New ICM/Guardian poll

I was hoping for a new Populus poll today, but instead we have the monthly ICM/Guardian poll. The topline figures with changes from their last poll a month ago are CON 41%(+2), LAB 27%(nc), LDEM 20%(+2). The poll was conducted on the 10th and 11th of July.

Both the Conservatives and Lib Dems are up slightly at the expense of “others”, who are down on 12%. This is the lowest they’ve been since the Telegraph began printing details of MPs expenses, though it’s worth noting that ICM never recorded the very high levels of other support that some other pollsters did in the first place – they peaked at 15% in ICM, compared to 23% with YouGov and 30% (!) with ComRes. We should wait to see some declines elsewhere before concluding that the tide has turning.

The other questions in the poll looked at spending and the renewal of Trident. The Guardian’s report says that more than two-thirds of respondents wanted to see the government cut spending, and only 42% of people wanted to see Trident renewed.

106 Responses to “New ICM/Guardian poll”

1 2 3
  1. Be interesting to see what Populus say but I think the scene is set.

    It seems that the protest over expenses may be backing off

  2. Tories above 40 was a bit unexpected. The last Yougov also had them on 40. It would useful to see what other pollsters show up in the next days/weeks before drwaing any conclusions as to the current state of the parties. I was rather please with my ever brilliant prediction though, I was only 1 point out on the parties.

  3. The important thing here are the changes from the previous ICM poll. They suggest that the next Yougov poll will be Cons 41, Lab 24, Lib Dems 19/20.

  4. I tend to pay more attention to YouGov and ICM myself but there should be a Populus out tonight for The Times.

    Will it support the fall off in the OTHER vote is what I want to know.

  5. I was expecting minor lab recovery to be at the expense of cons, seems like lab are/have scrabbled back to respectability but not by hurting cons, mebe UKIP voters really do all go back to cons afterall.

  6. Interesting to see that the LAB vote is unchanged. I’m very intrigued to see the next few polls as this will surely begin to shape what we might expect to see next year.

    My prediction for a 50-60 seat Conservative majority seems to be holding firm

  7. Does anyone know what the breakdown is of ‘others’…. I am becoming more interested in them that the G2+1 which are staying pretty stable

  8. The Guardian are leading with figures over nuclear

    disarmament. I have to query their findings, considering that a

    Populus poll last month had 48% of 2005 Labour voters said

    it would make no difference to their voting intentions. 11%

    said it would weaken their intention to vote Labour.

  9. If this signals the direction of the post Euro vote/post expenses polls , then it looks much better for Cons than Labour.

  10. Just as with the ICM poll on Afghanistan, and the one last week about public spending / taxes, it seems that if you define the question tightly enough you can squeeze out the useful but “undesirable” responses.

    For example, here the Grauniad have deduced taht the lack of support for Trident means that people now oppose an independent nuclear deterrent. Actually, that does not necessarily follow.

    I for one am in favour of Britain possessing an independent nuclear capability, but question whether Trident is necessaril;y the right solution for Britain.

  11. Good results for the Tories and LDs. It’s very serious news for Labour if they get stuck at around 27% rather than moving up to 30%.

  12. Surely much depends on the starting point for any comparison?In relation to the ICM poll at the end of May which put Labour in 3rd place on just 22% , Labour have gained 5% – the Tories 1 % – whilst the LibDems have lost 5%!

  13. If the others figures are correct then it looks like the travel back has happened for the Conservatives. For Labour, the travel would have to be BNP/UKIP/Plaid/SNP. How much will move back is vital as it appears we have reached what might be termed the core vote. The LibDems in the high levels is the different factor here and if it continues I think it ends the days of parties winning at around 45%. It might explain why the LDs have been attacking Cameron more in the past few weeks – maybe they feel that they have to protect their position in their shire seats.

    It might also explain the focus on issues like Aid as this is something that Cameron may hope will take votes away from the LDs. I wonder if we are in for vote shuffles between the three parties for a while as each tries to attenuate their messages to try and increase support.

  14. The most significant outcome is that most voters support spending cuts. So the economy is still a dominant backdrop to overall support for all parties. Labour’s fortunes are still attached to economic recovery.

  15. The other questions in the poll looked at spending and the renewal of Trident. The Guardian’s report says that more than two-thirds of respondents wanted to see the government cut spending, and only 42% of people wanted to see Trident renewed.

    If memory serves me well, back in the ‘Eighties polls suggested only 28% of the population wanted to keep Northern Ireland as part of the UK. And the constitutional status of Northern Ireland is…?

    Trident+ is the only solution on offer. When backs are to the wall people will understand this.

  16. This is a good poll for the Tories. Over the last month, Labour have had a reasonable press (certainly by recent standards). Tory spending plans have been under more scrutiny, as has Andy Coulson. There have been no horror stories about the economy. Even Mandelson has been off our screens. And yet their share of the vote remains unchanged while the Tories gain two points.

  17. Fluffy thoughts. Your logic is flawed. Trident and its successor make sense only as a “we’ll all go down together” if deterrence fails, or as a means of projecting power. The threats to our way of life today come from non-state agents or from states that do not have nuclear weapons. There is no particular reason why we should not be part of the US nuclear umbrella as is the rest of Nato. The sole exception – France – has gained nothing from being an independent nuclear power except an unwarranted place on the UN Security Council and this will be lost once the EU secures a place.

  18. It would appear that the polls are gradually returning to the position before the MP expenses affair, but Labour have not recovered as much as Tories relative to levels seen in first quarter.

    This weekend, it will be three months since Labour last scored 30% in any poll. Since the first Brown bounce collapsed in autumn 2008, Labour have only improved on their 2005 GE score (36%) in 3 polls, the last of which was on 16/11/08 (37%). The average over last 10 polls is only 24%, and over last 20 it is 23.05%. Next week, on the day of the Norwich North by-election, it will be exactly 18 months since Labour last led – by 1% at 38%/37% – in a single poll.

    Conversely, apart from the ComRes poll on 31 May this year (30/22/18), and the Harris poll on 16 June (35/20/16), Conservatives have been at or above the 36% with which Labour won in 2005 consistently since October 2007, and for most of that period have been above 40%. So far this year 17 polls have shown Cons below 40% compared to 44 at 40% or more, of which 7 at 45% or more.

    Only four polls this year (including that ComRs poll on 31 May) have shown the Con lead over Lab in single figures, compared to 5 with a lead of 20% or more. The average lead in the last 10 polls is 14.5%, and over the last 20 it is 14.1%. (The 10 before that had an average of 17.5%, but the 10 before those had an average of 13.2%).

    For the LDs, while they had been in the mid-low teens in December / January, they appear to have settled at +/- 20%. This is below their actual result in 2005, but at a level which suggests that they could achieve 22-23% when the GE is eventually called.

    We may yet see a last hurrah from the government, but it would seem that Labour are going to struggle to reach close to 30% at the next election, and may well drift back to low 20s. We could therefore see a close battle between Lab and LDs in terms of share of vote – but Lab would have to do significantly worse and LDs significantly better for the parties to get even close in terms of seats won.

    Given the above, it is not surprising that the city and civil service, not to mention the general public, are all assuming that there will be a change of government within a year.

  19. @Graham – apologies Graham but the poll you are talking about was likely a rogue (compare it to the poll averages for the period) and the starting place for labour is not actually that different than it is today.

    Good result for Con’s and Lib Dems – 3rd place for Labour is getting closer :-)

  20. @Paul H-J – Good post

  21. That Cameron will be PM in less than a year is now considered a done deal by almost everyone. The real question is how badly will Labour lose. The Weighted Moving Average is 39:25:19 so Labour support is holding firm but as noted previously I think that “Other” is more likely to swing to C than L.

    Also there will be further ministerial resignations in the coming months (how many have resigned this year – 8,9? I’ve lost count) and morale in the Labour party is rock bottom. The 0% debacle is hurting badly and Darling is concluding, rightly, that he has to get a grip because no-one else is.

  22. Good post Leslie.

    Trident is argued for mainly as it helps some people pretend that we are still a major player in world events (like, that’s why we are in Afghanistan). The money wasted on Trident can be much better spent elsewhere; pensions, fuel poverty, jobs, lowering national debt…

  23. @Jack

    “The money wasted on Trident can be much better spent elsewhere; pensions, fuel poverty, jobs, lowering national debt.”

    And also a much more effective conventional defense capability. Thanks for your comment!


  24. Liberals increase probably due to Nick Clegg: has ‘come out’ more into real issues and made these his own, not least Gukhas and now Afganistan (moving the news agenda in his direction in some cases, which isnt bad for a Liberal leader post Kennedy frankly).

    Paul H-J: good posting, but can we really expect Labour to drop in their core vote to such a position where the Liberals overtake them in the national share of the vote? I don’t see it personally because Brown has spent much time reinforcing that core base around the country…”labour investments…” artifical debate etc helps illustrate this.

  25. Kier, Thank you.

    NBeale, – I agree that Darling seems to be the only one in the cabinet who has any idea how bad the position is and that there is nothing to be gained by pretending otherwise.

    It may of course be that after a decade of being kept in the dark, most Ministers have no clue as to the connection between the PSBR and their own departmental spending allocation.

    It will be interesting to see whether the emergence of a back-bone will help Darling retain his seat. I don’t know whether there is a precedent for a sitting Chancellor to be defeated in his own constituency, but it certainly has not happened in the past 50 years.

  26. So, it’s all down to what happens to the ‘other’ votes … Granted, it’s difficult to see many of them going back to Labour (at least, among the ‘certain to votes’) but I think it’s still too premature to be saying that the Tories will win the election. Yes, it’s very easy for Labour to lose it – but winning is a much tougher matter, and the yawning chasm of a hung parliament still seems to me to be the most likely outcome.

  27. Paul H-J

    I don’t think a sitting Chancellor has ever lost their seat. I can find two 19th century ones who lost the seat they were sitting for, but it was in the era of general elections taking several weeks and people standing for multiple constituencies, so they were both returned for alternative seats. I can’t find any Chancellors who were properly ousted!

  28. Dean,

    Look at the bald numbers for votes received rather than %.

    1997: Lab 13.5m; Con 9.6m; LD 5.2m
    2001: Lab 10.7m; Con 8.4m; LD 4.8m
    2005: Lab 9.6m; Con 8.8m; LD 6.0m

    The Lab / LD gap in votes has reduced from 8.3m in 1997 to 5.9m in 2001 and just 3.6m in 2005.

    The gap is reducing at a rate of 2.3m per election, so on that basis, LDs will not overtake Lab this time. But it is not inconceivable. Twice before, the Liberals / Alliance reduced the gap between themselves and Labour by more than 3.6m.

    From 1970 to Feb 1974, Lab dropped 0.4m and Libs added 4.0m to give 4.4m reduction;
    From 1979 to 1983, Lab dropped 3.0m and Alliance added 3.5m to give 6.5m reduction and create a Lab/LD lead of just 0.7m votes – less than the Lab/Con lead in 2005.

    In order to overtake Lab in terms of votes LD need to add about 2-2.5m votes if Lab lose up to 1m votes.

    If however Lab lose more than 2m (and Cons lost 4.4m between 1992 and 1997), then the additional votes LDs need are correspondingly reduced.

    In 1997, Cons lost about 31% of their 1992 votes. If the same ratio were applied to Lab vote in 2005 (which is consistent with current polls) that would put Lab at around 6.6m votes, so LDs only need to add about 650k votes net to be ahead..

    Those are the mechanics. Do I think it will happen ? No.

    The main reason being that in 2005 LDs won 528k votes in Scotland and came within 400k of Lab, who fell below 1m in Scotland for only the second time since 1935 – the previous time being 1983 when Alliance took just under 700k, and the only other time since 1931 when Liberals had been within 0.5m votes of Lab in Scotland. While I can see Lab vote falling further in Scotland next time – perhaps by as much as 200k, I believe the LD vote will also fall in Scotland, and maybe by as much as 250k. They will obviously need to increase their vote in England by up to 1m to compensate for that, which I cannot see on current trends.

    Realistically, I expect the result next year to be:

    Con: 10.5-11.0m up approx 2m
    Lab: 7.0-7.5m down approx 2.25m
    LD: 5.0-5.5m down approx 0.75m
    Others 2.5-3.0m up about 0.75m

    Obviously plenty of room for movement, subject to polls and turnout, but the above translates to roughly C 41; Lab 27; LD 20 – and what a remarkable coincidence that is.

  29. Amazing how the polls change from one day to another.
    A little spin here a little spin there a new initiative a new policy and hey presto we are vinning, we are loosing. Just shows how gullible the electorate is.

  30. Interesting post Paul H-J.

    My own prediction (since last year) is for a Conservative majority of 53 seats. I can’t see it being a landslide based on recent polls.

    Your prediction of C 41:LA 27:LD 20 gives a majority of 72 seats.

  31. A drop from 36% to 24% for Labour based on the turnout number being the same would see a Labour fall in votes of 3.2 million.

    I think there is a strong possibility Labour will get about 6.4 million votes. The Lib Dems should gain at least 5.4 million votes. The Cons should receive at least 9.4 million votes.

  32. @COLIN 1.23pm

    Colin-could you find a name which shows that you aren’t me?


  33. @ Keir,
    It may well be that there was something a bit rogueish about the late May ICM poll – but others said the same thing about the Guardian’s June poll when it gave Labour 27%.
    Of course, if the latter suggestion is true and the real Labour level was more like 25% a month ago ,perhaps the July ICM poll does reflect an underlying rise in Labour support as well as for the other two parties!

  34. Graham,

    The convention is to count rise/fall in the vote from the last poll by the same company, or the last poll commissioned by the paper. You can make the polls say anything you like if you change the terms of reference enough.

    If you take the starting point from 1996 then Labour have lost nearly 30% according to this poll – ouch!

  35. Neil,
    It’s simply a matter of trying to establish where we are based on the best evidence available. Trends do require some analysis, which itself sometimes demands delving a bit further than the headline figures.

  36. Andrew,

    It depends what you mean by a “landslide”. A shift of over 100 seats in a single election would generally meet that description, so for Cons to gain a majority of 1 would effectively be a landslide. If you are thinking in terms of a Con majority over 100, then that implies 375 seats – a gain of 177 from 2005 – or 161 on the notional results for the new boundaries.

    However, even if you require a Con maj of 100 for your landslide, this poll is not far off it – just adjust the figures to C 43, Lab 27, LD 20 and hey presto.

    One cannot predict a landslide confidently since it only takes minor changes for for the numbers to shoot up – or down. Plus, the swingometer, while fun, cannot take regional variations into account. As the polls currently stand, a landslide is at least as equally possible as a hung parliament, probably more so.

    One reason I point people to the number of votes as well as share is that we will not have anything remotely like “Uniform National Swing”.

    There will be a fall in the total Labour vote. If Labour drop from 9.6m to c7.3m, that works out at about 3500 per seat. Likewise a Con increase of about 2m is around 3100 per seat.

    UNS would suggest any Lab seat with a majority of under 6500 over Cons will fall. That would indeed give Cons somewhere in the region of 150 gains and a majority of about 50-60.

    But actually, given the differential turnout which occurred in 2005 – whereby turnout was much lower in Labour’s heartlands relative to the “marginals” – it is more likely that we will see Lab vote drop by only a few hundred in some seats, and by 5000 or more in others.

    Quite where Labour lose their votes will have as much impact as where Tories can add them.

    The only things I am prepared to predict with confidence are:

    (a) Cameron will be PM after the next election; and
    (b) it will be a very interesting night.

    [(a) will happen even if Cons fail to reach 326 seats as Clegg would be nuts to keep Brown in power, assuming he could.]

  37. Thanks again for an interesting post Paul H-J. I am absolutely fascinated by the statistics of our GE system and agree with your point (b) – it is going to be an incredible run-up to the GE and a very interesting night.

    Your point about relatively small shifts in opinion polls leading to wild variations in results is also valid. A 41:27:20 split would give a Con maj of 68 but a relatively small shift to 44:23:21 would turn that to 190!

    I agree that a Tory majority of 50 would be an incredible result from the last GE but for me a landslide means 150 seat maj or more.

  38. Paul H-J

    Fantastic postings.

    I disagree with your point a however- it isnt all that cut and dry. Because when you wrote “(a) will happen even if Cons fail to reach 326 seats as Clegg would be nuts to keep Brown in power, assuming he could.”

    I would point out that the Liberals deal breaker will be some kind of PR voting system or federalism- and no Conservative leader will ever EVER introduce any of that. Clegg certainly wont be lending support to D.C in the even of a hung parliament that is in my opinion for sure.

  39. Poupuls Prediction:

    Con 42
    Lab 25
    Lib 21

    GE Prediction: (I will be right because I’m Brilliant at Poll Predictions – due to my “superb political emotional intelligence” )
    Con 44/45
    Lab 26
    Lib 21

  40. Mean’t Populus, sorry….

  41. I’m getting worried. No doubt I’ll be out trying to sway people from a wasted Tory vote. Their latest policy on Aid has enraged me so much >:(

  42. If it were a hung parliament, Tories would be (presumably) the biggest party, but would refuse the LibDems’ demands for electoral changes.

    Labour would refuse as well, I’d guess, leaving Tories to govern with minor parties if possible, or perhaps call another election to boost their numbers in the style of 74?

  43. Well I said there would be more resignations, but even I didn’t expect another one today. Brown treats Health and Defence with contempt – so people who understand, quit.

  44. NBeale,

    Have you noticed how it is the Ministers in the Lords who are quitting now ?

    Since they don’t need to spend more time with their constituencies nursing their majorities, what does that tell us about morale at the heart of the government ?

  45. New Labour will be lucky to hold onto 30% share of the vote at the next GE, or so Paul H-J’s stats tell me.

    Further on this poll, isnt this the first time since early June (4th?) that the Liberals managed 20% in the polls (of any major pollster)

    Richard Manns- that might be one course of action, but I personally think that there could be much to gain politically from having a minority government (of any party) after 2010- it might enable all kinds of reforms, and consolidations of our abused constitutional customs…

  46. Dean,

    The only way that Cameron will not be PM after the next election is if:

    (a) Brown retains a Labour majority (very unlikely); or
    (b) Labour, although short of a majority are by far the largest party and Brown can either cobble a coalition with minor parties (eg SDLP, Plaid etc) or survive as a minority administration; (unlikely) or
    (c) Lab + LD gives a majority – and Clegg is willing to support Brown continuing as PM – a reckless stance for LDs..

    (c) could theoretically happen even if Con have more seats than Lab.

    Cameron does not need LDs to support him should he fall short of a majority so long as Lab + LD is less than 325. The key figure for Cameron is 300. That is because NI + SNP/Plaid + sundry others are likely to be at least 25 (if not 30+) which would deny Lab+LD a majority.

    If Con are over 300, then in all probability, Lab will be below 275 unless LDs have done v. badly.

    Even if the figures are C 290, Lab 280, LD 50, others 30, it would be reckless for Clegg to keep Brown in No 10. In that case Clegg will be setting LD MPs up for slaughter since a further election is likely within a year, and the public will not forgive LDs for inflicting more uncertainty on the country.

    It only makes sense for LDs to support a minority Lab govt if Lab have more MPs than Con and a Lab-LD coalition could likely survive more than a few months. That realistically needs Lab to be at 300 or more with LDs on 35+ such that there is a working majority over 20.

    Remember that a repeat election within a few months, while putting all parties under financial pressure, could give LDs a bonus in terms of seats won from both Con and Lab as they will have much better information as to where to concentrate their limited resources.

    FWIW, I don’t think any of the above will apply since if Cons have made 75+ gains needed to get from 214 to 290, Lab will likely have fallen below 280, and possibly lower. That would have happened with a Con lead of 6-7% as opposed to the 14% current average lead.

  47. Graham has a point in that the 27% in the previous Guardian ICM for Labour was called, probably fairly, as overstating their support but within moe by many on here.
    The next poll will tell us if there has indeed been a small rise for Labour as well from others.
    Also, I was surprised others are as low as 12% and would not be surprised to see this over 15 in the next few polls.

  48. PAUL H J,
    I don’t think there is any way that Plaid or SDLP would help to install a minority Tory Govt. Beyond that, would the SNP wish to be seen by the Scottish electorate as having helped to hand power to the Tories? I can think of nothing more likely to guarantee that Labour regains the political initiative in Scotland!

  49. @Dean Thomson

    I far prefer our adversarial politics; politics is a trade and we pay them in votes, and to quote Adam Smith:

    “People of the same trade seldom meet together, even for merriment and diversion, but the conversation ends in a conspiracy against the public.”

    Cameron could take note (and probably has) of the words of Churchill when a less adversarial hemi-cycle was proposed for the rebuilding of the House of Commons:

    “…we shape our buildings and afterwards our buildings shape us.”

  50. Brown has zero management skills – there is no way he could ever lead a coalition. He is also gratuitously insulting to the Lib Dems. Betfair Spread Betting gives a 25% chance of a hung parliament (personally I think that’s much too high – nearer 10% I’d say) and an 87% chance of the Conservatives having the most seats. With a 7% chance of a Labour Majority this means that there is a 76% chance (19/25) if there is a hung parliament that the conservatives will have the largest no of seats.

    PS Betfair also gives the Tories a 92% chance of taking Norwich North. Brown could be gone by Christmas.

1 2 3