GE19 – The Difference a Boundary Makes

Though I was done with General Election coverage, eh? Well, so did I! I’m back with one final post, this time on the effect of prospective new Scottish Westminster boundaries. Scotland’s constituencies date back to 2005 (and use wards drawn in 1999 as building blocks) and have been overdue a refresher for a while now. I’ve written in the past about how the need for boundary changes is one of the big flaws in FPTP, and indeed contradicts some of the key arguments in its favour. This post will refrain from re-hashing that, but yes, a lot of the new boundaries are truly awful.

The past three UK Governments have tried to change boundaries UK wide, and cut the number of MPs from 650 to 600, to no avail. The Coalition’s attempt failed when the Lib Dems withdrew support for the proposals as tit-for-tat when enough Conservatives rebelled to prevent Lords Reform. The brief 2015-17 Conservative majority started the process over again, but Theresa May’s ill-judged snap election gamble came before the final report. When that report came at the end of 2018, the fact the DUP would lose out in changes that cut an MP from Northern Ireland effectively put the brakes on them given the Conservative minority was reliant on DUP support.

Now that there’s a much more comfortable Conservative majority which is unlikely to be brought down by Brexit, it’s likely changes will finally go through. Whether they’ll use the final proposals from last year or start the review over a third time, I’m not sure. Anyway, although the likelihood of boundary changes was in my mind over the past week, I hadn’t expected any translation of last weeks results onto alternative boundaries anytime soon. As it turns out, Electoral Calculus were very quick to do exactly that using the 2018 proposals.

I’ve used that excellent resource to inform this post (I may do my own calculations later, just for comparison), but do bear in mind the caveat that (as with my stuff on PR) this assumes every vote is identical to last week. Different boundaries wouldn’t prompt voter behaviour as dramatically different as PR would, but it would still be a bit different as party strategies and voter tactics would change to suit. The main thing to look at is of course the headline seat shares (vs current boundaries);

  • SNP – 47 (-1)
  • Conservative – 5 (-1)
  • Lib Dem – 1 (-3)
  • Labour – 0 (-1)

Notably, although Scotland drops 6 seats overall, the SNP would only have been one MP down. The Conservatives also lose a single seat, as do Labour for whom it’s rather more disastrous as that means they’d have none. In absolute terms the Lib Dems are hardest hit, losing all three of their mainland seats to be reduced to just Orkney and Shetland again. Let’s take a look at how things look for the non-SNP parties – there’s not much point in looking at where the SNP come first or second since that’s “literally every constituency in the country”.


The lost Conservative seat comes from the North East of the country. Boundary changes here effectively merge most of the Gordon constituency (SNP) with West Aberdeenshire and Kincardine (Con) to create a Gordon and Deeside seat that would go Conservative, and they’d hold an extended Banff and Buchan which takes in some of the rest of the old Gordon seat. They’d lose out in the Moray and Nairn constituency, as their tight 513 lead over the SNP in the current Moray is narrowly overturned by bringing in a few thousand voters from more SNP favourable Nairn.

Staying in the North East, extending the Aberdeen South boundary slightly northwards also counts against the Conservatives, whilst the Kincardine and Angus North seat would be a prospective marginal for them. The Angus South and Dundee seat would be well out of their reach however. Moving a bit further south, they emphatically lost the Stirling constituency last week and, although keeping it marginal, came substantially behind the SNP in Ochil and South Perthshire. With the urban portion of each of those seats (Stirling city and Clackmannanshire, respectively) being shorn off to leave a rural Stirlingshire, Strathearn and Kinross seat, they’d be a much tighter second there.

Finally, they’d effectively hold all three of their current border seats. Berwickshire, Roxburgh and Selkirk doesn’t see huge changes, nor does a (re-ordered) Galloway and Dumfries. Substantial changes however to what is currently Dumfriesshire, Clydesdale and Tweeddale. It loses the Tweeddale portion to become Dumfriesshire and Lanarkshire South East, which extends up to Carluke and Lanark. That change makes it a much closer marginal with the SNP, putting in play the prospect of Scotland’s longest serving Conservative MP losing his seat. Also in the south, a re-drawn Ayr and Carrick would be a closer marginal than the current (and Cumnock) seat is, thanks to trading the Doon Valley for Troon and Prestwick.


If Labour didn’t think they could do any more disastrously than they did last week, these boundary changes would prove them wrong. Edinburgh South is substantially redrawn such that the new version is only about half made up of areas from the current seat, with the other half of the constituency being drawn from parts of the existing South West and East seats. That tips the balance in the SNP’s favour, though as a marginal seat. The flip side is that as the remainder of the current South moves into a new East, that also becomes a marginal. Nearly winning two seats isn’t as good as actually winning one, however.

Elsewhere, East Lothian actually has no boundary changes whatsoever, and those that turn Rutherglen and Hamilton West into Lanarkshire West are so mild they barely affect the distribution of votes. That means those seats would be about as easy (or hard) to win back as they are at the moment. The new Midlothian and Upper Tweeddale (sub for Midlothian), Fife South and Mid Fife seats (Kirkcaldy and Cowdenbeath split pretty evenly between these) and Glasgow North East all become more strongly SNP, reducing Labour’s chances there too.

Liberal Democrats

All of the Lib Dems’ mainland seats are highly marginal, so it’s no surprise that boundary changes that require all of those seats to expand don’t help their chances. Edinburgh West is the closest run, as changes there effectively result in an 80:20 balance between the current West and areas from other more SNP inclined Edinburgh seats. In an actual election, I’d expect this to be the one the Lib Dems would fight strongest to hold onto as a result, so wouldn’t write it off.

North East Fife may look like a very small change on the map, but that belies how many new voters are added to it. The Levenmouth area of Fife is very strong for the SNP and the Lib Dems barely exist there. At present only a small chunk of it, Leven is part of the constituency. The proposals would expand to include the whole of Levenmouth, which therefore flips the seat to the SNP. Highland North covers the existing Caithness, Sutherland and Easter Ross seat that the Lib Dems have a very slender 0.6% margin in, so it’s no surprise bringing in a large chunk of SNP-favourable Wester Ross causes the seat to change hands.

And to finish up, what about that dramatic loss in East Dunbartonshire? Well, even had Swinson had won last week, that might just have delayed her defeat until these boundary changes.

The current East Dunbartonshire seat doesn’t include the Campsies or a big chunk of Kirkintilloch, which are included in the council area and aren’t as Lib Dem inclined as the Bearsden and Milngavie area is. Those areas would be incorporated into a redrawn Dunbartonshire East, and to make matters worse, a small portion (perhaps a quarter) of Bearsden transfers to a Dunbartonshire West seat, taking Lib Dem votes with it. The seat would still be marginal, but right at the upper end of that scale rather than the most marginal seat in the country. One that, all things considered, I’d doubt the Lib Dems’ ability to take.

Whether these boundaries are the ones we end up with for the next election or not, they are a reminder of the fundamentally undemocratic nature of FPTP. For the same votes to produce such dramatically different MPs based on where the lines are drawn should be unacceptable – though as ever I note my gratitude for the hard work of the Boundary Commission, who draw these to the best of their ability and not to benefit any party. It isn’t their fault the system and rules they have to work with are dire!