What are these doing to the seat calculators?

What are these doing to the seat calculators?

Could their rise be disguising the scale of the swing?

The big trend from almost all the polls in recent weeks week has been the increase in the share for “others” – UKIP, the Greens, the BNP and SNP/PC in Scotland and Wales. In some surveys they are now more than double the 8.2% that they got between them at the 2005 general election

A question for anybody wanting to bet on and/or predict the outcome is whether these historically high levels might be distorting the standard seat calculators that we use to project how many MPs the parties will end up getting?

With the exception of perhaps just one or two seats the impact of this dynamic in a first past the post system is to reduce the aggregate shares going to the main parties and it is on that last group of numbers that we seek to project seat numbers

Taking England alone it might be an idea for projection purposes only to regard the UKIP/GREEN/BNP effect as being broadly neutral in terms of impact on the main parties and do some rule-of-thumb calculations from there. So you would assume a 2005 level for others and try to extrapolate.

You could see a formula evolving where you take the difference between a current survey’s total for others and the 8.1% of 2005 and divide that between Labour, the Tories and the Lib Dems in the same proportions as the headline figures in the poll.

Thus the most recent ComRes poll had C39-L25-LD17 – an “other” aggregate of 19 points or 11 up from 2005. You would then apportion that figure, for seat projection purposes only, in accordance with the three party split.

On a simplistic projection using the Anthony Wells calculator that adds 14 seats to the Tory total. All of this is only a thought and I’m sure that I’m not the only one who has been giving it thought.

The challenge of factoring in the others is another reason why I’m increasingly more sceptical of the much repeated mantra that the “Tories need a 10% margin for a majority” argument. I think that the bar is probably a couple of notches lower.

Mike Smithson

Comments are closed.