Why AV reform would still be a big improvement on Westminster


1:21 pm - July 13th 2010

by Mark Pack    


      Share on Tumblr

The possible impact of the alternative vote (AV) on British politics is almost always talked about in the context of seat numbers and tactical voting, with a dash of talk about legitimacy courtesy of AV meaning that each MP has to end up with at least 50% plus 1 of the transferable votes.

Having a decent relationship between vote share and seat numbers is an important part of what a voting system should deliver (and the failure of first past the post to do that is part of what originally made me join the Liberal Democrats).

However, the votes / seats correlation is not the only factor.

If it was, national list PR would be the best electoral system. Rather, the votes / seats correlation is one of a number of factors, factors which at times can be contradictory – and which therefore help keep debates about the most suitable electoral system bubbling along.

So that’s one reason why my eyes tend to glaze over when another seats projection showing how AV might work appears. It’s not just about seats.

A second reason is that I’m dubious about the value of such projections even in their own terms. You don’t just change the way votes are counted when you change an electoral system. You also change the ways parties, the media and the public behave. So asking someone how they would vote if there was an AV election tomorrow is a rather false question because before any actual AV election we’ll see a different dynamic of electioneering.

It may well also be that when faced with an actual ballot paper the public’s behaviour will change in unexpected ways. We’ve already seen how the number of votes for parties other than the major ones shot up on the list element of the Scottish and London devolved elections – far more so than expected in advance.

The third reason is an extension of this. What really interests me about AV is the change in political culture it can bring about. Personally, I would prefer STV because it both brings multi-member constituencies (thus, for example, usually allowing voters to continue to support their preferred party even if there is an individual candidate they really want to vote against such as because of their record on a local issue) and also because it has preferential voting. AV may fail on the former but it too has the latter.

Under preferential voting most candidates hoping to win most of the time have to have an eye on appealing to the second preferences of those who cast a first preference for another party. That imposes a significant burden on the style of politics where you seek out any difference between yourself and another party and inflate it to baby-eating monstrosity levels.

Instead of politics where everything is black and white, AV with its shades of grey when you vote encourages shades of grey in campaigning.

That does not stop passionate disagreement where it’s justified, but it would be a healthy brake on some of the more juvenile styles of politics that we see all too often.

So whilst AV isn’t my first choice of voting system, what it should do for our political culture makes it a big improvement on our current system.

    Share on Tumblr   submit to reddit  


About the author
Mark works at Blue Rubicon and lectures at City University. He also edits Liberal Democrat Newswire - the monthly newsletter about the Liberal Democrats. He is co-author, 101 Ways To Win An Election and blogs here.
· Other posts by


Story Filed Under: Blog ,Our democracy ,Westminster

Sorry, the comment form is closed at this time.


Reader comments


1. Flowerpower

You don’t just change the way votes are counted when you change an electoral system. You also change the ways parties, the media and the public behave.

Fair point. But unless someone comes up with a way of modelling the likely effects, how can we make an informed choice in the referendum? If the outcomes are absolutely opaque, we’re being asked to take a leap in the dark. And the sane response to such an invitation is usually a flat NO.

I’m not in favour of AV because of the effects in terms of who gets into power, I’m in favour because it means I can vote for the party I want to and have that be visible.

I want a voting system where I can vote (for instance) Communist, then Stalinist, then Leninist, and then finally NuLab if none of those get in. Sure, my actual vote will probably still end up with NuLab, but people will know that there is support for Communist policies out there, and there’s a higher chance that they’ll more support the following time.

3. Chaise Guevara

“But unless someone comes up with a way of modelling the likely effects, how can we make an informed choice in the referendum? If the outcomes are absolutely opaque, we’re being asked to take a leap in the dark. And the sane response to such an invitation is usually a flat NO.”

An understandable concern, but one that in practice leads to rejecting progress on principle. Even if we do have a way of modelling the effects, can we trust it? Indeed, will we not see at least two very different models, one from the ‘yes’ camp and one from the ‘no’ camp? By definition, we cannot know for sure what AV will do until it’s implemented, but that isn’t a reason to stay with the current system forever.

‘That imposes a significant burden on the style of politics where you seek out any difference between yourself and another party and inflate it to baby-eating monstrosity levels.’

I suppose it does. By the same token it makes even more appealing the kind of politics where you hide any rational disagreements and differerences, denounce principles as indulgent, pretend that any kind of argument is evidence of ‘tribalism’, refuse to commit to anything in case it’s held against you, and basically try as hard as possible to be a two-faced blank and hope nobody notices you don’t stand for anything in particular and that the people who complain that they’re all the same stay at home as usual.

“But unless someone comes up with a way of modelling the likely effects, how can we make an informed choice in the referendum?”

By looking at the effects in other countries that have changed to or from AV. Papua New Guinea is a commonly-used example. The effects would likely be less extreme in the UK, at least at first, because we have a smaller number of major parties.

6. XerxesVargas

“We’ve already seen how the number of votes for parties other than the major ones shot up on the list element of the Scottish and London devolved elections – far more so than expected in advance.”

AV will undoubtedly result in more votes for the smaller parties. However, what it will not do is allow them any breakthrough. Under AV or AV+ the seats will still, by in large, go to one of the major parties.

AV will make almost no change to the electoral map of the UK. There will still be a focus on marginal constituencies and votes elsewhere will still count less than those.

So, how exactly, is AV better than FPTP?

7. Chaise Guevara

“AV will undoubtedly result in more votes for the smaller parties. However, what it will not do is allow them any breakthrough. Under AV or AV+ the seats will still, by in large, go to one of the major parties.

AV will make almost no change to the electoral map of the UK. There will still be a focus on marginal constituencies and votes elsewhere will still count less than those.

So, how exactly, is AV better than FPTP?”

Because it provides the realistic possibility of smaller parties gaining seats, over the course of a few elections, that they never would have got under FPTP.

Let’s say you want to vote Green, but are in a Labour/Tory swing seat. For tactical reasons, under FPTP you vote for whichever of those two you prefer. Under AV, you still do this, but only after ranking the Greens as your first preference. So all of those Green votes are counted instead of being sacrificed for short-term pragmatism.

Suddenly, the Greens may have a sizeable proportion of the local vote and, in a couple of elections’ time, a decent shot at winning the seat. Under FPTP, they’d languish with a percentage in single figures forever.

“AV will make almost no change to the electoral map of the UK. There will still be a focus on marginal constituencies and votes elsewhere will still count less than those.

So, how exactly, is AV better than FPTP?”

Assumptions, assumptions.

“If the outcomes are absolutely opaque, we’re being asked to take a leap in the dark.”

They’re not opaque though, or at least they can easily not be. You just need to do a study of who people would ideally vote for if they weren’t voting tactically and work out just how that would have an effect on a previous recent election.

My feeling is neither side want to do this for fear of undermining their own argument, either because people aren’t voting as tactically as we believe on the AV side, or because they are and we currently have a government that isn’t actually what the people would prefer.

@7: Let’s say you want to vote Green, but are in a Labour/Tory swing seat. For tactical reasons, under FPTP you vote for whichever of those two you prefer. Under AV, you still do this, but only after ranking the Greens as your first preference. So all of those Green votes are counted instead of being sacrificed for short-term pragmatism. Suddenly, the Greens may have a sizeable proportion of the local vote and, in a couple of elections’ time, a decent shot at winning the seat. Under FPTP, they’d languish with a percentage in single figures forever.

And even if the Greens don’t win the seat, the Tory and Labour candidates know that a lot of people are expressing Green 1st preferences, so they will be more willing to embrace Green policies in order to get the lower preference votes of those voters.

The ‘national list’ PR is one of the reasons I’m so firmly against PR. If we must change then the system then at the very least we need to keep the direct accountability of an MP to her/his electorate (rather than their party). If AV ensures the continuation of this and manages to stop the whiny ‘wasted vote’ (note: It’s not a wasted vote just because you didn’t ‘win’) then there may be something to be gained.

I have yet to see one reasonable argument against making votes more fair. However, I have in recent weeks seen some unreasonable arguments against fairer votes, among them:

1. “I am against fairer votes because it decreases the likelihood of my favoured party winning outright” or, “because parties I don’t like will have more power”;

2. “I am against fairer votes because it would lead to more compromises and fewer of my favoured policies”.

We can’t even make constituencies more equal without people moaning about how it disadvantages them or some other, quite possibly irrelevant group. On the other hand they don’t mind that their current advantage is another’s disadvantage.

There was even the argument that the votes of “the poor” ought to have greater value than “the rich” because the poor are more disadvantaged.

There seem to be a number of people who claim to be in favour of “democracy” but underneath they are only if favour of it so long as their favourites win.

“And even if the Greens don’t win the seat, the Tory and Labour candidates know that a lot of people are expressing Green 1st preferences, so they will be more willing to embrace Green policies in order to get the lower preference votes of those voters.”

You’ve hit it square on the head. In fact could we argue that the lurch by Labour and Lib Dems more rightward on their policies can be attributed to the sheer numbers of Tory support in the country under a FPTP system?

13. Chaise Guevara

“1. “I am against fairer votes because it decreases the likelihood of my favoured party winning outright” or, “because parties I don’t like will have more power”;

2. “I am against fairer votes because it would lead to more compromises and fewer of my favoured policies”.”

Which basically translates as “if we let people vote for who they want they’ll vote for the wrong people”.

I’m sick of this stuff too, UKliberty. Next person who tells me we should keep FPTP because the BNP might gain a few seats under a fairer system gets a poke in the eye.

One thing of interest to smaller parties, specifically the Greens, but also others hoping to breakthrough.

Short Money. Paid for to opposition parties based on them having seats in Parliament and their share of the vote nationally. So if Lucas retains her seat, and the Greens grow a bit more in other seats to potentially gain a few more elsewhere (my local Green candidate would’ve got my first or second pref under AV, no doubt at all there), then the number of 1st pref votes they get will give them Short Money, regardless of where the votes are cast.

For a very small party like the Greens, some money to fund policy research and development would be very useful for them I suspect, and disproportionately so.

Of course, the whole funding arrangement should be redesigined, but I don’t think that’s on the table short term.

But yes, the need to appeal across the lines, try to recruit 2nd, 3rd and even 4th preferences will make a significant change medium term; although I suspect it’ll take awhile to get used to.

15. Richard P

@MatGB, interesting point. I see that even under the current system, the Greens are already entitled to a certain amount of Short Money, totally £51,000 for May 2010-Mar 2011, about a tenth of what Labour’s getting (£5.2m). The SDLP gets £55,000, Plaid £62,000, the DUP £129,000, the SNP £146,000. ( http://www.parliament.uk/documents/commons/lib/research/briefings/snpc-01663.pdf )

16. Richard P

Sorry, obviously I meant a hundredth, not a tenth, of the Labour total – a big difference and one that also makes more sense.

Ah, see I didn’t know where to go look it up. Crucially, the Green vote got squeezed a lot in many seats up and down the country this time. Locally, for example, their vote in the locals more than doubled, but their vote in the general halved down to just over 1%, despite an awesome candidate.

Guarantee that locally they’d have got at leat 5% under AV, possibly even more, and if they’d managed to get to every hustings/public meeting, the candidate could’ve made it into real contention (3-way marginal already).

It’ll also, naturally, help UKIP on the right, especially in ‘safe’ Tory seats. And giving them some money to do some actual research. Then they might come up with some reality based policies…

“Fair point. But unless someone comes up with a way of modelling the likely effects, how can we make an informed choice in the referendum?”

Someone has.

http://rangevoting.org/BayRegsFig.html

“AV” is listed on this graph under a different name, instant runoff voting.

“Plurality” is the current system.

As you can see, any improvement hinges on how honest voters are; honesty is typically estimated to be about 10% (most voters are at least somewhat tactical), so the net-improvement of AV would be minimal.

If you want a REAL big improvement, approval voting and range voting should be looked at.

19. Chaise Guevara

Interesting. Why are the Greens and SDLP so close in terms of funding when the latter have three times as many seats?

Chaise, it’s a formula based on votes cast and seats gained. Greens got 3* the votes, despite being squeezed (although it appears not as much as I’d thought).

Aha, from Richard’s link

General funding for Opposition Parties – the amount payable to qualifying parties
from 1 April 2010 is £14,351 for every seat won at the last election plus £28.66 for every
200 votes gained by the party.

So, assuming we’re right that smaller parties will get increased first pref shares in the same way they increase in Euro and List elections like London Assembly elections, that’ll really help increase diversity in policy ideas.

Um, it’ll also hurt the Lib Dems a fair bit I suspect, although the seats gained proportion might go up even with decreased vote shar eon first prefs.

Anyway, slightly derailing Mark’s point, but it’s a complementary argument.

Dale, I’m happy to ‘look at’ any electoral system. But as things stand, only AV or the status quo are on the table. There isn’t enough legislative support for anything else, and if this reform is blocked, further reform is completely off the table.

22. Chaise Guevara

“Chaise, it’s a formula based on votes cast and seats gained. Greens got 3* the votes, despite being squeezed (although it appears not as much as I’d thought).”

Ah, makes sense. Thank’ee kindly.

23. Richard P

Re Approval Voting and Range Voting. The Electoral Reform Society has a page on Approval Voting ( http://www.electoral-reform.org.uk/article.php?id=59 ) and the arguments against it (e.g. “can give strange results and is very dependent on voter tactics rather than voter preferences”) seem to me compelling. Range Voting similarly encourages strategic voting. I suspect it would have much the same negatives as Borda counts, which “encourage tactical voting to an even greater extent than First Past the Post” ( http://www.electoral-reform.org.uk/article.php?id=75 ).

My own preferred system would be a variation upon AV+ (I think the main flaw with AV+ is that we need to add the facility to select a second pref on the list section, so if one’s preferred list doesn’t get enough votes, one’s list vote gets transferred to a different list; it is also important to have an open or at least partially open list, and finally I’d probably have a slightly larger list top-up than the deliberately small one suggested by Jenkins).

But we should support AV as a step in the right direction. Let’s not fool ourselves that it will bring about a revolution (Australia is one of the few countries with AV and pretty much has a two-party system – the Liberals and Nationals being very similar, in permanent alliance on national level and considering merger – in fact Australia’s lower house has less variety in party terms than the House of Commons has). But it greatly reduces the need for tactical voting and makes it difficult for a rightwinger to win merely on the strength of a split left vote (and, admittedly, difficult for a leftwinger to win on the strength of a split rightwing vote).

In addition, with AV, we can still push for a move to AV+ at some future point. It’s not going to be easy, though I suspect to push for a proportional system after AV has been defeated in a referendum would be even harder.

unless someone comes up with a way of modelling the likely effects, how can we make an informed choice in the referendum?

The moon on a stick.

25. Richard P

Mark says, “each MP has to end up with at least 50% plus 1 of the transferable votes.”

Not true, though perhaps it’s an acceptable simplification. We would almost certainly find that some MPs in close contests were still elected on a minority vote under AV. Not everyone would bother to express multiple preferences, and in some cases supporters of an eliminated candidate would express a second or even third preference for candidates who had already been eliminated or who would later be eliminated. As far as I know (though I’m not sure that Clegg has brought forward detailed proposals) there is no suggestion that we would hold a second election in such cases. The recipient of the largest minority of votes would be elected once all the lower candidates had been eliminated and all the possibilities for lower-preference-vote distribution exhausted.

Richard, Marks says “of the transferable votes” so he is actually correct. If you don’t express enough preferences, your ballot is eliminated, the winner has to have 50% of the votes left at any stage.

So not even a simplification, but an accurate description that perhaps doesn’t go into the detail some of us prefer, but is workable and accurate.

Richard: I was using “transferable” to mean those votes which end up being transferred to one or other of the final two candidates left in contention (which I think is the same as the point you’re making?). I thought I’d go for one word rather than a whole sub-clause 🙂

Ah yes, I see that now – thanks.

“Re Approval Voting and Range Voting. The Electoral Reform Society has a page on Approval Voting ( http://www.electoral-reform.org.uk/article.php?id=59 ) and the arguments against it (e.g. “can give strange results and is very dependent on voter tactics rather than voter preferences”) seem to me compelling. Range Voting similarly encourages strategic voting.”

There is no qualitative support for this argument; just someone’s intuition.

The simulation data I referred to earlier is a quantitative refutation of that intuition.

http://rangevoting.org/BayRegsFig.html

I think e-r.org’s confusion stems from the fact that with approval and range voting, you can very easily cause your second-favorite to win at the expense of your first-favorite; and sure, that’s bad. But it neglects the fact that under every other voting system (plurality, IRV, etc.), you can cause your least favorite to win at the expense of your first-favorite. The “tacticalness” involved is un-familiar, and so seems quite egregious, while the worse tactical choices under familiar systems are “just how it is”, and ignored.

I suggest looking Stephen J. Brams take on the issue.

Richard P,

Your comment from the Electoral Reform society, ““can give strange results and is very dependent on voter tactics rather than voter preferences”, is kind of backwards.

IRV gives much more bizarre counter-intuitive results.

http://scorevoting.net/IrvPathologySurvey.html
http://www.electology.org/debate/IrvPlurality

Dale’s link to the Bayesian regret figures shows a quantifiable measure of the sum of the negative effects of the misbehaviors of various voting methods. This means taking the frequency with which each misbehavior happens, times its negative impact in terms of decreasing voter satisfaction (“utility”). Score Voting and Approval Voting come out way ahead of IRV (aka “alternative vote”).

Most of the arguments about “weird results” with Approval Voting turn out to be flawed if you treat them as logical axioms and evaluate them for consistency.

@11 ukliberty: There seem to be a number of people who claim to be in favour of “democracy” but underneath they are only if favour of it so long as their favourites win.

Well said!

32. Richard P

Clay & Dale, thanks for the links. Are you opponents of PR, or can range voting be operated in multi-member constituencies?

Obviously some of rangevoting.org’s criticisms of Alternative Vote are stronger than others – I think we can avoid hanging chads by not using punch machines to vote, and I think computerised counting should be avoided anyway.

The non-monotonicity of AV is interesting. Someone on ukpollingreport recently denied that AV was non-monotonic, and I had a feeling they were wrong but couldn’t quite remember why. http://en.wikipedia.org/wiki/Monotonicity_criterion has a good explanation.

In http://www.rangevoting.org/Burlington.html it is argued that range and approval voting would probably have elected the Democratic candidate, whereas AV caused the more leftwing Progressive Party candidate to win because the Democrat was eliminated first.

I obviously haven’t had chance as yet to read all these materials so forgive me if this has been answered. But does it concern you that range and approval voting would seemingly have this pro-centrist or pro-moderate effect? Don’t we actually have a more effective democracy if we have leftwingers and rightwingers arguing it out? I know that both FPTP and AV have been accused of encouraging politicians to adopt centrist policies to maximise their chances – wouldn’t this be still truer of approval voting?

@17 MatGB: Guarantee that [the Greens]’d have got at least 5% under AV. […] It’ll also, naturally, help UKIP on the right, especially in ‘safe’ Tory seats

You’re right. If you look at the last election where Green and UKIP voters knew their votes wouldn’t be wasted, they got 9.1% and 16.5%.

Under AV, these parties would probably each get 4-10% of 1st preverence votes (they’d do a little less well in a general election than a European one, since voters take it more seriously).

@RichardP

Warren Smith, the Princeton math Ph.D. who did the Bayesian regret calculations, also invented two proportional methods: Reweighted Range Voting and Asset Voting. They are simpler than, and in some objective ways superior to, STV.
http://scorevoting.net/RRV.html
http://scorevoting.net/Asset.html

Interestingly enough, it turns out that Asset Voting was originally invented by an Oxford mathematician named Charles Dodgson (aka Lewis Carroll, the author). And RRV actually seems to have been originally invented by a Danish mathematician in the 1800’s (but totally ignored in his day).

A somewhat USA-centric broad take on PR is here:
http://scorevoting.net/PropRep.html

PR seems like it can be a very big improvement, if done correctly. I don’t think party list and MMP systems are very good forms of PR. I think STV is decent, but its single-winner form is pretty poor. But because so many elections are inherently single-seat elections, I think it’s still critically important to focus on single-winner voting methods, like Score Voting (aka Range Voting) and Asset Voting.

Your point about paper ballots is appreciated, but keep in mind that IRV (sorry, to me “AV” is Approval Voting) tends to result in more spoiled ballots, and it cannot be subtotaled in precincts because it is not additive.
http://scorevoting.net/IrvNonAdd.html

I’ve been told by a guy named Tony Kennick that this is less of an issue in the UK, for some complex reasons. I live in San Francisco, so I don’t know… he may be right.

I don’t focus too much on monotonicity, because it’s already encompassed (along with every other criterion imaginable) by Bayesian regret. However, our page on the subject is informative.
http://scorevoting.net/Monotone.html

@RichardP

Warren Smith, the Princeton math Ph.D. who did the Bayesian regret calculations, also invented two proportional methods: Reweighted Range Voting and Asset Voting. They are simpler than, and in some objective ways superior to, STV.
http://scorevoting.net/RRV.html
http://scorevoting.net/Asset.html

Interestingly enough, it turns out that Asset Voting was originally invented by an Oxford mathematician named Charles Dodgson (aka Lewis Carroll, the author). And RRV actually seems to have been originally invented by a Danish mathematician in the 1800’s (but totally ignored in his day).

A somewhat USA-centric broad take on PR is here:
http://scorevoting.net/PropRep.html

PR seems like it can be a very big improvement, if done correctly. I don’t think party list and MMP systems are very good forms of PR. I think STV is decent, but its single-winner form is pretty poor. But because so many elections are inherently single-seat elections, I think it’s still critically important to focus on single-winner voting methods, like Score Voting (aka Range Voting) and Asset Voting.

Your point about paper ballots is appreciated, but keep in mind that IRV (sorry, to me “AV” is Approval Voting) tends to result in more spoiled ballots, and it cannot be subtotaled in precincts because it is not additive.
See IrvNonAdd

I’ve been told by a guy named Tony Kennick that this is less of an issue in the UK, for some complex reasons. I live in San Francisco, so I don’t know… he may be right.

I don’t focus too much on monotonicity, because it’s already encompassed (along with every other criterion imaginable) by Bayesian regret. However, our page on the subject is informative.
See Monotone

> does it concern you that range and approval voting would seemingly have this pro-centrist or pro-moderate effect? Don’t we actually have a more effective democracy if we have leftwingers and rightwingers arguing it out?

Well, I think it’s complicated. “Moderate” will mean different things in different districts. A more urban area will probably elect someone with markedly different priorities than the winner from a rural agrarian area. And in that case, a voter who thinks there’s a downside to electing moderates can just appropriately reduce support for moderates, and then if enough other voters agree with him, parliament (or whatever) will have fewer moderates and more partisans arguing it out.

That probably is more of an academic argument than something that really addresses your practical concern. So yes, I would say that for multi-winner legislative bodies, it seems like PR is more appropriate. Whereas for e.g. mayor or prime minister (“single-seat” elections), it seems we just want to elect the “social utility maximizer”, even if he/she is a moderate.

Incidentally, you might enjoy these computer election simulations graphs, demonstrating the pro-extremist tendency of IRV. Could this be worse than being pro-moderate?
See IrvExtreme

> I know that both FPTP and AV have been accused of encouraging politicians to adopt centrist policies to maximise their chances – wouldn’t this be still truer of approval voting?

Well, to an extent, sure. No matter what system you use, candidates will want to make themselves appear to be in whatever segment of political issue space that gives them the best chances, whether that’s the centroid or some extreme.

The important benefit in Approval Voting is that tactical voting is less harmful. Say you sincerely feel W=10, X=8, Y=4, Z=0. If the two strongest candidates are Y and Z, then you’d of course want to start by voting for Y, like you would under FPTP (assuming you’re being maximally tactical). But then you could go ahead and vote X and W as well. Essentially you vote for the candidate you’d vote for with FPTP, plus *every candidate you like beter*.

That is tremendously important, because it means that if enough voters prefer W or X to the “frontrunners”, Y and Z, then W or X can win even if voters don’t think they can win. Since candidates don’t have to prove they CAN win, only that you SHOULD vote for them, they put lower priority on raising insane amounts of money and trying to get the nomination of a major party, and higher priority on talking about the issues.

It may be hard for you to understand the USA perspective on this stuff however, because I get the sense that the political system and the voters are more sane in the UK. Money maybe isn’t quite as all-important as it is here. I could be wrong though. 🙂

@RichardP

I was unable to post the rest of my lengthy comment. Feel free to email me, clay@brokenladder.com if I can send the rest to you.

Thanks

Richard P

But does it concern you that range and approval voting would seemingly have this pro-centrist or pro-moderate effect?

Why is this a bad thing? (aside from the obvious that if you aren’t centrist or moderate you won’t be happy)

39. Richard P

ukliberty,

“Why is this a bad thing? (aside from the obvious that if you aren’t centrist or moderate you won’t be happy)”

I can think of three possible arguments. It might be held that it is more important to give some people what they definitely want rather than to elect everyone’s second-best – in other words, majoritarianism versus consensus politics. (I realise that Alternative Vote is far from perfect.) So in the Burlington example cited above, the leftist candidate is seemingly the preferred option of the largest single minority, as well as receiving adequate support from the majority once second and later preferences are distributed, whereas the Democrat is (on the face of it) the third-placed candidate – some people would argue that the Democrat should be elected as the one least offensive to the largest number, but others might disagree.

Secondly it might be held that having more radical voices in Parliament allows left and right to battle it out, which we might prefer to being dominated by consensus-oriented moderates.

Thirdly there is the proportional argument. (Of course, Alternative Vote isn’t proportional.) It might be felt better, if you have a population that is partly rightwing, partly leftwing and partly centrist, to represent each group in Parliament so that each strain of opinion can have some input, rather than having centrists all over the place on the grounds that neither right nor left object to the centre as strongly as they object to each other.

There is (mentioned above by Clay) a proportional form of range voting that I’ve not had chance to read upon yet.

Phil,

Under AV, these parties would probably each get 4-10% of 1st preverence votes (they’d do a little less well in a general election than a European one, since voters take it more seriously).

I’d actually go slightly further. Traditionally psephologists have taken the lazy way out and assumed that lower turnout affected all parties equally, I’m 100% certain now that that’s not the case, especially given the last Euro results where the Labour vote disproportionately collapsed, and from looking at low turnout vs high turnout constituencies on my ongoing safe seats are bad analysis.

I posit that in the Euros, most if not all UKIP and Green supporters turn out and vote. I’m pretty sure of that for UKIP, as it’s “their” election, and fairly sure of it for the Greens. Therefore Greens would get lower end of your scale 1.2M is 4% of 29M, and UKIP would get about 8.4% on 2.5M.

Of course, we are all blowing smoke, it’s completely impossible to predict, but even on just 1.2M, Green Short Money would be nearly £200K, assuming they keep Lucas. That’ll give them some serious research money to spend, which would be very useful for politics generally. And that’s assuming they don’t do better in some areas, with, for example, more LD voters switching to them.

Clay/Dale, as I said above, while I, personally, am very interested in many other voting systems, and my preference for STV is because of the results it gives and the way it works, no other option is currently on the table nor is it likely to be before the referendum.

If we don’t win the AV referendum, no further reform is ever going to happen unless Greens and LDs between them get an overall majority. Ergo, not joinging into this discussion as, essentially, I think it’s derailing to talk about pipedreams when we have to concentrate on practicalities. In the next Parliament, and int he run up to the next General Election, this sort of discussion is very useful. It may even be useful for a discussion on the House of Lords, for which no system decision has been agreed yet. But for the HoC, it’s AV or no change.

A somewhat USA-centric broad take on PR is here:
(scorevoting.net)PropRep(.html)

PR seems like it can be a very big improvement, if done correctly. I don’t think party list and MMP systems are very good forms of PR. I think STV is decent, but its single-winner form is pretty poor. But because so many elections are inherently single-seat elections, I think it’s still critically important to focus on single-winner voting methods, like Score Voting (aka Range Voting) and Asset Voting.

Your point about paper ballots is appreciated, but keep in mind that IRV (sorry, to me “AV” is Approval Voting) tends to result in more spoiled ballots, and it cannot be subtotaled in precincts because it is not additive.
See IrvNonAdd

I’ve been told by a guy named Tony Kennick that this is less of an issue in the UK, for some complex reasons. I live in San Francisco, so I don’t know… he may be right.

I don’t focus too much on monotonicity, because it’s already encompassed (along with every other criterion imaginable) by Bayesian regret. However, our page on the subject is informative.
See Monotone

> does it concern you that range and approval voting would seemingly have this pro-centrist or pro-moderate effect? Don’t we actually have a more effective democracy if we have leftwingers and rightwingers arguing it out?

Well, I think it’s complicated. “Moderate” will mean different things in different districts. A more urban area will probably elect someone with markedly different priorities than the winner from a rural agrarian area. And in that case, a voter who thinks there’s a downside to electing moderates can just appropriately reduce support for moderates, and then if enough other voters agree with him, parliament (or whatever) will have fewer moderates and more partisans arguing it out.

That probably is more of an academic argument than something that really addresses your practical concern. So yes, I would say that for multi-winner legislative bodies, it seems like PR is more appropriate. Whereas for e.g. mayor or prime minister (“single-seat” elections), it seems we just want to elect the “social utility maximizer”, even if he/she is a moderate.

Incidentally, you might enjoy these computer election simulations graphs, demonstrating the pro-extremist tendency of IRV. Could this be worse than being pro-moderate?
See IrvExtreme

> I know that both FPTP and AV have been accused of encouraging politicians to adopt centrist policies to maximise their chances – wouldn’t this be still truer of approval voting?

Well, to an extent, sure. No matter what system you use, candidates will want to make themselves appear to be in whatever segment of political issue space that gives them the best chances, whether that’s the centroid or some extreme.

The important benefit in Approval Voting is that tactical voting is less harmful. Say you sincerely feel W=10, X=8, Y=4, Z=0. If the two strongest candidates are Y and Z, then you’d of course want to start by voting for Y, like you would under FPTP (assuming you’re being maximally tactical). But then you could go ahead and vote X and W as well. Essentially you vote for the candidate you’d vote for with FPTP, plus *every candidate you like beter*.

That is tremendously important, because it means that if enough voters prefer W or X to the “frontrunners”, Y and Z, then W or X can win even if voters don’t think they can win. Since candidates don’t have to prove they CAN win, only that you SHOULD vote for them, they put lower priority on raising insane amounts of money and trying to get the nomination of a major party, and higher priority on talking about the issues.

It may be hard for you to understand the USA perspective on this stuff however, because I get the sense that the political system and the voters are more sane in the UK. Money maybe isn’t quite as all-important as it is here. I could be wrong though. 🙂

> I think it’s still critically important to focus on single-winner voting methods, like Score Voting (aka Range Voting) and Asset Voting.

I meant to say “and Approval Voting”, as Asset is a PR method.

43. Richard P

Clay, I assume the above is the long reply that got cut off last night, which you’d earlier offered to email (which I hadn’t got round to taking you up on). Thank you, very comprehensive.

I think the prevalence of inherently single-winner elections is somewhat smaller in the UK. Many towns don’t have directly elected mayors, and the prime minister isn’t directly elected either (except as an MP in his or her local constituency).

“We now have a once-in-a-lifetime opportunity to right the wrongs of the archaic First Past The Post system.”
http://www.fairervotes.org.uk/

Really? I thought it was supposed to be a first step? One more reason to insist on an amendment to the bill, to give voters a third option, a proportional system.

“Papua New Guinea is a commonly-used example.”

Not in this part of Ealing.

46. XerxesVargas

@Chaise Guevara

Bit late now, i know, Sorry.

I probably should have said that I’m not for FPTP but full PR.

Any how

“Suddenly, the Greens may have a sizeable proportion of the local vote and, in a couple of elections’ time, a decent shot at winning the seat. Under FPTP, they’d languish with a percentage in single figures forever.”

So AV may help some smaller parties win a seat in 15-20 years time or so? Come on. Its nothing more than a sop and as I said will change very little – not nothing, just not much. In terms of the overall make up of Parliment it will still be the domain of the big 3.

@ Phil Hunt

“And even if the Greens don’t win the seat, the Tory and Labour candidates know that a lot of people are expressing Green 1st preferences, so they will be more willing to embrace Green policies in order to get the lower preference votes of those voters.”

But unless there is a national movement to the Greens, for example, then you will get a patchwork of influence. So one Labour/Tory MP may appear greener to win those votes. Whilst in the constituency up the road an MP from the same party has UKIP in first place. So he picks up the language of the euro-sceptic. But at a national party level they cancel each other out and nothing changes.

Plus you are still coming at it from an acceptance that the major parties will still be the major parties and dominant.

Xerxes, AV is slightly better than FPTP in terms of smaller party representation, for a start it elminates the ‘split vote’ problem, and it may allow smaller parties to develope in a stronger way, whereas FPTP specifically discourages them.

It’s not actually very good, but it’s the only game in town, if we reject AV, then that’s it, we’ve rejected electoral reform, it’s off the table.

Do I want more? Hell yes. Ander AV, we can show there’s strong desire for further change, if we reject it, any attempts at more change will be dead unless one of the big two parties commits to proper reform (like Labour did in 1997) and then actually carries it out (like Labour failed to do in 1997).

Crucially, if your preferred electoral system is STV based on local authority or county boundaries, as mine is, AV is a step in the right direction, but it’s also better than FPTP for smaller parties.

It’s this or nothing. I don’t want this, I want more, but I’d rather some reform than none, it’ll strengthen those parties and candidates that want more reform.

Importantly, it also allows, mid term, for the breakup of the Big Two, less likely to happen under AV than STV, but that’s also a longer term objective.

@MatGB

IRV does not better in terms of smaller party representation. It has been used in e.g. Australia (House of Reps) and Ireland (Presidency) since the early 1900’s, and has maintained two-party domination.

IRV does not eliminate the “spoiler” (aka “vote splitting”) problem. Burlington VT (USA) demonstrated this in their last mayoral election, where the Democrat would have beaten the Progressive if the Republican had been removed from the ballots. (The Republican was a “spoiler” and split the “prefers-Democrat-to-Progressive” vote.)

I don’t know what you mean “it may allow smaller parties to develope in a stronger way, whereas FPTP specifically discourages them.” In tactical respects, IRV is essentially the same as FPTP.

It’s not actually very good, but it’s the only game in town, if we reject AV, then that’s it, we’ve rejected electoral reform, it’s off the table.

That, unfortunately, seems to be a valid point.

49. roberto c

@Clay

That might all very well be the case, but you do know that 99.999% of people eligible to vote in the AV referendum don’t give a rat’s ass about all of that?

Any change towards something (even marginally) better than FPTP in even a single aspect (such as requiring each MP to have a majority of votes) should be welcomed by reformers when it is either that or the status quo.

If you whine and convince people it’s not worthing changing the status quo even a bit, then they won’t ever want to change the status quo a lot.

Clay, in Australia, it does not maintain a two-party dominance. There are two competing groups, but the Coalition parties compete against each other in the districts where they are strong.

That’s one of the reasons it was introduced, to allow them to do so. Sure, they’ve a pre-existing agreement to work together in most cases, but they don’t always.

Regardless, AV is the first step towards STV, and it is the only game in town. In the UK, there are 3 nationally effective with two challengers that will be able to breakthrough in some areas (UKIP and the Greens), plus in Wales & Scotland there’s a 4th effective party, and in Northern Ireland there’re 5 effective parties.

In that situation, AV will strengthen the pre-existing challenging parties (including, also, the Lib Dems) and thus help break the existing duopoly. And it is only a short term measure, if it last without further reform for more than two General Elections I’ll be surprised and disappointed.

Aside, I said I wouldn’t get into the other systems debate, but I’d like to have one point clarified, earlier you said

I think STV is decent, but its single-winner form is pretty poor

STV is always multi-member, by “its single winner form” are you referring to Alternative Vote? In which case I do, actually, agree with you.

I don’t like AV, it’s a minimal reform, but importantly, it’s a step in the right direction, which is what matters.

51. Richard P

“in Australia, it does not maintain a two-party dominance. There are two competing groups, but the Coalition parties compete against each other in the districts where they are strong.”

I’m not sure – it’s sufficiently close to a two-party that at least one pollster seems to ask about who will vote “Liberal/National” instead of asking about the parties individually ( http://www.angus-reid.com/polls/view/australian_opposition_maintains_lead/ ). The Nationals are currently running at less than 5% in most polls ( http://en.wikipedia.org/wiki/Australian_federal_election,_2010 ). The Greens don’t currently have any seats – I guess they may win some this time round, but as AV is non-proportional, it’s not certain.

Liberals and Nationals have actually already merged in Queensland. True, there are some states where they’re less close.

But the Nationals (the third party, if you count them as a separate party, which a lot of people don’t seem to) have only 10 seats in the 150-seat Parliament. That means Australia’s lower house is more two-party-dominant than the UK’s!

@Robert,

Regardless of whether IRV is (at least marginally) better than FPTP, many of the common claims in its favor are simply inaccurate. I was just clarifying that.

53. xerxesvargas

@MatGB

I don’t disagree with you at all and on balance I would rather AV than FPTP. But it is a pale version of electoral reform.

Richard P @39, good points. I’m very much inclined to agree although I haven’t yet grasped the value of left and right battling.

55. Chaise Guevara

“I don’t disagree with you at all and on balance I would rather AV than FPTP. But it is a pale version of electoral reform.”

Agreed. I have my qualms about PR (mainly that I don’t like the idea of politicians as party-selected yes men, and suggestions about how to avoid that seem painfully complicated), but I’m not going to argue against it, because it’s patently a fairer system. And there are also advantages to politicians becoming less autonomous (see the abortion limit debate awhile back, when MPs demanded they be free to vote based on their own consciences).

Still, AV will do for now, and we’ll be lucky to get that.

56. Richard P

Chaise, you say “I don’t like the idea of politicians as party-selected yes men, and suggestions about how to avoid that seem painfully complicated”.

Surely STV would avoid this (as would Reweighted Range Voting) because you’d continue to vote for individuals rather than parties. Working out the results might be complicated, but casting the votes wouldn’t be, surely?

In the case of PR lists, I see several options:
* we could have require parties to have an internally democratic process for selecting and ordering candidates (although personally, I wouldn’t favour compulsion)
* we would have primary elections to determine the ordering of candidates (personally again, I’d leave this up to the parties to decide if they wanted it)
* we could have open or partially open lists.

Open lists allow a free choice of candidates from among lists. In some variations, voters can choose from multiple lists.

Partially open lists are less democratic than fully open, so their only advantage is in their greater simplicity. You would get the option to vote for an individual candidate (pushing them up the list, if sufficient voters do the same). Another variation (formerly used in Sweden but no longer) allows voters to cross a candidate’s name off the list, pushing them down (if sufficient numbers of voters do the same).

Actually, it’s noteworthy that the voting reform debate so far has completely ignored certain questions:
* will we get voting reform for English local government? Scotland already has STV for local council elections.
* can we change the voting system for British elections to the European Parliament? Few if any British voting reformers are fond of closed lists, yet Great Britain uses them (Northern Ireland doesn’t) for elections to the EP. They could be replaced by STV or by a more open list system. (NI uses STV for the same elections, even though their three MEPs form part of the UK delegation.)

57. Chaise Guevara

“Surely STV would avoid this (as would Reweighted Range Voting) because you’d continue to vote for individuals rather than parties. Working out the results might be complicated, but casting the votes wouldn’t be, surely?”

Yes, but STV is basically AV, not PR. I think AV is great, slightly beautiful in fact, but it’s not representative.

“In the case of PR lists, I see several options:
* we could have require parties to have an internally democratic process for selecting and ordering candidates (although personally, I wouldn’t favour compulsion)
* we would have primary elections to determine the ordering of candidates (personally again, I’d leave this up to the parties to decide if they wanted it)
* we could have open or partially open lists.”

This first option is not democratic. The second sounds overcomplicated, though I could well be wrong. I confess I don’t follow the third one, not being a student of the voting process. Could you layman it up for me a bit?

As to your point on English government: I’m pretty sure we already voted against it, although that’s obviously not a reason not to vote again. I’m personally against devolution in general, but I can see why people get annoyed at the current disparities.

@Chaise Guevara

STV is basically AV, not PR. I think AV is great, slightly beautiful in fact, but it’s not representative.

Incorrect. STV is a multi-winner proportional voting method. IRV (aka “alternative vote”) is the single-winner (thus non-PR) form of STV.

But I think Reweighted Range Voting and Asset Voting are better and simpler.

“STV is basically AV, not PR. I think AV is great, slightly beautiful in fact, but it’s not representative.”

That’s inaccurate. STV is a form of PR. It’s not 100% proportional, and it tends to be less proportional than list-based systems, but it is pretty proportional. The degree of proportionality depends how many seats you have in each multi-member constituency. Three-member constituencies would yield rather poor proportionality, six-member very good. In the most recent elections to the Irish Dail, Fianna Fail won 47% of seats with 42% of the vote; Fine Gael won 31% of seats with 27% of the vote; Labour won 12% of seats with 10% of the vote; the Greens 4% of seats with 5% of the vote; and Sinn Fein 2% of seats with 7% (perhaps because that 7% is the proportion of first preferences, and maybe Sinn Fein are less likely to pick up second and later preferences than the other parties).

The fact that STV is multi-member whereas AV is single-member isn’t a mere technical difference. It is what allows STV to be proportional.

STV is actually more complicated than AV, because with STV, there is the concept of “surplus votes” where if a candidate gets more votes than she or he needs to win a seat, the second preferences of the surplus voters are distributed to other candidates.

In the case of List systems (or the list component of AMS and AV+ systems), there are various ways that voters can be given a say in which individual candidates are elected from the lists (rather than relying on the order determined by each party). Admittedly, none of the methods is fully satisfactory (since there is a trade-off where you either have a very democratic but very complex system, or a less democratic but simpler system).

In Finland you are asked to vote for a candidate from the list rather than for the list as a whole. The number of votes received by each candidate determines how high up he or she is on the list.

In Sweden you can choose between voting for the list as a whole or for an individual candidate. If 8% or more of voters vote for an individual candidate, that candidate moves to the top of the list. If more than one candidate has received 8% of votes, the candidates are re-ordered according to how many individual votes each received.

Other variations are used in Denmark, Norway, and the Netherlands.

60. Chaise Guevara

Happy to be corrected on this one. I’d been under the impression that STV was basically AV restricted to two preferences. Like I say, I’m not a student of this stuff.

Can you extrapolate on open/closed lists etc? I can guess at the meaning, but I suspect my guess is wrong.

61. Chaise Guevara

Arrgh! Be careful what you wish for…

Thanks Rich P. I’m going to take a proper look at that tomorrow. It’s late, and I’ve a had a couple of beers, and I’m in no fit state to take this in properly. For now, the only thing that seems clear is that I’ve misunderstood or misapplied STV. Thanks for taking the time to set things out.

62. Chaise Guevara

Crap. 60 is a response to Clay, not Richard P. Wasn’t expecting many people to be posting at this hour.

AV restricted to two preferences would be SV (supplementary vote), the system used in mayoral elections.

With a “closed” list system, the voter votes for a list, with no ability to express a preference for or against any particular candidate on the list. The ordering of the list is thus entirely in the hands of the party. So if the list gets sufficient votes for three candidates to be elected, those three will come from the top of the list (as ordered by the party).

The various open and partially open variants are attempts to give the voter a degree of control over the ordering of the list – or complete control in the Finnish case (though rather hampered by the fact that the ordering of the list is being determined by each voter having a single vote, with all the FPTPesque problems that potentially raises!).

Chaise, I suspect you’ve mixed up STV with Supplementary Vote, which is the bloody stupid voting system Labour lumbered us with in London. And this hour implies you think it’s late. Welcome to the internets, some of us are just waking up 😉

Also, it’s Thursday, and the Question Time panel is awful. Again.

Lords and ladies, to the average punter in the voting booth, AV is STV – the voting process is the same, you have a selection of candidates from each party to choose from (as many candidates from each party as there are MPs being elected in your multi-member constituency), you rank them in order, some get elected.

Yes, the fact that STV requires multi-member constituencies is what makes it proportional – the more MPs elected for each constituency, the more proportional the vote can be. The most proportional voting system would be one where there is one constituency and all 650 MPs elected in it.

@Roberto C

you have a selection of candidates from each party to choose from (as many candidates from each party as there are MPs being elected in your multi-member constituency

There is no reason that the number of candidates should be limited to the number being elected. That’s a completely arbitrary ceiling.

67. Richard P

61. No problem, Chaise, hope it is a little clearer now.


Reactions: Twitter, blogs
  1. Liberal Conspiracy

    Why AV reform would still be a big improvement on Westminster http://bit.ly/8XsJYi

  2. Chris Wiggin

    RT @libcon: Why AV reform would still be a big improvement on Westminster http://bit.ly/8XsJYi

  3. Andrea

    RT @libcon: Why AV reform would still be a big improvement on Westminster http://bit.ly/8XsJYi

  4. Kai

    RT @libcon: Why AV reform would still be a big improvement on Westminster http://bit.ly/8XsJYi

  5. Richard Davis

    RT @libcon: Why AV reform would still be a big improvement on Westminster http://bit.ly/8XsJYi

  6. Naadir Jeewa

    Reading: Why AV reform would still be a big improvement on Westminster: The possible impact of the alternative vot… http://bit.ly/brkYkt

  7. Dan Sutton

    I liked this article. http://liberalconspiracy.org/2010/07/13/why-av-reform-would-still-be-a-big-improvement-on-westminster/

  8. Mark Pack

    RT @libcon: Why AV reform would still be a big improvement on Westminster http://bit.ly/8XsJYi < Guest post from myself

  9. Rachel Smith

    RT @markpack: RT @libcon: Why AV reform would still be a big improvement on Westminster http://bit.ly/8XsJYi < Guest post from myself

  10. james blessing

    RT @libcon: Why AV reform would still be a big improvement on Westminster http://bit.ly/8XsJYi

  11. Kasch Wilder

    RT @markpack: RT @libcon: Why AV reform would still be a big improvement on Westminster http://bit.ly/8XsJYi < Guest post from myself

  12. Joe Jordan

    RT @markpack: RT @libcon: Why AV reform would still be a big improvement on Westminster http://bit.ly/8XsJYi < Guest post from myself

  13. Tweetminster

    "Why AV reform would still be a big improvement on Westminster" http://bit.ly/autoZT

  14. Gulam Hussain

    RT @tweetminster: "Why AV reform would still be a big improvement on Westminster" http://bit.ly/autoZT

  15. Oliver Cooper

    @markpack on @libcon says that AV is better than status quo. It is, but real reform would be even better. http://bit.ly/autoZT @democracyuk

  16. Rachel Gibbons

    Interesting comment thread too. RT @tweetminster: "Why AV reform would still be a big improvement on Westminster" http://bit.ly/autoZT

  17. Loren Nardini

    RT @tweetminster: "Why AV reform would still be a big improvement on Westminster" http://bit.ly/autoZT

  18. DAVID ROBERT WARREN

    RT @libcon Why AV reform would still be a big improvement on Westminster http://bit.ly/9Fexqu

  19. Kilsally

    RT @tweetminster: "Why AV reform would still be a big improvement on Westminster" http://bit.ly/autoZT

  20. ANN COOPER

    RT @tweetminster: "Why AV reform would still be a big improvement on Westminster" http://bit.ly/autoZT

  21. Robert Sprigge

    RT @tweetminster: "Why AV reform would still be a big improvement on Westminster" http://bit.ly/autoZT >> Alternative Vote changes things.

  22. Jack Allnutt

    RT @tweetminster: "Why AV reform would still be a big improvement on Westminster" http://bit.ly/autoZT

  23. Nick Murdoch

    URL sums it up. Good article: http://liberalconspiracy.org/2010/07/13/why-av-reform-would-still-be-a-big-improvement-on-westminster/

  24. LibLink: Why AV reform would still be a big improvement on Westminster

    […] can read Mark’s post in full here. Share this story with your […]

  25. Why AV reform would still be a big improvement on Westminster | Mark Pack

    […] Media & PR Why AV reform would still be a big improvement on Westminster Date: 15 July 2010 0 comments Tags: alternative vote, electoral reform Earlier this week I did the following post over on Liberal Conspiracy: […]

  26. Vote No To AV

    A quite weak and bland article from @LibCon: "Why AV reform would still be a big improvement on Westminster" http://bit.ly/autoZT

  27. Dinti Batstone

    Why AV reform would still be a big improvement on Westminster | Liberal Conspiracy http://t.co/9t4oPsv via @libcon

  28. The two electoral tests the Coalition should run | Mark Pack

    […] proportional representation for an elected Upper House. Significant though the impact of both the alternative vote and Upper House elections may be, there are two much smaller ideas the government should look to […]





Sorry, the comment form is closed at this time.