Thursday, July 20, 2017

Ipsos-Mori subsample adds to the weight of evidence suggesting the SNP are still ahead

Ipsos-Mori's subsamples (just like Survation's) tend to be very small and produce extremely volatile results.  However, for what it's worth, the subsample from the firm's first post-election poll is very positive for the SNP...

SNP 39%, Conservatives 26%, Labour 23%, Greens 7%, Liberal Democrats 6%

There have now been twelve subsamples from various firms since the election, of which seven have put the SNP in the lead, four have put Labour in the lead, and only one has put the Tories in front.  As Labour appear on balance (albeit not in today's numbers) to be the main challengers, it's particularly significant that the SNP have been ahead of Labour in eight of the twelve subsamples.

Wednesday, July 19, 2017

The crumbling of Colonel Calamity : evidence begins to mount that the Scottish Tories have slumped to third place

Just as a slight corrective to the recent string of good polling results for the SNP, there's a new ICM poll which has the SNP a little behind Labour in the Scottish subsample: Labour 35%, SNP 33%, Conservatives 25%, Liberal Democrats 4%, Greens 3%.  However, the potential error in an individual subsample is enormous, so those figures are very easily consistent with the SNP holding a small lead.  That still appears to be where the balance of probabilities lie - there have now been eleven subsamples from various firms since the general election, of which six have put the SNP in the lead.  Perhaps more to the point, the SNP have also been ahead of Labour in seven of the eleven subsamples.

That said, the battle between SNP and Labour seems to be close enough that it's not possible to say with absolute confidence which of the two parties is in front.  The Tories have slipped out of that conversation, and it now looks increasingly likely that they've dropped to third place.  If it gradually becomes an accepted fact that Scottish politics has reverted to a traditional SNP-Labour duel, what on earth will happen to the love affair between Colonel Ruth and the media, both north and south of the border?  I believe the line is "she was the future, once".

Monday, July 17, 2017

But we can still RISE now, and vote for a unionist party again...

So is there anything to be said for Cat Boyd of RISE "proudly" voting Labour at the general election?  Well, there's certainly something positive to be taken from the fact that she's admitted doing it.  There's been a tendency among the unionist commentariat to treat the 27% of people who voted Labour as if they were part of some pan-unionist bloc vote comprising more than 60% of the population.  The reality, as we already know from the opinion polls, is that Labour's support was a coalition incorporating people who voted Labour because of its stance on independence, and also people who voted Labour in spite of that stance.  It'll be very useful to have a high-profile example like Cat Boyd to illustrate that point.  This episode may also be helpful to the SNP on the list vote at the next Holyrood election, because RISE (or whatever succeeds RISE) will find it even harder to pitch for 'pro-independence tactical votes' now that their commitment to independence has been shown to be rather superficial.

However, there's an idea doing the rounds that we must show veneration and respect towards Ms Boyd for voting Labour as part of an alternative strategy for achieving independence.  That is, it has to be said, a bit silly.  Voting Labour in the hope of furthering the cause of independence is no more and no less irrational than voting UKIP in the hope of keeping Britain inside the European Union.  It's been suggested to me that I'm missing some incredibly sophisticated point here, ie. that pro-indy people voting Labour are starting a conversation with the party that will eventually lead to a change in its constitutional stance.  But voting is essentially a passive act - you're not entering into a dialogue with the party you vote for, you're simply endorsing them.  It doesn't matter if Labour are privately conscious of the fact that much of their support is pro-indy - the lesson they'll draw is that those people have already proved stupid enough to vote for them, and so they can just persevere with the same policy and expect the same results in future.  If you reward undesirable behaviour, don't complain if you get more of the same.  For the proof of that, simply consider the fact that a substantial minority of Labour's voters in the decades leading up to the 2014 referendum were solidly pro-indy.  That had no impact at all.

Entryism can sometimes be a viable tactic for changing a party's stance, but that involves actually becoming members and activists (and then trying very hard not to get expelled).  Merely voting for a party you disagree with and have no influence within is entirely counter-productive - and that really ought to be a statement of the bleedin' obvious.

Are there any circumstances at all in which voting for an anti-independence party can help independence?  I can perhaps think of just one.  In the closely-fought 1992 general election, Labour were firmly committed to the establishment of a devolved Scottish Parliament.  It was not unreasonable to take the view that devolution was a necessary first step if independence was ever going to happen (as Margaret Ewing put it, there was never going to be a "Big Bang"), so the priority had to be to ensure that devolution happened.  There were a very small number of Labour-Tory marginal seats in Scotland, such as Stirling, that were going to help decide whether there would be a pro-devolution Labour government or an anti-devolution Tory government.  There was therefore a case to be made that tactically voting Labour in a seat like Stirling was a constructive act for a pro-independence voter.

Nothing that happened in this year's election was remotely analogous to that.  Labour were not making any sort of constitutional offer at all, and there were no Labour-Tory marginals in any case.  If Cat Boyd voted Labour in an SNP-Labour battleground seat, she was helping an anti-independence party against a pro-independence party.  If she voted Labour in an SNP-Tory battleground seat, it was even worse than that, because she was harming both independence and Corbyn's chances of becoming PM.  It was, in short, a very foolish thing to do, no matter which way you look at it.

More good polling news for the SNP - this time from YouGov

YouGov are the only firm I'm aware of that claim to weight the Scottish subsamples of their GB-wide polls separately.  Given the small size of those subsamples, a huge margin of error still applies (something in the region of 8%) but nevertheless it was a cause for some concern when the first YouGov subsample since the election gave Labour a modest lead.  I'm relieved to say that has been reversed in the second post-election poll, which was released today: SNP 36%, Labour 31%, Conservatives 25%, Liberal Democrats 5%, UKIP 1%, BNP 1%.

Those figures are very much in line with the subsamples from the Opinium and Survation polls released on Saturday.  The situation now is that there have been ten Scottish subsamples from various firms since the election, with six putting the SNP ahead, three putting Labour ahead, and one putting the Tories in front.  The information we're going on is admittedly very limited, but it does look as if perhaps Labour have leapfrogged the Tories into second place, but haven't quite managed to overtake the SNP.

Elsewhere in the YouGov poll, there is plenty of other evidence of how Scottish public opinion continues to be radically different from opinion south of the border.  Across Britain, Theresa May has moved back into a small lead over Jeremy Corbyn on the question of who would make the best Prime Minister, but respondents in Scotland prefer Corbyn by a near 2-1 margin.  Across Britain, a narrow plurality feels that the UK is right to leave the European Union, but respondents in Scotland take the opposite view by a whopping margin of 56% to 33%.

SNP buoyed by two new polls

I'm not sure if it was Wimbledon or the Doctor Who announcement that distracted me, but I completely forgot to check on Saturday night whether there were any new opinion polls.  As it turns out there were two Britain-wide polls - one from Opinium and one from Survation.  Both had the SNP slightly ahead on the Scottish subsample...

Survation: SNP 33%, Labour 25%, Conservatives 24%, Liberal Democrats 14%, UKIP 4%

Opinium: SNP 35%, Conservatives 31%, Labour 29%, UKIP 1%, Liberal Democrats 1%

Survation's subsamples are always particularly small and not correctly weighted, but as it happens the recalled vote of the sample on this occasion is reasonably close to the actual result in June, so there's one less reason to be sceptical than usual.

In conjunction with the SNP's creditable near miss in the Elgin by-election, where the swing against them since May was negligible, I'd say these new figures move the balance of probability back towards the SNP having some sort of continuing lead in Westminster voting intentions.  There have now been nine Scottish subsamples from various firms since the general election, with five putting the SNP ahead, three putting Labour ahead, and one putting the Tories in front.  The SNP have never been lower than second place, whereas both Labour and the Tories have been third some of the time.

My best guess is that if a full-scale Scottish poll of Westminster voting intentions was conducted now, it would probably show the SNP with a narrow lead over Labour, with the Tories in third place.  I certainly wouldn't exclude the possibility that Labour have a small lead, but I don't think there's any real danger that the Tories are ahead.

Friday, July 14, 2017

Colonel says "Phew"! Huge scare for Ruth as SNP run Tories close in Elgin by-election

It always looked fairly predictable that the Tories would win yesterday's Elgin North by-election.  They had won the popular vote in the ward in May, their tails were up after winning the Moray parliamentary constituency in June, the local SNP were presumably a tad demoralised, and of course we know from long and bitter experience that Tory supporters are more likely to make it to the polling stations in low turnout local by-elections than supporters of other parties.  Given all of those disadvantages, it's really quite striking just how close the SNP came to pulling it off...

Elgin City North by-election result (first preference votes) :

Conservatives 40.0% (+7.1)
SNP 38.8% (+6.1)
Labour 15.8% (+3.9)
Independent - Monaghan 5.4% (n/a)

We shouldn't get carried away by the increase in the SNP's vote, because like the other parties they benefited from the much reduced vote share for independent candidates.  Nevertheless, the closeness of the result gives us a fair bit of reassurance that things have not worsened for the SNP since the general election in areas where the Tories are their main opponents.  (For what it's worth, there's also no sign of any Tory bandwagon effect in the Scottish subsamples of opinion polls.)  It remains to be seen what is happening in the SNP-Labour battleground areas.

One of the fascinations of local elections conducted under STV is seeing how Labour voters transfer when faced with a choice between SNP and Tory.  The answer in this case was pretty evenly : Conservatives 91, SNP 90.  If the SNP suffer significantly from unionist tactical voting in the next general election, it's unlikely to be in Tory-SNP marginals.  I have my doubts as to whether it will happen very much even in Labour target seats, because Tory voters will surely feel increasingly conflicted about helping a left-wing Labour leadership into power.

Thursday, July 13, 2017

The Tories begin their Great Attempt to Destroy Devolution

So it's official - the Tories have not only betrayed the promise that new powers will come to the Scottish Parliament after Brexit, but they have also announced that some of the existing devolved powers will be taken away.  Unless you count a minor change over powers relating to Antarctica which were devolved by mistake, this will be the first time that powers have been snatched back by Westminster since Devolution Day in 1999.  Don't let anyone fool you into thinking this is happening as an automatic consequence of Brexit - as things stand, the Scottish Parliament has total control over devolved matters except where limited by EU legislation.  For that to change, the Tory government has to effectively repeal parts of the Scotland Act, and that is what they have set about attempting to do today.

Is there any hope that the power-grab can be stopped in its tracks?  Under the Sewel Convention, the Scottish Parliament can withhold legislative consent for its powers to be removed.  We already know that the Supreme Court regards the convention as legally unenforceable (in spite of the fact that it's written into law!), so everything will depend on whether the UK government feels that it is too politically damaging to abandon Sewel.  Remember they will have an eye on the next independence referendum (regardless of whether that happens in two years or in fifteen) and will know that one of the big topics of debate in that campaign will be whether or not "The Vow" was honoured.  If Sewel is ripped up just two years after being written into statute, it'll be extremely hard to argue that the part of "The Vow" relating to the permanence of the Scottish Parliament was fulfilled.

The other big advantage the Scottish government have is that they appear to be of one mind on this subject with the Labour-led Welsh government.  We know that Labour no longer give a monkey's about protecting Scottish devolution, but because of the Welsh dimension there'll be pressure on them to resist anything that undermines devolution in both Scotland and Wales.  Now that we have a hung parliament, a united front between Labour and the SNP could open up the possibility of the Tory government suffering defeats on the floor of the Commons.

*  *  *

Hot on the heels of Julia Rampen's fearless and groundbreaking "Aren't Scottish Labour adorable?" series of articles, the New Statesman have served up a somewhat less innovative "The Nats are doomed!" piece from James Millar.  I just thought I'd do a quick run-through of the highly dubious points made in the article, and also the outright inaccuracies.

* "Many in the party repeat the mantra that they won the election in Scotland, but some sound like they are trying to convince themselves."

In all honesty, Mr Millar, they shouldn't be finding it terribly hard to convince themselves, given that they won the popular vote by a whopping 8.3% margin, and also won 59.3% of the seats. As I've noted a couple of times before, the scale of the SNP's triumph last month was roughly on a par with the UK-wide Thatcher landslide of 1987. It is actually perfectly possible to simultaneously acknowledge that a party won an election, and also lost some ground in the process. Consider for example the difference between the Republicans' showings in the 1984 and 1988 US presidential elections. In 1984, they carried 49 states and won 525 electoral votes. In 1988, that had dropped to 'only' 40 states and 426 electoral votes. The extent of the slippage was noted, but if anyone had tried to claim that the Republicans hadn't 'really' won the 1988 election, they would have been laughed at, and rightly so.

I can't remember if I've ever listened to one of Mr Millar's podcasts, but I did notice that he gave his post-election podcast the understated title of "SNP Apocalypse!"  The mind boggles as to what he would have come up with if the SNP hadn't won the election comfortably.

* "Another [MP] admits that that the result in June could’ve been worse. “If the election had taken place on the Friday rather than the Thursday, I’d have lost my seat. It was one-way traffic to Labour.”"

I've been quick to dispute the claims that the Scottish Labour recovery was a 'mirage', but it's important not to go to the other extreme either.  If you think back to the council elections in May, long before the Corbyn surge, it looked like Labour were competitive in a handful of parliamentary constituencies.  In June, they won a handful of parliamentary constituencies.  The situation was scarcely transformed out of all recognition in the intervening month. I've seen a number of SNP activists contradict the suggestion that there was significant direct slippage to Labour, so it does appear that Mr Millar is only reporting the private conversations that actually concur with his own preferred narrative.

In fairness, no-one can say for sure that an extra day wouldn't have made a difference in Glasgow East or Glasgow South-West...but those seats were so close that a good sneeze could have made a difference.

* "Not only has the group in Westminster been trimmed from 56 in 2015 to 35 just two years later, but many of the survivors have seen their majority slashed, some to double figures, Stephen Gethins’ majority in north east Fife is just two."

Which ignores the fact that the North-East Fife result was comparatively good.  Even when the conventional wisdom was that the SNP would win around 45 seats, it looked like North-East Fife would probably fall.  Holding that one against the tide was a considerable bonus.

* "Many in the party have never known a reverse before. The last time the party went backwards was 1979."

That, I'm afraid, is just complete and utter rubbish.  I could at this point give you chapter and verse on the occasions that the SNP have lost ground in European and local elections, but doubtless someone would come along and insist that there is a big difference between 'first order' and 'second order' elections.  So instead I'll just give you the examples since 1979 that are indisputably from 'first order' votes.

In the 1983 general election, the SNP's vote share fell from 17% to 12%.

In the 2001 general election, the SNP vote share fell from 22% to 20%, and they lost one of their six seats.

In the 2003 Scottish Parliament election, the SNP's constituency vote share fell from 29% to 24%, and they lost eight of their 35 seats.

In the 2005 general election, the SNP's vote share fell from 20% to 18%.

Conclusion?  You'd have to be very, very young not to be able to remember a time when the SNP went backwards in an election.  And in truth, if anyone out there wasn't expecting some kind of correction after a freakish election in which the SNP won 50% of the popular vote, they were being a bit naive.

Tuesday, July 11, 2017

Relearning the oldest truth : voting Tory doesn't produce results for Scotland

I've been having my latest technological meltdown, but 24 hours late, here's a quick note to let you know I have a new article on the TalkRadio website!  You can read it HERE.

*  *  *

After an insanely long wait, ICM have finally released the datasets from last week's GB-wide poll.  The Scottish subsample figures are: Labour 37%, SNP 32%, Conservatives 23%, Liberal Democrats 3%, Greens 2%, UKIP 2%.  The Labour lead can be partly explained by the fact that respondents who said they would vote SNP were sharply weighted down from 64 to 43 - that may have happened for good reasons, although if we assume YouGov are right that a disproportionate number of SNP-inclined voters simply didn't turn out on June 8th, weighting to past vote recall may start to underestimate the SNP's potential strength.

We've now had seven Scottish subsamples from various firms since the election, with three putting the SNP in the lead, three putting Labour in the lead, and only one putting the Tories in front.  I think all we can say with confidence now is that it looks like a tight three-way battle, and that the Tories probably aren't in first place.  I'm not convinced the Scottish Labour recovery will survive any return to public infighting between the Corbynites and the "moderates", so perhaps that's what we should be keeping the closest eye on.

Friday, July 7, 2017

Tories take a hammering in new YouGov poll

If the Tories imagined their narrow lead in the GB-wide Survation poll a few days ago was some kind of turning-point, they appear to have been mistaken.  We've since had an ICM poll (albeit one with radical methodological changes) putting them two points behind Labour, and tonight brings word of a YouGov poll putting Labour eight points clear - the biggest Tory deficit for many a year.

Of most interest to us, of course, is the Scottish subsample, and the news isn't that great : Labour 36%, SNP 31%, Conservatives 25%, Liberal Democrats 5%, Greens 1%.  This is the third subsample that's put the SNP in second place since the election, but it's the first one of those that I take remotely seriously, because the previous two were absolutely tiny Survation subsamples which you could tell were clearly skewed by looking at the past vote recall of respondents.  As far as I know, YouGov still weight their Scottish results separately, which means they should produce results that are slightly more accurate and stable than other firms' subsamples.  However, even when properly weighted, a subsample has a much greater margin of error than a full-scale poll - I make it roughly 8% in this particular case, which means YouGov's figures are entirely consistent with a small SNP lead.  I still think that's the most likely state of play, because in total we've had six subsamples since the election, with three putting the SNP in the lead, two putting Labour in the lead and only one putting the Tories ahead.  It's also significant that the SNP haven't been in third place in any of the subsamples, whereas both Labour and the Tories have.  However, even if the SNP do still have the advantage, it's plain enough that we're faced with a fairly tight three-way battle for the time being.

'For the time being' are the operative words, because it's entirely predictable that a major political shock like the one we saw last month will radically shift the opinion polls in the immediate aftermath.  Sometimes the change is superficial and temporary (for example the brief Conservative surge during the fuel crisis of 2000 which was completely reversed at astonishing speed), and sometimes it's meaningful and lasting (for example the Tory slump after Black Wednesday in 1992, which they didn't properly recover from for well over a decade).  We'll just have to wait and see which category the current situation falls into.  The irony is that if Labour have hopes of exploiting this apparent moment of relative vulnerability for the SNP and winning back a truckload of central belt seats, their own strength in the opinion polls may end up preventing them from doing so.  The Tories aren't going to willingly call an election unless they think there is a good chance of winning an overall majority, and as I've noted before, there is no realistic prospect of them being forced into an election by a vote of no confidence in the Commons for at least three years (unless Tory MPs defect to other parties).  The more I've thought about the parliamentary arithmetic, the more I've come to the conclusion that it's not totally inconceivable that the new parliament will stagger on for the full five-year term.

The Tories are very lucky that the exit poll was wrong about them only having 314 seats.  The difference between 314 and 318 may not sound all that great, but it could well be enough to swing the balance between a very short parliament and a very long one - which apart from saving the Tories' bacon, could also rescue the SNP from having to defend their own 35 seats for a good number of years.  Ideally Scotland will be an independent country by 2022 (Nicola Sturgeon's recent statement very much leaves that possibility open), but even if it isn't, Labour could be in a completely different place by then.  There are already plenty of signs that the truce between the Corbynites and the "moderates" is starting to break down.

*  *  *

Stormfront Lite's notorious Deputy Editor is a reliable source of painful prose and unwitting comic genius, and yesterday morning's effort was no exception -

"I know I’m not the only Tory who somehow hopes Ruth Davidson becomes an MP before the next Tory leadership election but I suspect she sees her role for the next few years as ensuring Scottish Nationalism really is killed stone dead and that can only be achieved in Holyrood and not Westminster."

Which suggests that Scottish nationalism hasn't "really" been killed stone dead thus far?  Well, quite.  It's a tad difficult to boast about killing Scottish nationalism when you've only just been beaten by Scottish nationalists in a sixth successive nationwide election.

The roll of shame for Ruth is : the 2012 local elections, the 2014 European elections, the 2015 general election, the 2016 Scottish Parliament election, the 2017 local elections, and the 2017 general election.  Every single one an SNP victory.

Tuesday, July 4, 2017

Scottish Labour may be dinosaurs, but it wasn't a dead cat bounce

As you may have seen on social media, there was a post on the LSE politics blog yesterday arguing that the notion of a "Corbyn bounce" in Scotland at the general election is a "myth" and a "mirage", and that the SNP should not be sidetracked into strategies to fend off Labour when the Tories are the real enemy.  It would be comforting to agree with that, and to tell ourselves there is only one unionist party worth worrying about, but I do think the conclusion is wrong.  The main flaw of the blogpost is that it treats the results of the 2015 and 2017 general elections as if they are the only pieces of information available to us.  When you look at it that way, it does appear superficially that the entire drop in the SNP vote can be explained by a natural 'correction' after the freakish result of 2015, and that the very small revival in the Labour vote (from 24% to 27%) was an inevitable side-effect of that readjustment, rather than being caused by the phenomenon that generated a much bigger Labour surge in England.

However, if you widen your gaze to take account of opinion poll evidence (and indeed the Holyrood election of last year), the picture suddenly looks very different.  There is overwhelming evidence that Labour's true recovery in Scotland was not from the low of 24% recorded in 2015, but from the much worse position that the party slumped to after that election.  Survation, who proved to be the most accurate pollster, had Scottish Labour languishing at just 18% as recently as mid-April, but by the end of the campaign that had jumped to 29%.  Other firms showed a similar trend (even if the exact figures were very different).  That sort of big shift is much more in line with the Corbyn surge that occurred in England, and given that it happened at exactly the same time, it's not unreasonable to suppose that it probably happened for much the same reason.  It can't really be explained by the correction in the SNP's vote share, because the SNP had dropped to the low 40s (and the Tories had risen to the high 20s) before the Labour surge even started.  It looks much more likely that the SNP suffered a drop from an unsustainable high, and then suffered a further small drop as a direct result of voters switching to Labour because of enthusiasm for the Corbyn project.

Even though the premise of the LSE blogpost is wrong, I do think it's correct to argue that the SNP shouldn't head off on a radical left wild goose chase to try to deal with the Corbyn threat.  If you're losing voters who are inspired by the prospect of a radical left government at Westminster, you can hardly counter that by offering a radical left opposition at Westminster (and an opposition with third-party status at that).  You can only compete by putting forward an alternative inspiring vision that Corbyn can't/won't offer - and that means a much greater focus on independence than we saw in the recent campaign.

*  *  *

Elsewhere in the LSE blogpost, a reasonable point is made about the SNP's "heroic" determination to conflate support for independence with support for remaining within the European Union, which may have cost the party votes from Brexit supporters.  That factor may also go a long way towards explaining the Tories' seemingly puzzling failure to defeat the SNP in places like Perth & North Perthshire and Edinburgh South-West - both constituencies that voted Remain by a much more emphatic margin than the rural north-east did.

But this is not just a dilemma for the SNP.  Could the Scottish Tory surge be a similar phenomenon to the SNP surge of 2015?  In other words, was it partly caused by temporarily energised Brexit supporters who were determined to reinforce their vote from the referendum last year?  If so, the Scottish Tories are very likely to suffer their own natural 'correction' at the next general election (as long as it doesn't take place in the near future).  Those new north-east Tory MPs, especially the ones with the narrowest majorities, probably shouldn't get too comfy in their seats.