Those supporting the “Ken is a certainty” position are looking at this as though it was a General Election where image and how things appear in the media are the critical factors. I’m looking at it as a local election where there’s a totally different dynamic.

One almost unknown aspect the UK election process is that after each election the authorities are required to make available a copy of the electoral register showing those people who actually voted. The party activists get this information and key it into their computers and then add information on voting intention gleaned from canvassing.

In a local election with a 35% turn-out all this data is magical. People are either voters in local elections or they are not and the good party machines know precisely who to target – those with a proven record of voting who have indicated in the past they are your supporters.

When the first election took place in 2000 no one really knew whether the electorate would behave as at local or a General Election. Thus the final week survey by ICM had 67% of Ken’s supporters saying they were “certain to vote”. The final turn-out was in the mid-30s and Ken got 12% less than the final poll with Norris getting 10% more. The Tories and LDs who vote when the turn-out is so low are the loyalists and are less likely to switch.

There is a direct link between Lab’s performance in all-London local elections and their ICM and Mori poll ratings.
In 98 Lab votes in London were 40.6 against a poll rating of 53. CON 29 polls-32 actual
In 2000 Lab votes in London were 33.2 against a poll rating of 48.5. CON 29.5-33.5 actual
In 2002 Lab votes in London were 34.1 against a poll rating of 48. CON 28 polls-34.3 actual

Thus histrorically Labour gets 13-15% LESS in London than its national opinion poll rating.
The Tories gets 3-6% MORE in London than the national opinion poll rating.

In addition the Lib Dems consistently get 3-5% MORE in London local elections than their national polls rating. In 98 it was 20.8 actual against 15% in polls; in 2000 it was 18.9% actual against 15 in polls; and in 2002 it was 20.6% actual against 17% in polls.

The March ICM poll was Lab 37; Con 35; LD 21.

Based on all the historical trends it is conceivable that in the June elections the all-London vote could split Con 38; Lab 25; LD 25. If the LD and Con supporters remained loyal to their parties then the Mayoral result could be very interesting. If Hughes gets second place he’ll win on second preferences.

In local by-elections in London this year the only seat to change hands was in Harringey where Lab lost to LD after seeing its vote cut by more than half on 2002.