What do we use cookies for?
We use cookies and similar technologies to recognise your repeat visits and preferences, as well as to measure the effectiveness of campaigns and analyze traffic. To learn more about cookies, view our Cookie Policy. By clicking "Accept" or using our site, you consent to the use of cookies unless you have disabled them.
More in Common have released their first MRP of the new Parliament which reveals a dramatic shift in Britain’s electoral dynamics. The model estimates that an election today would produce a highly fragmented and unstable Parliament with 5 parties holding over 30 seats. While Labour would still emerge on top, they would have barely a third of the total number of seats, and a lead of just six seats over the Conservatives, while Reform UK emerges as the third largest party on 72 seats.
The model is based on voting intention data collected between 31 October 2024 and 16 December 2024 from 11,024 adults in Great Britain.
“With potentially four and a half years to go, this model is not a prediction of what would happen at the next General Election. Instead, however, it confirms the fragmentation of British politics that we saw in July’s election has only accelerated in Labour’s six months in office. The First Past the Post system is struggling to deal with that degree of fragmentation, which is why our model shows so many seats on a three-way knife edge, and many being won on exceptionally small shares of the vote.
There is no doubt that many voters have found the start of the Starmer Government disappointing and Labour’s estimated vote share would drop significantly were there to be an election tomorrow. Far from the usual electoral honeymoon, our model estimates that Labour would lose nearly 200 of the seats they won in July’s election. While the new Government is still in its infancy it is clear that decisions such as means testing the winter fuel allowance and other budget measures have landed badly. The pressure from the public is now on the Government to deliver.
However, our model also shows the challenge the Conservatives face. Despite the models estimate of seat gains, they would be heading for their second worst share of the vote in history, and without making significant gains against Reform on the right, and the Liberal Democrats on the left, would struggle to come even close to being able to form a majority Government.
Instead, the big beneficiaries of the unpopularity of both main parties are Reform UK with our model estimating their number of seats will increase 14 fold. Strikingly, the Party appears to have established two particularly strong regional bases of support - one in South Yorkshire/North Nottinghamshire and the other in Tyne and Wear.”
Overall, there are 221 seats where the winner holds a lead less than 5 percentage points, where even a small swing could change the results.There are 87 seats in a statistical tie with the estimated winner less than 2 percentage points ahead of their closest rival.
In a sign of the fragmentation of British politics, in two out of every five seats the modelled winner has less than a third of the vote share. Across the country there would be a series of exceptionally close three-way races between the Conservatives, Labour and Reform UK (and the Scottish National Party in some races in Scotland).
This model also suggests some unwinding of tactical voting between progressive parties - Labour, Liberal Democrats, and the Green Party are seeing vote share rising in seats they are not realistically targeting but falling in seats they hold or are competitive in. This would likely change in the run up to an election.
The Conservatives would win an additional 102 seats if an election were held today, despite receiving their second worst result in the popular vote in history, and cede one to the SNP in Scotland. But they would be nowhere close to being able to form a majority Government. They would be in second place in 207 seats.
The model finds that Reform are set to make significant gains - particularly in South Yorkshire and North Nottinghamshire, Greater Manchester and Tyne and Wear, as well as their first ever seats in Wales. They would retain all of their five current seats, gain 67 seats and come in second place in a further 206 seats.
The model finds the Liberal Democrats would comfortably hold a majority of the seats they won from the Conservatives in the 2024 General Election with 58 seats - highlighting the difficulty the Conservatives would face in building back to a majority. Of the 14 seats they are estimated to lose, 13 fall back to the Conservatives and 1 to the SNP. However, they are likely to be close to their electoral ceiling, with an estimated second place finish in only 35 seats.
The SNP would be set for a recovery in Scotland with 37 seats - a gain of 28 from their 2024 results, reversing many of Labour’s gains in July but still below their 2019 total of 48 seats. 26 of their gains would come from Labour, one from the Conservatives, and one from the Liberal Democrats.
The model estimates that all of the current Independent MPs would retain their seats, and they would gain three seats: Ilford North, Birmingham Hall Green and Moseley, and Bradford West.
The model estimates that the Greens would hold two of their four seats, losing Waveney Valley and North Herefordshire to the Conservatives but retaining Brighton Pavillion and Bristol Central. The party would also come second in 20 seats.
‘Multilevel Regression with Post-stratification’ (MRP) uses data from a voting intention poll to model how people will vote based on their demographics, voting behaviour and information about their constituency. These results are then applied to the demographic and electoral makeup of each constituency to make a constituency-level estimate. The model is 'multilevel' because it uses both individual and constituency-level data.
The voting intention regularly published by More in Common is a national estimate based on a representative sample of at least 2,000 people. It indicates roughly how many people in Great Britain intend to vote for one party or another. This is simple to calculate and allows us to track changes through time.
But if you want to estimate a national seat count, this isn’t as useful. No political party performs equally well in every seat, because their supporters are not evenly spread across the country. For example, a 70-year-old man who didn’t go to university and lives in a small village has a higher likelihood of voting Conservative than a 25-year-old woman renting a flat in a major city. The benefit of MRP is the ability to use information about the different people who live in every constituency across the country to estimate how many people will vote for each party.
When we ask people their voting intention, some people say they don’t know. We push them to say who they would vote for if they were forced to choose, and we use this response as their expected vote. Some people, when asked to imagine that they were forced to choose, still don’t know who they would vote for. Using our MRP model, we’re able to make a better guess at how these “double don’t knows” might end up voting. When training the model to predict people’s voting intention based on their demographics, voting behaviour and information about their constituency, we excluded the responses of people who didn’t know who they would vote for (after the squeeze) from the training data. When we apply the model to all the voters in the constituency, it effectively means we estimate the votes of people who don’t know, according to how people like them (in terms of demographics and past voting behaviour) but who do know, intend to vote. So if someone lives in a rural area, is over 75 and voted Conservative in 2024, the model uses the fact that most over 75s in rural areas who voted Conservative in 2024 and do know who they’ll vote for say they will vote Conservative, to guess that if they do vote it will likely be for the Conservatives.
A key part of MRP is the post-stratification, which applies predictions to a dataframe of who lives in each constituency. Our post-stratification frame relies on demographic data from the census, which is then extended using data from the British Election Study Post-Election Random Probability Survey which allows us to model which types of people voted for different parties in previous elections.
As the British Election Study Post-Election Random Probability Survey for the 2024 General Election is not yet available, we have used our post-election polling to approximate the demographic characteristics of those who voted in 2024. This is the best currently available data that we have but we will update our post-stratification frame once the BES data for 2024 is available.
Our voting intention polls during the 2024 General Election campaign used the actual candidates who were running in the respondent’s constituency - this model assumes that all parties are fielding a candidate in each constituency. We also don’t have assumptions about tactical voting in this model as we would during a General Election campaign as these tend to not be useful without an imminent election.
With four and a half years before the next General Election must be called this model is unlikely to represent anything close to the ultimate result and should not be seen as a projection of the election.
As well as not knowing what might happen between now and 2029, we also don’t know which parties will stand in different seats, what tactical voting might look like and who will ultimately turn out to vote. What’s more, the degree of electoral fragmentation makes individual seat dynamics even more difficult to project than previously.
Instead this model provides a baseline for how the electorate has fragmented since the last General Election and what the implications of that might be for the make up of a future Parliament. We will continue to update it throughout the next Parliament and introduce new data as it becomes available.
MRP models are a good way to estimate how the parties might perform across different constituencies based on their demographic makeup. However, they don’t account for local factors that impact a small number of constituencies, such as a popular incumbent, well known or controversial council policy. These factors make it difficult to predict exact vote shares even in the best of times, but even more so when three parties are polling at over 20%, making three-way races more common. Therefore it would be a mistake to draw too much from the estimated vote share in an individual constituency.