The uniform swing was probably a good model in the old two-party system, but there are many signs that it simply doesn’t predict future elections very well any more.
Some time ago I wrote about some regional results that indicated that the LibDems were collapsing in the Tory-dominated south but doing well in the Labour-dominated north.
And now a new predictor tool, Resolver One, has appeared.
Basically, it uses a much more complex model that takes into account that different parties lose their voters to other parties in different proportions. For instance, Tory growth might come disproportionally from the LibDems; if so, constituencies with many LibDems are more likely to fall into the Tories’ hands than those with many Labour voters.
The result of this analysis is that they predict that the Conservatives will double their number of seats at Westminster, while Labour will hæmorrhage more than half their seats and the LibDems will lose almost as big a share of theirs.
If this prediction turns out to be true, it will be an almost unprecedented landslide.
I hope it turns out to be wrong, as I would much prefer a hung parliament.