PoliticsHome have YouGov’s regional figures for the past week here. These are the first properly weighted regional breaks from YouGov on data entirely after the first debate (the data release last week straddled it) and they seem to suggest that the Lib Dem boost is stronger in the North.

The swing from Labour to the Lib Dems in the North-East is 13%, in Yorkshire 11.5% and the North West 9.5. Compare this to the swing in the South East (5%) and London (7%). In Scotland there is very little sign of a Lib Dem advance, with the party on 25% compared to 23% at the last election (though this is still much better than the Liberals were polling in Scotland a few months back).

In the South West, the swing from Conservative to Liberal Democrat is 5.5% in this poll, suggesting around 9 Conservative losses to the Lib Dems, rather than the other way around.

On a uniform national swing, the data behind these figures would produce a seat distribution of CON 245, LAB 273, LD 100 – a hung Parliament with Labour the largest party. However, if we do regional swings based on this data, it produces a projection of CON 262, LAB 245, LD 111 – the Conservatives the largest party and an extra 11 Lib Dem seats. The difference is down to the Liberal Democrats gaining more Northern seats (most notably 6 extra in the North east, including all three Newcastle seats) and two extra in the South-West. The Conservatives would gain an extra 20 seats, partially offset by larger losses to the Lib Dems.

Of course, the normal caveats about sample size apply to the regional breaks – with 1000 or so respondents in each region the differences are not necessarily significant. There does seem to be a north-south divide in evidence though.

34 Responses to “YouGov regional figures”

  1. Bad news for us SW Tories, but generally shows that the marginals effect is still there, kinda sorta. It’s just changed quite a bit since the LibDem surge.

  2. The analysis of these regional polls has been somewhat muddled to say the least.

    The Tories were on 38/9 average a fortnight ago with YG so a lot of these votes have come from them. Comparing against the 2005 to calculate that the voter’s journey was from red- to -yellow is therefore wrong.

    Also, there is a lot of encouragement for the Tories in regional swing. In some areas the swing exceeded the national average by 2-3%. In fact in the areas it did not eg Scotland or SW and SE, arguably the Tories expected that anyway.

    Thus, this regional breakdwon contradicts Kellner’s article and clearly shows a Tory majoirty is still possible.

    Don’t get me wrong; possible is not the same as probable.

  3. The momentum still is very clearly with the Lib Dems. They retain what they have and pick up a few scraps from both the other 2. Clearly in 2nd place in the National Poll.

    You can see why Clegg is confident. Clegg may lead a Lib/Lab coalition on a 4 year joint program.

    You can see why GB is how he is.

    Why Cameron is trying to pick off normally Labour seats weakened by the swing to Lib Dems as nothing open to him in the South

  4. Well from comparing these regional numbers to the ones I got out from the YouGov daily numbers are the following:
    My Numbers: 34.6/26.6/32.4/6.1
    YouGov Poll: 36 /31 /28 /6
    So we some big difference here
    North – my numbers are pretty much spot on with a slighlty bigger shift from Con to Lib of 2% but that is marginal error stuff.
    ML/W – my numbers are completely spot on (small 0.3 difference in the labour vote)
    Scotland – Basically similar numbers with a slightly bigger shift from Lab to SNP of 2% in the regional polls.
    South – My numbers register a 19% vote share for Lab and a 33% vote share for Libs. The regional poll gives it 16% Lab and 36%Libs.
    So i can see that my numbers are over estimating Lab’s vote in the South and underestimating it in London.
    I will check the numbers again to see if its a human error on my part, but i have some questions.
    What period does the YouGov regional numbers breakdown cover? and is it possible that some constituencies that where attributed to the South in the daily polls come up as London in the regionals? How many polls does the regional breakdown cover? (I am only covering the latest 7). What is the MOE of the regional breakdown?
    Some of these questions might be stupid but am interested in seeing why i have such discrepancies. My numbers are all from YouGov (undoubtedly it could be as simple as human error ofc)

  5. Anthony

    Using the published figures on Electoral Calculus’s model I get Con 257/Lab 253/LD 108/Nat 11

    Is this different models or rounding? In any case it shows the sensitivity of the number of seats to the slightest movement at these sorts of figures.

  6. Focus on the margins

    It is ghere that the real picture will be seen

  7. Swing from red to blue

    NE 7%
    NW 5%
    York 5.5%
    E-Mid 4.5%
    W-Mid 7%
    Wales 7.5%
    Scot -0.5%
    SW 2.5%
    SE 4%

    For Tory’s interested in still achieving a majority they need a national swing of 6.9%.

  8. Do the models hold up even with the regional numbers? I suspect with such a large swing in the opinion polls to the “third” party perhaps only a seat by seat opinion poll would give a clear picture.

  9. Eoin:

    Haven’t read Kellner’s article: I take it that he clearly said a Conservative majority was not possible then?

    That’s quite surprising.

  10. Once again i confirm my prognosis. Labour is losing its base.

    Swings C2DE
    Lab-Con 7.9
    Lab-Lib 12.6
    Con-Lib 4.8

    These numbers confirm that. See where you would expect a core of C2DE voters to be, such as the north and parts of the midlands, and you see why Labour is in trouble.

  11. @Paul C,

    He said that there was no evidence of marginal boost. And that UNS actually hid an even bigger problem for the blues.

    On some of these %s blue dont seem too far away at all in my opinion. And they also show some regional boost. this is not the saem thing as margianl boost of course. But MORI showed 1.5% ICM showed 0.8% so it is not all a foregone conclusion.

  12. @Xiby,

    I have to caution you chap. A lot of these voters left red a long time ago. When were compare with 2005 we forget that in between a lot of these guys found a comfrtable home in the blue camp.

  13. Guardian/ICM poll: CON 33/LAB 28/LD 30/Other 9

  14. @xiby

    “some big difference here” with your numbers !

    These regional swings (i.e. marginals swings as well) clearly put us firmly in balanced parliament territory whereas you- correct me if I am wrong please- have been predicting a Conservative majority……??!

  15. *****Guardian/ICM poll: CON 33/LAB 28/LD 30/Other 9***

    MORE bad news for Cameron…..especially given that today Clegg has made it clear he will do a deal with a Labour party in third place with votes (and second with seats) as long as GB is OUT….just as I have been saying for the last 10 days…. ;-)

  16. Yougov Regional – Does the data support the theory that Lib Dems are gaining more where there base was low.

  17. @ Rob Sheffield

    No Rob i surely was not predicting a Con majority, but Cons with largest seat share, on the basis of regional shifts.

    The problem with my numbers are essentially London. here my numbers where off quite a bit. I had Lab in a worse off position then it actually has.

    Then i had Lab fairing marginally better in the South then it actually is.

    The rest of my numbers match the regional breakdown.

    I am still trying to see where the prob with my london numbers is tbh.

  18. All seems heading in the direction I’ve previously outlined.

    Cons: [“the more people see Cameron/Osborne, the less they like them”] in gradual and slight decline. Very unlikely to be reversed.

    Labour: Moving up slightly after previous lows. Brown sticking to policy and looking impressive to me.

    Lib Dems: Holding at a very high level and with the final leaders’ debate still to come.

    Tories will be hit, I believe, by massive tactical voting. Nick Clegg has positioned his party so that can they can either support Labour or ask for their support, depending on respective vote percentages.This will ensure that, even if Conservatives have most seats, they will be out-numbered by the combined Lab/Lib Dem mp’s. and won’t form the next government.

    It’s far too complex to predict seats accurately, but vote share could easily be getting close to 30/30/30/10. I’ve always felt this would be an awful election for the Tories: I’m starting to think it could be a lot worse than that.

    Longer term I think it’s quite ironic that the very success the Lib Dems are experiencing may mean that, in future elections, under a new voting system their influence is lessened as more minority parties take a greater share of both votes and seats.

  19. @Paul Croft

    “Nick Clegg has positioned his party so that can they can either support Labour or ask for their support, depending on respective vote percentages.This will ensure that, even if Conservatives have most seats, they will be out-numbered by the combined Lab/Lib Dem mp’s. and won’t form the next government.”

    This is clearly the current trajectory. It all dpends on the ‘devil-in-the-detail’ of the eventual seat tallies. That is VERY hard to predict because of regional variations and tactical voting (on all sides)……

    On Thursday night you will- minor squabbles and pointed figures aside- have two parties singing from basically the same songsheet and one out there….far far far to the right of the other two……….

  20. Somebody please remind me, what is the change to prior (Sat for Sunday Paper’s) ICM? Thanks :-)

  21. Fascinating.

    If these figures are replicated on the night, the election would appear to offer the chance for the Progressive parties in the UK to form a coalition and permanently change the electoral system.

    After which, it would be practically impossible for there to be a Conservative majority. Or for them to even be part of a coalition government.

    Wise tactical voting is needed by Liberals and Labour in order to achieve this. But it is definitely on the cards.

  22. @PaulCroft
    “the very success the Lib Dems are experiencing may mean that, in future elections, under a new voting system their influence is lessened as more minority parties take a greater share of both votes and seats.”

    Surely that’s what they want – to fulfil the wish of voters in a PR election

  23. I’ve come back here because they are driving me potty on the ‘tonight’s polls’ thread.

    There are problems facing us using these results.

    Let us assume they are a true reflection (take as read the results).

    Then we have to make an assumption, the same as we make for UNS, namely that they apply to every seat in region to which the swings apply. We must accept this assumption, because we accept it when we deal with the national results of polls in the same way.

    The first observation I make is that the above will not hold any more than it will for UNS for any seat, it’s just that there are fewer of them to iron out the swings and roundabouts of variances from URS (Uniform Regional Swing).

    If we look at the URS’s, There is a very wide variation between them and as a percentage of the 2005 % result achieved. For instance, Con goes up 3 in the WM from 39 but down 2 from 39 in the SW. That is a huge variation on a fairly close % (35 and 39) achieved by the Tories in each region. There are probably more strident examples.

    I just wonder whether the assumptions I made at the start of this contribution can be made. I mean what’s so different about the WM from the SW?

  24. Howard

    I agree. For some reason the regional breakdowns just don’t seem to ring true as an accurate finding of what is going on – the sum total nationally seems to produce nice consistent results, but once broken down they seem all over the place. Perhaps it is a mistake to use parts of polls designed for one purpose (a national poll) to extrapolate regional or localised trends – fun though it may be to try!

  25. Doubt 6 NEast seats will go LD – only Durham centrol a real prospect.

  26. Howard,

    I dont know if this helps or if it’s what you are driving at, but you have to remember to consider the regions as progressively changing from, say, SW to WM. There is no ‘real’ boundary.

    Thus, for example, the Con swing is weakest in Cornwall. Stronger through Devon – but still weak. And then gradually increasing as it enters the West Midlands. But it would be weaker in Gloucestershire than in Staffordshire.

    Such that where two CON-LAB marginals required the same swing to fall to CON, that would make the CON-LAB marginal in Staffs more likely to fall CON than a CON-LAB marginal in Glos.

  27. Wonder if Tories and Labour now realise how annoyed we all got with expenses and the blocking out of real democracy which FPTP gives?

  28. That’s a good suggestion David but what I would then ask is ‘why is that then? I cannot see any fundamental demographic reason for such a ‘;wave’.

    London is huge and is enveloped by the SE region. It has a fairly larger swing to Con than the SE.

    Sticking with Con (Lab loss seems fairly uniform) there seems to be a middle england bulge of Tory loyalty. As those seats for the main part are Tory already, perhaps that’s why the Tories end up with only a marginal swing of 4% as Peter Kellner reports.

  29. @Howard
    “I mean what’s so different about the WM from the SW?”

    Where to start… :)

    Possibly that in the South-West the Lib Dems traditionally poll in first or second place, rather than third.

  30. “perhaps only a seat by seat opinion poll would give a clear picture”

    Luckily we are going to get one of these next week :)

  31. Jon,
    I meant demographically. I suppose Wm is blue collar and the SW less so OK I’ll buy that (you are one thing or the other and not that airy fairy crowd – i can almost hear Adrian Chiles saying that). Don’t get London though.

  32. @Howard,

    My guess would be that the SW/WM difference is becuse the direction of tactical voting is quite different – any anti-Tory tactical voting will go to LDs in most SW constituencies, but Lab in the WMs… the Tories will benefit from anti-Lab tactical voting in the WMs, which won’t really occur in those SW seats where Lab is in 3rd place.

    That’s my view from here in Cornwall. I don’t get London, either…

  33. Anthony,
    What model are you using for regional swing seat calculations? Your figures for seat changes, compared with a uniform national swing, show quite a difference with these figures – Con +17, Lab -28, LD +11. But when I put these figures into Electoral Calculus’s regional predictor, I get a change from their prediction under UNS of just Con +1, Lab -6, LD +5.

    And when I put these regional figures into a spreadsheet of my own, I get Con+3, Lab -4, LD +3. I don’t think these regional results can really make as much difference as your own model implies – it is possible the figures for the national share have been ‘over-rounded’? When I weight the regional YouGov shares by their 2005 total vote (or by their total population, for that matter) I get overall figures more like Con 33.9%, Lab 27.4%, LD 30.5%. These are much closer to the regional figures than the ones you quote (they show Con and Lab neck-and-neck for seats, for instance).

  34. Well the polls are consistently showing Labour in third place . It will be interesting to see what will happen to their vote now Gillian “the bigot” has decided not to vote Labour. I suspect her views and not Gordon’s will chime with many disenchanted voters.