Referenda have been a controversial topic in the UK for the past decade and some of the most heated debate has revolved around the issue of repeating a referendum. In what circumstances is it legitimate to push for a second referendum on a topic? How soon after the first result can it be held? And when is it seen as undermining democracy?
Despite the title, framing this as ‘yes’ versus ‘no’ issue is not particularly helpful – it should instead be conceptualised as a question of a vote for institutional change versus a vote for continuing the status quo. Generally, a vote for institutional change is regarded as a permanent endorsement, whereas a vote for the status quo may be seen as indicating ‘not now’, whether that means a vote only ‘once in a generation’ (as argued with Scottish Independence) or even sooner. There is thus a constitutional asymmetry between the results and the mandates that they produce, and politicians should be aware of this when seeking to use referenda to settle nagging political questions or for party political issues.
In this blog I will examine the UK’s attitude to repeat referenda and the factors that influence it, drawing from a selection of referenda on the EU, devolution to Scotland, Scottish Independence and the electoral system.
Votes on the EU:
It is worth quickly looking at the first UK referendum on Europe under the Wilson government on whether the UK was to remain within the (then) European Communities. The UK had joined in January 1973, without a direct public vote on the matter, and renegotiated EC membership in 1974. In 1975, voters were asked to vote ‘yes’ or ‘no’ on the question:
“Do you think the United Kingdom should stay in the European Community (the Common Market)?”
For this plebiscite, the ‘yes’ vote represented the status quo option and 67.2% of voters were in favour.
While not widely seen as a repeat referendum, the UK did vote again on membership of the EU in June 2016. There were, as far as I am aware, no questions raised at all about the legitimacy of holding another referendum on this topic 41 years (nearly to the day) after the 1975 vote. It is quite obviously not seen as objectionable, by politicians or by the public at large, to reassess the status quo once ‘enough’ time has passed. This is particularly the case if the status quo has continued to evolve – with the EU changing and expanding massively over the decades. In 2016, the question read as follows, with ‘remain’ being the status quo option:
“Should the United Kingdom remain a member of the European Union or leave the European Union?”
After a divisive campaign, the final result was 51.9% in favour of institutional change.
Almost immediately after the Brexit result, calls began for a second referendum to be held, with an online petition reaching over 4 million signatures by the start of July 2016 (though some signatures were found to be fake). Various arguments were advanced to support this, including citing the close result, low turnout, the open-ended nature of the question, and rampant misinformation.
This was extremely contentious, with Prime Minister Theresa May warning in very strong terms that such a vote would “break faith with the British people” and that “[t]o ask the question all over again would be a gross betrayal of our democracy – and a betrayal of trust.” This echoes the thoughts of many that a second vote would not only be wrong, but illegitimate. It would not be democratic to hold a repeat referendum so soon after the first.
Even now, 7 years later, there are calls for a second referendum, though the Conservative government have been quick to dismiss them. It cannot yet be conclusively seen if the vote for institutional change will be regarded as permanent, but it currently appears very difficult (both politically and practically) to reverse.
Votes on devolution to Scotland:
A referendum was held in Scotland in 1979 to decide whether there was sufficient support for the creation of a Scottish Executive and Scottish Assembly with limited legislative powers. The question ran as follows:
Do you want the Provisions of the Scotland Act 1978 to be put into effect?
For this vote, ‘yes’ represented the vote for institutional change. While the majority of the voters supported the proposal (51.6%), there was a threshold requirement included in the Scotland Act 1978 that at least 40% of the eligible electorate voted ‘yes’. The turnout of 64% meant that the ‘yes’ vote did not meet this threshold, and the proposed Assembly was scrapped.
Unsurprisingly, this was not felt to have resolved the issue. The Campaign for a Scottish Assembly, a cross-party organisation, was set-up on the first anniversary of the failed referendum and helped birth the Scottish Constitutional Convention (SCC) to create a blueprint for change. Eighteen years of Conservative government at Westminster only served to stoke demand for a new constitutional settlement.
Following the emphatic Labour victory in 1997, though, devolution to Scotland (and Wales) was put back on the table, and a referendum was promptly legislated for.
The principle question on devolution ran as follows:
Parliament has decided to consult people in Scotland on the Government’s proposals for a Scottish Parliament:
I agree there should be a Scottish Parliament
I do not agree there should be a Scottish Parliament
A second question was included on whether the new Parliament should have tax-varying powers, and both were passed by significant majorities (74.3% and 63.4% respectively). In both cases, ‘I agree’ was the vote for institutional change.
Of referenda in the UK, devolution to Scotland is a unique case. It is not a situation where ‘no’ in 1979 was taken to mean ‘not now’, but rather where a slim majority for ‘yes’ nonetheless resulted in a ‘not now’ by disqualification. An 18-year period between votes, in which support for nationalism grew, and different, more developed proposals were developed, also helped prevent this from being viewed as reopening the question too soon.
Alternative Vote referendum:
In May 2011, the Conservative Lib-Dem Coalition held a UK-wide referendum on the Parliamentary voting system. The question ran as follows:
At present, the UK uses the “first past the post” system to elect MPs to the House of Commons. Should the “alternative vote” system be used instead?
For this referendum, ‘yes’ was a vote for institutional change. The result was 67.9% for the status quo, on a particularly low turnout of only 42.2%. It was regarded as a comprehensive and decisive defeat by all the political parties involved. Director of the No campaign Matthew Elliott said at the time, “I personally believe that this result will settle the debate over changing our electoral system for the next generation.” I will return to the slightly nebulous idea of ‘once in a generation’ votes later.
The decisive ‘No’ vote continues to be cited as an endorsement of first past the post (FPTP) and a rejection of any type of electoral system change. The Conservative government response to a 2016–17 parliamentary petition for proportional representation read: “A referendum on changing the voting system was held in 2011 and the public voted overwhelmingly in favour of keeping the FPTP system.” Then, in May 2023, Mayoral elections and Police and Crime Commissioner (PCC) elections were changed from using the Supplementary Vote (SV) system to FPTP. Home Secretary Priti Patel claimed that the change ‘reflects that transferable voting systems were rejected by the British people in the 2011 nationwide referendum’.
This is a confused mandate. The referendum was not framed as being the final answer on any type of electoral system change, just on AV. As Alan Renwick notes, the ministers at the time seemed to understand this since they introduced the Supplementary Voting system for PCC elections in the months after that referendum.
At the same time, while there are still campaigns for electoral reform, the AV system has fallen out of favour and is now rarely mentioned. The AV referendum can be seen an interesting example of a ‘no’ vote extending beyond ‘not now’ even by those campaigning for it, perhaps due to the scale of the defeat and the specificity of the proposal. Unlike most of the constitutional referenda held in the UK, the AV referendum was not open-ended and, unlike all other referenda mentioned here, was not merely advisory – it bound the government to the result. It appears that referenda can provide permanent and strong ‘no’ answers to political questions, but only when the result is far from close. And in those situations, politicians may seize upon it to widen what that ‘no’ can apply to.
Scottish Independence Referendum:
The issue of repeat referenda comes up the most frequently regarding Scottish independence. In 2014, David Cameron conceded a referendum following the electoral success of the SNP. The referendum question was simply:
“Should Scotland be an independent country?”
Here, ‘yes’ was the vote for institutional change, and it was defeated 55.3% to 44.7%. Needless to say, this has not resolved the question of Scottish independence, and the relative closeness of the vote became a rallying point for Scottish nationalists.
Much debate has specifically surrounded the term “once-in-a-generation” – used by Alex Salmond and other senior SNP politicians ahead of the vote, and mentioned in Scotland’s 2013 white paper. Salmond explicitly referred to Scotland’s devolution votes when expanding upon what constituted a suitable amount of time between referenda, stating: “If you remember that previous constitutional referendum in Scotland — there was one in 1979 and then the next one was 1997. That’s what I mean by a political generation.” This discussion provides interesting insight into the central question of when it is appropriate and legitimate to re-ask a question and re-assess the status quo, with even Whitehall agreeing on the idea of a generation (though disagreeing on exactly how long that might be).
To look at Northern Ireland, the Northern Ireland Act 1998, states that any referendum on Irish reunification can be repeated after a minimum of seven years.
This is a crucial example of the constitutional asymmetry of sovereignty referenda. Here, ‘no’ was quite explicitly stated to mean ‘not now’. The outcome may have looked quite different if it was widely perceived that the opportunity was ‘now or never’, however it is doubtful that the Scottish voters also expected the ‘yes’ vote for institutional change to be subject to another referendum inside a generation. Even if it the vote had been the reverse (55% in favour of independence), independence was expected to be permanent.
Doubtless to say, referenda still sit uncomfortably within the UK constitution. There are many unresolved questions about when they should be called, and for what reason, and few politicians (particularly within the government) have attempted to form coherent answers to these, as often vagueness can be politically advantageous. This blog does not attempt to answer those questions for them, but rather explores what we can draw from past referenda.
It is interesting to contrast the UK experience to the many countries that have held multiple repeat referenda – including Ireland, France and Denmark on key EU treaties, and Switzerland on a whole range of topics. For these, voters are rejecting the policy on offer and are part of a negotiation process, giving feedback to politicians who then seek compromises before putting it again to the voters. What the Swiss model achieves, in particular – through public input, parliamentary groundwork and regular votes – is a situation where “winners and losers” doesn’t apply in the same manner, as Bruno Kaufmann argues. Or, if there are losers, they are “happy losers”, included in the decision-making process.
It is much, much more difficult to find international examples where there was a result for institutional change that was then subjected to a repeat referendum. This points to the asymmetry of mandates – the constitutional mandate for institutional change, even if actually minimal, is seen as permanent and it is politically very difficult to attempt to undo or challenge a democratically sanctioned change.
Practically, this makes sense – institutional change, especially of the size typically addressed in referenda (the AV vote aside) is a huge undertaking which will likely take years and many resources to thrash out. The risk of it being undone or altered several years later is a political headache most in power would rather avoid. Nat Le Roux labels this “the issue of asymmetric irreversibility”. However, with no constitutional provision to trigger another referendum, it means that there are limited opportunities for the public to change their mind and express it effectively.
This asymmetry is something that politicians must keep in mind when deciding to call a referendum – it is not easy to stop or dissuade change through a referendum, unless the victory for the status quo is very decisive. A narrow victory can provide a rallying point for those campaigning for change, and encourage a wider perception that a future referendum is on the cards. Therefore, as Nat Le Roux concludes, “at least in the UK context, a government should not call an advisory referendum unless it supports the change which is being proposed.” Referenda should not be considered as a reliable mechanism to provide a final answer on tricky political issues.
Kelly Shuttleworth is a PhD student looking at constitutional conventions across the UK, New Zealand and Canada at the University of Auckland. She is a contributing writer for the Constitution Society. Prior to this, she worked at various research organisations in the UK, including the Institute for Government and the Bennett Institute for Public Policy, focusing primarily on devolution issues.
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.