Boundary-drawing policy issues

[Adrian]

Hmm. But because it's possible it doesn't mean that you should do it. That implies that there is something sacrosanct about wards, which there isn't. You're committing a kind of cart-before-the-horse fallacy of logic. As you're already aware, my view is different.

What are the advantages of using wards as building blocks?
- convenience

Er, that's it.

What are the disadvantages of using wards as building blocks?

1. In several cases, ward boundaries have already changed, and in lots more cases they will change during the lifetime of the parliamentary boundaries. So even the convenience argument does not hold in many cases. If the coterminousness of boundaries was so important, current (2016-2017-2018) boundaries should be used in the Review, and interim reviews should take place when ward boundaries change.

2. Ward boundaries are put in place for a particular purpose: local democracy. They are not intended to be used for other purposes.

3. Because they have to be of a certain size, ward boundaries often cut across communities and/or combine disparate areas. A strict adherence to ward boundaries by parliamentary boundaries compounds this problem. This is a great pity since a flexible approach in the Review can allow these problems to be mitigated.

4. Because of the arithmetic, strict adherence to the use of wards as building blocks makes it much more likely that there will be cross-border constituencies, orphan wards, and seats that split communities and/or combine disparate communities.

[islington]

Adrian

Thanks for this. We all approach this process according to our own lights, and (fortunately) there's no need for us to hammer out an agreed position. But I suspect we might not be so far apart as implied by our two posts above.

You refer to cutting across communities (either by splitting them or by combining disparate areas). And if this is the paramount consideration, then I agree that using wards may sometimes get in the way of it - as, indeed, may LA boundaries, which in some instances fly completely in the face of natural local ties.

But the legislative framework doesn't give local ties the same degree of eminence: it is more interested in things like LA boundaries (including wards) and minimizing change - it is in the latter context (and only there) that community ties get a mention (the gist being that if it's necessary to change a seat, try to do it in a way that doesn't break existing ties within the seat). So in legislative terms, community ties are a second-level consideration, whereas LA boundaries (including wards) and minimizing change are first-level considerations.

That said, the rules rightly allow the Commissions a high degree of discretion. So, despite the fact that the legislation makes no mention at all of a positive requirement to create seats that reflect local ties (as opposed to the negative requirement to try to avoid breaking existing ties within seats), it is open to a Commission to pay attention to local ties in drawing up a scheme and, as a matter of common sense, I am sure the Commissions do this where they can. But it's one thing to take community ties into account as a factor when creating a seat, even though the legislation doesn't specifically provide for it; it's quite another to give it priority over other factors for which the legislation does specifically provide. What I've sought to do, therefore, is to create seats that reflect local communities whilst also avoiding unnecessary divisions of LAs and wards and minimizing change. Much of the time, notwithstanding some notorious exceptions, most LA and ward boundaries, and most existing seats, do broadly reflect local ties; so all these aims go hand in hand and it's usually possible to achieve them all simultaneously. But where there's a conflict between them, I feel priority should be given to the considerations that are specified in the legislation (respect for LA/ward boundaries and minimum change) as opposed to one that isn't (a positive obligation to create seats that respect community ties).

You could say, if you like, that this is a relatively narrow view, based on the legislation. Fair enough: but if you prefer a more principled argument, I'd certainly want to make the case for drawing boundaries based on some kind of pre-existing unit, as opposed to a more 'free-hand' style of boundary that takes less account, or no account, of existing units. My concern here is that a 'free-hand' style opens up a virtually unlimited range of options and thus is much more open to subjective judgments, and possible political manipulation. I'm not saying it would necessarily lead to the kind of grotesque computer-generated gerrymanders that we see in the US, but I'm worried that it might open the door to moves in that direction - after all, boundaries of the type seen in some US states would not be possible if there were an obligation to stick to existing wards, or counties, or some equivalent unit. (There would still be gerrymandering, of course, so long as legislators are directly involved; but it would be less flagrant and the scope for it would be much more limited.)

There is also a question of transparency and public access. A system that works by combining existing units (such as wards) is relatively easy to understand and interested members of the public, such as ourselves, can work up plans that can, if necessary, be promoted in opposition to the suggestions of professional experts in the Commissions or in political parties. 'Free-hand' boundary drawing, however, would require much more granular data and more sophisticated mapping, so it would become far more difficult for anyone other than the experts to take part.

Wards seem to me, on the whole, to be the most suitable 'pre-existing unit' for our purpose. They are, as you say, devised for an electoral purpose (local democracy); so it does not seem unreasonable to utilize them for a different (but not dissimilar) electoral purpose (national democracy). And they have the great merit, compared with some rival units, of uniformly covering the entire country. But I think there would be a decent argument for using civil parishes, if they covered the whole country instead of mainly rural areas only; and there is certainly a case in Scotland for using community council areas rather than wards. Parishes and community councils have the advantage, over wards, that they are not subject to such frequent change, because I certainly agree that the use of wards presents a problem when the boundaries alter. This is, though, less of an issue now that seats are to be revised during each Parliament: although it will obviously be inconvenient for Birmingham (for instance) that the next general election (if it takes place on schedule in 2020) will be fought on boundaries based on wards that are no longer in use, the difficulty will apply for one GE only; whereas in the past, with much less frequent reviews, a mismatch between wards and constituencies might have persisted for decades. 

Finally (sorry for a long post), I ought to add that I recognize that there is an arguable case for ward-splitting in a relatively small number of cases: Glasgow, Edinburgh, Sheffield and Brighton for instance; possibly Birmingham and maybe one or two others that I can't call to mind at the moment. But since we now know that we can, without a single ward split, draw up 600 seats that are, at worst, serviceable*, then that's what I'd like to propose; whilst leaving the door open for respondents to make the case for ward splits if they wish. (And the test should be whether the split allows a substantial (not marginal) improvement on the best possible non-split plan.)

* My definition of 'serviceable': no worse than the least satisfactory existing seats (see separate thread for suggestions of which these might be).

[John Chanin]

I think we all agree that wards should be respected. We don't agree that this should trump sensible seats in community terms. We all have to accept that "minimum change" is the first thing in the Boundary Commission's mind, although we all grumble about it. So perhaps minimum change is the best argument for ward splits.......

Where wards are very large (Birmingham, Leeds, Sheffield, and of course the multi-member Scottish wards) forcing seats into a procrustean bed, as the BC did last time was extremely unsatisfactory. This is particularly the case for Birmingham, where new wards will be announced the same week as new parliamentary seats. Therefore sticking to old ward boundaries will not only produce absurd seats, but will of necessity cross the effective ward boundaries in literally dozens of places. Sheffield would be the same if old wards are used (although since the order introducing the new wards was prior to December 2015, I assume the new wards will be used here).

The more interesting argument for me is about seats that include wards from more than 2 local authorities. The BC has been extrememly and in my view unnecessarily hostile to this. The issue is whether the overall pattern of seats in an area produces equal size seats and reflects sensible communities. Otherwise why bother with a FPTP system?

[Adrian]

I was of course being somewhat disingenuous when I dismissed the use of wards with the single word "convenience"; convenience is an important factor, and several important things come under that heading, most of which Islington has reminded us of. And he is also right to remind me of the legal situation, which the Commissions are obliged to stick to rather more strictly than I would wish, although, since ensuring that the redistricting is legal is their job, people like us should not shy away from submitting the proposals that we would actually like to see enacted, so that they get an accurate view of public opinion, rather than second-guessing what the Commissions might find acceptable.

One thing I'd correct him on though is regarding minimum change. Although this is normally an overriding factor in their eyes, this first review since the change of law is somewhat different since they do have more latitude than usual, and they clearly demonstrated this at the zombie review.

[jamie]

Adrian What would it generally take to override minimum change? Is there a recognition that with the big changes in some regions allowance of seats there will be significant change? As someone who has focuses on making good seats (community, geography, local authorities etc), I have usually created seats with significant changes, sometimes creating seats with only around 50% of a previous constituency within them. Would this basically be a non-starter?

[islington]

The actual rule is as follows:

"A Boundary Commission may take into account, if and to such extent as they think fit—
(a)   special geographical considerations, including in particular the size, shape and accessibility of a constituency;
(b)   local government boundaries as they exist on the most recent ordinary council-election day before the review date*;
(c)   boundaries of existing constituencies;
(d)   any local ties that would be broken by changes in constituencies;
(e)   the inconveniences attendant on such changes."

The 'minimum change' elements ((c) to (e)) are the same at this review as at any other. I believe the Government originally proposed to suspend them, for this review only, on the grounds that the new strict numerical limits, and the cut in the number of seats, meant that the Commissions needed to start with a clean slate. But this provoked a backbench revolt and the Government backed down. So, in theory, the 'minimum change' rules have the same force as usual.

I say 'in theory' because, of course, in practice the features newly in force for this review mean that the Commissions will have no choice but to depart quite widely from the current map whether they like it or not. Adrian is thus technically incorrect in saying that the 'minimum change' rule is different this time, but in practical terms he is quite right. In my own (non-split) plans, for instance, only 85 seats survive unchanged (and this number includes those that are changed only to realign with new wards). I shall be interested to see whether the Commissions' proposals (not long to wait now) have more or fewer than this. 

* Note that there's an interpretation rule squirrelled away somewhere that says that the term 'local government boundaries' in (b) includes wards; also that the 'most recent election day' provision means that it is the old wards that will be used in Sheffield, so John's assumption is incorrect.

[John Chanin]

Sept 5, 2016 16:01:03 GMT islington said:

Note that there's an interpretation rule squirrelled away somewhere that says that the term 'local government boundaries' in (b) includes wards; also that the 'most recent election day' provision means that it is the old wards that will be used in Sheffield, so John's assumption is incorrect.

I thought I knew the rules, but obviously not on this one. I will have to go back and redo my scheme for Sheffield (and a couple of other places with new wards agreed between May and December).  However I will certainly be looking at ward splits in Sheffield.....and using new ward boundaries to justify them. 

[jamie]

islington
Are the rules quantifiable or will they be making subjective judgments on which seats are better?

[islington]

Sept 5, 2016 16:36:12 GMT jamie said:

islington
Are the rules quantifiable or will they be making subjective judgments on which seats are better?

Well, Jamie, I'm not claiming to be a special authority on this, and I suspect other members of this forum may be better qualified to reply.

But FWIW, my understanding is that the rule deliberately allows the Commissions a great deal of discretion. It says "may" take into account (rather "shall" or "must", which would create a more clearcut requirement), and the phrase "if and to such extent as they think fit" also allows a great deal of latitude. In theory, indeed, a Commission could argue that these provisions allow it to ignore completely the list of factors set out in (a) to (e).

In practice, however, the Commissions are clearly very strongly guided by the specific factors set out. I think they are wise to do this, because if a Commission sought to take advantage of the "if and to such extent that" wording to ignore completely one or more of the specified criteria, it would risk incurring politicians' wrath and might find itself working under more restrictive rules next time. The balance between these factors, however, is a matter of judgment, not generally capable of being objectively quantified. 

Why are the Commissions allowed such wide discretion? For two reasons, I think: first, because it is recognized that the job is inherently difficult; and second, to insulate the Commissions from the risk of legal challenge, which might otherwise cause uncertainty and delays and make it hard to stick to the timetable.

[jamie]

Sept 5, 2016 17:10:36 GMT islington said:

Sept 5, 2016 16:36:12 GMT jamie said:

islington
Are the rules quantifiable or will they be making subjective judgments on which seats are better?

Well, Jamie, I'm not claiming to be a special authority on this, and I suspect other members of this forum may be better qualified to reply.

But FWIW, my understanding is that the rule deliberately allows the Commissions a great deal of discretion. It says "may" take into account (rather "shall" or "must", which would create a more clearcut requirement), and the phrase "if and to such extent as they think fit" also allows a great deal of latitude. In theory, indeed, a Commission could argue that these provisions allow it to ignore completely the list of factors set out in (a) to (e).

In practice, however, the Commissions are clearly very strongly guided by the specific factors set out. I think they are wise to do this, because if a Commission sought to take advantage of the "if and to such extent that" wording to ignore completely one or more of the specified criteria, it would risk incurring politicians' wrath and might find itself working under more restrictive rules next time. The balance between these factors, however, is a matter of judgment, not generally capable of being objectively quantified. 

Why are the Commissions allowed such wide discretion? For two reasons, I think: first, because it is recognized that the job is inherently difficult; and second, to insulate the Commissions from the risk of legal challenge, which might otherwise cause uncertainty and delays and make it hard to stick to the timetable.

Thanks for the reply. I agree that it is preferable for the commission to have wide discretion but it does make predicting their response to a proposal difficult. In my Boundary Assistant plans, I've significantly changed seats where it isn't absolutely necessary in order to meet other criteria. Hopefully the initial proposal next week will shed some light on their thinking on the criteria and if minimum change is very important or not. 

[jamie]

With regard to naming, how do you decide which place goes first? Is it alphabetical, largest town, local authority name etc?

[jamie]

Are the commission willing to accept big changes across the region or just small changes here and there? Similarly, is one good counter-proposal enough for them or do they focus on mass objections?

[Adrian]

Sept 14, 2016 12:00:12 GMT jamie said:

Are the commission willing to accept big changes across the region or just small changes here and there? Similarly, is one good counter-proposal enough for them or do they focus on mass objections?

The answer to these questions is yes. They listen carefully to mass objections, but they often ignore them if they can't find a solution to them. (e.g. There are very few non-split-ward solutions in somewhere like Birmingham, and we know from experience that they'd rather look like idiots than listen to reason when it comes to that particular peccadillo.) Sometimes they rely on someone like one of the people on this forum to make a proposal that no-one else had thought of. At the zombie review they made wholesale changes in some areas, but apart from a few obvious cock-ups (e.g. Mersey Banks) it can be hard to predict what they will change.

[islington]

Lennon said:

"Thinking about these it occurred to me that whilst we all talk about 'orphan' wards as being a bad thing - the real issue isn't a lone ward, but a small number of electors from one LA district. In some areas it could easily be the case that 2 or 3 wards combined only have the same number of electors as a single Sheffield or Birmingham ward - should they still count as 'orphan' even if there are technically more than 1 of them? I am thinking instead of categorising "Orphan" Electors, where less than 15% of the Electorate of the seat is from a single Local Authority. I suspect that realistically it makes very little difference except in a couple of places, but seems to be more 'rigorous' to my logical mind in terms of a reason why it is an issue to be avoided. Thoughts and comments?"

This was in the 'Objective Measurements' thread, but I feel it belongs here.

I entirely agree with his point. Where a single Birmingham ward is included in a seat otherwise outside the city, as the BCE suggests for Sheldon ward and I propose for Hall Green and Oscott, it represents such a large proportion of the resulting seat that it is inconceivable that its electors would be ignored or neglected by the MP. In fact, I'd put the threshold at 10% of the seat, rather than the 15% that Lennon suggests.

Also on the subject of orphan wards, I feel that striving too hard to avoid them can cut across the statutory objective to have regard to LA boundaries. This applies particularly where an authority represents a logical and coherent unit and is roughly the right size for a seat, but is outside the permitted range (above or below) by a relatively small amount. A few examples, off the top of my head, would be: (just too small) Crawley, Eastbourne, Hartlepool, Worcester; (just too large) Chesterfield, Slough.

Such an authority might reasonably expect to form the basis of a seat (i.e. having regard to LA boundaries); but the obvious way of achieving this is likely to involve adding, or hiving off, a single ward so as to get the numbers within range. Although this will result in an orphan ward, probably well under 10% of the relevant seat, I'd argue that in the circumstances the offence is a venial one.

 

[minionofmidas]

Sept 26, 2016 10:28:51 GMT islington said:

Lennon said:

"Thinking about these it occurred to me that whilst we all talk about 'orphan' wards as being a bad thing - the real issue isn't a lone ward, but a small number of electors from one LA district. In some areas it could easily be the case that 2 or 3 wards combined only have the same number of electors as a single Sheffield or Birmingham ward - should they still count as 'orphan' even if there are technically more than 1 of them? I am thinking instead of categorising "Orphan" Electors, where less than 15% of the Electorate of the seat is from a single Local Authority. I suspect that realistically it makes very little difference except in a couple of places, but seems to be more 'rigorous' to my logical mind in terms of a reason why it is an issue to be avoided. Thoughts and comments?"

This was in the 'Objective Measurements' thread, but I feel it belongs here.

I entirely agree with his point. Where a single Birmingham ward is included in a seat otherwise outside the city, as the BCE suggests for Sheldon ward and I propose for Hall Green and Oscott, it represents such a large proportion of the resulting seat that it is inconceivable that its electors would be ignored or neglected by the MP. In fact, I'd put the threshold at 10% of the seat, rather than the 15% that Lennon suggests.

Also on the subject of orphan wards, I feel that striving too hard to avoid them can cut across the statutory objective to have regard to LA boundaries. This applies particularly where an authority represents a logical and coherent unit and is roughly the right size for a seat, but is outside the permitted range (above or below) by a relatively small amount. A few examples, off the top of my head, would be: (just too small) Crawley, Eastbourne, Hartlepool, Worcester; (just too large) Chesterfield, Slough.

Such an authority might reasonably expect to form the basis of a seat (i.e. having regard to LA boundaries); but the obvious way of achieving this is likely to involve adding, or hiving off, a single ward so as to get the numbers within range. Although this will result in an orphan ward, probably well under 10% of the relevant seat, I'd argue that in the circumstances the offence is a venial one.

 

And in the "too small" cases, the area involved will be very much oriented towards the town anyways. The "too large" ones are more problematic. Getting the hived-off area right is important - it doesn't look like the Commission did that in Slough.

The percentage at which an area is "inconceivable to ignore" is, sadly, very much contingent on factors the Commission is bound by law to ignore. A low-turnout working class area in an otherwise safe Conservative seat can be ignored up to a much larger point than any kind of area (including a similar one) in a true marginal. So a strict cutoff point is problematic unless it's set at the worst-case scenario end.

[Foggy]

Sept 26, 2016 10:28:51 GMT islington said:

Also on the subject of orphan wards, I feel that striving too hard to avoid them can cut across the statutory objective to have regard to LA boundaries. This applies particularly where an authority represents a logical and coherent unit and is roughly the right size for a seat, but is outside the permitted range (above or below) by a relatively small amount. A few examples, off the top of my head, would be: (just too small) Crawley, Eastbourne, Hartlepool, Worcester; (just too large) Chesterfield, Slough.  

I do so wish the extra ward from Wealden were not Willingdon. The train doesn't go there, and the bus links going west are at least as good as those heading north. Ideally it'd take in a ward or two from what's currently in the Lewes constituency instead.

The orphan ward that is most easily justified would be the one from the West Dorset DC area in the Dorset South seat. Of course it'd be possible to draw a seat within quota using only wards from Purbeck and Weymouth & Portland – which would then be non-contiguous.

[emidsanorak]

Given that Rule 5 (1) states:

"A Boundary Commission may take into account, if and to such extent as they think fit—
(a) special geographical considerations, including in particular the size, shape and accessibility of a constituency;
(b) local government boundaries as they exist on the most recent ordinary council-election day before the review date;
(c) boundaries of existing constituencies;
(d) any local ties that would be broken by changes in constituencies;
(e) the inconveniences attendant on such changes."

there is a balance between different considerations. It's difficult to say which is the better scheme when there is a mix of good and bad seats. It should however be possible to create a scoring system.

[islington]

The sudden outbreak of split wards of the '2019 electorates' thread has set me thinking about whether any rules, or guidelines, might be applicable here.

I have very serious reservations about splitting wards, of course, but I acknowledge that it's likely to happen as part of the review. And the evidence of the other thread is that split wards are like sugar almonds: you think, "Oh, I'll just have one to taste," then, "One more won't hurt," and before you know it, half the packet has disappeared. So here are a few suggested rules or guidelines, just off the top of my head.

1. Whole wards are the norm and splits the exception.
   
2. Any split must be justified, case by case, showing how it represents a significant improvement over the best possible non-split plan. The justification should be in terms of the stautory factors applying to boundary-drawing (respect for the existing map, for LA boundaries, &c).
3. Splits should use Polling Districts.
   
4. No ward should be split between more than two seats.
5. No seat should involve more than one split ward.
6. No 'orphan splits': that is, when the boundary of an authority is crossed, at least one whole ward should be taken.

The rationale for the last rule is that orphan wards are regarded as undesirable because of the risk that a single ward from an adjoining LA might be regarded as a peripheral and unimportant part of the seat to which little attention need be paid. A split ward, however, by definition will involve even fewer electors that a whole ward, so will be at even greater risk of being marginalized.

[Wisconsin]

Why a *significant* improvement? It needs only to represent an overall improvement based on all the factors.

You continue to treat ward integrity as the paramount statutory factor. It isn’t - it ranks equally with the other factors.

Parliament were made aware of the law and its likely consequences for split wards and criticism of the BCE for not splitting wards in previous reviews. The House of Commons approved the law and the BCE should take note.

[J.G.Harston]

Aug 8, 2020 12:38:30 GMT islington said:

1. Splits should use Polling Districts.
   

I can't agree with that. Polling districts are complete artificial fabrications, completely nothing to do with communities of interest, entirely for convenience of getting to a polling station. That's what got us a boundary running though the backs of gardens in 1983, and slicing through a block of flats in 2010. Also, polling district boundaries are often not actually lines on a map, but simply collections of addresses. A line may be drawn on a map as a convenience between two groups of addresses, but often than is nothing more than "west of here is XA, east of here is XB". If some houses get built in the middle of that area they very well may get assigned to the polling district that is most convenient for them resulting in the "line" moving.

For instance, if somebody appeared on the register living at Sheffield Airport, would they really really be put in Tinsley polling district, or would the line that arbitarily goes through the airport be "moved" north to put the new resident in Greenland five minute's walk away.

In the First Zombie Review, one of the best arrangements for Sheffield was YL's model that sliced the top off Burngreave along a natural boundary rather than the arbitary polling district boundary.