LE22: Constitutional Christmas Cracker

As one last little bonus for both 2022 and specifically May’s local elections, it’s time for a little treat I’ve been teasing on Twitter. Hopefully most of you are lazing around contentedly, perhaps a bit too full of chocolate, definitely unwilling to see another turkey (or nut roast) for at least a month. So what better time to kick off another round of constitutional bunfighting?


What This Is

An estimate of the balance between Pro-Union and Pro-Independence parties in each polling district (well, most of them) in Scotland at the May 2022 Local Elections, after eliminating Independent and minor party candidates and re-distributing their preferences as recorded. Islands councils are not included, as none of their wards pit parties from opposing camps against one another. For obvious reasons, the three mainland wards that were uncontested also aren’t included.

What This Is Not

An estimate of the support for the Union vs Independence in May 2022, at any level. It’s also not an estimate of support for the Union vs Independence now, or in 2014.

Why Not?

Basically, although it’s everyone’s favourite game to play constitutional proxy by tallying up votes between camps, it’s not as simple as that. For one thing, although the link between party and constitutional preference is strong, it’s not perfect. As investigated in this piece, the two least “settled” voter blocs split 4.3 (Green) and 4.8 (Labour) to 1 in agreement with their chosen party’s constitutional position at the Holyrood election last year, which still represents a significant number of voters.

We might expect those figures to be slightly less rigid at local level, given the greater importance of personal vote and campaigning. Indeed, my own personal favourite anecdote from standing for election in 2017 is the woman who on polling day marched up furiously to ask me “are you the only candidate standing?! I’ve had NOTHING through from anyone else! I’ll HAVE to vote for you then!” That’s not a voter thinking deeply on the constitution – or any policy area!

For another, elections aren’t just about how people vote, but whether they vote at all. Around about 85% of the electorate turned out for the 2014 Independence referendum, an astronomical level by UK standards. Last year’s Holyrood election was closer to 65%. This May’s local elections were in the region of 45%, not much more than half of the referendum turnout. 

We know that turnout isn’t the same across all groups of voters, but instead varies between them. Generally speaking, poorer and younger voters are less likely to vote, whereas wealthier and older voters are more likely. When turnout is higher, it means more of those low-turnout groups have voted. In the context of Independence, do you know how voters tend to split? You got it, poorer and younger voters tend to back Independence, wealthier and older voters tend to back the Union. That means of all the elections to treat as a proxy on the constitution, local elections are the least useful.

For a live example of the difference turnout can make, let’s look at the Glasgow North East seat in the UK Parliament, which swung very narrowly back to Labour in 2017. First here’s the 2015 vote:

Now the 2017 figures (note that I’ve deliberately scaled to number of votes rather than proportion, to illustrate the point):

As you can see, although Labour gained +9.2% of the vote in percentage terms, they only picked up 883 additional votes. That’s only around a tenth of the 8,581 the SNP lost. Even allowing for some growth for the Conservatives and Lib Dems, in total there were 6082 fewer votes cast than there had been in 2015. There’s an obvious conclusion to draw: a lot of SNP voters, energised by the referendum in 2015, were less fussed in 2017 and didn’t feel the need to turn out at all. That doesn’t mean they stopped supporting Independence or, crucially, that they wouldn’t turn out at a referendum to do so.

Particular Oddities

Overall, allowing for the absent areas, the estimated split between party camps ends up somewhere in the region of ~56-57% Pro-Union to ~43-44% Pro-Independence. That’s a slightly wider gap than was the case according to polling averages in May (which were about 47-48% Yes, after excluding Don’t Knows), illustrating those points about imperfect party vs constitutional alignment and turnout.

According to the estimated figures, this would mean Dundee, North Lanarkshire and West Dunbartonshire tipped for pro-Union parties, whilst Inverclyde did so much more heavily than their 2014 No vote when it was less than 100 votes away from backing Independence, even with the ward that would have swung it for the Union then being uncontested in May. Demographically speaking that’s deeply improbable, and if you think Dundee of all places was (narrowly) Pro-Union in May, I have some magic beans to sell you.

In the other direction, Falkirk actually ends up with a narrow Pro-Independence party lean despite a No lead in 2014, whilst Clackmannanshire and Angus are a hair away. Further areas like Argyll and Bute, East Renfrewshire, Perth and Kinross, and the Scottish Borders also had the Pro-Independence parties notably outperforming their 2014 Yes votes. Again, if support for Independence was actually lower in May than it had been in 2014, you’d really not expect such results.

Of course, we should expect that since 2014, different areas will have had a range of internal shifts. For example, there’s evidence of a greater alignment of 2016 Remain and Leave voters with Yes and No than at the time. That would mean we would expect heavily Pro-EU areas like Edinburgh and East Renfrewshire to be slightly more Pro-Independence than in 2014 even with another 45:55 result, whilst places like Moray and Dumfries and Galloway with higher Leave votes than average might have hardened for the Union. We could only possibly know for certain the degree to which this has happened with a referendum, however!

Okay, We Get It...

“You can’t fully treat any election results as a proxy for the constitution, can we see the maps? We’ll behave!”

No, you absolutely won’t, but I’ll show you the maps anyway, because they are fun. Just keep all of these caveats in mind and try not to be too daft.

Bring On The Maps!

Note that throughout this section, particular attention will be drawn to areas with an estimated Pro-Independence lean. This isn’t some nefarious ploy to big that side up, but instead reflects the fact that as the initial minority side, it’ll be the one that will have areas “swing” to it as we go through the maps.

Initial Estimate

Top 5 mainland council areas for Pro-Union parties: Dumfries and Galloway, Aberdeenshire, South Ayrshire, Scottish Borders, East Dunbartonshire.

Top 5 mainland council areas for Pro-Independence parties: Glasgow, Falkirk, Dundee (Pro-Union lean), Angus (Pro-Union lean), Clackmannanshire (Pro-Union lean).

Top 5 major towns for Pro-Union parties: Newton Mearns, Bearsden, Ayr, Elgin, Rutherglen.

Top 5 major towns for Pro-Independence parties: Cumbernauld, Alloa, Coatbridge, Arbroath, Falkirk.

Additional major towns with estimated Pro-Independence party lead: Glasgow, Glenrothes, Livingston, Motherwell.

This is the map as spat out by the initial estimates. Looking sub-council but more broadly than major towns, you can see particular concentrations of votes for the Pro-Union parties in Annandale, Roxburghshire, Caithness, North East Fife, Helensburgh and Lomond, western Edinburgh, deep inland Aberdeenshire and Angus, and the rural components of Lanarkshire, Ayrshire, Renfrewshire. Doing the same for Pro-Independence parties, you can pick out areas like northern Argyll, Skye, western Falkirk, the eastern ends of the cities of Edinburgh, Aberdeen and Inverness, plus northern Perth.

One really useful thing this map does, even though it’s indicative rather than authoritative, is show that of the mainland council areas, there isn’t a single one that doesn’t have at least a handful of districts which go against the overall local trend. Even Dumfries and Galloway has some areas, particularly in Dumfries, which showed a lead for Pro-Independence parties. Similarly, Glasgow and Dundee, the poster-children of 2014’s Yes vote, have areas which give the advantage to Pro-Union parties.

This should be a reminder, for those somehow still in need of it, that suggesting Scotland should be partitioned on the basis of any future constitutional referendum is an indescribably stupid idea. Leaving aside the various other problems with the plan, far be it from splitting off unified “Pro-Union” and “Pro-Independence” council areas, there’d be large and almost certainly pretty angry minorities in every part of the country fuming about their neighbours.

We can also compare with the Scottish Index of Multiple Deprivation (SIMD) to confirm the general point that more deprived areas tend to be more likely to vote for Independence. The relationship isn’t perfect – these things never are – but there are plenty of examples. The one Pro-Independence party lead in Bishopbriggs corresponds to Auchinairn, the part of the town with the highest levels of deprivation. The Pro-Union party led sections of Kilmarnock North cover the least deprived areas. Dundee in particular maps extremely obviously in this way, with almost every part of the city in the least deprived half seemingly leaning towards Pro-Union parties, and every part in the most deprived half the opposite way.

As noted earlier however, this is likely to be an underestimate of actual Independence support at the time of the election. With caveats mentioned earlier firmly in mind, let’s introduce some swing.

With +5% uniform swing to Pro-Independence parties

To start with, adding 5% takes us to a point where there’s a relatively narrow Pro-Union lead still. That’s broadly equivalent to, though it won’t be exactly, where support seemed to be lying in May.

Mainland council areas that newly swing for Pro-Independence parties (running total 10): Dundee, Angus, Clackmannanshire, North Lanarkshire, West Lothian, Argyll and Bute, Midlothian, West Dunbartonshire.

Major towns that newly swing to Pro-Independence parties (running total 23): Dundee, Musselburgh, Renfrew, Paisley, Clydebank, Kilmarnock, East Kilbride, Irvine, Cambuslang, Perth, Greenock, Airdrie, Inverness, Bathgate.

At this point, the Central Belt in particular has become a lot more blue. Falkirk and Clackmannanshire for example are already almost entirely Pro-Independence party leaning, whilst Glasgow has also lost a large portion of the Pro-Union party-leaning districts. Even in council areas that lean towards the Pro-Union parties, increasingly blue tints are evident in large parts of Highland, Dumfries, and northern areas of Aberdeen, East Dunbartonshire and East Renfrewshire. The deepest Pro-Union tints remain in Annandale, Roxburghshire and north western Caithness.

Again though, there’s a degree to which this could be an underestimate, but instead against current polling which is showing a relatively narrow lead for Independence.

With +10% uniform swing to Pro-Independence parties

If we add another 5% swing (for 10%) in total, we get a narrow Pro-Independence party lead, roughly in line with actual polling at time of publication (52.8%, per the arbitrary five-poll average).

Mainland council areas that newly swing for Pro-Independence parties (running total 23): Renfrewshire, Highland, Inverclyde, Perth and Kinross, North Ayrshire, East Ayrshire, Fife, Stirling, Aberdeen, South Lanarkshire, Moray, Edinburgh, East Renfrewshire

Major towns that newly swing to Pro-Independence parties (running total 35): Peterhead, Stirling, Dumbarton, Wishaw, Kirkcaldy, Hamilton, Dunfermline, Aberdeen, Dumfries, Bishopbriggs, Edinburgh, Kirkintilloch

The blue for Pro-Independence parties is now almost overwhelming in the Central Belt and southern Fife, with primarily the most rural components holding out, and most of those pretty faintly. Only Newton Mearns, Harthill, Helensburgh, Larkhall, Kilwinning and southern and western Edinburgh stand out as remaining areas of estimated Pro-Union party strength in the Central Belt, though Caithness, Annandale and south-central Aberdeenshire still generally retain their dark red tints.

Councils with Pro-Union party leads are nonetheless generally shot-through with Pro-Independence party districts now, particularly in coastal Aberdeenshire, Dee and Glenkens, Tweeddale, Clarkston and the north of Ayr. Even in Roxburghshire, which is slightly lopsided due to the Kelso ward not having had an SNP candidate at all in May, Hawick estimates to a narrow Pro-Independence party lead.

I have to say that at this point I’m pretty sure some of the estimates are outstripping what would happen in a real referendum – I’m not sure that the likes of Aberdeen or East Renfrewshire would swing to Independence if it won by less than the Union did in 2014! As noted earlier, that’d likely be counterbalanced by stronger leads for Independence in areas including Dundee, North Lanarkshire and West Dunbartonshire than this map indicates.

Effectively, where a real referendum would differ from this is that swing wouldn’t be uniform. Those areas that have Pro-Independence party leads also generally had lower turnout, and so would likely experience gains of more than the +5-10% range here as heavily Pro-Independence voters turn out. By contrast, the most Pro-Union leaning wards often already have higher turnout, so would have gains lower than +5-10% as they’d have fewer additional voters turning out.

As with all maps which simply colour geographic areas according to partisan leans, these three maps have all had the issue of big rural districts being visually dominant. Just by the very nature of polling districts, generally speaking the smallest contain the most voters and largest the least.

Mapped to Population Centres

So for one last little bit, let’s do the same kind of results and buildings overlay as I did with party-based data earlier in the year. This version takes us back to the initial estimate, which had the substantial Pro-Union party lead. For simplicity, because it looks a lot more awkward when using shades to convey strength this way, this simply shows whether a district leaned towards Pro-Union or Pro-Independence parties.

This moderates some of the deeply coloured rural areas, on both sides – think Annandale for the Pro-Union parties or northern Argyll for Pro-Independence. The Central Belt closeup in particular gives a much better indication of how divided the main population centres are, without large stretches of their sparsely populated rural surrounds visually crowding out the smaller but more dense urban areas.

Remember, you promised to behave...

Overall, what these maps give us is the general shape of support for Pro-Union versus Pro-Independence parties across the country in May’s local elections. As discussed earlier, due to the imperfect relationship between party support and constitutional preference plus much lower turnout, these are not precise estimates of how areas would vote in an actual referendum. Even if they were, they serve as a reminder that most parts of the country have very large minorities that lean in the opposite way to the local norm. Just like elections, just because a given area is associated with a particular partisan colour, doesn’t mean everyone that lives there voted that way.

And with that, I can now finally say that’s a wrap on 2022 at Ballot Box Scotland! It’s been a very busy but very enjoyable year, and I hope people have found my coverage useful and interesting. Have a very Happy New Year when it comes, and I look forward to resuming coverage in 2023 – which should, in theory, be a national election free year!

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)