European election notes, IV
So, the BNP. Let's drill down to what we all have hanging at the backs of our minds.

I've poked around a bit with the BNP voting figures, and there's nothing desperately leaping out at me. I did eventually try using only the voting figures for England (the BNP in Scotland and Wales is a strange beast, and skews things a bit), and comparing the BNP's vote share in any given region to their national vote share.

At first glance, this doesn't tell us anything interesting - they're much stronger in the Midlands and the North; this year, a voter in Yorkshire was 50% more likely to vote for the BNP than the national average, and one in Cornwall 40% less likely.

But, year on year, the most intriguing region is London. The BNP took 18,000 votes in London in 1999, 76,000 in 2004, and 86,000 in 2009. This seems at first glance like an inexorable increase from a very small amount, but it turns out that in 1999, a London voter was about a third more likely to cast a BNP vote than the national average. In 2004, they were about a quarter less likely, and that remained the same this year. The South-West had much the same trajectory; in 1999 a voter was 20% less likely than average to vote BNP, but is now 40% less likely.

The converse happened in the North-East, which went from 25% less likely 1999 to 35% more likely in 2009, and Yorkshire/Humber, which leapt from average in 1999 to 50% more likely in 2004, and remained there.

Again, I've no explanation for these. But there ought to be something worth looking at here - why did the BNP not pick up votes in London in the same way it did elsewhere? Can we figure that out, and replicate it?