Alternative Vote

I have been trying to get up to speed with what ‘AV’ is in order to consider how I would vote in the May 2011 referendum regarding whether we should change the way in which Members of Parliament are elected in the UK.

The current system, described as ‘first past the post’, is simple. The country is split into constituencies each of which has the right to return one Member to Parliament. All valid votes in the constituency are counted and the candidate with the most votes is returned.

A perceived problem with this system is that where there are many candidates contesting a seat, it is possible for a candidate to be returned who enjoys the support of significantly fewer than half of those voting in the constituency. Given that only a certain percentage of eligible voters even turn out to vote in a given election, the candidate’s support among eligible voters will be even less. So the constituency’s ‘representative’ in Parliament may only be supported by, say, under 30% of eligible voters and actively opposed by more than half.

As I understand it from a few google searches, the point of AV appears to be that in order to win, a candidate ought to be supported, to some extent at least, by an overall majority of those voting in the election. To seek to achieve this, the AV vote comprises a synthetic series of run-off elections.

A run-off election is where if no candidate has an overall majority, the candidate with the fewest votes is ‘knocked out’ and another round of voting takes place. This process is repeated until someone does get an overall majority.

The idea is that the candidate eventually returned must at least have appealed to the voters enough such that more than half of them have been prepared to express a preference for him or her as the ‘best of a bad lot’ in circumstances where the candidate that they really wanted has been cut for lack of support.

With AV, rather than actually having multiple rounds, the run-offs are compressed into one ballot paper, on which voters rank candidates in order of preference. The list of preferences enables it to be ascertained on a count (which is therefore more complicated than a ‘first past the post count’) who voters would have voted for next, once a more preferred candidate was ‘knocked out’.

What I am not yet clear about is what happens in AV if no-one gets to 50% of voters after all other preferences have been taken into account. This scenario could arise if everyone voted for a preferred minority candidate, followed by one of the main parties of Liberal Democrats, Conservatives or Labour as a second preference, but could not stomach putting either of the other two main parties as a preference on any basis.

So in the penultimate round, all of the small parties have dropped out and you might end up with 35% Labour, 33% Conservatives and 32% LibDems. The LibDem is accordingly eliminated as the weakest candidate, but there are no further preferences to assign to Labour or Conservatives.

I would have thought that there are two options:

1. Declare the vote void and start again.

2. Calculate the 50% out of the number of positive votes in the final round, rather than out of number of voters who filled out a ballot paper in the first place. Accordingly, on the above example, Labour win, with their 35% of the total original voters amounting to 51.5% of the positive voters counted in the final round.

Option 1. would seem to be highly inconvenient. However, assuming that 2. were to be what happens, that would appear to remove the legitimacy intended to be conferred by being able to say that the returned candidate was supported by at least 50% of those who troubled to vote in the election. We would be back to the same problem said to exist with the ‘first past the post’ system.

This entry was posted in UK politics. Bookmark the permalink.

3 Responses to Alternative Vote

  1. DBirkin says:

    The 50% rule only means 50% of the round in question. Just for arguments sake.

    If 1000 people voted but in the final round only 100 preferenced either of the two candidates, the winner would only need 51 people. This they would class as 51% of voters (in last round).

    AV doesn’t guarantee that the candidate win any more of the total vote than now.

  2. Lee Griffin says:

    Spot on, this is one kind of issue with how AV is presented…but let me try and focus you on a different persepctive.

    Take a (fake) FPTP result of 55% Labour, 45% Tory. you could say that is a >50% endorsement. Except the way votes are counted is to exclude spoiled ballots. In fact if we include spoiled ballots (those voting in such a way that they are not stating a preference for one candidate) that 55% could be lower than 50%. Imagine, for a second, that there are a whole range of voters out there spoiling their ballots. Hell, with only two candidates standing (Yep, I know, would never happen) the turnout would probably be lower than average too.

    Now look at what your situation of no further preferences are. Lib Dems eliminated, but none of them wish to carry their votes. In essence those Lib Dems are saying, for the final round “No candidate gets my vote, but I’m taking part in the process”. This is spoiling your ballot by another name. By removing the Lib Dems votes as spoiled you would end up with a >50% winner.

    Of course the main thing is that, of course, this could mean no better level of endorsement by voters than FPTP. But this should be obvious as a possibility, there is no guarantee that all those people that voted for someone different would ever put their vote for someone else. It’s highly likely they would, but possible that they wouldn’t.

    What it does do is give the opportunity for more accurately determining who is really endorsed >50%, and that’s a good thing…when the worst that AV can do is provide the same level of endorsement as FPTP if no-one in the order of things preferences, but provides a more accurate view of endorsement the moment at least 1 voter does…then that’s a good change in my eyes.

  3. tucola says:

    Absolutely: agreed it’s no worse than FPTP in this respect. I was just pondering on the fact that it seems potentially to be subject to one of the same problems, theoretically at least.