Gordon Brown’s AV Trap for the Lib Dems

Labour have been talking recently about a referendum on introducing a new system of voting. They have billed it “proportional representation”, but the system they will propose is the Alternative Vote system.

Under the AV system, instead of voters putting a X against a single candidate, they rank candidates in order of preference. When the votes are counted, first preferences are looked at first. If nobody gets more than 50% of the first preference vote, then the lowest candidate is eliminated, and their votes are redistributed to their second preferences. This process continues until a candidate goes over that 50% hurdle.

I thought it would be interesting to speculate on what the result of the last election would have been under such a system. Obviously, it is impossible to tell accurately what voters would have voted under a different system. But you can look at likely distribution of second and third preferences, and so predict what the outcome would have been.

I started with a dataset of all the election results for England, Scotland and Wales taken from The British Parliamentary Constituency Database, 1992-2005, Release 1.3 compiled by Pippa Norris, which you can find here on the Harvard University website.

I made some assumptions on the likely distribution of lower preference votes. If you’re interested, you can find details of the assumptions I made at the bottom of this post. (Of course, there is lots of room for disagreement on those assumptions, and different assumptions would produce a different result, but I tried to guess what real voters would do.)

With the help of Microsoft Access, I ran an analysis of what the result would have been in 2005 with the AV system and those assumptions I had made.

Here are the results (with the actual results that occurred in 2005 under first-past-the-post in brackets):

Conservative: 190 seats (197)
Labour: 359 seats (356)
Lib Dem: 66 seats (62)
Plaid Cymru: 3 seats (3)
SNP: 7 seats (6)
Others: 2 seats (3)

The astonishing thing about this result is that it is almost identical to the result produced by first-past-the-post!

If you believe this analysis (and maybe you don’t – such figure-crunching, well-beloved of Peter Snow, is never really robust) there are two conclusions to draw from this:

  1. AV is not a Proportional Representation system at all (although Labour have been promoting it as such)
  2. AV might make little difference to the election results in practice.

I think this sheds a good deal of light on why Labour have been doing this. This is all about party political calculation and not about fairness. The aim is to use that referendum promise to try and persuade the Lib Dems to support Labour in the event of a hung parliament after the next election, but not actually to give the Lib Dems a truly proportional system.

Will the Lib Dems fall for it?

=========================================================
Here are the assumptions I made about where second preferences would go:

England:

Labour Voters:
Conservative 20%
Lib Dem 40%
No second preference 40%

Conservative Voters:
Labour 20%
Lib Dem 40%
No second preference 40%

Lib Dem Voters:
Conservative 30%
Labour 40%
No second preference 30%

Scotland:

Labour Voters:
Conservative 5%
Lib Dem 10%
SNP 40%
No second preference 45%

Conservative Voters:
Labour 10%
Lib Dem 10%
SNP 60%
No second preference 20%

Lib Dem Voters:
Conservative 20%
Labour 30%
SNP 20%
No second preference 30%

SNP Voters:
Conservative 40%
Labour 10%
Lib Dem 10%
No second preference 40%

Wales:

Labour Voters:
Conservative 5%
Lib Dem 40%
Plaid Cymru 20%
No second preference 35%

Conservative Voters:
Labour 5%
Lib Dem 40%
Plaid Cymru 5%
No second preference 50%

Lib Dem Voters:
Conservative 20%
Labour 20%
Plaid Cymru 10%
No second preference 50%

Plaid Cymru Voters:
Conservative 5%
Labour 20%
Lib Dem 5%
No second preference 30%

About these ads

6 thoughts on “Gordon Brown’s AV Trap for the Lib Dems

  1. Nah, the way forward is multi-member constituencies, each returning two or three or [however many] MPs.

    The voter can choose whether to vote for a named candidate or a party generally, and after [the correct number of ] MPs have been elected in each MMC, each then chooses a sub-constituency whom he or she will represent in the same way as MPs do now.

  2. [...] I thought it would be interesting to speculate on what the result of the last election would have been under such a system. Obviously, it is impossible to tell accurately what voters would have voted under a different system. But you can look at likely distribution of second and third preferences, and so predict what the outcome would have been. via adamcollyer.wordpress.com [...]

Comments are closed.