Boundary update

This is largely for the sake of completeness, since as things stand I doubt the boundary changes will take place, nevertheless I thought I should really update the notional figures for the provisional boundary review.

Calculating how the votes cast at the 2017 election would have translated into seats using the proposed boundaries the Conservatives would have won 298 seats (19 less than currently), Labour would have won 244 (18 less than currently), the Lib Dems would have won 8 (4 less than currently), the SNP 30 (five less than currently), Plaid 2 (2 less than currently). As you can see, the Conservatives and Labour would lose about the same number of seats, but the Conservatives would have been nearer to an overall majority and once you’ve taken away Sinn Fein MPs, may have been able to avoid doing a deal with the DUP.

Regular readers will recall that before the election it looked as if the boundary reviews would have favoured the Tories more – I suspect this change is largely because the 2017 election happened to produce a lot of very marginal seats, and that small boundary changes have flipped some of these in Labour’s favour. If you look at how it affects the swings the two parties would need to win a majority it’s clear that the boundary changes would still help the Tories:

  • On the the new boundaries the Conservatives would need a lead of 2.8 to get an overall majority, compared to 3.4 currently
  • On the new boundaries Labour would need to be 3.6 points ahead to become the largest party, compared to 0.8 currently
  • On the new boundaries Labour would need a lead of 7.8 points to get an overall majority, compared to 7.4 currently

Some of you may be wondering why, if the boundary changes are about evening out the size of constituencies the result is still a system that seems to favour the Conservatives over Labour. This is not a sign of something being afoot – the four boundary commissions are genuinely independent – rather it’s because differently sized constituencies (“malapportionment”) is only one of several factors that can produce a skew in the electoral system, and the current Conservative advantage comes not from seat size, but from the impact of third parties and the Tory vote being more efficiently distributed. For example, when it comes to translating votes into seats huge majorities in safe seats are “wasted” votes. At the 2017 election there were 89 Conservative seats where they got over 60% of the vote, but 115 Labour seats where they got over 60% (and 37 seats where Labour got over 70%). None of this is set in stone of course – up until 2015 the system tended to favour Labour – if a party outperforms in marginal seats it can do better than uniform swing suggests, if it gains votes in safe or unwinnable seats then it would do worse.

The new boundaries are rather irrelevant if they never come into force – when the Boundary Commissions report in Autumn 2018 there then needs to be a vote in both the Commons and the Lords to implement their recommendations. That would have been challenge enough with a majority given that there is every chance of a few Conservative rebels. Without a majority it’s going to be very difficult indeed, especially since the DUP have so far opposed the changes (at the provisional stage the changes were thought to hurt the DUP and benefit Sinn Fein).

Nevertheless, for anyone who wants them notional figures for the 2017 on the provisional boundary recommendations are here.


An update on the boundary review. Back in September I published notional figures for the proposed boundaries in England & Wales. I’ve now updated those to include Scotland as well (this is partly because the Scottish boundary Commission published later, but it also took much longer to do – the Scottish Commission are much happier to split wards between constituencies, which probably leads to constituencies that better follow communities… but it makes it trickier to work out notional figures.)

Notional figures for new boundaries for England, Wales and Scotland

The partisan effects in Scotland are no great surprise. The SNP won 56 of Scotland’s 59 seats in 2015, so it was inevitable that most of the losses will be SNP. That aside, on the new boundaries they will be even more dominant. Orkney & Shetland is a protected seat so the sole Liberal Democrat constituency is retained, but Labour and the Conservatives will both see their single Scottish constituency disappear on the new boundaries.

Edinburgh South, the lone Labour seat in Scotland, is split between the new Edinburgh East and Edinburgh South West & Central seats. Both will notionally have an SNP majority of over 4000 – Edinburgh East will be a SNP-Lab marginal, with a SNP majority of 7.9%, Edinburgh SW&C will be a three-way marginal with the SNP in first place, the Conservatives in second place and Labour close behind them.

Dumfriesshire, Clydesdale & Tweeddale, the lone Tory seat in Scotland, mostly goes into Clydesdale & Eskdale, with the rest of the seat split into several much smaller parts. The new Clydesdale & Eskdale seat will have a notional SNP majority of about 5000. On paper the best seat for the Tories will be the new Berwickshire, Roxburgh & Selkirk seat, with a notional SNP majority of only 1.3% (though that’s an increase from 2015).

Now we have notional figures for the whole of Great Britain we can work out national totals and what sort of swings would be needed for parties to win a general election on these boundaries.

The 2015 general election had results of CON 330, LAB 232, LDEM 8, SNP 56, Others 24.
On the proposed boundaries the 2015 general election would have been CON 319, LAB 203, LD 4, SNP 52, Others 22. The Conservatives lose 11 seats, Labour lose 29, the Lib Dems 4 and the SNP 4.
Note that on the boundaries proposed for the abandoned review in the last Parliament the results would have been Con 322, Lab 204, LD 4 and SNP 50 – so this new boundary review is actually marginally worse for the Tories than the one that was blocked before the election.

I should add my normal caveat that these notionals are an accounting exercise – projecting how people voted in each ward, moving them into their new seats and totting up the votes. It does not take into account that some people might have voted differently in 2015 if they’d lived in different seats, for that reason I suspect it may slightly underestimate the Liberal Democrats (and it’s possible that the Greens might actually have saved their seat).

We can also look at what difference the boundaries would make to the leads each party needs to win an election.

  • Currently the Conservatives need to have a lead of 5.7% to get an overall majority (hence the 6.5% lead they actually got translating into only a tiny majority). On the proposed boundaries the Tories would get an overall majority with a lead of only 1.9%.
  • In contrast Labour currently need a towering lead of 12.6% to win an overall majority, and the boundary changes would move that target even further away, requiring a lead of 13.5%. To even be the largest party Labour would need a lead over the Conservatives of 4.7% (up from 3.9% on the current boundaries).

(One might reasonably wonder why, if the review makes nearly all the seats the same size, it still leaves the Conservatives in a better position than Labour. This is because different seat sizes is only one part of how votes translate unevenly into seats. The crucial part in explaining the present Conservative advantage is the distribution of the vote and the impact of third parties. The collapse of the Liberal Democrats and the growth of the SNP and UKIP means the system now favours the Conservatives. The Lib Dems are primarily strong in areas that would otherwise be Tory… but now win very few seats, UKIP have largely taken votes from the Tories, but this has not translated into many seats. In contrast the SNP are now utterly dominant in an area that previously returned a large number of Labour MPs. What this means if that if there is a Lib Dem revival or a Labour revival in Scotland the skew towards the Conservatives will unwind.)

These are only provisional recommendations – the boundary commissions will revise them based on the consultation period, so much of the detail will be tweaked before the final recommendations. It’s also far from a certainty that they will actually be implemented when they are complete. Earlier this month Pat Glass MP had a Private Members Bill which if passed would tweak some of the rules of the review, requiring the Commissions to start the process again from scratch and therefore probably delaying it beyond the election. I doubt the Bill will go far – it is nigh on impossible to pass a Private Members Bill in the face of government opposition. However, second reading did highlight some opposition to the boundary changes. Firstly, the DUP spoke against the boundary changes – there had been some speculation around conference season that there had been some sort of deal and the DUP were onside. They are apparently not. Secondly two Conservative MPs (Peter Bone and Steve Double) voted in favour of the Bill. It doesn’t take many rebels to stop the boundary changes progressing…


-->

A final post on boundary changes (at least until the Scottish proposals next month). This comes from a discussion I had with Mark Pack. Normally the thing we look at with boundary changes is what the party-partisan effect is, how the new boundaries would change the sort of swing that Labour need to win a general election. However, currently Labour are a very, very long way from the sort of polling lead they’d need to win a majority, so a small change in that figure really doesn’t make a lot of difference. More interesting in the current political climate is the effect it would have on Labour internal battle and any potential deselections.

The rules for how Labour will deal with re-selections after boundary changes are yet to be confirmed, so these are based on the rules set out for 2011 in the Labour rule book, on the assumption that Labour’s NEC will use similar rules this time round. A Labour MP has a right to seek selection in any seat that contains 40% or more of the electors in their existing seat. If an MP’s seat is divided up so much that no single seat contains 40% of their old electors then they’ll have the right to seek nomination in a seat with less than 40% of their old voters. If they are the only sitting MP to seek selection in a seat, they are nominated through the normal trigger ballot process. If more than one sitting MP seeks the nomination in a new seat there is a members ballot to pick between them.

Applying those rules to the provisional boundaries we can see where there may be contests under those rules. Note that this list is exhaustive, it contains every case where Labour MPs could compete against each other under the selection rules… but in some cases it will be easily avoided through either agreement (there are enough seats to go round) or retirement (an MP will be well over 70 come the general election and possibly considering retirement anyway). Of the 231 Labour members of Parliament in England & Wales, 142 of them should not face any re-selection difficulties connected to boundary changes – they may well see changes to their seat, but there is a single notionally Labour seat to which they have the sole right to seek selection. What about the other 89?

Avoidable Challenges

There are six places where more than one MP would have a right to seek selection for a seat, but where there are enough Labour seats to go round, so if MPs co-operate and agree between themselves who will stand where, no head-to-head challenge is necessary and no one is left empty handed. These are:
Alfreton and Clay Cross. Nastasha Engel and Dennis Skinner both have the right to seek selection here, but Skinner also has the right to seek selection in Bolsover, so a challenge seems unlikely.
East London. Mike Gapes’ seat is sliced up into tiny pieces, and if the NEC follow past practice he should have the right to seek selection in any of the successor seats. He is the only sitting MP with a right to seek selection in the new, ultra-safe, Forest Gate & Loxford seat so I imagine he will go there. If not, he could challenge Wes Streeting, Margaret Hodge or John Cryer (who could, in turn, seek selection in Stella Creasy’s Walthamstow)
Redcar. Andy McDonald and Anna Turley can both seek selection in Middlesbrough NE & Redcar, but McDonald is also eligible for the safe Middlesbrough W & Stockton E seat, so a challenge is avoidable.
Ashton Under Lyne. Jonathan Reynolds and Angela Rayner are both eligible, but Rayner is also eligible for the safer Failsworth & Droylsden.
Stockport. This is avoidable, but not without some pain for Ann Coffey. Andrew Gwynne & Ann Coffey are both eligible for the safe Stockport North & Denton seat. Ann Coffey is also eligible for the Stockport South & Cheadle seat, but that is far more marginal (that said, Coffey will be 73 at the next election, so may not stand).
Pontefract. Yvette Cooper and Jon Trickett are both eligible to seek selection, but Yvette Cooper also has a free run at Normanton, Castleford and Outwood.

Not Enough Labour seats to go round

The following seven areas have enough seats to go round, but one or more of them is notionally Conservative, so there may be a contest for the winnable seat or someone may be left in a seat that is notionally Conservative:
South London. Siobhain McDonagh’s seat is sliced up. Two of the successor seats, Merton & Wimbledon Common (a potentially winnable marginal) and Sutton & Cheam (no hope) are notionally Conservative, so she will have the choice of fighting one of them, or challenging either Chuka Ummuna or Rosena Allin-Khan.
South-East London. Erith and Thamesmead is split up into Erith & Crayford (a Tory seat) and Woolwich. The only option for a Labour seat for Theresa Pearce is to challenge Matthew Pennycook for the Woolwich nomination. Pennycook has the option of seeking the Woolwich nomination, or going up again Vicky Foxcroft for the Greenwich & Deptford nomination.
Coventry. Geoffrey Robinson’s seat becomes comfortably Conservative on new boundaries, but he has the option of going up against Jim Cunningham for the Coventry South nomination. He’ll be 81 by the next election, so I assume he won’t.
Nottingham. Vernon Coaker’s Gedling seat disappears. Half goes into the Conservative Sherwood seat, so there is the potential of a battle against Chris Leslie for the nomination in the Labour Nottingham East and Carlton seat.
Cumbria. The Workington seat disappears. Part of it goes into the very Conservative Penrith & Solway seat, which is unlikely to be attractive to Sue Hayman, leaving her the option of fighting Jamie Reed for the Whitehaven & Workington seat.
Wrexham. Susan Elan Jones’s Clwyd South seat is dismembered. Part of it goes into the elaborately named De Clwyd a Gogledd Sir Faldwyn seat, but that is notionally Conservative. The other part goes into Wrexham Maelor, where she would have to compete against Ian Lucas for the nomination.
Newport. The Newport seats are combined into one. Jessica Morden would also have the right to seek nomination in Monmouthshire, but that’s solidly Tory leaving one Labour seat between her and Paul Flynn. Flynn will be 85 come the next election, so the issue may well be resolved by retirement.

Straight two way fights

There are seven Labour seats where there are two Labour MPs who are eligible for that seat, and that seat only – meaning a straight fight is unavoidable unless someone stands down:
Sunderland West – Bridget Phillipson vs Sharon Hodgson
Newcastle North West – Catherine McKinnell vs Chi Onwurah
Wednesfield & Willenhall – David Winnick vs Emma Reynolds (though Winnick will be 86)
Stoke South – Rob Flello vs Tristram Hunt
Dudley East & Tipton – Ian Austin vs Adrian Bailey (though Bailey will be 74)
Neath & Aberavon – Stephen Kinnock vs Christina Rees
Cardiff South & East – Jo Stevens vs Stephen Doughty

More complicated fights

There are eight areas where there are rather more complicated fights… but where ultimately there are more Labour MPs than there are seats, so something will have to give:

Birmingham. Roger Godsiff’s seat disppears. He will have the right to seek election in four other Birmingham seats, putting him up against Gisela Stuart, Jess Phillips, Richard Burden or Steve McCabe. He will be 73 come the election though, so may choose to stand down.
Islington & Hackney. The change that got the most attention when the proposals were announced. Essentially Meg Hillier, Jeremy Corbyn, Diane Abbott and Rushanara Ali have to somehow share out the Finsbury Park & Stoke Newington, Hackney West and Bethnal Green and Hackney Central seats. Someone is going to get stuffed.
Rochdale & Bury. Debbie Abrahams, Ivan Lewis, Liz McInnes and Simon Danzcuk are in play, with Rochdale, Prestwich and Middleton and Littleborough & Saddleworth. If Danzcuk remains suspended from the Labour party then the problem presumably resolves itself.
Liverpool. Steve Rotheram’s seat disappears and he would be eligible to challenge Louise Ellman, Peter Dowd or Stephen Twigg for selection in their seats. Rotheram himself is standing for Liverpool mayor, so it won’t be an issue for him. If he steps down though whoever is elected in the subsequent by-election would face the same issue.
Bradford & Leeds. Judith Cummins seat disppears. She is eligible to seek selection for Bradford West (against Naz Shah), in Spen (against Jo Cox’s successor) or in Pudsey, where Rachel Reeves will likely also be seeking the nomination (Leeds West vanishes, but Pudsey takes much of its territory and becomes a notionally Labour seat)
Sheffield. Newly elected Gill Furniss sees her seat dismembered – she is eligible to seek nomination in Sheffield North and Ecclesfield (against Angela Smith) or Sheffield East (against Clive Betts).
Pontypridd. Owen Smith’s seat is dismembered and he will have the right to seek nomination in either Chris Bryant’s Rhondda & Llantrisant or Ann Clwyd’s Cynon Valley and Pontypridd. Ann Clwyd will be 83 by the next election, so it may be resolved by retirement.
Islwyn. Chris Evans’ seat also vanished, and he will have the choice of competing against Nick Smith in Blaenau Gwent or Wayne Davies in Caerphilly.

The deep blue sea

Fourteen Labour MPs do not have a notionally Labour seat they would be eligible to seek selection in. In some cases this is just because of a slight change to an already ultra-marginal seat (e.g. Chris Matheson in Chester notionally loses his seat, but there’s really little change from 2015), in other cases it leaves them with a very difficult fight:

Andy Slaughter would face a Tory majority of 14% in the new Hammersmith & Fulham seat
Gareth Thomas would face a Tory majority of 11% in the new Harrow and Stanmore
Joan Ryan would face a small Tory majority of just 3% in the new Enfield seat
Ruth Cadbury faces a 10% Tory majority in Brentford & Chiswick
Tulip Siddiq faces a 9% Tory majority in Hampstead and Golders Green
Alex Cunningham is only eligible for the nomination in Stockton West, with a 7% Tory majority
Chris Matheson doesn’t actually face much change, but Chester would have a 1% Tory majority on paper
Jenny Chapman faces a notional Tory majority of 1% in Darlington
Madeleine Moon’s Bridgend is merged with the Vale of Glamorgan to create a notionally Tory seat, but with a majority of only 3%
Alan Whitehead’s Southampton Test would have a 4% Tory majority on paper (Southampton Itchen would flip to Labour… but Whitehead doesn’t have the right to go there under Labour rules)
Melanie Orr would be eligible to seek selection in either Grimsby North & Barton or Grimsby South and Cleethorpes. Both, however, would be Conservative.
Holly Walker-Lynch faces a similar situation, under Labour rules she can apply for Calder Valley or Halifax, but they are both notionally Tory.
Finally, in the sorriest situation of all are Margaret Greenwood and Alison McGovern. They are both only eligible to seek selection in the new Bebington & Heswall seat… and even if they do get it, it’s now notionally Tory.

So, by my reckoning there will probably be around 15 re-selection battles where a sitting Labour MP faces up against another sitting Labour MP on the provisional boundaries, though remember that these are subject to change (and it only takes a small adjustment by the boundary commission to shift the number of voters from an old seat above or below 40%). It’s also worth noting that you don’t need boundary changes for a deselection – there is a normal trigger ballot process than can be used to deselect an MP and some of the speculation about deselections – Peter Kyle for example – is not due to Labour seats being merged together.


This is the third in a series of posts on the boundary review. There is a general overview of what is happening and why it’s controversial here, a summary of what the effects are and some of MPs who are losing their seats here. This final post has the full, seat-by-seat, estimates of how the votes cast at the last general election would have fallen out on the new boundaries in England and Wales.

Full notional results for England and Wales.

The changes in England and Wales result in the Conservatives losing 10 seats, Labour losing 28 seats, the Liberal Democrats losing 4 and the Greens losing Brighton Pavilion (though notional calculations like these risk underestimating the performance of parties with isolated pockets of support like the Greens and Lib Dems, so it may not hit them as hard as these suggest). The Scottish boundary commission don’t report until next month, but for obvious reasons the Conservatives and Labour can only lose a maximum of one seat each there, meaning that on these boundaries the Conservatives would have had a majority of around 40 at the last election.

The usual caveats I give for notional results apply – this is an accounting exercise, estimating what the ward level vote within each constituency would have been in 2015 (basing the distribution on the distribution at local elections) then reallocating the wards to their new constituencies and adding them back up again. If there is a radically different pattern of support in an area at local and national elections the figures might be misleading, if there are loads of independent candidates in any area (as in rural Wales, North Yorkshire or Cornwall) then the figures won’t be that accurate. If you know an area really well and you think the projections are wrong, then you are probably correct… but hopefully any such errors cancel out.

And a final caveat – this is purely a prediction of how the votes would have fallen out if the votes at the last election were counted on the new boundaries. They are certainly NOT a prediction of what would happen at the next election.


Overnight the English and Welsh boundary commissions released their initial proposals for the boundary changes – the English proposals are here, Welsh proposals are here. If you missed it, I discussed the wider background to the boundary changes and how they work yesterday. The initial proposals are very much as expected – we’ve known the broad strokes of the review since the rules were set in 2011 and the electorate numbers were released in the spring. The review will cut the total number of MPs from 650 to 600, including a reduction from 533 to 501 in England and a reduction from 40 to 29 in Wales. The English losses are disproportionately in the North, where the population is falling relative to southern England, and will consequently hit Labour harder than the Tories. Today’s proposals are about the details – finding out what the specific proposals are and what the specific impact will be.

As usual, I’ve worked out notional projected figures for what the new boundaries would mean at a general election (using a similar method to that used by Rallings and Thrasher for the official notional figures they will produce for the media once the final report it out). The changes are, as ever, a case of swings and roundabouts. A net total of 43 seats are abolished across England and Wales, but the impact is more than just those abolished and created seats, other seats may notionally change hands as wards are moved from one seat to another. Just because a seat held by a party is abolished, it doesn’t mean that party necessarily looses out as parts of that dismembered seats may push neighbouring seats into their column.

Even defining which seats disappear can be debatable (especially when there are contentious selection battles to come). In England there are around 35 existing seats that don’t really have a seat which is its clear successor, 12 of these are Tory seats, 23 are Labour. Looking at the knock-on effects elsewhere, there are around 16 seats where the Conservatives would notionally have won seats they don’t currently hold, 13 seats where Labour would have won seats they don’t currently hold. In Wales the changes are more drastic, given the sharp reduction from 40 seats to only 29. By my calculation Labour lose 7 seats in Wales, the Conservatives lose 4.

There are only three seats in England and Wales that the Liberal Democrats would have won on the proposed boundaries: North Norfolk, Westmoreland & Lonsdale and Ceredigion. I should add a caveat to this though – these projections are purely an attempt to work out what the result would be if the votes cast at the last election had been counted on these new boundaries. It’s purely an accounting exercise, estimating what the general election vote was in each ward, allocating them to their new seats and totting them up. It doesn’t account for the fact that people might have voted differently if they’d been in a different seat – for example, it’s possible that people who voted Tory in Penistone East ward in 2015 when the ward was in Penistone & Stocksbridge might instead have voted tactically for Nick Clegg had their ward been in Sheffield Hallam. For that reason I think such notional figures can sometimes underestimate the Lib Dems. Note that the Liberal Democrats have suggested they might win Cambridge on the new boundaries – this is not unrealistic at all, on my notional figures Labour have a majority of under 200 there.

Aside from Nick Clegg, several other high profile figures face seeing their seats abolished or made tricky to win. Looking at the Conservatives, George Osborne’s Tatton seat is abolished (the largest part of it goes into David Rutley’s Macclesfield seat), David Davis’s Haltemprice seat also goes, mostly into the new Goole seat, the successor to Andrew Percy’s Brigg and Goole. Priti Patel’s Witham seat forms the smaller part of a new Witham & Maldon seat, mostly made up of John Whittingdale’s Maldon seat. Jeremy Wright, the Attorney General, also sees his Kenilworth and Southam seat abolished. Some media reports have suggested Justine Greening may also be in trouble: she’s not, her seat has Wimbledon Common added to it but will still be very Conservative.

For the Conservatives, MPs in danger of losing their seat poses a risk to getting the boundary changes through, and it’s likely the party will try to use retirements, peerages and so forth to ensure no current Tory MPs lose out. For Labour MPs the boundary changes take place against a different background and MPs who have opposed Jeremy Corbyn may face difficulties in selections if there are local boundary changes. High profile MPs whose seats are abolished include Tristram Hunt in Stoke Central (divided between Stoke North and South), Ian Austin in Dudley North (his seat mostly ends up in the new Dudley East & Tipton, the successor seat to Adrian Bailey’s West Brom West), Vernon Coaker (whose Gedling seat is divided between Chris Leslie’s Nottingham East and Conservative-held Sherwood). Jeremy Corbyn’s own seat sees substantial changes: Diane Abbott’s Hackney North and Stoke Newington is abolished and split between a Hackney Central seat (mostly made up of Meg Hillier’s Hackney South) and a Finsbury Park and Stoke Newington seat (mostly made up of Corbyn’s Islington North), with a much smaller part going into a new Hackney West and Bethnal Green seat.

I’ll put up full notional results later on today, but looking at England and Wales as a whole, it looks as if the Conservatives would lose 10 seats, Labour would lose 28 seats, the Liberal Democrats would lose 4 and the Greens would lose one (by my calculations the new Brighton North would be a close three-way marginal, with the Conservatives narrowly ahead of Labour). That means at the last general election the Conservatives would have won a majority of 40, rather than the majority of 12 they actually got.

I’ll release full notional figures for each seat later on today.