'Is AV really so complex? Or is it just confusion marketing?' blog discussion

dtsazza

zagfles wrote: »

Under AV, second is clearly the worst place to finish. It doesn't seem right that the views of those voting for a candidate finishing third or fourth are better reflected than the views of those voting for the second place candidate.

That's one of those arguments like the "multi-vote" one, which seem like they might be true, but which I don't think stand up to scrutiny.

Every voter is still going to have a stake in the final round (except those who didn't fill out every preference and all of whose choices have already been eliminated, so they're definitely not well-represented). At this point there are two parties left, and everyone is voting for the party out of those two that they most prefer.

Clearly everyone is equally represented in this case. Just as they were the round before. And the round before that.

At heart this seems to lie upon a similar root misunderstanding to the multi-vote fallacy, that somehow you got "less say" than other people because you put down second and third preferences that turned out not to be needed. And just as with the other fallacy, it's not true that this gives you less representation. Arguably, if anything you're getting slightly more representation, because you're still in a position to further the chances of your favourite candidate. That's definitely a good thing, arguing that those who've gone through their second, third and fourth preferences are somehow better off is... odd.


And I think it's inaccurate to say that second is the worst place to finish - in one respect every place apart from first is equal in that it gets zero seats instead of one, and in another respect second is better than third or fourth in that it gives you a good platform to persuade people to vote for you next time round, as the nearest competitor.

zagfles

dtsazza wrote: »

That's one of those arguments like the "multi-vote" one, which seem like they might be true, but which I don't think stand up to scrutiny.

Every voter is still going to have a stake in the final round (except those who didn't fill out every preference and all of whose choices have already been eliminated, so they're definitely not well-represented). At this point there are two parties left, and everyone is voting for the party out of those two that they most prefer.

Clearly everyone is equally represented in this case. Just as they were the round before. And the round before that.

At heart this seems to lie upon a similar root misunderstanding to the multi-vote fallacy, that somehow you got "less say" than other people because you put down second and third preferences that turned out not to be needed. And just as with the other fallacy, it's not true that this gives you less representation. Arguably, if anything you're getting slightly more representation, because you're still in a position to further the chances of your favourite candidate. That's definitely a good thing, arguing that those who've gone through their second, third and fourth preferences are somehow better off is... odd.


And I think it's inaccurate to say that second is the worst place to finish - in one respect every place apart from first is equal in that it gets zero seats instead of one, and in another respect second is better than third or fourth in that it gives you a good platform to persuade people to vote for you next time round, as the nearest competitor.

Of course only the winner gets the seat, but the whole point of AV is the question "if you can't have your first choice, who do you prefer out of the others?". Every voter has a say in this, except the voters of the party who finishes second.

Second is obviously the worst place to finish. I've given enough examples in this thread, but here's another.

Say the votes when there are 3 left stands at

Labour 40%, LD 30%, Tory 30%.

Say the LD voters are 50/50 second choice Tory/Labour, and Tory voters nearly all prefer the LDs.

Will the Tory voters' views be better represented if the Tory finishes 2nd or 3rd?

If the Tory finishes third, Tory voters 2nd choice will count and they'll transfer mainly to the LDs giving them the seat. The Tory voters got their second choice by finishing third.

If the Tory finishes second, the LD 2nd choices will split between Tory & Labour, giving Labour the seat. The Tory voters' second preferences are ignored. The Tory voters didn't get their first or second choice. They got a worse result because they did better.

Same applies to Labour voters if the Tories were in the lead.

dtsazza

tonyhamm wrote: »

AV is an electoral system which gives voters an extra 'compromise' vote...

I'd prefer to call it "contingency" (i.e. "on the basis that this party doesn't get in, my next favourite is X"), but fair enough. And it's multiple cascading contingency votes, not just "a" vote.

which will largely benefit the LibDems.

It's difficult to substantiate claims like that, because of course we don't have any evidence of what people would have put as their second preference under AV in past elections.

What we can say, though, is that the second preference vote will largely benefit those parties that are the voters' second favourite (fairly obviously). That's pretty much the major benefit of AV, in my opinion. You can vote for your favourite party, but if it turns out they can't win, you're not disenfranchised but still have a say in who gets to represent you.

That's the problem with FPTP - the only votes that count are those for the top two parties. If you vote for any other party, you are literally wasting your vote, in the sense that it cannot influence the outcome of the ballot. As well as the tactical voting that this entails, it's not even clear who the top parties will be (c.f. Cambridge in 2010, where both Labour and the Conservatives said the other party couldn't win there) - so there's no guarantee, even with tactical voting, that you can have a say in who gets in. AV counters this by ensuring your wishes are counted at all stages.

The electoral process is that Labour voters will largely vote Labour 1st, then LibDem 2nd. Conservative Voters will largely vote Conservative 1st then LibDem 2nd. LibDem voters will vote LibDem 1st.
So it will largely benefit the LibDems and no-one else. (The FACTS ARE that the Liberals are the main party in Australia ruling for all but 14 of the past 62 years.)

Let's assume this is indeed the case. Then in constituencies where it's basically Labour vs Conservatives at the moment, this won't do anything to benefit the Lib Dems as they'll be eliminated before they get the transfer votes from either party.

In constituencies where one party has over 50% of the vote at the moment (or close to it and would be bumped over by smaller parties' transfers, e.g. UKIP and BNP are likely to transfer to Conservative before Lib Dems), then the second preferences don't matter as this party will win the seat outright.

In any constituency where Lib Dems would win under FPTP, there's no way then can get more than the seat they already have, so AV won't help them do any better here.

So in constituencies where it's closely matched, and Lib Dems come second vs. one of the other two - then yes, the Lib Dems may well benefit from this. But then so they should; if the votes are something like 37% Tory (=> Lib Dem), 34% Lib Dem, 29% Labour ( => Lib Dem) then currently under FPTP the Tories get this set. That's despite 63% of the electorate ranking Tories last (presumably), and preferring the Lib Dems as an AV vote. In fact, 100% of the electorate prefer Lib Dems as a first or second preference, so in the absence of any sort of PR, I think it's undeniable that a Lib Dem win is by far the fairest and most representative result.


Still, we're talking about specific parties again, which is getting off track from the mechanics of the systems. AV will provide a boost to parties that are currently losing in seats that have a very fractured set of first preferences, and where that party is a second choice for most everyone. I think that's the only fair outcome in these situations.

(It'll be interesting to see how the preferences of Labour and Conservative voters transfer in practice, though. I think many will not put the other party down out of an almost knee-jerk, historical response of them being "the enemy". However, if people act rationally, imagine that (e.g.) the Conservatives were not standing for this seat and think about who they'd prefer to win - I think the majority of Conservative voters would actually prefer Labour's policies and practice over the Lib Dems, if they're honest with themselves.)

Because of the new compromise votes, a candidate that tries to appeal to everyone, dispite carrying impossible to deliver opposite LibDem policies like Strong immigration AND a free-for-all immigration amnesty, will collect most 2nd preferences, therefore the nature of politics will change to become unprincipled, with more broken promises.

If a candidate promises impossible to deliver policies, these policies should be discounted in your decision-making process, and you should consider this a black mark against them for being unrealistic.

If an upstart bank offered you an account yielding 35% interest paid yearly - would you sign up instantly? Or would you read all the T&Cs in detail, ask around on forums what the catch is, get legally-binding written guarantees from the bank before tying your money up for a year? And when you couldn't find anything positive said about the account, would you go ahead and open it anyway, or would you recognise it for a scam?

Caveat emptor applies to voting as well. And as I've said above, this is nothing to do with the electoral system, because you cannot decide who to vote for (under FPTP or AV or anything else) until you've assessed all of the candidates. This assessment comes outside of the remit of the electoral system. It's as much a (potential) problem under FPTP as it is a (potential) problem under AV.

Now if your concern is that "but people aren't voting properly" - then I agree with you. People generally don't make a rational decision on the basis of what will further the progress of the country. But again, that's nothing to do with the electoral system, it's to do with the tenets of democracy meaning that idiots have as much of a say in how the country should be run, as professors of politics. (At least we have a representative democracy, though. )

Edit to note - I'm not saying that anyone who votes without dispassionately considering all candidates is an idiot. I'm just crudely delineating a spectrum from "least able to vote appropriately" to "most able to vote appropriately".

In Australia there are all sorts of tactics & gaming of the the extra votes that AV gives the main parties to crush smaller parties and reduce political competition.

Could you expand on this? Why do the smaller parties need "tricks" to be crushed if they're only getting a minority of support? Alternatively, how can they possibly be crushed if they're getting a majority of support (and thus under AV would win as soon as they had > 50% of the votes)?

dtsazza

zagfles wrote: »

Of course only the winner gets the seat, but the whole point of AV is the question "if you can't have your first choice, who do you prefer out of the others?". Every voter has a say in this, except the voters of the party who finishes second.

I disagree. You do get to express on the ballot paper who your second choice is (and your third choice etc.). The point is, it's not taken into account in the voting rounds because there were no rounds where that contigency ("if you can't have your first choice") applied.

It's like the old Yes example of "If they don't have a Mars, get me a Twix". If they do have a Mars bar, your contingency isn't going to fire (and the Twix option is irrelevant) but it would be strange to be bothered by this, because it was a contingency for an event that didn't happen.

Second is obviously the worst place to finish. I've given enough examples in this thread, but here's another.

Say the votes when there are 3 left stands at

Labour 40%, LD 30%, Tory 30%.

Say the LD voters are 50/50 second choice Tory/Labour, and Tory voters nearly all prefer the LDs.

Will the Tory voters' views be better represented if the Tory finishes 2nd or 3rd?

If the Tory finishes third, Tory voters 2nd choice will count and they'll transfer mainly to the LDs giving them the seat. The Tory voters got their second choice by finishing third.

If the Tory finishes second, the LD 2nd choices will split between Tory & Labour, giving Labour the seat. The Tory voters' second preferences are ignored. The Tory voters didn't get their first or second choice. They got a worse result because they did better.

Same applies to Labour voters if the Tories were in the lead.

That's true - it's that one situation where tactical voting does still apply under AV. In my opinion, it happens in a relatively modest set of circumstances (for a start the votes need to be relatively equally split three ways, and the distribution of second preferences needs to be notably asymmetric).

But the point of the referendum isn't "is AV perfect" (hell no, it's not) but "is AV a better system than FPTP?"

In the interests of full disclosure, it should be noted that under FPTP, Labour just wins outright. There's not even the slightest consideration as to what the second preferences of the Lib Dem and Tory voters were (which would be particularly bad if all of the Lib Dems had Conservatives as a second preference). That's why even in this example, I think AV comes out on top - it tries to allow voters the best chance of expressing their desires as to the eventual winner, and usually succeeds.

zagfles

dtsazza wrote: »

So in constituencies where it's closely matched, and Lib Dems come second vs. one of the other two - then yes, the Lib Dems may well benefit from this. But then so they should; if the votes are something like 37% Tory (=> Lib Dem), 34% Lib Dem, 29% Labour ( => Lib Dem) then currently under FPTP the Tories get this set. That's despite 63% of the electorate ranking Tories last (presumably),

Where do you get this from?? The analysis I posted earlier for second preferences in all the elections since 1980 show that LD voters preferred the Tories in all the elections the Tories won, and preferred Labour in all the elections Labour won - ie their second preferences were in line with general opinion. There is no in-built bias among LD voters to Labour - and certainly not now!

GooeyBlob

zagfles wrote: »

And with the rise of UKIP, and the LDs being deserted by their more Labour leaning voters, AV is probably worse for Labour than it is for the Tories.

I guess in the majority of cases it will depend on which way the UKIP and BNP votes split in individual constituencies.

JimmyTheWig

dtsazza wrote: »

I don't think it's automatically obvious that AV would cause more hung parliaments, and I haven't seen any evidence presented anywhere to support this.

I take your later point that we have no evidence to say who people would put their second preferences for, but I think it is quite clear that, generally speaking, Labour and Conservative voters will put the LibDems as a second preference.
Which will mean the LibDems will get more seats than they do now (some would say that this is a good thing as it is more proportional to their number of votes). Which, given that the LibDems are the third party in the UK, will mean a more equal three-way split in parliament. Which will mean a hung parliament is more likely.

Then, I don't agree that hung parliaments lead to broken promises. Also, I don't agree that a hung parliament takes power away from the voters

I think it can do, certainly.
Negotiations to form coalitions are done by the top people of each party, behind closed doors. Naturally, to enter a coalition you have to make compromises. If you voted for the candidate in your constituency who won the seat because of one or two of his party's policies, then those policies may get shelved in coalition negotiations, you really have had a wasted vote.


I don't know what the answer is, though, as I certainly agree with PR which would lead to this same situation.
Maybe parties should make these agreements before the election. E.g. announce that they will go into coalition with party x if they get so many seats or with party y if they get a different amount. If we go into coalition with party x then we will drop policies A, B and C. If we go into coalition with party y then we will drop policies D and E.
Or maybe if a coalition is formed we should have a referendum on the coalition's proposed policies?

irnbru_2

JimmyTheWig wrote: »

Or maybe if a coalition is formed we should have a referendum on the coalition's proposed policies?

To me this misses the fundemental point: you elect a representative and assign them the power to negotiate and vote on your behalf.

If you think your MP has gone totally against what they said they would do then you seek a recall election.

It's recall legislation that's needed too.

JimmyTheWig

Sounds good, irnbru.
That might solve a lot of problems in one go.

dtsazza

JimmyTheWig wrote: »

I take your later point that we have no evidence to say who people would put their second preferences for, but I think it is quite clear that, generally speaking, Labour and Conservative voters will put the LibDems as a second preference.
Which will mean the LibDems will get more seats than they do now (some would say that this is a good thing as it is more proportional to their number of votes). Which, given that the LibDems are the third party in the UK, will mean a more equal three-way split in parliament. Which will mean a hung parliament is more likely.

That is a reasonable argument, and I think you're right that the Lib Dems will be the most likely party to benefit. But there's an argument that tempers this somewhat, derived from Would AV have changed history/

Lib Dem voters have historically tended to favour the overall victor (i.e. the party that won the most seats). This means that in constituencies where the Lib Dems come third or lower, the "winning" party is going to get a boost, which will lead to it winning some seats that would be lost under FPTP. So while the Lib Dems are gaining seats, the leaders are pushed closer to the point where they have a majority. In fact, in four of the six elections featured, the winning party actually gained more seats under this simulated AV, and in five of the six they had a greater majority under AV.

At the parliament level, both systems are "first past the post" in that any party with more than 325 MPs has a working majority. Since there's no strong evidence that AV would reduce the leading party's number of seats (in fact mild evidence that it's likely to increase it a little), I consider that parties that would have reached this figure under FPTP are still likely to do so under AV.

Negotiations to form coalitions are done by the top people of each party, behind closed doors. Naturally, to enter a coalition you have to make compromises. If you voted for the candidate in your constituency who won the seat because of one or two of his party's policies, then those policies may get shelved in coalition negotiations, you really have had a wasted vote.

I don't know what the answer is, though, as I certainly agree with PR which would lead to this same situation.

Yeah, it's tricky. If a vote's split roughly 1/3 for each of the major parties, then it definitely can't be considered fair for any one of them to have a working majority, so some form of coalition would be required.

Part of the problem is that we ostensibly vote purely for our local MP to represent the constituency. So hey, what's the problem with a coalition? You got the MP you voted for, just like every other constituency did.

However in practice, most issues are decided along party lines, voting for policies put forward by the party leadership. As such I believe that the majority of people would be happy with a stuffed bear as an MP, so long as he was affiliated with the party they agreed with, and always followed their whip in parliament. (Not everyone, of course, and some MPs are great - but I think it's hard to contest that people feel that parliamentary issues are of greater relevance to their vote than local ones.)


(As an aside, I think PR could help here, because the issue is one of culture. We expect our elections to result in one party having a majority - but let's be honest, this is only equitable in a two-party system. In any system with three or more notable contenders, the odds of one party getting over 50% of support gets lower and lower.

We need to face up to the fact that the good old days of the past are gone, and some sort of shared-power governance is the way forward. Once we get over the "coalitions are weak" mentality, and a system that reinforces that, we can make the most of the situation. I believe that if we sat down and thought "let's assume that every from now onwards is going to result in a 'hung parliament' (what a loaded term). How should we adjust things to work optimally in that environment?," there are loads of frankly anachronistic processes, both explicit and implicit, that could be changed.

The current coalition isn't doing well, partly because of ideosyncracies, and partly because our political system isn't made to support it. At present the tendency is to point at the former and saying "see, coalitions don't work." Instead of being like Cindarella's ugly sisters and trying to cram our multi-party political landscape into a single-party majority result, adopting PR would be a step towards the latter approach.

And it gets rid of deliberate or historically accidental gerrymandering, which is a required step before any parliamentary result can be claimed to be "fair".)