Voting system geekery: London mayoral edition

I’ve just been along to vote in the elections for mayor of London.

It’s a kind of alternative vote system; you can pick a first and second choice (but not a third and so on), and after the first round of counting, if no candidate has 50% of the votes, they eliminate all but the top two candidates and reassign votes according to people’s second preferences.

I think that’s clearly an improvement on a straight first-past-the-post system [FPTP], but I can’t see that it makes sense to fix it at only two rounds of counting — rather than, to take the simplest alternative, eliminating the candidates with the least votes one at a time, reassigning the votes, and doing it as many times as you need to.

Electing a mayor is a somewhat different situation to a general election; some of the problems that general election reform would attempt to fix simply don’t apply.

So for example, there’s the question of proportionality: the number of MPs each party has in the  Commons is often wildly different to the percentage of votes they won nationwide. But there’s only one mayor, so that’s irrelevant to a mayoral election.

Also, since the whole of London is one big constituency, everyone’s votes count exactly the same; there are no safe seats where the voters can have little influence, or marginal constituencies that attract wildly disproportionate attention from politicians.

So some of the specific issues don’t apply. But the overall problem with FPTP is that it deals very badly with anything other than a two party system, and tends to entrench a two party system by default.

Having an alternative voting system solves part of the problem. It reduces the potential of spoiler candidates; what could be called the Ross Perot problem, of a minority candidate having a disproportionate impact because they attract just enough votes to swing the election. And it removes some of the bias against minor parties and new parties, since if you know you have a second choice, you can at least vote for a minor party without feeling that your vote is wasted. If you feel that the Green manifesto actually represents your opinions most accurately, but you’d rather have Labour than Conservative, you can vote Green without feeling that you are mainly helping the Conservatives.

However, cutting straight to two parties for the second round of voting still helps entrench the two party system. You can feel free to vote for a minor party for your first choice, but the tactical element just comes back in for the second choice, since there’s a strong incentive to try and guess which two parties are going to make the cut and vote for one of them, so that your vote counts for something. If ‘everyone says’ that Labour and Conservative are the two favourites, and your preference is, say:

Green > Lib Dem > Labour > Conservative > UKIP > BNP

then there’s a strong incentive to vote 1) Green 2) Labour.

That doesn’t seem ideal, but I don’t actually think it’s a major problem as long as there are two clear front runners. Being pragmatic with your second choice isn’t an outrageous compromise. But if the votes are reasonably closely split between three or more main candidates, then the whole thing breaks down again. Let’s say the minor parties have 10% of the vote between them, and Tory, Lib Dem and Labour are running in the polls at about 30% each: well, a tiny swing between any two of those will decide who gets through to the second round of voting, so we’re back to a tactical voting situation again.

Our hypothetical voter now has a strong incentive to vote 1) Lib Dem 2) Labour, and now they are compromising on their first choice.

And if there were four strong candidates, then the outcome would become even more random and the cut off of the top two for the second round of voting would be even more arbitrary.

It’s one of those things which is annoying because it’s so unnecessary. Why go straight down to two candidates? Why not have as many rounds of counting as it takes?

A thought on AV

I’m sure the whole world is on the edge of its seat on the eve of the UK’s referendum to change its voting system. I don’t particularly want to go over the arguments about voting reform again [previous posts for the truly interested], but I’m interested by the psychology of the supporters of first past the post. It is, after all, a truly terrible voting system for any election with more than two candidates.

There are two obvious reasons why people might support the status quo, one more cynical than the other. The cynical one is simple self-interest: FPTP tends to entrench a two-party system and creates a structural advantage for the two major parties. The other obvious reason is small-c conservatism. We’ve always done it this way, if it ain’t broke don’t fix it, if it was good enough for Winston Churchill it’s good enough for me, blah blah blah.

But some politicians seem to have a visceral dislike of it, and I wonder if part of that is the psychology of party politics. You don’t generally end up as an MP, and certainly not a government minister, without a deep tribal attachment to a particular party. The kind of people who go into student politics, who get jobs as researchers for MPs, who go door to door with leaflets: these are people who always know who they’re going to vote for, who always have a dog in the race. If their man doesn’t win, that counts as a loss. People who are active in party politics are like sports fans — Chelsea till I die! — and their experience of elections is completely different to that of the mass of the public.

Perhaps that’s why they seem to regard the use of second preferences as somehow cheating. They can’t relate to people who are ambivalent about all the parties, who often feel they are voting for the least bad choice, or who feel unrepresented by the system.

I think AV is a better system than FPTP because it eliminates spoiler candidates and tactical voting. But I think it also appeals to me personally because it is a better reflection of the way I feel about the parties. I don’t go into the voting booth with one clear automatic first choice; I go in weighing up a mix of different priorities. I don’t agree with any of the parties wholeheartedly, and I don’t like them very much, but some are preferable to others.

I might go in to vote feeling that Labour, the Lib Dems and the Greens would all be acceptable, with Conservative a distant fourth but still clearly preferable to the BNP, UKIP, the Christian Party, the Workers’ Revolutionary Party, or Respect; well, AV allows me to express that.

Depressingly it looks very much like the referendum is going to fail, so we are going to continue with a voting system that is designed for people with the absolute political certainty and strong partisan instincts, rather than wishy-washy people like me. Ho hum.

Yes to AV.

Just a quick pointer to a couple of my blog posts from last year’s election: ‘First Past The Post makes politicians lie to us’. And the follow-up: ‘FPTP makes politicians lie to us (hypocrisy update)’.

I would actually favour some kind of proportional representation — this one seems quite ingenious — but at least AV would be an improvement, because it would, I think, effectively get rid of tactical voting.

Serious political geeks will explain that tactical voting is technically possible in an AV system; but it’s much more complicated. It’s quite difficult even to explain how it works. Unlike FPTP. In a FPTP election with more than two candidates, tactical voting is so completely obviously part of the system that people don’t even have to think of it as tactical: there is an obvious and clear incentive to vote for a less-favoured candidate who seems to have more chance of winning.

Under AV on the other hand, anyone who wants to do tactical voting is going to have to do detailed research into voting patterns in their constituency and some careful mathematical analysis of different possible outcomes in order to work out how to maximise the effect of their vote. I don’t think many people will do that, and if they do, well, good luck to them.

Why I like STV

I’ve been mulling over the various flavours of voting reform available and I’m most tempted by Single Transferable Vote — which is, as it happens, also the system favoured by the Lib Dems at the moment, although that’s not particularly why I like it. No, what I like about it is that it allows you to express some nuance in your vote.

A quick explanation for those of you who haven’t been geeking out over voting reform over the past few weeks.

Rather than the current system with one MP per constituency, STV would have larger multi-member constituencies. So under one particular example that someone came up with, I would be in a constituency called ‘South Central London’, consisting of Southwark, Lambeth and Wandsworth, which would have 6 MPs. And so each party would have multiple candidates standing in each contistuency. And instead of marking a cross next to one candidate, the voter chooses candidates in order of preference – writing in 1 2 3 4 5 in the boxes.

Since there are 6 MPs, any candidate who is first choice for 1/6th of voters is elected. Let’s pretend here are 600,000 voters to make the maths easy; so 100,000 votes means you’re elected. If you got 120,000 votes, the 20,000 surplus votes would be redistributed to other candidates in the proportion that they were the second choice of all 120,000 votes. So if it was a Labour candidate, it’s quite likely that most of those 20,000 votes would be distributed to other Labour candidates. That process is repeated until all the seats are filled; but if at any time the votes are split such that no-one has 100,000 votes (which is quite likely with a lot of candidates) the candidate with the fewest votes is eliminated and their votes are redistributed in the same way.

Notice that the system doesn’t actually attempt to be proportional across the whole country, it is just locally proportional; but that’s good enough, really. Most constituencies would elect at least one representative from each of the major parties, so all the parties would be trying to win votes in all parts of the country; MPs would be tied to a local constituency, though a slightly larger one than we’re currently used to. It also effectively creates a certain minimum threshold for minor parties: for example, the BNP got about 2% of the national vote on Thursday, so strict proportionality would give them 12 MPs; but using STV, they wouldn’t have any MPs unless enough people voted for them in a particular constituency. So in my theoretical example, 100,000 people in South Central London would have to put the BNP as one of their choices for them to win any of the seats. Whether you think that kind of minimum threshold is a good thing or not is another question: it might be more democratic to let the fruitcake parties have a presence in parliament.

But the reason I like STV is that it actually allows you to vote with a bit more nuance. You have a choice of candidates from each party: if you’re an old fashioned socialist, you can try to find a Labour candidate who seems more left wing for your first choice. If you believe in small-government but you’re socially liberal, you might be able to choose Tory candidates who share those views and avoid voting for the ones who seem overly motivated by religion. Or you might want to vote across party lines: my small government socially liberal voter might vote for some combination of Conservative and Lib Dem candidates. If, like me, you don’t necessarily want the Green Party to be running the government but you do want a strongly pro-environment voting bloc in parliament, you could vote Green>Lib Dem>Lib Dem. Because although each voter nominally only has one vote, hence Single Transferable Vote, in practice you have a sort of mixed vote which you can share out among the candidates. You don’t know in advance quite how much of your vote is going to count for which candidates, because the maths is far too unpredictable, but it does at least make some allowance for the fact that most of us don’t completely agree with any of the parties.

FPTP makes politicians lie to us (hypocrisy update)

Yet another leaflet through the door, this one from the Conservatives, folded so that on one side it says, in great big shouty letters:

Honest Politics?

on the other side, it quotes the Lib Dem leaflet as saying

“Only the Lib Dems can beat Labour here…”

with a big arrow pointing at the quote and in even shoutier all-caps:

IS THIS ACTUALLY TRUE?

The leaflet points out, accurately enough, that it isn’t true, and quotes the Conservative candidate, who is of course outraged by this:

“I got involved in politics because I was tired of dishonest politicians. At a time when we desparately need to restore trust in politics, it is a shame that my opponents seem intent on misleading people and using dirty tactics.”

“What does this say about them? How will they behave if elected?”

All of which would be very reasonable, if she hadn’t put her own leaflet through my door with this quote on it:

“Only the Conservatives can beat Labour here”

Which is exactly the same kind of blatant attempt to mislead the voter. The hypocrisy of it is breathtaking, and frankly they should all be ashamed of themselves.

A pox on all their houses. This kind of crap is exactly why we need electoral reform.

First Past The Post makes politicians lie to us

Here are some direct quotes from leaflets delivered in this constituency:

Labour leaflets: ‘Only Labour can keep the Tories out here’ ‘It’s a straight choice between Labour and the Tories round here’

Lib Dem leaflet: ‘Only the Lib Dems can beat Labour here’

Conservative leaflet: ‘Only the Conservatives can beat Labour here’

All of them accompanied by bar charts that purport to show why their party is the only one that can fend off the dreaded enemy.

It’s not exactly inspiring to have politicians try to win your vote in this way. It doesn’t do much to counteract people’s disenchantment with politics. But more sinister is the fact that at least two of those statements are in direct contradiction with each other. In other words, someone is telling porky pies.

In fact, when you read their explanations carefully, it is hard to pin down any statements of fact which are outright falsehoods; but all of them are, I would say, intentionally misleading.

The actual situation is that this constituency has been a Labour seat since 1992, and at the last election the vote split like this:

45% Labour
24% Lib Dem
22% Conservative
7 % Green

The boundaries have changed slightly, but the best guess is that it won’t make much difference; it is almost certainly a safe Labour seat, but with the Lib Dems and Conservatives pretty much equal in second place.

The statement which comes closest to being true is ‘Only Labour can keep the Tories out here’. If your main priority is keeping out the Tories, then a Labour vote is the safest option. On the other hand, the statement ‘It’s a straight choice between Labour and the Tories round here’ is pretty close to an outright lie. And it is illustrated with a shocker of a graph:

Yup, that’s right, a local election leaflet showing the ‘share of the vote at the last general election’, and it’s not using the votes cast in this constituency, but the national vote share. Classy.

And the Lib Dems and the Conservatives directly contradict each other. So who’s lying? Well, both of them. Since they both got about the same number of votes last time, it is completely unclear who has more chance to overtake Labour this time. The Lib Dems have gained more in the polls since 2005 than the Tories have, but how it is likely to play out in particular constituencies is totally unknown.

And yet both leaflets have clear bar charts showing why they are the ones to vote for! The Conservative leaflet even has the Lib Dems in 4th place behind the Greens:

And when you look closely, that’s because the chart illustrates the vote for each party’s leading candidate in the council elections to one electoral ward in 2006. Only the leading candidate, out of three, in only one ward out of the eight that make up the constituency. The chart is accurately labelled, but incredibly misleading on a leaflet which is mainly about the parliamentary candidate.

The Lib Dem leaflet uses a less ridiculous but still dubious trick: it shows what purports to be a ‘bar chart’ which illustrates the last general election result in this constituency:

But it’s not really a graph in the mathematical sense at all: the length of the bars is not proportional to the number of votes cast. An accurate chart would have the yellow and blue bars almost exactly the same height, both about half the height of the red bar.

What’s most depressing about this is that it is completely normal. This is how British elections are fought, all around the country, every time. Tactical voting is so deeply engrained in our political culture that we expect to have to vote negatively, and candidates consistently bend the truth to present themselves as the tactical vote of choice.

I am sick of it. I just want to make a choice based on policy and vote accordingly.

More on voting reform

I think this statistic is very telling (from ten days ago, so the exact number has probably changed… but the point stands):

The Lib Dems climbed to a high of 33% in the voting intention polls this week, and it seems that this figure could be higher if Clegg’s party were perceived by the majority to have a significant chance of winning the General Election come May 6th. Just under half the country (49%) would vote for the Liberal Democrats if they were seen to have a reasonable chance of winning.

Ironically, if those 49% actually voted Lib Dem, they would win by a landslide. So we have a situation where if people actually voted for the party they wanted, the election would have a completely different result; but because they are guessing other people’s votes and trying to play the system, they end up cheating themselves.

It’s particularly stark at the moment because for the first time in decades we have three parties roughly even in the polls. But it’s always true at British elections that millions of people say they won’t be voting for their real first choice party because the electoral system would make their vote worthless.

It has been a fact of life in British politics for so long that we don’t even think it’s odd. And if it was simply an inevitable consequence of all democratic systems, well, fair enough. But since various ingenious people have come up with systems that allow people to vote for their real preference without their vote being wasted… doesn’t that seem like quite a good idea?

Close Menu