoubtedly make this case. My point in the blog is simply that if Parliament cannot produce a majority for any credible version of Brexit, then there may be no alternative but to refer this back to the public, with all options on offer. It would be messy, that is for sure. This is about finding the least-worst path out of the impasse. If the PM manages to get her deal through the Commons in January after all, then that will presumably settle the matter.

As Parliament looks almost certain to reject Mrs May's deal, shouldn't the referendum questions simply be "No deal Brexit" or "Remain". The legitimacy of the PM's deal has already been undermined by Parliamentary opposition. The exception might be if both the ERG and the various other Leave and Remain factions come together and agree on how the PM's deal should be modified, but at the moment this seams very unlikely. (Especially as some seem to think "no deal" is the best of all possible options.)
To my mind, the impossibility of getting an agreed modification to the PM's deal is that the Leave side have not produced a rational case for how trade might be arranged with the rest of the world, once the UK has left the EU. Most of the other countries are members of their own trade bodies that look very like the EU and which will become more like the EU as time goes on. Where is the space for an independent country to trade? Writers for ASEAN and MERCUSOR have already described the UK as "a small and isolated country from whom concessions may be won"

Seems logical and thoughtful. But how would any referendum be seen by the public? And what if the public favoured 'no deal'? Isn't there a more fundamental point that the 2016 Referendum was conceived solely to bring a measure of unity to the political party of government, which it failed to do (in fact the opposite)? Leave won by only 38 percent of the total electorate?

Isn't that a major issue too? And doesn't this process raise major questions for British parliamentary democracy?

Let's get the first referendum completed first, before we start a second! A second referendum can
only ask. DEAL OR NO DEAL?

Interesting but there is no mention of the importance of a threshold. In the first referendum only 38% of those eligible to vote actually voted leave (17.5m), that leaves 62% who did not vote or voted remain. If we want to truly identify the will of the majority of the people then we need 22m people to vote for it. Clearly to achieve a threshold of 50% would be difficult as it would require a very high turnout. This is the next issue, do we set a minimum turnout to validate the result? We also need to learn lessons from the first referendum with regard to control of campaign spending, and the independent verification of information to eradicate myths and misinformation. I suppose this underlines why referendums are 'blunt instruments' and , in my view should be advisory and treated as such. It's a pity the media has not highlighted the shortcomings of the first referendum , maybe it's not too late?

There seems to be a big elephant in the room that campaigners for a second referendum refuse to detail plans for. Namely what happens in the very likely event that a second referendum votes remain, by, say, a 5% margin. Do they really expect this to be deemed any more valid than the equally small opposition win previously. So, shall we have "best of three" ! - or, prior to a second brexit referendum we have to have a pre referendum referendum to agree on what % majority is an acceptable majority. This may sound flippant but can you even begin to imagine the chaos resulting from a remain win but smaller than the previous leave win. What is most worrying is that there are supposedly sane politicians recommending a second referendum. An alternative vote is not applicable, but a second , p.m deal / no deal vote following from the leave vote is the only conceivable , or rathet, workable, possibility.

In the preferential vote system, after rejecting the least popular result, A super majority should be required, and if that is not achieved, the fall back position should be taken, which must be 'revoke'.
If though the super majority choice is for an option which has already been rejected by parliament, then a general election would be required, to reverse that rejection.