New piece in
@prospect_uk
arguing that changes in electoral geography at the local elections suggest Labour is on course for a parliamentary majority, even if their poll lead narrows as expected:
Average of final polls so far: Con 43, Lab 34, LD 12
Uniform change projection: Con majority of 40
If 2017 style polling error, Con lead 4 points, majority 10
If 2015 style polling error, Con lead 17 points, majority 116
BBC Projected National Share of the vote (PNS) for the 2019 local elections:
Con 28
Lab 28
LD 19
Others 25
Only the second time all parties under 30%. Best Lib Dem share since 2010.
Off to work on the exit poll with a brilliant team. Many thanks to all the interviewers and others
@IpsosUK
collecting the data. Wish us luck!
To everyone registered, please vote and remember your ID.
Wondering if MRP projections of devastating Tory losses from proportional change are plausible?
@dibden_jake
and I have reviewed the historical precedents:
Without coalition partners Tories need a majority. That would require a 10-point swing between polls now and the election—bigger than any post-war Con gov in final 17 months. Bigger even than the 7.5-point swing that Margaret Thatcher enjoyed as a result of the Falklands War.
New post on variation between the polls and pollsters, with
@DanSnow96
tl;dr Differences in published vote-intention figures seem to be mostly due to different kinds of sample.
Con have lost 30% of their seats declared so far.
Gains, as approx. shares of Tory losses:
Lab 53%
LD 34%
Green 23%
Sums to more than 100% because Others down.
Based on 5891/8081 seats declared so far.
Benchmarks for today’s elections as local council elections
For all parties:
Net gains of seats and councils - good
Losses – bad
For voters (please vote & remember ID):
Turnout above 32.7% (from 2019) – good
Lower – bad
For all candidates:
Good on you for standing!
Many thanks to
@GuardianHeather
and
@ProfJaneGreen
for their great contributions to last night's retrospective on the local elections and look ahead to the general election
@TrinityOxford
. Recording with them and me here:
New article
@ElectoralStdies
with
@_johnkenny
@wouterpoortinga
, Gisela Böhm, and Linda Steg, on the links between climate change attitudes and voting in Europe:
Differences between party families aren't just about left-right and liberalism...
Averaging forecasts suggests Labour will win a majority of 200 with a 20-point lead over the Conservatives on votes.
See details of the Fourth Combined Forecast for the UK 2024 General Election, with
@_johnkenny
, Paul Furey, and Polina Ryzhuk at:
An explainer, by John Curtice
@whatukthinks
and myself, for the BBC Projected Nation Share (PNS) of the local election vote: what it is, why it is important, how it is calculated and issues this year ...
My blog, with John Curtice, on the meaning and method behind the Projected National Share of the vote (PNS) and projected House of Commons calculation for today's local elections:
A note, with John Curtice
@whatukthinks
, on the methodology for the BBC's local elections Projected National Share of the vote (PNS), and how it may differ from Rallings and Thrasher's National Equivalent Vote (NEV):
7-point lead would win Labour 315 seats, but SNP travails and improved electoral geography for Labour would likely raise that to c. 341, a majority of 32
The Conservatives have lost about a quarter (net) of the seats they were defending. The number of losses in each council is largely the product of how many seats they had to lose.
Averaging forecasts suggests Labour will win a majority of 194 with an 18-point lead over the Conservatives on votes.
See details of the Final Combined Forecast for the UK 2024 General Election, with
@_johnkenny
, Paul Furey, and Polina Ryzhuk at:
My take on the politics of the Brexit process and Tory party leadership.
Tory MPs face a choice between keeping May, who is likely to end up supporting a referendum, or probably a hard Brexiteer who ends up with a no-deal Brexit after failed renegotiation.
Delighted and proud to have helped with this
@UNDP
@UniofOxford
global survey of climate change attitudes covering 77 countries and 87% of the global population.
2023 Labour lead in the PNS is just short of those Labour achieved before Blair's 1997 general election success, but still better than all the other Labour leads in the PNS under Con government, including 2012 as the previous record since 2010:
New working paper on how governments in the UK tend to win elections, but lose if there is an economic crisis. Combination of that and post-crisis changes of PM explains 22 of 27 elections since 1922, and all since 1987:
My postmortem on the UK 2019 general election forecasts with
@_johnkenny
and
@RosieShorrocks
:
Polls did well, and uniform change from polls was a good guide to seats - better than betting markets and most other methods.
If the Conservatives continue losing at the same rate, and so lose 30% of their seats still to declare they will be down by -1025 when all the results are in.
Averaging forecasts suggests Labour will win a majority of 200 with a 20-point lead on votes (same as last week).
See details of the Fifth Combined Forecast for the UK 2024 General Election, with
@_johnkenny
, Paul Furey, and Polina Ryzhuk at:
@HzBrandenburg
@sarahobolt
@DeltapollUK
Thanks. You're right the conclusion to my
@DeltapollUK
blog is completely wrong due to a transcription error and writing too quickly. Sorry! May's Deal is the Condorcet winner. Blog post will be revised asap.
First Combined Forecast for UK 2024 General Election, with
@_johnkenny
, Paul Furey and Polina Ryzhuk:
Average of forecasts suggest Labour have an 86% chance of winning a majority, and are estimated to win with a lead of 20-points and majority of 188.
Labour have, since the PNS started in 1982, only managed to move from opposition into government once, in 1997. That was preceded by PNS leads of 10 points or more. Neither Neil Kinnock’s 8-point lead in 1990, nor Ed Miliband’s 7-point lead in 2012 proved to be enough.
@RosieShorrocks
@_johnkenny
Write up includes a plea for more Welsh, and especially Scottish, polls. The latest, so far as we know, are getting on to a month old and those nations have relatively more marginals, especially Scotland.
Honoured (and nervous) to be speaking at 5pm UK time at this:
Daily Press Briefing by the Spokesperson of the UN Secretary-General. Guests: Cassie Flynn (UNDP) and Stephen Fisher (University of Oxford) on the People's Climate Vote 2024
Final combined forecast with
@_johnkenny
and
@RosieShorrocks
Con: 341 seats, 8.9 point lead and 73% chance of a majority.
Write up includes full table of seats forecasts and final polls with UNS projections:
Both the confidence vote and promising to step down before 2022 make it more likely that Theresa May will pivot to a referendum: she has more freedom and less future career to lose.
AKA me digging myself further into a hole:
A governing party falling further where it started stronger is an extremely important development if replicated in a general election. It helps explain the scale of the collapse of the Liberal Democrats in 2015, Labour in 2010, and the Conservatives in 1997.
Only saw this great piece from
@BNHWalker
after I wrote mine. It makes the same main point but with different presentation of evidence, and a different take on the variation.
My
@DeltapollUK
blog conclusion is completely wrong due to a transcription error in graphic 3 and writing too quickly. Sorry! May's Deal was the Condorcet winner as you can work out from graphic 2. Blog post will be revised asap.
New blog with
@caprosser
on his Local Elections based forecasting model. The 2018 and 2019 local elections suggested a borderline Tory majority.
Con 329 seats. Just 10 seats lower than
@YouGov
MRP, and same predicted Lab seats (231).
Boris Johnson won 125 of the seats that Blair (notionally, given boundary changes) won in 2005 when Labour last had a majority. Those seats they won in common represent just over a third of each of their tallies.
#GE2019
.
@rafaelbehr
and
@AnushkaAsthana
, your helpful writing, on the political pendulum and attitudes to Boris Johnson respectively, are linked in my piece with
@MattLebo4
here:
My analysis in
@prospect_uk
is consistent with the excellent piece by John Curtice
@whatukthinks
(). Mine brings in additional factors: one bad for Labour (likely lead narrowing; but new electoral geography more than compensates
Averaging forecasts suggests Labour will win a majority of 192, and a 21-point lead over the Conservatives on votes.
See details of the Third Combined Forecast for the UK 2024 General Election, with
@_johnkenny
, Paul Furey, and Polina Ryzhuk at:
Incumbency effects and boundary changes are not discussed in the piece, but roughly factored into my electoral geography adjustment. Main point is I think the greater drop in the Con vote in Con seats will be more important for seat outcomes.
ICYMI my piece with Andreas Murr.
Since both Con and Lab MPs have voted on leaders, 12/15 elections won by the party leader with the biggest winning margin among own MPs.
Starmer margin on Lab MP noms (25.9%) larger than Johnson's first round 22.7%.
Very grateful for this award for the Peoples' Climate Vote. We are especially glad of the opportunity to work with
@cassie_flynn
@EriYamasumi
and others at
@undp
and
@BrowningEnviro
and elsewhere.
Please check out the report at:
Con vote dropped more where they started strongest, especially where they were defending a seat from 2019. In such places the swing to Labour since 2019 was five points greater than elsewhere. The same pattern is true on a smaller scale for swings from 2021 and from 2022.
My corrected and extended
@DeltapollUK
blog:
- No Condorcet cycle. Sorry about the excitement-generating error.
- May’s Deal is the Condorcet winner, but any of the three options (Deal, No Deal, or Remain) could emerge ahead.
TLDR - Not only is proportional change plausible, things might actually be worse (as in 1929, 1906, and Canada 1993). However, 1997, Labour in 1983, and LibDems in 2015 surprisingly may provide some hope for the Conservatives.