To what extent was there split voting in the Scottish Parliament election – and did it matter?

The Scottish Parliament election saw the SNP emerge victorious, with the Scottish Conservatives overtaking Labour to finish in second. The voting system in use is designed to ensure that a majority is very difficult to achieve, which worked against the SNP in terms of their final share of MSPs. Here, Sean Swan asks whether voters ‘split’ their votes, and whether this had a tangible affect on the outcome of the election. 


Credit: Scottish Government, CC BY 2.0

In elections to the Scottish Parliament each voter has two ballots, one to elect a constituency MSP on a First Past the Post basis, and one to elect regional MSPs by party list PR. Regions are comprised of between eight and ten constituencies, and each region returns seven MSPs. The regional seats are allocated using the d’Hondt method in which each party’s vote is divided by the number of seats it has won pus one. Because the seats won in the constituency election are included in this calculation, the more constituency seats a party wins, the harder it is for that party to win regional seats.

Opinion polls indicated that the SNP would win almost all the constituency seats. This would make it very difficult for them also to win regional seats. On the other hand, smaller pro-independence parties, such as the Greens who did not stand in most constituencies, could take a regional seat on about 6% of the regional poll. This fact suggested that a regional vote for the SNP might be wasted and pro-independence voters might make better use of their regional vote by voting Green.

In April Professor John Curtice, in a study commissioned by the Electoral Reform Society, suggested exactly this. He argued that the SNP would win a majority on the constituency seats alone. Consequently, in order to maximise the number of pro-independence MSPs elected, pro-independence  voters should split their vote by voting SNP in the constituencies but give their regional vote  to other pro-independence parties ‘such as the Greens or the leftwing party Rise’. Based on opinion polling in 2016 up to that point, which put the intended SNP constituency vote consistently in the 51% – 53% range, Prof Curtice was correct in this assessment.

However, the SNP, unsurprisingly, was opposed to this strategy. The SNP strategy was ‘Both Votes SNP’. Nicola Sturgeon cautioned voters that if they wanted “to see an SNP government re-elected […]do not think you can take a chance with one of your votes […] go to your local polling station and vote SNP with both of your votes.” Such caution was well founded as the danger was that the opinion polls could lead to complacency on the part of SNP voters who might be tempted either to not bother to vote or to play ‘a game of chance with the electoral system’ and give their regional vote to another party.. There may also have been some question as to the accuracy of the polls given the polling fiasco in the UK general election last year.

The actual SNP vote turned out to be 46.5% in the constituencies, dropping to 41.7% in the regions, yielding a split vote of 4.8%. That there was a split vote is obvious, there always is to some degree, but what is more difficult to determine is how much of this split can be attributed to the Curtice split vote strategy. Some estimation of the size of the split which was due to this strategy can be gained by comparing the difference between the SNP split vote this year and in 2011 (see Table One).

Table 1: Difference in the SNP’s constituency and regional share of the vote by region

Screen Shot 2016-06-08 at 17.31.51

Firstly, there was a larger SNP split vote this year (4.8%) than in 2011 (1.35%). Secondly, the main beneficiary of the increased split appears, as would be expected if the split vote strategy was in play, to have been the Greens. In 2011 the Greens got 4.4% of the regional vote; in 2016 this rose by 2.2% to 6.6%. Meanwhile the SNP’s share of the regional vote in 2011 was 44%, dropping by 2.3% to 41.7% in 2016.

This is strongly suggestive of the split vote strategy resulting in an increase of about 2.25% in the SNP split vote, to the benefit of the Greens. (RISE can largely be discounted from these equations as it failed to take off, taking only 0.5% of the regional vote nationally and even in Glasgow obtaining only 1%).

But what impact did this actually have? Hypothetically re-running the regional vote with the SNP vote uniformly increased by 2.25% and the Green vote decreased by the same percentage, produces the results given in Table Two.

Table 2: Actual regional results and hypothetical results had the 2.25% extra split in the SNP vote not occurred 

Screen Shot 2016-06-08 at 17.32.03

Note: 2.25% was added to the actually SNP vote and 2.25% has been subtracted from the actual Green vote

Without the Curtice split vote strategy the SNP would have taken an additional two regional seats. This would have given them a total seat count, including constituency seats, of 65 which would have produced an SNP majority in Holyrood. On the other hand, the Greens six regional seats would have been reduced to only two. In other words, Nicola Sturgeon was right to the extent that the split vote tactic cost the SNP a majority; on the other hand, it resulted in two more pro-independence MSPs being elected than if the strategy had not been followed. The consequences of the split vote strategy were thus significant, but the size of it, in the context of polls showing the SNP winning almost all constituency seats, was surprisingly small. All parties experienced some degree of vote splitting.

Historically the SNP vote does not split easily. In 2011 the difference between the SNP’s regional and constituency vote was only 1.4%, which was the smallest split of any main party. In 2016 all the other major parties experienced a reduction in the size of their split vote while the SNP slit not only increased but became the largest split vote of any major party (see Table Three).

Table 3: The degree of split voting by party in 2011 and 2016

Screen Shot 2016-06-08 at 17.33.37

Whether or not a split vote strategy is desirable depends on whether the desired outcome is to maximise the number of SNP MSPs or to maximize the number of pro-independence MSPs. Had the split vote strategy effect been much larger, say 25%, the SNP would have taken no regional seats and the Greens would have had about 16. In other words, it would have cost the SNP 4 seats but would have cost the unionist parties 10 seats. Whether or not we ever witness split voting on that scale probably depends on how polarised Scottish politics becomes on the constitutional question.

Note: this post represents the views of the author and not those of Democratic Audit – Scotland or the LSE. Please read our comments policy before posting. 

Sean Swan is a Lecturer in Political Science at Gonzaga University, Washington State, in the USA. He is the author of Official Irish Republicanism, 1962 to 1972.

A guide to making up your mind in the referendum on EU membership

Post a Comment