Written evidence from Dr Craig Prescott (EUR 98)

Public Administration and Constitutional Affairs Committee

Lessons Learned from the EU Referendum inquiry

Introduction
 

1. I am a Lecturer in Constitutional Law, now at the University of Winchester, having previously taught Public Law at the University of Manchester and King's College London. I hold a PhD in Constitutional Law from the University of Manchester, and referendums form a key aspect of my research.
2. The questions raised by the Committee do not yield many easy answers, but in summary, my conclusions are as follows;
• Direct democracy through referendums provides a useful supplement to representative democracy, particularly on constitutional issues. The outcome of referendums should be politically binding on government and Parliament.
• Referendums are most appropriate for fundamental constitutional issues. This means that EU membership is an entirely proper issue to put to the country in a referendum.
• The EU referendum is not legally binding on Parliament or the government, but this is due to more technical reasons and the nature of the referendum rather than through an intention not to be bound by the result.
• Legally binding referendums have their difficulties, and non-legally binding referendums are preferable.
• Nevertheless, should a referendum be legally binding, then there should be legal thresholds imposed on the result, and possibly the turnout. One problem is calculating the thresholds as the electoral registers have not fulfilled this role, and election law is not always enforced consistently.
• Post-referendum planning appears to be extremely lacking within government, but without further reform, following a similar process used for general elections when the Civil Service consults with the Opposition is inappropriate.
3. I have made no particular comments regarding the campaign or the administration of the referendum.

(A) Relationship between direct democracy and representative democracy

4. The issue is how to resolve constitutional questions in the UK, and establish the appropriate role for referendums. Many countries around the world make clear when a referendum is required, and some constitutions (such as Ireland) require any constitutional amendment to be approved by the electorate at a referendum. The lack of a codified constitution means that the UK lacks clear legal answers to this issue.
5. Representative democracy does not serve constitutional issues as effectively as it may do for other policy areas. Consequently, direct democracy through referendums on constitutional issues provides a useful supplement to the democratic process.
6. Representative democracy rests on the doctrine of the mandate. Political parties at elections present detailed manifestos from which the electorate choose, and the election of a party into government which grants that party a mandate to implement their manifesto.[1] This means that the electorate votes for which party best reflects their preferences. As voters are unlikely to agree with all policies contained in a manifesto, they must vote for a party knowing that they disagree with some of their policies.
7. It follows that voters will disregard the policies that matter least to them, and as opinion polls constantly indicate that the electorate are uninterested in the constitution,[2] they are extremely unlikely to cast their votes according to the constitutional policies of the political parties (to the extent that they are even aware of them). This means that elections are at best a blunt instrument on gaining the views of the electorate on constitutional issues.
8. It is also the case that, despite manifesto commitments, there is often substantial disagreement on constitutional issues within political parties and with some referendums (such as the 1975 EEC referendum) have been held to resolve these disputes. This is particularly the case within the Labour and Conservative parties as they are based on shared socio-economic perspectives and values. These principles can be difficult to translate to constitutional issues which raise different types of questions.
9. Representative democracy can also fail to give the electorate a clear choice on constitutional issues. For example, in the 2010 General Election, Labour, Conservative and Liberal Democrats all (albeit with varying degrees of enthusiasm) had manifesto commitments to House of Lords Reform, but there was no such consensus amongst the public. It is inconceivable that 88.1% (the aggregate share of vote for Labour, Conservatives and Liberal Democrats) of those voting supported that policy. Similarly, at the 2015 General Election, these three parties supported Britain's membership of the EU (again to varying degrees of enthusiasm) and won 88% of the vote, yet a year later a majority of the electorate wants to leave.
10. Given that the legitimacy of a constitutional settlement rests on the agreement of the people to be governed by those rules,[3] it is appropriate that representative democracy is supplemented by direct democracy to ensure that constitutional rules have the agreement of the people. Also, there is the principle that although the legislature has the power to legislate, it does not have the ability to transfer that power to legislate away (either ‘upwards’ to the EU, or downwards to devolved institutions) without specific consent from the people.[4]
11. Given that direct democracy is capable of providing a specific mandate for a constitutional issue, when the outcomes of representative and direct democracy clash, direct democracy should prevail. As Vernon Bogdanor has recently stated, ‘the referendum has now established itself as a third chamber of Parliament, issuing legislative instructions to the other two’.[5] The issue with the EU referendum is that the Commons has been asked to implement a policy that a majority of MPs are against. The logical resolution of this, is for either MPs to accept the result as politically binding, or hold a general election whereby a more representative parliament could be elected.[6] In either event, eventually the will of the people will be implemented.
12. David Cameron in his Bloomberg Speech encapsulated the inevitability of direct democracy prevailing over representative democracy when stating that

“democratic consent for the EU in Britain is now wafer thin... Those who refuse to contemplate consulting the British people, would in my view make more likely our eventual exit. Simply asking the British people to carry on accepting a European settlement over which they have had little choice is a path to ensuring that when the question is finally put - and at some stage it will have to be - it is much more likely that the British people will reject the EU”.[7]

(B) Questions to be decided by referendums

13. Given the discussion in the section above, it is clear that fundamental constitutional issues should be decided by referendums. This begs the question as to what qualifies as a ‘fundamental constitutional issue’ in a country with an uncodified constitution. However, the House of Lords Constitution Committee gave an indicative list.

•    the abolition of the monarchy,

•    the leave the European Union,

•    for any nations of the UK to secede from the Union,

•    to abolish either House of Parliament,

•    to change the electoral system of the House of Commons,

•    to adopt a written constitution,

•    to change the UK’s system of currency.[8]

14. As that last item on the list is a reference to the Euro, then all these questions involve the transfer of power either within Parliament or to and from it and making the list difficult to disagree with.
15. Although not intended to be a definitive, one notable exclusion from this list is devolution to the nations and regions of the UK. However, beyond highlighting the expectation that the creation of a new institution should require a referendum, a more detailed consideration of this point is beyond the scope of this paper.
16. It is clear that for the reasons highlighted above, it was entirely appropriate to hold the EU referendum.

(C) The Legal Status of Referendums

17. To date, there has been one legally binding referendum held in the UK. This was the 2011 alternative vote referendum. Other than the 1979 devolution referendums in Scotland and Wales, all other referendums held at the national or regional level have been non-legally binding.[9] The outcome of these non-legally binding referendums have been followed by the government and Parliament on each occasion. Of course, the outcome of the EU referendum still has to be implemented.
18. For a referendum to be legally binding, it also needs to be post-legislative. This is because there needs to be legal provision made in advance detailing what should happen if the proposal in question is approved. With the 2011 alternative vote referendum, the Parliamentary Voting and Constituencies Act 2011, provided for the holding of the referendum,[10] but also contained the provisions necessary to introduce the new electoral system should it be approved.[11] The legislation then placed a legal requirement on a Minister to either bring those provisions into effect or provide for their repeal depending on the result of the referendum.[12]
19. It is possible for a non-legally binding referendum to be post-legislative, as shown by the 2011 referendum on whether the Welsh Assembly should gain primary legislative powers.[13] However, most non-legally binding referendums have been pre-legislative, as further legislation was required to implement the change if approved. For example, the 1997 referendums were proceeded by a white paper setting out the government’s proposals,[14] and legislation was enacted providing for the referendum to be held.[15] The Scotland Act 1998 and Government of Wales Act 1998 were required to implement the result of the 1997 referendums and the provide the legal basis for the creation of the Scottish Parliament and Welsh Assembly.
20. Certainly, in the case of the 1997 devolution referendums, the choice of a non-binding, pre-legislative referendum was as much a political as it was a legal choice. In 1997, the Labour Government was reluctant to use parliamentary time enacting detailed legislation only for it to be rejected at a referendum, as happened in 1979.[16]

Scottish Independence and EU Membership Referendums

21. Given this background, both the Scottish independence and EU referendum could only have been non-legally binding, as they could only ever have been pre-legislative. It would be impossible to place in law in advance of the referendum, the precise details necessary to implement Scottish Independence or the UK’s exit from the EU. This is because these could only ever be placed into law at the conclusion of the necessary negotiations. In this sense, unlike previous referendums, these two referendums are of a different nature, as they were not seeking approval for a proposed constitutional change, but establishing the authority to start a process of constitutional change.
22. One possibility for both referendums could have been to place in law a legal requirement to commence the relevant negotiations by a specified date after the referendum. As regards the EU referendum, this could have resolved the ongoing debate about the role of Parliament in triggering Article 50 as Parliament would have explicitly authorised the government to start the process to negotiate Britain's exit from the EU.[17] A drawback would have been to limit the flexibility of government's (and indeed Parliament’s) response to the referendum outcome. Additionally, determining an appropriate deadline for invoking Article 50 would be difficult. Overall, the unpredictable events since the referendum, particularly the appointment of a new Prime Minister, have shown that the flexibility around invoking Article 50 has been essential.

Are Legally Binding Referendums Desirable?

23. The benefit of legally binding referendums is certainty about the consequences of the result, particularly when the legal changes required to implement the change are already on the statute book. However, as stated above, this is not always possible. Also, there are problems with legally binding referendums.
24. Firstly, it is possible that the referendum campaign could reveal an issue or problem with the legislation enacted in advance. The legally binding nature of a referendum would make it difficult to reconsider the legislation, as it has already been passed by Parliament. Further, the government may feel that it is risks opening ‘Pandora’s box’ should it look to amend the legislation, as Parliament may wish to reconsider other aspects of the legislation.
25. Secondly, a legally binding referendum (whether it is pre or post-legislative) could mean that the government and Parliament are legally required to implement change based on a narrow result in favour on a low turnout. At least a non-legally binding referendum gives Parliament an opportunity to consider the issue again.
26. Finally, it is debatable whether, under the doctrine of parliamentary sovereignty, any referendum can ever be completely legally binding, as it is always open to Parliament to repeal or amend the legal provisions that purport to make the referendum binding.
27. The best approach is to hold a non-legally binding referendum proceeded by a white paper detailing the proposals. This happened with the 1997 devolution referendums.[18] The white paper sets the terms of debate at the referendum, and if approved, Parliament, while accepting the result as binding, can consider the detail of the government's proposals. This proved particularly beneficial to the creation of the Welsh Assembly where amendments to the Bill altered the eventual structure of the new institution.

Overall

28. Non-legally binding referendums are to be preferred as they give flexibility to the government and Parliament in how to respond. This is particularly the case when the long term consequences are impossible to place in law due to the nature of the question asked. EU and Scottish referendums are examples of this.
29. I have used the phrase ‘non-legally binding’ rather than ‘advisory’ in this paper, as it follows that the views of the electorate should be more than mere advice, which Parliament is free to ignore.

(D) The Regulatory System for Referendums

Should Referendums Have Minimum Percentage Threshold?

30. This follows on from whether referendums should be legally binding. If a referendum is legally binding, then legal safeguards should be imposed to prevent a proposal being implemented with a small majority in favour on a very low turnout. These safeguards could either be a minimum percentage of votes in favour of change and / or a minimum turnout requirement.
31. In my view, neither is desirable. Traditionally, one of the benefits of the UK constitution has been its flexibility, but both of these potential safeguards tilts the rules of a referendum against change. A minimum threshold of votes in favour of change would have prevented the creation of the Welsh Assembly in 1997.[19] In hindsight, it's difficult to suggest that this would have been desirable, as the Assembly has since become an established part of the constitution.
32. Similarly, a minimum turnout requirement means that voters can defeat a referendum proposition merely by remaining at home. This flatly contradicts the sustained efforts being made to increase voter engagement with politics. A positive feature of both the Scottish Independence and EU referendums is that turnout and engagement with politics increased, arguably because the referendum rules were neutral and every vote counted equally.
33. Furthermore, both of types thresholds depend on the electoral registers being able to provide accurate figures in order to calculate the necessary thresholds. In addition to longstanding concerns about the accuracy of the electoral registers, the registers are not designed to achieve this task. Many people, for various reasons, are registered to vote at more than one address and so can vote more than once in local government elections. These registers are used for general elections and UK-wide referendums when you can only vote once. Voting twice is a criminal offence, but as Charles Moore showed in The Spectator, there are seemingly few measures to ensure that electors only vote once at referendums, as the system traditionally relies on the honesty of the electors.[20] For any threshold to be accurately calculated, provisions of election law such as this need to be significantly tightened and enforced more consistently.

(E) Post-Referendum Planning

34. It is hard to come to a conclusion other than the planning for after the EU referendum has been inadequate for it appears that none has taken place. The then Foreign Secretary stated that it was not ‘appropriate’ to make any post-referendum plans.[21] Usually, referendums have been held to seek the approval of a change proposed by the government. It follows that the government will have made plans following in order to implement government policy. This was not the case with either the Scottish independence or EU referendum, as government policy was to maintain the status quo. On one level, it follows that there should be no contingency planning for either Scottish independence or exiting the EU as this was not government policy.
35. Although in general, a core purpose of government is contingency planning. In very different context, the Civil Contingencies Act 2004 imposes legal requirements on government to plan for various major emergencies, most of which, are thankfully, unlikely. Yet it seems strange that this principle of contingency planning does not carry through to major economic and constitutional events that have a reasonable chance of occurring.
36. Whilst it would have been difficult to set out a detailed strategy for any EU negotiations,[22] an explanatory document developed within government, after consultation with outside expertise outlining the options and the process of the negotiations must surely have been possible. The difficulty for the government is that that any document could be subject to freedom of information requests. Any disclosure could be portrayed as the government contemplating leaving the EU which would have been damaging for the remain campaign.

Civil Service Consultation?

37. One possibility is for the Civil Service to adopt a model similar to that of general elections and hold consultations before a referendum. The issue is who to consult. One obvious group to consult would be the lead campaign group selected by the Electoral Commission. In the EU referendum this would have been Vote Leave. The problem with this, is that campaign groups are unlike political parties seeking election, in that their sole purpose is to win the referendum. They are not in an equivalent position to the Opposition who are presenting themselves as an alternative government who will have responsibility to implement their policy.
38. It is also difficult to graft onto the existing regulatory structure any sort of requirement that the lead campaign group should explain how change should be implemented. For example, with the EU referendum the ideal structure of any agreement to leave the EU is hotly debated. This may make it harder to qualify as the lead campaign group, as the law requires applicants to show that they ‘adequately represent those campaigning for the outcome at the referendum’. Requiring applicants to become the lead campaign group to explain in their applications how their position could be implement may place the Electoral Commission in an awkward position as they could be faced with making a political decision when choosing which group to appoint as the lead campaign group.
39. As suggested in paragraph 36, a preferable approach would be to consult with independent experts.

September 2016

              Page 2 of 6