UK parliamentary boundary review

[CumbrianLefty]

Thank you very much for doing these summaries

[President Punxsutawney Phil]

These summaries were great, thank you very much.
Thank you for your hard work.

[Filuwaúrdjan]

The addition of the Swansea valley to Brecon & Radnor may make Labour competitive again here forty years after they ceased to be after a previous boundary change, or it may give the Lib Dems a new source of tactical votes.

The local Party in Brecon & Radnor will want to give it a go with these boundaries, as will the one around Pontardawe. Genuinely no reason not to. Could be quite a messy contest.

Clwyd North and Clwyd East are both marginally Tory, while Wrexham moves a little in their direction.  The "Glyndŵr" area may provide enough Labour votes to make them competitive in Montgomeryshire & Glyndŵr, but they may need to get the votes of people in Montgomeryshire who have previously voted tactically for the Lib Dems.

Doubt there would be much partisan effect from the Wrexham changes: the Maelor is very Conservative, but the other area going into the constituency is generally pretty Labour in General Elections (local elections in Wrexham being every bit as Special and local elections in North Wales are apt to be). The Berwyns horror story of a constituency will certainly be a very messy contest: again, Labour have no reason not to give it a go, for all their lack of history in Montgomery.

[Filuwaúrdjan]

Just doing the numbers for the B-R-CT seat and if we break its electorate down by county, then we have:

Brecon - 47.0
Radnor - 29.0
Glamorgan - 24.0

However, if we add the Brecon part of the Swansea Valley to the Glamorgan part, then we find that the proportion of the electorate in the Swansea Valley comes to 35.1%. The proportion of the existing Brecon & Radnor in the Swansea Valley is 14.8%.

In the Cross Berwyns horror, then the proportions are 68.0% from the existing Montgomery and 32% from the existing Clwyd South. If we switch to counties, then the figures are 65.5% Montgomery and
34.5 Denbighshire.

[YL]

Just doing the numbers for the B-R-CT seat and if we break its electorate down by county, then we have:

Brecon - 47.0
Radnor - 29.0
Glamorgan - 24.0

However, if we add the Brecon part of the Swansea Valley to the Glamorgan part, then we find that the proportion of the electorate in the Swansea Valley comes to 35.1%. The proportion of the existing Brecon & Radnor in the Swansea Valley is 14.8%.

When Labour used to be competitive in Brecon & Radnor, before 1985, did they have any areas of strength in the bits which remained in Powys other than the Ystradgynlais area?  I can't really think of anywhere else in the current constituency which strikes me as a likely Labour stronghold, but I may be missing something, and they were still getting respectable votes in the 1980s and 1990s, which must have been coming from somewhere.

[Filuwaúrdjan]

When Labour used to be competitive in Brecon & Radnor, before 1985, did they have any areas of strength in the bits which remained in Powys other than the Ystradgynlais area?  I can't really think of anywhere else in the current constituency which strikes me as a likely Labour stronghold, but I may be missing something, and they were still getting respectable votes in the 1980s and 1990s, which must have been coming from somewhere.

Brecon, which has some large estates especially at its western end, was the main one and is one of those Marches/Mid Wales towns that is not the very tidy place people assume, but there was a degree of residual strength in some of the other towns, especially Presteigne, but also Knighton and (prepare for a surprise) Crickhowell, which hadn't gentrified much back then.

[Хahar 🤔]

When Labour used to be competitive in Brecon & Radnor, before 1985, did they have any areas of strength in the bits which remained in Powys other than the Ystradgynlais area?  I can't really think of anywhere else in the current constituency which strikes me as a likely Labour stronghold, but I may be missing something, and they were still getting respectable votes in the 1980s and 1990s, which must have been coming from somewhere.

Brecon, which has some large estates especially at its western end, was the main one and is one of those Marches/Mid Wales towns that is not the very tidy place people assume, but there was a degree of residual strength in some of the other towns, especially Presteigne, but also Knighton and (prepare for a surprise) Crickhowell, which hadn't gentrified much back then.

Unfortunately the residual strength in Prestiegne was destroyed by Geoffrey Fourmyle.

[Filuwaúrdjan]

At least according to the Electoral Calculus figures, Bridgend remains narrowly Tory...

Not a chance, ftr.

[President Punxsutawney Phil]

What would this review look like if it was done on, say, late 2000s rules?

[YL]

At least according to the Electoral Calculus figures, Bridgend remains narrowly Tory...

Not a chance, ftr.

I’m a bit suspicious of the EC notionals in Wales (and in some other places), hence the caveat, but they are the only ones out there so far.  Looking at some past local election results, I think I agree with you and think that the new Bridgend is notionally Labour.  As discussed in another thread its Tory MP is sniffing around chicken run opportunities.

[YL]

What would this review look like if it was done on, say, late 2000s rules?

Welsh over-representation would not have been corrected: each part of the UK would have retained more or less its existing number of seats.  In particular England would have seen fewer new seats and more abolished ones.

There would have almost certainly been less change.  There would probably have been no constituencies crossing 1974 county boundaries in England (with the possible exception of the one between Merseyside and Cheshire) and almost certainly no split wards in England.  New unitary authority boundaries in Cheshire, Dorset, Northamptonshire and possibly Bedfordshire would presumably have been respected (but not Cumbria: that change was after the enumeration date) though the details would depend on the electorate figures.  And there would be lots of little changes actually made which were unpopular and would not have been needed without a 5% rule (e.g. removing Cherry Hinton from Cambridge) so probably wouldn’t have happened.

I might play with this scenario at some point.

[CumbrianLefty]

In other words, it would be better overall.

Never forget the 5% variation limit arose from a lengthy grift from the likes of Policy Exchange about how the existing arrangements gave Labour A hUgE aNd PeRmAnEnT iNbUiLt AdVaNtAgE - and the original proposal for only 600 MPs had similar overtly partisan motivations.

[JimJamUK]

In other words, it would be better overall.

Never forget the 5% variation limit arose from a lengthy grift from the likes of Policy Exchange about how the existing arrangements gave Labour A hUgE aNd PeRmAnEnT iNbUiLt AdVaNtAgE - and the original proposal for only 600 MPs had similar overtly partisan motivations.

Has there ever been any actual evidence of how quota variation limits and the number of MPs supposedly benefit Labour? Much more important as far as I can tell was the separate quota for Wales, along with the relative decline in Labour voting urban/post-industrial areas.

Of course, in reality both the existing and proposed boundaries currently have a major bias against Labour owing to its poor vote distribution, which only looks to be moderately curtailed at the next election.

[YL]

In other words, it would be better overall.

Never forget the 5% variation limit arose from a lengthy grift from the likes of Policy Exchange about how the existing arrangements gave Labour A hUgE aNd PeRmAnEnT iNbUiLt AdVaNtAgE - and the original proposal for only 600 MPs had similar overtly partisan motivations.

Has there ever been any actual evidence of how quota variation limits and the number of MPs supposedly benefit Labour? Much more important as far as I can tell was the separate quota for Wales, along with the relative decline in Labour voting urban/post-industrial areas.

I've never seen anything convincing.  As you say, there was the over-representation of Wales, and there was also the fact that the slow pace of reviews meant that the electorates boundaries were based on were often quite out of date even when they were first used, but the main ways the Commissions used the greater electorate flexibility were to avoid crossing major local government boundaries and to avoid splitting wards in big cities (e.g. the current boundaries in Leeds, where there are some big seats made of five wards and some small ones made of four) and neither of those has any systematic partisan effect that I can see.

[Filuwaúrdjan]

What essentially happened was that a) certain peculiarities of the 2005 election were widely misunderstood and that b) David Cameron acted, as he often did, in an extremely petty manner about something that had annoyed him.

So for a) the issue was that in 2005, Labour won a large majority despite only a small lead in the popular vote, and in fact won a majority of seats in England despite (very) narrowly trailing in the popular vote there. This was taken by many Conservatives as a sign of 'bias' in the electoral system created by the way in which electoral boundaries were then drawn. Which was nonsense: what had happened was that Labour had an absolute shocker of an election amongst more liberal-minded voters as well as Muslims and students due to various controversial decisions made during the 2001-5 Parliament, but also had another outstanding election with Lab-Con swing voters. Vote-shares and majorities dropped and sometimes collapsed in constituencies where they had previously been very high, but at the same time some seats that had been unexpectedly gained in 1997 were actually held. The result was an unusually large mismatch between seats won and votes polled.

Meanwhile for b) it happens that Oxfordshire missed out on gaining an extra seat at the Fifth Periodical Review by a literal handful of electors, and that this infuriated and enraged David Cameron who was, of course, an Oxfordshire MP.

[CumbrianLefty]

And as far as the reduction in Welsh seats is concerned, a Labour government would have agreed to that in due course just as they did for the Scottish seats cut that came in for the 2005 GE.

The other changes, not so much.

[EastAnglianLefty]

It used to be the case that inner urban seats had smaller electorates, whilst suburban seats had lower electorates due to housing shifts. Similarly, poorer industrial areas tended to depopulate, whilst wealthier constituencies attracted new migrants.

Stagnant housing construction has meant that fewer people are moving out of the inner suburbs to the outer suburbs; urban cores have begun to grow again and with some rapidity; and the Conservatives have done much better in poorer industrial areas in the last couple of elections. Those have removed a lot of the "structural" advantage to Labour.

It's also the case that boundary changes usually look advantageous to parties that have just won an election, because they're more likely to have a seat with a large majority near an opposition party seat with a small majority. They always look particularly bad for the Lib Dems because their vote is highly contingent and based on tactical voting, but this is almost always misleading.

[afleitch]

And as far as the reduction in Welsh seats is concerned, a Labour government would have agreed to that in due course just as they did for the Scottish seats cut that came in for the 2005 GE.

The other changes, not so much.

Yeah. Post Government of Wales Act 2006 (and the referendum)  there would have been a reduction in Welsh seats and as with Al's point on Oxfordshire, a number of Tory shires were already at 'extra seat' level by 2010.

[YL]

What would this review look like if it was done on, say, late 2000s rules?

The quotas would have been
74785 for England
71983 for Northern Ireland
68197 for Scotland
56757 for Wales

I suspect some political pressure to sort out the discrepancies might have appeared when these numbers were published.  But assuming they were accepted...

... I had a look at allocation by county in England outside London.  Note that here "county" means
- anything with a county council
- any Metropolitan county
- Berkshire
- any unitary authority not in Berkshire or a Met county

County boundaries, in this sense, were not supposed to be crossed, but it was possible to make exceptions.  I don't think the BCE's practice in the Fifth Review was entirely consistent, but I think it can be approximated by saying that any county whose county average would be more than about 10% from quota will be considered for grouping, but that grouping will only be done with counties which were part of the same 1974 county.  (This last condition had no justification in the rules, but it seems to fit with the BCE's actual approach.)

The allocation of seats to a county or group of counties is based on which choice gets the county average closest to the quota.  For this reason the electorate needed to get n+1 seats is slightly less than n+1/2 quotas.

Based on the above, the following counties would remain with their existing groupings and see the same number of seats as they do now: Cornwall (6), Devon inc. Plymouth and Torbay (12), Wiltshire (5), Swindon (2), Portsmouth (2), East Sussex inc. Brighton & Hove (8), Essex inc. Southend and Thurrock (18), Suffolk (7; just missing out on an 8th), Norfolk (9), Hertfordshire (11), the three Bedfordshire unitaries (6), Warwickshire (6), Worcestershire (6), Herefordshire (2), Shropshire inc. Telford & Wrekin (5), City of Leicester (3), City of Nottingham (3), Nottinghamshire (8), Derbyshire inc. Derby (11), Lincolnshire (7), West Yorkshire (22), North Yorkshire (6), City of York (2), Warrington (2), Hartlepool (1).

The following counties would gain a seat: Somerset (6), Gloucestershire (7; NB this is an extremely close decision), Berkshire (9; also very tight), Bucks inc. MK (8), Oxfordshire (7), Surrey (12; another tight one), West Sussex (9), Kent inc. Medway (18), Cambridgeshire inc. Peterborough (8), Leicestershire inc. Rutland (8).

The following counties would lose a seat: Staffordshire inc. Stoke (11), South Yorkshire (13; another very tight call), Merseyside (14; note this can be three for Wirral and 11 for the rest, so there is no need to threaten to cross the Mersey), Greater Manchester (27), Lancashire inc. Blackpool and Blackburn (15), Cumbria (5), Tyne & Wear (11).  The West Midlands Met county would lose 2 seats (26).

Then there are some more complicated decisions.  The "Avon" area doesn't justify an extra seat if you look at it as a unit, but if you look at the unitaries separately it does, though they'd probably have paired Bristol (4.49 quotas) and North Somerset (2.20 quotas) for 7 seats.

The "Humberside" area clearly loses a seat, but the question is whether to split it into the Lincolnshire part (3.23 quotas) and the Yorkshire part (5.92 quotas) or put the Isle of Axholme with the East Riding.  (I would generally favour the former, as "Humberside" is an abomination.)  Perhaps there is a case for doing what the review actually did, and grouping Axholme with Doncaster, breaching a 1974 boundary but avoiding the negative consequences of the other choices for where to put it; if you do that, South Yorkshire does not lose a seat.

Some unitaries can be separated from their previous partners.  Southampton (2) can be separated from Hampshire (14) and Darlington (1) from Durham (5), the latter area losing a seat overall.  Cheshire East didn't exist when the Fifth Review was operating, but can be treated separately for 4 seats, with Cheshire West & Chester paired with Halton for 5.  (Well, on the numbers it can.  It's not easy to draw a nice map.)  The Dorset and BCP unitaries similarly didn't exist, and can be separated with 4 seats each.  The situation in Northamptonshire is similar, but North Northamptonshire's average is close to the 10% deviation, so they might not have been split.

On the numbers, Northumberland (3.34 quotas) should lose a seat or be grouped with somewhere else (northern Tyne & Wear, presumably, though of course that would breach a 1974 boundary) but they have been nice to it in the past because of its large size, and I don't think the situation has changed very much.  So I assume it would keep its 4 seats.

I assume that they wouldn't have given the Isle of Wight a second seat, because they hadn't in the past under these rules, though according to the general principles used it would have made sense to do so.

[President Punxsutawney Phil]

Here's a mock-up of what this might entail (tried to avoid crossing any district boundaries)
Northumberland
yellow 60283
mauve 62580
red 64593
blue 61958

North Tyneside
orange 77382
green 79033

Newcastle upon Tyne
pink 63164
purple 62443
light teal 65450

Gateshead
yellow 74674
mauve 69945

South Tyneside
red 56947
blue 58423

Sunderland
orange 66339
green 68963
pink 72688

Durham
purple 81296
light teal 74679
green 80426
mauve 78992
red 78140

Darlington 78498
Hartlepool 71228

Stockton-on-Tees
green 71804
pink 71977

Middlesborough, Redcar and Cleveland
yellow 69398
light teal 66256
purple 65440

[CumbrianLefty]

Post 2015, It's better to think of it as an advantage for the "progressive" parties than for Labour specifically.

It's not a significant advantage though and has more to do with demographic trends.

The game's the game ultimately. The Tories have a built in advantage in terms of councils but you never hear about that.

And councillors, due to the "double tier" areas being mostly Tory-inclined shires. So when they do fall behind on that metric - as in May - its a sure sign things aren't going too well for them.

[YL]

North Tyneside
orange 77382
green 79033

Newcastle upon Tyne
pink 63164
purple 62443
light teal 65450

Gateshead
yellow 74674
mauve 69945

South Tyneside
red 56947
blue 58423

Sunderland
orange 66339
green 68963
pink 72688

They weren't usually as reluctant to cross Met borough boundaries as that, but worked with a county level allocation, which here gives 11 seats for Tyne & Wear.  So I think they'd have left the Sunderland and North Tyneside seats unchanged, as you have, but grouped Newcastle, Gateshead and South Tyneside for six seats; this might have involved South Shields and Jarrow & Gateshead East seats much like the ones we actually got (but perhaps both extending a little further west, given the slightly higher quota), Blaydon gaining one Gateshead ward, a recreated Tyne Bridge covering the centres of both Newcastle and Gateshead, and expanded Newcastle North and Newcastle East seats.

Or they might have tried crossing the Tyne between the Blaydon area and Newcastle's western fringe instead, but the Tyne Bridge idea is the one with precedent.

[Harry Hayfield]

The official notional election results will be published by Professors Thrasher (Sky News) and Rallings (ITN) sometime later this year, my guess is certainly before the end of the year, but thanks to the House of Commons publishing a similarity index, a friend of mine has sent me the following unoffical calculations

Unoffical Notional Electrion 2019
Conservatives 371 seats
Labour 200 seats
Scottish National Party 48 seats
Liberal Democrats 9 seats
Plaid Cymru 2 seats
Green Party 1 seat
Speaker 1 seat
Northern Ireland Parties 18 seats (DUP 8, Sinn Fein 7, SDLP 2, Alliance 1)
Conservative majority of 94 / 100 when allowing for Sinn Fein and Speaker(

Based on 2019, the Conservatives gain 6, Labour lose 2, the Liberal Democrats lose 2 and Plaid lose 2

[CumbrianLefty]

Which, if true, is rather less beneficial for the Tories than some previous boundary reviews.

[Conservatopia]

Which, if true, is rather less beneficial for the Tories than some previous boundary reviews.

Yes, this was widely predicted to be the case as the 2019 election lead to more Tory seats in non-traditionally Tory areas.