Finance and Constitution Committee

Meeting date: Wednesday, September 4, 2019

Official Report 622KB pdf

---

Referendums (Scotland) Bill: Stage 1

Agenda item 2 is to take evidence on the Referendums (Scotland) Bill. I welcome our first panel. Professor Justin Fisher is head of the department of social and political sciences at Brunel University London, and Dr Alan Renwick is deputy director of the constitution unit at University College London. Thank you for the written submissions that you have provided to the committee. I also thank others who have done similarly. The submissions are very helpful.

We will go straight to questions. The written evidence that we have received highlights the need for a minimum period of at least six months between the referendum legislation coming into force and its being implemented. In your views, does that six-month period apply to the legislation for each referendum or does it apply from the date of any framework bill for referendums being in place? I ask that question because it seems to me that almost all the required rules and regulations would be in place as a result of the passing of a framework referendums bill and, therefore, that the intention behind the Gould convention of a six-month period will have been all but met. That is my perspective, but you might have a different viewpoint, which I would like to understand. Who would like to take the floor first?

Thank you for inviting me. I agree that the Gould principle is that the rules of a referendum should be clear at least six months in advance, and that that implies that, if all the rules are in place and the only matters to be decided subsequently are the question and the date, the Gould principle would not be broken by setting a referendum somewhat less than six months in advance of the poll.

That said, it is clear from experience in Scotland and throughout the United Kingdom that allowing a decent amount of time for any referendum is really important. Above all, it is vital in order to allow voters time to hear from the campaigns, reflect on the arguments and come to their judgment. It is also important, as Justin Fisher in particular has pointed out in his written submission, so that there is fairness between the campaigners in their preparations for the poll. It is important for administrative purposes, as well.

Although the Gould principle applies just to the rules of a referendum, I encourage the view that doing referendums slowly and carefully is always the better approach to take.

I would echo Alan Renwick’s thoughts on that. It is important to recognise that referendums are relatively unusual mechanisms and that they will, almost by definition, attract an element of controversy. Therefore, it is important not to rush important considerations in the preparation for a referendum. Considerations include agreement on the question’s wording, which can be very difficult. As I say in my submission, although the designation process worked very well in the Scottish independence referendum, it worked very badly in the European Union referendum in 2016.

The lesson that we are learning—before our very eyes, at Westminster—is that rushing things is not a good way to proceed if the process is to be orderly. Therefore, my advice—which might seem to be unusually lengthy, but is in line with Alan Renwick’s—is that it would be better to be conservative about the time period than to try to rush things through.

I am interested in the actual referendum period. I appreciate that there is a timeframe for the passing of the necessary legislation. Dr Renwick, in your submission you say that the referendum period should be

“determined on a case-by-case basis”.

What bearing does the subject of the referendum have on the period? You made other points about separating out deadlines for registering participants from those for the controlled campaign period. How do you get to the point at which the judgment about what is too long or too short a referendum period is based on good principles?

I clarify that I did not say that the referendum period should be decided on an ad hoc basis; I said that it would not be unreasonable to think that some variation might be appropriate. In saying that, I was thinking that there are some issues for which a particularly long referendum campaign might be appropriate. For example, it is widely agreed that it was actually a good thing that the campaign in 2014 was as long as it was, because it allowed the issues to be dealt with in considerable depth.

The more important point is that a minimum referendum period should be defined. At the moment, as far as I can see, the bill does not do provide for that. Internationally, the absolute minimum period that is generally seen as being at the limit of what is acceptable is four weeks. Therefore, in my submission I suggest that a referendum period of at the very least four weeks should be allowed for. That takes account of the fact that the bill proposes that designation of lead campaigners should happen before the referendum period. That means four weeks for applying for designation, 16 days for determination of the designation and four weeks for a campaign, which gives a total of 10 weeks, which mirrors the UK-wide referendums framework in the Political Parties, Elections and Referendums Act 2000—PPERA.

A number of people have suggested in their submissions that the minimum period should be longer than that—that it should be 10 weeks, with designation taking place earlier than that, which would certainly be better. An absolute minimum of four weeks is fundamentally what is needed, but a minimum of 10 weeks would be better.

I share Alan Renwick’s view. I do not see a case for rushing. Referendums are comparatively rare and are generally on issues that are of great importance, so we should not try to push things through in a hurry.

As I have outlined in my submission, even the PPERA framework can disadvantage sides in a referendum in which there is more than one competitor for designation. The lesson that emerged very strongly from the 2016 EU referendum was that that should be avoided, which therefore elongates the period before the controlled period proper. It might be that the controlled period could be relatively short, but it is very important that the preparations be done over a longer period.

I am conscious that the International Institute for Democracy and Electoral Assistance—International IDEA—recommends that legislation

“should allow for an adequate period for the campaign”.

Obviously, practice has varied, over the piece. In relation to the alternative vote referendum, there was three months from the passing of the legislation to polling day. I do not remember anybody complaining about that.

The Venice commission talks about the need for

“fundamental aspects of referendum law”

to be in statute and for more detailed technical parts to be in regulations. Where do issues of timing fit in? Are they “fundamental aspects” that should be in statute or should they be in regulations? Is there—Dr Renwick touched on this—scope for looking at such matters case by case?

I think that the experience of the 2016 referendum suggests that leaving Governments to decide the period between the legislation and the referendum date can cause immense difficulty. That is precisely what we saw in respect of the leave side’s campaigns in the build-up to 2016.

There is an opportunity to learn from the experiences of major referenda in these islands. It strikes me that the very good lesson has been learned that Governments will, if they are given the ability to introduce a shorter period between the legislation and the referendum date, use it. That does not seem to be democratically very healthy.

The key thing is that the minimum period be specified in primary legislation. If it were deemed to be appropriate to extend that period on a particular case, that would not cause problems for me. A minimum period is fundamental: in his evidence, Justin Fisher sets out a very strong case for extending that minimum period beyond the PPERA minimum period.

My final question is a more general one. The bill is derived from past practice and is relatively full of examples of good practice. Where could it be, as I think someone said in evidence, more forward-looking and a bit less conservative? Where is there room for democratic innovation?

I will talk about the regulatory aspects; innovation is more Alan Renwick’s bag. We do elections in this country pretty well, because we do them regularly. We do not do referenda in this country consistently well, because we have not had many of them.

In 2014, we had a very successful referendum and, in 2016, we had a very unsuccessful referendum, from the perspective of the workings of the legislation. We need to learn from that. One of the big lessons of the 2016 referendum was that campaigners, whether they are registered or unregistered, were able to use the legislation to render such things as spending limits virtually meaningless. That cat is out of the bag. If we have more referenda, there is, unless that loophole is closed, a real danger that problems will arise.

Relatively few problems arose in 2014. In light of the experience of 2016, if steps are not taken on spending limits, I fear that there might be difficulties in Scotland. This is an opportunity to address some of the regulatory issues that I highlight in my written evidence, and specifically those on the spending limits for permitted participants and non-registered participants.

We will come back to spending limits. Your written evidence says that the bill is not fit for purpose, so we need to delve into that a bit more.

Are you finished, Angela?

I am interested in hearing Dr Renwick’s views on the scope for democratic innovation.

The bill is based strongly on a 20-year-old Westminster legislative framework. Given that this is an opportunity to establish a new legislative framework, it would be good to think hard about what changes would help Scotland to be at the forefront of democratic practice, rather than being a couple of decades behind democratic practice.

On the regulatory side, in addition to Justin Fisher’s points are many big questions around the rise of digital campaigning. A lot of work on that is going on in Westminster and Whitehall, where people all accept that existing rules are not fit for purpose. It is important that the Scottish Parliament pushes forward on that front.

09:15  

Beyond regulation, in my evidence I set out the scope for deepening the democratic quality of referendums in Scotland in two principal areas. One is the quality of information that is available to voters during a referendum campaign. In particular, lessons can be learned from Ireland, where citizens assemblies preceded the referendums, most notably on same-sex marriage, on abortion and on repealing the eighth amendment. There is lots of evidence—we might be able to talk more about it with Theresa Reidy in a couple of weeks—that the citizens assemblies helped the quality of debate in the subsequent referendums. They surfaced the range of arguments, identified arguments that were weaker or stronger, and led to much higher quality debate during the campaigns.

My other issue is how to avoid the mess that Westminster is now in with respect to Brexit. The difficulty of the Brexit referendum—in terms of its structure, an independence referendum would be the same, in essence—is that it was on a basic principle, not a precisely worked-out plan. We call that a pre-legislative referendum. A difficulty arises if people vote for a principle, then the version that is worked out through subsequent negotiations is remote from what was promised during the referendum. My view, and that of last year’s report by the independent commission on referendums that we ran, is that if referendum promises on a broad principle do not match what is subsequently delivered, there ought to be a second referendum that has been baked into the process from the start, so that voters are clear what they are voting for. That would ensure that there is a proper democratic mandate for what subsequently happens—it would ensure that Parliament does not get mired in debates about what is the will of the people, and can stand above that question. Therefore, a referendum would not end up with the mess that people on all sides must accept Westminster has got itself into.

We have opened up quite a wide area. I will focus on the more technical bits of the bill, but I say to Professor Renwick that there is a distinct difference with what happened in Scotland’s referendum in 2014; a white paper existed prior to the referendum, but there was not one prior to the EU referendum. Regardless of whether we thought the white paper was good, at least it was there.

Yes.

I will not be drawn on the white paper, convener, although I am quite tempted, given that introduction.

We have covered one aspect of the bill’s regulation-making powers—namely, the very broad powers that ministers would enjoy to specify the referendum period, if the bill were to be enacted in its current form. It seems to me that at least two other aspects of the bill would give ministers very significant powers. Dr Renwick has just said that we should not be too wedded to PPERA as a model because it is nearly 20 years old, but a key difference between the bill and the UK legislation is that, unlike in UK law, ministers would have the power under the bill to call a referendum. Is that an advance, or is it a retrograde step?

It is very clearly a retrograde step. It is clear from international practice that referendums should not be used to circumvent the representative process but should supplement it. Calling a referendum is a big decision. In recent years, we have seen the extent to which a referendum can have fundamental effects on politics, society and how we all think of ourselves. The calling of a referendum should be subject to the greatest possible scrutiny in advance. I would advocate that, as in the Irish case, that scrutiny should take place not only in Parliament but more widely, and should include citizens assemblies and other processes. However, at the very least, the calling of a referendum should be scrutinised in Parliament, where it should be subject to the greatest possible scrutiny.

That is possible only if a referendum can be called only through primary legislation rather than secondary legislation. Overwhelmingly, that is what international practice indicates should happen. The vast majority of referendums are post-legislative referendums, whereby a bill has gone through Parliament and a proposal has been made to put an issue to a referendum. That way, Parliament is not circumvented. I think that it would be highly undesirable for the Parliament here to allow itself to be cut out of the process.

Do you agree with that, Professor Fisher?

I have nothing further to add. Alan Renwick makes a very strong case, and I agree entirely.

Dr Renwick, the main example that you refer to in your written evidence is the power in section 37, whereby ministers will be able to amend the act by regulations. You say that that power should be

“defined more tightly or removed.”

If we were not to remove it but to define it more tightly, how should we do that? What tightness of definition would you advocate?

The principle should be that such powers should not exist unless there is a specific justification for them. An example that I give later in my evidence is a power to set up a citizens assembly in relation to a referendum. Because we are talking about a relatively innovative way of doing referendums, I would not advocate putting into the legislation a requirement that there must be citizens assemblies, but it would be sensible to empower ministers to set up that kind of thing.

There might be other aspects of the detail of the regulation of referendums that it would be sensible to deal with in secondary rather than primary legislation, but I would follow the advice of the Electoral Commission on that—I would not want to define that boundary.

I would like to clarify an issue that was raised by Angela Constance in her questioning. I do not want to retread the ground that has been covered on specifying the referendum period, but I think that Angela Constance said—I am sure that she will correct me if I am wrong—that nobody objected to the referendum period for the AV referendum in 2011 being three months. However, that was a post-legislative referendum, not a pre-legislative referendum. Am I right in thinking that, in your judgment, that would make a material difference on the question of timing?

I am not sure that that would necessarily make a difference, but it could do.

The Electoral Commission’s view was that a longer period would have been preferable in 2011, because the legislation for the AV referendum included a range of regulations pertaining to the campaign, so it did not entirely conform to the Gould principle. There should have been a longer period in that case.

Whether anybody objected at the time is neither here nor there. The experience of 2016 suggests that, with the benefit of hindsight, the short referendum period was unwise and created significant difficulties for campaigners and for the fairness of the campaign.

That was very clear and helpful. Thank you very much.

I want to pick up on Adam Tomkins’s second question about the use of regulations. The issue of what powers ministers have in this area is not black and white.

Sometimes, powers are there to deal with incidental or unexpected circumstances that arise. Although the EU referendum was mostly legislated for through primary legislation—the European Union Referendum Act 2015—important parts of the EU referendum, including the date of the referendum, the referendum period and the periods of reporting and spending during the campaign, were legislated for through secondary legislation.

I understand from Dr Renwick’s evidence that having regulation by ministers, rather than dealing with issues in primary legislation, would be unsatisfactory, but in any legislation there are areas in respect of which any unexpected circumstances or incidental matters must be tidied up.

Yes, but the difficulty that has arisen with the PPERA framework is that, because it is now 20 years old, we have had an accretion of various bits of additional rules in the primary legislation and in the secondary legislation for each referendum, and now is an opportunity for the Scottish Parliament to take all that and provide a consolidated set of rules in primary legislation. You are right: circumstances change, and rules need to change, and it is difficult to pass primary legislation on all those matters.

I am not saying that we would do this, but if the committee was to consider recommending to the Government a super-affirmative process—to use a Scottish Parliament term—around regulations, would that help?

Yes, but there are some matters that are currently proposed for regulations that in my view should not be in regulations.

I understand that.

Just to be clear, we are talking about section 37 here, not section 1.

Yes. At the moment, the section 37 provision allows any aspect of the bill—the act, as it would then be—to be amended by secondary legislation, and that seems to me to be inappropriate.

In your view, would that be inappropriate even with a super-affirmative procedure?

Yes.

I differ from Dr Renwick here in that there is a danger of characterising PPERA solely in the context of referendums. PPERA may be 20 years old, but in respect of elections it has been remarkably successful. There have periodically been amendments to the legislation—for example, adjusting donation declaration caps on loans. That has been done very successfully and there is a great deal of satisfaction among participants in elections with the way in which they are run.

On PPERA, it would be reasonable to say that, although elections have been successful under that 20-year-old framework—and I definitely agree that it needs to be updated in respect of digital campaigning—the referendum aspects are rather different, because referendums have been so infrequent and have thrown up issues that were probably not well thought through when the original legislation was passed.

I will pick up on section 37. It seems that the policy intent is effectively to have a keeping-pace power in order to keep pace with best practice in legislation concerning referendums and policy. Is it possible to characterise the issue as the policy objective not being met by the wording of the bill? Could the provision in the bill be modified and more tightly defined, so that it is clear that the powers are specifically for keeping pace with best practice and could not be abused by a future Executive?

For me, there are two principal difficulties with the provision. One is the generality of the section, which applies to the whole of the bill—there are no elements of the bill that are protected. I do not have the wording in front of me, but the other issue is that the bill says that changes can be made to the act in relation to

“any modification (or proposed modification) of any other enactment”.

To me, that seems an extraordinary power to confer on ministers—that they can propose changes to the act based on proposed modifications of other legislation, whether or not those modifications have gone through. The bill does not indicate what stage the proposal must have reached.

I would defer to lawyers on this point—I am not a lawyer—but it sounds as though a minister could propose some change to another enactment and, on that basis, would be able to propose, via secondary legislation, a change to this act. That seems a strange power to give to a minister.

09:30  

Could the wording be tightened in such a way as to avoid that dubiety? We could make it clear that the power would be used in relation to legitimate proposals by reputable bodies, for example, and that it is specifically to allow the act to keep pace with best practice without the need to resort to primary legislation.

One change would be to simply remove the words “or proposed modification”. On whether it would be desirable to make provision in relation to recommendations from the Electoral Commission, for example, you might want to pick up that issue with the Electoral Commission, as I do not have a strong view on it.

Do you want to add anything to that, Professor Fisher?

I agree with Alan Renwick on the issue. It would seem extraordinary to confer those powers. Even with the tightened wording that you suggest, the definition of “legitimate” would be a matter for debate. Most issues that come up for change are fairly major, such as the issue of digital imprints, and those can be dealt with through primary legislation. I share Alan Renwick’s concerns. The provision seems to me to be an unnecessary part of the bill.

I want to go back to Dr Renwick’s remarks concerning the powers that are conferred on the Executive earlier in the bill. You described that as “a retrograde step”, and I think that you said that it would “circumvent the representative process”. That is quite strong language. To be precise about language, the literal meaning of “circumvent” is “to go round”. However, the proposal in the bill is for the affirmative procedure to be used, which would necessitate a minister coming before a committee to seek its approval, with the instrument having already been scrutinised by the Delegated Powers and Law Reform Committee. A vote in the chamber would then be required. Do you accept that the phrase “circumvent the representative process” is slightly too strong a description? Further, is there an argument that the use of the super-affirmative process might allay some of the concerns that you have expressed?

It is the circumvention of the full representative process. The processes of primary legislation are set out as they are for a reason. They set the degree of scrutiny that is deemed appropriate for major decisions. A decision to hold a referendum is a major decision, so it should be subject to the greatest level of scrutiny in the representative system. No form of scrutiny of secondary legislation matches the scrutiny that is given to primary legislation. If it did, what would be the point of doing something through secondary legislation rather than primary legislation? It seems to me that the reason for doing something through secondary legislation is that it is easier to do, but that is not how the representative system is set up; it is set up to provide checks and balances. In less healthy democracies and in non-democracies, there is a long history of strong leaders using referendums to get round Parliaments. Scotland should not be going there, and it should not open itself to the suspicion that it could be going there.

You have again talked about going round Parliament but, to be clear, the proposal would have to go through Parliament.

I have one final question. In effect, the bill proposes to capture all the aspects of referendums, shy of the question, the date and the referendum period. When previous referendums have been legislated for by primary legislation, that legislation has contained a range of aspects. PPERA covers campaign rules but not process, whereas the bill covers process.

In effect, all there would be to decide would be the question, the date and the period. There would still be an opportunity for full parliamentary scrutiny; a committee could take evidence, both written and directly from ministers, and there could be a debate in Parliament. I am trying to understand the Government’s motivations. Is it more about giving certainty by having a defined time period rather than any nefarious motivations, which perhaps your remarks imply?

The biggest decision that is left unsettled by the legislation is whether to have a referendum. Within that, there is the question of what the question will be, what the timing will be and so on. However, the decision on whether to have a referendum is a fundamental one. There are three differences between the process for any secondary legislation and the process for primary legislation. First, there is less scrutiny of secondary legislation; secondly, secondary legislation is unamendable; and, thirdly, symbolically, it is somehow saying that it is of secondary importance. I do not think that any of those differences are desirable in the context of any referendum.

I am sorry to jump back a wee bit, but I want to follow up on a point that Adam Tomkins raised. Dr Renwick, you talked about the distinction between pre and post-legislative referendums. I entirely understand your point, but I am not quite clear where you think that should lead us. Are you arguing that pre-legislative referendums are, in principle, bad practice and should not happen or that there should be a formal threshold for when a pre-legislative referendum requires a further confirmatory referendum? In the Scottish context, I am thinking about the 1997 devolution referendum. It was pre-legislative, but I do not think that there was a general view that it was in any way flawed or illegitimate simply on the basis that that was how it happened.

You are right about that. What we have set out, and what the independent commission on referendums set out in its report last year, is that if, in calling a pre-legislative referendum, the Government sets out in detail in a white paper what it expects to happen, as happened in 2014, and then delivers on that, there is no need to have a further referendum. Broadly speaking, that is what happened in 1997, too; there was a clear plan, which was delivered on. However, if that is not delivered, democracy requires there to be a further referendum.

If the Government thinks that the matter is sufficiently important that it requires the consent of the electorate as a whole, such consent should be informed consent—it should be informed by an understanding of what the proposed change to the status quo will be, once that proposal has been worked up.

So, the objection is to an undefined principle. I think that a lot of people would agree that the UK’s current political situation stems from a mismatch between the simple proposition to leave and what is deliverable. However, there are also those who would say that they did not vote for a deal and that they have not been betrayed by no deal. It is a matter of interpretation whether the promise has been met. It seems difficult to codify that.

Yes. The independent commission proposed that it would be for the Parliament that called the referendum to determine whether what had been agreed fitted the proposal that was set out before the referendum was held, because you cannot have the courts deciding that. It is a fundamentally political question. In part, this is an attempt to defuse some of the awkward politics that Westminster has got itself stuck in. Moreover, it is an attempt to ensure that there can be informed consent on these fundamentally important questions. However, you are right that the messiness of politics is the messiness of politics, and you cannot excise that entirely from the process.

Okay. Thank you.

On that matter, Dr Renwick, in paragraph 42 of your evidence you say:

“The dangers could be greatly reduced if the Scottish government produced a detailed plan for independence (as it did ahead of the 2014 vote)”.

We discussed that earlier.

You go on to say that

“In the event of a vote for independence”—

others might not agree, but this is a bill about referendums and not necessarily about an independence referendum—

“it would allow the Government to claim a mandate for its plan both before the Scottish Parliament and in ... negotiations ... with the UK”.

Will you expand on that?

Bearing in mind that not all yes campaigners in 2014 supported the Government’s white paper.

No? That is a shock to me.

We formulated and developed the proposal for double referendum requirements, including in discussion with the Parliamentary Assembly of the Council of Europe, which adopted the proposal at the start of January in its recommendations for referendums across the board. When I am thinking about this, I am not thinking specifically about the Scottish independence referendum. However, I am very conscious that, in any particular place, the proposal will be thought about in the context of a particular referendum—and here that means the independence referendum.

I am conscious that the political implications of such a proposal are likely to be of concern to members of the committee, and it is important to make the point that, although it might appear on the surface as though requiring a two-referendum process might make things harder for the side that wants change, there are at least two reasons why that is not the case. The convener referred to the first reason. The standard argument against a two-referendum process is that it weakens the negotiating position of the people seeking change. If the UK Government knew that any proposal had to go to a second referendum, and if the proposal was not very strong, the UK Government might try to give Scotland a bad deal in the negotiations. The mandate point—if the Scottish Government had a mandate for its particular version of independence and for negotiating that particular version—is a counterpoint to that.

In addition, we have seen, through the Brexit process, that the idea that you cannot have a democratic process at the end of the negotiation of something as fundamental as either Brexit or independence does not hold up. You need to have some sort of democratic scrutiny process there, and therefore there is the danger that the other side will try to gain in the negotiations. In paragraph 42, I was just trying to counter that argument against our proposal.

The second reason, which I outlined in the final paragraphs of the evidence, is that if you have a two-referendum process, you make it much more likely that there will be a majority for change at the first referendum, because people know that there is always a second chance to vote if they do not like what they get in the end. Having a two-referendum process rather than a one-referendum process does not seem to have implications in relation to who will win. However, it does mean that you have a better democratic process.

Does Justin Fisher want to say something about that?

Just briefly. Alan Renwick’s proposal has a lot to commend it. It is important to remember that as a country—whether we are talking about the UK, Scotland, Wales or England—we are still feeling our way on referendums. Up until 2016 we got lucky. The referendum in 1975 was supposed to settle the issue of European Economic Community membership. It manifestly failed, and so the idea of a second referendum to try to settle such an important constitutional issue such as independence or withdrawal from the European Union seems to me to be fundamentally sensible in a country that is not very experienced at running referendums.

Let us get back into some of the technical aspects. I think that John Mason has questions about spending.

09:45  

Yes. Professor Fisher, some of your comments about spending limits were quite strong—in fact, you said that the limits are “virtually meaningless”.

Your written evidence includes figures, and I have to say that I struggle to get my head round some of them. It refers to

“designated lead campaigns (£1,500,000); permitted participants (£150,000); nonregistered campaigns (£10,000)”.

Is your main problem the relationship between the different spending limits?

Yes. In a sense, all limits for the designated campaign are, by definition, arbitrary, but once you have settled on a figure you need to have a relative spend for the non-designated campaigns, just as we do for third-party spending in general elections.

So you consider that the system for spending on general elections works better than the system for referendums.

Yes—without question. The lesson of 2016 is that the £7 million spending limit for the designated campaigns was dwarfed by the amount of spending on each side.

How will we fix that?

There are two ways to do that: reduce the amount of the permissible spending for the registered participants that is non-designated, and significantly reduce the spending for non-registered participants. The clear lesson from 2016 is that those who wish to exploit loopholes in the spending limits will do so.

Yes. I was involved in a group as part of the yes campaign in 2014. I was struck by the fact that salaries are not included in the spending limits—I think that that issue also appears in your written paper. There is the question about having to include the cost of any work that is contracted out but not having to include the cost of a staff member who someone already employs. If there is a gap between the spend and the limit, maybe it is the spend that is artificially low in some cases.

I think that the staff issue is an anomaly. The issue is the communication with voters through campaign materials, whether those be hard copy or digital—that is where the real abuse lies and where we saw a significant problem in 2016.

We could take the view that there should be no spending limits, but, if we do that, all bets would be off. For the past 20 years, we have had a principle of setting spending limits at national level. If we have such limits, we need to ensure that the designated campaign in a referendum or the candidate in a general election has primacy. We have had in place that principle for the candidate ever since the 1880s. Up until the 1990s, people who were not candidates could spend only £5. That limit was tested in the European Court. The figure was subsequently increased to £500.

It strikes me that, in referendums, the relative maximum spend for non-designated spend is out of whack, and that is particularly acute in the Scottish case where the non-registered campaign limit of £10,000 is the same as the limit that is in place for a UK referendum.

Despite the fact that it is a smaller area.

Yes, despite the fact that the electorate is much smaller. In relative terms, £10,000 spend for 1,000 voters and £1 million for 500,000 voters is considerably greater proportionately than is £10,000 spend for 7 million voters.

You say that spending on each side should be as equitable as possible. In one sense, I get that. On the other hand, am I right in saying that the remain side spent a lot more than the leave side but the leave side still won? The amount of money that you spend does not necessarily equate to getting the result that you want.

No. The amount that is spent does not tell you anything about the quality of the campaign. Vast sums of money could be spent on a terrible campaign. Indeed, the remain campaign did precisely that, as did the yes to AV campaign in 2011. Having the money does not mean that it is spent properly. You cannot legislate for that. However, you can legislate to set upper limits on how much should be spent. Indeed, the referendum legislation goes further than the election legislation in trying to ensure equity by having an element of public funding, to ensure that both sides are adequately funded. That does not exist for general elections

Would you want that to be in the primary legislation rather than in secondary legislation?

I think that it would be sensible to legislate for that and update it as and when required.

Does Dr Renwick have anything to say about that?

No. Justin Fisher is very much the expert on that. The one point that I would make is that the bill does not, of course, allow for public funding for designated campaigners in Scotland.

I think that somebody else will come on to public funding. Is that right, convener?

Yes.

I want to dig down a bit more on the spending issue. From what I understand, the Scottish context in 2014 was a bit different from the context that we have just been hearing about with regard to what happened in the 2016 referendum. If we look at the Electoral Commission figures for spending during the 2014 referendum campaign, which I have in front of me, we can see that both sides spent around 95 per cent of the limit and the combined spend of all other campaigners was less than half of that on the no side and around a quarter of that on the yes side. In the Scottish context, it is therefore not clear to me what the concern about a

“risk of multiple nonregistered campaigns challenging the primacy”

of a designated campaign is based on, because, if we reflect back on past practice here, we see that that did not happen.

You could take that view, but I would counter that by saying that the cat is now out of the bag about how to get round the legislation. That was very clearly demonstrated in the 2016 referendum. Scotland may be different, but that may be an anomaly in comparing the two referendums.

It seems to me that there could be any number of issues on which there is a referendum in Scotland and the sort of behaviour that we saw in the 2016 referendum campaign could be attracted. It would be folly not to learn from the difficulties that were experienced at the UK level just a few years ago.

Did you look at the Electoral Commission figures for Scotland?

Only in respect of the number of non-registered participants.

But not the detailed spending pattern.

No. My evidence is based on a large study that I undertook for the Electoral Commission on how things worked for campaigners. That seemed to me to be particularly pertinent.

Okay. Does John Mason have any other questions?

Yes—just on a minor point to finish off. People do not have to produce receipts for spending under £200. Is that a minor or a major problem?

I think that it is a minor problem. There is a real danger in overregulating what is essentially a voluntary activity. If very small receipts have to be declared, people will be driven out of politics, and the costs will be increased substantially, because much more would need to be spent on compliance. A level has to be set, and £200 seems appropriate.

That is great. Thanks very much.

Alexander Burnett has a question on Electoral Commission issues.

My question is about your views on the Electoral Commission’s role in the question testing for referendums—particularly Dr Renwick’s point in his submission that

“Question testing should take place even when a question has previously been tested or proposed by the Commission.”

Absolutely. I think that the process should always be applied, irrespective of whether the question has been tested before. As I and several others have pointed out in written evidence to the committee, preceding the Brexit referendum, the Electoral Commission looked at the issue twice. There was a private member’s bill previously; the second time round, the commission produced stronger recommendations that led to a change in the question.

It is clear that circumstances change, and the degree to which a question meets the intelligibility test may also change. Fundamentally, that has to be allowed for.

I go back to comments that we both made earlier. The only impartial reason that I can see for not having the Electoral Commission test questions is that that would speed things up, as it takes 10 to 12 weeks to do that. However, speeding things up is not a virtue in referendums, so I do not think that that is a good reason for not testing questions.

I completely agree with Alan Renwick on that issue. It is pertinent to note that polling companies constantly review their questions because the questions rapidly go out of date in respect of people’s understanding of what they mean. For something as important as a referendum, and given that referendums do not take place regularly but are one-off events, it would be sensible to question test on every occasion.

Thank you—that is very clear.

Professor Fisher, you said earlier that it is important to get agreement on the questions. Aside from the importance of the process being robust, do you agree that the wording of section 3(7), which excludes the role of the Electoral Commission if the questions have already been used in a previous referendum, could undermine confidence in the questions if there is not broad agreement and consensus on them?

Going back to my previous response, I think that excluding the Electoral Commission from the process would be inadvisable, so I would not recommend that at all. It may well reduce confidence—however, that is pure speculation. On whether the question remains fit for purpose, it seems to be a leap of faith to say that, if it worked the last time, it must work this time. Would a time limit between referendums need to be set to allow for such a decision to be made? The provisions seem a little loose, and excluding the Electoral Commission would remove an important part of the referendum process.

Alan Renwick, in your written evidence you suggest that the 28-day purdah period be changed

“to cover the whole of the natural campaign period”,

although certain announcements would continue during that period. Currently, the 28-day rule on purdah is clear and is understood by most voters. Would this change create confusion?

No. Why do you think it might create confusion?

Because, at the moment, there is a cut-off point after which, in theory, nothing should be published by the Government within the 28-day period. You are suggesting a “long and thin” model whereby some announcements could continue over the whole referendum period.

There should still be a cut-off point—it is just that it should cover the whole referendum period. For the 2016 referendum, for example, the 28-day period clearly covered the final four weeks of the campaign but the referendum covered a 10-week period, and there was much active campaigning going on during those first six weeks. It would make no sense at all to have different periods for these things. That would create confusion.

We have a situation whereby a regulated campaigner would incur expenses for making a claim but the Government could make a claim that was not regulated at all during the initial period of the campaign, and that makes no sense at all. Whatever the actual referendum period is, that should be the period to which the restriction on publications by Governments should apply.

If purdah covered the whole period, what impact would that have on normal, day-to-day Government business?

If the scope of the provision were unchanged, the danger—particularly for a broad topic such as independence or Brexit—is that it would impinge on a Government’s normal business. That is why you need to not just lengthen the period but narrow the scope so that it covers only communications that promote a particular outcome in the referendum. I know that the Electoral Commission has been working on detailed proposals on exactly how to define that scope. In essence, it would be based on the sorts of communication that count towards expenses for registered campaigners.

10:00  

Given that the 2000 act will still apply across the rest of the UK when the bill goes through, how would the UK Government abide by that if it had an interest in a referendum that was taking place in Scotland?

I presume that, if it was a referendum that was subject to a section 30 order, there would be something similar to the Edinburgh agreement to get that agreed between the two Governments. I confess that I have not thought through what would happen if a referendum was not subject to a section 30 order and the UK Government was interfering in things. Clearly, that would create difficulties.

That is why I asked the question. It would create difficulties. It would create an uneven playing field for the two Governments. Is there any way in which we can resolve matters? Surely, the easiest way to do so would be to replicate what currently happens under the 2000 act.

The danger of doing that is that we would be allowing the tail to wag the dog. It is not obvious why the UK Government would intervene in a referendum that was not subject to a section 30 order.

I agree that there is a potential difficulty here, but the bill should be trying to regulate referendums in Scotland as well as possible. The current provisions regarding publications by Governments are just not working—it is manifest that they did not work in 2016—and they should not be allowed to stand. Obviously, I am arguing—along with the Electoral Commission and many other people—that PPERA should be changed, too.

Has the UK Government given any indication that it intends to look at the 2000 act?

It has indicated that it is looking at aspects of the 2000 act, particularly in relation to digital campaigning. There are many people in the UK Parliament who think that such provision ought to be included in that review, but I cannot comment on what the current UK Government thinks.

I must admit that I have never been entirely comfortable with the idea of excluding Government from the referendum process. The provision in PPERA was introduced largely because of the concerns of a member of the then Neill committee about the conduct of the 1975 referendum. In one sense, it seems slightly perverse that a Government that wishes to propose a referendum should not be able to have any say in it.

It is also worth saying that, in the 2016 referendum, the Government complied entirely with the law when it came to distributing the booklet—I think that I was the only person in the UK who did not receive one—but there was still a great deal of controversy about that. The period in which Governments can make any claims is set. By definition, that is slightly arbitrary. Wherever that time point is set, Government will push up against it—of course, it has an interest in the outcome.

There is a significant regulatory challenge. You outline clearly the one that would be posed by a UK Government intervening in a Scottish referendum, albeit that, politically, it would seem very unwise for it to do so. It is an extraordinarily difficult area on which to legislate, and I would be cautious about excluding Government from certain areas of business during a referendum campaign for what is quite a substantial period of time. I do not have a neat answer to your question, but it raises the issue of being careful what we wish for.

Patrick, I think that you want to ask questions about a national database—is that right?

Yes, thank you, convener. In their submissions, Dr Renwick and Professor Fisher both talk about establishing a national database to make it easier for campaign organisations—in particular, smaller, less well-resourced campaign organisations—to check the permissibility of donors. Have you thought of alternative ways of achieving the objective of removing that barrier for campaign organisations? Have you considered any possible unintended consequences of the database approach? I am thinking of practical consequences, such as a possible mismatch between who was on the national database and who was on the locally held electoral register, or political consequences.

Over the past decade or so, we have seen two or three attempts to establish national databases for one purpose or another—most recently, two years ago, when the Scottish Government abandoned the proposal to turn the national health service central register into a national database. Whether the concerns are legitimate or illegitimate, there are real problems of perception about how national databases potentially change the relationship between citizens and the state. Given that trust is really important in a democratic exercise, is there a danger that the proposal will be interpreted in a way that will change people’s level of trust in the system?

I accept that there is a danger of that. As a minimum, you could ensure that all local authorities keep data in the same format, to enable the merging of that data. We are talking not about holding dual databases but about a system whereby a newly established referendum campaign will not be required to go round every local authority in the country in which the referendum is being held.

Would having local authorities use the same format be reasonably easily achieved?

It would mark a fundamental shift from local authorities effectively doing what they like to the Electoral Commission running elections more centrally. It would require a shift, and there has been some resistance to that, but quite why there is such resistance is beyond me. It does not seem unreasonable to insist that the registers, even if they are kept at local authority level, are all in the same format, because they are critical to compliance in relation to donations. The local authority registration existed before the compliance requirements in respect of donations were introduced by PPERA. It seems to me that one piece of legislation has not kept up with the other.

Dr Renwick, do you have anything to add?

No.

Related to the notion of compliance and the need to check the acceptability of donations are the recent allegations that were made against the Brexit Party about the way in which it was raising funds from smaller donations that were below a threshold and about whether they might have been coming from overseas donors. Is there a case for our going further when questioning which donations a campaign body is responsible for and for our checking and reporting on where they come from?

It is a question of threshold. The essential principle is that donations should be made by persons who are registered to vote in the United Kingdom. At present, there is a threshold that means that not all donations need to be checked. In the light of the accusations against one particular party, there may be a case for lowering that threshold, but I go back to the point that I made earlier in respect of receipts. There is a balance to be struck between effective regulation and overregulation, which would need to be discussed between all the parties.

Thank you.

I have one additional question, which is on digital imprints. One of the big changes in campaigning in recent years has been the use of technology—particularly social media, which in any campaign can give rise to issues about how its use is controlled and regulated and how spending on it is monitored. Are there any experiences relating to digital imprints and digital campaigning from referendums internationally that we can look at, which could be used to strengthen certain provisions of the bill?

I think that digital imprints are required in New Zealand, where there have been some difficulties in working out exactly where to apply the rules and where not to apply them. In my written evidence, I raised the issue—which I think the Electoral Commission also raised in its submission—that we do not want to capture individuals who are just expressing their views on Twitter. There have been difficulties around that area in New Zealand, and it is clearly a point that needs to be thought about quite carefully.

Scotland was, of course, the first part of the UK to issue a regulation requiring digital imprints, for the 2014 referendum. That provision was widely welcomed, but the Electoral Commission concluded afterwards that there had been some problems with a lack of clarity around where the threshold should have been applied. The bill proposes a slight addition to the wording of the provision relating to that in the 2013 act, but it still does not really address the issue. It still does not make it clear that someone who is expressing a personal view and not being paid for it does not have to provide an imprint. That needs to be clarified in the bill.

As Alan Renwick has said, the need for digital imprints is quite clear, although I would demur from the Department for Digital, Culture, Media and Sport’s statement that the legislation is therefore not fit for purpose. All the empirical evidence at this stage is that the actual electoral impact of digital campaigning is somewhat overstated. That is likely to change as other campaign methods become more important.

As Alan Renwick says, it is difficult to capture things like organised Twitter campaigns, but I do not think that legislation should seek to do that, because, in many ways, it is no different from simply trying to regulate ordinary conversational campaigns. There comes a point at which we have to say, “If this is an advertisement, it should have an imprint,” but there are some things that simply cannot be regulated. You cannot regulate people chatting in a pub, and you cannot regulate people expressing a view on Twitter, even if collectively they do a lot of it.

It seems to me, therefore, that we should have digital imprints not necessarily because there is a desperate need for them now but because there will be a significant need for them in the future. Always remember that you cannot regulate every aspect of a referendum or an election.

I am a little bit uneasy with the distinction that you have made between formal advertising and online conversation. Surely the change that social media brings about is the ability for organised funded campaign organisations to capture online and social media conversation and to use it as the platform for their message. Surely we need to think about a way of achieving the regulatory effect in that space.

How would you do that? In one sense, that is no different from a group of Liberal Democrat party members meeting in a pub, chatting with people and saying, “The Lib Dem view on this is X.” They are having a public meeting.

Online conversation seems to me to be something that goes beyond what one could reasonably regulate. You might say, for example, that any official Twitter account for Labour, the Conservatives or whoever, should be subject to regulation. It will have a digital imprint because the account says “Labour” or “Conservative” on it. What is to stop people using a fake account?

What if social media are being used, as in recent examples, to boost massively an organisation’s false claim about Turkey’s EU membership, or the impact of EU membership on the ability to protect polar bears—claims that are very clearly organisationally derived campaign messages—and those claims are unaccountable?

I understand the dilemma, but I am not sure how different that is from any of us spreading disinformation through everyday conversation with colleagues. It is a question of scale rather than principle.

If it is a question of scale, would it be reasonable to suggest that the threshold for regulation should be the reach of a social media account rather than whether it is social media or paid media?

10:15  

We have imprints, which have originally been on campaign materials, so that we can track and audit spending. If an organisation takes out an ad on Facebook or Twitter, there can be a digital imprint to match the spending. However, when the Labour Party, the Scottish National Party or whoever engages in a Twitter conversation, that is cost free.

I accept that financial accountability is one aspect, but in order to achieve accountability for telling the truth, is it not also necessary to know whom a message is coming from, because that allows us to judge whether an organisation is being honest in its campaigning?

I think that you are equating official campaigns with truth telling.

I am saying that they are not always the same—that is the problem.

I recognise the issue, but there is a danger of going down a road on which we try to regulate anything that moves. There is a point at which regulation can become ineffective if the people who are pushing social media are getting around the legislation. We would not want to create the equivalent of the Dangerous Dogs Act 1991. There is a real risk of regulation falling into disrepute if it tries to cover everything but ends up doing something very badly.

I appreciate that the matter is difficult; I think that it needs more thought.

A lot of thought is being given to that. The Electoral Commission is working hard on the question, and the UK Cabinet Office has been thinking about it, and said under the previous Administration that it would be introducing proposals on how to do digital imprints later this year. Quite a lot of thinking is happening.

Members are right to identify that there are problems on both sides of this difficult issue. It seems to me that we do not have enough transparency. A lot of campaigning is opaque and we do not have the balance right, but exactly how to shift it is a difficult question that requires engagement with the organisations.

I have a couple of questions for completeness, just to make sure that we cover all the ground. The Electoral Commission submission points out that the bill

“does not require campaigners to give details about what they spent money on, or when, where and how they spent it.”

It recommends amending the bill

“so that campaigners would be required to include this information in their spending returns”

on the basis that this is proportionate and

“does not impose an unreasonable burden”.

Is that a reasonable suggestion and is there an international perspective on it?

That is certainly a reasonable suggestion. The UK goes further than most countries in relation to the degree to which spending is already regulated in referendums—it is already doing better than most—but it is still not doing well enough. The Electoral Commission’s recommendations have been developed over many years: they are well thought out, and I would take them seriously.

That is not an unreasonable suggestion at all, given that the data will already have been collected by the organisations. Every organisation will have a sense of where and when it spent money, so I do not have any objection to the suggestion.

I have a final question—again, on the Electoral Commission. It has raised concerns about allowing campaigns that spend over £250,000 up to six months to provide an auditor’s report. It points out that once it had carried out compliance checks, the information would not be publicly available until about nine months after polling day. Its view is that the information needs to be available to voters

“as soon as possible after a referendum, while it is still a live issue.”

What is your view on that?

There is always an issue about when information is published. The question of it being a live issue is important. For a referendum, however, that would be rather difficult because there would be no subsequent referendum unless Alan Renwick’s sensible proposals were introduced.

I can see why there would be an argument for shortening the period, but I do not have a firm view on how long it should be. However, we need to recognise that once a vote is cast in a referendum, the point at which there is moral outrage might be two months or it might be nine months later, and will not change the outcome unless there was an illegal act.

The horse will already have bolted.

Quite.

The independent commission on referendums took evidence on the issue and came to the conclusion that three months would be a reasonable cut-off date.

It is worth adding that the organisations involved might have ceased to function nine months after the referendum. Therefore, it would be sensible to shorten the period, so that they do not incur further cost unnecessarily.

I have a question about thresholds, which are not really touched on in any of the submissions. Petition PE01754, which has been lodged with Parliament, suggests that a two thirds majority should be required for constitutional change to be agreed to, as is the norm for many organisations, from sports clubs to political parties. Do you have any views on that?

That would be a very bad idea. To have a majority vote for a proposition and then be told that that majority does not have any standing inflames passions and does no good to the subsequent political processes. Very few countries have supermajority requirements for referendums. The only such referendums that I am aware of have been in Canadian provinces.

Quite a few places have turnout thresholds or what we call electorate thresholds, as was used in Scotland in 1979. Turnout thresholds are clearly undesirable and a bad idea because they encourage people who are in danger of losing to suppress turnout in order to invalidate the vote. Electorate thresholds—in which the percentage that must vote in favour of a proposal is set; it could be 40 per cent, for example—do not have such clear theoretical problems. However, it seems to me that use of an electorate threshold was discredited by the 1979 experience, so you would be a brave politician to recommend introducing one in Scotland.

Generally, my view is, as I have indicated, that a referendum should be part of a multistage decision-making process; it should not be a one-stop way of getting to a decision on a major issue. The safeguards should be in the rest of the process—not in any superwacky threshold on the referendum vote itself. Having in place such things as proper parliamentary scrutiny and a citizens assembly process ensures a good democratic process, and if the majority wants change at the end of that detailed process, that would be the democratic decision.

I would not put it quite as forcefully as that. I see a case for a supermajority for fundamental constitutional change. However, I think that the dual referendum proposal that Alan Renwick makes is a better safeguard and is more defensible. Of course, if people know the rules beforehand, a supermajority is, arguably, defensible. However, on balance, the dual process is a better one.

This session, which has been quite long, has been more than useful, so I am grateful to the witnesses for coming along to provide evidence, which saw us move from supermajority processes to superwacky processes—which is now a new part of our lexicon, as far as the committee’s report is concerned.

We will have a brief suspension before we move to the next panel.

10:24 Meeting suspended.  

10:30 On resuming—  

I welcome our second panel. Michael Clancy is the director of law reform at the Law Society of Scotland, and Jess Sargeant is a researcher at the Institute for Government. Thank you both very much for your helpful written submissions. We will go straight to questions.

Good morning, panel. You obviously heard the earlier evidence session. I have a general question about your overall assessment of the bill. An apparent theme in all the evidence that we have received is that the bill, although it is based on previous good practice, could be a wee bit more forward looking in some areas. Do you agree or disagree? Do you have any comments in that regard?

I agree with that general assessment. The purpose behind the bill is very good. At the moment, Scotland does not have standing legislation that would apply to future referendums; legislation was applied ad hoc for the 2014 referendum. It is important to have standing legislation, so that the rules are known in advance and are not developed for each separate referendum.

As you have mentioned, the bill is based on the UK’s regulatory framework, which was introduced in 2000. Obviously, some updating of that legislation needs to be done, and the bill incorporates a lot of the changes that have been applied to UK referendums since 2000. That is positive.

There is scope for further updates. However, as the previous panel mentioned, a lot of the legislative solutions to regulating the new types of campaigning are not obvious yet. Unlike PPERA—the UK’s regulatory framework, which has been unamended since 2000—the bill needs to become living legislation, with opportunities to amend it when legislative solutions need to be made available.

As we say in our submission, our main concerns are the regulation-making powers in the bill and the fact that the referendum question and period would be specified in subordinate legislation rather than in primary legislation. Although there might be reasons for doing that, it is important that any future referendum and its result command maximum legitimacy. Therefore, there should be maximum scrutiny of that referendum proposal, so we propose that those matters be covered in primary rather than secondary legislation.

The committee can be confident that the Law Society shares many of those views. As we say in our submission, the bill is a good idea. We say that the policy objectives that have been set by the Government would be met by the bill and

“We agree that the provision of a standing legal framework will enable future referendums to be legislated on in a consistent and timely manner.”

However, there are concerns about the proposed regulation-making powers and the powers for the Electoral Commission. Also, with a view to the future, we think that lessons could be learned from what has transpired in relation to the European Union referendum. In particular, the views of the Electoral Commission and the Information Commissioner’s Office, and some of the investigations of Westminster committees, should be drawn on.

I am interested in Jess Sargeant’s comments about making the bill a “living” bill. Obviously, there are opportunities to amend legislation once it has been passed. That tends to be done through regulation-making powers. Do you have any thoughts on how to keep standing legislation up to date with best practice?

That is difficult, because it often requires legislative time. One of the reasons why the UK framework has not kept up in that way is that the Government has not been prepared to give time to amend it. The bill contains a regulation-making power for ministers to update regulation on the basis of Electoral Commission recommendations. That is how the Scottish Government has approached the problem of being able to update the legislation. In Westminster, referendums usually require primary legislation. If the regulation-making powers were taken out of the bill, the opportunity to pass primary legislation to set the question and the referendum period every time there was a referendum would provide a legislative vehicle through which to make updates.

As a coda to that, I say that the power to modify in section 37 has provisions for Scottish ministers to make regulations

“in consequence of or in connection with any modification (or proposed modification) of any other enactment relating to ... the conduct of referendums ... elections”

and the “entitlement to vote”.

Clearly, that is an attempt to future proof but—as we heard earlier about concerns regarding the phraseology or proposed modification—that stretches future proofing to not just what has happened but to what is perceived or imagined might happen at some point. An amber light should be placed there and we should proceed with caution. As Jess Sargeant mentioned, that provision is supplemented by giving

“effect to recommendations of the Electoral Commission”.

If we flip to section 2, we see that it will allow regulations under the eventual act to make “incidental, supplementary, consequential, transitional” and all the other usual elements of modification, and to “modify any enactment”. In our evidence, we highlight that that does not specify the bill, but any enactment would include that element, once it is enacted. Therefore, there are two provisions that address potential modification of the legislation.

Okay. Colleagues will come to the question how we balance the ability to future proof legislation and the appropriate role of Parliament.

I turn your attention to our discussion with the first panel about timing. When we look at what has gone before, there has been variation in practice in relation to time between legislation passing and polling day, and there has been variation in practice in terms of the referendum period. There has been discussion about whether we should or could vary the length of the referendum period, depending on what the referendum is about. What are your thoughts on the issues of timings? What are the pros and cons in respect of minimum and maximum periods for the controlled part of the campaign?

As we highlight in our evidence, a problem is that the bill does not have a minimum specified period between the designation of referendum campaigners and polling day. That means that the period could be incredibly short and that the lead designated campaigners would not be able to use their entitlements, such as a higher spending limit, for long, which would have massive impacts on the ability of the campaigners to make their cases and on public debates about the issue. That needs to be rectified in the bill. The Electoral Commission recommends that a minimum of 10 weeks be stated in the bill. That is my primary issue with the bill.

What is your view about a minimum period? Should it be four weeks or 10 weeks?

I would take the Electoral Commission’s advice that the minimum period should be 10 weeks. That has been practised for most UK referendums. As you mentioned, some periods have been longer but a minimum of 10 weeks is a good period of time.

With regard to the amount of time between passing the legislation and polling day, referendums should not be rushed. The Electoral Commission recommends that legislation should be clear

“at least six months before it is due to be ... complied with”.

If it related to the regulation of campaigners or the designation process, that would mean before the designation process began, which would be about nine months before the referendum. There is not a magic number. The Electoral Commission has a lot of experience of running referendums and I would be keen to listen to its recommendations.

I again make the point that referendums have huge potential to change the future of the country, as we are seeing right now, so maximum time should be given to make sure that they are run correctly.

Thank you. Does Mr Clancy have anything to add?

We made no particular comment about referendum periods in our submission. I would leave it to the experts.

I asked the first panel about the minimum period of at least six months between the referendum legislation coming into force and its being implemented. The distinction is that what is being discussed now is the framework legislation, which produces rules and regulations about things, such as spending, and covers all the technical aspects. If the Government has it right—and that is up to the committee to determine in this process—and the bill were to be in place by the end of the year and then there was a referendum, of whatever form, nine months or a year after that, would the framework legislation provide enough of the bedrock for that six-month rule to be the process for the future?

The Electoral Commission’s recommendation is partly so that people who are required to regulate campaigners or hold the poll are aware of the legislation that they are required to comply with. Another element is public debate and discourse, because the public need time to prepare to for a referendum. Campaign groups need time to decide whether they want to apply for designation as lead campaigners to help to organise the best debate possible. Those are some reasons for its recommendations, and there is also the case for a long period of time between a proposal for a referendum being decided and polling day.

Under the bill, such a referendum could be made available by regulations under the affirmative procedure, so we would be talking about a period of 40 days, which seems quite a short period of time for the idea to be out there. We must be very careful to ensure that there is sufficient time for people to become acquainted with it. As Jess Sargeant has indicated, the Electoral Commission has views on the issue, which we would not demur from.

That takes us neatly to a question from Adam Tomkins.

Good morning. Section 1 would, if enacted, confer on the Scottish ministers the power to provide for referendums in Scotland by regulations. What are your views on that power?

As currently framed, that power is an inappropriate way to proceed. We have indicated in our commentary on section 1—combining it with section 3, as there is a measure of crossover—that legislation setting the date for the referendum and the question or questions to be asked should take the form of an act or, at the very least, a Scottish statutory instrument that is subject to the super-affirmative procedure, but that would be a very sub-optimal position. As the phraseology “super-affirmative procedure” is bandied about, I put it on the record that the Scottish Parliament’s provisions for the procedure mean that there has to be pre-legislative scrutiny on the draft, a formal consultation on the draft order and approval in the chamber. The issue crops up occasionally.

10:45  

It was mentioned this morning that subordinate legislation is not of sufficient stature for great questions about a country. We can debate that back and forth, but the idea that statutory instruments cannot be amended is a significant feature here, which is why we draw attention in our submission to the provisions in legislation where statutory instruments can be amended—namely, the Census Act 1920 and the Civil Contingencies Act 2004. Therefore, the provisions to make amendments have to be built into the bill that would allow for such regulations to be made.

Again, if the committee were to choose to support the super-affirmative procedure, we would strongly recommend that the approach includes the capacity for amendments to be made to that subordinate legislation during its passage.

I agree with most of those comments. In our evidence, we propose that any referendum question should be in primary rather than secondary legislation. That, in part, is to allow it to go through the full scrutiny process of primary legislation, rather than the process for secondary legislation, which is slightly less rigorous. It is also because, if the statutory instrument were unamendable, as Michael Clancy mentioned, the question of whether there should be a referendum is conflated with what the question for the referendum should be.

The Electoral Commission will comment on the intelligibility of the referendum question, although there is a slight exception for questions that have been tested before, which I imagine we will talk about later. However, the commission will not comment on whether it is the right question. That job should be for the Parliament, but approving the question via a statutory instrument would not give Parliament the opportunity to look at that.

MSPs might have to choose between passing a question that they think is sub-optimal because they want to hold a referendum, and not having a referendum at all.

Thank you very much—that is very helpful. The IFG’s view is not that section 1 should be amended so that we use super-affirmative procedures, which is a suggestion made by the Law Society of Scotland, but that section 1 should be removed, so that future referendums require primary legislation. Have I got that right? I want that point to be clear.

Yes.

You may have heard my questions to members of the previous panel. Do you share their concerns about the whole area of spending limits? As I have mentioned, I do not think that a lot of people understand the terms “designated lead campaigns”, which are allowed to spend £1.5 million, “permitted participants”, who are allowed to spend £150,000, and “non-registered campaigns”, which are allowed to spend £10,000. Does that concern you? Should we be doing something about that?

We have not looked into the detail of that, which I know Justin Fisher has done. The previous panellists and the Electoral Commission are those who have had the experience of that.

I suppose that we are in the same position. We highlight in our submission that the Venice commission “Code of Good Practice on Referendums” states that

“funding must be transparent, particularly when it comes to campaign accounts.”

Again, there are experts in that field—we heard from one of them this morning—who are better placed to detail precisely what the figures, the accounting position and relative quotient should be.

Jess Sargeant has commented on the idea of public funding, so I want to ask about that. Perhaps you could explain to us why you think that public funding might be required. It appears that, for the previous European and Scottish referenda, sufficient money was raised on both sides, so there would not be a need for public funding.

You are right to say that, in the Scottish independence and EU referendums, the campaigns were able to raise sufficient funds. However, the UK regulatory framework makes provision for public funding for referendum campaigns, so the designated campaigns received such funding.

Although, inevitably, the Referendums (Scotland) Bill is being talked about in the context of possible referendums that might happen in the short term, the bill will cover any referendum that is held in the future in Scotland. Such referendums might not attract levels of donations that are high as those that, say, an independence referendum would attract. That would be particularly problematic if a lot of business groups and political parties were all aligned to one side in a referendum, in which case the other side might struggle to raise funds and put its case to the public.

Was that not the case in the European referendum? Most businesses and unions were on the one side.

I am not saying that that situation is inevitable in all circumstances. The reason why the UK framework contains provisions for public funding goes back to when the Committee on Standards in Public Life was first asked to look at the issue in 1999. It drew experience from the Welsh referendum in 1997, in which one side struggled to attract significant funding. The committee, campaigners and the public felt that that had a real impact on debate, in that people were not able to hear both sides of the argument and consider the question. As we know, that referendum had quite a low turnout, which is perhaps a reflection of the question and the demand for the referendum, but I think that, in part, it was a question of the quality of debate during the referendum.

We do not make a firm recommendation that we should definitely have public funding; that is for Scotland to decide. I note that the decision not to have public funding for the 2014 referendum followed a Scottish Government consultation that found that the public did not want it. We need to bear in mind that the bill will cover all future referendums, and we might want to think about providing for that eventuality.

I understand that argument. However, I fear that there is a lack of confidence in politics among the public. I was at a by-election the other week, and people were questioning whether any of my expenses were covered by taxpayers’ money. Obviously, they did not want them to be, and I was able to say that, basically, all the money had been donated by me. Would the downside of allowing public funding be that it would lead to a negative reaction in the public, who would say that taxpayers’ money was going to a political party or campaign?

I totally understand that argument. That is the risk with policy proposals such as those relating to public funding for referendum campaigns. However, that argument needs to be balanced against the potential for having a poor quality of debate in future referendums. As I said, we do not come to a firm view on that matter. It is one of the notable differences between the bill and the UK regulatory framework, so it is worth commenting on and thinking about.

Would a compromise position be that the assumption should be that public money would not normally be used, but if it became apparent that one side was not able to put its argument, allowing public funding would become the exception or fallback position?

That is one possible approach. The slight difficulty with it is the question of who would decide whether such funding was allocated, and whether the decision could be perceived as having a political purpose. It is a difficult issue. I do not propose to come down firmly on one side or the other, but it is worth talking about.

One of the big changes in recent years has been the growth in online campaigning. The bill makes provision for digital imprints to be used for appropriate online campaigning. There are challenges in understanding what requires a digital imprint and in monitoring that properly, and it is important that things do not get out of control. What do you see as appropriate online uses that require digital imprints during a referendum?

At the risk of repeating some of what has been said, I think that there is a distinction to be made between paid political advertising and organic advertising. It is clear to me that online advertising that has been paid for—in parallel with published advertising that has been paid for—should definitely contain an imprint.

I think that organic advertising that is shared peer to peer on social media should not require an imprint. As the Electoral Commission said, there is still some thought to be given to the issue, and there are difficult trade-offs to be made about the regulatory burden. As was mentioned by the first panel, the Scottish 2014 referendum was the first referendum in which imprints were applied to digital materials, and we should learn as many lessons from that as we can.

In our submission, we highlight the report of the UK Parliament’s Digital, Culture, Media and Sport Committee on online harms. In that report, the committee makes various recommendations to the UK Government, which so far have not been taken up, about making sure that a category relating to digital spending on campaigns is introduced and recommending that information about all political advertising material should be searchable and public.

This committee heard evidence from the first panel that the UK Government has indicated that it will make a technical proposal for a regime on digital imprints later this year. That is clearly gathering a head of steam and we recommend that the Scottish Government takes it into account.

Patrick Harvie, you are interested in database issues.

I think that both witnesses were in the room when we were discussing the issue with the previous panel, but the concerns that I have with the idea of a national database are partly practical—is there is a danger that there could be a mismatch between a national database and the locally held register, and whether there could be any practical consequences of that. My concerns are also partly political—what risk to the perception of democracy would we be running by creating a national database, given the concerns and controversies that have arisen since the national identity database, the NHS central register and the proposals from the Scottish Government regarding the national entitlement card? Some of those schemes have been very controversial, so would this scheme risk running into the same issues of perception?

Inevitably, there will be people who will contest any change in that area, particularly if it touches on the exposure of details about individuals. The important thing is that, at the moment, we are talking in theory without any concrete proposal, so if one was going to say anything to those who are going to create concrete proposals, it would be to make sure that a proper balance is set between protecting data and, at the same time, allowing scrutiny and accountability. Beyond that, it is quite difficult to say much more.

Presumably, the concern about the objective of enabling campaign organisations to check the permissibility of donations is one that we all share. Do either of you have suggestions about other ways of achieving that objective that do not necessarily go down the route of a national database? Is that something that you have looked at?

Unfortunately, it is not something that we have particularly looked at, so I will defer to the experts on that one.

Not to worry. Thank you.

I have a question about the role of the Electoral Commission in relation to question testing for referendums. Both the Law Society of Scotland and the IFG have expressed concerns about that. The IFG specifically talked about the proposed exemption and said that

“there is only one circumstance in which this exemption would apply”,

that being if the 2014 independence referendum question were asked again. The exemption would seem completely unnecessary, given that the bill is meant only to cover devolved issues. Jess Sargeant, would you like to elaborate on your views on that provision?

Obviously, question testing by the Electoral Commission comes with cost and time implications, but it is considered to be one of the most rigorous processes of question testing in the world. From all previous referendums, we can say that, although people might have problems with the reason for the question being asked, everyone has faith that the question that they are being asked is not biased and is fair—there is consent between both sides on that. There may be reasons for not testing again, such as timing or cost, but I think that they are trumped by the legitimacy aspect of testing.

11:00  

There is no reason why a question that has been tested before cannot be further improved. More experience of the question or debate might mean that new issues come to light. It is important to test the question in the current context, because things might have changed. Question testing takes about 12 weeks and the cost is quite marginal relative to the whole cost of a referendum. The reasons for exempting certain questions are not strong enough to justify the inclusion of the exemption.

We have criticised section 3(7), which provides that the section would not apply to involvement of the Electoral Commission

“in relation to a question or statement if the Electoral Commission have—

(a) previously published a report setting out their views as to the intelligibility of the question or statement,

or

(b) recommended the wording of the question or statement.”

Again, it is a question of context. Is a question set once a question set for ever? Have conditions changed? We think that questions should be scrutinised in the light of the conditions at the time when it is proposed to pose the question.

The Electoral Commission uses stronger language:

“The Commission firmly recommends that it must be required to provide views and advice to the Scottish Parliament on the wording of any referendum question”.

The commission is the expert in this field and we defer to it.

A live example of a question being improved is the one that was initially proposed in the European Union Referendum Bill in 2015. It had been previously recommended by the Electoral Commission, based on question testing of a private member’s bill. The question was:

“Should the UK remain a member of the European Union, yes or no?”

Although the question had already been tested, the Electoral Commission tested it again. In that process, it found that some campaigners had concerns about “remain” being in the question and they recommended that the word be changed to the “remain/leave” format that we ended up with. There are opportunities for further improvement, and something that looks like an attempt to evade that could be particularly problematic for a future referendum.

The Electoral Commission suggested the question that was used in 2014.

Yes—that is the point.

If the Electoral Commission were to test that question, would you expect that the process would look at the circumstances, including the fact that people in Scotland are familiar with the question and that it is regularly used in polls? Given that backdrop, would that be part of the testing? To introduce something different now might confuse people. Obviously, I will put that question to the commission as well.

I would certainly expect the commission to do that. Part of its question-testing process is to assess intelligibility, which includes putting the question in front of people and getting them to vote. The process is all about whether people understand the question that they are being asked and give the answer that they want to give to indicate their preferences. The commission does that by testing a variety of questions with people. If its evidence showed that people were confused by a referendum question that asked the same question, in essence, as the 2014 referendum question but was worded differently, it might recommend that the same question be used.

I will wait to find out what the Electoral Commission says in answer to that question.

Over the past 20 years, the Political Parties, Elections and Referendums Act 2000 has been in place and we have had three referenda with a 28-day purdah period in place. What is your view on the suggestion of a longer purdah period?

The recommendation of the Electoral Commission and the independent commission on referendums is that the purdah period should be extended to cover the whole of the referendum period rather than just the last 28 days. The IFG does not have an institutional view on that.

There is a slight added complexity when two Governments are involved in referendums. There would need to be an agreement if the purdah period were extended by one Government, as it is likely that you would also want that to apply to the other Government as well. Obviously, the difficulty is that Scottish law could not bind the UK Government, so it would have to be in the form of some other agreement.

As I say, the IFG does not have an institutional view, but an extension of the purdah period has certainly been recommended by people who have looked at this quite thoroughly.

We have not taken a view on the purdah period at all.

Is that because you are broadly satisfied with the existing rules?

No, it is because we have not considered it as being an issue.

The Electoral Commission and the independent commission on referendums recommended that the period should be not only lengthened but narrowed in scope. At the moment, the purdah period applies to all publications relating to the referendum which, for a functioning Government, could be difficult to sustain for a longer period of time. You would need to look at both those recommendations and generally review the purpose of the period and whether it is doing exactly what it was intended to do, which is ultimately to prevent Governments from using public funding to intervene in referendum campaigns.

I suggest that the committee works with the Electoral Commission—essentially, with the experts—to see how that might best be achieved.

Having that guidance in place, you said that there would need to be some form of agreement between the two Governments in any referendum where there was interest from both parties. What if that agreement was not forthcoming?

That is a difficult question and it is not one that I have an answer to. I would hope that in good faith, there would be some sort of agreement. For example, in the 2011 Welsh referendum, the Electoral Commission asked the Welsh Government and the UK Government to extend the 28-day purdah period to cover the whole of the referendum period. The Welsh Government did that; the UK Government did not but it agreed to remain neutral and it did not take a position in the campaign, so there are alternatives. Agreement might be slightly more difficult in other contexts but certainly, in the case of a second independence referendum, some form of agreement would need to be discussed as part of the negotiations on the process.

Okay.

I thank our witnesses for coming along today and providing us with some useful evidence.

11:07 Meeting continued in private until 11:43.