Colin Rallings and Michael Thrasher yesterday published their now traditional pre-local elections set of projections. Based on performances in council by-elections, they are projecting the Liberal Democrat vote share in May to be 16% – the same as in their earlier prediction which I covered in Liberal Democrat Newswire #75.
Is that good, bad or indifferent? Here’s the context which tells you the answer to that…
First, what that 16% means. Each year the round of seats up for election varies, which means simply comparing vote shares from one year to another is not comparing like with like. If one year, say, lots of Labour-leaning urban areas have elections but the next year they don’t, then an apparent drop in the Labour vote may simply reflect the change in seats being contested rather than an actual change in level of support.
That is where the “national equivalent vote share” comes in. It is an adjusted vote total which allows for the changing set of seats up each year, so that you can make like-for-like comparisons between different years.*
Here’s the Liberal Democrat performance over the last few years on this measure:
So if the Lib Dems were to get 16% that would be a welcome change in the trend over the last few years and would be slightly up (+1%) on the last time the bulk of this year’s seats were up for election, namely 2012. It would also be well up on the party’s current poll ratings and on the last couple of years of local elections.
However, as I wrote in that edition of Liberal Democrat Newswire:
The Thrasher and Rallings predictions have consistently over-estimated the eventual Lib Dem vote share, doing so on all seven occasions I have data for.
In the last Parliament, that over-estimate was 2 points, so overall it’s touch and go on these figures whether the Lib Dems should expect to be up or down a little on 2012, but it looks much more certain the party will be up on the last couple of years.
As for what that means in terms of seats, Thrasher and Rallings predict 40 Lib Dem net gains in the council elections.
Although they’ve consistently over-predicted the Lib Dem vote share, due also to errors in predictions for other parties, their record on predicting Lib Dem seat changes is very varied, and in fact the last two came in at being worse than reality turned out to be. But the variance across the years is so huge, it’d be unwise to read too much into those last two on their own.
Moreover, the Lib Dems have made net gains only once in the last 9 rounds of local elections (note how that predates Coalition) so my instinct is that it would be unwise for Lib Dems to set their expectations purely by the Thrasher and Rallings seat predictions.
Or in other words, the evidence so far is that the party is recovering from last May, but there is clearly a very long way yet to go in order to get back to where we were at the start of the last Parliament. That recovery won’t happen in just one year.
* There are two such national equivalent vote share calculations made each year, one by the BBC and one by Thrasher and Rallings. They use a different methodology – one looks at vote totals and one looks at vote changes, to put it crudely. So the fact that they consistently tell the same story is good evidence that the calculations really are properly allowing for the variation in which wards are up which year.