1) Labour’s lead continued to fall

For the main horse race – who is in the position to win the next general election – the key point is always the lead in the polls, and throughout 2014 Labour’s lead over the Conservatives continued to fade away. In 2013 it fell from around ten points to around six points. This year the trend continued, with Labour’s lead fading from six points to just under two points. Given the complexities of the Lib Dem collapse, the rise of UKIP, a large number of new incumbents and the separate race in Scotland it is dangerous to rely upon on uniform national swing, but include all those factors and I think we are now into hung Parliament territory.


Labour’s lead has almost wholly been down to falling Labour support rather than increasing Tory support. The Conservative share of the vote started the year at around 32% and ended the year in roughly the same place, Labour supported started the year at around about 38% and finished the year at around 34%. While there is always some churn between different political parties and there will be some people who have moved from Labour to Conservative, it’s certainly not the main factor – rather what we’re seeing is an anti-government vote that had previously been going to the Labour party by default is now finding many homes and showing itself in rise of the Green party, the SNP and UKIP. The public’s lack of confidence in Ed Miliband and Labour isn’t manifesting itself in them running back to the Conservative party for safety… it’s manifesting itself in them going off to find more attractive oppositions to vote for.

This has in many ways been the pattern of the 2010-2015 Parliament. The Conservative party’s vote fell to around 32% early in the Parliament and has stayed there, seemingly immune to events, announcements, people or policies. Labour inherited a substantial lead early on thanks to the Liberal Democrat collapse and have watched it be nibbled away by rivals.

2) The Greens finally woke up

So to those rivals. The rise of UKIP has been covered by everyone, the most remarkable story of the Parliament. The increase in support for the Greens is a newer development. Earlier on this Parliament I was frankly surprised that the Greens were not doing better. They had elected their first MP, the government were implementing unpopular austerity policies and Labour were constrained in their opposition to cuts by a desire to establish their own economic credibility. Elsewhere in Europe radical left-wing parties were benefiting from an anti-austerity vote, yet here it wasn’t happening. The Greens were marooned on around 2%.


This year it finally did, and looking at the charts it appears to be the European elections – with the coverage and campaigning that it implies – that sparked the Greens into life. The European elections pushed them up to 4% or so, and since them other polls have shown them building on that, in many cases getting their highest levels of support since their first breakthrough back in the late 80s. What impact that will have at the election beyond providing a home for some people who might otherwise have voted Labour or Liberal Democrat is a different question – in 2010 the Greens managed to breakthrough and win a seat despite having a derisory national share of the vote. There has been an Ashcroft poll in one of their most viable targets (Norwich South), but it showed Labour well ahead, so it is possible that the increase in Green support may not translate into any extra seats.

UKIP meanwhile have managed to keep the bandwagon rolling onwards through 2014. There was an expectation that their support would peak after the European elections and then go into decline, but things were thrown off course by the defections and by-elections of Douglas Carswell and Mark Reckless which kept the party at the forefront of politics through the autumn and allowed them to finish the year with higher support than in January. I would still expect their support to be squeezed as we get closer to the election, as the race focuses more upon the binary choice between a Conservative and Labour led government but events, such as further defections now a by-election is no longer unavoidable, could easily push that off course.

3) Economic confidence began to stall

We started 2014 with people being increasingly positive about the state of the economy, but still pessimistic about their own finances, and pondered whether the improving economy would filter through to people feeling more positive about their own finances. What actually happened in 2014 was that perceptions of the improvement in the economy peaked over the summer and have now started to falter – the economic statistics, GDP growth and unemployment, may have remained strong, but public perceptions have started to go back down again. In August YouGov found 50% of people thought the economy was showing signs of recovery or on it’s way to recovery, by December that had fallen to 40%.


The impact of this is difficult to call as the economy is something of a two edged sword. Improving perceptions of the economy have gone hand in hand with a growing Conservative lead on the economy and that has remained steady… so far. If falling perceptions of the economy eat into the Conservatives lead on economic competence it will damage them. On the other hand, as the economy has improved it has fallen down the list of issues people consider important and become a less salient issue; if people worry more about the economy and it dominates their thinking more it may help the Conservatives…

4) Concern over immigration overtook the economy

Between 2007 and 2013 polls were consistent in showing that the public thought the economy was the number one problem facing the country. This year we saw it overtaken by immigration, and in the most recent few MORI polls it has been fighting with health for second place.


These top three issues are each strongly associated with a party – the public invariably trust Labour more on the issue of the NHS, over the course of last couple of years the Conservatives have built up a solid lead on managing the economy and UKIP generally lead the other parties on the issue of immigration. It is in the clear electoral interests of the Conservatives to have an election dominated by the economy, for Labour to have an election dominated by the NHS, for UKIP to have an election dominated by the issue of immigration. Over the last year things the issue agenda has clearly been moving in UKIP’s favour.

5) We found out where the Lib Dems were doing well and badly

In previous round ups like this I’ve always ended up saying how Lib Dem national support is in a dire state, but their MPs may or may not survive due to their personal votes and tactical voting. Through 2014 though we have got a lot more data on where the Lib Dems are doing well and badly and where they may be able to withstand the tide against them. At the last election the Liberal Democrats won 57 seats. Thirty-three of these are English and Welsh seats with the Conservatives in second place, and twenty-six of those we have Ashcroft polls for. Twelve are English and Welsh seats with Labour in second place, and we have Ashcroft polls for eleven of those (all but Bristol West). A further eleven are in Scotland, where the impact of surging SNP support remains to be seen and where we will hopefully have some Ashcroft polling later this year. Finally there is the unique Lib Dem vs Plaid seat of Ceredigion.

The average swing in the LD-v-Con seats is a modest 2.2 points from LD to Con, enough for the Conservatives to take seven seats. However, because the majorities and the swings aren’t evenly distributed there were actually ten seats where Ashcroft found the Conservatives ahead and three more (St Ives, North Cornwall and Torbay) where it’s too close to call). The line on the chart below is the swing needed for each Lib Dem seat to fall, the bars the swings recorded in the Ashcroft polling (when Ashcroft has done more than one poll in the same seat I’ve averaged them)


In the LD-v-Lab seats it is a different story, the average swing is a towering 12 points from LD to Lab, enough to win all the seats at a trot. Again, there is some variation from seat to seat, but this is only enough to save two of these seats – the once unassailable Old Southwark and Bermondsey, and Birmingham Yardley where John Hemming seems to be bucking the trend.


While national polls are never going to tell us too much about Lib Dem seat numbers, based on the Ashcroft polling The Lib Dems look set to loose around twenty seats in England and Wales, plus however many in Scotland (and given their dire performance in Scottish polls and the 2011 Holyrood elections that’s unlikely to be pretty). I think a fair assessment is that the Lib Dems start 2015 looking set to loose about half their seats at the election.

6) The SNP lost the war, but are winning the peace

Which brings us to Scotland. The aftermath of the referendum has been stark in terms of Westminster voting intention, with all polls since mid-October showing substantial SNP leads, ranging between 16 and 29 points. Thanks to the electoral system, if anything even nearly approaching this happens at the general election it will have a huge effect on seat numbers.


For the last thirty years the electoral system in Scotland has worked heavily in Labour’s favour – they have enjoyed around 40% of the vote in Scotland, the rest being split between the SNP, Lib Dems and Conservatives, who have all struggled to get more than a quarter of the vote. That has translated into a consistent block of forty-plus seats for Labour. If that flips round in other direction, with a large lead for the SNP, we can expect the electoral system to deliver a similar boon for the SNP. There are still many unknowns about the Scottish vote – the post-referendum SNP is a new development, we don’t know if it will last, nor do we really know how the vote will be distributed and how the SNP surge in support is distributed. The Ashcroft Scottish polling will at least tell us more on that front, at present we can only say that things looks very good for the SNP in Scotland, and very worrying for Labour.

279 Responses to “Six opinion poll findings from 2014”

1 2 3 4 6
  1. Happy New Year to all.
    @Robin Hood – your posts are interesting and, might I say, confident… but I am not so sure what qualifies your analysis. As posted above, swingback, as the ‘back’ refers to, implies a swing from opposition back to incumbant. This has not yet happened. If anything, Con appeared to fall back a bit post Autumn Statement. We are certainly going to see, as Pressman has made clear, everything and the kitchen sink thrown at Ed.
    I spend a fair amount of time campaigning in a couple of marginals in London. OK, it is London. But for the first time, I have had people saying ‘voted Con last time, not again’. I am just not sure where this swingback is going to come from. But maybe it will. We wait and see………….
    (electoral calculus has Labour 5 seats short of a majority, even though most seats in Scotland are lost – as per Mike Smithson)

  2. http://www.thetimes.co.uk/tto/news/

    “Teenagers are more likely to be victims of crime than any other age group and see assault, robbery and even rape as part of growing up, a report has found.”

    I wonder if the tidal wave of violent crime that is being inflicted on the 11-19 age group is connected to the ratio of young males to young females in a particular manor (as more young males competing over fewer young females was the reason for the extreme levels of violence where I grew up) and I wonder if this tidal wave of unreported violent crime has been having an effect on polling in some way?


    “On the other hand, as the economy has improved it has fallen down the list of issues people consider important and become a less salient issue”

    1) All my rellies on zero hour contracts hate it because they are working full time jobs on the same basis they would have done agency work in the past. Now the agency work level of job security is fine when it’s mutual consent but not so great otherwise. Young people growing up now will come to see that as normal as gang rape but older people who remember when full-time jobs came with some financial security aren’t going to see it as an improvement.

    2) Other issues over taking economy in the second spot could be because people think the economy is improving or because those issues are perceived to be getting worse faster than the economy e.g. the NHS was always going to start peaking sooner or later when what has been happening in the poorest areas became more widespread.

  3. LEFTY

    Your alternative is therefore a Greece exit from the EZ in order to regain monetary sovereignty.

    Ironic when you consider that Greece managed to disguise its poor public finances in order to get into the EZ.

  4. Your alternative is therefore a Greece exit from the EZ in order to regain monetary sovereignty.

    Ironic when you consider that Greece managed to disguise its poor public finances in order to get into the EZ.

    ….with the assistance of Goldman Sachs…..

  5. AW
    Marvelously clear and well thought out analysis. Thanks, and HNY for 2015.

    “In 2010 Con and Lab combined got 65% and there’s nothing to suggest they will improve upon that this time, only this time the Lib Dems are on a quarter of what they had in 2010, with the 18% (churn adjusted) going to UKIP, SNP and the Greens.”
    Suggestive impressions that the 65% will increase are perhaps that the swing to minor alternative opposition parties are mid-term statements against both main parties, and may swing back; secondly, in respect of Labour that, small though it may be, a swing from 31 ish to 34 ish in the past four months may emerge as an upwards trend, while Con have been stuck on 31 ish. This gap may widen as the bipolar choice of who you want to form the government and who runs the best campaign take shape.

  6. Talking of immigration being seen as the no.1 issue now and the help that gives UKIP, here is a snippet from a Daily Telegraph article this morning:

    ‘Douglas Carswell, who earlier this year became Ukip’s first MP in the Commons after defecting from the Conservatives, has warned that his party must show an “inclusive” face and not blame immigrants for Britain’s woes if it is to become a serious force in politics.

    He said that a said that a dislike of foreigners was “not merely offensive but absurd”.

    “No Ukip candidate should ever make the mistake of blaming outsiders for the failings of political insiders in Westminster,” he said.

    “There has never been anything splendid about isolation.

    “It was our interdependence that put the Great into Great Britain – and it is what sustains our living standards today. In such a world, a dislike of foreigners is not merely offensive, but absurd.” ‘


  7. “Suggestive impressions that the 65% will increase are perhaps that the swing to minor alternative opposition parties are mid-term statements against both main parties, and may swing back”

    “This gap may widen as the bipolar choice of who you want to form the government and who runs the best campaign take shape.”
    Or it may reverse. Who can tell?

  8. Happy New Year All!

    Little political news, so have not been posting. Pleased to see that Hollande is starting to see the folly of over taxing the rich and has dropped his 75% tax.

  9. “Palestinians sign up to join ICC ”

    Didn’t even know they played cricket.

  10. guy monde

    “HNY to all”

    A bit lazy.

    Find it bizarre that some don’t recognise that “swingback” is when votes which initially swung from Governing party[ies] to the opposition SWING BACK.

    Snot happening.

  11. Agree with R&D – its been obvious ever since the lib dem collapse that all the previous ‘rules’ around voting behaviour (i.e “swingback”) are now redundant.

    The emergence of UKIP and the recent rise of SNP and Green means we are in ever more uncharted waters.

    However i think we can can assume that some of the what the tories and lab have lost to the smaller parties will return – but how much and from where and too whom is the 326 Seat question.

    Interesting point about how the tories have performed – marrooned in the low thirties for years – whilst labour have been more volatile. This may suggest that labour have more chance of recovering their vote to a more significant degree. Also interesting how the lib dem collapse looks to be less significant in tory seats than labour ones. Then you have the effect of the SNP – that looks pretty nailed on to me as LOC scottish voters may belive that a SNP block in westminster will mean more powers for scotland – and a more social democratic government

    Given all this I think something like Tory 33 Labour 35% Lib Dems 10 UKIP 12 Green 5 might be the out come but with the SNP stopping a labour majority.

    Intersting times! Will PR be back on the agenda?

  12. CMJ

    Indeed so -they were paid a handsome fee to fool the suits in Brussels & get them into the EZ.

    Perhaps they will be re-engaged by a far left administration to get them out of it :-)

  13. From Bruges,
    A very happy new year to everyone ,certainly it will be interesting in many ways.

  14. “No Ukip candidate should ever make the mistake of blaming outsiders for the failings of political insiders in Westminster,”

    Quite. Apart from any other reasons blaming immigrants instead of the political class would roughly halve the maximum vote ceiling imo.

  15. Pups

    “HNY to all

    A bit lazy.”

    Make allowances. I only have two typing digits and a max of 10 if I was cleverer. You have 40 even without your dad.

  16. As others have said, 2015 is a fascinating year in prospect, on so many fronts.

    While I’m loathe to make predictions, I will predict that the vast majority of predictions for 2015 will turn out to be incorrect.

  17. I have been having some (slight) second thoughts about my earlier criticisms of Robin Hood’s version of the Swingback thesis. This occurred because I have finally been able to track down his original (Nov 11 2012) UKPR post on the topic – which for reference can be found rather unexpectedly at:


    My earlier criticism had been based on the fact that Swingback didn’t seem to be going to plan between a point 10 months out from the election (when he made his July 2014 projections) and 5 months out (when he revisited his projections last month). Specifically, historical figures suggested that an anticipated swingback of about 5% over the last 10 months should have been reduced to about half that figure earlier this month. As we haven’t seen a 5% reduction in Labour’s lead since July, I argued that swingback was not following the designated script and should therefore be treated with scepticism as a basis for trying to call the forthcoming election.

    What I had missed was that before becoming rather anaemic over the last few months, Swingback substantially *overshot* Robin’s predictions between November 2012 and July, and is currently more or less on target if you set aside the July recalculation.


    On Nov 11 2012 at 6.45 pm @ Robin Hood wrote:

    Mid-term Con poll lead: -8.8%
    Con lead at subsequent general election: 3.3%
    Con ‘swing back’: 6.1%”

    On December 7th at 5.54pm (and on the present thread) he wrote:

    “This yields an average pro-Conservative “swing back” of +2.55625% in the final five months.”

    So, between Nov 2012 and December 2014 3.5% of the originally expected Swingback should have been ‘in the bag’ (3.5% being approximately 6.1% minus 2.55625%).

    According to the polling averages Labour’s lead was about 10% in Nov 2012 and referring to Anthony’s first table above it has been reduced to 1.7% this month. This represents a Swingback of 4.15% over this period, which is pretty much in line with what he was expecting over this two-year phase of the election cycle.

    So, what has happened is that Swingback has more or less gone to plan over the last two years by first dramatically overshooting expectations and then proceeding at a much lower pace than in preceding elections.

    For those trying to anticipate what will happen over the last four months, the question is whether we should focus our attention on the full period (for which RH’s version of Swingback is still on target) or concentrate more on VI shifts *within* this period (for which Swingback has been erratic and now appears to be at a standstill).

    Personally, I am still inclined to place greater emphasis on the more recent patterns as – to my mind – these better reflect the current political climate (economic confidence fading; immigration shifting up the rankings etc.). But I did want to acknowledge that I wasn’t aware of some of the longer-term patterns in several of my earlier posts.

  18. @Unicorn

    The 1.7% figure you quote is the’average’ Labour lead for December as a whole.The final figure is being shown as 3%.

  19. @ AW/Others

    Just had my virus checker kick in twice today and once yesterday while on UK Polling report.

    In both cases today the block was on http:/ and then a load of other letters and numbers that go off the page.

    Obviously can’t be sure it is not something on my machine but given it hasn’t happened on any other sites it’s pointing to a UKPR ad.

  20. @ RosieandDaisy

    “Find it bizarre that some don’t recognise that “swingback” is when votes which initially swung from Governing party[ies] to the opposition SWING BACK.”

    It would certainly help if we could all agree on a precise definition of the term. You are obviously right that over the past year the coalition parties have not been the overall beneficiaries so in that sense it is correct to say that it is not happening.

    One reason for this may be that all other parties have lost support to Ukip.

    However, because election seat outcomes are decided on the basis of the *margin* between parties and not on the absolute count itself, most of the formal analyses have used the term to refer to changes in the Labour/conservative VI difference and not to the individual performance of any one party.

    I can attest that this is certainly true of Robin Hood’s numerical projections – as I have just been going through his past posts with a fine tooth-comb. In one point he talked about the marginal *seats* swinging back in the Tory direction.

    Since this is the kind of ‘swingback’ that will make the difference in the election I think it makes sense to use the term as Robin Hood does. (The Tories could benefit from this kind of Swingback even if they remain becalmed at 31%).

    Alternatively, we could talk about ‘Swingback of absolute support to governing party/parties’ (of which there is no recent evidence) or ‘Swingback of *relative* support’ (patchy evidence for Tories – see previous post: also the LibDems have recently gained against Labour – with whom they compete for certain seats).

  21. @ SHEVII

    I don’t know if it is related to your virus problem but I have been getting some links to pretty unsavoury sites when browsing UKPR pages. This seems to happen when I am using an iPad and spool fast down the comments to get to the bottom. At these points my finger-placements are not very accurate and I seem to activate ads that are much less anodyne than those headlined in the designated advert slots. Again, this does appear to be a UKPR-specific problem and it would be good if it could be put right.

  22. @ Graham

    “The 1.7% figure you quote is the’average’ Labour lead for December as a whole.The final figure is being shown as 3%.”

    True – but this doesn’t materially affect the point I was making. If you assume that the end-of-year lead has settled down to the most recent average of 3%, then between Robin Hood’s first (Nov 2012) prediction and now Labour’s lead has dropped from 10% to 3%, representing a 3.5% Swingback over this period. As it happens, this is *precisely* the Swingback figure he predicted for that 2-year period.

    So – irrespective of the exact current Labour lead averages – my original point stands. Overall Swingback over the last two years has been almost exactly in line with (Robin Hood’s) expectations. However, over this period Swingback has itself been swinging around a lot, making it rather difficult to anticipate what will transpire next.

  23. @Unicorn: “Swingback substantially *overshot* Robin’s predictions between November 2012 and July, and is currently more or less on target if you set aside the July recalculation.”

    There’s been no ‘swingback’. Labour’s lead may have fallen but those votes haven’t swung back to the governing party (or parties) – they’ve gone elsewhere, e.g. to the Greens, SNP and UKIP. You may argue that the effect will be the same (although I suspect you’ll be wrong – e.g. a Labour to UKIP movement is clearly less beneficial to the government than a Labour to Tory one) but you need to call it something else.

  24. Happy new year to all – Hope Ann in Wales is enjoying her time in Bruges and congratulations to Unicorn for being one of the rare posters on this site who goes on thinking about issues on which s/he has pronounced and is prepared to modify their position.

    On ‘swingback’ whatever its definition, I can see why Greens might move towards Labour come the election. They are unlikely to win many seats themselves and a Labour government is probably the best chance of getting them what they want.

    I don’t however see why the SNP should do the same. If they vote SNP they may well get SNP. They won’t let in Cameron since Nicola Sturgeon has said she won’t support conservatives and they are more likely to get a Labour government which should pursue the kind of broad policies they want and be forced to pay particular attention to the views of the SNP and Scottish interests.

    Can Amber or some other Scottish luminary tell me what arguments might dissuade SNP voters from behaving as above?

  25. OK … Three posts in a row from me. Time, methinks, to imbibe some mead instead…

  26. I side with the argument, that what we are seeing is not a true “swingback”. After all, the votes have NOT swung BACK to the tories and remained with the opposition parties.

  27. What we are seeing is not ‘swingback’ but could be called ‘swingaway’ in that all we can say for sure is that voters seem to have moved from Conservatives (since 2010) and from Labour (since 2012). With a political system locked into a two-party dominance what we can reasonably expect in the next few months is ‘swing back’ to both main parties: in what proportion will determine the outcome I think.

  28. @Charles

    Think the main argument in Scotland will be over whether voting in SNP MPs increases the chances of Conservatives being able to hang on as largest party.

    This is a possible (even likely?) consequence of a large block of SNP MPs being elected.

    It then comes down to some fairly technical discussions about Constitutional precedent. As I understand it in a hung parliament the incumbent Prime Minister has the first chance of forming a government whether or not they have the largest number of seats.

    However if it is clear that they cannot form a majority then the Queen should invite the leader most likely to be able do do so to form the government. In this case we know SNP are more likely to support Labour so voting SNP shouldn’t prevent a Labour government forming in most circumstances.

    A potential complicating factor is the Lib Dems. In 2010 Nick Clegg said he would only deal with the largest party. If he maintained that stance and you had a result like 285 Con, 275 Lab, 40 SNP, 25 Lib Dem,10 UKIP, 23 Others then it would be interesting constitutionally.

    Lab + SNP + PC + Green + SDLP + Sylvia Hermon would come to around 323, not quite enough to form a government if Clegg was only dealing with the largest party.

    Con + LD + DUP would come to around 319 again not quite enough to form a government assuming it’s impossible for Conservatives to do a deal that keeps both Clegg and Farage happy.

    So voting SNP could in theory reduce the chances of a Labour Government but only in a situation where Labour would have had to do a deal with Clegg anyway.

    In such a situation I’d expect the Lib Dems (possibly under a new leader) to do a deal with Labour, maybe along C&S grounds.

  29. Alec is right about predictions being well dodgy now. Too close, too many variables and too many marginals based on different sets of criteria.

    Will the UKIP vote drop and if it does by how much and will it split evenly or in favour of one of the two parties?
    Don’t know.

    Will the LD vote in the seats they are defending pick up when they present their electorate with a two horse race and keeping out the Tories/Labour (delete as applicable) and will this depend on whether they are defending against Tory or Labour?
    They are more likely to defend seats they are holding against the Tories than against Labour but the Tories have more low hanging fruit so don’t know if the Tories will get those. Also don’t know if Labour can grab some of the fruit at the very top of the tree. Certainly don’t know how this balances out between gains for Tories and gains for Labour overall.

    Will Green voters go back to/choose Labour in the marginals when presented with the same choice as per LD vote above?
    Don’t know.

    Will SNP voters go back to Labour in seats where SNP are 3rd and under normal conditions might let in the Tory/LD even if the polls say the SNP could snatch them on current UNS?
    Don’t know.

    Will UKIP get 2 naturally Tory seats or 10 and might they get one or two from Labour?
    Don’t Know.

  30. @All (RE: Swingback)

    I was going to take a peek of some polls at regular periods over this government’s past term, and compare 2010 Con voters that moved to Lab and vice versa and look for swing (in either direction).

    Sadly, the first poll I looked at (6th Nov, 2010) did not have that breakdown, so does anyone have 2010 to present data on that matter?

    For all we know a 2% swing from Con to Lab has been and come back, but the current polls are favouring neither party.

  31. No sign of swingback since the end of August when Populus put the Tories ahead.

  32. @ Charles

    Can Amber or some other Scottish luminary tell me what arguments might dissuade SNP voters from behaving as above?
    Not an easy one to answer without straying into partisan territory or kick-starting another Scotland discussion – which has, unsurprisingly, rather worn out its welcome on a UK politics site.

    Happy New Year. :-)

  33. @Northumbrian Scot – Thanks. That is food for thought.

    In the light of what you say i think that Labour should go as far towards doing an implicit deal with the Greens, Plaid and SNP before the election as it possibly can e.g. it should be strong on rebalancing vis-a-vis the regions and kick Trident into the longest possible grass. I would also try and steal a little of UKIP’s cloths, scrapping HS2 for the foreseeable future and agreeing about the pain that certain sectors of society are suffering but disagreeing about the diagnosis,

    In the case of the SNP voters the hope would be that they could see that they could get the key things they wanted out of Labour and might risk them if they let the conservatives become the largest party.

  34. @ Unicorn

    To quote Graham:
    No sign of swing-back since the end of August when Populus put the Tories ahead.

    You’ve gone back and performed a very fair analysis on Robin Hood’s swing-back model/ theory. But it doesn’t seem to address the swing-away which we’ve seen lately, or have you dealt with that and I missed it?

  35. Northumbrian Scot

    My recollection is that Clegg said he would talk first to the party with the biggest mandate not exclusively which is why he could send his team for sham negotiations with Labour presumably to extract more from the Tories.

    I agree with your main point though a genuine question is would Con most seats but Lab+SNP be a bigger block end with a minority Con Gov’t (even minority LD/Con).

    Lis – will understandably talk up this possibility but need to be carfeful not to overdo, as scaremongering can be counter-productive.

    The SNP, though, will argue fairly that more SNPs MPs means more pressure on UK Lab to do deal with them that suits their views re Scotland more.

    As an aside the politics would be fascinating with the SNP saying you have a minority Tory (led) Government as Labour would not be reasonable with them; whilst Labour would say the SNP made unreasonable demands.

    In the credibiity stakes most seats Lab or Con in E&W may be important?

  36. Charles
    I don’t think there’s a lot of mileage in disagreeing about the diagnosis, but plenty in prescribing a different course of treatment . Such treatment will in any case help with various maladies which have befallen the country and ‘natural’ Lab supporters in particular (whether they are actually Lab supporters or not)
    When you say kick Trident into the grass, I presume you mean refuse to engage with any thoughts of scrapping it?

  37. Agreed Jim Jam and Charles some interesting possibilities.

    I think in the post referendum world we live in there will be more media focus on who wins most seats and votes in both England & Wales and England.

    Could see some newspapers (Express and Mail especially) talking about Labour “lacking a mandate to govern England” if they don’t have most English seats or votes.

    Constitutionally this is nonsense but it will need addressing by Ed as EVEL proposals will be knocking around.

    Some combination of Grand Committee / Lords Reform / regional devolution will probably form the Labour response.

  38. @NorthumbrianScot

    I believe the constitutional position is that Cameron remains the PM after the election unless he decides to resign, which he would do either on losing a vote of confidence, or (more likely) if he thinks he is going to lose a vote of confidence.

    So Cameron certainly has an advantage over Miliband: there may be minor parties that are prepared to do a deal with either of them, but Cameron gets first go purely because he is PM already. So there could be a situation in which both Cameron and Miliband could secure a majority in a vote of confidence by doing whatever deals are necessary, but Miliband never gets to try because Cameron does not have to resign. The situation would only be clear-cut if a majority (of those not abstaining) declare they are definitely for one and definitely against the other.

  39. This post is about using figures from 2014 (like AW’s) in an attempt to forecast the outcome of the election. I take it that everyone who visits this site has an interest in what the polls tell us about the future (and in particular about the likely outcomes of elections and referenda) and not just in the polling tallies themselves – whether past or present. So, here goes…

    Summary (for those preferring to skip quickly onwards)

    Assuming that there is a general interest in predicting real voting behaviour I am going to outline some ways of playing around with the workings of the Electionforecast model – adapting it to explore scenarios that might be of interest to UKPR contributors. I hope the model-builders see this as an acknowledgement that their (rather labour-intensive) work is valued and used by others. The upshot is that the future may look a little more ROSY (get it?) for Labour than alternative calculations might suggest.

    As I understand it (and please feel entirely free to correct me if I am mistaken) the Electionforecast model consists of (at least) two basic components: (î) an integrator or aggregator of polling and actual voting data and (ii) a further feature to make projections about a future event (e.g., a general election). The integrator uses weighted national and constituency polls and regional crossbreaks to rebalance the settings on a daily basis. Within the public domain at least, I am not aware of any other facility that draws upon such a wide variety of polling evidence. This is the kind of analysis that Nate Silver does with such brilliant accuracy for American elections.

    The forecasting component uses historical data to introduce regression-to-the-mean adjustments to their “Nowcast” VI tallies and in this way to forecast what the vote-share and seat tallies are projected to be at the time of the forthcoming election.

    In my own personal view it would be difficult to improve upon what the first component does. As constituency polling evidence accumulates the Nowcast should offer an increasingly accurate summary of the state of play around the 650 parliamentary seats. (Granted, everything could be thrown into disarray by ‘events’ and developments like the unpredicted pre-2010 Cleggmania. But factors of this kind would sink any other model just as emphatically.)

    If the polling-data-aggregator seems rather good, I think many UKPR posters would take issue with the workings of the forecasting component. Basically this assumes that parties whose current VI-share is higher than their 2010 vote-share will lose a proportion of this support as the election approaches. Similar, parties underperforming compared with 2010 are predicted to see a steady improvement. So, in today’s instantiation of the model, the Conservatives are predicted show a 3.4% gain in VI between now and the election and for the LibDems the corresponding projected gain is 4.2%. Drops are predicted for all other parties: 0.6% for Labour, 6.0% for Ukip, 2.1% for the Greens, 0.6% for the SNP and there are also adjustments for each of the smaller parties. For the first three parties and the Greens the changes are introduced almost uniformly across the constituencies. For example, for the Tories in 99% of the constituencies the change is either +3 or +4. For SNP and PC the changes are uniform where these parties field candidates (implying a projected SNP drop of 6% in Scotland). For Ukip the projected losses range much more widely across constituencies (from 0% to 7%). Overall, the authors, rule-of-thumb *seems* to have been to halve the difference between each party’s current VI and its 2010 vote share.

    Taken together this gives their current forecast of a hung parliament with Labour (just) ending up with the largest number of seats. Specifically, their predictions today are:

    Conservatives: 288 Seats
    Labour: 277 Seats
    LibDems: 25 Seats
    UKIP: 3 Seats
    Greens: 1 Seat
    SNP: 35 Seats

    What happens if we don’t ‘buy’ any or all of the authors’ forecasting adjustment figures? Well, it happens that a simple spreadsheet can be used to test the effect of different sets of swing assumptions. So, for example if a new forecasting theory holds that two parties will each improve by 2% and four others will *drop* by 1% each, then the Nowcast output can be adjusted to implement these changes (always assuming uniform swingback across all constituencies). Note that this is just a rough-and-ready tool for asking “What if?” questions, and there is no way of reproducing the probability distributions that are an important feature of the original model itself.

    A few weeks ago – largely for purposes of discussion – I posted a comment suggesting that (linear) trends established over the 2014 calendar year might just remain unchanged for the rest of the period leading up to the election. Using my original regression equations (posted in comments at 6.14 pm on November 24th and at 11.28 am on November 25th) this model would predict the following VI changes between January 1st 2015 and the election:

    Conservatives: a small VI reduction of 0.34%
    Labour: a larger VI reduction of 1.88%
    LibDems: a further VI reduction of 1.07%
    UKIP: a further VI *rise* of 1.54%
    Greens: a VI rise of 1.33%

    I didn’t do calculations for SNP at the time because – until recently – separate VIs for them were difficult to extract. However, using the second trend in YouGov’s recent post “5 public opinion trends from 2014”, a rough estimate of the further SNP VI increase by May would be 13%.

    Lastly, specific VIs for smaller parties were not reported, and so there is no obvious way of calculating trends. So, for convenience I assume no change is expected for any other party.

    Feeding these adjustments into the modified version of the Electionforecast model gives the following prediction for the general election:

    Conservatives: 269 Seats
    Labour: 275 Seats
    LibDems: 18 Seats
    UKIP: 8 Seats
    Greens: 2 Seats (picking up Bristol West)
    SNP: 56 Seats

    These projections might seem fanciful. But the point of the exercise is to show the extent to which the projections depend on the precise swingback or regression-to-mean assumptions used in the calculations. If instead trends continue as in the past year, the Labour will emerge with most seats and this – plus a large left-wing SNP parliamentary presence – would make a substantial difference to the political climate in discussion for forming a government.

    If anyone has a personal set of expected swings they’d like to see tested in this way, then I’d be happy to plug the figures into the spreadsheet and report back with the outcome. Note that – in order to make sense – the anticipated gains should roughly match the expected falls. In this balancing exercise SNP changes can be compensated for by a 57/650 = approx 9% adjustment to VIs for UK-wide parties (on the assumption that their influence will be restricted to the Scottish constituencies). For example, a 10% SNP rise would be balanced by a 0.9% fall over parties fielding candidates in seats across the entire country.

  40. Good stuff there, Unicorn, but I’m not sure I understand this bit. Have you got the figures wrong somewhere or is it me being dense?

    “Taken together this gives their current forecast of a hung parliament with Labour (just) ending up with the largest number of seats. Specifically, their predictions today are:
    Conservatives: 288 Seats
    Labour: 277 Seats”

  41. Interesting post on PB,


    if the Tories can’t win back UKIP voters in large numbers or get LibDems in Tory/Labour marginals to vote Blue then they are in deep trouble.

    I’d go for deep trouble, because you tend to lose if you try to fight a “War on Two Fronts” and pretty much everything the Tories might do to win back Kippers on the EU or Immigration would be anathema to most

    Labour just need not to be the Tories to hold on to the LibDem’s who swing to them after 2010 and still who have stuck with them so far.


  42. @ Norbold

    Sorry. Today’s Electionforecast figure are correct and my summary put the wrong party in the lead. It is, of course, the Tories who are 11 seats ahead in their current forecast.

  43. Happy New Year everyone!

    Regarding coalitions – here’s something that has been puzzling me. If you are a small party and have watched how the LibDems have been annihilated because of their coalition, why would you be so eager to line up to have the same thing happen to you?

    Has there ever been a coalition that hasn’t hurt the smaller party? The Liberals didn’t come out well from their coalition with the Conservatives in the 1930’s either.

    In Ireland, the Labour party looks like it will suffer from their coalition with Fine Gael. In Germany the Free Democrats have died, and the SPD also suffered badly from their grand coalition with Merkel’s party.

    Are the small parties so keen because the desire to get their hands on the red boxes is stronger than their overall survival instincts?

  44. Unicorn
    “It is, of course, the Tories who are 11 seats ahead in their current forecast.”

    Umm, electionforecast today has Lab 11 seats ahead of the Cons with 288 to 277.


  45. @ Amber

    I hope your ‘swing away’ question has been partly dealt with by my long post at 3.54 pm.

    My ‘alternative’ electionforecast projection was based on the (probsbly implausible) assumption that the 2014 linear trends will remain unchanged. That leaves Labour 6 seats ahead of the Tories on May 7 and (from memory) leaves the SNP with all but three of the Scottish seats.

    If you think that Labour’s fortunes are now on the up you could choose a different figure for the expected Labour VI change between now and May and I could tell you what effect this alteration would have when grafted onto the first part of the Electionforecast model.

  46. @Amber – Happy new year! I enjoy your posts whether or not they happen to be on Scottish matters, So, hopefully, you will have plenty to comment on in future!

    @Guymonde – I think that some parts of the country have poor services, a chronic shortage of housing, high unemployment and/or rotten working conditions and low wages., particularly for unskilled men. I think that immigrants are wrongly blamed for causing these things and that given our demographics we can’t afford to exclude them anyway.

    So I would concentrate on dealing with the problems for which immigrants are blamed rather than on posturing about toughness on immigration, a position on which Labour are always going to be outflanked anyway. I doubt this will persuade back the die-hard members of UKIP, but hopefully it might make a few feel that their concerns had been heard.

    As for Trident, I don’t see the use for it. It costs a lot of money and I don’t understand the circumstances in which we could deploy it , As it seems to be some kind of national virility symbol I would not overtly scrap it,without getting something in return in international negotiations. However, I see no point in pouring good money after bad when there are a lot of better uses to which it could be put,.

  47. @ Bramley and @ Norbold

    Apologies! The table summarising today’ s Electionforecast projection should have read:

    Conservatives: 277 Seats
    Labour: 288 Seats
    LibDems: 25 Seats
    UKIP: 3 Seats
    Greens: 1 Seat
    SNP: 35 Seats

    I hope I have now got that right.

    This now means that the Electionforecast (’Regression-to-mean’) and my own (’continuing 2014 trends’) projections are not that different for Labour and the Conservatives. Labour are ahead by a few seats in both cases. The main difference lies in what is projected to happen to the smaller parties.

  48. I dont think mathematical forecasts/formulas based on previous voting behviour can tell us anything at all this time around.

    How can you possibly scientifically calculate the LOC 2010 lib dems voters disgust at Clegg? Or the UKIP voters anger/bigotry with regards to immigration? And how this might influence their vote in individual seats all with their own individual variables such as the the parties previous performance – and voters perception of who might win it and incumbency factors?

    We have the polls and our own interpretations what is motivating (or demotivating) voters. Algorithms wont count for toffee come may.

  49. @Unicorn

    With 18 seat for NI, that prediction is 3 seats short of 650 (I always count the speaker and others in the 650).

    My prediction of May 2015 is less complimentary to the big two:

    Lab 275
    Con 275
    SNP 50
    Lib Dem 25
    NI 18
    UKIP 3
    Green 3
    Respect 1

    Just another wrong prediction to add to the list of wrong predictions. Interestingly, the SNP can’t ‘prop up’ either Con or Lab. Rainbow, anyone?

  50. I think one thing we can be sure of, SG, is that Respect won’t go into coalition with the Conservatives!

1 2 3 4 6