The city mathematician, Martin Baxter (above) , has produced his latest prediction based on appyling the changes in the latest polls, weighted for sample size, to what happened in 2001 on a uniform national swing.

Baxter’s latest vote shares are: LAB 39.09: CON 33.4: LD 19.32. His calculator produces this predicted House of Commons – LAB 388: CON 176: LD 52.

There are three things that could go wrong with this prediction: the poll figures might be overstating Labour’s margin; the seat distribution might not follow the uniform models; and the Lib Dems might do disproportionately better in their target seats – all factors that happened four years ago although then the seat distribution worked very much in Labour’s favour.

In March 1992 the polls showed average Labour lead of 5%. The actual on polling day CON + 7.6% a Labour drop 12.6%

In March 1997 the polls had an average Labour lead of 27%. The actual on May 1st was 13% a Labour drop of 14%.

In April 2001 the average Labour poll lead was 19.6%. The actual on June 7th was 9.3% – a Labour drop of 13.6%.

Today’s Labour poll lead is less than 6% – what will the actual be on polling day?

Even if the Labour over-statement is only half what it was at the lowest of the last three elections then the two main parties are neck and neck. And then there’s the Lib Dems, who always seem to suffer from a massive understatement this far out, to factor in.

Another calculation to bear in mind is to work out what happens by taking the uniform national swing from the 1997 result and not 2001. Using the Hill & Knowlton 2001 calculator Labour gets 28 seats less than Baxter while the Tories gain 26 and the LDs put on 2.

Interestingly the current spread betting price buy prices are quite close to the Hill & Knowlton calculation. IG Index spreads are LAB 351-358: CON 193-203: LD 68-72. Tradesports currently have just under 5/2 on the Tories getting more than 205 seats.

Nobody is brave enough to predict anything other than a Labour victory with a working majority – but what will be its size? Punt at your peril!

© Mike Smithson 2005