Only a few days to go: We’re raising £25,000 to keep TheyWorkForYou running and make sure people across the UK can hold their elected representatives to account.Donate to our crowdfunder
My name is David Rossiter. I do not want to outdo Charles, but I have studied and published on the process of redrawing boundaries for about 40 years. I was the lead researcher on a Leverhulme-sponsored study on the work of Boundary Commissions in the 1990s, and was responsible for much of the modelling for the McDougall Trust report on the impact of the Parliamentary Voting System and Constituencies Act 2011 in 2014.
Q Thank you both for joining us. I salute you for your combined seven or eight decades of work on these matters. It strikes me that you are extremely well placed to help us have a very down-to-earth conversation, and to remove some of the high-falutin’ terms that get used sometimes in these matters.
I noted that you and your late colleague, Professor Ron Johnston—we send you our condolences on his loss—looked into claims of bias in prior reviews. You were very clear that there is a function here for levelling the playing field by ensuring updated and equal boundaries. Could you please go into that?
Thank you for your words on Ron. Do not take it amiss, but I think both David and I would, in some respects, prefer it if Ron were here to talk to you in person. I mean that in the best of possible senses.
Your question about bias is very interesting. Obviously, it has been the cause of some concern. There has been a particular party political concern about the extent to which the system has become substantially biased in Labour’s favour. Part of the concern is around constituency size effects, which the current legislation and the 2011 Act deal with.
You heard earlier today—I think John Curtice also discussed it this morning—that there are two things to bear in mind. First, we are talking about bias between Conservative and Labour. As long as we have a first-past-the-post system, there is in-built bias against small parties with equal vote shares. The Conservative-Labour bias in particular does have an element around the constituency size effect, which the legislation largely removes. Most of the bias that has caused comment and concern in recent years has come from other sources that are nothing to do with the constituency size issue. They are to do with things such as preferential abstention rates, third-party effects in different seats, and in particular the efficiency with which parties’ votes are spread.
In the last few elections—every election since 2015—the relative Conservative-Labour bias has run in favour of the Conservative party and not Labour, largely because the Conservatives have become much more efficient in how they campaign and where they win the votes. To that extent, the legislation deals with one of the sources of bias. However, as a few witnesses this morning pointed out, that is one of the smaller components of the bias picture, and the bigger elements of bias are not really dealt with by this legislation—and I suspect cannot be dealt with by any legislation.
The change to a UK-wide quota quite clearly deals with the fact that there were higher levels of representation in Scotland up to 2005, and still are in Wales. If you look back to when the current constituencies —the ones you are representing—were first defined using 2000 data, there was no bias at that time in favour of either party in terms of the size of the seats. The 10 largest seats defined at that stage included Hornchurch and Upminster as well as Croydon North; one was Conservative, one Labour. If you look at the 10 smallest seats, again, there is a completely equal mix. So for every Hexham, there was an Islington South and Finsbury. It is not that the commissions were unable to provide equality at the date of enumeration—that is, the date they have to work to. It is the demographic change that took place in ensuing years that has caused the big disparities that were more evident in the 2005 and 2010 elections than in 2000.
That demographic change was already slowing down in the 1990s, and over the past decade it has effectively ground to a halt. That process is no longer continuing. From that point of view, the pre-2011 legislation was able to deal with an awful lot of the difficulties that come from differently sized seats. The issue was: how, if at all, can you deal with the fact that certain areas grow in size and certain areas reduce in size? Reducing the period between reviews—the Bill suggests eight years—seems the best way to achieve that.
Q Thank you very much. You concluded with an argument in favour of regular reviews and, I suppose, getting on with it. As you pointed out, the age of the data that currently holds sway is in itself an argument for moving ahead to the first of a new series of reviews, and establishing a series from there.
Q I would like to start by passing on my condolences to you both after the unexpected death of your colleague Ron Johnston. There is an argument to be had about where the balance lies between drawing constituency boundaries that look like the communities that people recognise around them, and the electoral quota and the flexibility to stray either side of it. The Bill proposes a variance in the electoral quota of 5%. What do you think the number should be to strike that balance between community and constituencies of equal size?
I guess we can break that down into two constituent parts. One is whether we should have a principle of priority within the rules, as in the 2011 Act and in the Bill, with some notion of equalisation of electorates being the top criterion rather than the medium criterion, to avoid some of the confusion and tension of the earlier rules. To that extent—Dave may feel differently about this—I would certainly endorse the notion of having an equalisation rule as the top priority.
The second element of this is where to draw the tolerance. Should it be 5%, 1% or 10%? On that point, I think you have a rather more open debate on your hands. Dave referred, when introducing himself, to the work that we did for the McDougall Trust in 2014, looking at the process around the sixth review—the first under the 2011 legislation. In that work, we tried to estimate how much disruption different tolerances would cause in the system—how much breaking of ties and breaking up of existing seats there would be. Inevitably, there will be quite a lot, both in the first review under the new rules and in any subsequent revision. However, on our estimates, if you set the tolerance at around 7%, 8% or 9%, disruption is reduced, and you do a better job of maintaining existing ties and links.
Yes, equalisation is important, but the question is what tolerance you should work to, and how wide you set that tolerance. Our estimates suggested that 8% starts to get you into the compromise zone and makes life a bit easier.
Yes. I am afraid that it is probably a rather technical point, but it is quite important, in terms of the effect that the rules will have on future reviews. The 2011 Act created the UK quota and laid down the rules for allocations to countries and regions, but if we look at registration statistics over the last 20 years, we can see how those national and regional entitlements vary over time. We know that in an average eight-year period, we would be likely to see about eight changes to either national or English regional entitlements—that is between each pair of reviews. With a fixed Parliament size, that would necessarily mean that four new seats would be created in the UK and four abolished.
In the case of an abolished seat, you will have to redistribute 60-odd thousand electors to neighbouring constituencies. That in itself will take most, if not all, of those neighbours over quota. Any seat over quota will need to lose one or more wards to compensate for the addition. The process continues in this way, much like ripples on a pond, until all seats are within the 5% tolerance. Several of the affected seats will need to become participants in the process, even though they were within quota; they act merely as transit stations.
You can think of the scale of the impact of this process, which is required by the 2011 rules, as inversely proportional to the level of tolerance. As a rule of thumb, which is always useful in such circumstances, dividing 100 by the level of tolerance give you a rough idea of the number of seats that will be affected. By contrast, under the previous rules, which allowed the commissions far more discretion, the process would affect just a handful of seats and would typically be contained within a county. In the fifth review, Cornwall gained a seat, but that had no knock-on effect whatever on Devon. That is simply not possible under the current rules.
If we assume eight changes of entitlement in eight years, and if we take the existing 5% tolerance, the rule of thumb would suggest that, every eight years, 160 seats will require significant and often major change. To that has to be added the 100-plus other seats that have drifted outside what is a much tighter quota than has ever existed before. This is something that I have not heard mentioned as part of what the 2011 Act effectively ensured. The critical point to take away is that the interplay of the rules with such a tight tolerance will effectively guarantee a major redrawing of constituency boundaries at every subsequent review.
Q Dr Rossiter, I hope I am not putting words in your mouth, but would a way to solve that involve some level of tolerance over the total number of constituencies that need to be drawn up? The Bill fixes that number at 650. Is there an argument for giving the commission the flexibility to go as close to 650 as possible while respecting community ties?
Back in 1998, we wrote a proposed new set of rules that would have achieved what I think would have been a rather better way to work—I would say that, wouldn’t I? We felt at the time that the differences between national quotas, and the discrepancies between constituencies across England, were too large. We suggested that a new set of rules could say, “Yes, we’ll have a UK-wide quota, and we will have a target size for Parliament of whatever number of seats you wish.” It is 650 in the present case. We then said that a commission should be restricted to no more than 10% variance around the UK-wide quota, but that it should aim to get constituencies as near to that quota as was practical. That would give commissions the extra latitude that they would need to avoid many of the difficulties that were so evident in the 2013 and 2018 exercises.
At the same time, we would make it clear that electoral equality is a very important thing to aim for, and it should be the goal in all circumstances. I believe that having a degree of flexibility is extremely important, and I fear that not having it will inevitably cause consequences further down the line.
One of the areas that I was quite pleased to see in the Bill was a re-examination of how the inquiry and hearings are held, because that is problematic.
However, there is still a bit of a challenge for the public hearing process, because the areas in which those hearings now operate are just so incredibly large. There was some discussion earlier in your deliberations about ways in which the process might be improved to allow greater flexibility in local discussion. But you must remember that you are talking about entire regions, and about entire countries in Scotland and Wales, and people can turn up at a hearing in one corner of the region or country to talk about a seat in quite another part, and the chances of having a meaningful conversation about those proposals are remarkably small.
I am not sure that I have a clever proposal for you, but I think that is something to worry about; the extent to which those hearings really produce helpful information in all bar a few cases would be a concern that I have. I cannot suggest a fix for you, but if you want to look at something, that is another area that it is worth just having a bit more thought given to it.
Q I suppose that this question is for both of you. If you think back to previous reviews that have taken place—admittedly, obviously, under different rules—to what degree of magnitude do you think this review will end up changing the existing constituencies?
Big is the very short answer. This is liable to be one of the most disruptive reviews that we have seen for quite some time. As Dave mentioned earlier, on our estimates you are looking at major disruption again, and again, and again, into the future, especially if you hang on to that 5% tolerance. So, this will be big. Further reviews will also be big, so this will become a feature of the system going forward.
If I can just add to what Charles has said, when we did our 2014 exercise we estimated that approximately half of seats would experience major change at this first review, but we based that on 2010 data, because that was the data that was available at that time. So, we were looking at rectifying changes that had taken place over 10 years, plus the change to the rules. We will now be looking at an exercise that has to rectify the changes over 20 years and I think that we will be looking at something like two thirds to three quarters of seats experiencing very significant change at this coming review.
Contrary to what I think are some of the optimistic views that were expressed earlier in proceedings, I see little chance of county boundaries remaining intact in large parts of the country. I think that most county and unitary authority boundaries will need to be breached. I also think that many more constituencies will be split across local authorities, and vice versa, and many more seats will have orphan wards in them.
Again, looking at this in an historical context, there have not been that many reviews that have had to deal with 20 years of changes, so it is probably not too helpful to concentrate on the disruption this time round; it was always going to be like this. I think that what is much more worthy of consideration in terms of legislation is realising the longer-term implications of it, because the danger is that if these changes are not realised, you only have to go back to the 1954-55 debates in Parliament, when MPs suddenly realised what had happened in the previous legislation and said, “We do not want our constituencies changed on this basis. Why are we having all this change?” Four years later, legislation was introduced to reduce the need to change to meet an arithmetical standard. My fear, obviously, is that that will be repeated.
Q That is interesting. Out of interest—I could go and check this now, but I do not have it in front of me—on the 2018 review, which obviously did not happen, for various reasons that we have discussed already, what percentage of seats underwent what you would consider major revision?
I do not have that figure to hand. One of the problems is that this affects different parts of the country differently, so, for example, during the 2018 review, the south-east of England was little affected because it was set to lose only one seat during that review. Now that we go back to 650 seats, because of the growth in the south-east of England, the south-east will gain seven seats. Gaining seven seats inevitably results in a huge amount of change.
So, it can be helpful to look at what happened in 2013 and 2018 as exemplars of what results from this, but this is the problem: the devil is always in the detail. It is always in the specific geography of the area. It is always in the specific number of electors—whether a county, for example, has an integer entitlement or a non-integer entitlement. I have near me the example of East Sussex. East Sussex at the moment is entitled to eight and a half seats. With a 5% tolerance either way, that will mean that the East Sussex boundary has to be bridged. Kent is perfectly okay. West Sussex is perfectly okay. Therefore, in sorting out the problem in East Sussex—this is all provisional on 2019 data not changing an awful lot—we will need to see something that goes across the county boundary in one way or another.
Until we know the final figures, we will not be able to be absolutely certain on any of these issues. At least half of seats were changed during the 2013 and 2018 reviews, and when I say that the forthcoming review would be between two thirds and 75%, that is simply a reflection of the fact that it is trying to deal with that extra amount of time. What seems surprising is that maintaining 650 seats does not necessarily help a huge amount. It helps slightly, but not a great deal, in minimising the disruption that is going to happen. I hope that that is helpful.
Q Thank you very much, Professor Pattie and Dr Rossiter, for coming before the Committee. I have a couple of questions that I want to explore with you. You may have seen in previous oral evidence there has been some discussion of the idea of the building blocks for constituencies, whether those are used by polling districts or wards. Can you offer a view on that? Perhaps Professor Pattie would start off.
This is an interesting issue, isn’t it? The issue here again is obviously over, partially, the practice of splitting wards—which clearly can be done—and partially the pragmatics, if you like. I know you have had lots of evidence already about data sources, software availability, etc. I will leave that to people who are more expert in handling those data systems, but clearly that causes an issue. I think I would raise just two points, here. First of all, harking back to our 2014 McDougall Trust report, we did try there to estimate the relative effects on disruption of playing around with the tolerances versus playing around with ward splitting. Ward splitting certainly helped to reduce the amount of disruption, but in our estimates it did not reduce disruption anything like as much as widening the tolerances moderately. The second thing you have to bear in mind here is that we are talking about disruption to communities. Remember how the Boundary Commission’s local government wards operate. It tends to be quite strong on the idea that, in building the ward suggestions, it is trying to represent people, so when you split a ward, arguably you are splitting a community—you are doing the very thing that you are trying to avoid, to avoid the thing that you are trying to avoid, if that makes sense. You end up in a strange circular process in which you disrupt a community to save a community. Where the white line is on that is anyone’s guess, but ward splitting is neither technically a global panacea, nor conceptually a panacea, precisely because in splitting a ward, you might well be splitting a community.
Q Continuing the theme of geography, although I appreciate that you will not necessarily have the amendment paper in front of you, I have tabled new clause 5, which looks specifically at the highland constituencies and that limit of 12,000 sq km. I have asked this question of other witnesses before the Committee. Can you offer any thoughts on ways in which to manage constituencies so that they are slightly more manageable for Members? I think that most people would agree that having a constituency of 12,000 sq km is somewhat unsustainable. In my name, I have tabled a new clause to say that it should be 9,000 sq km, for example. Do either of you have a view of that, in terms of the management of constituencies?
At the risk of sounding flippant, the Durack division in Western Australia is 1.63 million sq km. The north highlands is large, but there are much larger seats out there. It is how you strike the balance, I guess, but where it is can be tricky. I would not want to minimise the workload of an MP, in particular working in any area as large as the north highlands. Where one draws that line is a judgment call. I do not think that you will find an easy answer. To use a phrase much bandied about at the moment, I do not think that this is an area where one can defer to the science, because there is no clear science to this.
Q This is for both our witnesses, but I will start with Dr Rossiter. Do you agree that reaching electoral equality is important not just between regions but within regions? I will take the example of between regions first.
At the moment, Wales has an electoral quota of about 54,500, as opposed to about 72,000 in the north-west. Within Greater Manchester, where I am an MP, the number ranges from about 63,000 to 95,000. To take the concept that you just put forward of not splitting communities, in my borough are two seats that are prettily evenly divided: mine is Heywood and Middleton, and the neighbouring one is called Rochdale. From the sound of things, they are self-contained communities, but, in reality, I represent about a third of Rochdale. If you were not to split the communities, my neighbour would represent 103,000 people to my 57,000. Taking that to the logical extreme, do you not accept that, at some point, you will have to split some communities in order to achieve electoral equality?
Beyond that, talking about disruption in future reviews, would you accept that, to a degree, splitting wards would minimise that, reducing the amount of absolute disruption? Most of the disruption that will come from this review relies on the fact that the electoral figures we are using are 20 years out of date.
If I take your second point first, I do not think that the difficulties that are going to come with the current review will be of such a scale that anything really can be read into them—too much should not be read into that, if you see what I mean. To take your first point, the commissions have always been capable of producing constituencies that are very close to quota. The problem you are identifying—these large differences in constituencies—has largely come not because of an observance of local ties, but from demographic change within and between regions. I am totally comfortable with the concept of trying to achieve equally populated constituencies—I have always thought that should be aimed for. My concern is the unintended consequences of a set of rules, which I think is the territory we have entered.
In terms of principles, absolutely every person’s vote should be treated as equal in so far as that can be achieved in a constituency-based system. There is no reason why either between or within areas that should not be achievable. Where local authority boundaries have to be crossed to achieve that, I have no problem with that. I remember writing a paper back in the 1980s about how we needed to look at crossing London borough boundaries, which were being observed as almost sacrosanct at the time, causing quite significant difficulties and an over-representation effect.
What I think we are looking at is how you strike the right balance. I do not disagree at all with where you are coming from and what you are trying to achieve; it is just that by placing in a rule as strict as 5%, you are removing a degree of discretion that will not benefit anybody either politically or in their sense of connection with a constituency and their MP.
To add to that, the point I was trying to make earlier was not that one must never split communities. That is going to happen, and it always has happened under the boundary review process; there have always been communities split. My point is to recognise that splitting wards in itself is not a solution, because that may involve another form of community split. But we must also remember—Iain put this nicely this morning when he described the different directions in which community can run, depending on how it serves different people’s interests—that community is very much in the eye of the beholder. I am sure we all recognise, even in areas that we know well, that we could quite quickly generate quite a few different views of what a local community really was. They are often genuinely held. So, one should not be too—how can I put this?—precious about community versus size. I think David is absolutely right: the issue is where to strike the balance and how one achieves that as relatively painlessly as possible.
Q To take that thread, if we are not being too precious about communities, why is a 7% or 8% variance better than a 5% variance? Surely it is better to get closer to the mean.
Well, you still have equalisation and a fairly tight parameter in terms of the size of seats, but one does not have to artificially flex things too much. You are trying to strike the balance between the rules of equalisation and rule 5 conditions. One is trying to hit that balance point between equal electorates and not too much disruption.
Q I take the academic point, Professor Pattie, but I think it would be quite a hard sell on the doorsteps to tell some of my constituents that 20,000 extra voters are required because it will save somebody a bit of work.
Q I am not sure which of the two of you I am aiming this question at, but how much does locality and the experience of living in a community influence the way people vote? Does it bring outcomes where people vote collectively in a similar pattern?
You heard evidence from John Curtice this morning on this and I would not disagree with him. There certainly is evidence that people are influenced by the context in which they live and by what is happening around them both in terms of the economic and political environment and in terms of the climate of opinion around them. People who in a sociological sense look very similar, but live in different areas, can go in very different ways much more akin to other people within their area. Is it the biggest influence on people’s voting? No, probably it is not. Does it have an effect? Yes, it does.
Q Could it lead to frustration? We do not keep clearly identifiable communities with common characteristics integral within parliamentary representation, but their voices could be lost because they cannot vote collectively in response to the experience of living in their particular locality.
That is rather harder to argue, to be honest. The extent to which people would see themselves as acting for their local area in a constituency sense is quite a hard one to argue. People have a sense of “my area” [Inaudible], but is that the constituency? That is much less obvious.
Over the years, for my sins, I have attended an awful lot of what were local inquiries and I have listened to a very large amount of evidence put forward about local ties. I tend to agree with Professor McLean, who gave evidence this morning, that one person’s local ties go in diametrically opposed directions to another person’s local ties, depending on their political preferences, so I am not at all upset at the idea that arguments in terms of local ties might take a lower role in the hierarchy. In fact, I think that that is a sensible thing.
What I do think—this also goes back to the previous question—is that where you have got local government boundaries and existing seats, you have inconvenience when you cross those. I listened to the evidence earlier from the person representing the electoral registration officers and I have also heard evidence given at inquiries from Members of Parliament who have repeatedly referenced the difficulties that they have when they have to deal with multiple local authorities. People deal with four. or, as recommended in one of the recent reviews, five local authorities for one Member of Parliament. That aspect of discretion is something that the Boundary Commissions over years and years have shown great ability to recognise. Again, I come back to my point: that is where their discretion and their ability to address those concerns is being curtailed.
Q I was going to come on to that, so you have segued to it nicely. To come within the 5%, or whatever we finish up setting in this piece of legislation, requires either taking a piece of a ward—going to sub-ward level within a local authority area—or going across that local authority boundary. Would you suggest that it is better to go to a sub-ward level and stay within the local authority area, rather than having constituencies span two or more local authority areas?
Speaking personally, it would depend on the evidence in the particular case. I do not think that one is necessarily better than the other. I have noticed, when we have been looking at this, the significant help that increasing that tolerance by very small amounts will provide. As soon as you go from 5% to 6%, you have a big payback from going up by that one percentage point. That payback increases to around 8%, which is why we came to the conclusion in our previous report that a figure of 8% would be much more helpful. Beyond 8%, the advantage begins to flatten off, because you are reaching a point at which any sensible commission can reach solutions.
In all this, we accept and understand entirely that the 5% introduction was not an attempt to be cussed or anything of that nature. It is simply that 8% and 5% are not worlds apart. If you are able to achieve far more when you adopt one rather than the other, you have to wonder why you would want to go for the lower figure, unless there is some major negativity in that regard. Again, as people who write published papers, we have to do our research, and we have looked for anything that would support 5% in any of the previous discussions regarding the 2011 Act and so on, and we have struggled.
I know you have had reference to the standards related to the “Code of Good Practice in Electoral Matters”, and there seems to have been some confusion over what that says. I am not sure whether that confusion has been sorted out; I was very surprised by what I heard the other day. I think there is probably an understandable source for this confusion, because an earlier edition of an OSCE publication did indeed say that a 10% tolerance—quite reasonably taken to mean no more than 5% either side of the norm—should be aimed for, but that was never referenced in that version of that booklet; a subsequent edition of that observer handbook has come out, and that reference is no longer in there.
Probably the best statement of what is best in this area is the OSCE’s “Guidelines for Reviewing a Legal Framework for Elections”, which specifically endorses the “Code of Good Practice” and states that proximate equality—no more than 10% between electorates—should be the aim, but interestingly goes on to say that
“frequent changes in the boundaries of constituencies should be avoided”.
If we are looking for international standards on this, there is a clear suggestion that going right down to 5% is not necessary, and in so far as it causes change to boundaries, we would not fall foul of OSCE reports. They all seem to find that UK elections fail in one respect or another, but at least we would not fall foul on that.
Are there any other questions from Committee Members to put to our two witnesses? If there are no other questions, I would like to thank Dr Rossiter and Professor Pattie for the evidence you have given us this afternoon. We are very grateful. I am grateful to Members for their cooperation during this virtual session in these somewhat unusual proceedings.