Mark my words, Labour will panic before this election is over

The mood will turn. The day will come in the next few months when the opinion poll gap will start to close and Labour will panic. Headlines such as “Rishi’s Coming Up Roses” will be written in the Conservative press.

It is a law of nature that journalists and the wider politickerati cannot maintain a single note for long. Random variation and arbitrary events more or less guarantee that Keir Starmer will stumble or that Rishi Sunak will earn unexpected approval, and that public sentiment, running for so long against the government, will run, even if only briefly, in its favour.

It may not amount to much. The prime minister’s hopes that tax cuts and rising real earnings would reap the sustained gratitude of a relieved electorate seem unlikely to be fulfilled. The “time for a change” cry overrides crude economics just as it did in the run-up to the 1997 election.

But it feels significant that Andrew Bailey, the governor of the Bank of England, told the Financial Times yesterday, “We have an increasingly positive story to tell,” and hinted at interest rate cuts to come. If enough voters do feel better off later this year, it does at least give Sunak permission to rally.

You can see how it might happen. A fall in Labour’s opinion-poll lead could focus undue attention on outliers on the pro-Tory side, just as too much attention has been paid recently to outliers among pollsters that tend to be pro-Labour.

Journalists, bored with writing about Labour landslides, Canada-style meltdowns and Tory recriminations, will revel in writing the opposite. “I told you Keir wasn’t up to it”; “Tory strategists plot path to shock victory”; “Labour divisions reopen as poll lead slides”; “Lib Dem collapse means the blue wall will hold”...

There will be enough news stories, news analyses and commentary for everyone, even if the underlying facts haven’t changed very much.

As the election nears, increasing attention will be paid to the wide disparities in predictions for the outcome in seats. If you fed the average shares of the vote from this week’s opinion polls into the Electoral Calculus model, for example, it produced a Labour majority of 260. If you fed the same numbers into a uniform swing model, such as the one made by Ben Ansell of Nuffield College, Oxford, it produced a Labour majority of 90.

Uniform swing has always been a better guide in the past. Perhaps something in the way votes turn into seats has changed, but we won’t know that until after the election.

Meanwhile, if Labour’s poll lead goes down, it might approach the level – by my calculations currently a lead of about nine points – at which on a uniform swing Labour’s majority would disappear altogether and Starmer would face a hung parliament.

Starmer and other shadow cabinet pessimists are acutely conscious of this. I understand that Rachel Reeves, the shadow chancellor, is worried about the implications of a small or non-existent majority for the chances of getting radical planning reform through parliament. It was only a couple of years ago, after all, that Labour was posing as the “anti-developer party” across southern England.

Starmer and Reeves are right in any case to assume the worst. The great new book edited by Iain Dale, British General Election Campaigns 1830-2019, contains a brilliant chapter on the 1970 election by Michael Crick. That was an election that everyone (except possibly Ted Heath) assumed Labour would win, and which produced a big election-night shock.

With the opinion polls looking favourable, Harold Wilson consulted his inner cabinet about the date. “Everyone agreed?” he asked. Nobody dissented. “Right, then no one will be able to claim the virtue of hindsight.” Crick then adds: “When the rest of the cabinet arrived for their regular meeting a few minutes later, they were told nothing about the decision and told to say that no discussion on an election date had even taken place.”

It was the first election to be dominated by opinion polls, which had “become a prestige operation for newspapers”, says Crick, “and for the first time they were regularly fed to hungry broadcasters the night before as a form of promotion for the papers which commissioned them”.

But they were wrong. How much was down to polling error and how much to a late swing to the Tories has “never been properly explained”, but Crick recounts the symbolic moment just after midnight on the election results TV programme “when a graphic artist was seen painting extra blue numbers on the Conservative side of the BBC swingometer while Bob McKenzie carried on speaking – so unprepared had the corporation been for a Tory success”.

Every time it happens – and it happened again in 1992 and 2017 – journalists promise solemnly never again to allow election coverage to be so driven by opinion polls. Yet here we are, taking what is now an average Labour lead of 21 points as an enduring fact when it is no such thing.

Of course, the error may go the other way. The Tories may collapse altogether, be overtaken by Reform, and Starmer may find himself with a majority greater even than the 492 won by the National government in 1931.

One thing we ought to be sure of is that the way the election looks now is not the way it will look at midnight on polling day. Someone will, metaphorically, be painting extra numbers on the BBC set.

Another thing that is almost certain, though, is that between now and then there will be at least one news cycle in which the media will declare: “Starmer has thrown it all away.”