Why #BothVotesSNP Matters

You can bet good money that no social media discussion on the list vote will happen without someone asserting that ‘the SNP only won four seats in 2016 so a list vote for them is wasted’. This is usually coupled with the claim that the SNP can ‘only win’ in a few regions. This endlessly repeated social media factoid fossilises opinions, preventing those who hold them from looking into the real reasons for the SNP’s 2016 list result.

They singularly fail to understand that the number of list seats won is not limited by having a constituency landslide. Even if a party wins every single constituency seat, it can still win seats on the list if its percentage share of the vote is similar or higher than its constituency share. The notion that SNP has ‘maxed out’ and ‘can’t win’ is false — modelling proves just so.

As ever, words mean nothing without data to back them up. So let’s look again at 2016, and the claims that SNP can’t win in more regions and can’t win more than four seats — assumed to be a ‘plateau’, putting a ceiling on SNP hopes and feeding the narrative that an SNP list vote is wasted.

The 2016 Election

The 2016 election results are publicly available online, so anyone can model the results for themselves. That’s the advantage of conducting proper analysis using actual data — numbers don’t add up differently for you than for me, so an objective reality behind the superficially appealing social media factoids can be identified, that is independent of ‘opinion’.

Returning to the ‘just four seats’ narrative, what too many Yessers fail utterly to understand is that the four seats result in 2016 is a direct consequence of the SNP list vote being too low— over 5% lower than its constituency percentage share — and not because it has ‘maxed out’ on the list, being somehow ‘prevented’ from winning more seats by the electoral system itself.

This modelling is (as is any) naturally hypothetical, but is provided specifically to demonstrate that the SNP can and will win extra seats, if its list percentage doesn’t peel off to minor parties and stays in line with its constituency percentage (i.e #BothVotesSNP).

The table below shows the potential result if the SNP list percentage was the same as its constituency percentage in 2016. (I have only used the Greens as a source of SNP votes, for the toughest scenario — if the SNP had also taken votes from, say, Labour, its result would have been enhanced even further.)

#BothVotesSNP produces over double the actual 2016 list result — ten seats — proving that there is nothing inherent in the AMS algorithm that somehow ‘prevents’ the SNP winning list seats directly on account of its success in the constituencies.

To see just what the arithmetic can potentially produce, now imagine if all the Green vote went to the SNP (as above, keeping the unionists the same, so as not to give SNP an easy time in this modelling). The table below shows the result. (Remember here, we’ve not trying to predict a real result — the focus is entirely on determing if the arithmetic of the AMS algorithm somehow ‘stops’ the SNP winning list seats).

We now see seven additional seats for the SNP, again winning in all but two regions, and again putting the indy bloc up one seat:

Even in the two regions the SNP didn’t win a seat in under this projection, it is in second place in both for the last list seat:

Current Polling

That was 2016, but what about now? Politics has moved on since 2016. Here’s the baseline projection from the latest Panelbase poll from June-July:

The SNP list percentage share of the vote in this poll is again over 5% lower than its constituency percentage share. But what if they were equal — if people did follow #BothVotesSNP?

The SNP would now triple its list seats to twelve — winning in seven out of the eight electoral regions. So much for the assertion that the SNP ‘can’t win in my region’.

Lastly, just to prove again, as with the 2016 modelling, that the arithmetic of the AMS algorithm doesn’t prevent the SNP winning list seats, let’s imagine all the indy vote went to SNP:

Now the SNP wins thirteen seats. (Overall, the indy bloc is down a single seat but, whether we like it or not, unionists will look only on the fortunes of the SNP as an indicator of support for independence.)

As you see, now the SNP would have 70 constituencies (projected from the poll) plus 13 list seats = 83 out of the 129 seats in Holyrood. That’s 64% of the seats in the Scottish Parliament. The best way to that fabled ‘indy supermajority’ beloved of the list party advocates? #BothVotesSNP

Obviously, in the real world, whether these #BothVotesSNP scenarios come to pass depends entirely on votes actually received, but they’re given here purely to knock on the head the persisent indy social media myth that the mechanics of the AMS itself somehow ‘prevents’ the SNP winning list seats as a result of a constituency landslide. Data analysis conclusively proves that there is nothing inherent in the ‘system’ that puts a lid on SNP list seat aspirations — only its percentage share of the vote does that.

The Importance of #BothVotesSNP

It is thus critical for the SNP’s list fortunes that supporters adhere to #BothVotesSNP and vote for the party in the list as well as in the constituency.

Too many fail to realise that the SNP’s 2011 success resulted from an almost perfect #BothVotesSNP vote — there was just 1.2% difference between the SNP’s constituency and list percentage shares — while the 2016 loss of its majority resulted from the atrophy of the SNP list vote, as supporters fell for spurious ‘advice’ to vote tactically for another party on the list.

SNP supporters must ignore ill-informed social media factoids about an SNP list vote being ‘wasted’, and resist the seductive snake oil lure of the new fringe list parties, whose platform is based on false notions of tactical voting and/or utterly unrealistic assumptions that they will somehow take, for example, ‘just half’ of the SNP list vote.

Data modelling proves that the SNP can win more list seats, it can win in more regions. And it’s a vastly more efficient use of your list vote than a new list party trying to get enough votes to win a single seat.

Holyrood 2021 will be most important election in our lifetimes: the British government and the other unionist parties will judge our desire for independence by the success (or otherwise) of the SNP alone. A strong (and majority) SNP government is therefore essential. Take nothing for granted next May: be under no illusion, your list vote is not a ‘second’ vote to gamble with— use it wisely.

‘Votes per Seat’

One last point: in their efforts to portray an SNP list vote as ‘wasted’, list party devotees often make reference to the half baked metric of ‘votes per seat’ (ignoring the fact that this depends not only on the constituency seats already won, but also the list vote share).

A previous article examined the concept of the ‘efficiency’ of list votes, determining the number of votes needed to get more indy bloc seats from the ‘status quo’ of four SNP seats (predicted by the Panelbase poll). While the first extra seat — a fifth SNP seat — needs just 35,000 more votes, a new list party would need three times as much, over 100,000 votes to win that extra seat. If there were two competing fringe parties (e.g. AFI and ISP) splitting the ‘tactical vote’, then over 200,000 votes would be needed nationally just to ensure one of them wins one seat!

[This is nationally based, as there is no way to know in advance in which region a list party may or may not win a seat.]

As mentioned, if the SNP list percentage catches up with its constituency percentage (similar to 2011 — #BothVotesSNP), then it will break out of the apparent ‘plateau’ of four seats caused precisely by its list percentage being a full 5% lower than the constituency vote share.

The overall mean SNP votes per list seat ratio has now fallen to 101,698 from 285,376:

Modelling from this same poll, we saw that if votes went from the SNP to just one list party (AFI, ISP & RISE: fight it out amongst yourself who is The One True Indy List Party, please), then 103,000 votes were needed for very first seat. So even using list party supporters’ own (and entirely facile) measure, the SNP now trumps the list party option in terms of the ‘votes per list seat’ metric.

But, with the ISP’s refusal to place itself under the AFI’s stewardship, we now can look forward to two fringe parties vying with each other to take SNP votes. No one can know which will win first where. Nationally, with an even split in our poll model, it will take over 200,000 votes for one of them just to get a single seat!

As I say, the very notion of votes per list seat is a facile analysis of just one part of the AMS process, but as too many Yessers are obsessively fixated on this metric, please reflect on the following ‘votes per list seat’ ratios from a #BothVotesSNP scenario:

SNP (13 seats) — 101,698
Party X (1 seat) — 103,808
Parties X & Y (1 seat) — 202,562

#BothVotesSNP doesn’t look so ‘wasted’ now does it?


Shortly after this article was completed, a new Survation opinion poll was published, showing fully an 11% gap between the SNP’s constituency percentage share and its list percentage share.

Above, we saw the consequences to the SNP’s list prospects from a 5% gap, so what does a whopping 11% gap get us? Here’s BallotBoxScotland’s projection:

Answer: three fewer SNP list seats, and three more unionists. A further lesson on the need for #BothVotesSNP — with the SNP doing so well in the constituencies, failing to vote for the party in the list too will only strength the position of unionists.

Follow ListVoteSense on Twitter.

List of articles by ListVoteSense.