h1

Is Labour’s boost confined to places where it doesn’t matter?

October 7th, 2008

Could the move to Gord be in the heartlands – not in the marginals?

Firstly the caveats: As I’ve been saying for weeks polling during the conference season is almost always very odd and we need to wait until the end October surveys at the earliest before we can start drawing conclusions. Secondly looking at sub-samples in polling data can often deceive.

But on Saturday the Guardian’s polling writer whose judgement I respect, Julian Glover, produced an analysis on the latest national ICM voting intention survey which identified a trend, which if backed up in the coming weeks and months could have a dramatic impact on how national vote shares convert to seats won at the next general election. For Glover suggested that the recent increase in Labour support might be being confined to its heartlands.

    If this indeed does prove to be the case then the much-quoted 10% vote margin that the Tories need to achieve a commons majority might not be the requirement after all.

Calculations, like we see in the standard poll share>seats calculators are based on a uniform national swing. If there are disproportionate moves in specific areas or types of seat then we might be getting the wrong picture – which for a spread-betting man like me could be very expensive.

The move back to Labour in its heartlands might explain the vast disparity that we saw on Sunday with between the standard national survey and the ICM Labour marginals poll. The latter had the Tories doing considerably better and Labour doing considerably worse in what will be the key LAB>CON battlegrounds.

What’s taken the Glover thesis is a step forward has been the detailed data from the latest YouGov poll. Reproduced above is the regional break-down from that poll and a comparison with the last national YouGov poll before the local and London elections on May 1st this year.

The big disparities are in the north and even more so in Scotland. Just look at the shares in the October survey and compare them with April. You should note that the Scottish shares are not just a fluke from one poll. In the previous survey taken after Brown’s conference speech the Scottish and northern numbers were in the same area.

YouGov’s Scottish shares in its past two polls have been 42% and 43% which is somewhat higher that the 39.5% that Labour got there at the general election. This can only mean that elsewhere Brown’s party is polling disproportionately worse. There are, of course, very few LAB>CON marginals in Scotland.

This is one we are going to have to watch.

Mike Smithson






Comments are closed.