h1

How the Tories could win with a 7% poll deficit

October 27th, 2004


We publish this article not because we think that it will happen but to demonstrate how with support for the two main parties in the 30s a totally different dynamic will apply at the next election and issues like polling accuracy and the tactical unwind will become even more critical. Last time Labour was so far ahead that inflated leads did not matter. Now they do.

Today’s ICM poll in the Guardian – LAB 37% CON 31% LIBD 23% – should make quite welcome reading for Tony Blair. But if it is as accurate as the final week survey that they did for the Evening Standard and published on the Tuesday before the 2001 General Election then on a uniform swing calculation Labour is 23 seats short of a majority.

CON 34.7% LAB 32% LIBD 22.8% ICM October poll with June 2001 “corrections”
LAB 301 CON 255 LIBD 59 seats Martin Baxter uniform national swing seat calculator
LAB 284 CON 289 LIBD 44 seats. Andy Cooke tactical unwind seat projection.

Of course we think the Cooke projection underestimates the Lib Dems and this analysis is grossly unfair to ICM. They did two polls in the final week in 2001 and the second one was more accurate – though it still under-stated the Tories and over-stated Labour. But the Standard survey had a much bigger sample and the interviewing started at the same time so should have been more accurate.

    In the past five weeks the Labour’s poll share has been all over the place – down as low as 28% and as high as 39% – while the Lib Dem and Tory figures have been much more stable. Labour seems to do much better with ICM, the Lib Dems seem to get better ratings from Populus and YouGov tended to show higher Tory shares until UKIP emerged in the early summer.

What this mean is that risk-averse gamblers, particularly on the spread markets, should proceed with extreme caution. We believe that there will be money to be made because the pollsters are not infallible and that there will be an element of tactical unwind. There is no point in betting at the moment – only move if the spread prices move in a totally perverse way as they did after the July by-elections.

Sporting Index LAB 338-346: CON 202-210: LIBD 72-77 (No change)
IG Index: LAB 338-346: CON 208-216: LIBD 71-75 (No change)

We also believe that the Labour price of 1/6 to win most seats should be avoided. The party is highly likely to come out on top but the risks are greater than the returns at this price.






Comments are closed.