AV - how to vote?

[ArchimedesUK]

Hello all. The local group for the Yes campaign (Sheffield says Yes! to Fairer Votes) has a website at fairervotessheffield.org.uk. There's a few videos on there that explain the main faults of using FPTP to select a winner from more than 2 candidates and how AV eliminates some of the main faults.

 

I'm supporting Yes because AV will eliminate tactical voting which forces people to choose between supporting the candidates they like and the candidate most likely to beat the person you dislike the most. It also masks the true support of candidates or parties an area, e.g. it makes it look like there's fewer Labour supporters in Sheffield Hallam than in reality because they're voting LibDem tactically to keep out the Conservatives. This of course makes Lab voters more likely they'll vote tactically again next time around.

 

AV lets you vote honestly for the candidate(s) you like without giving any advantage to those you dislike. Tactical voting under AV (e.g. putting your 2nd preference 1st, and 1st as 2nd, etc. in some effort help your 1st preference) will only hurt your 1st preference because giving them less 1st preferences increases the likelihood of them being eliminated sooner. When you factor in the possibility of unknown numbers of other voters maybe doing the same thing or trying to counter it, being honest about your preferred candidates really is the best policy.

[HeadingNorth]

Tactical voting under AV (e.g. putting your 2nd preference 1st, and 1st as 2nd, etc. in some effort help your 1st preference) will only hurt your 1st preference because giving them less 1st preferences increases the likelihood of them being eliminated sooner.

 

This is not always true. There are scenarios where a candidate can be eliminated because too many people voted for him as 1st choice instead of 2nd choice, where if they had put him as 2nd choice he would have won the seat.

[evildrneil]

If people vote for more than one, doesn't that force the parties even closer together politically, to try and pick up those 2nd and 3rd preferences?

 

Is that necessarily a bad thing?

[andikay]

This is not always true. There are scenarios where a candidate can be eliminated because too many people voted for him as 1st choice instead of 2nd choice, where if they had put him as 2nd choice he would have won the seat.

 

How would that work?

[HeadingNorth]

How would that work?

 

I'd have to dig out the book and quote their examples. I don't remember the numbers offhand.

 

AV is subject to quite a few potential paradoxes. This is probably the worst - imagine the headlines if someone loses the election because 2,000 too many people voted for him.

[andikay]

If people vote for more than one, doesn't that force the parties even closer together politically, to try and pick up those 2nd and 3rd preferences?

 

I’d say it’s a good thing; the main parties would have to try and address some of the problems that give people a reason to vote for parties like the BNP.

[ArchimedesUK]

I’d say it’s a good thing; the main parties would have to try and address some of the problems that give people a reason to vote for parties like the BNP.

 

The ippr (Institute for Public Policy Research) in their paper "The Worst of Both Worlds - Why FPTP is no longer fit for purpose" found that the last general election was decided by 460k voters (1.6% of the electorate) across 111 swing constituencies. Because of competing for such a small target audience the main parties tailor their message to these few, blur their party differences and can ignore everyone but those swing voters.

 

With AV the number of very marginal seats is doubled (source: the new economics foundation) and all seats become more competitive which gives a greater number of people more influence in deciding their MP and ultimately the outcome of elections. With AV candidates will have to reach out beyond their core voters to secure a broader level of support within the community.

 

As well as reaching out to those people you never needed to talk to before, this would pull policies into a more central position from either extreme of the political spectrum, while at the same time trying not to alienate too many of your core voters.

 

A candidate offering something that the majority of their constituents like or would accept is surely better than offering something that only a minority would love.

[ArchimedesUK]

I'd have to dig out the book and quote their examples. I don't remember the numbers offhand.

 

AV is subject to quite a few potential paradoxes. This is probably the worst - imagine the headlines if someone loses the election because 2,000 too many people voted for him.

 

FPTP suffers the same faults. It's possible for a party to get fewer seats than their vote share would suggest either because their support is too spread out across the country, thus being insufficient to win a seat in any area, or being so clumped together that you definitely win a few seats but don't have enough voters elsewhere to contest other areas.

 

 

The thing about tactical voting with AV (ably demonstrated by attempts to explain it on this thread) is that it's so complex as to be impractical and any attempt to achieve it would require such managing of voters as to be illegal.

 

A candidate asking voters to vote tactically under FPTP would just ask: "A Vote for X is a vote for Y - Vote Z instead"

 

Under AV the message might be "A vote for X is a vote for Y, could a few people vote Z instead but not too many please as that might make a vote for Z a vote for Y instead..." or something. So basically just vote honestly about your preferences.

[HeadingNorth]

FPTP suffers the same faults.

 

It does not. Under FPTP, the person with most votes in a constituency is always the winner. No other result is ever possible, and gaining extra votes can never be a bad thing.

[ArchimedesUK]

It does not. Under FPTP, the person with most votes in a constituency is always the winner. No other result is ever possible, and gaining extra votes can never be a bad thing.

 

As I said above in the rest of my statement - if your support is too concentrated nationally then you win only a few and if it is too spread out nationally you can't contest enough seats. It's possible for the winner of the election nationally to get the largest vote share but not get the largest amount of seats (e.g. 1951 General Election when the Conservatives got 250k fewer votes than Labour but won 26 more seats). Since all you need is 1 more vote than your nearest rival to win, anything significantly more than that indicates your supporters may be too clumped together and that you may have wasted campaign money that could have been used to win extra support elsewhere.

 

Under AV the winner in a constituency is always the one with the most (and at least 50%) of votes. This may or may not always be the candidate with the most 1st preference votes, depending on whether or not they have broad appeal within the area. The winner could be the person who came second in terms of 1st preferences but because they are more widely liked within the community they gain enough 2nd and 3rd preferences to pass the 50% mark.

 

(It's theoretically possible for the person who came 3rd in terms of 1st preferences to win - but only if the difference between 1st and 3rd is very slim and votes for 1st and 2nd place combined come to less than 50%. In practice it's only happened once in Australia, in 1970 - but that was under a different form of AV (the full preferential system) to the one we're voting on (optional preferential system), under which a person who came 3rd or lower has never won.)