Working the swing out from 1997 – not 2001

The standard way of working how many seats a party will get with a given share of the vote is to use one of the seat calculators, most commonly Martin Baxter’s, which makes its prediction based on applying the suggested uniform national swing to what happened in each of the seats in 2001.

But what happens if instead you compare today’s vote shares with what happened in 1997 when a much younger Tony Blair (above) led New Labour to its stunning landlside victory?

For 2001 was an extraordinary election for Labour which managed see its margin over the Tories decline by more than a quarter but still be returned with a majority of 160+. This was because the party did so much better in its marginals and the fact there was more tactical voting.

But what happens if these two factors do not happen again on the same scale which seem likely? Why not apply recent poll shares to the 1997 General Election result. This is possible thanks to the Hill & Knowlton 2001 calculator which is still available on-line.

When doing this you have to factor in the fact that there will be fewer seats this time because of the Scottish boundary changes.These are reckoned, on a notional basis, to have “lost” Labour 10 seats and the Tories, Lib Dems and the SNP one seat each.

So taking the February ICM poll of LAB 37: CON 34: LD 21 you get the following from Hill & Knowlton with comparisons on doing the same calculation with Baxter. LAB 336 (-33): CON 215 (+26) : LD 63 (+9).

Thus instead of Baxter’s prediction of a Labour majority of more than 90 Hill & Knowlton, adjusted for Scotland, gives you a Labour majority of just 25.

Which one’s right? Hard to say but the special factors that saw Labour get so many seats for its vote in 2001 won’t exist to the same extent as last time. Making the calculation, as H&N does, on the 1997 result puts today’s uniform national swing predictions into perspective – vital for all those spread-betting.
© Mike Smithson 2005