In my post on Thursday night, I pointed out three remarkable things about the Perth City South by-election result: a) that, on first preference votes, the SNP had moved from second place into first, b) that the SNP share of the vote had increased by 6.4%, and c) that there was a slight swing from Tory to SNP. Well, here's a fourth remarkable thing. In spite of the surge they enjoyed, the SNP still 'only' took 32.1% of the first preference vote, and almost all of the remaining two-thirds of votes were cast for unionist parties. Given the perception of a unionist bloc vote that wants to stop the SNP at all costs, you'd therefore expect the SNP to have suffered a lopsided defeat once the lower preferences of eliminated unionist candidates were redistributed. But that simply didn't happen. Even after several rounds of redistributions, the SNP were still agonisingly close to winning the seat - they were only beaten by 154 votes.
The simplest way of demonstrating what happened is to look at how the Lib Dems' votes transferred once it became a straight contest between SNP and Tory. Apart from a very small number of votes that had been transferred from the Greens at an earlier stage, almost all of these Lib Dem votes can reasonably be described as 'unionist party votes'.
Liberal Democrat transfers :
So almost two-thirds of this supposed unionist bloc failed to express a clear preference for the Tories over the SNP, and almost one-fifth actually expressed a preference for the SNP over the Tories. Obviously the high number of non-transferable votes can be partly explained by unfamiliarity with the voting system, but nevertheless, even among those Lib Dem voters who did use their lower preferences, more than one-third backed the SNP. The fact that more Lib Dems broke for the Tories than for the SNP explains why the Tories managed to squeak a victory - but unless the original first preference result had been extremely tight, that wouldn't have been enough to swing the balance. You're not going to see the Tories overcome first preference deficits of 8% or 10% on this pattern of transfers.
I'd suggest all of this could pose a problem for the Tories at the next Westminster general election. Assuming the 29% of the national vote they managed this year proves to be 'Peak Tory' (and there are many reasons for thinking it probably will), they're going to be looking to buck the trend in seats they already hold by appealing to Labour and Lib Dem supporters to cast an anti-SNP tactical vote. It may be that not enough people are going to be receptive to that message - and the problem could get a lot worse if the Tory government goes on to become anything like as actively disliked in the north-east and the south as the Major government was in the 1990s. I'm increasingly optimistic that the SNP can win back at least some seats from the Tories, whenever the election is held.
In SNP-Labour battleground seats, it's obviously a very different story, because most Tory supporters are for the moment obsessed enough with the constitution to think Labour are preferable to the SNP. But for how much longer will that be the case? Richard Leonard's elevation to leader could prove to be something of a watershed for unionist tactical voting, because Tory supporters will no longer be able to tell themselves that Scottish Labour is more centrist than the Corbyn-controlled UK party, and thus 'safer' to vote for.
We know that Labour are going to take every opportunity at the next election to peddle the fiction that voters need to abandon the SNP for Labour if they want to see a non-Tory government. Well, Tory supporters are going to hear that message as well, and some of them may even start to convince themselves that a tactical vote for the SNP could be the most practical way of preventing a Corbyn administration. At the very least, they may become more conflicted about whether the SNP or Corbyn is the greater and more immediate threat, which could lead them to simply revert to a non-tactical vote for their own first-choice party.
* * *
After the SNP's 'curate's egg' performance in by-elections this week, I was hoping that the two new Scottish subsamples from YouGov would offer some clues as to what the state of play really is, but in fact they've just muddied the waters even further, because they're completely contradictory.
YouGov (a): Labour 36%, SNP 33%, Conservatives 22%, Liberal Democrats 5%, Greens 3%, BNP 1%, UKIP 1%
YouGov (b): SNP 38%, Conservatives 25%, Labour 24%, Liberal Democrats 7%, UKIP 2%, Greens 1%
So we have the first YouGov subsample since the summer to put Labour in the lead...and then the first YouGov subsample since the general election to put Labour as low as third. Across all polling firms, twenty-two of the last twenty-four subsamples have shown an SNP lead - but the two that didn't have both been published within the last week. Make of that what you will.