During Nicola Sturgeon’s speech to conference on Saturday (which noteably, the BBC did not have any mysterious problems with transmitting, unlike last year’s spring conference), she tried to play down the polls, noting that the SNP have only 6 MPs in Westminster, and “anything over 11 is record-breaking”. Yet it is hard to ignore the Big Numbers. So, I thought it might be worthwhile doing a quick retrospective of all the polls since I took time out just before Xmas, to try and see in one overview how the numbers have changed, in the space of the last 3 months, in particular – but also looking back to its origins.
So this will be long, but bear with me…
Apparently all this began sometime in the second half of October – but no one is sure precisely when. It was a scenario that no one envisaged after a ‘No’ vote (least of all Labour)…therefore no one was really asking the question too much in polls in the immediate wake of the Referendum.
John Curtice, a psephologist who was no friend to either ‘Yes’ or the SNP, was first to divine the signs from the entrails in an ICM/Guardian online poll published on the 27th December 2014. He broke ICM’s data down into four different categories of seat, showing that Labour’s decline was the most massive where its victory margins had previously been the largest over the SNP (by more than 25 points) – so although Labour’s vote across Scotland dropped by 16 points, in what Curtice referred to as traditional Labour ‘citadels’ in its ‘heartlands’, it fell by 22 points, against an SNP rise of 26. The greater the faith placed, the greater the sense of betrayal, perhaps.
That combination would be sufficient to wipe out majorities that were always assumed to be impregnable, and Scottish Labour’s Westminster caucus would be left shrivelling to just three MPs. “We are prospectively looking at the collapse of citadels that have always been Labour since the 1920s,” said Curtice. “That will seem incredible to some in England, but to those of us who paid close attention to Alex Salmond’s 2011 landslide at Holyrood, it would merely be the next chapter in the political transformation of a nation.”
A PanelBase poll conducted over the 17th and 18th January reported the SNP to Labour percentage share starting to separate at 41% to 31% – mostly the expressed reservations were due to both the oil price and Jim Murphy’s continuing unpopularity as the new leader of Scottish Labour.
Four days later, both Survation and IPSOS/MORI showed the SNP on 52%, with Labour on 24%. Curtice noted that seven polls over the preceding couple of months had shown a 46%/26% split, representing a 21% swing from Labour to SNP since 2010, and his average projections gave SNP 46, Labour 27, Conservative 13, LibDem 5, Scottish Greens 3. Which projected SNP to rise from their current 6 seats to 46, and Labour to fall to 9 from their current 40 Scottish MPs.
80% of Yes voters were apparently planning to vote SNP, and pollsters started noting similar trends for Holyrood 2016 voting intentions: on the 22nd January, an IPSOS/MORI poll for STV had SNP with more than double Labour support for both constituency (53, or +8.4 on 2011 results, to 24, -7.7) and regional (48, +4, to 22, -4.3) vote next year.
But the real bombshells were still to come.
On the 3rd February, Lord Ashcroft, the Conservative peer who seems to spend enormous amounts of money on polling (a thousand people per constituency, rather than a thousand people across Scotland or the whole UK) released a set of results showing a breakdown of the percentage vote of the top parties as Labour 31%, Conservative 31%, LibDem 8%, SNP 4%. It is worth pausing to reflect on that for a moment: UK-wide, the SNP were registering a seismic (given that Scotland has less than 10% of the UK population) 4% of the vote. But the big shock was in his first set of results, looking at sixteen marginal Labour and LibDem constituencies with strong ‘Yes’ votes, to see how much of that might be translating into support for the SNP. His data from 16,000 Scots indicated that 15 would go to the SNP with swings of over 20%, and an overall swing of 25.4%.
The following day, a YouGov poll for The Times seemed to echo this, with the Scotland vote breakdown of SNP 48%, Labour 27%, LibDems 18%, Conservatives 4%, prospectively translating into 48 seats for the SNP, 11 for Labour, and none for anyone else. Interestingly – and often forgotten – the prospective LibDem losses to the SNP in Scotland make the continuation of the current Conservative/LibDem (often referred to as ‘ConDem’ for numerous reasons…) coalition far less likely – although Labour campaigners that argue for tactical voting for LibDems to keep the SNP out (step forward Robert MacNeill) do not seem to care too much about this. The same YouGov poll also indicated that 52% would now vote for independence, with 48% against.
The same day, Peter Kellner, the YouGov President, noted that in January he had predicted a partial comeback for Labour to win 31 of their current 41 seats in Scotland, but given that the polls had refused to shift with Murphy’s appointment, he had reduced his prediction from 31 to 24, and gave this warning to Scottish Labour: “If Murphy cannot trim the SNP’s lead from 20 points to six to eight points, Labour could end up with as few as 10 to 15 seats”.
In the wake of Ashcroft’s first results, IPSOS/MORI had been busy, and on the 17th February presented their polling regarding prospective electoral deals between Labour and the SNP: 56% of English Labour voters supported the SNP deal, with only 25% opposed. A Survation poll released three days later showed that 35% of Scots wanted a Labour-SNP coalition government, with only 19% wanting a solely Labour one. Amongst the ‘also-rans’, 7% wanted another five years of the ConDem coalition, and a quaint 8% – perhaps rooted in the 1980s – wanted a LabLib coalition. Aw, bless.
At the end of February, Survation repeated their poll from the 19th September – results day for the Referendum – and the comparison was truly fascinating. From the 19th, the figures were Labour 38%, SNP 34%, Conservatives from 15-18% (no, I don’t understand this indecision, either). By the end of February, using the same methods for the same questions, Survation were getting SNP 45%, Labour 27%, Conservatives 13-15% (again – search me…). Interestingly, using the Electoral Calculus website tool on the 19th September figures to represent a resurgent Labour come May 7th, prospectively gave the Conservatives 2-3 seats in Scotland, increasing from their solitary, smaller-than-the-number-of-pandas-in-Scotland 1. ‘Vote Labour, Get Conservative’, apparently – in this scenario, a late swing back to Labour could actually be enough to make a Conservative government MORE likely – although you won’t be seeing that on any Labour election literature…
The second bombshell came on the 5th March, with Ashcroft’s second poll results released, this time of a thousand constituents in eight ‘No’ voting ‘safe’ seats for SNP opponents. All 5 Labour seats showed large swings to the SNP, with only East Renfrewshire’s Scottish Labour leader Jim Murphy holding on by 1%. Of the 2 LibDem seats, Charles Kennedy, the former party leader, was projected to go, leaving only Alistair Carmichael as the sole extant LibDem MP in Scotland, up on Orkney and Shetland. The 1 Conservative seat, David Mundell in Dumfriesshire, Clydesdale and Tweedale was projected to be a dead heat (Conservatives and SNP both 34%, Labour 18%, LibDems 7%). Overall, this suggested over 50 seats to the SNP. UK-wide Ashcroft’s polling showed LibDems and SNP now tied on 5% of the entire UK vote – but LibDems are contesting 650 seats, and the SNP are only contesting 59…
It is worth noting that arguments have been made that Ashcroft’s polling system may be slightly underestimating SNP support, as it asks the voting intention twice, and only uses the second answer: psephologists posit that the second answer can be used to give an ‘apologetic’ response, through a feeling of having to ‘balance’ the first more honest answer with a different second one. But, even accepting Ashcroft’s data at face value, the New Statesman’s electoral calculator went further with the data, projecting 56/59 seats going to the SNP. Which would leave Alistair Carmichael for the LibDems up in Orkney and Shetland, and Labour with Glasgow NE (Willie Bain) and East Renfrewshire (Jim Murphy), a 25% swing to SNP in West Dunbartonshire taking it from Labour’s Gemma Doyle.
An ICM poll for the Guardian between 13-19/3/2015 showed no sign of this support slipping, with SNP 43%, Labour 27% (+1 since December), Conservatives 14%, UKIP 7%, LibDems 6%, Greens 3%.
And, finally, on the 31st March, a ComRes poll for ITV News of Labour-held constituencies across Scotland showed a 19 point swing from Labour to the SNP since the 2010 General Election – with SNP support in those 40 Labour constituencies now at 43%.
Which all rather begs the simple question – why?
Well, I suspect the answer lies in Ashcroft’s second set of polling data, from across the rest of the UK, predicting a dead heat for Labour and Conservatives on 272 seats, with Labour having fallen from their peak above 40 down to the low 30s, where the Conservatives have been stuck since mid 2012, and the LibDems on below half of their 2010 vote share, with UKIP falling back slightly from their peak at the end of 2014. Polls show 70% of Scottish voters believe that Labour has ‘seriously lost touch with ordinary working people’, and that two thirds believe UKIP is a party for ‘oddballs and extremists’. The LibDems opposed VAT rise and austerity until they were in government, when they supported both, so are little trusted. A leaked memo even shows that the Conservatives are only seriously contesting 2 seats – including the one that they currently hold – and for that David Mundell is avoiding mentioning the fact that he is a Conservative Party candidate on his leaflet, perhaps thinking it is not a vote-winning strategy…
It continues to oscillate tightly backwards and forwards, on the UK-scale: after the leader’s had their non-debate (the ‘battle of the TV interviews’) last Thursday, Labour had a 4% lead on the Sunday – then Ashcroft’s polling the following day gave the Conservatives a 2% lead. In terms of seat forecasts, ‘Elections Etc’ and ‘Election Forecast’ have generally had the Conservatives 6-8 seats ahead of Labour, with ‘May 2015’ and The Guardian putting the lead at slightly less, at 4-5. Polling gives Conservatives 276-286 seats, with Labour on 271-280 seats, and the LibDems on 22-26 – so if the SNP can muster 40 seats in Scotland, it makes it electorally impossible for the Conservatives to gain power, even with the LibDems help. With a vote for the SNP, it avoids the dangers of Labour swinging even further to the right once in government, as happened so tragically under Tony Blair.
So Sturgeon made it very clear that she would not be supporting the Conservatives under any circumstance whatsoever in November. As the polls had yet to leap up to their current levels, nobody paid a blind bit of notice – until last weekend when Alex Salmond reiterated the same point, and that the SNP would vote down a Conservative Government, and uproar hit the right wing tabloid press down south. But it looks as though neither Labour nor the Conservatives are going to get much more than 33% of the vote (is it worth pointing out that ‘Yes’ managed significantly better than that in Scotland?…). With such a widely-promoted tie below majority government level, perhaps, just for once, Scots know that their votes would not just be making up the numbers, but actually for ONCE have a chance of making a difference to the political landscape of Scotland – and even the rest of the UK.
There is also less of a sense that one just has to accept whatever London hands down anymore, and that there are no alternatives and no way out: a recent study by the University of Edinburgh indicated 69% of Scottish voters believe that Scotland WILL become independent (although, coyly, no timeframe was offered or asked for…). And with this self-belief, this sense of empowerment, perhaps there also comes a desire to flex those muscles…and just once try voting for something different from the consensus, at the one time when it really might make a difference.
So, Glasgow NE and East Renfrewshire as Labour’s remaining seats…that means that there would be as many Labour MPs as pandas, in Scotland? Now, there would be a thing…
“the most important thing for the SNP in every Westminster election is to achieve the thing we failed to achieve since 1974 and that’s to achieve relevance in a Westminster election. And listen, folks: nobody can say we are anything other than relevant to this election campaign.” (Alex Salmond, 29/3/2015)