You are viewing j_rentoul

John Rentoul

John Rentoul is chief political commentator for The Independent on Sunday, and visiting fellow at Queen Mary, University of London, where he teaches contemporary history. Previously he was chief leader writer for The Independent. He has written a biography of Tony Blair, whom he admired more at the end of his time in office than he did at the beginning.

"The Independent's must-read man" - Daniel Finkelstein

You can contact John in the comments area or email him at

Previous Entry | Next Entry

How high is Cameron's mountain?

Posted by John Rentoul
  • Tuesday, 24 February 2009 at 10:05 am
I am grateful to Mike Smithson at and Anthony Wells at UK Polling Report for making me go back to my spreadsheets. They pick me up a point I made at the weekend

The quirks of the system mean that the Conservative share of the vote has to be eight percentage points larger than Labour’s to win a Commons majority. But I suspect that the opinion polls still overstate Labour’s support. In the last four election campaigns, Labour has been overestimated by an average of five points. If that is still true, the Tories need be only three points ahead in the opinion polls to win.
I am saved only by the "if". Smithson and Wells are quite right to point out that the record of the pollsters was better in 2005 than in the previous three elections. The average of all the polls during the last election campaign overestimated Labour support by 1.8 percentage points, and underestimated the Conservatives by 0.4 points, thus overestimating Labour's lead by 2.2 points.

(Source: my own analysis of UK Polling Report historical data. I have taken all the polls during the campaigns rather than simply the final one from each pollster. I know that this introduces campaign effects, such as that of the leak of the Attorney General's long advice that focused that last week of the 2005 campaign on Iraq, but I think it provides a more meaningful picture from our vantage point, before an election campaign has even started.)

I had thought that Labour was overstated in 2005 by 5 points, because my original data excluded YouGov, for which it was the first full campaign. As you can tell, YouGov tended to report lower Labour figures than other companies. Interestingly, this is no longer the case. As the following table shows, YouGov still reports the highest average Conservative figure, but it is currently ComRes, pollster for The Independent and The Independent on Sunday, that reports the lowest average Labour figure.

                      Con      Lab       LD
ComRes     39.3     30.3     17.4
ICM               39.4     32.0     19.4
YouGov        40.6     32.8     15.5
IpsosMORI  40.5     34.0     16.0
Populus      38.9     32.7     17.1

(Average, comparable months, March 2007 to January 2009.)

A commenter at also pointed out that some estimates suggest that, on a uniform swing, the Conservatives need to be 10 points ahead, rather than eight, in order to secure a majority in the House of Commons. This is certainly what Martin Baxter's excellent Electoral Calculus website suggests, if you try feeding, say, Con 40, Lab 30, LD 21 into his prediction machine. Glad to hear from anyone with further and better particulars.

So, it would seem that the Conservatives need to have an average lead in the polls during an election campaign of six to eight points in order to expect to win a majority. Their current 14-point average lead still gives them a cushion.