Is there a tension between Parliamentary Democracy and referendums?


Nat Le Roux on parliamentary sovereigntyText of Nat le Roux’s presentation on Parliamentary Sovereignty at the McDougall Trust workshop, 9th November 2016:

The governing norm of our constitution is the principle of Parliamentary Sovereignty, which means that Parliament is the supreme legislative authority in the UK and can enact or repeal any law.  The same legislative process applies to nearly all laws because we have no separate category of constitutional law which is entrenched with special amendment procedures.

So where do referendums fit in that model?  If parliament is sovereign, referendums (unless they are treated as purely consultative exercises) would appear to be at best an unnecessary confusion.  Yet referendums are becoming more frequent and, on the whole, the legitimacy of these referendums is generally accepted by both politicians and the public. So, Parliamentary Sovereignty notwithstanding, there seems be an emerging constitutional norm that some legislative changes require direct popular endorsement.

At the same time, there are no coherent constitutional rules which tell us when a referendum is required, or indeed when it might be inappropriate. This is in contrast to many other democracies, where the circumstances in which a referendum should be held are more clearly defined, often in written constitutions. In principle, a British government can call (or not call) a referendum on any subject at any time.  To take a recent example: in 2011 the Coalition government held a referendum on reform of the voting system, fulfilling a political bargain between the coalition partners, but it could equally well have changed the voting system by ordinary legislation without a referendum.

Because of this flexibility, the history of referendums in the UK is characterised by political opportunism. They have commonly been used, or promised, to resolve intractable internal disputes within political parties. Labour’s promise of a referendum on the voting system under John Smith in the 1990’s is an example: the shadow cabinet was split over PR and a promise of a future referendum (which Tony Blair incidentally never fulfilled) was a convenient way out of the impasse.

An Act of Parliament is required to hold a referendum, but governments have in practice never found it difficult to secure a majority for a referendum and many such bills, including the bill authorising the 2016 EU referendum, have enjoyed cross-party support. It seems there is no political advantage in opposing the calling of a referendum.

The fact that each referendum requires its own separate legislation also allows the government to vary the franchise and the campaign rules opportunistically. There are some technical rules in the Political Parties, Elections and Referendums Act 2000, but there are no consistent fundamental principles for the conduct of referendums in Britain,  again in contrast to the position in other democracies.

The timing of referendums has been similarly opportunistic. There was no referendum in advance of the UK joining the European Economic Community in 1973, but in 1975 Harold Wilson held a referendum on continued membership, following a renegotiation of the UK’s position within the Community. This was the first UK-wide referendum.

Unlike the Danes, British voters were not offered a referendum on the Maastricht Treaty in 1992, although it fundamentally changed the terms of EU membership. David Cameron’s promise of a second European referendum in early 2013 was not in response to any recent significant change in the terms of British membership. Nor was there much evidence of widespread popular demand for it, despite many years of vociferous lobbying by those politicians who wanted us to leave the EU. The pledge was intended to placate the right of the Conservative party and win back UKIP voters ahead of the 2015 general election.

It is interesting that the three major referendums held while David Cameron was Prime Minister shared a common and rather curious feature. In all three cases (AV, Scottish independence and EU membership) Cameron initiated or conceded a referendum as a device with specific political objectives, and then campaigned actively against the referendum proposition and in favour of the status quo. Rather like the words of the hymn Jerusalem, Mr Cameron’s referendums posed a series of questions which invited the answer ‘no’. By contrast, referendums in other democracies, and previous referendums in the UK, have more typically sought popular consent for a change to the status quo which the government supports. That is an important point to which I will return later.

So the principle of what we may term the opportunistic referendum seems clearly established, and there is no reason to believe that a government of any complexion will in future voluntarily surrender this very useful political device.  That perhaps requires one qualification. Once a referendum has been held on some issue, a precedent is created which may make subsequent legislation on the same issue politically very difficult without a second referendum.  So if we continue to hold fairly frequent referendums, the latitude of future governments may over time be somewhat reduced. But even that is far from certain:  voters who rejected directly elected mayors in referendums in some parts of England are now having these mayors imposed upon them, in a slightly different form.

In a parliamentary democracy, some types of referendum have the potential to create an intractable constitutional difficulty. If a majority of elected representatives hold one view on a matter, and a referendum shows that a majority of the public hold the opposite view, which manifestation of democratic legitimacy should trump the other?

In Britain, Parliamentary Sovereignty is traditionally regarded as the governing norm of the Constitution. In principle, it would seem to follow that a Parliamentary majority can always overturn a referendum result. The political reality, as we are now seeing, may be different: those who voted with the majority will understandably assume that, while a referendum may have been merely advisory in law, Parliament should nonetheless defer to the ‘will of the people’.

That is certainly the view of the present government.  At the Conservative Party Conference last month, Theresa May made her position on the matter very clear:

Those people who argue that Article 50 can only be triggered after agreement in both Houses of Parliament are not standing up for democracy, they are trying to subvert it 

She went on to elaborate a novel constitutional proposition which my colleague Andrew Blick has christened the ‘May Doctrine’. According to the Prime Minister, so strong is the mandate delivered by the referendum that (paraphrasing Andrew’s recent paper)   Parliament, the devolved legislatures and the courts cannot question it and have no legitimate role in initiating its implementation. That is a matter entirely for the Executive, who are also solely responsible for interpreting voters’ views on the priority which should be given to various objectives in the Brexit negotiations.

The May Doctrine seems to me a very questionable proposition, in two respects.  First, it rests on an acceptance that the people have delivered their verdict with, in the Prime Minister’s words, ‘emphatic clarity’.  Well, it does not seem particularly emphatic to me. The majority in favour of Brexit was narrow: 51.9%, in contrast to the 67% who voted to remain in the European Community in 1975.  Putting it another way: on a relatively high turnout, 37% of the electorate actively supported leaving the European Union.

Pro- Remain commentators have highlighted other reasons for viewing the referendum result with some scepticism: the mendacious nature of the campaign, the fact that a significant element in the Leave vote was a generalised protest against an out-of-touch elite, that many Leave voters did not properly understand the consequences and so on. You may agree or disagree with these assertions. However we need to exercise some caution here: no doubt many voters were badly informed, but that is also true in a general election. We cannot know why people voted as they did, and so should not overstate our scepticism.

So the referendum result is a mandate, but it is weak mandate for an irreversible step (I will return to the issue of irreversibility later on) and certainly does not justify the sort of hyperbole we heard at the Conservative conference.

Second, the May Doctrine implies that a referendum result gives the government a constitutional trump card which allows it to override or bypass Parliament. That is a very novel, and I think dangerous, constitutional doctrine.

It was not widely reported, but when Theresa May dropped her support for withdrawal from the European Convention on Human Rights at the time of her accession, she explained her change of heart on the grounds that she did not think there was a parliamentary majority for withdrawal.  Politically, her change of stance was the price of support from one element in the parliamentary party, but her explanation is interesting.

The government’s clear view is that Article 50 of the Lisbon treaty can be invoked using the prerogative power. So why then can the prerogative not be deployed to withdraw from the Convention – another international treaty – without parliamentary endorsement?  The only logical explanation for the apparent inconsistency is that we have a new constitutional principle:  that a referendum can authorise the government to exercise the prerogative power in circumstances where it might otherwise be unreasonable to do so.

Last week the Divisional Court rejected the government’s view on the triggering of Article 50. The core of the judges’ reasoning was that when Parliament enacted the European Communities Act 1972, and subsequent legislation (quoting from the judgement at para 94):

Parliament intended EU rights to have effect in domestic law, and that this effect should not be capable of being undone or overridden by action taken by the Crown in exercise of its prerogative power.

That seems to me a very compelling argument, and entirely consistent with the established constitutional limitations on the exercise of the prerogative which date from the 17th century. The pro-Brexit newspapers of course took a rather different view.

The government is appealing to the Supreme Court, but ( at least according to some sources) does not expect last week’s judgement to be overturned. Perhaps in anticipation of this, the Prime Minister’s argument seems to have shifted somewhat. In her Sunday Telegraph article last weekend, Mrs May tells us that:

 Parliament voted to put the decision about our membership of the EU in the hands of the British people. The people made their choice and did so decisively.

The clear implication of this is that the parliamentary vote to hold the referendum was also a vote to implement its result. That is plainly wrong in law, but I suggest it is also a highly misleading representation of the decision MPs thought they were taking in June last year. This is from the House of Commons Library Briefing Paper issued to all MPs ahead of the 2nd reading debate:

This Bill requires a referendum to be held on the question of the UK’s continued membership of the European Union (EU) before the end of 2017. It does not contain any requirement for the UK Government to implement the results of the referendum, nor set a time limit by which a vote to leave the EU should be implemented. Instead, this is a type of referendum known as pre-legislative or consultative, which enables the electorate to voice an opinion.

If MPs had been told at the time that a future government would interpret a Leave vote as a binding instruction to invoke Article 50, and to do so without a vote in Parliament, I suggest that the debate on the Referendum Bill might have been a rather less anodyne affair.

While we wait for the Supreme Court’s ruling, and putting aside the specific issue of the scope of the prerogative power, I believe we can draw at least one obvious conclusion from recent events: that there is no consensus either amongst politicians or the public on the proper relationship between Parliament and the popular will as expressed in a referendum result.  People’s opinions on the issue are mainly a reflection of their views on the merits of Brexit; views which are being expressed in increasingly intemperate language.

I suggest it is unsurprising that there is no consensus on the principle. The question contains an inherent paradox: there is an unavoidable potential for conflict if a political system has two parallel and independent sources of legitimate democratic authority, and no constitutional rules which determine their relationship. That issue has interested political theorists for decades, but in Britain it has never previously been confronted in practice.

So why has this issue not come up before, since we have held quite a number of referendums over the years?  I suggest for two reasons.

First, the problem only applies to ‘advisory’ or pre-legislative referendums, where the consequences of a vote for change are not laid down precisely in advance.  It does not apply in post-legislative referendums such as the AV referendum of 2011, which sought popular endorsement for a new voting system for which Parliament had already legislated in the Parliamentary Voting and Constituencies Act 2011. If the referendum result had been in favour of AV, the new voting system would have been introduced without further recourse to Parliament. There was no potential conflict of democratic legitimacies because MPs have determined in advance the exact legal consequences of both possible referendum results.

The second reason that we have avoided a conflict of legitimacies is that, until 2016, all major referendums have either upheld the status quo, or produced the result sought by the UK government of the day, or both.

Including the 2016 EU referendum, eleven referendums have been held to date either in the UK as a whole or in one of the constituent nations of Scotland, Wales and N Ireland. All have been on constitutional issues and the great majority have been about transfers of sovereignty.  However the individual question was phrased, each effectively offered a binary choice between the status quo and some alternative arrangement, defined with varying degrees of precision.  (The only exception was the 1997 devolution referendum in Scotland, where there was an additional question about tax raising powers for the Scottish Parliament.)

There have been three UK-wide referendums, including the 2016 EU referendum. The previous two both upheld the status quo.  The 1975 referendum on the European Community was not a vote to join: Britain had been a member since 1973 and the referendum was about whether we should stay in.  The result of the 2011 AV referendum was similarly for no change.

In addition there have been a total of eight referendums on devolution and sovereignty in Scotland, Wales and N. Ireland. The first three of these were in the 1970s. In 1973, voters in Northern Ireland voted to remain part of the UK and not to join the Republic of Ireland. In 1979, Wales voted against devolution, while Scotland voted by a bare majority in favour of devolution. However a threshold requirement that was applied in the legislation was not met, meaning that the proposition fell. In the subsequent four referendums, there was a vote for change. N Ireland voted in favour of the Good Friday Agreement in 1998, Scotland voted in favour of a Scottish Parliament with tax raising powers in 1997, and Wales voted in favour of devolved powers in 1997 and again in 2011.  The eighth and most recent of these referendums was the 2014 Scottish independence referendum, which upheld the status quo.

So, excepting the 1979 Scottish referendum, there have been four referendums in total which endorsed change. Three of these were held in the early years of the 1997 Labour government, which had a substantial parliamentary majority. In each case, although the referendums were pre-legislative, the government was a strong supporter of the proposed changes: the creation of Scottish Parliament and a Welsh Assembly, and the implementation of the Good Friday agreement.

That leaves us with the 2011 referendum on the extension of powers of the Welsh Assembly as arguably the only example before 2016 where a referendum result has gone against the status quo without strong support for change from the UK government of the day.  If so, it is a very unconvincing example: the proposed changes were of a rather technical nature and arguably did not merit a referendum at all, there was substantial support for change in the Welsh Assembly itself, and an acceptance at Westminster that this was a matter for the Welsh electorate.

Thus the 2016 EU referendum is the only one of the eleven major referendums to date where the electorate voted in an advisory referendum both against the status quo and contrary to the recommendation of the government of the day.

This outcome is historically unprecedented, and explains why the potential conflict between Parliament and the views of the electorate as expressed in a referendum has been brought into sharp focus for the first time.

This leads on to a broader question. What is, or ought to be, the proper role of referendums in the British political system? Are they, as many Bremainers now seem to think, just a thoroughly bad idea, or are the some types of legislative decision which do properly require an additional – or indeed alternative – direct popular mandate?

Looking at the question from first principles, a referendum seems to me quite a good method for testing changes in social attitudes on matters such as, for example, same-sex marriage. These are questions where everyone’s opinion is of equal value, where many voters will have relevant personal experience, and where MPs may be out of touch with public opinion.

Interestingly, and in contrast to many other countries, there have in fact been no referendums in the UK on issues of this type, with the exception of the local referendums on Sunday opening in Wales in an earlier era.

Why is this? The obvious reason is precisely that, on many of these issues, the views of MPs have historically tended to be out of step with those of the public. Throughout the second half of the 20th century, MPs tended to have more progressive views than the electorate on matters such as the death penalty.  More recently, that relationship may have reversed itself on some questions.  To take one recent example: polls suggest that 80% of the public are in favour of a more permissive approach to assisted dying, but in 2015 a substantial majority of MPs took the opposite view in a free vote. Needless to say, there is no likelihood of a referendum on that issue, which would present MPs with an unpalatable choice between ignoring the popular will and voting against their own conscience.

The questions that voters have been asked in referendums, and with increasing frequency, are nearly all about sovereignty or the devolution of powers. (The AV referendum is the only clear exception).  Sovereignty referendums, which are nearly always advisory or pre-legislative, present two particular difficulties.

The first is the issue of asymmetric irreversibility. In sovereignty referendums, there is an important difference between the consequences of a vote for the status quo and a vote for independence, or withdrawal from an international organisation.  A ‘Leave’ vote, if implemented, is effectively irreversible: a ‘Remain’ vote leaves open the possibility of future referendums on the same issue.  That is quite unlike a general election, where whatever the result, voters have an opportunity to change their minds five years later.

Because of the asymmetrical nature of the possible outcomes, many constitutional theorists believe that sovereignty referendums should require some form of threshold or super-majority to overturn the status quo.  We did adopt this idea on one previous occasion: the legislation which provided for the 1979 referendum on Scottish devolution set a threshold for implementation of 40% of the registered electorate, a deliberately high hurdle which was not achieved.

So why did we not have a threshold or supermajority provision in the EU referendum?  After the event, 4 million people signed a petition in support of that idea, but it did not really get an airing at the time of the decision to hold the referendum.

It is fairly clear why David Cameron did not entertain the idea: UKIP and the anti-EU Conservatives would have cried foul, thus defeating the object of the exercise, and the PM was in any case very confident – along with most other people – that Remain would win with over 50%.

Perhaps more surprisingly, no MP raised the point in the debate on the second reading of the EU Referendum Bill on 9th June 2015, which was largely focused on whether 16- and 17-year olds should have a vote (although Lord Norton, to his credit , did raise it in the Lords’ debate).

The second peculiarity of advisory sovereignty referendums is the open-ended nature of a vote for change. There is another asymmetry here:  in the 2016 EU referendum, the nature of the status quo – ‘remaining a member of the European Union’ – was precisely knowable, even if it was complex. It consisted of the set of relationships between Britain and the EU which currently applied, together with the limited changes to the UK’s terms of membership which David Cameron had secured in his recent negotiations.

The nature of the proposed change – ‘leaving the European Union’ – was not known or knowable by anyone on polling day, however profound their understanding of the issues. That is because, as is becoming ever more plain, there are many alternative possible relationships between Britain and the EU, with very different economic and social effects.

It is worth noting in passing that the position at the time of the 1975 referendum was quite different. Britain had only joined the European Community (a far less complex entity that the EU later became) two years earlier, so most voters would have had a very clear idea of what life was like outside it.

Uncertainty about the nature of change does not apply equally to all sovereignty referendums. I believe most Scots in 2014 understood quite well what becoming ‘an independent country’ meant, despite competing narratives around some matters such as the future Scottish currency.

The meaning of ‘leaving the European Union’ was far less clear because EU membership is multi-stranded, and voters were being asked to consider a complex set of relationships as if it were a single thing. The government which called the referendum did not know what it meant either. Indeed it explicitly declined to give the matter any consideration, because it was generally assumed that a vote for the status quo was a foregone conclusion.

So we are now in a rather curious position. The people, who according to David Davis last week are ‘sovereign’, at least in this matter, have ‘instructed’ those in authority to leave the EU, and that instruction has been generally accepted as binding by politicians on both sides, (although not by the SNP).

However the referendum result tells us very little about what alternative arrangements the public would like to see.  The government, and many others, believe that concerns about immigration played a major part in the result, and that ‘taking back control of our borders’ should thus be at the centre of any new agreement with Europe. However even this is tendentious, since many voters may have believed that immigration could be dramatically reduced without material damage to their own standard of living: a YouGov poll in August indicated that 60%  of Britons would not restrict European immigration if it cost them personally a single pound of income.

Nor is it clear to whom this open-ended ‘instruction’ has been given. Has it been given to the Executive alone, or to Parliament, or to some combination of the two?  The question was barely considered before the referendum: now it is at the centre of debate.

So what general conclusions should we draw about the role of sovereignty referendums?  I will give you my own view about the Scottish and the EU referendums, but it is necessarily subjective, and I am sure many of you will not agree with me.

First, I believe that Scottish independence properly requires endorsement in a referendum. It is a decision beyond the competence of either the Scottish or the Westminster parliament. However it does not follow that David Cameron was under a constitutional obligation to facilitate a referendum in 2014, just because the SNP government in Edinburgh wanted one.  And if a referendum was to be held, it should – in my view – have had a threshold requirement. But because it did not, I fear that any second independence referendum must also be on the basis of a simple majority.

Second, I think it is clear that the UK government was not constitutionally required to hold a referendum to confirm our membership of the European Community in 1975. Earlier governments had made treaty commitments involving a comparable loss of sovereignty without a referendum:  for example, the agreements which allowed US forces to be permanently based in Britain.  The 1975 referendum was held mainly to resolve differences within the Labour government: a motive with a familiar ring to it.  Nonetheless, it became politically impossible afterwards for any subsequent government to leave the EU without another referendum.

However, it does not follow that a government which did not wish to leave the EU was in any way obliged to offer us that referendum.  David Cameron’s timing was entirely opportunistic. He was a risk-taker by temperament and, until last June, a successful one: to a betting man, Mr Cameron looked rather like a punter whose luck finally ran out in the sixth race of a seven-race accumulator. More seriously for the rest of us, he is also responsible for a major act of constitutional recklessness, because the conflicts of legitimacy which would manifest themselves in the event of a vote to Leave were entirely predictable.

Let me end with a few general thoughts.

First a question: Could we have held a referendum on EU membership without risking a constitutional train-smash?  Well, under some counter-factual circumstances, perhaps.  To take the obvious case, if a government had been elected on a mandate to negotiate withdrawal from the EU, the electorate could have been offered a choice in a subsequent referendum between the status quo and a fairly well-defined set of alternative arrangements. That might have worked out all right, although there would have been significant difficulties of timing and choreography. However in the actual circumstances of the 2016 referendum, I find it very difficult to see how the messy constitutional consequences of the unexpected result could have been avoided.

Second, not all referendums create potential conflicts of democratic legitimacy.  Post-legislative referendums are innocuous in this respect. It is advisory referendums which are constitutionally dangerous, and even these are unlikely to cause fundamental difficulties if the government supports the proposed change. In practice, serious problems arises only in very specific circumstances: where the government holds an advisory referendum in which it supports the status quo, and loses. So it would seem reasonable to conclude that, at least in the UK context, a government should not call an advisory referendum unless it supports the change which is being proposed.

Third, it seems to me entirely sensible that any referendum on an irreversible transfer of sovereignty should have a threshold or super-majority provision. However, as things now stand, that is rather unhelpful advice. A second referendum on Scottish independence is quite likely, and a second EU referendum is at least a possibility.  If either of these were held, it would be politically impossible to apply any rule other than a simple majority, because of the precedent set by the previous referendum on the same question.

Finally, underlying all of this is the fundamental problem of an absence of constitutional norms around referendums. This vacuum allows governments to call referendums for reasons of short-term tactical advantage, without adequately considering the constitutional consequences. They have done so with increasing frequency, and that tendency is only one example of a more general problem, which I believe has now become acute. The British constitution is now dangerously unstable, because of the series of uncoordinated changes which have been made, largely on the hoof, by governments since 1997.  This could not happen in most other democracies, because they have written constitutions which lay down special procedures for constitutional change. Let me leave the last word to my old tutor, Sir John Baker:

A written constitution may not be the best solution for this country, and it may not be practicably attainable, but if we are to carry on without one it would be good if some means could be found of protecting the wisdom of successive generations against the fickle politics of the moment.




This publication presents the personal views of the author and not those of The Constitution Society, which publishes it as a contribution to debate on this important subject.