Scottish Westminster Projection - First Of Twenty Nineteen

Which Side Are You On? 🔊

There have been 22 full Scottish polls fielded between June 2017 and January 2019. The only ones I use for my Scottish projections as subsamples of GB-wide or UK-wide polls are volatile and unreliable. Though sometimes I can't help quoting some notable subsample on Twitter when it's really tremendously good for the SNP, but with the strongest of caveats. The National also quoted one not long ago but their caveats bordered on the inaudible. Good one anyway lads. So here is what the true full polls say:

Of course a cropped screengrab from Wikipedia is not the easiest of reads. So let me translate this to a better looking representation: trendlines based on a three-point rolling average.

The trend is clear and illustrates the truth about Scotland's state of mind, whatever the Unionists say. The SNP is the first party and is far from being in a downward spiral. Quite the opposite in fact with recent polls also the ones offering better prospects for the SNP than we have seen at the end of last year.

The Sky Is No Limit 🔊

Let's see what all these polls have to offer for the SNP: voting intentions from each of the 22 polls and the projected number of seats. Note that we again see one of oddities of FPTP at work. In some cases the SNP is predicted to bag a bigger share of the popular votes but fewer seats. Which is not as odd at looks at first glance because the seat projections rely not just on one party's share of the vote, but more importantly on how all parties' vote shares change relative to each other. Just remember the 2017 GE delivered 20 marginal seats in Scotland, most of them really close and most of them SNP-Labour marginals, and it's easy to see how the SNP could be up 1% of the vote but down 2 seats.

Of course paradoxical situations are the exception rather than the rule and the big picture is definitely good for the SNP. First of all voting intentions over the last 19 months never fell below 36%, that is barely 1% lower than the 2017 result. Then only four polls out of 22 have the SNP losing seats, and these anyway look like outliers when compared to polls fielded over the same periods. Besides the point is quite moot as the most recent of all polls predicting the SNP losing seats is a year old and the trend has become much more favourable since. But the SNP is not the only player in this game and here is what the 22 polls predict for the whole Scottish representation in Westminster.

As I hinted earlier SNP's low points are also Labour's high points as the SNP-Labour marginals would be the first to fall if Labour managed to significantly close the gap with the SNP. Think Glasgow East, Glasgow North West, Inverclyde or Motherwell and Wishaw among others. Then of course it would take the stars aligning in a most peculiar way to actually achieve this, so it is by far not the most likely scenario and certainly not what recent polling predicts.

Late last year John McDonnell boasted that Labour were breathing down the SNP's neck but nothing can be further from the truth. Quite the opposite as SNP's lead over Labour is now about 15% when it was 'only' 10% in 2017. And Labour aren't even breathing down the Conservatives' neck in Scotland, which should be happening if they had built a really convincing case. Of course this does not mean the SNP should take anything for granted as they're prone to do because MPs just won't re-elect themselves. Key to success is as always a strong turnout.

59 is the limit at least in theory, and odds of reaching it are close to non-existent. But the SNP should raise the bar as high as realistically feasible and that is in the vicinity of 50 under current polling and possibly 55 in the near future if the SNP do all the right moves on all the best campaign themes. Back to the 2015 tsunami or close might sound far-fetched but it possibly is not and anyway nothing says even an uphill battle can't be won. All it takes is to fight it.

Light The Sky On Fire 🔊

As we saw earlier the most recent Scottish polls are definitely good for the SNP. The last one was fielded by YouGov from 21 December to 4 January as part of their mega-poll for People's Vote. With almost 3k respondents it certainly delivers a pretty accurate picture of where Scottish public opinion stands right now and the voting intentions are good news for the SNP in many ways.

This would amount to roughly a 4.5% swing from Labour to the SNP and/or 3% from Conservatives to the SNP. Not bad for a party some claim has lost touch and is declining. The problem with polls though is that they always include Green, UKIP and Others is their prompts as if these would field candidates in every constituency. This was not the case in 2015 and 2017 and is also very unlikely to happen at the next GE.

So here we have 6% of the vote that would probably shrink to less than 1% on Election Day and are up for grabs. Best estimate is that Green votes would go to Labour and SNP, UKIP votes to Tories and Others to about everyone. Which would roughly make the 40-25-21-8 shares in the poll morph into 42% SNP, 26.5% Tories, 22% Labour and 8.5% LibDems at the actual election.

My model takes care of this with the assumption that minor parties would field candidates at the next GE only in the constituencies where they stood in 2017 already. So the reallocation of votes is done automatically on a seat-by-seat basis. Which is how all prediction models work anyway. Based on the early January voting intentions, here is the distribution of seats we would get (by SLLM rating and summary):

For the record Election Polling finds the same result while Electoral Calculus projects 45 SNP, 9 Conservatives, 4 LibDems and 1 Labour. So everybody agrees on a significant success for the SNP though short of the 50 seats that would make it a true triumph. Aye I do believe we have to set the bar that high. For the moment. Until we raise it again.

Winds Of Change 🔊

This time eleven Scottish seats would change hands: six from Labour to the SNP and five from Conservatives to the SNP. Labour losses are only to be expected as their 2017 intake mostly won by small margins in constituencies with a strong SNP showing in 2015. Besides Labour steadily and conclusively fail to make their case in Scotland. Scottish Labour MPs are conspicuously silent in Westminster and the only one with a noticeable presence on social media is Ian Murray, the one that should know to shut the fuck up before making an arse of himself again and again.

The SNP making serious inroads into Tory territory is excellent news as these seats are the key to a major success. What the SNP need to stress is how disunited Scottish Tories are. When one of them trashes another's public positions as 'rubbish' while being dubbed 'SNP gain' by his colleagues, you have definite proof there is something rotten in Toryland. So five seats lost may very well be only just the start, as the updated data about SNP target seats show.

The data above are the 2017 results for the 24 non-SNP seats. Highlighted in yellow are the ones projected to switch at the next GE and pale yellow indicates those close to switching. Four more Tory seats. LibDem MPs appear much safer as voting intentions show LibDems going up, not spectacularly but enough to keep them relatively safe from the SNP surge. Then I think everybody should pay more attention to Edinburgh South. If other election results are any indication, the seat is actually closer to a three-way marginal than a Labour stronghold. It's common knowledge that Ian Murray was propelled back to Westminster by massive tactical voting and not all of his campaign funds came from Labour donors. Some day Tories will smell blood and realize they have a shot at gaining the seat. Then will be popcorn time.

On The Threshold Of Fire 🔊

Current polling would leave only nine marginal seats in Scotland instead of twenty after the 2017 GE. With eight of them SNP-Conservatives battlegrounds as Labour is again sent down the pipes by voters. And the four potentially re-elected but in the Danger Zone Conservative MPs may indeed find themselves in a worse situation than the math says. If any of them made a durable impression on his constituents, it's obviously a negative one and I would not be surprised to see at least the Two Rosses flushed into oblivion at the next GE no matter what raw numbers say.

Alternate scenarios right now are extremely positive for the SNP as even the worst case would have them 6 seats up. Only one SNP seat (North East Fife) could possibly switch as it was the closest result UK-wide in 2017 and LibDems doing well in polls makes a gain a credible outcome. SNP should take the threat seriously as this was a Liberal and then LibDem safe seat for close to 30 years before the SNP gained it in 2015. Also don't forget that the almost coterminous Holyrood constituency has been held by LibDems for 15 of the 20 years since the first Scottish Parliament election. If even Willie Rennie can unseat the SNP there then the Westminster seat could really be in jeopardy.

The only caveat here is that my projection is based only on the last YouGov megapoll published three weeks ago. But adding the earlier Panelbase poll (nine weeks old by now) to the mix would not greatly alter the big picture as the weighted average of both would still have the SNP between 40 and 48 seats.  So still a major success if one step back from a genuine triumph.

Ride The Tiger 🔊

Finally the usual breakdown of seats by party from worst to best for the SNP. Note that seven of the eight residual Tory seats are in the South or in the North East. Try to factor in how the farmers in Dumfriesshire and Berwickshire reacted when hearing that £160mil of EU funding had been stolen from them by the English Government to be 'reallocated' to English farmers. Or how fishermen in Fraserburgh and Peterhead felt when openly sold out by the English Government during Brexit talks. Guess that takes some Tory MPs several steps closer to the woodchipper than projections can say.

Starting form the voting patterns shown by the latest YouGov poll, the SNP would need a further 2% swing to reach the symbolic 50 seats and bagging 45% of the popular vote would deliver 54 seats. This would have looked like an impossible hurdle to clear just a few months ago but Brexit just 59 days away makes such scenarios more credible. All it takes is Scottish communities getting a clearer and fairer assessment of the damage Brexit will inevitably do, and choosing the MPs who have always made it their priority to defend Scotland's best interests first and foremost.

Let's Go Together 🔊

What we need right now is an united Yes Movement and everybody within it accepting that the SNP is the key player simply because they're they only ones who can get MPs elected as long as we stay in the UK and the only ones able to form a Scottish Government seeking Indy. Works both ways though: the SNP can't thrive without the Yes Movement just as the Yes Movement can't thrive without the SNP.

What the SNP must do next is simple: remember what went wrong in 2017. First and basic point is not to be dragged into debating devolved matters and how the Scottish Government handles them. This is not legitimate GE campaign material and I still remember that painful part of the 2017 debate when Nicola Sturgeon let herself be dragged into it and Patrick Harvie was the only one objecting it had no place in a GE debate. Second and more important: put Independence squarely at the center of the campaign as I already argued ten days ago. This is legitimate GE material as a majority of SNP MPs is a key component of the triple-lock mandate to seek Independence.

Remember that the SNP lost 477k voters in 2017 compared to the 2015 result. It is likely part of them switched to other parties, mostly to Labour. But the bulk actually chose to abstain because the SNP did not get the Independence message through loud enough. Best educated guess is that about 250-300k of these abstainers could have been held with a different campaign, taking the SNP vote to 42-45%. Abstainers probably cost the SNP 7 to 10 seats. So you know what to do next time. Putting Independence squarely on the table can't harm, quite the opposite.

Until then stay tuned for further polling and updates on post-Brexit martial law.

Wha daur meddle wi' me

© Paul Kantner, Grace Slick, Marty Balin 1974

Westminster Projection - First Of Twenty Nineteen

What Goes On?

New year and new polls but the new batch look even more puzzling than what we had at the end of last year. The Prime Minister of England suffered a crushing and humiliating defeat over her not-that-much-of-a-deal Plan A. And since she has no Plan B worth mentioning the UK is on a collision course with the Brexit Death Star. Yet that same Prime Minister that was deemed unfit to lead the Nasty Party by 117 of her own was next declared fit to lead the UK by all of them but two, plus all of the Blackmailing Unionist Party. But fit to lead where exactly?

Early 2019 polling shows that public opinion are just as ambiguous on where they actually stand as Jeremy Corbyn on Brexit. At least Corbyn can be credited with two successes: Labour were not strongarmed into supporting the Mayxit Deal and they successfully whipped their MPs into opposing it almost unanimously. So Labour avoided both first-class drubbing scenarios tested by YouGov in December. But where public opinion goes from there is far from clear as polls again show both major English parties losing votes from 2017. And there's no way the Westminster Weekly Freak Show Formerly Known As PMQs will restore any kind of public opinion confidence in anyone.

Public opinion can't fully trust a factionalized party they fear will make a pig's ear of Brexit no matter what the recipe, so Tories lose votes. But neither can public opinion fully trust a factionalized party they fear will make a pig's ear of Brexit whatever the recipe, so Labour lose votes. This is compounded by the leaders' approval rating. The most recent Opinium poll found May with a net -20% rating and Corbyn with a net -30%. Interestingly Opinium finds Nicola Sturgeon with a net -13% overall including English voters, but a net +2% in Scotland where it really matters. And 'preferred Prime Minister' polling still delivers 'None of the above' as first choice with May leading about 60-40 among those who express an actual preference.

Unfortunately pollsters offer only a two-way choice between the current Prime Minister of England and the current Leader of the Official English Opposition. So we'll never know who would emerge as the really preferred prospective PM in an STVish contest. Sturgeon anyone?

Beginning To See The Light

Labour will undoubtedly find some comfort in polls showing them closing the gap with Tories. But it's far from a reversal of fortunes, yet. Labour's electoral map connoisseurs surely know that differential turnout and deep sinkholes definitely play against them and that even a 3% lead is not enough for a majority. Sure there is light at the end of the tunnel for Labour, but right now it feels more like they're stuck in Dover staring at the Calais end down yonder. Not the time yet to have Wetherspoons' Surrey Sparkling delivered to Victoria Street and Kings Manor.

Unsurprisingly with the current sequence of polls the trend is unkind to both Labour and Tories. My current poll-of-polls includes the last six fielded between 13 and 18 January. Super-sample size is 10,824 with a 0.91% margin of error and it predicts a tie. Exact weighted average is 38.27% for Tories and 38.13% for Labour, so Tories are projected 43,881 votes ahead UK-wide. That's how precise British polls are these days. Naw they aren't, just kidding.

We have seen this kind of pattern before with Labour and Tories both leaking votes in every possible direction. Conservatives probably mostly towards UKIP and Labour towards Greens, and both to LibDems and SNP too. One of the oddities of current polling is that 'Others' are doing well too, projected to double their vote share in Great Britain. But 'Others' are that kind of amorphous nebulaish ensemble ranging from the West Dunbartonshire Community to the Monster Raving Loony Party through the Village Green Preservation Society. So they could probably field 375 candidates overall as they did in 2017 and fail to get even one elected.

Though of course local factors have already had and still can have a decisive impact in a few constituencies. Clues: Tatton 1997, Wyre Forest 2001 and Blaenau Gwent 2006 by-election. This time just look more closely at East Devon. Just saying.

Men Of Good Fortune

The seat projection has UK again trapped in the already infamous stalemate where no viable governing coalition is likely to happen. Unless of course Jeremy Corbyn finally accepts to boldly go where no Willie Bain has gone before and invites himself to Bute House for some friendly chat with the First Kingmaker of Scotland. Just don't take this never happening for granted as extreme situations call for extreme responses and anyway DUP would not have enough seats to propel English Socialism 2.0 to Number Ten even if you bribed them £10bn.

By a strange twist of fate Jezza might even find himself willing to lend an ear to a request for yet another Section 30 Order. Because such a tight result would actually land England in the statistically extremely unlikely spot where losing 59 Scottish MPs would not favour the overall winner but the runner-up: Labour. With just one Labour MP gone but eight residual Tories out, Labour would become the first party and a Lab-Lib coalition would bag a three-seat majority in Scotlandless and SinnFéinless Rump Commons. Food for thought.

The English Labour Party would find the prospect of Pesky Jocks' Independence even more appealing if they managed to get their act better together in Wales in London and beat currently dismal local polling, which has them barely holding all their Welsh seats and losing seven in London because of their not-fully-thought-through Brexit stance. Just holding all their London seats would take them to 285 seats in post-Indy Commons, and the hypothetical Lab-Lib majority would rise to a more comfortable fifteen seats. Very little needed indeed to achieve this.

How Do You Think It Feels?

On current polling the body count would be 56. But the detail of gains and losses shows a mixed bag and nothing Labour should actually celebrate wildly. The Scottish Branch Office losing six seats to the SNP is only the fair and balanced reward for their neverending show of incompetence and irrelevancy. But losing six seats to Tories and one to LibDems in Jezza's own London backyard would undoubtedly be a harder pill to swallow.

The full manifest of fatalities would deliver some interesting PortiBalls moments. Labour could pride themselves on unseating no fewer than 13 Tory MPs holding various government positions. Biggest prize of course would be decapitating DWP with both Amber Rudd and Universal Credit Minister Alok Sharma being shown the door by voters. Solicitor General for England and Wales Robert Buckland and Minister of State for Agriculture and Dark Money George Eustice also going down would be good news. Then Labour would have some lesser prizes with three Whips, five PUSSs and one PPS off the payroll.

But Labour too would suffer some painful losses with Shadow Secretary of State for Scotland Lesley Laird unseated along with three Shadow Ministers and two Shadow PPSs. Not that Laird will be missed much North Of The Wall as she is just as unconspicuous as an actual shadow in a dark room. And in case you wonder what a Shadow PPS actually does just don't ask. The true answer might hurt the feelings of the persons concerned. Just remember that a government PPS's job is basically bag-carrier and gossip-gatherer, and not even on the payroll but supposed to act as such on divisions. In other words, a glorified intern unpaid to rat.

Set The Twilight Reeling

Four dozen seats overall would fall into the Twilight Zone where the swing needed to switch them would be lower than the margin of error, including 22 of those projected to change hands. Of course you shouldn't always listen to what the math said as local factors can have a huge impact on the result too. Best educated guess is that Plaid Cymru would hold Ceredigion and DUP would hold Belfast North by the smallestish of margins even if the model has them changing hands. Closer to home I have a hunch such Scottish Tory luminaries as Ross Thomson, David Duguid, Douglas Ross and Paul Masterton will bite the dust anyway as they have alienated enough of their constituents to trigger a backlash even among their own 2017 voters and reverse the effects of tactical voting back then.

And of course 'within margin of error' tells only part of the story as it points only to seats more likely to defy the model and go a different way than predicted. If you broaden the spectrum to include all seats rated as 'Marginal' or 'Lean' by commonly accepted definitions, then 141 seats overall can now be considered competitive, roughly the same number as in 2017.

Competitive seats would include 62 projected for Labour, 53 for the Conservatives and 9 each for the SNP and LibDems. 165 Conservative seats and 199 Labour seats qualify as safe, including Labour's deep sinkholes that generate nothing but wasted votes by the 100ks and put Labour at a disadvantage compounded by unfavourable differential turnout. The SNP is projected with 11 safe seats and LibDems with 9.

The Valley Of Unrest

Right now even the extreme scenarios based on the reallocation of ultra-marginals fail to deliver a path to strong and stable government. As we have already seen with a tied popular vote the best you can get is a weak first party and wobbly minority governments, even with the broadest of reasonably feasible coalitions. With the added irony (or absurdity) of current rolling average delivering Conservatives as a first party, but a Labour-led coalition likely to get the most votes in Commons. So more than ever this is the perfect recipe for an instant snap election to sort out the mess.

This kind of situation demonstrates that we have probably reached the limits of what the Westminster System can offer. It was never meant to go beyond a two-party powerplay and started to fall apart 35 years ago when the Alliance bagged 4% of the seats on 25% of the vote. Recent polling of alternate scenarios has shown that history would repeat itself as LibDems could get at most 8% of the seats even if they managed to finish second on the popular vote with >25%.

Something has to be done to break the pattern and I am absolutely not advocating Proportional Representation which would only make matters worse. So you can think either Instant Runoff provided you keep single-member constituencies, to avoid the sort of STV-based free-for-all you see in the Northern Ireland Assembly. Or a true Mixed-Member System like the one I advocated some time ago to replace the current Additional Member System in Scotland. Anything that would both be more representative than pure FPTP while avoiding the inbuilt ungovernability of PR. Food for thought for another referendum?

That's The Story Of My Life

I again compared my results with what you get from Electoral Calculus and Election Polling, my two competitors readily available online, with the same polling data. The difference between the projections is 10 seats or less. As I pointed out already Uniform National Swing tends to favour the incumbent as it generally dampens the effects of small or moderate swings while Proportional National Swing tends to amplify the impact of swings, and the bigger the swing the bigger the differences in projected seats. So what we get here shows nothing out of the ordinary given the different way the three models process raw data.

In the run up to the 2017 GE seat projections by various predictors differed by as much as 25 seats two or three weeks before the election and it rose to a massive 60 in the last week when YouGov changed their prediction model. It is worth remembering that YouGov, despite all suspicions of a pro-Tory bias on their part, were the only ones to predict a hung Parliament even if their seat numbers did not match the actual result, most notably wildly overestimating the number of SNP seats. But of course everybody did back then, me included.

I also went though all my projections since I started this blog last July. Sixteen projections overall and most point to a hung Parliament where finding a majority would be trickier than in the present one.

Labour are projected to become the first party in only three cases out of sixteen. But the Conservatives bag an outright majority without DUP in only one case. Otherwise the continued Tory-DUP coalition does no better than a possibly workable minority (> 315 seats) in only two other cases and otherwise falls short. The oppositions don't really do better with an hypothetical Lab-Lib coalition able to venture a minority government in only one projection out of sixteen. Numbers are merciless: a coalition of all current opposition parties ousts the Tories in fourteen out of sixteen projections, and Labour's pigheaded refusal to talk with the SNP is the main and probably only reason why Jeremy Corbyn's chances at becoming the next tenant of Number Ten are slim to none.

Ride Into The Sun

Contrary to common wisdom I think that this year's events so far have actually strengthened Theresa May. Of course she lost the meaningful vote on the worst result ever for a Government in recorded parliamentary history. But the day after the very Tory MPs who considered her unfit to lead the Nasty Party backed her almost to the last man to declare her fit to lead the UK over the cliff unfettered.

To paraphrase Margaret Thatcher, May could rightly say her greatest success is Jeremy Corbyn. Many in England, and some in Scotland, had high hopes when Corbyn became English Labour's Suryong (aye, you may want to Google that one). Then you have to wonder if he actually is the Leader of the Opposition, or was it Angus Robertson and now Ian Blackford? Some days even Anna Soubry fills these shoes with a Cheshire Cat grin.

Now I believe that, contrary to all expectations including mine, the Maybot has finally earned her 12-month stay of execution, well now just 10.5 months but that will be enough for her to finish the dirty jobs. As long as DUP continue to prop her all prospects of a snap GE are gone, all it takes is £1bn a year until 2022.

Not exactly what I would have put a tenner on just a month ago but politics is all smoke and mirrors and upsets after all. Besides I think Labour don't want a snap GE that much and the trends of GE polls easily explain why. Or else they wouldn't have made such a clusterbùrachfuck of the vote of no confidence, which was never predicted to succeed and left Labour both proud they can brag they tried (like in 'look Mum no hands') and relieved it failed.

On an unrelated note you might remember what I said earlier about far-right blooming in the putrid non-hostile environment created for them by extremist Brexiteers and the gutter press. Head of counter-terrorism Neil Basu has expressed concern about the rise of tensions fuelled by far-right propaganda and taken the matter to Commons. Can't say we haven't been warned.

Meanwhile Jeremy Corbyn seems to have catch22ed (or does caught22 work better?) himself into a corner with his own brand of Brexit Shambles. On one side you have frontbenchers threatening to resign and the fear of electoral backlash. On the other you have frontbenchers threatening to resign and fear of electoral backlash. So Socialist Middle Ground now is support for a vote on the vote but no support for the vote. Of course the only thing you should take for granted is that Labour's position can still evolve depending on prevailing winds. Though the end result would probably look like running away from marshswamps straight into a bear pit buried in quicksand, as People's Vote is again proven to be so toxic even a sixty-foot titanium-clad bargepole fails to be shield enough from the radioactivity. Told you I told you so.

Until next time stay away from BBC and Herald Scotland but don't forget to fire up the cooshite detector.

Cha togar m' fhearg gun dìoladh

© Lou Reed, 1969

And be thankful I spared you any teethgnashing live version of 'Sister Ray'