Busting the myths of the Alternative Vote

On paper, it was a good thing: give people the chance to vote democratically about how they want to choose their representatives in the future. But I increasingly feel that the Alternative Vote campaign has shown British politics at its very worst. All I want to do is shout ‘shame on all of you’.

The campaigning leading up to the referendum on AV this week has become very dirty business indeed, dominated by vested interests. The very people who stand to benefit the most (or lose the most) are the ones leading the campaigns.  And of course there is the shamefully undemocratic way in which the referendum was negotiated – no public consultation, no review of all the options – instead just the offer of AV which anyone with the inkling of an interest in electoral reform will know is not proportional representation.

As statistician here at nef I find it so frustrating that the basic statistical implications of a switch to AV have simply not been discussed.  This has frustrated me so much that I have actually spent well over two weeks of my time this year (my own time with no funding!) designing and running simulations on the likely effects of AV on the results of general elections right back to Margaret Thatcher’s victory in 1979.  The results were published a few weeks ago for our report The Voter Power Index and showed that AV would mitigate some of the extremes of the First Past the Post (FPtP) system but it would in no way eliminate the problems. 

So consider this blog as a myth-busting blog as most of the ‘so-called’ facts being bandied about are certainly statistical exaggerations and many are simply un-truths. Below is a graph that summaries the changes to AV might have made in terms of the outcome of the 2010 general election.

Then let’s address some of the other so called ‘facts’.

• AV will be fairer: well slightly but it is well short of being a fair system. The Conservatives and Labour between them got 65% of the votes cast in England, Scotland and Wales, yet ended up with 87% of the MPs.  Under AV this would drop to 82%. Meanwhile Liberal Democrats got 23% of the vote but only 9% of the seats – which would rise to 13% under AV.  So a small step but hardly a giant leap.
• AV will allow parties to win from third or fourth place: theoretically possible but statistically unlikely – in our simulation of the 2010 election no candidate won from third place and only 41 seats had a winner from ‘second place’
• AV will create more extreme results: In six out of the seven elections studied the working majority for the sitting government would have been reduced, only in the 1997 Labour landslide would the majority go up, even then by only 10 seats (from 157 to 167)
• AV would always create coalitions: Only in the 2010 simulation did a coalition scenario emerge, all the other six simulations still showed working majorities.
• AV would allow more extreme candidates to get elected: The British Election Survey taken immediately after the 2010 election asked a representative sample of voters how they would have voted using an AV system.  This survey categorically shows the opposite effect – second preferences moderate people’s voting – ie the second preference of, say, a BNP voter is likely to go to the more moderate UKIP, and the second preference of a UKIP supporter to the more moderate Conservatives.

So there are the facts as I see them. I think it is a great shame they haven’t been discussed more as the changes that AV might make are broadly measureable and so could inform the debate. The greatest shame, however, is that the British Electorate have not been treated with more respect by the politicians who hold the power.