AV: the proposed Alternative Voting system for England

I am in the process of making up my mind  about how I shall vote on the AV system in the UK on May 5th and. I shall set out my current position here. This is not going to be one of my more entertaining blogs, I have to say, and it might remind some of you of an elementary maths class at some points.

Let the number of candidates be C and the number of voters be V. Each voter is allowed to place the candidates in rank order of preference on the ballot paper. Voters do have to vote for at least one candidate in order for their vote to be valid, but they do not have to vote (or rank) all the candidates.  The First Round of Counting (1stRC) is based upon the candidates who have been ticked as Rank #1. To win, a candidate must have more than 50% of the votes, V. Where this does not happen, the losing candidate with fewest votes at the first round (1stRC) is eliminated. I shall now explain what happens next.

The procedure moves to a Second Round of Counting (2ndRC) based on the remaining (C-1) candidates. These candidates keep the votes they earned in the first round (1stRC). To these votes are added the Rank #2 votes from the voters whose votes were eliminated at the first (1stRC) round because they had voted for the loser, providing of course that they had exercised their right to make a Rank #2 vote. Those that did not may be referred to as the Abstainers (A). These new totals for each of the remaining (C-1) candidates are inspected and if any candidate at this point has more than 50% of the total votes cast, (V-A), that candidate is declared the winner. I am assuming I am correct in stating that the calculation is based on the total (V-A) and not the original first round (1stRC) total of (V), since we must be talking about votes actually cast.

This process of reallocation continues through subsequent rounds until a winner can be declared. Matters must grind to a halt when there are only two candidates remaining; the winner is, obviously, the candidate with the most votes. I do not know what would happen in the unlikely situation where there was a tie at this juncture, but I can’t be worrying about that sort of thing right now.

It is my view that the maximum impact of a redistribution of votes will be where the range of scores in the first round (1stRC) is (approximately) as narrow as the number of candidates C.  If one thinks about this, another way to put it is to say that the number of votes scored by each candidate approximates to V /C, give or take a few. I mention in passing that the larger the turnout the more insignificant will the differences  be between roughly equal voting shares, if taken as a percentage of the vote.

The case of unequally distributed votes at the point of reallocation  for Rank #2 and beyond

I move on now to consider the case where a second round (2ndRC) has been triggered from an approximately even distribution of votes in the first round (1stRC). It occurs to me that a bias could occur if there were some extraneous factor that resulted in the losing candidate’s votes being redistributed by Rank #2 in a grossly uneven way across the remaining candidates. In this case, the candidate favoured would win, when on the basis of Rank #1 there would not have been much to choose between any of them. It might help, at this stage, if I generate an imaginary example to illustrate what sort of situation might cause this to happen. In this example I shall consider what party organisers might advise their own supporters to do in terms of their Rank #2 preference and also what they might aim to do in terms of persuading the supporters of other candidates in terms of naming them as their Rank #2 candidate.

Let us imagine that influence is brought to bear upon the voters who support a particular candidate as Rank #1, in terms of their choice of candidate at Rank #2 level. This might be engineered through the party machinery or the media, possibly with calculated and deliberate intent.  Let us imagine a three horse race in a given constituency between Labour, Tory, and the Greens. The voters seem to be equally enchanted (or disenchanted) by all three candidates, as far as can be discerned from the opinion polls. I shall now go on to explore this scenario in a little more detail.

At the first round (1stRC), on the basis of Rank #1, each candidate receives approximately one third of the vote, but Labour narrowly loses. So under AV the Rank #2 of all the Labour voters is re-distributed. And now imagine that the local Labour party and workers agreed to advise their supporters to vote for the Green candidate in Rank #2. It doesn’t matter what the Tories or the Greens said to their supporters in this example, since their totals will be determined by Rank #1.

At this point the calculations are made, and because of the strong instruction by Labour to their supporters about Rank #2, the bulk of the losers’ vote goes to Greens, and the Green candidate gets in by what looks like a rather good majority. Is this fair? Clearly it is a matter of opinion. In terms of the empirical facts, one can see that a well disciplined and tightly-led party could maximise its chances of getting their favoured next-best candidate through in the second round (2ndRC), in situations where they knew they were running approximately evenly in the opinion polls. This would be a case of minimising the collateral damage of defeat. They might also target one of the opposition parties to persuade them to place them in second position in Rank #2. This would amount to horse-trading prior to voting day. If Labour supporters agree to place the Greens as Rank #2 and, reciprocally, the Greens agree to place Labour as Rank #2, they could together maximise the probability that the Tory candidate (in this example) would lose at the second round (2ndRC) if they had not already been eliminated at the first round (1stRC).

I have so far focussed upon a pre-ballot coalition strategy for maximising the chances that a mutually undesired contender might lose. It is more difficult to articulate a recipe for a candidates own success at the second round (2ndRC).  It would be necessary to persuade the supporters of the opposing parties to name one as the Rank #2 candidate, as opposed to the other candidate(s).  My guess is that this would involve roughly the same actions and activities that the candidate would embark upon in order to win their vote as the Rank #1 candidate. I suppose one strategy could be encapsulated as urge the supporters of the other candidates to vote for one as Rank #1 beneath their preferred Rank #1, in order to prevent the other opposition candidates from getting in on the grounds that they might be undesirable in all kinds of ways.

My own view

AV is not proportional representation and it will not encourage a larger turnout than first-past-the-post. If anything, its complexity may dissuade voters from voting. It also could be argued that voters should be given a fresh ballot paper at every stage so that they could make a new and informed decision based upon the latest group of candidates as a whole (each time one will be eliminated and so a different voting problem is effectively generated for the voter ). It is not proposed to do this and it is unlikely that it ever will be, on the grounds of cost alone.

The effect of AV will be maximal where the vote is split evenly across candidates. The redistribution of losers’ Rank #2 votes cannot disguise the fact that the election must be a close-run thing in order for this to happen. It is possible that the AV system will encourage a lot of propaganda, wheeling and dealing between parties prior to election day, above and beyond what already exists in the first-past-the-post system.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

This site uses Akismet to reduce spam. Learn how your comment data is processed.


%d bloggers like this: