GE24: Final Predictions and Estimates

Hard as it is to believe, polling day is finally here. If you’d told me just two months ago that I’d be saying “finally” on the 4th of July, I’d have been astonished – after all, at that point we were expecting an autumn election at earliest, with some wondering if Sunak would even drag it out until the legal limit of January 2025. Yet one absolutely soaking surprise election call later, which really set the scene for our miserable non-summer thus far, here we are.

As is Ballot Box Scotland tradition, let’s take one last look over the state of play. Where has polling landed as the country goes out to vote? How has it shifted over the course of the campaign? And, crucially, what might that mean in seat terms? Remember too that there is a huge amount of information available on the BBS GE24 Hub beyond what’s in this piece, including complete results of the 2019 election on the new boundaries.

Polling

Final Average

The final shape of polling is much as you’d expect if you’ve been following the campaign at all closely: a Labour lead over the SNP, but not a massive one, whilst the Conservative vote has been cut substantially. That has been the pretty consistent story not just across polling in aggregate but in most of the individual polls included in this final average, with one exception – Savanta, who provided the last Scotland-only poll of the entire campaign, which makes a significant difference to the average.

They have the SNP a few points higher than everyone else, and Labour correspondingly lower. Indeed, this is the first poll since around February to put the SNP ahead at Westminster. Without it Labour would have a 5.2% lead over the SNP, rather than the final 3.6% here. As top boffin Mark McGeoghegan outlines here, that’s a small enough lead that allowing for error it’s still distinctly possible that the SNP could be closer to or further away from Labour. Even if the gap is wider on the day, it’s a far cry from the roughly 20% advantage Labour have had over the Conservatives GB-wide.

As ever with polling, it’s incredibly hard to say who has it right or wrong. In 2017 for example, the polling average overestimated the SNP by about 4-5%, and in 2019 it underestimated them by roughly the same amount. Unfortunately, the last Scotland-only polls by Ipsos and Opinium wrapped up about four weeks before polling day, leaving a bit of a gap in the final polling spread. It’d have been useful to have both of those firms, if only for the purposes of playing everyone’s favourite game of “who called it the closest”. 

Swing versus 2019

Relative to 2019 then, Labour look set to roughly double their share of the vote, whilst the SNP’s is cut by near a third and the Conservatives nearly halved. The Lib Dems may be breathing a very small sigh of relief, as for a good while it looked like they might lose quite a lot of ground, but they are only very marginally down.

The numbers with the biggest questions marks over them however are Reform UK and the Greens. As both only stood in a small minority of seats (about a quarter and a third, respectively) in 2019, these aren’t “true” swings, nor will their final national results be properly comparable either. Reading Green results is also complicated by the fact that although they are standing a record number of candidates, they’re still missing a quarter of constituencies. 

Polling Through the Campaign

Whilst the tracker on the GE24 Hub covers all of 2024, I’ve zoomed in here on the campaign period, starting with the first point in the average after Sunak called the election on the 22nd of May. At that point, Labour’s lead over the SNP was a full 6%. The SNP recovered the tiniest smidge over the course of the campaign, before settling back barely any higher than they started, but Labour too have slipped a little bit, in line with a GB-wide dup during the campaign.

The Conservatives have also ended up worse off, with the steepest decline coming when Reform UK really began to take off after Farage’s shock announcement that he was standing in Clacton. That brought them to within touching distance of the Lib Dems, whilst the Greens continue to seemingly languish in near-irrelevance whenever a Westminster vote comes around.

Seat Estimates

Before we get into the projection, let’s lay out four big caveats that come with this. I am;

  • ASSUMING that polling is roughly accurate
  • NOT Predicting a specific number of seats as definite
  • NOT Predicting which party will win key battlegrounds
  • NOT Predicting all other seats are safe and definitely won’t switch
Final Seat Projection
Estimated Seat Ranges

The ranges here for the SNP and Labour are pretty enormous. As I will never stop pointing out, this is simply down to our ridiculously antiquated and fundamentally unrepresentative, and thus undemocratic, voting system. If we actually had a proper democracy with a solid form of PR, I’d have estimated the SNP’s range of potential seats at between 16 and 21, compared to a range of 17 to 23 for Labour. However, we are where we are.

What’s most notable here is that there is an incredibly limited envelope for the SNP to beat Labour in seat terms in my view. Only if the SNP had their absolute best result and Labour their worst could the SNP squeak ahead. Labour’s expected performance is so strong that even on the middle estimate they win over half of Scotland’s seats, whilst their high estimate would give them bang-on two-thirds.

For the Conservatives, my assumption is that they will hold up relatively well compared to their English counterparts. That’s simply because their competition in Scotland is an also-flagging party of government, and in rural Scotland where they are strongest, Labour simply aren’t in the running. Finally, note that the Lib Dem middle and high estimates are the same. In effect, it’s hard to see where the Lib Dems pull off any wins on top of their 5 core seats, though they may be able to position the other two primary Highland seats for wins at the next election.

What if it was Proportional?

I’m not spending long on this bit, as there’ll be much more to come on the Proportional Representation front after the election. However, I did think it would be useful to make the explicit comparison. This is an estimate using my preferred nationally proportional (for parties above 3% of the vote) model, adjusting vote shares slightly to reflect different voter behaviour as seen in the BBS exclusive poll last week. This would primarily redistribute Labour’s artificially inflated share of seats to the Conservatives, Reform UK and Greens.

Mapping the Marginals

There’s something of a geographic balance here that Edinburgh South West slightly ruins, as otherwise it’d be the case that all of the tossup constituencies were in the North East or the South West (ish – East Renfrewshire is contiguous with Ayrshire, so visually it looks neat). There’s quite a significant difference in character between the patches to, with those North East tossups between the SNP and Conservatives, whilst those in the South West are largely between the SNP and Labour. The one exception is that Dumfries and Galloway comes out as a three-way marginal in my model, with 5% capable of swinging it for any of those parties.

Outside of that, there are some slight differences between BBS and some other estimates around the true centre of the country – Perthshire. Whereas some others are viewing Stirling and Strathallan as marginal, I’ve got it more firmly in the SNP’s camp. That’s because it’s got a very stark split between urban and rural components, and I cannot see the latter going particularly heavily for Labour whilst the SNP are polling at the level they are nationally.

Note that if you try and tally up the figures here assuming all leans and tossups go for or against parties in their low and high estimates, your figures won’t match. That’s because I don’t consider it likely that a party would miss out on or emerge triumphant in every lean. For example, although I’ve only got two Conservative leans, I don’t think it’s likely that they’ll lose both and be wiped out entirely. It is, however, still possible!

Labour vs SNP Contests

As outlined in detail in this piece, I expect Labour to sweep most Central Belt seats from the SNP simply due to each party’s spread of votes. Indeed, I’ve marked out most of the Central Belt as being Likely Labour, with only a handful of Leans in some of what have historically been more strongly SNP areas, including both West Lothian seats plus East Kilbride and Strathaven. Just like down south, it’s inevitable that Scottish Labour will have a huge amount to celebrate this year.

On the other hand, the reason most of Ayrshire lies within Tossup territory is that the Conservatives made really big advances across the county in recent years, but it’s also generally highly urbanised. That makes it ripe for a Labour revival but means they are starting from further behind than in the likes of Renfrewshire and Lanarkshire. Edinburgh likewise isn’t expected to be a complete Labour walkover because it’s typically been the most politically diverse and thus divided part of the country, which has until this point benefitted the SNP as the party with the least competition on their side of the constitutional aisle.

Conservative vs SNP Contests

Forget the other constituencies for a moment: all eyes this election are on Aberdeenshire North and Moray East. Douglas Ross’ astonishingly ill-judged decision to shunt aside his hospitalised colleague David Duguid at the last minute to stand for this seat has already ended his leadership of the Scottish Conservatives. It may yet end his political career entirely, if you assume given all of the evidence that he’s unlikely to want to permanently commit to Holyrood. Most models have his seat, the most Pro-Brexit seat in the country, on a knife-edge

Beyond that, whether the Conservatives can hold onto their other Aberdeenshire seats will depend on two things. Firstly, whether voters locally are more scunnered with them or the SNP. Secondly, just how many tactical votes they can garner from Labour and the Lib Dems. In another in-depth analysis piece, I set out part of why I believe the Conservatives simply cannot rely on a reciprocal flow of votes from their Pro-Union counterparts.

As much as Douglas Ross has tried to make out this election is a referendum on the SNP’s time in office at Holyrood, that is obviously not the case. This is an election to the UK Parliament, where his party has been in government for 14 years, and where all indications are the overwhelming majority of people across the UK want rid of it. In an election in which the constitution has been notable by its near-absence, I’m not convinced folk who held their noses to vote Conservative against the SNP in 2019 will do so again after what happened over the past 5 years.

Lib Dem vs SNP Contests

In another quite significant contrast to England, the Scottish Lib Dems are unlikely to make any dramatic gains. There’s nowhere up here that they are best placed to beat the Conservatives, and precious few where they are on track to beat the SNP. I reckon the Lib Dems have three seats on absolute lockdown. Their Orkney and Shetland stronghold is a given, and the fact some MRPs had it going SNP was enormously eyebrow raising in my view. Added to that, they should easily hold Edinburgh Western and “gain” North East Fife, bearing in mind that on new boundaries it’s notionally SNP. They’ll have dug into it something fierce though.

Their only two other prospects for wins are Mid Dunbartonshire and Caithness, Sutherland and Easter Ross. The former was a stroke of luck (or letter writing), as the first two proposals to redraw what was the East Dunbartonshire seat would have been fatal to their chances. Instead it got minimal tweaks, and is a very strong possibility to pick up. The far north seat ends up in the SNP column in 2019 terms due to the roughly 200 vote Lib Dem majority being swamped by over 20,000 votes being added from an SNP seat on the boundary changes. Nonetheless, the Lib Dems had already worked a closely matching Holyrood seat hard, and therefore are in a good position to gain it.

Smaller Parties

There is no prospect whatsoever of any of Reform UK, the Greens, Alba or any of the micro-parties picking up MPs. Whilst Reform UK and the Greens would be in line for seats if we used a proportional voting system (the Green vote share potentially being twice as high if so), we don’t, so they aren’t. Instead, it’s going to be a matter of seeing what their vote shares are like and where.

For Reform UK, they’ll want to live up to what has been pretty decent polling despite an almost complete lack of infrastructure and ground game. Basically, it’ll be a national tide of intense media coverage that lifts that particular vote.

On the other end of the spectrum, the Greens have never done well at Westminster and don’t expect to start, nor has it historically mattered to them that they haven’t. They will at least want to try and hold some deposits in Edinburgh and Glasgow, having failed to do so back in 2019.

Lastly, Alba are almost certainly going to lose every deposit they have, but a good result might be if they could scrape across 5% in Cowdenbeath and Kirkcaldy or Alloa and Grangemouth at least. They are also only standing in a third of constituencies, which will further deflate what is likely to be a very small national share.

Get out there and Vote!

With that, there’s nothing else to do but go vote. Polls are open from 7am until 10pm. Remember that for the first time you must bring photo ID in order to vote. If you are one of the people who received an incredibly late postal ballot, you can take it to your local polling station and hand it in there. If you arrive very late to vote, as long as you are starting the process by 10pm, you will be allowed to finish it.

As the night unfolds, check back here on the Ballot Box Scotland website and over on Twitter for updates as they come in. We’re not expecting the first Scottish seat to declare until about 1am, so there’ll be three quiet hours between polls closing and the first result. And as ever, if you value my work, please consider chipping in a donation if you can afford to do so, through the link below.

If you find this or other Ballot Box Scotland output useful and/or interesting, and you can afford to do so, please consider donating to support my work. I love doing this, but it’s a one-man project and takes a lot of time and effort. All donations, no matter how small, are greatly appreciated and extremely helpful.
(About Donations)