Graphic with decorative image, the blog title and a photo of the writer

How Scotland Voted in the 2024 UK General Election

Published: 10 July 2024

by Mark McGeoghegan, Doctoral Researcher, University of Glasgow

Going into last Thursday’s election, I doubt many would have predicted the enormous scale of the Labour victory in Scotland. Looking across the polling, there was very little sign that they would sweep the Central Belt in quite so spectacular a fashion, nor that the SNP vote would hold up so poorly.

Very few predicted that the Liberal Democrats would win Inverness, Skye and West Ross-shire. Likewise, it was not broadly expected that the only seat the Conservatives would lose would be Aberdeenshire North and Moray East, contested by the Scottish Conservative leader, Douglas Ross.

In the end, the SNP’s 39 seats lost – leaving them on just nine – was a far worse performance than models suggested ahead of the vote. Labour’s 37 seats slightly outperformed their predicted best-case scenario, and the Liberal Democrat’s six seats, overtaking the Conservatives on five, was also beyond what one could have reasonably predicted from the polls:

A table showing the UK General Election results in Scotland

The results, and the dramatic overhaul of the Scottish electoral map, suggest a polling miss. Yet voting intention polling in Scotland got the national vote shares pretty much spot on. A LOESS regression fitted to the Scotland-specific polls, Scottish subsamples in British polls, and implied Scottish vote shares from multiple regression and post-stratification models predicted the actual vote of every party to within one percentage point, except for the Liberal Democrats whose vote it understated by just 1.2 points. These errors are nowhere near large enough to account for the gap between the expected seat result and the eventual result on Thursday.

A table showing the Scotland predicted vote accuracy

The explanation lies in a mix of efficient swings and electoral geography. Scotland’s post-2014 independence referendum electoral geography set up Westminster elections which were, essentially, three separate contests. The SNP vote was very flatly distributed across the country. In 20 of the 59 seats Scotland had, the SNP vote in 2019 did not vary from their national vote by any more than four percentage points, and a further 15 did not vary by any more than nine percentage points.

On average, the SNP’s constituency vote shares varied from their national vote share by just 9.6%. In comparison, the Liberal Democrats' vote shares varied from their national vote share by 62%, Labour by 54%, and the Conservatives by 39%, meaning that their vote shares were more concentrated in a smaller number of constituencies.

This meant that in the south and northeast of Scotland, the contest was between the SNP and the Conservatives. In the Central Belt, it was between the SNP and Labour. And in areas with historically strong connections to the old Liberal Party, the contest was between the SNP and the Liberal Democrats.

Against a divided opposition and with a large lead in national vote shares – 19.9 points in 2019 – a geographically well-spread voter coalition was a strength for the SNP, meaning they could win seats everywhere. With that lead flipped to a 5.3 point deficit, their voter coalition looks less ‘well-spread’ and more ‘thinly-spread’, meaning they could now lose everywhere. In every seat in which the SNP did not win last week, they came second to a party with a more geographically concentrated vote.

Scotland's Changing Electoral Geography

A map of the 2024 General Election results in Scotland
A map of the 2024 General Election results in Scotland

Electoral geography is part of the answer, but not the whole answer. The swing against the SNP was 15 points Scotland-wide, but this varied by constituency. The SNP’s swing was proportional, which means that – in general – the SNP lost a greater share of the vote in constituencies in which they had won a greater share of votes in 2019. 

A graph of the SNP Swing by Notional 2019 Vote Share


There were exceptions to this pattern. The SNP’s vote share held up better in the rural northeast and seats like Perth and Kinross-shire (their lowest swing) and Angus and Perthshire Glens. All of these seats have historically strong links to the SNP, covering areas the SNP won in 2005 and 2010, before the independence referendum.

There were also seats where the SNP’s swing was disproportionately large. Na h-Eileanan an Iar and Alloa and Grangemouth were clear outliers, and they lost more than 20 points in both Falkirk and Edinburgh South West. Losing more votes in the parts of the country where they were strongest compounded the weakness posed by the SNP’s geographically distributed vote share.

The Labour swing, unlike the SNP’s, was non-linear. It followed a generally ‘quadratic’ pattern, which means that they gained fewer votes in constituencies where they had the least chance of winning and in the constituencies where they had been strongest in 2019. But they gained votes at a disproportionate rate in ‘middling’ constituencies where those votes did the most work for them, maximising the value of each vote gained in terms of seat gains.

Again, there were some outliers here, which overlap with the SNP’s. Labour’s swing was highest in East Renfrewshire, Falkirk, Central Ayrshire, and Alloa and Grangemouth. All four are seats where Labour came from third place to win and, like the SNP’s positive outliers, partially or entirely cover areas represented by Labour before the independence referendum.

A graph of the Labour Swing by Notional 2019 Vote Share


The Conservative swing was also quadratic, focused in seats they had previously come in second, spread across the Central Belt. The key exception was Aberdeenshire North and Moray East, where Scottish Conservative leader Douglas Ross was defeated by the SNP.

A graph of the Conservative Swing by Notional 2019 Vote Share


The Conservative and Labour swings are related, with the Conservatives losing most votes in seats with a historical connection to the Labour Party, and where Labour were well-placed to defeat both the Conservatives and SNP. In contrast, the Conservatives lost fewer votes where they were better positioned to defeat the SNP than Labour – though they nevertheless saw a heavy swing against them in these constituencies, too.

The Liberal Democrat swing was proportional, like the SNP swing by running in a positive direction. They lost small numbers of votes across a very large number of seats where other parties were better positioned to defeat the Conservatives and the SNP. But they gained a large number of votes in the constituencies where they were strongest according to the notional 2019 results. The key positive outlier here was Inverness, Skye and West Ross-shire, where they more than doubled their vote on the back of a strong local campaign. Again, the Liberal Democrats’ best performances came in areas with the greatest historical links to the party.

A graph of the Liberal Democrat Swing by Notional 2019 Vote Share


Not only were the SNP systemically disadvantaged by their geographically dispersed voter coalition and the more efficient concentration of their opponents’ voters, but the swing against the SNP and towards Labour, in particular, could not have been better designed to convert the 15.7-point SNP-to-Labour swing into Labour gains.

In the process, Scotland’s electoral geography shifted. The SNP and Conservatives retreated into their historical heartlands and a more traditional electoral geography asserted itself, as Labour reclaimed the Central Belt and the Liberal Democrats rose to the top in North East Fife, Mid Dunbartonshire, and Inverness, Skye and West Ross-shire. 

Why did Scotland vote as it did?

The election in Scotland, as elsewhere in Britain, was fought primarily on the questions of competence and delivery. Pre-election polling found that just 15% of Scots thought the UK Government could be described as ‘competent’ or deserved to be re-elected, while 87% thought that ‘Britain needs a fresh team of leaders’. At the same time, 32% thought that the SNP Scottish Government deserves to be re-elected, and just 28% thought that government could be described as ‘competent’. 70% agreed that ‘Scotland needs a fresh team of leaders’.[i]

13% not agreeing that Britain needs a fresh team of leaders, and 30% not agreeing that Scotland does, are values eerily close to the Conservative and SNP vote shares. While 40% of Scottish Labour voters said that they were voting ‘in a way to best affect which party becomes the UK government’, and just 13% ‘to show my support or opposition to the Scottish government in Holyrood’[ii], it would be wrong to suggest that this election did not contain a degree of swing against the SNP on the grounds of their own perceived incompetence in addition to the desire to oust the Conservative UK Government.

This pattern of motivations, both anti-Conservative and anti-SNP, makes sense given the swings we see above. On one hand, in seats Labour could win, voters swung disproportionately away from the Conservatives and towards Labour, maximising Labour’s chances of winning those seats from the SNP and building a larger Labour majority at Westminster. In Conservative-held seats where Labour stood no chance of winning, voters were less likely to swing from the Conservatives to Labour. And the SNP swing suggests that, outwith their historic heartlands, 2019 SNP voters were perfectly happy switching to Labour to both win a Labour majority at Westminster and to punish the SNP at Holyrood.

Noticeable by their absence in this election were the issues of Brexit and independence. In 2019, 55% of Scots said Brexit was a ‘very important’ issue in deciding who to vote for, and 34% said independence.[iii] In this election, just 4% mentioned Brexit and 17% mentioned independence.[iv] In 2019, these two issues were key to the SNP voters coalition. 42% of voters fell into the pro-independence plus pro-Remain ‘tribe’, and voter identification with constitutional positions was far stronger than their identification with political parties.[v]

As the only major pro-independence and pro-Remain party, constitutional loyalties won the SNP the 2019 UK General Election in Scotland. The neutralising of those issues in the five years since deprived the SNP of the primary glue holding their coalition together, allowing issues around competence and delivery to re-assert themselves.

What next?

The results of the 2024 General Election in Scotland have redrawn the Scottish political map. With independence and Brexit off the agenda, and competence and delivery issues driven to the top, Scottish voters swung strongly against the country’s two incumbent governments in a manner that maximised the SNP’s losses and Labour’s gains.

The redrawn map is also largely to Labour’s benefit. Of the 25 seats now classified as ‘safe’, won by a margin of 15 points or more, twenty are held by Labour (the other five are Liberal Democrat). Of the thirteen classified as ‘leans’, won by between ten points and 15 points, ten are held by Labour (two by the SNP, one Conservative). And of the 19 marginals, seven are held by Labour and seven by the SNP.

In other words, to regain a plurality of Scottish Westminster seats, the SNP would need a uniform national Labour to SNP swing of around seven points – with the caveat that, as we’ve seen in this election, such large national swings tend not to be uniform across constituencies. Historical experience would suggest that such a swing is unlikely within a single Parliament, and that most Scottish Westminster constituencies are likely to be represented by Labour for at least the next decade.

The one caveat I would apply here would be to note that, since the independence referendum, the average Scottish constituency has flipped between parties 2.19 times and the nation has now been represented at Westminster by 183 different MPs.[vi] Taken alongside persistently weak party identification and the neutralising of constitutional issues that did attract voter loyalty, this all suggests a substantially more volatile electorate than we are used to – and therefore, greater potential for more ‘big swing’ elections than in the past.

A map of Scottish Westminster Constituencies by Marginality

But the next Scottish election is not to Westminster, but to Holyrood in May 2026. Those elections will come in 22 months, and barring a substantial shift in the political environment will be fought on the same grounds as the election last week – competence and delivery.

The same polls that correctly predicted the Scottish result currently have Labour and the SNP neck-and-neck, but that could shift in either direction. Several authors, including myself, have speculated that we may return to a pattern of more nationalist Scottish voters voting Labour at Westminster and SNP at Holyrood, a pattern we saw from 1999-2011. But that was during a period of Labour government and SNP opposition, and a competence election with the SNP 19 years into government and struggling with high levels of dissatisfaction does not seem like the ideal ground for the resumption of such a voting pattern.

This election was a cataclysmic defeat for the SNP at Westminster and is a shot across the bow of their Holyrood group. The voting system at Holyrood means that a replication of this national vote at that election would not mean as heavy a defeat, but it would almost certainly eject the SNP from power.

What happens next depends on how the SNP and Labour react to this result. In the SNP, internal recriminations have already begun. Labour, on the other hand, have the advantage and the political momentum to press that advantage home. A Labour victory and return to power in Edinburgh in 2026 currently seems like the most likely outcome.

[v] The Referendum That Changed a Nation, p. 157

[vi] Thanks to Jonny Kiehlmann for the calculations here.

MPs Job-Share: A Practice Worth Trialling

Northern Ireland and the 2024 General Election Results

A smart electorate makes its views known

Election 2024 analysis: The future for Labour in Wales

How Scotland Voted in the 2024 UK General Election