Recent Posts

Worth noting that previous 'Find out Now' MRPs during the current Parliament have also shown rather low LibDem gains - or in some cases, losses.

en.wikipedia.org/wiki/Opinion_polling_for_the_next_United_Kingdom_general_election#Seat_predictions

I suspect that their seat-by-seat findings do not include any mechanism for accounting for tactical voting.

In seats, the Channel 4/ Find Out Now MRP predicts:

LAB: 380 (+178)
CON: 174 (-191)
SNP: 51 (+3)
LDM: 21 (+10)
PLC: 4 (=)
GRN: 1 (=)

Labour Majority of 110.

It's interesting to try to work out which seats would actually change hands. Firstly note that Plaid and the Greens hold their current seats, while SNP make a net gain of 3. Beacuse of this, the most probable figures for Scotland would be Lab 4, LD 4, Con 0, with the SNP taking all the other 51. This means just 3 Lab gains from the SNP, who offset this by gaining all 6 of the Tories' current Scottish seats.

This in turn means that, to achieve those 178 extra seats, the Lab gains in England and Wales would need to go on down to number 210 of their target list (below). This is Reigate where a swing of 17.22% would be needed. And Runnymede and Weybridge would fall too at 17.14% - as would places such as Huntingdon, Woking Isle of Wight, Devon Central and both Southend seats. It's noteable that many of these seats require larger swings than the 15% swing implied by a Lab lead of 17%. Perhaps something of a 'winner's bonus' after all? My expectation had been for Labour to underperform soewhat in turning votes to seats, as their vote has become increasingly inefficiently spread in the past 3 General Elections.

www.electionpolling.co.uk/battleground/targets/labour

A few days later and R&W's latest South West England cross-break shows..... another 18 point Lab Lead ! Today's is Con 28, Lab 46.

Again, small samples, etc, but this is the 9th of the last 11 R&W polls to show a Lab lead in SW England (from 30 points behind at GE2019). Their  average over Aug and Sept (11 polls) is now a 7 point Labour lead for the South West region. 

Using UNS (so a 9.5% swing), the Tories would narrowly hold Monmouth, but lose all their other Welsh seats.

Wales has a very high concentraion of narrowly Tory-held marginals. There are 8 Welsh seats which Labour would take on a 3.5% swing, comparied to just 5 for the entire South oF England (London + East+ SE+SWEngland). 

www.electionpolling.co.uk/battleground/targets/labour

I've had another look at those R&W cross-breaks for South West England, and this really isn't as unusual (for them) as I had thought. Having checked all ten R&W polls since the start of August, 8 of them have shown Labour ahead, albeit often narrowly, and on samples of around 100-150.

The South-West average across all 10 R&W polls since the start of August is Con 33%, Lab 38%, which would be a 17-18 point swing on GE2019 ( where the South West voted Con 53, Lab 23).

For comparison, YouGov's last 5 polls show a 16-point C>L swing across the combined South of England regions (SE+SW+E).

Today's YouGov again shows a huge swing in the South of England sub sample (of 584 voters), with Con 36, Lab 37, compared to Con 55, Lab 23 at GE2109. And this is right in line with the other 4 polls they have done over the past month. Their 5-poll averages for the South are:

Con 36% (-19)
Lab 36% (+13)
LD 14% (0)
Green 8% (+6)
Ref 4%

Comparisons are to GE2019. And in an actual election, Labour may be able to squeeze more from the Green VI than the Tories can from RefUK.

The other trend that I'd highlight is that YouGov show a larger Con>Lab swing with ABC1 voters (13%) than C2DE (10%). This is not ideal for Labour as most of the English and Welsh marginals are working class towns. However, it probably spell 'double trouble' for the Tories in places such as Bournemouth.

Having kept an eye on the geographical cross-breaks, those Southern Labour gains do look consistent with the polls.

YouGov have had the Tories and Labour close to level in the South of England across their past 5 polls, whereas at GE2019 the Tories had a 32 point lead. So that would be a 16 point swing. And it's a similar position in the English Midlands, although the impled swing in the North of England is much in line with the GB average. 

Conversely, Labour appear to be making less progress in Wales and in London (the latter partly because they are already so dominant), and little to no progress in Scotland. 

To Add:
One other trend to bear in mind: YouGov's cross-breaks show Labour making more progress in the ABC1 half of their sample than with the C2DEs - although they lead in both. Because of this, Id' see the likes of Colchester, Hexham and the Bournemouth seats as more likely gains than South Thanet. 

I am a bit suspicious of Savanta ComRes's poll, as it included a number of questions regarding whether respondents supported the Government's Enengy Price Freeze, and other very leading questions. These are highlighted in the Daily Mail report which was linked here this afternoon.

If these questions followed after the Voting Intention question then that's no problem. However, if they came before it, this would be bound to boost the apparent support for the tories. ComRes have yet to publish their tables and questions, so there is no way of telling at present whether this is genuine bounce or just a push-poll.

Yes, you are.

Maybe you didn't look at the Wakefield result in detail, but this is where there appears to have been LD>Lab movement.

Despite the LDs recently polling above their GE2019 level, their vote share was reduced from 3.9% to 1.8%. They finished in 7th place.

And this is in a constituency which had been a known Con/Lab marginal in 2019, so the LD vote was already squeezed. 

Obviously, this did not contribute much to the Labour majority, as the absolute numbers are so low, but it follows the pattern of other by -elections, such as Sidcup and Old Bexley where the LD vote is strongly squeezed ( to less than half of GE2019 levels) where Labour are challenging the Tories. 

Travel to work is the most common form of rail usage, so of course much of the growth must indeed be work travel.

But if you look at the extent of that growth - a doubling of passenger numbers from 30 years ago, and an increase from 5% to 10% (in miles travelled) for rail as a mode of transport, it is clear that we have seen a change in public behaviour and transport choices. I guess that it is possible that WFH will disproportionately affect rail usage - as longer commutes are more commonly done by rail, but we await confirmation of this. I would expect that ongoing high fuel prices would affect car commuting just as much.

jib

The most recent report (from 3 months ago) put rail usage at 80% of pre-covid levels.

www.railjournal.com/news/british-passenger-rail-numbers-reach-80-of-pre-covid-levels/

It would only take a relatively small continuation of the past 30 years' growth in the industry to take us back to (or above) 2018-2019 levels of passenger use. And '80% of pre-covid levels' is already a higher level of usage then we had from around 1920 to 2010. 

Rail makes up about 10% of all passenger miles in the UK, whereas Cars make up around 75%. I'd say that is more than a 'minor' part of transport provision, as it is unclear how we could manage without it, especially for longer distance journeys and urban transport. 

It's interesting to note that the proportion of overall passenger transport done by rail is continuing to grow, whereas other forms of transport (such as bus and car) are declining.

assets.publishing.service.gov.uk/government/uploads/system/uploads/attachment_data/file/942425/rail-factsheet-2020.pdf

pjw1961

"I would put money on the Conservatives to be the largest party at the next GE, as a minimum. "

The bookies make the Tories marginally favourites for Most Seats at the next GE. They are are 4/5 on with most companies, with Labour Evens.


www.oddschecker.com/politics/british-politics/next-uk-general-election/most-seats

I don't think that there is any doubt that YouGov's headline figures are more volatile than those of other pollsters - and have been for many years. I did once find an old Anthony Wells article which explained this, but can't lay my hands on it at the moment. One important feature of their polls is that they do not ask a 'likelyhood to vote' question as several other pollsters do. But they also have the highest proportion of any pollster of respondents who state that they 'would not vote'. 

Incidentlly, Techne continue to be the opposite with figures that seem to be constantly stuck in the same narrow range.