Saturday, August 07, 2010

Alterative Vote: What does 50% mean?

I've recently written a few pieces answering various questions about the operation of the Alternative Vote, including ConservativeHome CentreRight: The Alternative Vote in Australia – what does it produce?. There's another question that's come up a bit, namely:

"Who are the '50%' needed to win election?"

This is actually more complex than it sounds because an awful lot of the comment on this is confusing several distinct definitions. For example Nick Clegg, in evidence to the Commons Select Committee, stated:

" also means that people elected to Westminster know that, through the redistribution of the votes, they have a mandate of 50% or more of people in their community..."
(Quoted in The Constitution Society: Briefing Paper: Alternative Voting, page 3)

Clegg's statement is as confused as much of it. By "community" does he mean:
  1. Every single person living in a constituency, including children, eligible but unregistered adults and ineligible adults (mainly non-citizens)?
  2. Every single adult living in a constituency...?
  3. The full registered electorate, including many "ghost voters" who are still on the register despite having moved?
  4. The "actual" registered electorate, excluding the "ghost voters"?
  5. Voters who turn out and cast ballot papers?
  6. Voters who turn out and cast valid votes?
  7. Voters whose ballot paper is still the count at the crucial stage electing an MP?
  8. Voters whose ballot paper is still in the count at the "Two Party/Candidate Preferred" stage?
(The "Two Party/Candidate Preferred" stage is little known in the UK and in the numerous AV elections I've counted over the years I've never seen it done when a winner is already clear. It involves taking the count all the way to either the last two surviving candidates or the candidates from the two main parties to see the full distribution of support, even when the election has been decided at an early stage.)

Sure some of these are relatively close definitions but there's a big gap between others. Non-voters are a part of the community, but they're not part of the election process and I've not noticed anything in the proposed referendum that's going to change that. Similarly "ghost voters" are effectively out of the process (though get counted in the official turnout statistics). And there's no clear evidence as to whether AV will actually lead to an increase in turnout or what the impact will be on the level of spoilt ballot papers. There will be all manner of claims made in all directions on those points, but the one thing that can be said for sure is that compulsory voting is not on the table.

The crucial "50%" actually refers to a stage in the count when one candidate gets more votes than all other candidates still in the count. If votes don't transfer then a candidate can get elected with less than 50% of all valid votes cast. This isn't just hypothetical. The Australian state of Queensland uses the "optional preferencing" system that the UK will be offered, and although turnout is compulsory you can still get members of the Legislative Assembly elected with less than 50% of the vote. In the last state election 16 of the 89 seats were won with this - and that's when the votes are transferred all the way to the Two Party/Candidate Preferred. (Psephos: Queensland election of 21 March 2009) There would probably have been many more cases if the Liberal and National parties had not merged and instead fielded separate candidates in many seats.

Closer to home over the years I've been involved with many many AV elections for a large variety of organisations. When there were quite a few candidates (usually four or more) it was quite common for large numbers of voters to fail to use their transfers, with the frequent result that the winner again had significantly less than 50%.

To explain how this works, a constituency might have the following first preferences:

Conservative 40
Labour 35
Liberal Democrat 15
UK Independence Party 6
Green 4

Nobody wins on the first round. Transfers might go as follows:

Exclude Greens

Conservative 40
Labour 36
Liberal Democrat 17
UK Independence Party 6
Non-Transfers 1

Exclude UK Independence Party

Conservative 43
Labour 37
Liberal Democrat 18
Non-Transfers 2

Exclude Liberal Democrat

Conservative 49 ELECTED
Labour 45
Non-Transfers 6

You'll see that the Conservative has won with less than 50% of the total votes cast but it's still more than 50% of the crucial stage. And if turnout is only 60% of registered voters then the winner has the support of only 29.4% of voters on the crucial round, 24% of first preferences and an even lower proportion of the entire population.

These figures are rather different from the image being thrown about by some AV supporters that the system is going to result in MPs suddenly having huge mandates and the support of the "majority of the community". The choice at the forthcoming referendum is not so clear cut as it may first seem.

1 comment:

RodCrosby said...

It's a false comparison Tim. AV always elects a majority winner from the available information. In your example, if the elimination was done by separate runoffs, would the ultimate winner not receive >50% of the "vote"? People who abstain by truncating their preferences abstain just as surely as if they had never voted, and we never count abstentions in calculating the winning percentage under FPTP, nor should we under AV.

Also, under AV the winner cannot receive fewer actual votes in the final round than he (or the loser) did in the first, which is at least possible with separate runoffs.

AV is more strategy-proof than runoffs, since in the latter a voter can change their vote between rounds.

And AV passes more desirable qualities than runoffs. It passes Mutual Majority and Independence of Clones, which runoffs don't. Supplementary Vote is even worse than Runoff, since it also fails Condorcet Loser and Majority Loser.


Related Posts Plugin for WordPress, Blogger...