Stat-pummelled readers will be glad to know that this is the last article we plan to write about the vagaries of the AMS electoral system, and how it might apply to next year’s Scottish Parliament election, for some time. This one also shouldn’t be full of tables and figures, so strap yourself in and let’s get this job finished.
What our analysis yesterday and on Sunday concluded was that it’s extremely hard to “game” AMS by voting tactically – which is unsurprising because it was deliberately designed that way.
But advocates of a so-called “Yes Alliance” aimed at maximising pro-independence MSPs argue that there’s a “sweet spot” in which list votes for Yes parties other than the SNP can tilt the balance in the Holyrood chamber.
The reasoning is that with current polls suggesting that the Nats will win 70 or more constituency seats, the AMS divisor mechanism will reduce their list vote so severely that it’ll be too low to have a chance of winning any list seats.
Therefore, runs the theory, any list votes cast for the SNP will be wasted and should instead be “lent” to the Greens, RISE, Solidarity or other parties of the pro-Yes left in order to defeat Unionist parties.
It seems an attractive case. It has the advantage of “truthiness”, which means that it’s easy to get over in a couple of sentences, it sounds logical, and it takes quite a lot of time and detail to explain the flaws in the reality.
By rank carelessness, we seem to find ourselves in a position where that’s our job.
The argument’s great appeal is that in an abstract theoretical sense it’s true – there IS a statistical point where tactical votes could deliver more pro-Yes seats. The fatal weaknesses are that (a) the window is incredibly narrow and to either side of it you do more harm than good, and (b) it’s absolutely impossible to predict in advance where the point will be and tailor your vote accordingly.
To understand it, we need to start with first principles, namely the fact that under AMS, if you get 100% of the list vote you get all the list seats, even if you’ve already got all the constituency seats.
Now, obviously nobody’s ever going to actually get 100% of the list vote, but what the above fact tells us is that it IS perfectly possible to get list seats even in the case of a constituency landslide. To find out where that stops being the case, we just need to work out exactly where the cutoff point below 100% is.
Unfortunately, there’s no way of doing that, for several reasons.
1. You don’t know what percentage of the vote the landslide party will get.
Polling is a snapshot, not a prediction. Even in the last month before the 2007 Holyrood election, polls found SNP leads of anywhere from 12% to just 2%. In late August of 2006 – almost exactly the same distance we currently are from the 2016 election – a poll had Labour 8% in front.
Basing any sort of 2016 voting strategy on current polls, then, is idiotic. Even days away from the vote, let alone months, it simply won’t be possible to reliably say what any party’s list vote will be. (You’d think the mass failure to predict the Tory majority in May would be proof enough of that, but seemingly not.)
2. A constituency landslide doesn’t prevent list seats.
Even on current polling, though, SNP list seats look probable. The party got list seats in all but one Scottish region in 2011, even where they won most or all of the constituencies – they got one in North East Scotland on 52% of the vote despite winning EVERY constituency seat, and three list seats in Highlands & Islands on just 47% of the vote despite winning six out of eight constituencies.
Earlier this week the Electoral Reform Society projected, based on current polling figures, eight list seats for the SNP on a 54% vote share, even after sweeping up 71 of the 73 constituency seats.
(If you get one list seat in a constituency landslide you’re also more likely to get more, because the ongoing effect of the divisor is much less dramatic.)
3. You can’t predict local factors.
Even amid this year’s overwhelming SNP victory at Westminster, one MP from each Unionist party resisted the Nat tsunami – Ian Murray in Edinburgh, David Mundell in the Borders and Alistair Carmichael in the Northern Isles.
As we discovered on Sunday, any single constituency seat can affect the list outcome in unpredictable ways. In a hypothetical example using totally random figures, we saw how a Conservative constituency win brought the Tories no overall gains, but gave Labour an extra seat at the expense of the SNP.
4. People don’t actually like voting tactically.
The “SNPout” campaign in May’s UK general election was highly motivated, organised and funded, and was also relentlessly publicised and supported by a sympathetic media. Yet its effect was almost zero, despite the fact that in most Scottish constituencies it was incredibly easy to tell which party was best placed to defeat the SNP.
At the end of the day, people are simply reluctant to vote for any party other than the one they really support. You’ll be very lucky to get as many as 5% to do it, and for tactical list voting to start to work you need figures closer to 40%.
Tactical voting is hugely more effective in FPTP elections than AMS ones like Holyrood, but even with every possible advantage the Pouters failed dismally. A tactical Yes vote at Holyrood would be orders of magnitude more difficult.
5. The tactical vote itself is split.
And of course, none of those advantages will apply next year. The pro-Yes but non-SNP vote will be divided among several parties. The small number of voters prepared to vote tactically in the first place will have to decide whether their list vote goes to the Greens, Solidarity, the unknown factor of newcomer RISE, or someone else.
We already know from Monday’s article what happens when an “anti-X” vote can’t agree which direction to attack from – as well as the hilarious slapstick farce of the Pouters, we also saw how SNP and Green votes cannibalised each other at the 2014 European elections and let UKIP steal a Scottish seat.
We apologise if we’re repeating ourselves. And once again, we’re not telling anyone how to vote – if you want a RISE or Green or Solidarity MSP, vote for them.
But “the sweet spot” is a fantasy. It can only be identified in retrospect – standing in the polling booth you have no way of knowing what your vote will do. You may as well lob a brick into a bouncy castle blindfolded and hope it hits a child molester.