We’re a nation that’s gone so far up its own arse it can’t move from the crap everywhere
(Peter Boyd, Waking The Dead: Anger Management, 2004)
© Tony Banks, Phil Collins, Michael Rutherford, 1986
You can’t understand the picture if you’re inside the frame
(Peter Boyd, Waking The Dead: Final Cut, 2003)
There is an interesting paradox in current voting trends, and one that is quite counter-intuitive even for me. It might well prove easier for Labour to form a government in a hung Parliament than for the Conservatives. Bear in mind that, whatever the exact voting intentions and with Sinn Féin excluded for the usual reason, there will always be a bloc of 80-85 MPs who would be neither Labour nor Conservatives, and could exert massive influence on the post-election events. Of these only the DUP could be expected to support a Tory government, or perhaps not. So the likely scenario is that the Conservatives would indeed need an outright majority, or else they might face a Coalition Of The Unwilling ready to vote down each and every bill. While Labour could count on support from the SDLP and some sort of agreement with Plaid Cymru, the Greens, the LibDems and the Alliance Party, who have de facto become the local branch of the LibDems in Northern Ireland. Meaning Labour could count on an additional 15-20 votes and form a viable minority government with as few as 280-290 MPs of their own, as in this case the Conservatives would be down to 275-285. Of course, all these scenarios can happen only if Labour and the Conservatives are really close in the popular vote, somewhere between a tie and Labour leading by a couple of points. Which has happened at the end of last year in many polls, so let's just see first what the trends look like now, two months into the Year Of The Snoek, shall we?
What the 2021 polls tell us is that we're back to the situation we had in midsummer, which is definitely more of a trogglehumper than a night's dream for Labour. Personal image matters too and people seem to have forgotten what a flushbunking blatherskite Boris has been over the last twelve months. Next thing you know, we're gonna have a statue of Saint Boris Slaying The Virus in the middle of Parliament Square. Just the kind of narrative that definitely works in Leafy Eastern Surrey, though admittedly less so in the Red-To-Blue-To-Red Wall counties. But even there, you can see signs that lingering doubt has given way to benefit of the doubt. Interestingly The Guardian, never short on contradictions, now advise Labour to devote more efforts to South England, pretty much repeating my earlier arguments that New New Labour will win the next election if they crawl back to the same level of support New Labour enjoyed in the Home Counties, even at their low point in 2005. Which is definitely not what the last batch of polls predict, as Labour is left bopmuggered by the number of voters who have whiffswiddled back to the Tories since the world-beating Snoek Deal, and even more so since the First Minister of England is rollicking in his deepfreeze-ready Pfizer Honeymoon. And that's the moment Keir Starmer chooses to go natterboxing rommytot about how giganticus weapons of mass destruction are, which will probably not really help either in Hampstead or Glasgow. If he was really paying attention, Sly Keir should feel scrotty over where public opinion stands now, based on the weighted average of the last four published polls, conducted between 19 and 26 February. Super-sample size is 6,307 with a highly theoretical 1.23% margin of error.
Before we proceed any further, there's a necessary and long-overdue caveat, about the SNP's vote share in GB-wide or UK-wide polls. Which is, believe it or not, overestimated, or rather represented above its actual level because of the pollsters' weighting mechanisms. According to ONS data, Scotland accounts for 8.2% of the UK's population, or 8.4% of Great Britain's. But higher voter registration means Scotland accounts for 8.5% of the UK's electorate, or 8.7% of Great Britain's. Then higher turnout at the 2019 election meant that Scotland accounted for 8.6% of all votes cast in the UK, or 8.8% of all those cast in Great Britain. And now our problems start. Based on the four main pollsters' published data for their most recent survey, here is the share of Scotland in their original sample, and then after excluding undecideds and weighting by likelihood to vote. I included only each's last poll here, but you can trust me it's all the same in past polls, and sometimes even worse, depending on the relative weight of undecideds that are scratched from the final 'headline result'.
YouGov publish only one set of numbers, which I assume represent the original makeup of their sample, without any hint on how weighting might alter that. So the makeup of the original samples is something of a mixed bag, but more or less accurate or inaccurate depending on your mood. But the post-weighting share of Scotland is definitely and conclusively overestimated by three of our main pollsters, and quite possibly also by the fourth one even if we don't have the smoking gun here. Which is why you see the SNP vote skyrocket from 3.98% in 2019 to fantastically implausible levels like 5-6%, while low-to-mid 50ish% of Scottish vote is more like 4.2-4.4% of the GB-wide vote. And this also gets you biffsquiggled and wondering why the SNP don't systematically bag all 59 fucking seats when their vote share is so razztwizzlery high. Fortunately, real Scotland-only polls are here to remind us that the SNP really do better than in 2019, though it might not always be such a landslide as you might think at first glance. For example, this week's full Scottish poll from Survation has the SNP on 48% of Commons voting intentions, which would be something like 4.1% GB-wide, and yet my current Poll'O'Polls says 4.2%, which would be some 51% of the Scottish vote. Just saying, so that you don't get your hopes too wildly high, especially when the trend sees the SNP slowly losing votes.
If you want the high life, you have to embrace some serious lowlife
(Claudia Baxter, Silent Witness: River's Edge, 2016)
© Tony Banks, Phil Collins, Steve Hackett, Michael Rutherford, 1976
It strikes me that the human race is supposed to be more evolved
If you look at all this, you've gotta ask yourself if, as a species, we are on the right track
(Peter Boyd, Waking The Dead: The Hardest Word, 2004)
What polls tell us now is that the Conservative lead over Labour would be cut by almost half compared to 2019 and back to a 2015ish level. My model says it would take Boris just a few seats shy of a majority, but far from what is needed to pull a bonking stonker this time. Such an outcome would definitely not mean that Britain's Got Less Talent, as most of the fatalities here would be pibbling notmuchers from the 2019 intake, whose main contribution to the political debate has been the perfect triangulation of arse-licking incompetence. The only really noticeable Tory fatality here would be Graham Brady, Chair of the 1922 Committee of unruly backbenchers and a born-again Covidiot trying to outfox Nigel Farage on his turf. As it stands, Conservatives would gain only four seats: Alyn and Deeside, Dagenham and Rainham, Canterbury from Labour; Westmoreland and Lonsdale from the LibDems. The latter if they find a candidate who can outbigot Tim Farron on Christian values and homophobia, which shouldn't be too difficult when you think of it. Of course there are also even odds on Tories snatching Ayr, Carrick and Cumnock from the SNP. Just don't tell Douglas Ross and Andrew Bowie as they would drool at the prospect of a three-seat gain.
Odds are Bozo could even bribe the DUP back into some shade of supportivish neutrality, if he has enough pocket change left after spaffing so much on contracts awarded unlawfully to cronies and the publican next door. All it would take is reneging on the world-beating Christmas Gift To Britain Deal, which wouldn't matter anyway as nobody remembers what was in it, and most Ministers never even knew. Or he could try and convince Arlene Foster that duplicating the Swindon Roundabout under the Isle of Man is not such a plexicated idea as it looks, and could even be quite a ringbeller as long as it totally squishes Irish Reunification by not having any subaquatic connection to Dublin. The alternate scenarios, once I reallocate the marginals to the runner-up, are again quite interesting. The Conservatives' best case scenario would still see them lose two dozen seats, which proves that duplicating the 2019 upset is not that easy, even with some help from Astra-Zeneca. Labour's best case scenario would again trigger lots of speculation in the few days before Not-So-Hardy Keir decides not to extend an olive branch to the SNP after all. This is far from the 'get the popcorn out' scenario I mentioned earlier, where Keir could gamble a minority government without the Pesky Jocks. On these numbers, Keir would have to show his baws, and listen to Clive Lewis instead of Ian Murray and Anas Sarwar. But of course we already know Keir will never go that far, don't we? Better to never be First Minister of England that to make any concession to the Nats. Aye, right.
For your complete information, I have now re-engineered all my models, not just the Scottish one, to accommodate regional crosstabs when the pollsters provide them. When we have them, they're based on the Senedd and Holyrood electoral regions in Wales and Scotland. London is split between Outer and Inner, which has some taste of Middlesex vs Labourgravia, by Redfield and Wilton, who happen to be the only ones conducting London-specific polls since the 2019 election. Redfield and Wilton also crosstab their polling data from the rest of England with the eight ONS regions, while other pollsters stick to the cruder North-Midlands-South divide. I won't claim this is perfect and faultless, but is does look like an interesting approach to different changes in voting patterns all over the UK since the last election. As I mentioned earlier, factoring in the regional crosstabs helps the Conservatives against the SNP in Scotland, which could help SNP HQ realise that not all landslides do happen. It's fairly neutral in Wales, though in can help Plaid Cymru in some cases. In London, the Middlesex Effect helps the Conservatives against Labour. There's a strong opposite pattern in the rest of England, with the Conservatives only getting entrenched deeper in the Leafy South, while Labour come back with a meaner vengeance in the North and part of the Midlands. You can see below how my tweaked model compares with the untweaked version, and also with what my competitors predict with the same current set of voting intentions data.
What we have this time is the unlikely and seldom seen alignment of the polls, when all projections more or less based on uniform swing deliver a very similar number of seats, give or take la handful. I know from several other simulations that it is way more likely that my three competitors will deliver results quite different from mine, and then the regional tweaks make the gap even bigger, sometimes up to 25ish seats mostly switching from the Conservatives to Labour. The differences in algorithms and underlying data might not make a massive difference with some aggregates of polls, but it is not always true when relying on individual polls. As I said earlier about my Scottish seat projections, larger samples tend to level out the outliers. All you need is one outlier in one direction and another one going the opposite way, and it all folds back neatly into regression to the mean. On top of that I still need to assess more precisely whether the differences in the algorithms (using a mix of proportional swing and uniform swing instead of just uniform swing) matter more that the tweaks to the underlying data. Might make for some interesting case studies with various combinations of polls, though I have a hunch this might be conclusive only if we have a murder of polls going in the same direction, rather that the often contradictory results we have had so often in the past.
If you're a lion with a degree of moral sensitivity, then you just stopped being a lion
(Doctor Greta Simpson, Waking The Dead: The Hardest Word, 2004)
© Steve Hackett, Phil Collins, 1977
Enjoy Gary O'Toole doing something Phil Collins never ever did
Singing and drumming at the same time, which is quite a feat
So, making mistakes can make you popular? Who knew?
(Clarissa Mullery, Silent Witness: Betrayal, 2019)
Polls don't only show better electoral prospects for the Conservatives as a whole. Boris Johnson's personal ratings have also improved over the last two months, as well as his standing in Preferred Prime Minister polling. It's like people are sending the message that they don't mind eating snoek and no longer buying books from France for the next three years, as long as they all get their jags and then can go to the game and sling tins at the linesman again. Which might be uncomfortably closer to the truth than you might want to believe, for Englanders at least. According to the aggregate of all polls, the First Minister of England's net approval rating has improved from -11% in November, -8% in December when we had some of the first hints of 'Fucking Brexit Getting Fucking Done', -7% in January after the not-so-convincing Bozo Deal with the EU, -1% in February after twelve weeks of vaccine rollout. Boris Johnson has also skyrocketed back to a convincing lead over Keir Starmer in the Preferred Prime Minister polls, which definitely look bad for Sly Keir. There's a double whammy in there, with both a number of Starmeristas switching to undecided, and a number of undecideds rallying around Boris. I won't even lecture you on the importance of cuddling undecideds, in England just as in Scotland, as I have made that point repeatedly already, haven't I?
The next question is whether there will be a Very Tory Coup against Boris Johnson, something along the lines of November 1990 déjà vu all over again. Though you have to understand most Tory MPs would think twice before becoming part of that. They all know that Boris is an evil ruthless vindictive bastard. If you staged a coup against him, you'd have to win or be relegated to the Ninth Circle, so standing orders would have to be "shoot to kill and take no prisoners". A number of Tory MPs from the 2019 intake never thought they would be elected, even in their wettest dreams. They think they owe it to Bozo, so they're ready to be his cannon fodder, and wouldn't even try and explain their support by anything more intellectual than "because it's Boris". Never mind this lot are mostly complete shitwanks like Ben Bradley, and pretty much the Trashcan Men to Johnson's Randall Flagg. Numbers matter here, not brains, and these and other Boris Cultists probably make up about a third of the Conservative Parliamentary Party. Then another third have already challenged Boris from the right, broadly on the basis of libertarian and traditional English values, including painting the closing of the pubs as an unacceptable breach of civil liberties. This lot, hardly more intellectual than the first one, would probably have supported Michael Gove before he got his wings and claws clipped, though they probably dispunge Sleekit in private, but now they know better than to fuck with Boris. And finally, the last third are more like traditional One-Nationers still believing that conservatism has a compassionate side, but are also business-friendly. Since their natural leader, who would be Theresa May, has pretty much burnt herself out already, they would be the logical fanbase for PR-savvy Rishi Sunak. Though a divided Tory party might have some more surprises in store, including the unlikely return of Jeremy-In-The-Box Hunt. Or the much more likely scenario where the Covidiots dump the burnt-out vermicious Gove and switch allegiance to Rishi "Let's Open Everything Wide" Sunak, which would be one fucking fuck of a problem for Boris.
Now there is also a lot of ambiguity and contradictions in Rishi's levels of approval in recent polls. The most recent favourability survey by YouGov has him on a +13% net rating, which makes him the only Cabinet member with a net positive and a shitload of leagues ahead of the others: Priti Patel on -37%, Dominic Raab on -19%, Matt Hancock on -23% and even Boris himself on -12%. But YouGov is definitely the least Boris-friendly of all pollsters, so don't read too much into that one poll. Rishi is also doing a lot better here than Keir Starmer, who is on a net -10%. Oddly this 23-point lead does not translate into a massive advantage in Prime Minister polling, though there are multiple caveats here. You can't compare these directly with the Johnson-Starmer chart above as only one pollster, Redfield and Wilton, has tested the Sunak-Starmer option. While six pollsters test the Johnson-Starmer option regularly with somewhat different results, so the signal from Redfield tends to be lost in all the noise. Redfield and Wilton are also the most consistently Johnson-friendly in this niche polling, so I wouldn't rule out Starmer actually doing better than Sunak if any other pollster asked the same question, which would indeed be quite counter-intuitive if you take the favourability ratings at face value. But then Rishi hasn't had the same wealth of opportunities as Boris to outfox Keir in Commons, has he? Perhaps Lindsay Hoyle should now make Chancellor's Questions a regular feature, like every other week instead of the highly predictable and borderline boring PMQs, and we would see what happens.
You probably remember I said two months ago that Boris wouldn't last till the end of this year. Which was based on polls predicting a hung Parliament where Labour would be tied with the Conservatives on the number of seats, and might even end up as the first party. Now the last batch of polls says we are very far from that, because hopscotchy Boris is basking in the spotlights of a world-beating Covid vaccination scheme, and people have pretty much immunised themselves against all his poppyrot. So Oor Feisty Bozo has got himself a stay of execution, and he'd better make the most of it while it lasts, like awarding the next English school meals contract to Wetherspoons while everyone is watching the squirrels who survived Jo Swinson. Then the litmus test might come sooner that Boris wants or expects, when he has to take a principled stand and say something intelligible in the incoming debate about 'Covid passports', and I definitely can't think of any answer that wouldn't leave him crodsquinkled and at odds with one half of his own party or the other. Then Boris might try and make the point moot by sending Matt Hancock on a fact-finding mission about Gwyneth Paltrow's vagina-scented cure for Long Covid. Not available in Northern Ireland. Terms and conditions apply. Or he might take a leaf from Theresa May's playbook and kick that can of worms down the road, until the vaccine rollout is completed and public opinion agrees that segregation is the right thing to do. Which might not be as far-fetched as we would like to think. Never underestimate the genetically-inherited propensity of Englanders for discrimination and erecting barriers where none exist. Sunlit uplands are great as long as you don't let everybody in, aren't they?
These our actors, as I foretold you, were all spirits and are melted into air, into thin air.
And, like the baseless fabric of this vision, the cloud-capp’d towers, the gorgeous palaces,
The solemn temples, the great globe itself, yea, all which it inherit, shall dissolve,
And, like this insubstantial pageant faded, leave not a rack behind.
We are such stuff as dreams are made on, and our little life is rounded with a sleep.
(Prospero, The Tempest, Act IV, Scene I)
© Tony Banks, Phil Collins, Peter Gabriel, Steve Hackett, Michael Rutherford, 1974
No comments:
Post a Comment