Pollster Find Out Now and election experts Electoral Calculus have run an additional poll on the upcoming district council elections on 5 May 2022.
The poll asked residents of the 201 district and unitary councils up for election, whether and how they intended to vote on Thursday. Areas up for election include London, Scotland, Wales, parts of the north of England, and elsewhere.
The new poll of 2,148 GB adults was conducted between 27-28 April. This poll was combined with the earlier poll in April to produce a Dynamic MRP prediction of the local elections.
The number of councils in England and Wales predicted to be controlled by each party is shown in this table
|Party in Control|
|NOC / Coalition||37||26||−11|
Note that there are four new unitary authorities this year: Cumberland, North Yorkshire, Somerset, and Westmorland and Furness.
The prediction is that Labour will gain around 16 councils and are not expected to lose any.
The Conservatives might lose half-a-dozen councils, though they might also gain a couple (including the new North Yorkshire unitary authority). Losses could be bad in London, where Barnet and Wandsworth were already expected to go to Labour, and now the City of Westminster could also be gained by Labour.
The Liberal Democrats should pick up the new Somerset unitary authority.
This result would not be as good for the Conservatives as it looked earlier in April. The Partygate fines seem to have decreased the likelihood of Conservative voters turning out for the local elections. Indeed, 51% of people who voted Conservative at the last general election are likely abstain at the local election, compared with only 36% of people who voted Labour.
In terms of the number of wards won in England and Wales, the prediction is:
(Note that implied results have been inferred for comparison purposes for the wards in the new unitary authorities and the districts with new ward boundaries. These implied results have been improved since the previous analysis, which has resulted in a different baseline.)
Labour are predicted to win over 3,500 seats, while the Conservatives are predicted to win fewer than 1,000 council seats in England and Wales. That is quite bad for the Conservatives and worse than it appeared earlier in April before Partygate re-erupted.
Conservative losses and gains:
Liberal Democrat gains:
Scottish local elections are run using a proportional representation system which makes it more difficult to estimate the composition of councils in Scotland.
Martin Baxter, CEO of Electoral Calculus: "The renewed Partygate focus has made a poor situation for the Conservatives even worse by persuading even more Conservative supporters not to turn out at the local elections. The results could now be bad for Boris Johnson, especially if the Conservatives lose many hundreds of council seats and key flagship councils like Wandsworth or Westminster." And "These results would also be good for Keir Starmer, who needs to get his first real electoral gains since becoming Labour leader. The opposition doesn’t normally win a general election without winning other elections first, and Labour have to show that they can do that."
Chris Holbrook, CEO of Find Out Now: "Our poll suggests it will be a rough night for the Conservatives on Thursday, in particular the loss of the City of Westminster will be symbolically difficult for Boris Johnson. It might have other Tories jostling for position to replace him, even more than some in the media suggest they are already."
He added: "Local election results don’t always translate to general elections, of course. The Tories will try to argue that this is a blip driven by Partygate. Labour will suggest it is a true shift in opinion due to scandals and the cost-of-living crisis."
Our prediction is that Labour will win over three times as many council seats as the Conservatives. The last time that happened was 1996 when Tony Blair was leader of the opposition and John Major was Conservative PM.
In terms of equivalent national vote share (which is the usual measure of comparison for local elections), our prediction for 2022 is Con 24%, Lab 39%, LibDem 15%. That gives a Labour lead of 15%. The last time Labour had a lead that big in local elections was also in the mid-1990s when Blair was challenging Major. Blair went on to a landslide victory in 1997.
The last time the Conservatives had a lead over Labour of the same size (15%) in local elections was 2009 when opposition leader David Cameron was challenging Labour PM Gordon Brown. Cameron went on to be PM the next year.
But, as a cautionary note, the Conservatives under Michael Howard also had a big lead (+11%) over Labour at local elections in 2004 but went on to lose the next general election in 2005. Also William Hague had a lead of 9% in the local elections of 2001, but lost the general election on the same day.
This table shows the predicted winning party for each council, or 'NOC' if no party will have overall control of the council on its own.
|Barking and Dagenham||LAB||LAB|
|Basingstoke and Deane||CON||CON|
|Blackburn with Darwen||LAB||LAB|
|City of London||OTH||OTH|
|City of Westminster||CON||LAB|
|Hammersmith And Fulham||LAB||LAB|
|Isle of Anglesey||NOC||OTH|
|Kensington and Chelsea||CON||CON|
|Kingston upon Hull||LAB||LAB|
|Kingston upon Thames||LIB||LIB|
|Neath Port Talbot||LAB||LAB|
|Newcastle upon Tyne||LAB||LAB|
|North East Lincolnshire||CON||CON|
|Nuneaton and Bedworth||CON||CON|
|Reigate and Banstead||CON||CON|
|Rhondda Cynon Taf||LAB||LAB|
|Richmond Upon Thames||LIB||LIB|
|Vale of Glamorgan||NOC||CON|
|Westmorland and Furness||(NEW)||LIB|
Modern polling analysis often uses statistical regression techniques to get more accurate and geographically detailed results. Also called MRP (multi-level regression and post stratification) they have been used successfully by Electoral Calculus and other pollsters to predict general elections, local elections and the 2019 European elections.
These techniques work by spotting patterns between people's demographic characteristics and their likelihood to vote for various parties.
Both because of their larger sample size, and the more advanced statistical analysis, regression polls are often perceived as one of the more accurate ways of measuring public opinion both nationally and for specific geographic areas.
A variation of the approach, called Dynamic MRP, allows polls taken at different times to be combined to give a more accurate result which captures changes in opinion over time. Dynamic MRP was used to analyse this poll and the earlier poll together.
Clients who have benefited from insights from regression polling include well-known UK political organisations, non-departmental public bodies and campaign groups.
Ward-by-ward predictions for the local elections are available to purchase from Electoral Calculus. Please contact us for details.
Find Out Now conducted two linked polls:
Find Out Now and Electoral Calculus are both members of the British Polling Council and abide by its rules.
Data tables are available for the first wave, and the second wave.