by Dr Adrian Worton
Today marks the two-weeks-to-go point in the election campaign. So we will pause our series on individual parties to see if there has been any movement in the data since our last update. We will also be reflecting on the need for tactical voting. For the latest probabilities for each constituency, scroll down to the green box midway through this article.
The main reason there may have been a change is the release new MRPs from Ipsos and YouGov giving the latest constituency-level polling predictions. Their seat projections are largely in the same ballpark as our's have been, with Labour estimated to end up on between 425 and 462 seats, whilst the Conservatives on 99 to 123 [footnote 1].
Today marks the two-weeks-to-go point in the election campaign. So we will pause our series on individual parties to see if there has been any movement in the data since our last update. We will also be reflecting on the need for tactical voting. For the latest probabilities for each constituency, scroll down to the green box midway through this article.
The main reason there may have been a change is the release new MRPs from Ipsos and YouGov giving the latest constituency-level polling predictions. Their seat projections are largely in the same ballpark as our's have been, with Labour estimated to end up on between 425 and 462 seats, whilst the Conservatives on 99 to 123 [footnote 1].
Stirling & Strathallan remains absent from the data. Investigating the first YouGov MRP I have decided it is most similar to Edinburgh East & Musselburgh, so for now we are using the odds from the latter for Stirling & Strathallan. This means the total seats for each measure will be 1 higher than last time as they add up to the full 650 rather than 649.
What are the main changes to our update on the 16th June? The answer is: not a lot.
The lack of change is a bit surprising as there are a fair few differences between our previous projections and the latest MRPs. For example, both Ipsos and YouGov are projecting Reform UK to pick up more than 1 seat (3 with Ipsos, 5 with YouGov). However they remain favourites in only one seat with the bookmakers (Clacton). This possibly suggests that for a gambler there is strong value in backing Reform in Ashfield, which Ipsos rates as "Strong Reform".
Despite a lack of change overall, there is quite a bit of movement on individual seats. The most is in Frome & East Somerset, where the Liberal Democrats were previous rated as having a 1% chance but are suddenly favourites in a three-way marginal with 44% to the Conservatives' 30% and Labour's 26% [see footnote 2].
What are the main changes to our update on the 16th June? The answer is: not a lot.
- Looking at who is the favourite for each seat, the Liberal Democrats are up 5 seats, with the Conservatives down 3 and the SNP down 1.
- This is reflected in the expected seat measure too. The Lib Dems are up by 3.73, whilst Conservatives are down 3.63. Nobody else has moved by more than 1.
The lack of change is a bit surprising as there are a fair few differences between our previous projections and the latest MRPs. For example, both Ipsos and YouGov are projecting Reform UK to pick up more than 1 seat (3 with Ipsos, 5 with YouGov). However they remain favourites in only one seat with the bookmakers (Clacton). This possibly suggests that for a gambler there is strong value in backing Reform in Ashfield, which Ipsos rates as "Strong Reform".
Despite a lack of change overall, there is quite a bit of movement on individual seats. The most is in Frome & East Somerset, where the Liberal Democrats were previous rated as having a 1% chance but are suddenly favourites in a three-way marginal with 44% to the Conservatives' 30% and Labour's 26% [see footnote 2].
Constituency profiles
Select your constituency from the drop-down list below to view a quick factfile and our latest probabilities for its winner.
If you aren't sure of the name of the constituency you want, you can use the UK Parliament tool here to search by postcode.
Select your constituency from the drop-down list below to view a quick factfile and our latest probabilities for its winner.
If you aren't sure of the name of the constituency you want, you can use the UK Parliament tool here to search by postcode.
Tactical voting
Another development this week was the launch of tactical voting recommendations by the anti-Brexit campaign group Best for Britain.
The idea is to provide the best-placed candidate to beat the Conservatives. In 2017 and 2019 they published recommendations to try and deny them a majority. Specifically, this is by maximising the seats taken by all parties seen as pro-EU: Greens, Labour, Liberal Democrats, Plaid Cymru and SNP (no Northern Irish recommendations were given). For the purposes of this, those five parties are effectively treated as one identical bloc.
For example, below is Best for Britain's recommendation for Richmond & Northallerton:
Another development this week was the launch of tactical voting recommendations by the anti-Brexit campaign group Best for Britain.
The idea is to provide the best-placed candidate to beat the Conservatives. In 2017 and 2019 they published recommendations to try and deny them a majority. Specifically, this is by maximising the seats taken by all parties seen as pro-EU: Greens, Labour, Liberal Democrats, Plaid Cymru and SNP (no Northern Irish recommendations were given). For the purposes of this, those five parties are effectively treated as one identical bloc.
For example, below is Best for Britain's recommendation for Richmond & Northallerton:
However, the upcoming election sees the chances of a Conservative majority as nonexistant. No method we have used, nor anything anybody else has published, has the party getting even half the seats they would need to form a majority (325).
Even if the aim is not to stop a pro-Brexit majority, but just to minimise the number of pro-Brexit MPs, the Conservatives are not competitive in many seats.
Based on the review of the 2017 election, Best for Britain stopped making predictions for safe seats. So if you search for the ultra-safe Labour seat Bootle this time, you get the following message:
Even if the aim is not to stop a pro-Brexit majority, but just to minimise the number of pro-Brexit MPs, the Conservatives are not competitive in many seats.
Based on the review of the 2017 election, Best for Britain stopped making predictions for safe seats. So if you search for the ultra-safe Labour seat Bootle this time, you get the following message:
This is because "the incumbent MP is polling at over 15% ahead". Similar seats with one clear front-runner return the same advice.
But if Best for Britain see the aforementioned five parties as interchangable, then surely it only matters how close the Conservative (or any other pro-Brexit) candidate is to winning. Take my constituency for example:
But if Best for Britain see the aforementioned five parties as interchangable, then surely it only matters how close the Conservative (or any other pro-Brexit) candidate is to winning. Take my constituency for example:
Dunfermline & Dollar is a new constituency made up of three seats won by the SNP at the last election [footnote 3], with Dunfermline & West Fife the previous constituency that provides most of the new region.
Labour finished second to the SNP in the three previous elections in Dunfermline & West Fife, and held the seat before that. It is inarguably a SNP-Labour marginal, with little chance of a pro-Brexit party winning (1% in our model). Ipsos estimate the Conservative vote share at 10%, YouGov 6%. Even at the last election, which went great nationally for the Conservatives, they finished third here, with their share of the vote 23% behind the SNP.
Dunfermline & Dollar is guaranteed to go to a party Best for Britain rates as pro-EU, so why do they need to give a recommendation as to which one voters should pick if they see them all as interchangeable?
Given the Conservatives are struggling to hold onto even their safest seats (as we documented yesterday), then the notion of them winning any seat they didn't take in 2019 seems pretty remote. And indeed, here is a histogram showing the Conservatives' probability of winning seats they didn't take in 2019 (not including new constituencies):
Labour finished second to the SNP in the three previous elections in Dunfermline & West Fife, and held the seat before that. It is inarguably a SNP-Labour marginal, with little chance of a pro-Brexit party winning (1% in our model). Ipsos estimate the Conservative vote share at 10%, YouGov 6%. Even at the last election, which went great nationally for the Conservatives, they finished third here, with their share of the vote 23% behind the SNP.
Dunfermline & Dollar is guaranteed to go to a party Best for Britain rates as pro-EU, so why do they need to give a recommendation as to which one voters should pick if they see them all as interchangeable?
Given the Conservatives are struggling to hold onto even their safest seats (as we documented yesterday), then the notion of them winning any seat they didn't take in 2019 seems pretty remote. And indeed, here is a histogram showing the Conservatives' probability of winning seats they didn't take in 2019 (not including new constituencies):
There is precisely one constituency (Gordan & Buchan, previously just Gordon) that the Conservatives didn't win in 2019 that they are expected to win this year, with our model giving them a 68.4% chance of winning (Ipsos and YouGov both give the Conservatives a very narrow lead in this seat).
There are two more SNP-held seats (Angus & Perthshire Glens and Perth & Kinross-shire) where the Conservatives have a chance of winning of just under 20% (specifically, 19.0% and 17.1%). The remaining seats are very unlikely to turn blue, with the majority rated as having under 1% chance of doing so.
My tactical voting advice is clear: if you want to limit the Conservatives' seats (or Reform/DUP), you can use our constituency profile tool above to see if they have a realistic chance of winning, then vote for the nearest challenger. Other than that, vote with your heart.
But furthermore, I would argue that Best for Britain's recommendations fall short of their intended objective. If their aim is to increase ties with the EU then there is a clear difference between the manifesto commitments of Labour (left below) and the SNP (right below):
There are two more SNP-held seats (Angus & Perthshire Glens and Perth & Kinross-shire) where the Conservatives have a chance of winning of just under 20% (specifically, 19.0% and 17.1%). The remaining seats are very unlikely to turn blue, with the majority rated as having under 1% chance of doing so.
My tactical voting advice is clear: if you want to limit the Conservatives' seats (or Reform/DUP), you can use our constituency profile tool above to see if they have a realistic chance of winning, then vote for the nearest challenger. Other than that, vote with your heart.
But furthermore, I would argue that Best for Britain's recommendations fall short of their intended objective. If their aim is to increase ties with the EU then there is a clear difference between the manifesto commitments of Labour (left below) and the SNP (right below):
To be clear, I am not saying that either position is better in my opinion. But the SNP's should be much better in Best for Britain's opinion. So if they are to take a position in marginal seats between non-Conservative parties, they should be backing the ones which match their aims, rather than which one is polling better.
By just recommending who is polling better, that strongly benefits Labour, who are riding very high in the polls. Generally speaking, it strongly locks in the status quo and actively tries to stifle change.
A more obvious pro-EU tactic would have been to recommend votes for Green/LD/Plaid/SNP candidates when they have a chance of beating Labour, to put pressure on Labour to move to a more pro-EU position.
The Best for Britain voting recommendations just do not make sense to me. That is, unless, they have a vested interest in Labour doing well. It is not exactly a grassroots campaign. Their chair is the chair of Virgin Group and their CEO is a former executive of a business lobby group. It is not unreasonable to think that the tactical voting recommendations are taking more into account than its public aims.
A more charitable explanation is just that its recommendations are effectively done by algorithm, and that not much thought has been put into whether recommendations are necessary. Errors in the recommendations do suggest a lack of quality. But as a well-funded team of "researchers, data scientists, strategists, and activists" (Wikipedia), this just doesn't seem very likely. Our work here on TGIAF is entirely done by a two-person operation, in our spare time. It was made almost from scratch a few weeks ago (with a holiday in the middle), and despite this I have been able to see the flaws in Best for Britain's tool.
Summary
Despite new polling, our odds-based models seem fairly resistent to change. There is a case that there could be money to be made by using the latest data. But the overall picture for the Conservatives remains one of gloom, as their prospects slip further and further.
Meanwhile, the idea of tactical voting in an election in favour of the party that is set to have a majority in the region of 200 seats is preposterous, and raises serious questions.
We will return to our series looking at individual parties with a look at the Lib Dems and Reform, those best of friends.
By just recommending who is polling better, that strongly benefits Labour, who are riding very high in the polls. Generally speaking, it strongly locks in the status quo and actively tries to stifle change.
A more obvious pro-EU tactic would have been to recommend votes for Green/LD/Plaid/SNP candidates when they have a chance of beating Labour, to put pressure on Labour to move to a more pro-EU position.
The Best for Britain voting recommendations just do not make sense to me. That is, unless, they have a vested interest in Labour doing well. It is not exactly a grassroots campaign. Their chair is the chair of Virgin Group and their CEO is a former executive of a business lobby group. It is not unreasonable to think that the tactical voting recommendations are taking more into account than its public aims.
A more charitable explanation is just that its recommendations are effectively done by algorithm, and that not much thought has been put into whether recommendations are necessary. Errors in the recommendations do suggest a lack of quality. But as a well-funded team of "researchers, data scientists, strategists, and activists" (Wikipedia), this just doesn't seem very likely. Our work here on TGIAF is entirely done by a two-person operation, in our spare time. It was made almost from scratch a few weeks ago (with a holiday in the middle), and despite this I have been able to see the flaws in Best for Britain's tool.
Summary
Despite new polling, our odds-based models seem fairly resistent to change. There is a case that there could be money to be made by using the latest data. But the overall picture for the Conservatives remains one of gloom, as their prospects slip further and further.
Meanwhile, the idea of tactical voting in an election in favour of the party that is set to have a majority in the region of 200 seats is preposterous, and raises serious questions.
We will return to our series looking at individual parties with a look at the Lib Dems and Reform, those best of friends.
Footnotes
[1]
|
Ipsos projects Labour to be in the range 439-462 and the Conservatives 99-123. YouGov gives a single seat estimate, which is 425 for Labour and 115 for the Conservatives.
|
[2]
|
Frome & East Somerset is a constituency to watch. YouGov project the Lib Dems to take 38% of the vote, with the Conservatives second on 24%. But Ipsos have Labour top with 31%, followed by the Conservatives on 28% and the Lib Dems on just 21%. For what it's worth, Best for Britain back the Lib Dems in this seat. Suddenly, the rapid changes in this seat's odds make sense.
|
[3]
|
Strictly speaking, Neale Hanvey was suspended by the SNP before he won Kirkcaldy & Cowdenbeath, but close enough to the election that meant he still was listed as the SNP candidate on the ballot. His suspension was later lifted, before he eventually defected to Alba.
|