Both votes X and wasted votes: is it time to change Scotland’s voting system?
“Both Votes Labour,” voters were urged. “Lend Solidarity your vote on the PEACH paper”, Sheridan roared. Even newcomers were in on the act, asking to make your “2nd Vote Women’s Equality”. The social media Indy-Sphere was all a-twitter with the relative merits of casting your regional vote for the Greens, or whether that was some crafty unionist plot to undermine the SNP. More so than any election since 2003, when the Greens ran their hugely successful “2nd Vote Green” campaign, the mechanics of our electoral system defined the campaign strategies for all parties in this year’s Scottish Parliament election.
The SNP felt their majority was fragile, even in the face of polling suggesting an almost clean sweep of constituencies – and besides, no party willingly concedes votes to another – hence they were pushing a strong “Both Votes SNP” message. Labour, used to seeing significantly less on the list than the constituency vote, knew that this time they couldn’t afford that luxury. And the Greens, although not majoring on the topic, were keen argue that a list vote for the SNP would be unlikely to elect any more MSPs. For many voters this was all a shade bewildering – and for many activists, rather uncomfortable.
So, with all results declared, just who was right? Comically enough, everyone could make quite a convincing argument.
For Labour, they did indeed see their traditional dip on the list. They placed 2nd in the constituency vote, with 22.6%, but only 3rd in the regional, with 19.1%. Had all of their constituency voters followed the party’s advice, they may still have come third but would have suffered a less crushing defeat. So, it’s clear that a failure to go Both Votes Labour cost them dearly.
Likewise, the SNP were stung by a series of surprise constituency losses, and further failed to convert a few seats seemingly ripe for the picking – East Lothian, where former Labour leader Iain Grey was elected on a slender 151 majority in 2011, being the most remarkable. Having failed in spite of all that polling to manage a majority on constituencies alone, their 41.7% of the list vote nationally proved insufficient to recoup their losses. With only two of the eight regions returning any regional SNP MSPs, their argument that the only way to achieve an SNP majority was to give the party both of your votes was somewhat vindicated.
And for the Greens, bitter disappointments in Glasgow, Central and North East regions seemed all the worse with the knowledge that there were over a hundred thousand SNP votes in each region doing absolutely nothing, a mere few thousand of which would have seen Tory and Labour MSPs lose out to some fabulous Green women – Zara Kitson, Kirsten Robb and co-convenor Maggie Chapman.
The fact each party can make a strong case for their own particular interpretation of the Additional Member System (AMS) results should be a sign that the system is inadequate. Interestingly, for all that it’s often said that AMS is “complex” – it’s actually a comparatively simple form of proportional representation – the main problem is instead that it’s unpredictable.
If you are havering about “splitting” your vote between two parties, all you have to go on for the effect that is going to have is gut feel and hearsay, as the exact impact will depend on just how many other voters do the same as you, and you just can’t account for that. So trying to game an AMS result really can be a fool’s errand. It doesn’t help that with a different system for every election we have here, voters can find themselves confused as to which is which!
The most recent Scotland Act grants the Scottish Parliament powers over its own electoral system. Both the SNP and the Greens have a policy that favours the Single Transferrable Vote (STV) system, which we currently use for local elections – which seems ideal, as voters are already familiar with it. Although Nicola Sturgeon apparently declared at a hustings the matter wasn’t a priority, with the two parties having a majority, it’s definitely something that could be done. Many of my fellow Greens will know I have been typically – ahem – frosty towards STV for national elections, but the bizarre regional vote shenanigans this election have tipped me towards it.
STV’s main advantage over AMS is that it allows voters to fully and accurately express their feelings on a whole range of parties, with minimal risk of either wasting your vote or “splitting” any particular bloc of voters, be that on a left-right or independence-unionist axis. An SNP voter who quite likes the Greens could give them a later preference, safe in the knowledge that if their preferred pick is overly successful, a portion of their vote would go the Greens. Likewise, a Green minded voter that was worried their vote wouldn’t count could give another party their later preferences, and if the Green candidate was unsuccessful, know they’ll get at least someone they were reasonably keen on once the vote transfers. Voting in STV is dead simple – cliché, but it’s as easy as 1, 2, 3.
However, preference systems aren’t entirely the same thing as proportional systems. The complex workings of STV give partial proportionality simply by making it hard to win all the seats, but especially if you’re only using 3-4 member seats, as in our councils, that can still lock out parties with reasonable but evenly spread support. Although with great showings in Glasgow and Lothian we know the Greens 6.6% of the vote would be concentrated enough to return some MSPs under STV, it’s mathematically possible that a more even spread could see that share of the vote return no MSPs whatsoever, whereas even on a bad day AMS would see that translate to a fair few.
With that in mind, my personal preference is for some variant of the open list PR systems used in Scandinavian countries. Denmark, Sweden, Norway and Iceland all have similar electoral systems that produce extremely proportional results. These systems see most MPs (135 in Denmark, for example) elected from multi-member constituencies with the remaining number (40 for the Danes) allocated on the basis of the national vote share for all parties crossing a minimum % of the vote. The clever bit is that those MPs still all represent one of the constituencies – the maths for this can be complicated, but the national result is both easy to understand and representative of the national mood. Again, voting in such a system is as easy as pie – you put a single cross next to either a party meaning you’re happy for them to elect whoever, or a candidate meaning you support both that specific candidate and their party.
When 2021 rolls around, we should be able to look forward to an election campaign fought exclusively on the substantive policies and issues, rather than founded on bickering over how best to try and game the system. Either of the alternatives I’ve talked about could help deliver that, and with established consensus between at least two parties already on some kind of reform, perhaps this can be an early win for round two of minority government in Scotland.
Alan,
STV in Britain, in NI and Scottish locals, always transfers fractions.
Ireland uses proportions of next preferences but random for further preferences down. But I must say I have never seen any evidence that this has any implications on the results…
The electoral system can be modified by the Scottish Parliament under the provisions of the Scotland Act 2016. However, it requires a supermajority backing a proposal to do so.
http://www.legislation.gov.uk/ukpga/2016/11/section/11/enacted
Horror example: If the SNP changed their policy, and the Tories agreed, they could use a straight First Past The Post system.
Adam is correct to note the STV difficulties involved in population density. Larger constituencies are fine for urban areas, but murderous in the Highlands and Islands.
Aye get it changed to first past the post, at least everybody understands how that works, we basically sit back and watch the corrupted de handt system let mps in that very few voted for.
The problem with the Scottish AMS (and with the Welsh too) is the lack of proportionality brought about (a) the small amount of list seats coupled with a fixed number of MSPs and (b) splitting the country into 8 parts. It is technically easy to fix, the system would then simply be a list system with some constituencies to determine some names (but not party strength).
Such a system would in my view be preferable to STV. But politically fixing AMS would not be easy. However, neither would we get 6-8 members per STV constituency.
In my view the more realistic question is therefore: Keep the existing system, warts and all, or switch to Irish STV (3-5 seats per constituency, with an average about 4).
And please don’t answer the question by what would be better for which party, then you’ve already lost the argument for any change.
Yes, a reasonably quick and easy fix to AMS would be to effectively replace it with MMS as used in New Zealand – basically the same system, but with the list seats apportioned nationally, and “overhang seats” to correct for situations where any given party wins too many constituency seats compared to their vote share.
The problem with a single national list however is that it completely breaks the local link, and even if an open list, makes it very hard for voters to exercise adequate control over exactly who makes it into parliament. Of course, much like the way the Danish system allocates the nationally apportioned MPs back to individual constituencies, so too could we apportion the list seats nationally but allocate them to the regions, so that’s not a total roadblock.
As a rule, I tend to prefer list systems over preferential as they are generally more proportional, but I’ve come to the conclusion that mixed systems are worse than preferential ones.
ok well i prefer the additional members system and would like to see it adopted for all election, but the point is there is not one voting system you can come up with that people will not try and game. it just is nt going to happen so this is not a reason to reject any system.
examples of gaming in stv elections in ireland include, 1, guessing how many candidates you will win, if you stand to many you risk not winning as many stand to little the same you need to guess, 2, read in the irish independent a complex scheme to convince voters in a constiuency not to first preference the most popular candidate of a party but to preference a less popular one in hope of winning extra seats for that party, so the more complex stv just makes the gaming more complex. produces a less proportional result than addititional members and leaves people with out a consituency msp-mp (im sorry but only proporitionality justifies breaking that and am gives it you with out doing so) finally stv leads to candidate of the same party competing agains t each other weakening parties in an effect aken to the usa primary system.
This is absolutely true, but AMS is easier to game than STV is.
You’re quite right to highlight the difficulties that can be faced by parties who over or under reach in their number of candidates – for example, it’s likely that Glasgow SNP missed out on a councillor in Govan because they stood three candidates, electing only one. Likewise, there are a few wards elsewhere where parties stood only single candidates who then managed a vote share that could potentially have elected two.
I think the “scheme” you referred to is actually just perfectly common “voter management”, an integral part of STV campaigning. The ideal situation with STV is to distribute your votes such that you elect people by getting as close to the quotabas as possible, so you don’t end up needing to pick up too many transfers or risk losing too many by having a surplus that doesn’t “perfectly” flow to the other candidates in your party.
This is exacerbated in Ireland by their really, really bad rules for transferring surplus votes. Rather than “fractional transfers” formed from counting up all the next preferences, they only take a sample of the next preferences and transfer the surplus based on whatever that sample says. That led to interesting scenes in a few constituencies in Ireland this year, where candidates who seemed to have dropped out based on surplus allocation basically demanded they take a new sample just in case that got them through! Fortunately, since we already use automated counting for STV in councils, we’d likely do it at national level too.
Finally, I think it’s a positive that STV (sometimes – one of my problems with it is that it doesn’t ALWAYS) offers voters a choice in which candidates from a party to vote for. That’s a real positive compared to having only one candidate or a closed list, where that candidate or the most likely candidate from the list isn’t an appealing prospect to someone otherwise inclined to vote for a given party.
The constituency link can be, in my view, over-egged. It’s only real purpose is to give constituents one specific person they can shout at, because if you don’t share that representatives views on any particular issue, they just ignore you. That’s why FPTP is so utterly, crushingly bad – most people end up represented by someone who doesn’t agree with them and have no other options.
The Scandinavian system has merit, but as compared to the regional open-list AMS variation recommended by the Arbuthnott Commission, it has one defect. At the next election, the constituency whose Green was a best near-winner may be a different constituency. How can such a Green MP personally earn re-election? Isn’t it a bit random?
The biggest advantage to STV is that we already use it for Council elections. It feels like we spend almost as long explaining the voting system each year as we do putting forward our policies. If Holyrood elections were done using STV, that would be three years in our election cycle (Holyrood, Council, year off) without having to get to grips with a different system.
I wonder if the SNP or even Labour will have much appetite for a change to a system that would deny any chance of a majority without a majority of the vote. Also have always wondered why the Holyrood constituencies weren’t kept aligned with Westminster, and more regional MSPs added when the Westminster ones were last redrawn, but I can’t think of a good reason that they should be aligned.
Finally, wondering just how many Green MSPs we might have with STV. Even with 8 member constituencies, we will struggle to get the 11% of 1st preferences to be elected in the first round, and are then at the mercy of getting enough transfers.
But surely we would want STV with around 6-8 members per constituency?
To aid proportionality, I’d say 6 is about right as the number to achieve that without inflicting absurdly long ballots on voters. However, if we do that with our current total number of MSPs, those constituencies would be quite large, which might not mean that much additional benefit for smaller parties unable to consistently run campaigns across them.
I think 4 would maybe – maybe! – be okay. 3 definitely too few.
Of course, all my many criticisms of STV stand, I’ve just come to agree it’s likely a healthier system than AMS.
And world peace and heaven on earth and cake for everyone too please.
I ask again, current system or 3-5 member STV? If you cannot answer that the debate is a bit pointless…