Lessons Learned from the EU Referendum Contents

Conclusions and recommendations

Chapter 2: Referendums and representative democracy in the UK

A competitive or complementary relationship? Referendums and Parliament

1.The UK is, in principle, a representative democracy. The referendum in the United Kingdom has been viewed by some as a device which is both alien to our constitution principles and a potential threat to parliamentary sovereignty. However, referendums have become part of the UK’s largely uncodified constitution, with over a dozen referendums, since the early 1970s, on issues ranging from the future of Northern Ireland to the voting system used for Westminster elections. On the basis of these precedents, we agree with the House of Lords Constitution Committee’s judgement in 2010 that, if referendums are to be used, they are most appropriate as a device for resolving questions of key constitutional importance, and we consider them appropriate, when issues cannot be resolved through the usual medium of party politics. (Paragraph 14)

2.If the results of referendums are to command the maximum of public support, acceptance and legitimacy, then they must be held on questions and issues which are as clear as possible. Voters should be presented with a choice, where the consequences of either outcome are clear. There is bound to be uncertainty arising from what might be termed a “bluff-call” referendum, like the 2016 EU referendum, but this should not limit how the participants campaign on either side. The UK Government initiated the process which led to the referendum, despite being against the suggested proposal, and with the aim of using a negative result to shut down the debate about the question at issue. Moreover, the referendum was confined to a tight question, on the basis of a clear binary choice. There could, however, have been more positive efforts to explain, and therefore to plan for, the consequences for voters in the event of either outcome. This would have required providing impartial consideration of the outcome which the Government clearly did not want.  (Paragraph 19)

3.Parliamentary sovereignty, and the associated principle that no Parliament can bind a successor, makes the concept of a legally binding referendum impossible in theory. However, it is clear that, in reality, referendums are seen by the public as conferring an obligation on parliamentarians to deliver the result. Parliament has delivered this, and the EU (Notification of Withdrawal) Bill completed its passage through both Houses, and received Royal Assent on 16 March 2017. (Paragraph 23)

4.In other countries, referendums are not conducted on the basis that a Prime Minister must resign in the event of losing a referendum. A more responsible conduct of the Government’s case in the run up to the referendum, and proper planning for a Leave vote, would not have opened up so much new controversy nor left the Prime Minister’s authority and credibility undermined. Using a referendum as a “bluff call” in order to close down unwelcome debate on an issue is a questionable use of referendums. Indeed, it is incumbent on future Parliaments and governments to consider the potential consequences of promising referendums, particularly when, as a result, they may be expected to implement an outcome that they opposed. (Paragraph 24)


5.The relationship between the direct democracy of a referendum and the established expectations of representative democracy requires careful management. We agree that it is of the highest importance that the referendum process is seen to be fair, by both sides, and that the result is agreed to, even if not with, by both sides. This is particularly important given the public’s expectation, post-referendum, that the result will be accepted and implemented. (Paragraph 25)

6.To achieve this level of acceptance and legitimacy, referendums, therefore, need to be designed in such a way as to provide the utmost clarity for parliamentarians, campaigners and, above all, the electorate. Referendums should be limited to matters which lend themselves to a binary question. Confusion as to the possible consequences of a referendum result serves only to heighten the potential tensions between referendums and representative democracy and risks increasing the public’s disenchantment with politics. Referendums are the creations of Parliament and the Government. Parliament and the Government are therefore accountable and must take responsibility for the conduct of referendums, and the fairness of the question, and there should be proper information about, and planning for, either outcome. (Paragraph 26)

Chapter 3: The regulatory framework for referendums

The Political Parties, Elections and Referendums Act 2000

7.The reinstatement of the restrictions on government set out in section 125 of the Political Parties and Referendums Act (PPERA) was of critical importance to the fair conduct of the referendum and the legitimacy of its outcome. Parliament was right to resist the then Prime Minister’s desire that the government qua government should be able to promote its views on the referendum issue right the way through the referendum campaign. This would have constituted the use of public resources for political campaigning and would have made a nonsense of one of the purposes of PPERA, which is to create parity between the two campaigns. Fears that the application of section 125 could have potentially problematic consequences for the conduct of government proved groundless. Sir Jeremy Heywood conceded that the application of section 125 did not cause huge difficulties for business as usual. Regardless of the legal advice, or the concerns of ministers, the attempt to disapply section 125 of PPERA left the Government open to the impression that it was seeking to manipulate the referendum process. (Paragraph 41)

8.The provisions of section 125, while imperfect, have been successfully applied in numerous referendums since 2000. There is no evidence that section 125 created any of the threats to good governance that the Cabinet Secretary feared during his appearance before PACAC in July 2015. The purdah provisions of section 125 of PPERA play a key role in the fair conduct of referendums and must continue to do so in future referendums. (Paragraph 42)

Reforming the regulatory framework for referendums

9.PACAC supports the Law Commissions’ proposals to consolidate the law relating to referendums. The basic regulatory framework provided by PPERA would be strengthened by the establishment of a generic conduct order that can be applied for future referendums. Such an order would end the current practice whereby each referendum results in a process of ‘reinventing the wheel’ with referendum-specific provisions required which, in the case of conduct, essentially duplicate the provisions that were made for previous referendums. Consolidation in the form of a new generic conduct order would streamline the process of calling a referendum. It would provide greater stability in the constitutional and legal framework for referendums, and would also provide greater clarity for parliamentarians, campaigners and the electorate. (Paragraph 51)

10.Section 125 of PPERA provides important protection for the fairness of referendum campaigns, but the 28 days specified is too short a period in the context of the much longer official campaign period. The absence of a longer purdah period enables the Government and other relevant public bodies to produce promotional material during most of the referendum campaign period, a situation that is far from satisfactory. (Paragraph 60)

11.The Electoral Commission has long argued that the restrictions on the publication of promotional, referendum-related, material by the Government should be extended, beyond the current 28 day period, provided by section 125, to cover the full referendum period. Nothing but the Government’s political intentions are served by maintaining the 28 day purdah period. PACAC recommends that the Government should bring forward proposals to extend the section 125 restrictions so that they are in force for the full duration of a referendum period of ten weeks, as recommended by the Electoral Commission and so many other respected authorities. (Paragraph 61)

12.Section 125 was originally drafted some 16 years ago. Since that time campaigning and publishing have both become increasingly digital in nature. As a result, terminology and provisions that may have been appropriate in 2000, may be less effective at regulating campaign activity in 2017. PACAC, the Electoral Commission and the Government all had different legal advice and interpretations as to: a) whether the eureferendum.gov.uk website represented publishing for the purposes of section 125; and b) whether the steps taken to remove links to the website satisfied the exception provided in section 125(3)(a). This underlines the need for section 125 to be reviewed and amended so as to better reflect the increasingly digital nature of our democracy. (Paragraph 68)

13.PACAC recommends that the Government bring forward consultative proposals for the redrafting of section 125. These proposals should provide greater clarity as to the status of online publications, for the purpose of the section 125 restrictions, and what constitutes “specifically seeking access” to materials, under the terms of the exception laid out in section 125(3)(a), in respect of online material available from government websites. (Paragraph 69)

14.Section 125 has been in force for over 16 years and has operated, either directly or through similar provisions (e.g. via the Scottish Independence Referendum Act 2014), for a number of high-profile referendums. In light of this experience and the issues, detailed earlier in this report, that arose regarding section 125 during the EU referendum, there is now an opportune moment to explore the effectiveness of section 125. (Paragraph 74)

15.PACAC recommends that the Government undertake a wider review into section 125. Such a review should take into account PACAC’s recommendations that the length of the controlled period be extended, and for the provision to be redrafted so as to provide greater clarity on the question of what constitutes publishing. PACAC agrees with the Electoral Commission that this review should also aim to provide a tighter definition of the kinds of activities that should be restricted during the controlled, and referendum, periods. It should also explore the sanctions that should apply for the purposes of upholding section 125 and the investigatory powers that the Electoral Commission should enjoy to police the application of section 125. (Paragraph 75)

16.PACAC is concerned that campaigns are struggling to operate within the existing system of working together rules. These rules should provide clarity for campaigners, so that they can cooperate with one another in a way that is clearly consistent with the letter and spirit of the campaign spending rules. Instead, we heard evidence that the current lack of certainty regarding the working together rules has resulted in an “unnecessary freezing” of activity among campaign participants who wish to avoid any risk of breaching the rules. PACAC, therefore, welcomes the suggestion, from the Electoral Commission, that it will return to this issue after it has settled the full spending returns from the referendum campaigns. (Paragraph 81)

Chapter 4: The Electoral Commission and administration of the referendum

Administrative issues that arose during the referendum

17.The Register to Vote website crashed on the evening of Tuesday 7 June. The Government has stated that this was due to an exceptional surge in demand. It is clear that the level of public interest in the referendum, allied to the sheer numbers of duplicate applications and confusion as to whether individuals needed to re-register to vote for the referendum, created a much higher demand for the Register to Vote website. Duplicate applications pose an unnecessary administrative burden on electoral registration officers and are an equally unnecessary drain on the time of electors themselves. PACAC therefore endorses the Electoral Commission’s recommendation that the Government should develop an online service to enable people to check whether they are already correctly registered to vote. While PACAC is aware of the technical issues that would need to be overcome to deliver such a service, it would be of invaluable assistance in preventing the Register to Vote website from collapsing due to high levels of demand again ahead of future elections and referendums. (Paragraph 92)

18.The Government has consistently argued that the Register to Vote website’s collapse was the product of a significant, last minute spike in applications to register to vote, magnified by the large number of duplicate applications. However, the Government clearly failed to undertake the necessary level of testing and precautions required to mitigate against any such surge in applications. It is worrying that when testing identified issues in system performance, mistaken assumptions meant that these issues were not investigated further and corrected. (Paragraph 96)

19.The Register to Vote website is a key component of the transition to Individual Electoral Registration. The nature of the electoral cycle is that public interest and engagement grows in close proximity to election and referendum polling days. It is, therefore, of the utmost importance that the website can withstand spikes in the number of applications, not least as such spikes are likely to emerge near, if not on, the registration deadline. PACAC supports the recommendations of the Equal Experts report, in particular in ensuring more frequent performance tests of the website are conducted, including those that test it destruction. (Paragraph 97)

20.It is disappointing that the local electoral administrators who play a key role in delivering the referendum on the ground, including the implementation of the extended registration deadline, had to rely on news sources for updates on the Government’s decision to introduce emergency legislation. PACAC therefore welcomes the Electoral Commission’s recognition of the need for a more effective communication strategy between themselves, the Government and electoral administrators. PACAC recommends that the Government and Electoral Commission develop new guidelines, in consultation with bodies such as the Association of Electoral Administrators, on the level and quality of information provided to electoral registration officers and other administrators both prior to, and during, referendum campaigns. (Paragraph 101)

21.PACAC does not rule out the possibility that the crash may have been caused by a DDOS (distributed denial of service attack) using botnets. Lessons in respect of the protection and resilience against possible foreign interference in IT systems that are critical for the functioning of the democratic process must extend beyond the technical. The US and UK understanding of ‘cyber’ is predominantly technical and computer-network based. For example, Russia and China use a cognitive approach based on understanding of mass psychology and of how to exploit individuals. The implications of this different understanding of cyber-attack, as purely technical or as reaching beyond the digital to influence public opinion, for the interference in elections and referendums are clear. PACAC is deeply concerned about these allegations about foreign interference. (Paragraph 103)

22.We commend the Government for promoting cyber security as a major issue for the UK. We recommend that Cabinet Office, the Electoral Commission, local government, GCHQ and the new government Cyber Security Centre establish permanent machinery for monitoring cyber activity in respect of elections and referendums, for promoting cyber security and resilience from potential attacks, and to put plans and machinery in place to respond to and to contain such attacks if they occur. We recommend that the Government presents regular annual reports to Parliament on these matters. We recommend that Cabinet Office, the Electoral Commission, local government, GCHQ and the new government Cyber Security Centre establish permanent machinery for monitoring cyber activity in respect of elections and referendums, for promoting cyber security and resilience from potential attacks, and to put plans and machinery in place to respond to and to contain such attacks if they occur. We recommend that the Government presents regular annual reports to Parliament on these matters. (Paragraph 104)


23.On the whole, the referendum was well run and competently administered by the Electoral Commission. We pay tribute to the work of the Electoral Commission’s staff and local electoral administrators across the United Kingdom in ensuring that the referendum could be delivered effectively, within a compounded time frame, even with the added pressures that arose as a result of the crash of the Register to Vote website. (Paragraph 132)

24.The Electoral Commission’s dual role as a regulator and key delivery agent for referendums could pose potential difficulties. However, while we note suggestions that these roles should be divided between separate bodies, it is clear that the Electoral Commission is the only body, at present, which is capable of discharging both roles. (Paragraph 133)

25.The Electoral Commission’s dual role has resulted in the Chair of the Commission acting as Chief Counting Officer during referendums. Our inquiry explored whether it would be more appropriate for these roles to be separated out. However, we are again persuaded that the Chair of the Commission was the only person with the capacity, resource and expertise to co-ordinate the delivery of the referendum. (Paragraph 134)

26.The process of designating the two official campaigns could be improved to provide greater clarity and transparency by the Electoral Commission. In particular, it was unfair that designation was so late, so that the Leave side could not plan ahead and commit to spending while competing for designation, but the uncontested Remain side was not disadvantaged in this way. We recommend that the Electoral Commission undertake a review of the designation process to examine where greater transparency could be achieved. This review should include consultation with campaigners from each of the campaigns that sought designation during the EU referendum. It should address whether earlier designation would have been fairer, and whether there should be a more explicit fit and proper person test for those applying for designation. (Paragraph 135)

Chapter 5: The machinery of Government during the referendum

Guidance and the Civil Service Code

27.During the run-up to the EU referendum, there were many occasions when it appeared to many that civil servants were being drawn into referendum controversy. This damaged the reputations of the Civil and Diplomatic Services for impartiality. It is a matter of some regret to this Committee that the Government did not accept the recommendation of our predecessor Committee, PASC, in relation to the provision of a new paragraph in the Civil Service Code, which would have served to clarify the role and conduct of civil servants during a referendum campaign. This proportionate and sensible step, would have served to promote clarity and a clear understanding of the role for civil servants at the very outset of the EU Referendum campaign, thereby avoiding some of the controversies which arose during that campaign. (Paragraph 141)

The suspension of collective responsibility and the guidance given to Civil Servants

28.Relaxing Cabinet collective responsibility on the question of the UK’s EU membership was a sensible decision. However, the Cabinet Secretary’s guidance to Civil Servants on their conduct during the referendum, was ambiguous and caused unnecessary uncertainty as to the support civil servants could provide to Ministers supporting a Leave vote. We recommend that, in the event of a future referendum where Cabinet collective responsibility is relaxed, any guidance should be more precise as to the support that civil servants should not provide to dissenting Ministers, including a list of official departmental papers that would not be made available to those Ministers. All Ministers should continue to receive factual information in relation to the work and remit of their departments. (Paragraph 155)

The publication of referendum-related government documents

29.The Government was legally obliged to publish reports on aspects of the UK’s membership of the EU. However, the presentation of these reports, particularly those from the Treasury, and the decision to spend £9.3m on sending a leaflet, advocating a Remain vote, to all UK households, were inappropriate and counterproductive for the Government. These incidents strengthen the case for the purdah provisions of section 125 to be extended to the full length of a referendum campaign. (Paragraph 161)

The Government’s policy on Civil Service contingency planning

30.While the Government did not support a Leave vote, they nonetheless had a constitutional and public obligation to prepare for both outcomes from the referendum. In 1975, Whitehall undertook contingency planning for a possible vote in favour of withdrawal from the European Communities and there was no adequate reason for a refusal to prepare for either eventuality in 2016. Though we were relieved to learn that work was undertaken within the Civil Service on the potential implications of a Leave vote, Civil Servants should never have been asked to operate in a climate where contingency planning was formally proscribed by the Government. Such preparation would negate the need for the Prime Minister to resign. The UK Government failed to learn the lessons from the Scottish referendum in this respect. (Paragraph 172)

31.We recommend that in the event of future referendums civil servants should be tasked with preparing for both possible outcomes. While we recognise the important distinctions between General Elections and referendums, these preparations should include pre-referendum contact between the two designated campaigns and the Civil Service, along the lines of pre-General Election contact between opposition parties and the Civil Service. It should be reasonable to presume that the sitting Prime Minister and his/her administration will continue in office and take responsibility for the referendum result in either eventuality. (Paragraph 173)

Conclusion: the machinery of government and the EU referendum

32.While it is perfectly legitimate for the Prime Minister and government to take an official position during a referendum campaign, the fairness, and legitimacy, of a referendum rests on a careful and restrained use of the machinery of government. Unfortunately, many of the Government’s actions in the run-up to the referendum appear to have increased public distrust. As Sir Jeremy Heywood has acknowledged, the use of the machinery of government during the referendum contributed to a perception that the civil service were, in some way, biased. That any such perception exists is deeply regrettable and was entirely avoidable. We recommend that the Government heed the lessons from this referendum of the implications of the use of the machinery of government during referendums on public trust and confidence in the institutions of government. (Paragraph 178)

11 April 2017