Let the great boundary rejig commence

[Harry Hayfield]

You must be logged in to read this quote.

And because that was the Conservative Party policy, that means all the constituencies have to be 77,658 (and not one elector bigger or smaller than that number). This means that the regions split as follows:

Northern Ireland: 15 (rounding down to the nearest whole seat 15.06)
Scotland: 49 (rounding down) 50 (rounding up) 49.8
Wales: 29 (29.12)
England: 491/492

Allow me to start the ball rolling with my own constituency of Ceredigion. Based on that average, Ceredigion is 18,615 electors too small and needs to be merged with somewhere. I suspect that they might say that the preserved counties must be adhered to so that rules out merging with Dwyfor, Montgomeryshire, Brecon and Radnorshire, so that leaves only Carmarthen East and Preseli Pembrokeshire. Without knowing the exact electorates of the wards of Carmarthen East (which I should point out are being rejigged at the moment) I think that a merger with Preseli Pembrokeshire seems the most likely (as there has been a historical link with that constituency between 1983 and 1997) but would hope that the constituency would be called Ceredigion and the Preselis, or Ceredigion and Preseli and not just Ceredigion and Pembroke North as it was between 1983 and 1997

[Harry Hayfield]

The Welsh version of the Electoral Reform Society has been having a go at this and will be publishing their report on Wednesday in Cardiff. I have asked for the report to be e-mailed to me following publication (and will post once I get it) but they have raised an interesting statement, namely that according to the Government of Wales Act, the number of constituency AM's in the Assembly MUST equal the number of MP's elected. This means that if these plans do indeed go ahead (as expected), then the Assembly will shrink from it's current 60 to as few as 44 (when the Richards Commission called for a long term goal of an Assembly with 80 members).

[Harry Hayfield]

Welsh Electoral Reform Society Recommendations for Wales
(published June 9th 2010)

Name of constituency   Con   Lab   Lib Dem   Plaid   UKIP   Green   BNP   Ind   Others   Totals   Electorate   Turnout   Winner
Barry and Penarth   18,106   16,387   8,112   2,433   1,419   485   0   186   306   47,434   73,652   64.40%   Con
Blaenau Gwent and Tredegar   2,781   16,612   4,198   1,981   623   0   1,227   696   4,675   32,795   78,027   42.03%   Lab
Breconshire and Montgomeryshire   14,085   3,304   15,520   1,839   994   181   0   152   409   36,485   73,630   49.55%   Lib Dem
Bridgend   10,213   15,024   7,870   2,514   796   0   1,072   0   0   37,489   75,991   49.33%   Lab
Caerphilly   6,038   16,700   5,587   4,996   893   0   1,450   985   14   36,662   78,773   46.54%   Lab
Cardiff Central   10,637   13,184   12,753   1,437   912   532   0   191   341   39,988   74,574   53.62%   Lab
Cardiff North East   13,568   15,702   10,680   1,589   1,043   479   0   236   361   43,658   75,345   57.94%   Lab
Cardiff West   13,851   17,175   8,066   2,335   1,125   607   0   106   165   43,428   74,061   58.64%   Lab
Carmarthen   11,546   11,885   5,156   9,011   1,173   13   0   134   16   38,934   75,798   51.37%   Lab
Ceredigion and Rhaeadr   7,325   3,406   16,990   8,634   980   528   0   34   66   37,963   74,042   51.27%   Lib Dem
Conwy and Abergele   13,623   8,667   5,765   5,561   759   0   18   54   194   34,642   74,374   46.58%   Con
Denbigh, Llangollen and the Vale of Conwy   12,256   12,428   5,762   3,543   764   0   711   21   80   35,564   74,357   47.83%   Lab
Flint and Rhyl   12,629   15,043   4,982   1,978   571   0   834   0   76   36,114   73,720   48.99%   Lab
Gwynedd and Machynlleth   6,367   5,220   4,325   10,863   774   0   0   788   31   28,369   75,318   37.67%   Plaid
Llanelli   7,203   14,349   4,590   10,804   1,063   0   118   0   0   38,127   75,776   50.32%   Lab
Merthyr Tydifl and Ystrad Mynach   4,034   15,421   8,002   3,495   891   0   1,338   1,303   111   34,596   78,860   43.87%   Lab
Mold and Shotton   12,868   15,638   6,949   1,617   928   0   1,247   0   0   39,247   75,949   51.68%   Lab
Monmouth   18,945   12,211   9,526   1,140   1,017   445   283   0   30   43,598   78,292   55.69%   Con
Neath and Aberavon   4,681   16,768   5,183   3,531   615   0   1,284   556   338   32,955   74,811   44.05%   Lab
Newport   10,524   14,672   8,710   935   924   238   1,176   0   58   37,237   75,491   49.33%   Lab
Pembroke   16,827   12,691   5,414   3,883   1,001   0   0   145   0   39,961   75,339   53.04%   Con
Pontypridd and Aberdare   3,845   15,315   6,302   5,011   1,037   110   0   135   250   32,004   76,594   41.78%   Lab
Rhondda and Ogmore   3,300   17,213   4,727   4,712   564   41   313   1,648   94   32,612   75,838   43.00%   Lab
Swansea East and the Vale of Neath   5,861   15,564   6,960   4,863   806   120   1,286   81   96   35,637   77,901   45.75%   Lab
Swansea North and Loughdor   8,123   15,936   7,478   2,280   758   221   1,345   52   25   36,217   74,988   48.30%   Lab
Swansea West and Gower   9,345   13,539   10,561   1,870   695   272   927   252   120   37,581   77,775   48.32%   Lab
Torfaen   9,213   16,527   6,474   1,826   922   449   1,492   1,594   0   38,497   78,434   49.08%   Lab
Vale of Ely   11,178   15,772   8,758   2,776   1,234   479   84   0   374   40,657   73,196   55.55%   Lab
Wrexham   9,003   12,505   7,744   2,322   787   0   1,124   0   0   33,486   74,020   45.24%   Lab
Ynys Môn and Bangor   7,223   10,322   3,105   8,805   1,035   0   0   1,526   123   32,140   74,167   43.33%   Lab

Could I ask someone more skilled than me to table that please and could I also ask for a named current ward map of Wales so I can show which wards go into which seats? (please e-mail me direct with map)

[Harry Hayfield]

I haven't drawn many Welsh ward maps but tend to be fairly useless in my experience - some Welsh wards are absolutely tiny.

Also I can't see this on the ERS website - do you have any further details?

One more thing:

You must be logged in to read this quote.

WTF??!?!?!??  If you're going to combine Cardiganshire with anywhere presumably the Preseli or Machynlleth would be a better idea - at least there aren't a load of mountains in the way.

The Welsh part of the ERS e-mailed me the designs of the new constituencies and they were embaroged until a meeting today in Cardiff. If they have not yet appeared on the website, then pass. As to the idea of Ceredigion and Rhaedr, the wards in question are:

Ceredigion
All wards
Powys
Blaen Hafren, Llandinam, Llanfair Caereinion, Llanidloes, Rhiwcynon, Llanyre, Nantmel, Rhaeadr Gwy and the northern half of the electoral division of Llanafanfawr
Pembrokeshire
Cilgerran, Clydau, Crymych, Dinas Cross, Newport, Scleddau

[Harry Hayfield]

What's all this hogwash about you cant redo the boundaries cause the list of electors is out of date.
Do you use registered electors to determine boundaries? I can't think of anything less democratic. Canada uses citizens, all citizens (old, young). You could also use citizens over 18 (IE can vote).

The agrument that Harriet Harman used in the Commons yesterday is that what is the point of rejigging the boundaries if not everyone who can vote is allowed to vote (and cited all the data listed). My response would be "that did not stop you rejigging the boundaries of Scotland for the 2005 elections, and the rest of the UK for the 2010 elections"

[Harry Hayfield]

OK, it's time to go public.  Here's my plan for the North West:



Click on the map for all the details.

Could you list / send me a "how similar each seat is to the last one" tally and I'd be able to create a set of notionals for those suggestions. By that I mean, Wirral : 100% Wirral West + 7% Wirral South

[Harry Hayfield]

OK, it's time to go public.  Here's my plan for the North West:



Click on the map for all the details.

Notional North Western Summary
(Source: Andrew's allocations, Election 2010 Results from the Guardian)
Votes Cast
Labour 1,208,658 (40.25%)
Conservatives 933,978 (31.11%)
Liberal Democrats 634,547 (21.43%)
United Kingdom Independence Party 97,991 (3.26%)
British National Party 64,125 (2.14%)
Independents 32,975 (1.10%)
Green Party 15,486 (0.52%)
Other Parties 5,787 (0.19%)

Party Wins
Accrington and Blackburn North
Altrincham and Sale West
Ashton under Lyne and Stalybridge
Birkenhead
Blackburn and Darwen
Blackpool North and Fleetwood
Blackpool South
Bolton North and Turton
Bolton South
Bootle and Kirkdale
Burnley and Nelson
Bury North
Bury South and Crumpsall
Cheadle
Chorley
City of Chester
Colne and Clitheroe
Congleton
Crewe and Nantwich
Crosby and West Lancashire
Eccles and Worsley
Eddisbury
Ellesmere Port
Fylde
Garston and Halewood
Halton
Hazel Grove
Heywood and Middleton
Hyde
Knowsley
Lancaster and Morecambe
Leigh
Liverpool, Riverside
Liverpool, Walton
Liverpool, West Derby
Lunesdale and Wyre
Macclesfield
Makerfield
Manchester Central and Moston
Manchester, Didsbury and Wythenshawe
Manchester, Gorton and Hulme
Manchester, Withington and Levenshulme
Oldham East and Saddleworth
Oldham West and Royton
Ormskirk, Maghull and Skelmersdale
Preston
Rochdale
Rossendale
Salford
South Ribble
Southport
St. Helens North
St. Helens South and Whiston
Stockport
Stretford, Urmston and Sale East
Tatton
Wallasey
Warrington North
Warrington South
Westhoughton
Wigan
Wirral

[Harry Hayfield]

Pretty much as I was expecting except for Lancaster and Morecambe.  The two predecessor seats (Lancaster & Fleetwood and Morecambe & Lunesdale) both have Tory majorities of less than 1000, and the best Tory areas of both seats are in the new Lunesdale and Wyre.  I would have expected Lancaster & Morecambe to be safeish Labour. I imagine that Bolton North & Turton and Bury North would have been very close.

Lancaster and Morecambe
BNP 422 Con 19,737 Green 1,274 Ind 96 Lab 18,972 Lib Dem 7,915 UKIP 1,768

Bolton North and Turton
Con 19,323 Green 76 Ind 271 Lab 23,196 Lib Dem 7,196 UKIP 2,162

Bury North
BNP 2,092 Con 20,386 Green 153 Ind 307 Lab 19,500 Lib Dem 9,225 UKIP 1,405

[Harry Hayfield]

Any more regions in the pipeline?

[Harry Hayfield]

Welsh Allocation Averages
Clwyd 331,525 / 75,716 = 4.37 = 4 seats
Dyfed 279,972 / 75,716 = 3.69 = 4 seats
Gwent 348,532 / 75,716 = 4.60 = 5 seats
Gwynedd 180,376 / 75,716 = 2.38 = 2 seats
Mid Glamorgan 393,121 / 75,716 = 5.19 = 5 seats
Powys 102,601 / 75,716 = 1.36 = 1 seat
South Glamorgan 332,608 / 75,716 = 4.40 = 4 seats
West Glamorgan 293,034 / 75,716 = 3.87 = 4 seats
Total 29 seats

[Harry Hayfield]

and don't forget, if you would like your notionals worked out, simply PM me with the similarity to each old seat (for instance: Blackpool South (95% Blackpool South, 4% Wyre, 1% Blackpool North) and I will post the new notional makeup of your area.

[Harry Hayfield]

We seem to be making rapid progress. These are based on the 598 seat electorate calculations correct? And don't forget that for me to calculate these seats, express them as %ages o the old seat (for instance 65% of Leicester South, 17% of Leicester East)

[Harry Hayfield]

If I wanted to neither split Redditch nor wholly redraw rural Worcestershire nor have more than one constituency cross into Warwickshire, I had no choice but to draw these districts somewhat undersized, which forced large and sometimes somewhat unfortunate districts in Warwickshire. Broadly it's a return to the pre 2010 map, including the oddball Rugby & Kenilworth pairing.

Redditch 79,087
Loses the more rural wards in Worcestershire, but gains a larger rural territory in Warwickshire instead: Tanworth, Studley, Sambourne, Alcester, Kinwarton, and Bidford & Salford wards. 73% of the seat are in Worcestershire.
Stratford on Avon 74,711
Remainder of district, Leam Valley ward in Rugby
Warwick & Leamington 78,037
Regains all the Warwick district wards it lost in 2010 (but not the bit of Stratford district that it used to include as well): Lapworth, Leek Wootton, Cubbington, Radford Semele (however you pronounce that)
North Warwickshire 77,541
Regains the Arley & Whitacre and Hartshill wards, and thus exactly as in 97-10.
Nuneaton 77,378
Loses these, gains Bulkington, Wolvey, Fosse, Avon & Swift, and even Earl Craven & Wolston wards from Rugby
Rugby & Kenilworth 78,657
The motheaten ugly remainders of Rugby and Warwick districts.

The one alternative I could try is pairing Kenilworth with Stratford, tossing a lot rural country southeast of Warwick into the Rugby seat (and a bit into the Warwick seat too.)

I used to live in Stratford constituency and my personal feelings are that residents in Warwickshire would strongly object to being clumped in with Redditch.

[Harry Hayfield]

We have clearly come with a substantial number of ideas, however it is now difficult to tell what are proposals, what are actual committed constituencies and how those committed constituencies would turn out.

Therefore could I suggest that people who have made committed constituenciues (i.e constituencies that they are happy with and make electoral sense) to pm me with their constituencies (and how similar they are to old constituencies (i.e constituency name = 55% of 2010 constituency, 32% of 2010 constituency)) and I will post in a new thread entitled "US Atlas Forum Notionals 2010", how these new constituencies would rack up.

[Harry Hayfield]

Now that we have no elections until the Mid Terms in November, I can now concentrate on the UK rejigged boundaries. All I need is now similar (in percentage terms) the new seats are to the 2010 boundaries and I can whip up a set of notionals whenever you fancy them?

[Harry Hayfield]

When the Assembly elections are over, I'll try and see what I can manage for Dyfed (as was) but have to admit that the chances are very high that Ceredigion will have to cross a border in order to shape up. My own personal hopes are for a Ceredigion and the Preselis or a Ceredigion and Montgomeryshire West and fears are for a Ceredigion and Dinefwr.

[Harry Hayfield]

Based on those boundaries, and the last election results, what are the expected changes in seats?  Has anyone calculated that?  Or better still, produced a map?

I am working through them now (am up to Edinburgh). The changes are (so far):
Dumfries: Lab WIN

Suprises:
Angus East and Kincardine: SNP 32% Con 31%
Every time Angus undergoes boundary changes it seems to turn into a Conservative target. In 1992, the SNP won the seat with a majority of 954 (2.01%) which under the boundary changes in 1997 became an SNP majority of 473 (1.07%). In 2001 the SNP won Angus by 3,611 (10.31%), boundary changes made it an SNP majority of 3,719 (9.47%) and now in 2010 Angus was an SNP majority of 3,282 (8.65%) and now (thanks to Kincardine) it becomes an SNP majority of 408 (0.98%). Is there any reason for Angus having Conservative areas being added to it?
Cupar and St. Andrews: Lib Dem 34% Lab 30%
The successor to Fife North East has now come into Labour's sights thanks to the addition of Glenrothes and therefore (if these proposals pass) I would expect Sir Menzies to stand down at the 2015 election (he would have been an MP by then for 28 years anyway)

[Harry Hayfield]

Scotland Initial Proposal Notionals
Scottish Labour 896,463 votes (42.41%) winning 37 seats
Scottish National Party 421,467 votes (19.94%) winning 6 seats
Scottish Liberal Democrats 390,243 votes (18.46%) winning 8 seats
Scottish Conservatives and Unionists 354,356 votes (16.77%) winning 0 seats

GAINS
Dumfries: Labour WIN
East Dunbartonshire and Kilsyth: Labour WIN
Midlothian and Tweedale: Labour WIN

Potential Targets
Aberdeen North: SNP swing of 7% to gain
Argyll, Bute and Lochaber: Lab swing of 5.5% to gain, SNP swing of 7% to gain
Cupar and St. Andrews: Lab swing of 1.5% to gain
Deeside and Gordon: SNP swing of 7% to gain
Edinburgh West: Lab swing of 2.5% to gain
Angus East and Kincardine: Con swing of 0.5% to gain
Banff and Buchan: Con swing of 4.5% to gain
Dundee East and the Glens: Lab swing of 2.5% to gain

[Harry Hayfield]

Projection based on September average polls and initial estimates for England and Scotland
Labour 38% (+8% on 2010)
Conservatives 36% (-2% on 2010)
Liberal Democrats 11% (-13% on 2010)
Others 15% (+6% on 2010)

Labour 276 seats (+65 seats)
Conservatives 259 seats (-34 seats)
Scottish National Party 16 seats (+10 seats)
Green Party 2 seats (+2 seats)
Liberal Democrats 1 seat (-43 seats)
New Winning Line will be 301 seats, so far 554 seats calculated. 16 seats in Northern Ireland and 30 seats in Wales yet to be calculated

[Harry Hayfield]

The two Green wins are the new Brighton, Pavillion (gain from Lab) and Norwich South (gain from Lib Dem). That is a forecast based on the average polls in September 2011. To know the notional tallies (based on my estimates of Scotland and the Guardian's calculations) work backwards

[Harry Hayfield]

What the polls seem to be saying is Con -2% on the general (in essence virtually unchanged), Labour are +9% and the Liberal Democrats -13%. The majority of that -15% is from Lib Dem to Lab, but the SNP, UKIP and Greens are also up (SNP +2%, Green +2%, UKIP +1%) with the BNP and the Others -1%. This suggests to me that in Lib Dem seats with a majority of less than 22%, the Lib Dems will lose the seat. In 2010, there were only eight Lib Dems elected with a majority of more than 22% (Bath, Fife North East, North Norfolk, Orkney, Ross, Hallam, Westmorland and Yeovil) and of those eight only Bath, North Norfolk, Orkney, Yeovil remain in place and only Orkney is forecast to remain Lib Dem on a national uniform swing.

[Harry Hayfield]

These are the seats that I have as Lib Dem and their majorities (based on the data from the Guardian and my Scottish calculations)

Abingdon and Oxford North 0% over Con
Argyll, Bute and Lochaber 10% over Con
Bath 21% over Con
Bermondsey and Waterloo 19% over Lab
Berwick and Morpeth 6% over Con
Berwickshire, Roxburgh and Selkirk 11% over Con
Bideford and Bude 8% over Con
Birmingham Yardley 0% over Lab
Bodmin and Newquay 1% over Con
Bristol West 24% over Lab
Caithness, Sutherland, Ross and Cromarty 25% over Lab
Cambridge 13% over Con
Cheadle 6% over Con
Cheltenham 10% over Con
Colchester 15% over Con
Cupar and St Andrews 4% over Lab
Deeside and Gordon 14% over SNP
Eastbourne 6% over Con
Eastleigh 6% over Con
Edinburgh West 5% over Lab
Glastonbury and Wincanton 4% over Con
Guiseley and Yeadon 0% over Con
Hazel Grove and Poynton 2% over Con
Hornsey and Wood Green 8% over Lab
Inverness and Skye 22% over Lab
Kendal and Penrith 9% over Con
Kingston and Surbiton 13% over Con
North Devon 9% over Con
North East Somerset 4% over Con
North Norfolk 20% over Con
Norwich South 0% over Lab
Orkney and Shetland 51% over Lab
Richmond and Twickenham 2% over Con
Sheffield South West 5% over Lab
Sheffield West and Penistone 19% over Con
Solihull 2% over Con
Southport 7% over Con
St Ives 3% over Con
Taunton 11% over Con
Teddington and Hanworth 11% over Con
Torbay 8% over Con
Truro and St Austell 3% over Con
Willesden 1% over Lab
Yeovil 24% over Con


The polls since the general election suggest that the following seats are at risk:

Seats with a majority over Con of less than 11%
Seats with a majority over Lab of less than 21%

So looking at that list, the following should be expected


Abingdon and Oxford North 0% over Con

Argyll, Bute and Lochaber 10% over Con

Bath 21% over Con
Bermondsey and Waterloo 19% over Lab

Berwick and Morpeth 6% over Con
Berwickshire, Roxburgh and Selkirk 11% over Con
Bideford and Bude 8% over Con

Birmingham Yardley 0% over Lab

Bodmin and Newquay 1% over Con

Bristol West 24% over Lab
Caithness, Sutherland, Ross and Cromarty 25% over Lab

Cambridge 13% over Con
Cheadle 6% over Con
Cheltenham 10% over Con

Colchester 15% over Con
Cupar and St Andrews 4% over Lab

Deeside and Gordon 14% over SNP

Eastbourne 6% over Con
Eastleigh 6% over Con

Edinburgh West 5% over Lab

Glastonbury and Wincanton 4% over Con
Guiseley and Yeadon 0% over Con
Hazel Grove and Poynton 2% over Con

Hornsey and Wood Green 8% over Lab

Inverness and Skye 22% over Lab

Kendal and Penrith 9% over Con

Kingston and Surbiton 13% over Con
North Devon 9% over Con
North East Somerset 4% over Con

North Norfolk 20% over Con
Norwich South 0% over Lab

Orkney and Shetland 51% over Lab
Richmond and Twickenham 2% over Con

Sheffield South West 5% over Lab

Sheffield West and Penistone 19% over Con
Solihull 2% over Con
Southport 7% over Con
St Ives 3% over Con
Taunton 11% over Con
Teddington and Hanworth 11% over Con
Torbay 8% over Con
Truro and St Austell 3% over Con

Willesden 1% over Lab

Yeovil 24% over Con

[Harry Hayfield]

Whenever a national poll is published, the first thing I do is see whether that poll publishes the full data set (as a result I only track polls by Com Res, ICM, Ipsos-MORI, Populus and Angus Reid). I then enter the number of people voting for each party and generate the % shares. In the most recent poll I have from Com Res, this gave Con 37%, Lab 36%, Lib Dem 12% and Others 15%. Over the course of a month I average all the polls in that month (which in September gave Con 36%, Lab 38%, Lib Dem 11% and Others 15%.

Usually I use UK-Elect when forecasting elections but for elections where the seats are brand new (as is the case at the moment) I used a ratio system to make my forecasts which works like this. In 2010 (under these proposed constituencies) the Conservatives won 37.69% of the vote, the September average has them on 35.81% (a drop of 1.88%). This means that the Conservative vote has fallen by 5% of it's 2010 total. So I then say to every constituency "Right, take 5% off the Conservative vote". I then do this for all the other parties as well. Lab +30%, Lib Dems -53%, SNP +151%, UKIP -4%, Green +326%, BNP -34% and Others -21%.

So for instance in a seat like Cambridge, that gives the following:
Green 16,203 (29% +22%)
Lab 15,855 (29% +5%)
Con 12,190 (22% -3%)
Lib Dem 9,272 (17% -22%)
UKIP 1,143 (2% unchanged)
Others 402 (1% unchanged)
Green GAIN from Liberal Democrat

and at the same time allows for the 2% margin of error that all polls have

[Harry Hayfield]

Anthony Wells has just published his Scottish numbers and they match mine exactly (which if I may say so I am very pleased about indeed, partly because a) the Boundary Commission published how much of each seat was in the new seat and b) I couldn't remember which way round the calculations went. I think the Scottish Boundary Commission should be held up as an example to England, Wales and Northern Ireland as to how to produce boundary review data (and I am very tempted to ring up the Welsh Boundary Commission today and ask if they can release their data in January in a similar manner

[Harry Hayfield]

I read somewhere that had the new boundaries been used in the 2010 election, Caroline Lucas wouldn't have gotten her seat.

You are indeed quite correct.

Brighton, Pavillion and Hove
Lab 16,472 (31%)
Green 14,612 (28%)
Lib Dem 11,502 (22%)
Con 8,653 (16%)
UKIP 1,118 (2%)
Others 180 (0%)