Yet another poll predicts a Labour landslide – Starmer’s greatest fear now is that people won’t bother to vote

Rishi Sunak and his election campaign chief, Isaac Levido, ought to be pleased. The latest pre-general election “megapoll” is much better for the Conservatives than the last one, offering a few tiny glimmers of hope for the party as they face down an almost certain – and, indeed, massive – defeat.

For starters, while Survation’s MRP poll in The Sunday Times at the weekend projected the Tories being reduced to 98 seats, last night’s YouGov MRP poll would only cut the Tories to 155 seats. It would still be the worst result for the Conservatives since the party took its modern form in 1834; worse than the 156 seats held against the Liberal landslide in 1906.

But at least this latest poll would give Labour a majority smaller than Tony Blair’s in 1997, even if it also gives the Tories fewer seats than the 165 the party won then.

And it puts Labour only 17 points ahead in national share of the vote, whereas the average conventional opinion poll puts Keir Starmer’s party 20 points ahead.

This would be no consolation to Jeremy Hunt, the chancellor, who retained his new Godalming and Ash seat in Survation’s model, but who loses it to the Liberal Democrats in YouGov’s.

We are getting used to these MRP polls, which use large samples and fancy modelling to estimate seat-by-seat results. It is no longer surprising to have reputable polling companies predicting that the next election will be a poly-Portillo event. Michael Portillo’s defeat as a cabinet minister in his Enfield Southgate seat was a one-off sensation in 1997 – but if Hunt loses his seat this time, he might be joined by Grant Shapps, Penny Mordaunt, Mark Harper, Alex Chalk and Michelle Donelan, among others.

There is nothing special about MRP (short for multilevel regression and post-stratification), except the level of detail it produces. The conventional opinion polls convey the same message – some of them a bit better for the Tories, some of them worse. But MRP polls tend to attract more attention because they project the result in individual seats, including those of cabinet ministers, and they offer precise figures for Labour’s majority: 154 with YouGov, 286 with Survation.

Conventional polls produce the same kind of estimate, depending on how you convert vote share into seats. Electoral Calculus, for example, produces a Labour majority of 268 on the current average of conventional polls. It uses a proportional swing model, whereas a uniform national swing calculation, which has been more reliable, produces a Labour majority of 128.

YouGov has just changed its algorithm to make it less proportional and more like uniform swing. But if Tories were hoping that this obscure debate among psephologists would save them, all it does is reduce Labour from an absolutely huge majority to one slightly less than Blair’s.

There are few other possible straws for the Conservatives to clutch. The new constituency boundaries will not save them. The net benefit to the Tories of equalising electorates to take account of population changes is about seven seats (the Lib Dems lose three, Labour two and Plaid Cymru two). That is trivial against projections of Labour majorities of 100, 200 or more.

Sunak and Levido can hope to squeeze the Reform vote, trying to make the argument that a vote for Reform is a vote for Starmer as prime minister. The trouble is that most Reform voters don’t care: they simply want to express their disgust at the Tories, especially over immigration, legal and illegal.

And it is beginning to feel too late for the obvious way for Sunak to recover lost ground: that things might get better over the next nine months, with tax cuts and possibly the Rwanda policy helping to win voters over. It seems just as likely that voters will feel that things are only getting worse.

There is one other thing, however, that might help Sunak limit Tory losses, and that is a possible paradoxical effect of all these polls suggesting a Labour landslide – one that would give Starmer a “super-majority”.

If it becomes accepted that Starmer is heading for a big win, the voters might shy away from delivering it. We saw some of that suspicion of handing a leader too much power in the 2017 election campaign, when Theresa May started with a 20-point lead and predictions of a landslide win. There were other reasons she ended up with no majority at all, but a desire to “clip her wings” was one of them.

There wasn’t much of that sentiment around in 1997, mainly because the opinion polls were not trusted – they had pointed to a Labour win in 1992 when John Major was re-elected. So, though the numbers said Blair was heading for a landslide, it came as a big surprise on the night, including to Portillo.

This time, the assumption of a Tory meltdown might take hold, and floating voters might hold back from ushering in the one-party “Starmer supremacy”, as this week’s Spectator has it – while left-ish voters might think it safe to vote Green.

The magazine’s cover is a cartoon of the House of Commons, with the Labour government benches packed like a football terrace, looming over a nearly deserted opposition side. No wonder Tim Carter, the former press officer for the Labour Party, warned that this kind of premature landslidism risked putting people off voting Labour: “Don’t fall for it from the left or the right, both using similar tactics in an attempt to suppress the Labour vote.”