- The first major seat-level MRP poll under new constituency boundaries suggests Labour are set for a 140+ seat majority with 470 seats. The Conservatives are projected to win 129 seats but while Labour’s lead is significant, it is smaller than our last baseline MRP poll conducted in Autumn 2022. This is after Rishi Sunak wins back some Tory voters disaffected following Liz Truss’ premiership.

- New analysis assuming that Reform UK will stand aside in Conservative marginals, shows that Labour could drop to 401 seats and the Conservatives would rise to 201 seats.

- Undecided voters remain the third largest voter bloc of the electorate after Labour and the Conservatives. Analysis shows that when you redistribute these voters based on demographic calculations, 370 seats is probably the high end of Labour’s predicted seat haul, while the Conservatives would be on 232 seats. 

- When combining both of the above scenarios, Labour could fall as low as 316 seats which would be hung parliament territory. In this scenario, the Conservatives win 286 seats.

- Analysis reveals the new key electoral battlegrounds with Scotland, South East England, and the East of England home to over half of Labour’s marginals. Voter trends in Scotland are dominated by the fall in SNP vote share, which has led to a subsequent rise in undecided voters since our last MRP.

The data below is taken from a 10,102 poll performed by Focaldata, field dates 23 April - 9 May 2023, with subsequent analysis performed by Best for Britain. All data is available in the appendix at the bottom of this web page.

THE POLLING

‘If a general election were held tomorrow, which party would you vote for?’

Since our last MRP in Autumn 2023, Labour’s national vote share has come down 7pp, and the Conservatives have gained 5pp. This could be through a combination of a general national swing against Labour and the return of Don’t Knows to the Conservatives, as our ‘Wavering Wall’ report in December 2022 predicted.

Around one-eighth of poll respondents said that they ‘Don’t Know’ how they will vote, making them the largest constituency of the electorate after Labour and the Conservatives. You can see the headline polling results below:

*The baseline MRP by Focaldata predicted Labour winning 469 seats and the Conservatives 128 seats. 'Don't Know' was predicted to win two seats, namely Aberdeenshire North and Moray East; as well as Boston and Skegness. Given in a General Election scenario 'Don't Know' would not be on the ballot paper, we reallocated these seats to the party in second place. This meant that we awarded Aberdeenshire North and Moray East to the Labour Party; and then Boston and Skegness to the Conservatives in all scenarios and analyses that we performed.

MARGINALITY

206 of 632 seats in Great Britain are marginal. We are defining a marginal as a +/- 5% gap between the leading and second placed parties. Of these 206 seats, 165 are in England, 38 are in Scotland and 3 in Wales. 

Of these marginal seats, Labour are either the first or second placed candidates in 197 of them. This is up from 133 seats in Autumn 2022. There are also four seats where Labour are within +/- 5% as a third placed candidate. Also, 155 of these seats are direct Labour / Conservative marginals.

The maps below show where these marginals are located across GB, with the first map demonstrating the first placed party in each marginal seat, and the second map showing the projected-runner up. You can also look at a final map, showing third-placed parties in seats considered to be ‘three-way-marginals’.

LABOUR'S MARGINS

When looking more closely at the marginal seats the Labour Party is placing either first, or second in, we can see over half of them are  located in the South East of England, the East of England, and Scotland. The table below shows the amount of Labour marginals within each region - and their percentage in relation to all 197 of Labour’s marginals.

We have analysed 392 new parliamentary seats that are at least 80% similar to their “predecessor” constituency, according to the Boundary Commissions for England, Wales, and Scotland. 

These seats are largely representative of the entire set of 632 seats outside Northern Ireland (the average Labour vote share in the successor constituencies is 36.5%, identical to their average share in all of GB’s constituencies, excluding the speaker’s seat). 

When looking more closely at trends within these seats over time, it is clear that Labour’s margins (including don’t knows) have tightened since our last round of polling, with Labour’s median margin falling in every region. Areas such as Yorkshire and the Humber, the North West of England and the West Midlands are three areas where this impact is being felt most. The fall is lesser in London, Southern England, and the North East, but there is a fall everywhere. The graph below demonstrates this with box plots for each region.

It remains the case, though, that if there was an election today, Labour's seat haul would be very high. The scatter graph plots below show the change in Labour's margin from our Autumn 2022 MRP to now, versus their initial vote share in Autumn 2022, for seats that they were winning in the Autumn 2022 MRP. 

In essence, our analysis has found that because Labour’s margins have been so high - although they are now falling it’s not had a huge deal of impact on the seats the party wins. 

To demonstrate this, the graph below shows seats that Labour were projected to win in our Autumn 2022 MRP. Labour’s margin in Autumn 2022 is on the x axis, and the change in their margin between then and the Spring 2023 MRP is on the y axis. 

The graph above shows that Labour’s margin is falling in the majority of seats, but their initial margin was high enough to keep from losing too many seats. Though there are some seats that have flipped since Autumn 2022 (points to the left of the black dotted line), Labour retain many of their seats. 

This could explain why Labour is predicted to win 47 fewer seats compared to the baseline prediction from our Autumn 2022 MRP - and demonstrates that margins are shrinking across most seats.

SCENARIO 1: REGRESSIVE ALLIANCE

In 2017 and 2019, UKIP and the Brexit Party (now Reform UK) stood aside for the Conservatives in marginal seats. 

Under the assumption that, faced with a Labour victory, Reform UK would once more decide to stand aside for the Conservatives in marginal seats, we performed analysis to model the impact of this scenario. We assumed that 100% of Reform UK’s vote would transfer to the Conservatives in marginal seats.

When doing this, the results were:

SCENARIO 2: REDISTRIBUTING UNDECIDED VOTERS BY EDUCATION PROFILE

Analysis of the age and education profiles of undecided voters carried out in our last MRP report, the ‘Wavering Wall’, found that, in England, over two thirds of undecided voters were likely to favour the Conservatives.

Our analysis of data in this MRP however  has found changes to the demographic profiles of undecided voters. These are as follows:

• Undecided voters are now younger as a whole. This is likely to be a result of older voters returning to the Conservatives (in England), thus bringing the overall age of undecided voters down. 
• 69% of undecided voters are women (63% in the previous round).
• There has been a marked increase in undecided voters in Scotland. This is likely to be a result of those who previously intended to vote for the SNP now being ‘undecided’.

The education profile of undecided voters continues to skew heavily to poll respondents who are non-university educated (see the figure below). For this reason, we have continued to focus on a modelling scenario where undecided voters are transferred to other parties based on education profile. 

When we allocate undecided voters according to the relative similarities in education profiles in England, the Conservatives receive around 60% and Labour receive around 25% of undecided votes. Labour benefit from a slightly smaller transfer proportion going to the Conservatives this time around, consistent with a number of undecided voters returning to the Conservatives since Autumn 2022.

The voting transfers we have calculated are different in Scotland and Wales for each party, but undecided voters still break most in favour of the Conservatives in these countries. This is a significant change in Scotland when making a comparison against our MRP in Autumn 2022, where we calculated that undecided voters would break more favourably to the SNP than any other party.

The tables below the transfer proportions in Autumn 2022 and Spring 2023, and the overall changes between them. You can find out more information about the statistical analysis we performed on undecided voters in the appendix.

Table 1: Education Profile Redistributions (Autumn 2022 MRP)

Table 2: Education Profile Redistributions (Spring 2023 MRP)

Table 3: Education Profile Redistribution Change between Autumn 2022 MRP and Spring 2023 MRP (+/-)

Modelling this redistribution at seat-level, Labour wins 370 seats and the Conservatives win 232 seats. 

COMBINING SCENARIOS 1 & 2

We decided to model a final scenario, where Reform UK stands down in marginal constituencies that the Conservatives are contesting and we redistribute undecided voters by their education profile.

This scenario gives the following results, which would likely result in a hung parliament:

SCENARIO RECAP

An MRP poll is a snapshot in time. The constituency level data Focaldata returned to us shows that if there was a General Election today, Labour would still win 470 seats, with the Conservatives on 129 seats.

We then performed our own analysis. This found that if Reform UK stood aside for the Conservatives in their most marginal seats, as UKIP and the Brexit Party did in 2017 and 2019 respectively, Labour’s seat number would be reduced to 401. We have also calculated that if we transfer voters by education profile, Labour’s seat haul is reduced to 370 and the Conservatives would be on 232 seats.

In a worst-case scenario for Labour, in which undecided voters split between the parties in proportion to how similar their education profile is to the voters of those parties, as well as the Reform UK stand-downs, Labour’s seat haul is reduced to 316, with the Conservatives on 286 - a hung parliament. The charts below show these headline figures, and we will explain our methodology on this page for each scenario in further detail.

Table 1: Baseline MRP Model

Table 2: Reform UK Stand Aside Model

Table 3: Education Profile Transfer Model

Table 4: Education Profile Transfer + Reform UK Stand Aside Model

SCOTLAND

Labour is looking strong in Scotland. Our MRP projects that they would pick up 31 Scottish seats if the election were held today. 

When comparing the results from respondents in Scotland to our 10,010 person poll in Autumn 2022, and our 10,102 person poll in Spring 2023 we can see that the SNP’s vote share has fallen 8pp from 38% to 30% while Labour’s vote share has stayed the same at 28%. 

Deeper investigation as to how vote shares have changed since our Autumn 2022 MRP strengthens the assumption that Labour is benefitting from a large transfer of votes from the SNP to the ‘Don’t Know’ category, rather than a rise in Labour’s own vote share.

The figure below shows that the share picked up by ‘Don’t Know’s’ is higher than the share picked up by Labour, for an equivalent fall in SNP share. This means that in many cases, Labour’s lead in Scotland exists by default, rather than because they’ve directly won a large number of votes from the SNP.

The second figure shows that, in contrast to the SNP, Labour is  doing better in places where there are more DKs in Scotland. The SNP is losing some voters directly to Labour, but the takeaway is that, similar to what was going on in Autumn 2022 in England and Wales with Conservative voters, many of the disaffected SNP voters are simply undecided now, rather than committed to Labour. 

CONCLUSION

• Our data shows that the Labour Party is on track to win 470 seats - and a 140+ seat parliamentary majority on new constituency boundaries.

• But, our analysis which models a range of scenarios, puts Labour’s majorities under varying degrees of risk. 

• One model, that assumes Reform UK will stand down for the Conservatives in marginal seats, gives Labour 401 seats, and the Conservatives 202 seats.

• Another model, based on similar analysis to our wavering wall report assesses what happens when you transfer undecided voters by their education profile. This gives Labour 370 seats, and the Conservatives 232 seats.

• When combining the above scenarios into a ‘worst case scenario’ model - where undecided voters break in favour of the Conservatives, and Reform UK stands down in marginal seats, the model predicts a hung parliament where Labour get 316 seats and the Conservatives get 286 seats.

• Voter trends in Scotland are dominated by the fall in the SNP vote share which has led to a rise in undecided voters, thus meaning the Labour Party are currently on track to win in a number of seats in Scotland.

THE POLL

Poll of 10,102 GB adults by Focaldata commissioned by Best for Britain Limited conducted between 20 April 2023 and 9 May 2023.

Data Tables:

• Constituency-level MRP results can be found here. This includes the Spring 2023 MRP constituency-level results as well as Best for Britain's scenarios including one where Reform Uk stand down in Conservative marginals, one which redistributes ‘Don’t Know’ voters by education profile; and one scenario which combines the two. This sheet also includes details of marginal seats, taken from our 'Baseline Scenario'. 

• Focaldata poll results for the 20 - 26 October and 28-30 October 2022 polls can be found here. 

• Focaldata MRP results for 20 April - 9 May 2023 can be found here.

• Focaldata data tables for 20 April - 9 May 2023 can be found here.

NEW CONSTITUENCY BOUNDARIES

The UK’s boundary commissions have undertaken a review of UK Parliamentary constituencies and have published their proposed changes. The new boundaries are set to be adopted automatically in Summer 2023 and to be used at the next UK General Election.

Published revised new boundaries from each of the boundary commissions can be found below: 

Boundary Commission for England

Boundary Commission for Scotland

Boundary Commission for Wales

Our poll included respondents from GB only, so Northern Ireland has not been mapped.

Appendixes

Appendix A: What is MRP?

Appendix B: Vote Transfer methodology

Appendix C: When is the next UK general election?