Some months ago, I considered writing about the pros and cons of tactical voting by independence supporters on the list vote, but then the subject was dealt with by others with strong opinions and I decided not to get involved in a potentially acrimonious debate.
First it is worth trying to estimate the possible scale of tactical voting, specifically the case of SNP supporters voting for the Greens on the list vote. Across the whole of Scotland, the SNP won 1,059,897 constituency votes and 953,587 list votes, a difference of 106,310. The Greens gained 150,246 list votes and 13,172 votes in the three constituencies where they fielded candidates. A very simplistic extrapolation of the Green constituency votes is that if they had contested all 73 constituencies they might have won 73*13172/3 or more than 320,000 votes. Given that the Greens will have contested constituencies where support for the Greens was relatively high, one might guess that the true figure would less than half of this. This is still consistent with the hypothesis that there was no significant tactical voting, and that almost all the Green list votes were cast by people who either voted Green with their constituency vote, or else would have done so, had there been a Green candidate for their constituency, but mostly voted SNP instead.
I have created a spreadsheet into which I can put the election results for a region (constituency seats won by each party and list votes). I can then transfer list votes from the SNP to the Greens and vice versa, keeping everything else unchanged, and see how this would have affected the allocation of list seats. Unfortunately, the most detailed listing of the results that I have found is on Wikipedia (https://en.wikipedia.org/wiki/Scottish_Parliament_election,_2016), and even that is not complete.
The SNP would have needed another 16258 votes to win a list seat; the Green list vote was only 12722 and hence if every Green voter had switched to the SNP it would not have been enough to give the SNP a list seat. However, if 1813 (just over 1.4%) of the SNP voters had voted Green instead, the Greens would have won a list seat at the expense of the Tories.
If 11171 (nearly 48%) of the Green voters had voted SNP instead, the SNP would have won a list seat at the expense of the Greens; if 6136 (5.3%) of the SNP voters had voted Green instead, the Greens would have won a second list seat, at the expense of the Tories.
Highlands and Islands
If between 3608 and 7786 (24.4% to 52.7%) of the Green voters had voted SNP instead, the seat that the Greens won would have gone to the Tories instead; only if more than 7786 had switched would the final list seat have gone to the SNP rather than the Tories. If between 3388 and 7566 (4.15% to 9.3%) of the SNP voters had voted Green, the list seat that the SNP did win would have gone to the Tories instead; only if more than 7566 had switched would the SNP have gained a second seat.
Wikipedia does not give the list vote for the SNP. At least 21678 SNP voters would have had to vote Green instead to have given the Greens a third list seat at the expense of the Tories.
Mid Scotland and Fife
If at least 736 (4.1%) of the Green voters had voted SNP instead, the seat that the Greens won would have gone to Labour. Even if every Green voter had switched to the SNP it would not have been enough to give the SNP another seat. 18787 (15.6%) of the SNP voters would have had to vote Green instead to give the Greens another seat at the expense of the Tories.
North East Scotland
Wikipedia does not give the list vote for the Greens. However, the SNP would have required nearly 35000 more votes to win a list seat.
Wikipedia does not give the list vote for the Greens. However, the SNP would have required 14120 more votes to win a third list seat at the expense of the Tories.
If 1914 (11.1%) of the Green voters had voted SNP instead, the SNP would have won a list seat at the expense of the Greens; if at least 18547 of the SNP voters (13.7%) had voted Green instead, the Greens would have won a second list seat, at the expense of the Tories.
Could tactical voting, in the form of SNP supporters voting for the Greens on the list, have cost the SNP any list seats? In three regions (Central Scotland, Mid Scotland and Fife, North East Scotland) it is impossible; in three more (Glasgow, Highlands and Islands, South Scotland) it is highly improbable; in West Scotland it is just possible, but the seat did go to the Greens.
Could more tactical voting have increased the number of pro-independence MSPs? Yes, in Central Scotland, possibly in Glasgow, perhaps even in West Scotland and Mid Scotland and Fife, all regions where the SNP did not win any list seats. On the other hand, in the Highlands and Islands more tactical voting could have caused the SNP to lose a list seat to Labour.
Overall, I think that any tactical voting by SNP supporters on the list was limited and probably had no impact on the outcome of the election in terms of numbers of seats; any attempt to blame, even partially, the failure of the SNP to win another outright majority on tactical voting would be wrong. The real worry is that not enough people voted for pro-independence parties, 47.08% on the constituency vote and 49.42% on the list vote, including RISE and Solidarity. The increase in the Conservative vote is also worrying, even if some of it is due to desertions from Labour and tactical voting on the Unionist side. I wonder if this is a consequence of the Scottish election being overshadowed by the possibility of a Brexit.
One hypothetical scenario intrigues me, although I would not advocate it as I think it would be both unethical and risky. What would have happened if the SNP and the Greens had agreed that the SNP should ask their supporters to give their list votes to the Greens, while the Greens had not contested any constituencies but asked their supporters to give their constituency votes to the SNP? Is there anything, other than public opinion, to stop two parties trying to game the Scottish electoral system in this way, to get around the fact that any party which does well on the constituency vote will suffer for it when it comes to the allocation of regional seats?