Election algebra

I wrote a piece on trust in major political parties for the Sussex Express. Here's some evidence for a few of the claims that I made about the June 2009 elections.

"The press, and politicians have been worried that loss of trust in politics would mean low election turnouts, or a boost to minor parties."

See, for example, The Guardian

"Neither happened in June"

See graph - note the drop of over a million Labour votes. Other parties exchanged a couple of hundred thousand votes each. That's not to be dismissed, but all the four largest parties here, except Labour, polled better than in 1999. The BNP and Greens did get more votes, and 2 MEPs each, but their total vote numbers weren't enough to qualify as a radical breakthrough.

"the low Labour vote let two BNP members get elected."

Although their vote increased nationally, it actually declined in the two regions where they got elected. It's good to see the BNP vote declining in those areas. The Labour vote was down 180,000 in Yorkshire - 11,000 of those votes would have kept the BNP out there. They lost 240,000 in the North West region, 60,000 of those votes (or 5,000 UKIP or Green votes) would have kept the BNP out there. To be fair, other major parties saw significant drops in these two regions, too.

It would be interesting to try to work out some tactical voting strategies for the Euro elections, but it's much more complex than in First Past the Post. You'd need some pretty accurate polling results in your region to work out where your vote would be best placed to keep out the BNP. The answer is you need to place it with a party that has close to identical support, or close to double or treble the support of the BNP. It's definitely a drawback of the D'Hondt method that you can't express preferences.

These were the BNP votes:

	2004	2009
North West	134,959	132,094
Yorks & Humberside	126,538	120,139