You might remember that I calculated at the end of last week that if the Savanta poll of SNP members happened to be bang-on accurate, Kate Forbes would need exactly 76% of Ash Regan's second preferences to become First Minister. Intuitively that would seem attainable for her, given that - almost by definition - Ash Regan supporters tend to be extremely unenthusiastic about Humza Yousaf. However, the 76% figure assumes that every Regan supporter will use their second preference, which in the real world will not be the case. The percentage of second preferences required by Forbes will effectively creep up and up with every Regan voter who does not transfer at all.
So the key to the puzzle of this leadership election may be the percentage of SNP members who can be expected to use more than one preference. And it struck me that a strong pointer might be found in the detailed results of previous SNP internal elections. But then I ran into a snag, because as far as I can see there have only been five all-membership SNP internal elections in this century (and quite possibly ever). The 2004 leadership election is no use at all, because Alex Salmond won an absolute majority of votes and therefore no second preferences needed to be taken into account. Nicola Sturgeon also won an absolute majority in the depute leadership election in the same year, and it's maddeningly difficult to find full results from any of the three depute leadership elections over the last decade - the most details are available for 2014, and seem to imply that the vast majority of Angela Constance's votes did transfer on the final round, but I couldn't find any certainty of that.
But then it occurred to me that leadership elections from other parties might be just as useful, because there's no obvious reason why Labour members would be any more or less likely to use second preferences than SNP members. However, since Labour scrapped the electoral college and introduced one person, one vote, there hasn't been a single Labour leadership election which wasn't won by an absolute majority on the first count. So the best and most recent example I could find was actually the 2006 Liberal Democrat leadership election, which had three candidates, none of whom won a majority on the first count.
52036 valid votes were cast in the first round, of which 12081 votes were for Simon Hughes, the candidate who finished third and last. (As you may remember, Hughes at one point had been the outright favourite to win, but faded after a controversy about the dishonesty of some historical comments about his sexual orientation.) Once Hughes was eliminated and his votes were redistributed according to second preferences, there were 51325 votes left in play for the second and final count, in which the leadership was won by Sir Menzies Campbell. That implies only 711 of Hughes' voters did not transfer to one of the other candidates - which amount to less than 6% of the total Hughes support. That's a strikingly low figure, and if that can be assumed to be typical of a ballot of any political party's members, it's good news for Kate Forbes. If only around one in twenty of Regan's voters are likely to drop out on the final count, it means the target for Forbes to win does increase from 76%, but not by all that much.
* * *
If you'd like to help Scot Goes Pop continue in some form, donations are welcome. The simplest method is a direct Paypal payment. My Paypal email address is:
jkellysta@yahoo.co.uk
If you wish, you can add a note saying "for the fundraiser", although even if you don't do that, it'll be fairly obvious what the payment is for.
If you don't have a Paypal account, last year's fundraiser is still open for donations HERE.