An absolute (yellow) belter
Scottish results could be absolutely crucial to the formation of the next government, and a 'yellow belt' of potential SNP vulnerability may determine the outcome.
Of all Britain’s political corners, perhaps none are so electorally volatile, fascinating, and potentially significant as Scotland. Particularly so since 2015, when a tidal wave of yellow votes won the Scottish National Party (SNP) 56 out of the country’s 59 constituencies.
Their advance was significantly rolled back in 2017, when Nicola Sturgeon’s party were reduced to 35 seats following gains for each of the Conservatives, Labour, and the Liberal Democrats. Then comes along another General Election in 2019, and the SNP recover and move forward once again, capturing 48 constituencies and building fresh momentum for a second referendum on Scottish independence.
The constant to-ing and fro-ing of Scottish electoral fortunes has left us with a nation where 22 of the 59 constituencies are currently held with majorities under 10% (requiring just a five-point swing to change hands), and a whole host of incumbents with short records of electoral success. Outside of Yorkshire and the Humber, nowhere else in the United Kingdom contains a higher proportion of marginal seats.
Given how many seats are ‘up for grabs’ north of the border, not enough attention is being paid to what could potentially be a crucial set of constituencies at the next election. Especially when we consider that current models of vote intention polls put the Labour Party ahead in terms of Commons seats, but very much in need of third party support to govern.
If this pattern holds and comes to fruition by the time voters next head to the ballot box, then the outcome in Scotland becomes potentially crucial to the makeup of the next government.
Today’s post will focus on how the political battlefield might shape up in Scotland at the next General Election, examining results in the existing constituencies and also looking ahead to the new proposed boundaries as outlined in the Boundary Commission for Scotland proposals.
To start, we will consider where the ‘pro-Union’ (anti-Scottish independence) parties should cast their eyes to make seat gains and push SNP back once more. We’ll take a look at a ‘yellow belt’ of constituencies across the centre of Scotland where the Nationalists are particularly vulnerable.
The second section will explore seats where the Scottish Nationalists could look to target and break the 50 constituencies mark again, pushing further their case for a second referendum.
State of the union
For the pro-Union parties (the Conservatives, Labour, and the Liberal Democrats)1, there are a couple of ways we could look for potential gains at the next election. An obvious starting point for them would be to try and retake the new equivalent seats for each of the 14 constituencies that the SNP gained in 2019.
However, some of these constituencies are set to look more SNP-friendly after the review than they currently sit, making that particular task just that little bit trickier. Others, such as Glasgow North East and East Dunbartonshire, are being removed altogether and spread across other existing SNP territories.
Nonetheless, high on such a ‘strike back’ list for Labour would be the new East Lothian Coast and Kirkcaldy and Cowdenbeath constituencies. Meanwhile, the Conservatives would be taking a long look at Ayr, Carrick & Cumnock, Aberdeen South, and East Renfrewshire. For the Liberal Democrats, the new Highlands North seat is mostly comprised of their existing Caithness, Sutherland, and East Ross seat, which makes it an obvious starting point for Ed Davey’s party.
But what if we take a more nuanced approach, along the same lines as James Kanagasooriam used to identify ‘red wall’ constituencies, and look for places where the pro-Union vote might be ‘underperforming’ in terms of where the underlying support for anti-Scottish independence parties might be?
If we combine details from the existing seats regarding (a) current majorities, (b) political and demographic predictors of the combined ‘pro-Union’ vote, and (c) the gaps between second and third placed parties, we can build a list of short and longer-term target constituencies on new and existing boundaries for each of the Conservatives, Labour, and the Liberal Democrats.
Loose links in yellow belts and braces
Firstly, some key factors tend to correlate well with SNP electoral performance in elections since 2015. For instance, the (estimated)2 ‘Yes’ vote from the 2014 Scottish independence referendum is a positive predictor of SNP vote share, while a higher prevalence of older voters in a constituency is a negative predictor.
If we reverse this logic and look at factors which should predict larger anti-SNP votes – or, in other words, a strong ‘pro-Union’ vote – we can then identify areas in which the combined anti-independence vote may be underperforming (relative to those predictions). These are constituencies where, in effect, the SNP vote has significantly outgrown the latent support for independence (at least, according to 2014 referendum levels).
Secondly, we can evolve slightly the demographic-based model to trim down the ‘yellow wall’ model into a more achievable list of (short and longer-term) targets. Namely, while the anti-independence vote may be underperforming slightly in a given constituency, we probably should not consider the seat to be particularly vulnerable if the SNP enjoy a majority approaching 30 points (or more).
However, while the SNP do indeed have healthy majorities across many Scottish constituencies, to bring in an idea developed by Hannah Bunting in her PhD thesis, marginality is but one component in how we ought to define a ‘competitive’ seat.
While the effectiveness of tactical voting in terms of altering electoral outcomes is often played down, there is convincing evidence from the 2021 Scottish Parliament election to suggest that supporters of anti-independence parties are willing and able to vote tactically to try and deny the SNP a win at the constituency level.
With that in mind, we can include non-marginal constituencies where the Union vote is underperforming as potential targets if there is (1) a clearly identifiable second placed pro-Union party, and (2) a sizeable vote left to squeeze between the remaining two pro-Union parties.
We can then combine this list with a more traditional target list of marginals (those which would require less than a five-point swing to change hands) and produce a total of 20 constituencies in which the SNP would be vulnerable (to varying degrees) if an election were held on current boundaries.
This ‘yellow belt’ of constituencies runs from coast to coast across central Scotland, stretching from Gordon up in the Northeast to Argyll in the West. It branches down through Glenrothes and Edinburgh to reach Lothian in the Southeast, and includes pockets of Glasgow its surrounding counties in the centre. Ayr, Carrick, and Cumnock – a previous Labour stronghold briefly held by the Conservatives from 2017-2019, is its South-westerly most point.
Taking into account the proposed changes, the ‘yellow belt’ seats - broken down by which pro-Union party is (currently) best placed to take them - for an election held on new boundaries would shape up like this:
‘Yellow belt’ seats on new boundaries by (notional) second placed party
Of course, these ‘best placed’ metrics are based purely on the party notionally second in the new constituencies as it stands now. As with so much in British politics right now, things can change quickly. We may find for instance that Labour establish themselves as principle challengers in Renfrewshire and Edinburgh, or that the Liberal Democrats return as the primary challengers in the North and Highlands. We shall see.
Can Nicola turn the yellow screw?
For the SNP, they finished as runners up in each of the 11 constituencies they didn’t win in 2019. That’s right – there is not a single constituency in Scotland in which the Scottish Nationalists are not in either first or second place. There are no Con-Lab or Con-Lib Dem marginals on Scottish soil.
The proposed boundary changed would already see two such constituencies where the SNP are second placed removed and integrated with a seat they did manage to win.
This includes Scotland’s second-tightest marginal, Caithness, Sutherland & Easter Ross (Lib Dem majority of 0.6%), which is forming part of the new ‘Highlands North’ seat. Notional results3 already hand that new constituency to the SNP. Elsewhere, another current Lib Dem seat which Sturgeon’s party narrowly lost in 2019 is coming back (notionally) into their fold – Fife North East.
If implemented, the boundary review would leave just nine seats in Scotland not notionally held by the SNP.
Top of Nicola Sturgeon’s post-review hit list will be two Conservative seats: Gordon and Moray South (notionally Conservative) and Aberdeenshire West & Kincardine (set to be unchanged in the review, current Conservative majority of 1.5%).
Beyond that, the constituencies of Dumfries and Galloway, Edinburgh West, and Dumfriesshire, Clydesdale & Tweeddale appear next in line on the 2019 majority list. Each of them will change slightly in the boundary review, but not by much.
Of these, arguably the Liberal Democrat held Edinburgh West provides the best opportunity for an additional gain, with a much smaller estimated Leave vote (29%) and a higher estimated Yes vote in 2014 (39%) than the other two. David Mundell (MP for Dumfriesshire, Clydesdale & Tweeddale since 2005) for instance has already proven particularly tricky to unseat, having survived the initial 2015 wave and the 2019 SNP rebound.
On a particularly bad night for the Government though, the SNP could well capture the two Dumfriesshire seats above and indeed all six of the remaining Scottish Conservative constituencies. On something around a five-point swing, each of them would be likely to fall - even on new boundaries.
Looking past those Conservative constituencies, only two seats would remain in the way of an SNP ‘clean sweep’ of Scotland - Edinburgh South (Labour’s sole Scottish seat), and Orkney and Shetland. Both remain quite unlikely Nationalist gains in the near future - though the latter perhaps somewhat less so.
Of course, these are not the only pro-Union parties who contest elections in Scotland. They are however the largest and only anti-Scottish independence parties to currently hold parliamentary seats in Scotland.
These estimates come from BBC Election Team data, and are derived from models produced by David Denver and Jonathan Mellon. The original counting areas for the referendum results were local authorities, rather than Westminster constituencies, hence the need to model and estimate the Yes vote at the constituency level.
Notional results are calculated by modelling what the results would have been at the previous election had they been fought on new (or otherwise different) boundaries. Notional results are extremely useful and important in guiding our expectations for starting positions and predicted vote shares for each of the new parties. Examples for the new proposed UK constituency boundaries include Electoral Calculus and Election Maps UK.