Wednesday, 5 January 2011

AV, Spread of Opinion, and how it ISN'T a cause of coalitions

I've posted about this before, that the voting system under a single member constituency system isn't the cause of coalitions, nor is it the defense against them...that in fact it's spread of opinion nationally, and how groups of people congregate in their constituencies under a political landscape that has more than 2 parties achieving sizeable support.

In the comments I gave a worked example, so I thought I'd give it it's own post for clarity. Take a rather basic example of 5 equal sized constituencies, they have the following (very) rough make up....

1: Tory 60%, Labour 30%, Lib Dem 10%
2: Tory 40%, Labour 30%, Lib Dem 20%
3: Tory 30%, Labour 45%, Lib Dem 25%
4: Tory 25%, Labour 30%, Lib Dem 45%
5: Tory 20%, Labour 45%, Lib Dem 35%

The Tories achieve a total vote share of 35%, Labour 36% and Lib Dems 27%

Under FPTP this returns Tory - 2, Labour - 2 and Lib Dem 1. Hung.

Under AV, we (for the sake of this example) assume the second preference always goes to Lib Dem from Tory or Labour, but from Lib Dem it always goes to Labour.

The new result under AV would be Tory - 1, Labour - 2, Lib Dem - 2. Hung as well, though potentially with a much strong "massive" coalition of Labour and Lib Dems.

Now see what happens if the constituencies change to the following. Note that I'm not changing the proportional share of voted "nationally" (or across all constituencies), only how they're share is balanced in each area...

1: Tory 44%, Labour 41%, Lib Dem 10%
2: Tory 56%, Labour 39%, Lib Dem 10%
3: Tory 25%, Labour 45%, Lib Dem 30%
4: Tory 15%, Labour 30%, Lib Dem 55%
5: Tory 35%, Labour 25%, Lib Dem 40%

FPTP would now return Tory - 2, Labour 1, Lib Dem - 2

But under AV we'd now have Tory - 1, Labour - 1, Lib Dem - 3.

Notice how in the first example the hung parliament is the same proportion as in the second example, just with different people in power. But in the second example AV actually moves us AWAY from a hung parliament.

So there you have it. AV can cause coalitions where FPTP would normally have delivered a majority, but they can also cause a majority where FPTP would have delivered a coalition.

With boundary changes coming that we are still to find out the effects from, and several years of political wrangling in an obviously tense climate, it'd be insane to try and predict right now whether EITHER voting system would return a coalition or not at the next election.

In short, it's not practical (or honest) to claim that AV is a coalition making system.


  1. Good article. However don't expect the anti-AV campaign to stop making this claim, they have no genuine arguments, so with out these false ones they would have to remain silent!

  2. It's also irrelevant! I don't like coalitions but if that's more representative of the wishes of the people (which I think is/should be the real argument) then that's what I have to live with :)

  3. Oh absolutely, if people vote for it, they should get it. I don't think the public is really anti-coalitions, just perhaps anti-THIS coalition. But the fact remains that simply by implementing AV, you don't guarantee an increase in coalitions compared to FPTP. History shows the opposite, though that history is irrelevant now too.


Got something to say about my post? I'd love to hear it!

Try to keep it civil, I don't delete comments unless obliged to or feel the thread is getting too out of hand, so don't make me do it.