The Constitutional Question Dominates: How Scotland Voted in 2021

The constitutional question now dominates Scottish politics. And on the answer to that question the country is both evenly divided and sharply polarised. This is the key lesson of the 2021 Scottish parliament election.

For once the polls had a good election. The average standing of the parties in the final polls were all nearly within a point of the eventual outcome – the biggest errors were a two-point overestimate of the Green vote and a two-point underestimate of SNP strength on the list ballot, and doubtless one of these errors was a consequence of the other. We can thus regard the evidence that they provide on the role that people’s attitudes towards the constitutional question played in determining how they voted –  and, indeed, the level of overall support for independence versus staying in the UK – as reasonably reliable.

Dominated by the constitutional question

As we noted in our blog on polling day, according to the final polls as many as 88% of those who say they would currently vote Yes to independence were proposing to vote for the SNP on the constituency vote, while just 8% of those who back No shared this inclination. Meanwhile, no less than 91% of current Yes voters indicated that they would vote for one of the three pro-independence parties on the list vote, while, again just 8% of No supporters did so. These figures are broadly replicated by a post-election poll conducted by Savanta ComRes and reported in Scotland on Sunday. This found that among current Yes supporters, 87% voted for the SNP on the constituency vote, while 88% backed the SNP and the Greens on the list. The equivalent figures for No supporters were 8% and 9% respectively.

These figures are even higher than the equivalent ones for the impact of Brexit on how people voted in the December 2019 UK election. On that occasion, 83% of those (across Britain as a whole) who were in favour of Brexit voted for one of the parties that was advocating the implementation of Boris Johnson’s Brexit deal, while 83% of those opposed to Brexit voted for one of the parties that was prepared to hold a second EU referendum. There can be little doubt about the depth of the political fault line that the constitutional question now represents in Scottish politics.

Indeed, the importance of the constitutional question in voters’ minds was emphasised by the way in which unionist voters were willing to ignore the traditional left-right divide, and back whichever of Conservative or Labour was best placed locally to defeat the SNP.  As Table 1 shows, this was apparent above all in those constituencies where one of the opposition parties was defending the seat against a SNP challenge. In seats that the Conservatives were defending, support for the party rose while that for Labour fell. In contrast, the opposite happened in constituencies that Labour was defending. Meanwhile the Liberal Democrats appear to have been able to squeeze both Conservative and Labour support in the two mainland constituencies that they were defending.

That these patterns are evidence of voters backing whichever unionist party was best placed to defeat the SNP locally is corroborated by the evidence of Savanta ComRes’ post-election poll. This found that as many as 30% of those who voted for one of the three unionist parties on the constituency vote did so because they ‘voted for this party to stop another party in my area winning’. True, this is not quite the same as saying that they voted for a different party than the one they would otherwise have supported, but given that only 7% of SNP voters reported the same motivation for their vote, it does suggest that tactical considerations played a significant role in how some unionist supporters behaved.

A divided country

More difficult to prove but also difficult to deny is that the intensity of feeling about the constitutional question was also reflected in the record high turnout for a Scottish Parliament election of 63% – despite the limitations on conventional campaigning occasioned by the pandemic. The turnout in the 2014 independence referendum is of course testament to the perceived importance of the issue in voters’ minds. In any event, if people’s attitudes towards the constitutional question did indeed determine how nearly nine in ten people voted in an election in which there was a relatively high turnout, we might well regard the level of support for the pro-Union and pro-independence parties as a useful check on the message of the opinion polls that the country is more or less evenly divided on the issue.

Such a calculation does indeed corroborate the evidence of the polls that Scotland is divided down the middle on the issue. On the constituency ballot the three main unionist parties won just over half (50.4%) of the vote. But on the list vote the three main nationalist parties claimed just over half (50.1%).  In truth, far from demonstrating that Scotland clearly does or does not wish to revisit the question of its constitutional status, as many of those on both sides of the argument would like to assert, the only clear message from the ballot boxes was that the country is divided down the middle on the issue.

How the electoral system favoured nationalists

That, of course, may leave you wondering why with 72 seats, the two pro-independence parties at Holyrood enjoy an overall majority between them of 15 seats, matching the previous record high in the 2011-16 parliament and more than enough to secure the passage of any legislation for holding a second referendum. The answer lies in the workings of an electoral system which was deliberately crafted to be somewhat less than proportional.

Two features in particular are important. First, with nine constituency seats and seven list ones in the average region, there can be too few of the latter to correct the disproportionalities created by the outcome in the former. At this election, this happened in four regions, in all of which the SNP secured one more seat than their theoretical entitlement. True, one of these was won at the expense of the Greens, and so the net effect was only to add three to the pro-independence tally at Holyrood, but it still gave the SNP a substantial advantage.

Second, the D’Hondt formula that is used to allocate seats is more favourable to larger parties than to smaller ones. An alternative method would be to use the Sainte Laguë formula which treats larger and smaller parties equally. If all the seats at Holyrood (both constituency and list) were to have been allocated in line with that formula (in proportion to the parties’ share of the list vote), the SNP and the Greens would have had 65 seats between them, an overall majority of just one.

Yet the disproportionality could have been even greater. The tactical voting illustrated in Table 1 stopped the SNP winning five constituency seats, four of which (Aberdeenshire West, Dumbarton, Eastwood, and Edinburgh Southern) were in regions where the SNP did not win any list seats and where thus any additional constituency wins would not have been reversed by the party being allocated one less list seat. If the SNP had managed to win just one of these four seats, the party would have secured an overall majority on its own.

Holding another referendum?

So, if the country is evenly divided on the principle of Scotland’s constitutional status, what can we say about whether voters believe that another referendum should be held or not? In truth, this a question on which the country appears to be just as divided as it is on the substantive question – not least because most Yes voters, while hesitant about holding a referendum in the next year or two, are clear that one should be held within the next five years, while No voters take the opposite view. Panelbase, for example, have in three polls undertaken in March and April found that 99% of Yes voters believe there should be referendum during the new parliamentary term, while 91% of No supporters take the opposite view. YouGov do not find quite so stark a division but still report that as many as 71% of Yes supporters support holding a referendum in the next five years while 67% of No supporters are opposed. Nevertheless, given the views expressed by most Yes supporters, it would seem to be politically impossible for the new Scottish Government not to attempt to hold a referendum at some point – thereby setting up the prospect of a constitutional clash whose consequences are uncertain but the risks of which – for both sides in the debate – are substantial.


Addendum – the demography of party support

The Savanta ComRes poll also provides a first glimpse of some of the demographic variation in how people voted. It suggests there were sharp differences in support for the SNP by both gender and age, with women and younger people being more likely to support the party than men and older people. In contrast, the academic British Election Study found little difference in the level of support for the SNP by age group in 2016, while what had been a long-standing tendency for the party (and for independence) to be more popular with men than women was somewhat still in evidence. Plenty here to ponder and analyse in the weeks and months to come as more data become available, but these statistics would seem to indicate that the sharpened political division in the pattern of party support is accompanied by sharper – and new – demographic divisions in Scottish electoral politics too.



Avatar photo

About the author

John Curtice is Professor of Politics at Strathclyde University, Senior Research Fellow at ScotCen and at 'UK in a Changing Europe', and Chief Commentator on the What Scotland Thinks website.