A key feature – and perhaps, flaw – of First Past the Post elections is watching the competitions in swing seats. Ahead of this election, I ran my Ballot Box Battlegrounds series investigating the 16 seats that, based on 2016 results, were most likely to change hands. These were the seats requiring a swing of 5% or less between the top two parties to flip. As we know, very few of these seats ended up with a new party in place after the election. The SNP were the only party to gain any, and even then just three of them.
As quiet as the overall results may have looked in those terms, there were still plenty of interesting things going on in other seats. Every election sets the stage for the next, so as the last piece of major analysis of the results, it’s worth looking at what will be the key battlegrounds in 2026 – albeit we can probably expect some boundary changes before then.
The Obligatory Map Chat
Starting with the map on the left which shows whether seats became more or less marginal, there are some patterns to pick out. The seats which remained out of the SNP’s hands this election by and large became safer, perhaps indicating some degree of coalescence of the pro-Union vote behind the incumbent. Shetland however clearly bucked that trend, as what was formerly one of the safest seats in Scotland plummeted into the realms of the marginal. Ettrick, Roxburgh and Berwickshire was the other exception, though it remains a pretty safe seat.
Unsurprisingly given their dominance of the constituency vote, seats that the SNP’s share increased in correlate pretty well to seats that have become less marginal. The exceptions here are in the South – some of that obviously relates to them narrowly gaining two constituencies here, but even in the seats they held they lost some ground to the Conservatives.
Moving now to the right hand side, and the “not much change” nature of this election is shown by the fact we had 16 marginals before the election, and still have 16 now. 12 of these are the same seats, whilst the other 4 have been shuffled out for new marginals. Again, we see a slight strengthening of position in the remaining non-SNP seats, as 7 of the 11 (two-thirds) are now marginal, versus 10 of the 13 (three-quarters) previously.
That’s tipped the balance of who is most at risk in the marginals from the other parties to the SNP, who now account for 9 of the 16. Of course, the risk here is relative – even if the SNP did lose all of those seats, and failed to make any of them back up via the lists, they’d still be easily the largest party at Holyrood.
Lopsided Tactical Voting
Taking both sides of the image into consideration, we can probably say the SNP are now mostly at the upper limits of what constituencies they can credibly win. We heard a lot during the count and afterwards about tactical voting in support of non-SNP incumbents, and I can’t see that weakening over the coming years. In fact, Shetland is the only one I’d say they are at all likely to pick up – though that would be a momentous political moment nonetheless.
On the other hand, predictions of the SNP’s imminent demise have been circulating since they first took office in 2007, so it’d be foolish at this point to say they are set to go backwards in 2026. They may also benefit from the fact that tactical voting in seats they held seemed to work out a lot more in Labour and the Lib Dem’s favour than it did for the Conservatives.
There are quite a few places where the Conservatives made headway against the SNP, but lots of these weren’t seats they were likely to or needed to win. These included Clydesdale; Carrick, Cumnock & Doon Valley; Clackmannanshire & Dunblane; Cunninghame North (a lot of C’s here); and Stirling. By contrast, the picture in key targets and existing marginals was pretty mixed.
They made massive gains in the north of Aberdeenshire, slashing big SNP majorities in Aberdeenshire East and in Banffshire & Buchan Coast into marginals, and likewise made up a bit of ground in Moray. But in the Perthshire and Angus constituencies, the SNP reinforced their position – even though the Conservatives had won or come close to winning all of the equivalent seats here in the 2017 UK election.
More successfully, the Lib Dems are now pretty immovable in three of their four constituencies. They also made gains in Caithness, Sutherland & Ross at a time they were slipping behind the Conservatives in the rest of the Highlands. That big decline in Shetland has very little to do with a lack of tactical voting – Labour and the Conservatives were already below 10% of the combined vote in 2016. Instead, it suggests a genuine and large shift in support directly to the SNP – or, perhaps, to the SNP’s candidate, Tom Wills.
And for Labour, the piling up of tactical votes came in exactly the places it needs to. Almost every constituency in Central, most in Glasgow, and a few key seats in West are now more marginal than before this election. That suggests to me that Conservative voters were perhaps more content to vote Labour in constituencies than the other way round. However, this hasn’t actually put Labour in striking distance of gains – these seats are more marginal, sure, but they haven’t actually become marginal.
Final State of Play
Finally, let’s look at the precise list of 16 marginals we now have. The seats that are newly marginal at this election are highlighted in bold.
- Ayr, SNP over Con 0.4% (was 3rd, Con over SNP 2.0%)
- Banffshire and Buchan Coast, SNP over Con 2.3% (wasn’t marginal, SNP over Con 23.0%)
- East Lothian, SNP over Lab 2.6% (was 6th, Lab over SNP 3.0%)
- Dumbarton, Lab over SNP 3.9% (was 1st, Lab over SNP 0.3%)
- Aberdeen South and North Kincardine, SNP over Con 4.3% (was 14th, SNP over Con 8.5%)
- Perthshire South and Kinross-shire, SNP over Con 4.4% (was 8th, SNP over Con 3.9%)
- Aberdeenshire East, SNP over Con 4.6% (wasn’t marginal, SNP over Con 16.8%)
- Eastwood, Con over SNP 5.2% (was 9th, Con over SNP 4.4%)
- Shetland Islands, LD over SNP 6.8% (wasn’t marginal, LD over SNP 44.3%)
- Caithness, Sutherland and Ross, SNP over LD 7.0% (wasn’t marginal, SNP over LD 12.1%)
- Galloway and West Dumfries, Con over SNP 7.1% (was 10th, Con over SNP 4.5%)
- Moray, SNP over Con 7.7% (was 15th, SNP over Con 8.6%)
- Aberdeenshire West, Con over SNP 8.1% (was 4th, Con over SNP 2.6%)
- Edinburgh Southern, Lab over SNP 8.9% (was 5th, Lab over SNP 2.9%)
- Dumfriesshire, Con over SNP 9.9% (was 7th, Con over SNP 3.4%)
- Angus North and Mearns, SNP over Con 9.9% (was 13th, SNP over Con 8.4%)
Of the four new additions to this list, it’s worth pointing out that two of them were in the Bonus Round of my Battlegrounds series. That covered seats that weren’t strictly marginal but which could still change hands. Banffshire & Buchan Coast in particular looked numerically like a “safe” seat but I was right to expect it to be more hotly contested this time. Aberdeenshire East is a rather unsurprising addition to the list, but Shetland has taken me by surprise, given the by-election wasn’t marginal.
Another notable change here is how safe Edinburgh has become. In 2011 there were a lot of very close calls there, and in 2016 four of the six seats in the city were marginals, including the second closest seat in the country – or in other words, a quarter of all marginals were in Edinburgh. This time the capital was home to just a single marginal, and it’s a pretty secure one as marginals go, both because it’s 14 out of 16, and because Labour are clearly well embedded in the south of Edinburgh by this point.
Of the three Edinburgh seats that have ceased to be marginal, only Western is massively safe – a stonking Lib Dem majority of 21.1% there, whereas Central is 11.3% and Pentlands is 10.2%. That does however mean Central is the only one of the SNP’s gains not to be a marginal. The other absentee from the list this time, Perthshire North, only just squeaked out – John Swinney’s majority extending to 10.1% there, up only slightly from the 9.8% last time.
And that’s a wrap on the 2021 Scottish Parliament Election! Or, rather, a wrap on pieces about what actually happened. Long time followers of BBS will know that I’m fond of a hypothetical “what if the voting system was different” piece – that will come in due course.