Recent Posts

I think there is a little bit of comfort for the Tories in that their vote share was "only" down by 30% of the 36% they got in 2019. Polls currently show the Conservaive VI down by 40%+ on GE2019. Surprisingly , the Labour vote rose by almost as much as the Tory vote fell. 

Both RefUK and the Greens will be disappointed to have lost their deposits, and made little advance (compared to BXP) on 2019. It suggests that those pollsters who don't prompt for them, and therefore get lower figures, are closer to the mark than the likes of YouGov and PP who include them in their prompt. 

The gaps (averaging minus 3) between the Con losses and Lab gains in the 'regions' are consistent with the overall polling figures. Compared to GE2019, the LDs have lost about 3 points in GB polling, the Greens have gained 2-3 points and RefUK are 4 points higher than the BXP achieved 3 years ago. The regional figures reflect this, but in varying ways. For example, it looks like RefUK have picked up more in the Midlands. In other regions, LibDem losses and Ref + Green gains appear to have been roughly equal. There is bound to be a certain amount of 'noise' and approximation in my figures, as they are based on 6 x small samples of about 100.

In London, it looks to me like its the Greens who have gained rather than the LDs. R&W's most figures show the LibDems down by just a point in London at around 14%, but in the past 6 YouGovs the LDs average 8 points, down from 15 in GE2019. The most recent actual London poll, from Survation 4 months ago, was better than this for the LDs putting them on 13% (but still 2 points down on 2019). 

The polling effects of Liz Truss's 7-week spell as PM were really not good for the LibDems: looking at the YouGov cross breaks, she appears to have weakened LibDem vote retention, and this has since then only partly recovered.

In the 4 YouGovs immediatly before Truss became PM (so 5 Aug - 1 Sept) the YG LibDem2019 cross-breaks averaged:
Con 6%
Lab 30%
LD 53%

The 4 YouGovs from the start of the crisis she caused to her resignation (29 Sept-21 Oct) YG 2019LibDems averaged:
Con 1%
Lab 51%
LD 40%

And now the 4 most recent YouGov 2019LibDem cross-breaks average:
Con 7%
Lab 39%
LD 48%

 These latest figures show the Tory VI for the entire 18-64 age group as just 18% - so less than half the 37% that YouGov believe they held at GE2019.

And the 50-64 age cohort is now slightly closer to the 18-24s than it is to the 65+. 

But there are implications for Labour in these figures, too. While they are picking up far more Leave voters than they did in 2016, this has made almost no impression on the proportion of the Labour VI who believe that leaving was the 'Right' decision. YouGov have not asked this question in their latest poll, but typically the percentage of 'Right' in their current VI stands at around 13-15% compared to around 11-12% of their 2019 voters. 

TO ADD:

From YouGov today "Which Leave voters have turned against Brexit?
Adam McDonnell  -February 03, 2023, 10:55 AM GMT+0

Those who have changed their mind are notably younger than those who still think it was right to vote to leave the EU "

yougov.co.uk/topics/politics/articles-reports/2023/02/03/which-leave-voters-have-turned-against-brexit


If this sounds familiar to any on UKPR2, this is what I posted here on 15 October last year:

" YouGov tracker - "In hindsight, do you think Britain was right or wrong to vote to leave the European Union?" They have asked this three times in Sept and Oct, with results of 34/53, 35/52 and 34/52 for 'Wrong'.

This question has consistently seen much less movement than Voting Intention polls, but has overall been slowly drifting towards 'wrong'. This is despite around 75% of those who voted Leave still thinking it was the right decision. But as I've noted before, the movement is there. And views on the issue have become ever more age-divided.

Over 65s still think the decision was 'Right' by a 64/36 margin which matches exactly the way the same age cohort voted in 2016 (excl DKs). Those aged 18-64 now divide 70/30 for 'Wrong' , so an 18-point swing compared to the referendum itself where those aged under 65 voted narrowly to Remain by 52/48."
Last Edit: Oct 15, 2022 at 5:21pm by James E

2019 Tory Voters and the Remain/Leave axis

Many thanks for the kind comments on my R&W "Blue Wall" polling, and the implications for tactical voting. I touched on the matter of the differential movements within the 2019 Tory vote, but thought it was worth giving some more detail, as the assumption that the Conservatives' Remain-supporting seats are the most vulnerable to lost support seems to have become ingrained on the basis of the 2017 and 2019 General Elections. My main source for this is Deltapoll, who provide separate 'Con Remain' and 'Con Leave' cross-breaks, but there is support for this from other pollsters, such as YouGov.

Taking the past 5 Deltapolls in Dec and Jan, their Con2019 cross-breaks average:
Con 70%
Lab 16%
LD 3%
Ref 6%

But looking at the Brexit vote cross-breaks its:

Con-Leave
Con 68%
Lab 16%
LD 3%
Ref 9%

Con-Remain
Con 76%
Lab 17%
LD 3%
Ref 1%

So the Con>Lab movement is much the same for both groups, but the Tory losses are 8 points higher with their Leave voters. The main difference is that more Con2019 Leavers are defecting to Reform UK. I should add that Deltapoll's figures for RefUK are relatively low (4% typically) as they do not prompt for them, unlike YouGov and some others. A prompt for RefUK would only be likely to emphasise this trend, as their support is overwhelmingly (90%+) from Leavers.

And there's a similar picture from YouGov's overall Remain and Leave cross-breaks. Across their five most recent polls, YouGov have the Tories taking an average of 44% of 2016 Leave voters (down from 74% in 2019) and 13-14% of Remain voters (down from 29% in 2019). This means that their share of Remain voters has declined by 29%, whereas their share of the Leave vote is down by 41%. I should add that their overall vote-loss is around 43% of their 2019 share, as they are faring worse still with those who did not vote in 2016 (especially those aged under 25).

So it looks to me like it's time to abandon the idea that the Tories' Leave voting seats are relatively 'safer' for them. The only advantage that the Tories have in Leave voting seats is that most of these had a low LibDem vote in 2019. The polls show Labour taking 25-45% of  LD2019, and it is this which explains why the swing is larger in the "Blue Wall" than the "Red Wall", rather than the 2019 Tory voters. 

Thanks for this, and to moby for posting the detailed seats projection.

"Blue Wall" is such a silly name for these seats, most of which are somewhat vulnerable tory-held seats, but the findings are nonetheless very interesting indeed, particularly in relation to tactical voting. For the LibDems, the 42 seats polled include nearly all of their prime target seats, but anything from 3 to 12 of these are seats where Labour might also stand a chance.  These seats are all Con-held, 26 with the LibDems in 2nd place and 16 with Labour were second in 2019. All have at least a 42.5 Remain vote, and taken overall they were generally more Remain than Leave in 2016.

Today's figures which are not much different from the previous 5 polls since Oct are:

Labour 42% (+21)
Conservative 32% (-18)
LibDem 19% (-8)
RefUk 4% (+4)
Green 4% (+2)

(Changes from GE2019)

There are a some significant details here: firstly, the Con>Lab swing at 19.5% is 3 points higher than implied by R&W's GB-wide poll this week, and their recent average of a 21% Lab lead. This is happening despite the Conservatives actually holding a slightly higher share of their 2019 vote in these 42 seats than other polling suggests (64% compared to 55-60%). This is consistent with the findings of both YouGov and Deltapoll, both of which now show the Tories retaining more of their Remain voters from GE2019 and losing more of their GE2019 Leave voters.  This trend seems to have gone unnoticed. See in particular Deltapoll's Con Remain and Con Leave subsamples. 

On the other hand, in these seats Labour are actually gaining a larger share here than in most GB polls: 19-21 points in the Blue Wall compared to the 15-17 points in R&W's GB polls. As R&W's detailed tables for 2019 Con, Lab and LD voters are really very similar for their "Blue Wall" as for GB as a whole, this seems to simply be a product of the different mix of voters. Labour are not proportionately taking any more Con or LD votes here than elsewhere:  the 21 point rise in the Labour VI is due to them having more Tory2019 (@50%) and more LD2019 (@27%) voters to target. 

TACTICAL VOTING AND THE 3- WAY MARGINALS

Seat projections for these 42 southern seats give Labour 24 gains (8 of them from 3rd place), LDs 6 gains, and Conservatives 12 holds. I think these figures are derived from applying UNS to the 2019 results, and if so this produces some rather unlikely results. For example, Wokingham becomes:

Con 32% (-18)
Lab 31% (+21)
LibDem 30% (-8)

John Redwood would no doubt be laughing all the way back to the opposition benches if this happened. But in reality this is of course highly unlikely. It's a seat where the LDs finished 27 points ahead of Labour in 2019, and would become the better tactical choice in the event that the Lab vote dropped by just a couple of points (or the LD vote rose). But R&W's figures show the LDs losing 25-30% of their 2019 vote to Labour (even in this sample with many LD targets) and other pollsters such as YouGov and Opinium put this even higher (typically 40%). And this needs to be taken into account. 

On the basis of these figures, and the previous 5 polls in this sample, I would suggest that the LibDems are only the better tactical choice in seats where they were 20 percentage points or more ahead of Labour in 2019. While this may seem a high bar, it is of course less than the 29 point net LD>Lab shift implied by these figures or the average of around 27 points in the previous R&W 'Blue Wall' polls. So the 20-point rule would factor in combined Labour losses and LD gains of 7 points when an actual election comes - say Lab down 5 points to 42% and LDs up 2 to 10%. 

moby

Tom Newton Dunn is wrong to claim that "Most movement is in left-leaning areas".

The top 20 seats with most movement since 2016 include three Tory-held seats, Walsall North, Soton Itchin, and Thurrock. There are also a couple of Labour-held marginals, such as Wolverhampton SE or Dagenham and Rainham.

I'd characterise the seats with most movement as Urban and very working-class, but generally not as left-wing places. The findings confirm the long-standing pattern from YouGov's 'Hindsight' tracker where the movement is higher with C2DE (who were 35/65 leave in 2016) than with ABC1s (who were 55/45 Remain).


Incidentaly, there is plenty of "bregret" even in those constituencies which they describe as 'least bregretful'. For example, Boston and Skegness, (74/26 Leave) now divides as 37% who 'Agree' that Britain was wrong to Leave the EU, and 41% who 'disagree'. By my calculations, that's actually a larger movement (22% swing on 2016, excl DKs) than the GB figures as a whole, which show a 66/34 majority (or 18% swing) for 'Agree that Britain was wrong to Leave the EU' (again excl DKs). 

The present Conservative MP for Henley might not be easily amenable to Johnson's designs...

en.wikipedia.org/wiki/John_Howell_(politician)

"Howell first became MP for Henley at the 2008 Henley by-election, when the newly elected Mayor of London Boris Johnson resigned as MP.....
Howell was opposed to Brexit prior to the 2016 referendum. He voted in favour of the withdrawal agreement of the UK from the EU, and advocated against a second referendum to ensure the referendum result was honoured.


At the Conservative Party Conference in 2017, he was quoted by the Henley Standard as saying: "My message to Boris is to keep his bloody mouth shut!" regarding Johnson's demand that the post-Brexit transition should last "not a second longer" than two years. A year later, at the party conference in 2018, when Theresa May was reportedly being undermined by Johnson, The Guardian reported Howell as saying: "As far as I'm concerned Boris can just fuck off."  "

Omnisis
@omnisis

🔴 Lab 50% (NC)
🔵 Con 26% (+2)
🟠 LibDem 7% (-1)
🟡 SNP 3% (-1)
⚪️ Ref 7% (+2)
🟢 Grn 6% (+1)

Disapproval of Prime Minister Rishi Sunak’s performance has hit its highest (lowest?) point since he became Conservative Party leader, with nearly half of voters now saying he’s not performing…

👍 Approve: 27% (+1)
👎 Disapprove: 47% (+6)
🤷‍♂️ Don’t know: 26% (-6)


Despite Chancellor Jeremy Hunt claiming that Brexit ‘can be an incredible success’, voters don’t seem to share his optimism in our latest Brexit sentiment tracker:

* All *

🇬🇧 Stay Out: 29% (-2)
🇪🇺 Re-join: 49% (+2)

* Exc DKs*
🇬🇧 Stay Out: 37% (-3)
🇪🇺 Re-join: 63% (+3)

The LibDems have persistently fared better in Local Elections and Local By-Elections for the past 30 years. 

The norm is that they receive 7%-10% more of the vote (compared to Westminster polling) and this needs to be factored into our 'expectations'. 

As posted previously ( 9 Dec):
"Our former host on UKPR1 wrote on this in 2009 : "Can local by elections predict general elections?"

Anthony Wells puts the level of LibDem overperformance and Labour underperformance as around 6% to 10%, and I suspect that about the same applies now.



Examples are:


"Sum of local by-elections Jan-May 2005: CON 33%, LAB 26%, LDEM 31%

Actual general election result 2005: CON 33%, LAB 36%, LDEM 23%


Sum of local by-elections Jan-Jun 2001: CON 32%, LAB 30%, LDEM 26%

Actual general election result 2001: CON 33%, LAB 42%, LDEM 18%


Sum of local by-elections Jan-May 1997: CON 28%, LAB 37%, LDEM 30%

Actual general election result 1997: CON 31%, LAB 44%, LDEM 17%"



So Labour underperformed by 7, 12 and 10 points in pre-GE local by elections, and the LDs overperformed by 13, 8 and 8 points. The Conservatives' share is quite close to the subsequent GE results, though. Incidentally, the LDs overperform (and Labour tend to underperform) in all Local Elections: another UKPR article from 2012 puts the LibDem overperformance in scheduled Local Elections at 7 points on average."

A little point about the widely-maligned People Polling: even with their figures of Lab 50%, Con 21% reported yesterday, their figures since the start of November are within 1.5 points of YouGov - at least as measured by the Labour lead.

Its an average 23.25% Lab lead with YouGov (in 12 polls) compared to 24.72% average in 11 PP polls.


And looking more widely across the rather stable polls we've had in the last 3 months, PP are only 3.7 points higher than the overall polling average of a 21% Labour lead. Comparing all pollsters who have published figures since 1 Nov gives us these average Labour leads:

People Polling (11 polls) 24.7%
Ipsos Mori ( 2) 23.5%
YouGov  (12 ) 23.2%
Omnesis ( 11) 23.1%
Average of all 85 polls: 21.0%
R& W (13) 20.6%
Techne (11) 20.4%
Deltapoll (9) 18% - but see below
Sevanta (6 ) 17.2%
Opinium (5)  16%
Kantar (2) 16%

Also, single polls from Focaldata (25%), Survation (23%) BMG (18%).
Deltapoll may have changed their methodology in Nov or Dec, as their more recent polls average 14% compared to 21% in November.

@isa

"RFM appear to be the main beneficiary of the decline in CON VI, which, I suspect, is likely to be almost as damaging to CON prospects in the battleground seats as direct switching to LAB.

I would be grateful to read the thoughts of jimjam , James E , pjw1961 etc. on how they feel this apparent switch from CON to RFM might play out in such seats if replicated at the next UKGE. Many thanks in anticipation."

It's worth remembering that the Brexit Party only fielded candidates in Labour-held seats in 2019, taking an average of 4.3% of the vote in the 275 seats they contested (but 2.0% overall in GB). Much of the Con/Lab battleground for the next election is these same seats, where polling such as R&W's "red-wall" suggestes that RefUK are only gaining 2-3 points at present. My expectation is that the Tories will be able squeeze a point or two from this, but probably more if Reform UK now decide not to field candidates against those Conservtives they approve of.

However, in those seats which were already Tory-held from GE2017, Reform UK can only cause damage to the Tory vote share if they stand this time. I could see this making quite a difference in the longer range Con/Lab battleground: places such as Loughborough, Worcester, Stevenage and the Northampton seats - plus a lot of longer range seats in the South of England.

One further point about Reform UK: they have had two disappointing results in by-elections in late 2022, barely improving on the Brexit Party's 2019 vote share even though the Conservatives were really not in contention (both were easy Lab holds). I would expect them to fall away in the same way in a General Election, particularly in closely-contested seats. 

As I've touched on those differing regional swings, these are the current average Con>Lab swings in Redfield &Wilton's regional cross breaks over their most recent 14 GB polls going back to 30th Oct, and with an average (and fairly consistent) Con>Lab swing of 17%

East England 24%
South West 22%
West Midlands 21%
East Midlands 21%
South England 19%

Yorkshire & H 15%
North West 14%
North East 13%
Scotland 9%
Wales 9%
London 6%

Re R&W's so-called 'Blue Wall'.

Compared to GE2019 it's:

Lab 40% (+19)
Con 30% (-20)
LD 21% (-6)
Ref 6% (+6)

So a somewhat higher swing at 19-20%, compared to 16-17% implied by most GE polls (including R&W's most recent).

This supports the view that the swing in higher in the South - as evidenced by R&W, YouGov and Opinium. These selected Tory-held 42 constituencies, which include 26 with the LDs in 2nd place and 16 with Lab in 2nd, show a slightly better LD performance than R&W's GB polls. However, the difference is fairly marginal. The cross-breaks for Oct-Jan (5 polls) show slightly less LD> Lab movement, and just a little more LD retention. But the LDs are only holding 59% of their 2019 voters compared to 54% for the Conservatives and 87% Labour.