The reaction to some of his pledges was “a sharp intake of breath”, Keir Starmer said on Thursday. He welcomed that because “nobody can doubt the scale of our ambition, nor its urgency”. I fear he may be misinterpreting the responses. People are not saying to him, “That will be really difficult, but we admire your determination to push the system.” They are saying, “Are you out of your mind?” In his speech in Stoke, the Labour leader started to put numerical targets on some of his promises on crime. Two in particular were striking. He promised to halve knife crime, and to halve violence against women and girls.

Everyone should want these things, but how do you change human behaviour on this scale? Starmer said about knife crime: “We know so much of this is about prevention, about pulling young boys back before they get in too deep. It’s about good youth work, neighbourhood policing, mental health support – in every school. We’ll do all that.” That is admirable, but it will cost money. About which there was nothing in the speech, and when Starmer was asked about it afterwards, he referred to “efficiency savings”. On violence against women, he only got as far as the “first step”, which was “modernising the police”. He said: “You can’t defeat misogyny without robust policing, but you can’t have robust policing without defeating misogyny.” That is a soundbite, not a policy. If you look up the party’s policy, it does have one. A “green paper” called Ending Violence Against Women and Girls, was published in 2021. Its opening words are: “After a decade of underfunding...” It promises new offences, new minimum sentences, more support for women at risk, better education, more specialist units, and something about the internet. All to be paid for, it would seem, from “efficiency savings”. As for “halving the level of violence”, how is that to be measured? Over what period is it to be achieved? How was the target set? Has anything remotely like it ever been achieved anywhere in the world? How was it done? Basic questions; no answers. Something seems to have gone wrong in Starmer’s stately progress towards election preparedness.

He started well last year, setting out modest proposals, to be paid for by credible sources of new revenue. Abolishing non-dom status was to pay for Wes Streeting’s NHS workforce plan. VAT on school fees was to pay for breakfast clubs for primary schools. I assumed that there would be more like this, building up to a fully costed, stress-tested and believable programme of incremental improvements symbolising the larger ambitions of a Labour government. Instead, Starmer delivered a terrible speech last month about “missions”, setting himself an unrealistic target that means betting against the economic performance of the US, Germany, France and Canada. The first of his missions is to “secure the highest sustained growth in the G7”. From breakfast clubs to taking on the laws of economic history in one hallucinatory leap.

In the same speech, he repeated Labour’s promise to generate all of the nation’s electricity from green sources by 2030. I have found nobody who knows anything about energy who thinks that this is possible. Yet, at precisely the time a party that finds itself unexpectedly ahead in the opinion polls might start preparing for government by asking searching questions about the few policies it has, Starmer chose to reaffirm an impossible promise – and to add to it a gamble on other countries’ economic indicators, over which he has no control. Now, a month later, he has compounded his errors with promises to reduce knife crime and violence against women – with no new resources except imaginary “efficiency savings”.

The sudden other-worldliness of Labour’s promises makes Rishi Sunak’s impossible promise look easy by comparison. When the prime minister announced his five promises at the beginning of the year, most of them were derided as predictions of what was likely to happen anyway (inflation halving, debt falling, the economy growing).