There’s been some interesting discussion on LabourList about whether Labour should – yet again – pick Ken Livingstone as their candidate for London Mayor. Declan Gaffney has made use of some very interesting research into the demographics of London elections (published here).

That demographic analysis deserves a long second look; and once you do, the picture isn’t pretty for anyone who wants Ken Livingstone to be elected as Mayor of London.

The most striking feature in the demographic facts and figures is the polarisation of London politics between 2004 and 2008. The study looks at the election results, ward by ward, and looks for correlations with demographic data from the census. In 2004, 16 of those demographic variables were strongly correlated with voting for Ken Livingstone. The pattern doesn’t throw up many surprises: in areas with a younger population, a greater number of people of Black African or Caribbean heritage, few car-owners, high unemployment and deprivation – Ken Livingstone did well.

Moving forward to the 2008 election, the pattern is even more striking. All of the factors which strongly correlate with Ken’s vote in 2004 did so again, often more so. But a whole host of other correlations emerged in 2008 – perhaps very concerning ones for Labour strategists. There is a much clearer economic divide in the 2008 results. Compared to 2004, Livingstone was much more dependent on people living in local authority or housing association housing, doing worse amongst people with mortgages and private tenants. The correlation between Ken votes and unemployment and social deprivation goes from strong to massive. People in social groups D and E voted for Ken more strongly and those in AB and C1 voted against him much more strongly as well. (Interestingly, this appears to differ from the national picture: Labour support between 2005 and 2010 General Elections fell by most amongst C2 and DE and only a little amongst AB and C1, according to Ipsos Mori.)

Multi-car households voted strongly against him, while the carless strongly backed him; a reflection of the economic and class divide, or an indication of anger about congestion charging and bendy buses?

In short – the demographic picture is of Ken losing the 2008 election by losing the middle ground, and drawing strength from Labour’s strongest support on council estates.

In many ways it’s the same picture as in the recent General Election. Labour’s core vote in inner London turned out strongly; not people with huge affection for the Labour party, but still determined to do their part to try to get the devil they knew back in power. That’s why both Lib Dems and Conservatives found it difficult to gain ground in their targets from Labour, and why Labour fared so well in the London council elections on the same day.

There’s no doubt that Ken Livingstone is a good candidate to get Labour’s core support to the polls in 2012. But what can Labour in London do to claw back support from ABC1 demographics, the employed, car-drivers, and the people who don’t live on estates?

Many of these people have voted Labour in the recent past. Many of them will have liberal social values which, in the late 1990s, were best voiced by Labour. Writing them off now would be an epochal mistake for London Labour.

But winning back their support be a difficult job – and many will doubt whether Ken Livingstone is the man to do it.

While Labour’s standing as a party has plummeted, so has Ken’s. Of course he still has admirers. But if what people say on the doorstep is any guide, they tend to be the same Labour core voters. They aren’t the people that Labour need to reach out to win the 2012 Mayoral Election.

Fortunately for the Lib Dems, I think a Ken victory is likely, in a selection overshadowed by the leadership contest. We can hope that Labour will continue to barricade themselves amongst their core vote and that there’s an opportunity for the Lib Dem campaign to win the support of a middle class disenchanted with both Ken and Boris.