Forecasting local elections is a fool’s errand. But let’s have a go anyway. There are numerous predictions already out there, some of which, regrettably, smack a little of expectation management. Having watched and reported on local elections since 2013, I, alongside my colleague Andrew Teale, noted some predictions identifying particular boroughs as easy Tory or Labour gains when, on the ground, nothing could be further from the truth. Bolton, for instance, was labelled by one forecast as an easy Labour gain, but to achieve that the party would have to win every seat on the council up for grabs this year, which has never happened in the borough’s history.
This week I sought to model what will happen across Great Britain on Thursday. My calculations reference the latest opinion polls and feature a regression model of ward-by-ward results for the seats up this year since 2015.
While modelling any local election is difficult, this year’s are doubly so. A large portion of the seats in Wales and Scotland, for instance, feature significant support for nationalist candidates, something opinion polls have a hard time picking up. Most of the seats up in England were last contested in 2018, a year before the collapse in the Labour’s base. In essence, the arithmetic has changed quite drastically. For England, what that means in practice is that comparing directly with 2018 and nothing else — assuming that because the Tory vote has fallen since 2018 Labour will make gains in these seats — is nonsensical. Today, while Labour’s vote is similar in size to 2018, the components that make up its base are drastically different. In 2018 Labour was winning scores of seats in Grimsby, but from 2019 onwards the Lincolnshire borough has voted decisively for the Conservatives.
The forecast
My forecast, which attempts to take this into account, predicts that the Conservatives will lose more than 200 council seats across Great Britain on polling day (Thursday 5 May). They will suffer net losses of 63 in London, 38 across the rest of England, 83 in Scotland and 22 in Wales.
Labour, meanwhile, will make a net gain of 35 council seats in London but a net loss of 16 across the rest of England. That net loss, however, will be more than compensated for by net gains of 87 and 41 in Scotland and Wales respectively.
These figures suggest Labour will still lose some seats when compared with 2018 but that it will overtake the Conservatives and to have the second largest number of councillors in Scotland.
For the Liberal Democrats, my model suggests a net gain of 34 seats. The Greens are forecast to achieve net gains of 35.
In Scotland, the Scottish National Party looks set for a middling set of results. The central forecast expects them to lose 12 seats but it is possible they may shed as many as 25, or even, on a good night, make a net gain of three.
These figures should be treated with caution. I have provided both a central forecast and a range of reasonable probability. The ranges (best and worst) in the graph above help to illustrate the level of uncertainty in local elections such as these. The Conservatives are on course for a slew of losses and Labour are expected to do better than at the 2019 general election. But the seats being contested in England muddy that impression, since most were last up before Labour’s rout – before the party’s base took flight. That context, when analysing the results over the next few days, is key.