Voting Systems and Simplicity
voting · 10 comments · 300 words
it has been said that democracy is the worst form of Government except for all those other forms that have been tried from time to time… Winston Churchill
I read this wonderful post from ThreeBallot voting.
I've read it through twice, and I only think I understand how it works.
For all its flaws, First Past The Post is really easy to describe. You can explain how it works in a sentence. "The person with the largest number of votes for them is the winner." I reckon you can do it in two words: "Biggest wins."
Every other form of voting is complicated.
The attacks against paper-based FPTP are simple to understand - and simple to defend against. Here's a 7 page paper from Princeton, and a follow-up 10 pager, discussing some of the subtle and complex attacks against ThreeBallot.
Even if you think ThreeBallot solves some problems, it is undeniable that it is complicated to teach and tricky to defend.
Recently, I was a voter in a ranked-voting election. All of us marked our favourite candidate as "1", the next best as "2", etc. Except for one voter, who thought we were assigning points! They gave a "5" to their favourite candidate and a "1" to the candidate they disliked.
In most democracies, people only engage with a ballot a few times per decade. Any complex skill atrophies unless regularly practised.
Is there a voting system which is simple to understand and robustly democratic?
I think approval voting is nice: check boxes for the candidates you like, don't check boxes for the candidates you don't like. Degenerates into FPTP if everybody ticks exactly one box. This, called "bullet voting", is a form of tactical voting, and is a good idea if you have strong preferences about which candidate should win (as opposed to not really caring which of A, B, C win as long as it's one of them). If not, it gives parties who aren't the two most popular a chance by letting people vote for more than one party (at the same rank).
Range voting extends that by letting you give a numeric score, but figuring out how much more (or less) you like the Greens than Labour can be hard, so it's easier just to give e.g. 10 to candidates you like and 0 to those you don't, and then it's approval voting again.
One other nice thing about FPTP is that it satisfies the participation criteria: it's always better to vote honestly than to not vote at all. Many Condorcet systems (and other popular replacements, like instant runoff) fail this.
Marcus Downing says:
"One other nice thing about FPTP is that it satisfies the participation criteria: it’s always better to vote honestly than to not vote at all."
Except I live in a region that's dominated by one party. I always vote of course, on principle, but my vote has never mattered because they've returned the same MP for the last 15 years, always with a comfortable margin. That MP has no incentive to listen to local people, because it won't affect their future electoral chances at all. It may be that if all the disenfranchised voters actually voted together they'd overturn that; but it's unlikely. It can be hard to persuade people that voting matters when their vote isn't represented at all.
The pros and cons of voting systems aren't in how mathematically correct they may or may not be, but in how it affects human behaviour. A more fair voting system would engage people in politics, and would encourage politicians to care about the people they represent.
Max Wainwright says:
“The electoral system is proportional. This means that the number of seats allocated to parties, for example in the Riksdag, by and large are the same as the parties’ share of votes received.”
https://www.val.se/servicelankar/servicelankar/otherlanguages/englishengelska/electoralsystem/generalelections.4.1dac782216e1e29d7891861.html
The exact allocation of seats is more complicated, but nothing that voters really need to consider (https://en.m.wikipedia.org/wiki/Sainte-Laguë_method)
@edent says:
It's the "but nothing that voters really need to consider" which worries me. If you can't understand it, you can't audit it. If you can't audit it, then it is ripe for corruption.
Max Wainwright says:
You can understand it, it’s not that complex. You can audit it, but it shouldn’t affect how you vote. Unlike FPTP, which is insane (but simpler, i agree).
Max Wainwright says:
What I meant was: it’s nothing you need to consider as you vote. It’s too complex to use as a guideline for voting, but not to understand after the votes have been counted.
@edent says:
If I can't understand how the vote is counted, how can I understand how to effectively cast my vote?
When the London Mayoral elections were run, I saw lots of people misunderstanding how to use their second preference. For example, some people put their first preference second and a minor party first because they wanted to send a message to the winner that they shouldn't take their vote for granted.
Max Wainwright says:
The point is to vote for who you want to get in. You can’t know what is the efficient way because you don’t know how others will vote.
I don’t think labelling “voting for whom you like the least” as efficiency is accurate. If the other party wins your vote gives them zero seats - not so efficient.
Ok, but we don’t have preferences in Sweden, I don’t see the relevance?
@edent says:
I was trying to show that any complex counting schemes leads to confusion in how voters will behave.
The Sainte-Laguë method you mentioned is likely to fall into that category. We both agree that FPTP is a bad voting system. I'm trying to understand if the complexity of better systems outweighs their utility.
Max Wainwright says:
People here in Sweden in general don't know about how exactly the seats are allocated, and if they do they don't care. So I'd say it doesn't influence how they vote – meaning you can hardly say it leads to confusion.
If by "confusion in how voters will behave" you mean confusion for e.g. statisticians or the media, is it a requirement for an electoral system to promote predictability in voters? That the system itself should promote that predictability? Why?
I think the complexity is overrated. The more advanced pages covering our elections beforehand will show not just the predicted numbers but also what allocation of seats that will lead to. It's trivial to be transparent about this stuff.
And I would say doesn't complicate auditing in any real way. The difference in complexity between the Sainte-Laguë method and FPTP might is large in relative terms, but in practice there's hardly any difference. We have computers to do the computing.
Don't know if it's just because it's what I grew up with, but it really does not seem complicated to me or anyone I know.