The monthly ICM poll for the Guardian came out today, with topline figures of CON 28%(-3), LAB 33%(+1), LDEM 14%(+3), UKIP 14%(nc), GRN 5%(-1). Tabs are here. There’s a movement from Con to Lab since ICM’s previous poll, but nothing that couldn’t be normal margin of error – the broader picture still suggests a very small Labour lead, with no strong trend (in November the average Labour lead was 1.6%, so far this month it’s 1.3%).

This is the first time since September that any poll hasn’t shown UKIP ahead of the Liberal Democrats. The previous one was also ICM, and apart from one unusual Populus poll in July you have to go right back to March to find polls from other company doing the same. ICM consistently show the highest level of support for the Liberal Democrats for methodological reasons. This is largely because of how they handle don’t knows – when people say they don’t know how they’d vote, ICM look at how they voted at the previous election and assume that 50% of them will end up voting the same way. This is based on recontact surveys of don’t knows after previous elections and has made polls more accurate in the past… but of course, we can’t know until May whether that will still hold true under the sort of political realignment we seem to be seeing this Parliament. Without the reallocation ICM’s figures today would have been LDEM 11%, UKIP 15% – so it would have been quite a high Lib Dem figure anyway even without the adjustment.

Ipsos MORI’s month political monitor is also due. Today’s Evening Standard reported some figures, but I assume they are saving up the voting intention figures for tomorrow. The data so far is here, and shows people’s optimism about the economy in general (as opposed to their personal finances) dropping to its lowest since last year. I wrote about similar findings from YouGov at the start of the month, so this does appear to be more than a single poll; people’s confidence in the economic recovery does seem to be faltering a bit.

149 Responses to “ICM/Guardian – CON 28, LAB 33, LDEM 14, UKIP 14, GRN 5”

1 2 3
  1. @Spearmint,

    Latest Ashcroft marginals shows a swing from Con to Lab of just 3% – less than what the national polls are showing.

  2. @Spearmint

    But my point is it wasn’t an Ashcroft poll. And the Ashcroft one in Doncaster N was always going to be pretty dumb and pointless too (even if that mangled data for a moment made it appear otherwise).

    As a domiciled taxpayer, rather than a (former) non-domiciled non-taxpayer, I want his money back.

  3. I don’t think labour will be bothered about the Ashcroft poll, they seem to change radically poll by poll. if you don’t like one poll just wait a short while and another one will be great news.

  4. @ BRAMLEY
    Using YouGov’s First Verdict, with a large sample (4,335), I asked:
    “Which party would you most like to win in your constituency? Please choose the party you like best, regardless of whether they have a chance of winning?.”
    I followed up … just one example of real significance: a third of the Green vote went to Labour.

    So what? The first question isn’t a standard VI question, but one that I would expect to show a higher Green share than standard VIs.

    And we have just had the Ipsos-Mori poll, with the Greens higher than in any other national (‘next General Election’) poll than since 1989 using standard VI questions, anyway. If Greens lose 1/3 of that share, they would still be on course for a record share for them at the General Election, by a mile.

  5. @Paul M,

    Agreed. Marginal polls are subject to MOE too, and the latest isn’t radically different from the national situation, just a 2% Tory boost or so. Could just be a case of MOE like the latest IPSOS Mori and ICM polls.

  6. @Spearmint
    “Also those Plaid and Green votes in Carmarthen W. & S. Pembrokeshire and Warwick & Leamington look squeezable.”
    Who do you see doing the squeezing?
    I don’t understand why you picked these constituencies. Was your tongue in your cheek?
    CW&SP 2010 4232 Plaid but no Green candidate.
    W&L 2010 only 693 votes for Green and (not surprisingly) no Plaid candidate.

  7. @ Bramley

    I haven’t seen the YouGov poll, but I’m guessing that their sample was national, not just in marginals (LAB-CON and maybe LAB-LIB, LAB-NAT, LAB-UKIP), where tactical voting would happen. In those seats, the Green share is much lower than nationally (Ashcroft’s poll today had it at 4%).

    Also, if you look at the difference between Ashcroft’s two questions in those constituencies, the difference is only 0.5 points, so one-ninth of the Green vote in the first question.

    Whether that is a house effect, random variation, or a difference between marginals and the national picture, isn’t clear.

    It’s also worth remembering that the same thing applies to the Lib Dems.

  8. … in the hope that once again low-information voters will fall for it.

    How dare anyone challenge the SNP in their newfound and hard-earned monopoly over low-information voters in Scotland.

    Phil Haines: when I look at the oddities of the ICM methodology in relation to the Lib Dems and UKIP from a relatively anecdotal perspective, I can understand why they are sticking to their guns.

    Regardless of whether the LDs were to get 5% or 15% in May 2015, it wouldn’t surprise me in the slightest if somewhere in the region of a quarter of those who ultimately vote Lib Dem would be very reluctant to admit it (even before we talk about the number of voters who would admit to doing so, but be clear that it was a purely tactical vote against a party they like the idea of even less).

    Equally, as well as UKIP are now doing, and despite an increasingly strong consensus that their average in the polls reflects the proportion of voters who want to vote UKIP as their first preference, discussion over UKIP VI continues to focus on how many of their supporters will vote for them knowing that UKIP are extremely unlikely to win their constituency.

  9. Re: Ashcroft, one thing to bear in mind when looking at marginals vs national swing… Because of the situation in Scotland, the GB swing looks a lot worse for Labour than the England and Wales swing, which is of course where all the marginals are. What that means is that if you compare the marginals to that swing, the swing is smaller than across E&W. Whether that is tactical/incumbency effects that Ashcroft isn’t picking up, or whether it’s that marginals had fewer 2010 Lib Dems to take (as Steve Fisher suggests) isn’t clear. But what is clear is that the swing in the marginals is consistently smaller than in E&W as a whole.

    I’ll post some detailed analysis in the coming days, but first I need to beat the Christmas traffic and get home!

  10. @Chrishornet

    It is well researched fact that the more information a person had the more likely they were to vote Yes in the referendum.

  11. Ah yes, balance. That easy to define, objective measure of coverage and debate.

  12. I have now decided to vote for the party that will fix the NHS, after my recent stay and coming out in a worst state than when I went in I can truly say the whole thing is a shambles.

  13. Since he beginning of October over 10 different pollsters have conducted 118 voting intention opinion polls. Only two of those polls have recorded Labour at less than 30% and those were the two monthly IPSOS/MORI polls. None of the other 116 polls have done so.

    Hmmmm, as I think they say around these parts.


  14. @Chris Hornet

    The ICM adjustment was designed in an era of three (or even two and a half) party politics to establish the intention of voters who were reluctant to disclose their intention and who (mainly) instead answered “don’t know” or refused to say.

    Nowadays we’re in an era of five party politics (or six in Wales and Scotland) with much reduced party loyalities and I suggest that a “don’t know” response is much more likely in such circumstances to mean just that. For no party is that more likely to be the case than for a 2010 Lib Dem. Their positioning in the political spectrum plus their old mantle of the “none of the above” party opens up numerous choices. Is that DK between Lab/LD? Or Lab/UKIP Or LD?UKIP? Or LD/Green? Or Lab/Green? Or Con/LD? Or Con/UKIP? Or just that a respondent knows that they’re not going to vote LD anymore but hasn’t thought much about the alternatives? And then add in a whole load of permutations for the nationalists. Lab/SNP? Plaid/Lab? Plaid/Lab/Green? UKIP/Plaid? etc.

    So (a) the don’t knows are more likely now to be just that and (b) they’re less likely to revert to their former party than previously. So adding back 50% to their previous party and 0% to the rest seems bizarrely crude. It’s a bit like the pollsters are preparing to fight the last war rather than the next.

  15. Just to expand on my last comment – because it seems more pithy than I intended it to be – it was nigh-on impossible to achieve true balance in the IndyRef debate, even for those who genuinely wanted to do so.

    It’s relatively easy to give each side an equal opportunity to put forward its argument (though there are many divergent opinions on whether that happened, and if not who it favoured). But it’s far more difficult to determine how to equitably treat both sides in a debate between the status quo and a huge change to it.

    We had reams of data on how Scotland has done under the union (much of which can be and was used as a case for independence), but no modern data on how Scotland has done as an independent nation. That was always going to create an asymmetrical discussion.

    Coupier’s liberal use of the word “fact” illustrates that as strongly as anything? Anyone from either side of the debate could point to facts about the union, and argue that these were good or bad, but by definition any discussion about an independent Scotland being either better or worse than the union was hypothetical. Therefore, how on earth do you define “information”?

  16. It’s unquestionably crude Phil, and I certainly agree that 50% is now too strong a weighting.

    All I’m saying is that until we actually see a FPTP campaign in which VI does not shift disproportionately towards those parties that can win significant numbers of seats, a reasonable argument can be made for them sticking with their basic principle of strong reallocation.

    The other point I would make is that it remains the case that a supporter of a nationally growing smaller party that has no hope of winning the supporter’s constituency, is more likely to say “I’m going to vote for [super cool alternative party]” and then change their mind later for tactical reasons or having learnt something about the party that they’re not so keen on, than say “I don’t know” but then go on to vote for the super cool alternative party.

  17. Looks as if Ken McIntosh Labour MSP for Eastwood has been persuaded to stand down in the New Year. Murphy will fight that seat in a by-election meaning Murphy will be at Holyrood before the GE. And a new Labour candidate will be found for Eastwood. That must be a GE opportunity for the Conservatives.

  18. Couper,

    When the price of oil goes up (should that be if) I would share you optimism about what it would mean (for an independent Scotland) for the UK.

    First, Osborne may have got round to implementing the Wood Review recommendations. If so there will be better regulation. The industry will be co-operating and trying to maximise recovery. Technical advances are likely to make small fields economical. Co-opedration may help to reduce costs.

    The PWC study in 2011 found that the industry in that year contributed around £30 billion in taxes. In an independent Scotland that would not all come to Scottish revenues but most of it would. From memory, Professor Kemp suggested that from what was currently in production, independent Scotland would get 90% of oil revenues and 60% of gas revenues. At a guess that might be somewhere in the region of £20 to £24 billion. A very large slice of the Scottish budget. Much of this, but not all, is unaffected by what the price of oil does.

    In 2011 the taxes on the oil/gas industry was around 80%. The industry postponed (cancelled?) around £6.8 billion of investment.
    With a more stable fiscal policy and taxes set lower oil and gas exploration and production can pick up.

    I read that the developing countries will continue to increase fossil fuel use. Hope so.

    Of course, in an independent Scotland we might not be starting from here. We might not have needed a Wood Review. I am teasing. We might have made the situation worse than it is. Whispers – doubt it.

  19. @ Dave,

    Labour, obviously. Plaid just don’t seem to be getting much of a surge in the new era of six party politics, and anyone who’s voting for Leanne Wood’s Plaid Cymru is definitely an anti-Tory voter.

    Warwick & Leamington has a very high Green vote compared to the other Lab/Con marginals in the sample.

    And both parties have limited resources and won’t be getting much national press attention, which makes it hard for them to mount a strong campaign in seats that are obviously unwinnable. Ukip will probably be squeezed too, but they’re better resourced and they get far more media coverage so the main parties may have a tougher time with them.

    I wouldn’t count these two seats as firm Tory holds at this point.

  20. @Bluebob

    “I have now decided to vote for the party that will fix the NHS”

    Does this mean you won’t, after all, be voting for those who farm out services to Capita and their like? Might I recommend the Greens?

  21. @ Alec,

    You may want to have a look at this week’s Scottish edition of That Which Must Not Be Named.

    I dunno if it’s a good idea for Labour to bang on about the SNP’s oil revenue overestimates, but it’s definitely the strategy they’ve adopted…

  22. @Couper

    It wouldn’t surprise me if Jim just pops his name on the top of ‘the list’.

    (Is it a prerequisite that list members must battle for a constituency before being offered a list seat?)

    If so, that’s the one down side of the list system. They should tweak it so that list placement is based on narrowest defeat at the top and so on. At this rate, we could have a First Minister (if Jim got elected by list and won the election) without anyone actually voting for him.

  23. Lots of fascinating material in Lord Ashcroft’s latest set of polls of marginal seats. I’ll need to do quite a few calculations before I can make sense of it all.

    However, at least one result struck me because of the way in which it chimed with an analysis I had posted about a month ago. I had argued that Labour was likely to underperform the UNS projections in a number of the current battleground seats. The reason for this is that I noticed that several of these seats had relatively low LibDem vote shares in the 2010 election. Since Labour’s current VI partly comprises ex-LD voters, the party’s fortunes in individual constituencies is likely to vary depending on the size of the 2010 LD vote. In seats with a significant LD presence, there are presumably numerous itinerant ex-LD voters available to boost the Labour VI. Where the LibDems did less well the Labour VI boost was likely to be less impressive.

    Because the LD presence was relatively low (<15%) in several battleground seats I argued that Labour might perform less well than the UNS model suggests.

    One of the seats I had identified as fitting this pattern was Harrow East. This seat should fall to Labour with a swing (to Labour) of about 3.6%. Given the current polling averages of +2 for Labour, this should be a labour seat according to the Uniform Swing model. Indeed, if Labour’s national figures are currently being held back by big losses in Scotland, and if Labour is therefore more than two points ahead in England, it would have to be classified as a secure Labour gain. (I note also that yesterday evening Electionforecast had this seat down as a Labour gain).

    However, as this is a seat that had a <15% LD vote in the 2010 Election, it was a seat that I identified as being more than usually difficult for Labour to bag. I was therefore very interested to see that the Ashcroft snapshot had this seat categorised as a Tory hold.

    (It is a pity he didn’t look at Watford, as this was a rare example of seat in which I calculated that Labour would do *better* than the UNS projection – based on the available of ample 2010 ‘wandering votes’)

  24. MORI poll repeats the now regular differentiation between Scotland and England as to absolute certainty to vote in the forthcoming GE, that we have become accustomed to across various pollsters in recent months – Scotland 82% : England 56%.

    However, it seems that not all Scots are equally engaged with all aspects of politics.

    Paul Hutcheon reports on the less than overwhelming response to the LiS leadership election from union members.

    “Scottish Labour did not provide turnout figures for either the members’ section, or for the unions.

    Nor were figures provided for the total number of people who voted in each part of the college.

    Such a basic disclosure, I am told by party sources, could prove embarrassing.

    However, some unions have embraced transparency and coughed up their turnouts. Here they are

    1. ASLEF: 21%
    2. Community: 12%
    3. Unite: 10.2%
    These numbers are abject – around 90% Unite’s political levy payers did not vote– but at least these affiliates have provided basic information.”

  25. Typo last post (last para) “If so,” should read, “If not,”

  26. Typo last post (last para) “If so,” should read, “If not,”


    At least SLAB had an election this time.

    The Brown non-election was slightly different, in that he was the only candidate, rather than the most popular figure in the party. The impression I got from the election was that another candidate might cause division, whereas Sturgeon was actually popular.

    I still think that they should hold some form of confidence vote in the event of no other candidate, and at least demonstrate the level of support for the candidate. It would be shocking to think anyone got elected to such a post with less than 50% confidence (unlikely, but always reassuring to see the data).

  27. Could we be seeing the UK bubble start to burst and a small sign of life on the lib dems heart monitor?

  28. By UK I of course mean UKIP


    The lists don’t work that way if a list candidate resigns the next on the list takes his/her place. Murphy couldn’t get into the Scottish Parliament on the list until 2016. McIntosh and Murphy could in effect swap jobs – McIntosh stands down and fights Eastwood in May and Murphy stands in a by-election for Holyrood in the New Year.

    Both seats Westminster and Holyrood have Conservatives in second, whether the Conservatives will bother putting up a fight I am not sure.

    The Ipsos-Mori Scottish crossbreak SNP 52% Labour 15% – a reverse bounce. (joke)

  30. There is more spin than substance to Ashcroft’s claim that there is ‘moderate Tory’ cheer in the latest poll. There are very few signs here of a Tory majority and largest party still looks like being a struggle. The persistence of their ‘polldrums’ even after set-pieces, such as the Autumn statement, which was expected to generate some momentum, has got to be worrying them. One does have ask the question: what will it take to get their VIs to rise anywhere near what they require for a majority?

  31. These numbers are abject – around 90% Unite’s political levy payers did not vote– but at least these affiliates have provided basic information.

    For what it’s worth OldNat, I don’t think it’s a politically shrewd idea for the SNP to attack trade unions, even in cases where doing so is entirely justified. That said, I agree with you. If EM’s policy of the political levy no longer being compulsory has taken effect then those figures are not only derisory, but truly bizarre.

    Union affiliated or not, Labour supporting or not, I’ve never understood the concept of having a vote and not making use of it (be that for or against strike action or in a leadership election). I’ve argued for four years that Ed Miliband is leader of the Labour party because of non-Labour trade union members deciding that voting for him would help their favoured parties.

  32. @ Chris Hornet,

    If Labour trade union members can’t be bothered to vote for their leader in Scotland at a time when the party is in crisis, many trade union members don’t even bother to vote on strikes that will stop their pay, and 30% of the Labour Party members who pay £45 a year for the privilege of voting in leadership elections didn’t bother to cast a vote in 2010, what is the likelihood that non-Labour trade union members sat down, compared all the candidates on the Labour leadership slate, calculated which one was most likely to help their respective parties (which would differ from trade unionist to trade unionist), and collectively decided to throw the election to Ed Miliband?

    Honestly, have you been talking to [email protected] H0dges?

    If your reading of Ashcroft is true, what does it tell us about Labour?
    I would say that their possibilities are equally “dicky”.

    They cannot both win. The point is, it does not put us any farther forward. The loose talk yesterday about “Labour stabilizing” ect ect, was the usual over zealous reaction that habitually occurs when a slightly optimistic poll is published.

  34. “I would say that their possibilities are equally “dicky”.”

    You may well say that.

    And only the polls and the voting imbalance in the constituencies would suggest that you are wrong.

  35. missis minty

    I expect the more devious TU members voted for Dane Abbott [the next London Mayor……….] in 2010.

  36. Unicorn – totally agree.

    Labour to take net 20-30% LD votes more than Cons per seat.

    The LDs are already squeezed below 23% by definition in virtually all Lab/Con marginal seats.

    Even without Ashcroft polls as Evidence in Harrow East 25% (using midpoint) of the LD Vote share is only 3.75% – gving Labour just just over half the swing it needs.
    Labour need more net direct Con-Lab (or less losses to others) in this seat (and others if the LD share in 2010 was below 23%) than UNS suggests.

  37. Spearmint – on the one hand you have the one candidate you’ve actually heard of: the heir apparent to Blair in most ways, minus the ultimate responsibility for taking us to war with Iraq. On the other you have a slightly weird-looking guy who is being rammed down your throat by the union and just happens to be the Blairite’s baby brother.

    Hardly rocket science for those of that mindset, and given how close the final result was we’re not talking about that many people.

  38. PAUL
    What polls exactly tell me I am wrong? The one on this thread would it be? What about the latest IPSOS which gives the same figures in reverse. Tories 3 points in front. So what does that prove do you think.
    Also, you better hope that Superjim works in Scotland over the next 5 months. Furthermore, when will Milibands ratings begin to look anything like a PM waiting to happen.

  39. Ipsos Mori poll

    Con 32% (nc)
    Lab 29% (nc)
    UKIP 13% (-1)
    LD 9% (nc)
    Green 9% (+2)

    Just one poll. But a three-way contest for third place? Have both LD and Green been simultaneously this high before?

    I think Labour squeezing back tactical voters depends a lot on any perception that the voter’s preferred party might beat UK nationally for the important third place position. Tactical voters might see it as a return on investment worth some risk. I’m not saying that average voters think this through, but tactical voters are, well, tactical.

  40. I think Watford would be a very interesting one for Ashcroft to poll more often.

    Obviously I’d say that even if it were a boring seat, but my main reason is that according to his November poll the Lib Dem candidate’s personal vote is worth 13 points (based on the difference between what party you’d vote for and what candidate you’d vote for in your constituency – with Labour and the Tories quite close to one another there’s relatively little scope for tactically voting Lib Dem). That’s the sort of discrepancy that needs to be mapped with four or five post-selection polls to determine just how big Thornhill’s personal vote is, and which of the big two parties it might hurt if she doesn’t win.

  41. Found my own answer. It has happened or nearly happened before but always Ipsos Mori.

    For example 12 July:

    UKIP 12%
    LD 8%
    Green 8%

    Perhaps this is primarily a house effect. In polls other than Ipsos Mori the LD vs. Green zero sum has been much stronger.

  42. @RolandGatinoise

    My focus on the Tories is purely on the basis that they are doing worse than I would have expected by this point in the electoral cycle. You are right – Labour doesn’t look like getting a majority either, but I had never expected them to.

  43. ChrisHornett
    Have you ever belonged to a trades Union?
    Frankly the idea that individuals dance obediently to the tunes of their Union’s executive is laughable.

  44. Frankly the idea that individuals dance obediently to the tunes of their Union’s executive is laughable.

    Where did I imply otherwise?

  45. ChrisHornett
    Or are you saying non-labour members of affiliated unions opted in and voted for EM to stop Labour winning the GE?
    Could they really be that bothered?

  46. @Spearmint
    W&L I can’t see 693 Green votes making much difference to a Conservative majority of over 3000. I suppose you are basing “a very high Green vote” on the poll, which gives them 7%.

    WC&SP Judging by the 2014 MEP voting, Plaid is already squeezed, and not by Labour. (I’m basing that not on the results for Wales, but on the local count)

    On my conviction that working with percentages is liable to give a distorted picture compared to working with actual figures, I’ve recast Ashworth’s Carmarthen poll to a base of 40000 voters, the actual 70% turnout in 2010.
    T2 is his Table 2 with a sample size of 1000
    T3 is his reworked ‘turnout weighted’ Table 3 with a base of 493, (rounded to 500)
    T7 is Table 7 with the %s taken as %s of 40000

    CON LAB LD UKIP PC G BNP Ind Turnout
    2010 16649 13226 4890 1146 4232 – – 364 40507
    T2x40 7840 8320 1320 5040 4200 1080 160 480 [40000]
    T3wx80 12160 11680 1520 6720 5280 1280 240 560 [40000]
    T7 13200 11600 1600 5600 6400 1200 – 400 [40000]

    It is interesting to see that without any adjustment the T2 figures bear little resemblance to 2010. It is hard to believe that the Con. vote would be halved.
    Successive adjustments in T3 and T7 get closer to realism for the two main parties, but have less effect for smaller parties, though I can’t see Plaid increasing its vote by 50%. The more ‘adjustment’ the bigger the Tory vote.
    There was no Green or BNP candidate in 2010, and there is unlikely to be either in 2015. In 2010 there was an independent, and I have used Ind for ‘other’.
    The BNP ‘voters’ were probably taking the mickey.
    (I live in West Wales, though not in this constituency.)
    I hope the table layout stands up to posting!

  47. Or are you saying non-labour members of affiliated unions opted in and voted for EM to stop Labour winning the GE?

    In most cases there wasn’t an opt-out from the political levy in 2010. That process was initiated by EM after he became leader.

  48. It doesn’t!

  49. I’d have thought the most logical explanation was EM was meant to draw votes from Balls to let DM through the middle but it went wrong.

    “I don’t think labour will be bothered about the Ashcroft poll, they seem to change radically poll by poll. ”

    Well that’s the thing. Are Ashcroft’s polls all over the shop because they’re all over the shop or are they picking up what the national polls aren’t i.e. the early stages of political balkanization.

1 2 3