Starmer may be on course for No 10 and coalition of chaos

The Conservatives did so badly in the local elections that their own spin backfired. Greg Hands, the party chair, said he was expecting to lose 1,000 seats, with the transparent intent of claiming a good result if it turned out to be not quite as many. With one council to finish counting on Monday (Redcar and Cleveland), the Tories have lost 1,061 seats.

But Keir Starmer is engaged in a spin operation of his own. The results were good for Labour, but he insisted: “Make no mistake, we are on course for a Labour majority at the next general election.” He was contradicted by John Curtice, the doyen of psephologists, who said that if Thursday’s votes had been reflected in a general election, the result would probably have been a hung parliament.

Curtice heads a team of number-crunchers who produce an estimate for the BBC of how the whole country would have voted if there had been elections in Scotland, Wales, London, and the other parts of England that didn’t vote on Thursday. Their final estimate was a Labour lead over the Conservatives of nine percentage points in share of the vote.

He was supported by Michael Thrasher, who carried out a similar analysis for Sky News. Thrasher’s latest estimate is a Labour lead of seven points. These estimates both fall short of the “double-digit lead” that Curtice set in advance as the threshold at which Labour would be able to say it was on course for a parliamentary majority.

Extrapolating from local votes to produce a national picture is hard enough; projecting that picture in turn onto a notional general election requires a host of further assumptions, giving Labour spinners plenty of scope to quarrel with the experts. There is no way that the Liberal Democrats will win 20 per cent of the vote in a general election (Curtice) or even 18 per cent (Thrasher), says Labour: that vote would transfer in Labour’s favour when people knew they were choosing a government for the country. I am afraid this is a faith-based rather than an evidence-based assumption.

A better argument by Labour spinners is that Scotland has moved Curtice’s “double-digit lead” target in their favour. The party’s gain in support at the expense of the Scottish National Party means that it can hope for a 20-seat bonus in the next parliament. If we are generous to Starmer, we could say that, if the local elections had been a general election, they might have produced a range of outcomes – from a hung parliament, in which the Tories were the largest party, to a tiny Labour majority.

However, such an artificial exercise can take us only so far. Things will happen between now and the general election, and in a general election people will vote differently. The predictive power of local elections is low, but there is one potentially significant fact that Stephen Fisher, one of Curtice’s colleagues, has pointed out. In all but one of the sets of local elections held under Tory governments since 1983, Labour has performed worse at the general election later that year or the year after.

This cannot be an iron rule, but the exception to it is interesting: in May 2017, Labour did disastrously badly in local elections, only to recover dramatically in the general election a month later, depriving Theresa May of her majority. That does not seem to me to be a useful model for what might happen at a general election next year. On the contrary, I would say it is likely that the outcome for Labour in the general election will be worse than the seven- to nine-point lead it secured on Thursday. Which is why I have been writing about hung parliaments every now and then for months, when I have not been writing that the chances of Rishi Sunak retaining power are higher than most people seem to think.