Tuesday, 26 April 2011

Is AV better than FPTP ...revisited

I was recently pointed towards Gowers' Weblog by @lipscombe1 (James Lipscombe) ..to be honest it was something I have come across in my AV saga before, read a few lines of, chuckled to myself about, and then left.

It does seem however, like the 'Voting Power index' and the 'the Jungle Animals had an election video' that the Yes campaign..or at least it's campaigners are taking it quite seriously.

So let's get in to it.

I will not be copying over all the text and then reviewing it as I believe his post is far too long and would only make it worse. I may however copy over bullet points which in combination with the link to the original post here you can see for yourself to check I haven't used anything thing out of context. It isn't my aim to do so and as always my comments box will be unmoderated (except for when I am attacked with foul language rather than my posts of course..and perhaps even then I might let it slide if it makes me laugh).

Ok, the post opens with a brief background of the referendum, past referendum's etc and informs the reader that we will be treated to a more mathematical slant of Blog, which I am always in favour of (maths is a lot easier to debate).

A Little History is where we hit our first snag.
It becomes very obvious that this is going to be a FPTP is not PR section..which is a valid point..however in the context of FPTP vs AV it isn't.

The second point is comparing NATIONAL % of votes when we do not have a national based system but a constituency based system, this is pretty misleading for two reasons. You can have two areas where a seat is hard fought and won under fptp, compared to a safe seat. This will throw the national results..secondly and most importantly, the constituencies are not all the same size.
You can have two constituencies where 90% of the population vote for someone, but because the constituencies are different sizes it would appear that a seat for A needs X votes, where a seat for B only needs Y votes.

This is a valid criticism none the less but that has nothing to do with FPTP and is just as likely to happen under AV. What is good though is that the other part of the act that brought us the referendum will also aim to fix this problem.

The usual suspects are called upon, Thatcher and Blair. I personally think this is a hook to get anti thatcherites and anti blairites into the debate in a kind of "FPTP was the reason we had X" fashion.

What he has conspicuously left out of course is the poll results from the time that could be useful when comparing both systems historically.

To save time, I'll concentrate on the last election mentioned in this section, 1997.

Labour, 43.2% of the vote, 418 seats 64% of the seats

Conservatives, 30.7% of the vote, 165 seats 25% of the seats

Liberal Democrats, 16.8% of the vote, 46 seats 7% of the seats

Now, that looks quite damning doesn't it and I can see why a Yes campaigner might like to use that.

However, let's look at the same thing again, but with AV seat % according to research.

Labour, 43.2% of the vote, 418 seats 64% of the seats Under AV 471 seats 68% of Seats

Conservatives, 30.7% of the vote, 165 seats 25% of the seats under AV 70 seats 10% of seats

Liberal Democrats, 16.8% of the vote, 46 seats 7% of the seats Under AV 115 seats 18% of seats

Source being BBC findings which Roy Jenkins also backs up

So instead of giving Labour 21% too many seats, it gives them 25% too many.
Instead of giving the Conservatives 6% less of the seats, its gives them 21% less.
Instead of Giving the Lib Dems 10% less seats it gives them 1% more.

If you add all the results together, this election is 10% less proportional under AV than under FPTP.

I feel I should make it clear here that I am not saying one system will be more or less proportional, they are not proportional systems and some of the elections he mentioned will be more proportional under AV, some less.

The point I am trying to make is that this argument belongs to those promoting STV, (even though it would still produce some disproportionate results) or national lists.

This point is also missed by him in his section
Does FPTP reflect the will of the voters?> in which he ironically says about the election we have just looked at , the 1997 election of Tony Blair "However, it is not reasonable to say that the size of Blair’s majority — he had almost two thirds of all seats — reflected the national mood."

He then makes a similarly flawed comment on the party with the most seats not getting the most votes, this statement suffers from the same problems we have discussed.

To be honest with you readers, it was at this point I chuckled and clicked the X in the corner of my page. I have seen many of these arguments in the last 8 months and I have replied to so many. I have not read the rest of the article, so I will read through it now and post some more :)
OK, he has admitted that AV is not proportional so the top bit of his blog is irrelevant..interesting tactic "..there is no obvious reason to expect AV to be proportional."

Ah, the wasted vote/tactical vote argument.
Here, is another Yes campaign line. That if you don't vote for the winner, your vote is wasted.
However I feel that a vote is NEVER wasted. If I want to vote for a party, that is my democratic right, if i lose, it's not wasted, calling it wasted is very cynical.

The phrase "It's not the winning, it's the taking part that counts" should be the motto of any democracy lover.

He then talks about monotonicity with AV, i.e that sometimes it is possible for you to benefit your candidate of choice by voting against them.
He judges that this is unlikely to catch on..well I can agree with that, but it's likely to go on as you will see here

Very importantly we have to think about what people gain from voting tactically under both FPTP and AV and the cost of doing it.

Under FPTP, people tend to vote tactically when they are beyond reasonable doubt their candidate of choice will lose. So they vote for someone else, the cost is your favoured candidate is even less likely to win. You have made a compromise.
Under AV, when you vote tactically you risk causing your candidate to lose but you only do it if you are beyond reasonable doubt your candidate of choice will lose, if it pays off, you candidate of choice will win. The cost is much more worth it. No compromise has to be made.

This is why, although less likely, I believe tactical voting to be worse under AV.

Of course, depending on what definition of 'tactical voting' we use, if it was for example "having your vote go to a party that isn't the best standing, in your opinion, in order to stop another that you dislike from winning" we could describe AV as a whole as institutionalised tactical voting.

The FPTP name is brought up. I can answer this, the winning post is when your votes are X+≥1 where X is the number of votes of the second place.
It is not a fixed post in the terms of votes needed, but nor is 50% of final round votes+1 so there is no real change or advantage with AV..

Boy, this blog is long...

I will skim through some bits "It was the FPTP system that allowed Mrs Thatcher to rule for eleven years with large majorities and well under half the votes" ...as above, in some circumstances it would have got 'worse' under AV with Thatcher so another slightly misleading statement there.

"Under AV there is a real chance that Labour would have won the 1983 election", not according to the research..it suggests Labour would have lost 20 more seats...

"it is unlikely that John Major would have had his shock victory in 1992, for instance" again not according to the research

"Probably in 2005 it would have left Blair a little more punished for the Iraq war" not according to the research

We then get on to the counters to No2av arguments..
1. the horse race, i know where he is coming from on this one, though i will say there are races like what he has described as silly, where you see how far you can run in a set time (or set amount of votes) and the racer that has run the more laps or distance is the winner.

2. Av being unfair.
This is a tricky one, i am not exactly sure what no2av mean by unfair..what I mean when I say unfair is this...not all ballots are treated the same, you can have some ballots where every preference is used towards the result and you can have people that only have their first preference used. Kind of like this video http://www.youtube.com/watch?v=BFbYwHAqrC4

3.this leads on from the 2nd nicely. Think of a system, where instead of preferences, you have X ballots, but multiple rounds...every time you vote, the least popular candidate is eliminated and everyone that voted for that candidate gets another vote to use on adding to the remaining candidates vote total
This system quite clearly gives some people more votes than others.

It is what I have referred to as DB voting, named after myself ;) ...It will produce the same results AV every time, because it is the same thing but worded differently. We have got into the realm of semantics here. Clearly though, having your second preference ignored if you first doesn't win while other people have their 2nd, 3rd 4th and 5th preference influence the result is not equal or fairer.

4.This works if you don't believe people's first choice should, if we are being objective, be more valuable than a 4th preference. I do believe that, so I see an unfair advantage.

5. I agree in part with this one, Av doesn't help the BNP but where I disagree is that it will not help the extreme issues become mainstream.

Yes, I am sure many extremist people vote for mainstream parties, but there WILL be more of these votes up for grabs (SEE George Foreman Grill analogy)

6. The cost i am not going to comment on, I do not think no2av are lying, i do think they are wrong...and especially now all the fuss has been made, counting machines will not be needed...there will be a cost in education but this point only really counts if we do not consider AV better than FPTP, so is more of a reason to back up someone already voting no then to convince people who think AV is the best system.

7. Av is not simple..ranking is.

"Labour, A
Conservative, B
Lib Dems, C

where and (let us suppose)

Those lucky BNP supporters now get their second-choice votes counted. Let’s suppose that and that and are the percentages of votes transferred to Labour, the Conservatives and the Lib Dems, respectively. So after the first elimination the position is

Labour, A+E
Conservatives, B+F
Lib Dems, C+G

Suppose now that the Lib Dems are third. Now we have and and the standings are

Labour A+E+H+K
Conservatives B+F+J+L

Let us think carefully what and stand for. is the number (expressed as a percentage of the total number of voters) of Lib Dem voters who transfer to Labour, which means Lib Dems who either put Labour as their second choice or put the BNP as their second choice and Labour as their third choice. is the same thing but with “the Conservatives” replacing “Labour”. is the number of BNP supporters who put the Lib Dems second and Labour third, and is the number of BNP supporters who put the Lib Dems second and the Conservatives third."

Put your hands up if that is simple?

We may think it is simple, I think playing piano or guitar is simple...but not everyone will..i would go so far to say, most people won't like it.. hence the increase in spoiled ballots even in OPV in Australia.

8. It is fair to say that some politicians will become more extreme, others will become more safe..depending on what is needed.
He also makes a point that under FPTP you can elect someone that most people dislike, I agree..this can of course happen under Av, as long as they aren't the MOST disliked. I doubt very often the condorcet loser does win under FPTP though, not with the likes of the BNP and other such "healthy grill waste" parties.

9.This is an odd one, hung parliaments are believed to be more likely with AV, most independent sources will say this, from 4 Fact Check, BBC, FullFacts.org etc.. they will become more likely anyway but will get more so with AV. That is not the odd bit though, the author argues my point..that in 1997 AV could have led to a bigger swing to Labour..even though he criticised FPTP for the swing being too much already.

10. Is true, but will change over time, like who benefits under AV in Oz has, but at the moment the lib dems would benefit from AV.

11. I agree with the author here, you shouldn't vote not to hurt Clegg..you should, in my opinion, vote no because it is fairer.

12. Not too swayed either way here, only way to dislodge a government is to take away it's MPs, only way for a MP to lose in FPTP is for someone to get more votes than them, under AV someone can get more votes than them, but because the new party is a polar opposite to the other non-elected party, they second preference the incumbent.

An example would be say a Lib Dem MP.. he is unpopular, so labour over takes him in the first round, but the tory voters, while unhappy with the lib dem are unlikely to preference Labour highly...in fact they may see another Labour Mp as a threat to a possible Tory government, so they tactically preference Lib Dems, meaning that the Lib Dem is safe even though unpopular because at least they're not tory/labour..this would be unlikely under FPTP.

13.Ahh, this is pretty much the Voter Power Index argument. The more marginal your constituency the more value your vote has as it has an increased chance of changing the result.

This is only true of course if you view your vote in isolation of everyone else's.

To say for example that in a game result that ended A 20 vs B 21 ..the last score for B was worth the most, is completely in accurate.
If you took away any of B's goals, that last goal would crash in its perceived value.

It should be plain to see that every goal is important and every goal counted as 1 goal.

The same can be said of votes.

Every vote is worth one vote for one person under FPTP. Everyone can change the result by '1'

A nice way to demonstrate this is to go to the GE2010 results page and look at you constituency and look at your chosen party's result. Everyone changed that number by '1'. That is all ..and that is every voter.

14. I also don't go for the "it's been done this way for ages so let's keep it" but i also don't go for the "it's been done this way for ages so let's get rid of it" argument.
I want to keep FPTP because it is, in my opinion better than AV.

I liked this post, I feel no need to change my mind, but at least they're arguments.

I hold many of the authors views on what is wrong, but most of the time completely disagree with the cause or remedy.

I would hate to think that we sleep walked into a Yes vote because "anything must be better".
Yes, sometimes FPTP is crap, I have many blog posts where I make this clear, but AV doesn't fix these problems and can make some worse.

I'd encourage everyone to vote NO in May, AV is not a step in the right direction, it is a step to split the reformer vote.


  1. I thank you sir!

    My post (in the other place) on the subject is not nearly as persuasive or entertaining.


  2. This comment has been removed by the author.

  3. This comment has been removed by the author.

  4. Just on the 'wasted' vote issue.

    I think you may have misunderstood what the Yes campaign means here (that sounds condescending - but I just think it's what's happened).

    Your argument is that a vote that isn't for the winner isn't a wasted vote - it just wasn't for the winner. That is fair enough - until we consider constituencies.

    In Scotland in the 2010 GE, the Conservatives won 17% of the vote and ended up with just one seat. So for a Tory voter in Scotland - a vote for the Tories really was a wasted vote - it didn't end up helping the Tories one bit.

  5. 258 seats in the 2010 election didn't get Labour in to government ..are all the votes for them wasted?