Thread by @MattSingh_

By-election night thread! Polls have just closed at the #Wakefieldbyelection and the #TivertonandHonitonByElection. Analysis follows

As usual THERE ARE NO EXIT POLLS

Load up on coffee as it's going to be a late one. BBC suggested earlier that declarations could be around 4am...

In terms of likely outcomes I'll have much more on this later, but little doubt that #Wakefield (for which polls have been published) is going back to Labour, whereas #TivertonandHoniton (where none have) is murkier, but Lib Dems I've heard from seem much happier than Tories...

I’ll be talking a lot about swing, and it’s notable that the national picture is somewhat murky too (part random variation aka margin of error, part systematic differences between pollsters aka house effects) but on average GB polls are showing about a 9% swing from CON to LAB

Likewise historical comparisons – though note that win/lose stats you'll hear are highly sensitive to which seats happen to be up for election and when. For example:

https://twitter.com/GoodwinMJ/status/1540015801968869378

So what is par for the course? The rule in by-elections is that governing parties lose vote share – the question is how much – on average it's been 13.5 points since 1983. Swings tend to be smaller when the challenger is the main opposition party, and bigger when it's vs the LDs

BBC now saying Wakefield might declare a bit earlier. We can hope...
bbc.co.uk/news/live/uk-p…

The first indications we'll get are turnout. You'll hear a lot of rubbish about this, so repeat after me: THE TYPE OF SEAT MATTERS. Wakefield is a lowish turnout seat (64% in 2019), Tiverton is a higher turnout seat (72% in 2019). A good rule of thumb is 2/3rds of GE turnout

So a par turnout (a benchmark, not a prediction) would be low 40s in Wakefield and high 40s in Tiverton and Honiton. Who benefits from higher or lower turnout isn't entirely clear. Conservative voters are clearly grumpier, so low turnout could mean it's them staying home. However

Higher turnout could mean they're coming out to vote for someone else, which would be even worse for the Tories. So take your pick...

Looking specifically at #Wakefield, Labour needs a swing of less than 4% to retake the West Yorkshire seat. This is a long-time Labour, Leave-voting, coalfield seat, and part of @JamesKanag's original Red Wall analysis (see link)
journals.sagepub.com/doi/10.1177/20…

And we have polls! In fact this is the first by-election since Rochester & Strood in 2014 where we’ve had them from more than one pollster. (Some incorrectly claim polls from both Survation and Focaldata at Hartlepool 2021, but the latter were an MRP projection, not a local poll)

And despite using completely different methodologies, they showed very similar numbers. By-elections are a pain in the neck to poll, and the polling was a few weeks ago, but if the indicated swing of 14 or 15% from Con to Lab were to materialise...

...it would be the biggest by-election swing in that direction since Wirral South in 1997. This should not be overinterpreted – it is still not of the same order as some of the swings Labour was getting at the equivalent point in that parliament – but then that is a very high bar

Labour's 4 biggest postwar by-election swings were all during that period. So worth keeping these sorts of figures in mind for comparison:

https://twitter.com/johnrentoul/status/1539882222995603457

Candidates not in the running may not affect the outcome, but can very easily impact on the margin of victory...

https://twitter.com/DamianSurvation/status/1540084744037306369

In any case, it seems that the long period of Labour weakness in by-elections, from the rise of UKIP until the end of the vaccine bounce, has come to an end. This is what an opposition party should be doing in by-elections, and probably even better…

https://twitter.com/lewis_baston/status/1539929319916584961?s=20&t=sfr-aK28q_FbNnFVd0v6mQ

Not sure what this translates to in reality. Worse than the polls? Or were they expecting to do better than the polls?

https://twitter.com/LouisDegenhardt/status/1540093865100775427

See also in Tiverton and Honiton. Tories sounding very gloomy about that one too:

https://twitter.com/charliemansell/status/1540108574667509760

In #TivertonandHoniton, unlike Wakefield, no polling has been published. I'm aware of one private poll, but so on in the campaign that even if I could share the results they wouldn’t be very predictive – the swing in this type of contest usually grows during the campaign

In any case, the Lib Dem briefings that it was neck-and-neck sound a lot like they were expectation management. Which is understandable because getting a 20% swing and losing really would not be a poor result...

But the fact is, the Lib Dems got a bigger swing in Chesham and Amersham (when the Tories were doing very well nationally) and a MUCH bigger swing in N Shrops. So an LDs win on a big swing is to be expected, the question is how big. Here's the record book:

https://twitter.com/JohnRentoul/status/1539882327995826181?s=20&t=7gL8RKjjD-yNK3-qZUqCqg

Coming up...

https://twitter.com/KeejayOV3/status/1540109066525097986

In the #TivertonandHonitonByElection Lib Dems aren't just coming from third in 2019 and 2017, but in 2015 they actually came FOURTH behind Labour and UKIP

Like I said, this would be better for Labour than they've been managing recently. But do bear in mind that Ed Miliband got SIX by-elections swings over 10%...

https://twitter.com/iainjwatson/status/1540111823688581120

Wakefield turnout 39.1% – pretty close to what you'd expect in a by-election in this type of seat

https://twitter.com/MattSingh_/status/1540092442824916993

Tiverton 52%... So a bit above the benchmark, but unremarkable. Turnout is not the story tonight...

https://twitter.com/MattSingh_/status/1540092442824916993

Watch how Labour does in Tiverton and how LDs do in Wakefield. By-elections, because of the intensity of the campaigning, tend to see a lot of tactical voting, but if they really crash, that implies tactical voting on steroids and worth keeping and eye on

https://twitter.com/marklobel/status/1540128623448973314

This doesn't sound like it's going to a recount

https://twitter.com/EdwardJDavey/status/1540143256704585729

Maybe Ed Davey should take the big hammer to the "polling" numbers the Lib Dems were briefing to the Guardian

#Wakefieldbyelection result sounds like about a 13% swing

12.7% swing to Labour – their biggest since Middlesborough in 2012

29.9% swing to Lib Dems at the #TivertonHonitonByElection – not quite North Shropshire, but still a huge swing

Labour lost its deposit in Tiverton and won Wakefield on a decent swing. Lib Dems lost their deposit in Wakefield and won on a huge swing in Tiverton. This is industrial scale tactical voting, and it's a big deal...

https://twitter.com/MattSingh_/status/1540137002414358528

Not to anything like the same extent. But isn't need to. Even a relatively modest amount of Lib-Lab tactical switching would flip a lot of seats in a GE

https://twitter.com/Bharat_LearnAca/status/1540172386145804289

IMO this is the most significant takeaway. 12.7% swing to Labour in Wakefield is the sort of by-election result they were getting 2010-14, suggesting it's got its act together in by-elections (and in general). 29.9% to LDs in Tiverton and Honiton is one of the biggest ever.

My (highly caffeinated) thoughts for @CapX on what to make of the by-election results, whether or not comparisons to the early 2010s make sense, and why tactical voting is such a big deal
capx.co/hammered-in-ho…