Of all the anti-obesity policies that have been tabled in recent years, calorie labelling is the least objectionable. It is not regressive like the sugar tax or anti-competitive like the advertising ban. If it's nannying, it is of the gentlest sort; a nudge rather than a shove. It would have been a nightmare for small businesses to implement, but they were given an exemption when the government made calorie counts mandatory in cafés, pubs and restaurants two years ago.

You could argue that the number of calories in a meal is basic information that consumers have a right to know, even if it doesn't make any difference to what they eat. That is just as well because we now know that it doesn't change what people eat. An evaluation funded by the National Institute for Health and Care Research was published this week and concluded that "the introduction of the mandatory kcal labelling policy in England was not associated with a significant decrease in self-reported kcals purchased". Monitoring the behaviour of 6,578 customers in pubs, fast food outlets, cafés and restaurants before and after the introduction of mandatory calorie counts, the authors found that the number of calories consumed actually rose from 909 per meal to 983 per meal.

The lead researcher Megan Polden said: "People did appear to be more knowledgeable about the number of calories in their meal. But that didn't unfortunately lead to any change in the number of calories that people were actually consuming." The most obvious explanation is that people have a pretty good idea of whether a meal is "healthy" or not without needing to know exactly how many calories are in it. When people are eating out, they want something tasty and are not particularly worried about how many calories are on their plate.

This study therefore provides a valuable service by disproving the idea that people eat the "wrong" food because they are ignorant. The policy worked insofar as it made people better informed, but making people better informed didn't alter their decision-making. From the perspective of "public health" campaigners, it was (another) failure.

Once again, a naive public health policy has disintegrated upon contact with reality. In light of this failure, and given how unpopular calorie labelling is with eating disorder charities, will the law now be repealed? Of course not. Every anti-obesity policy has been a failure but they remain on the statute books. Repealing ineffective laws would be a sign of weakness. The only answer is more laws. The medicine isn't working? More medicine! The authors of the evaluation claim, with hint of desperation, that calorie labelling could have worked if it was introduced "alongside other policies". Henry Dimbleby was on the airwaves on Monday demanding health warnings on breakfast cereal.

And yet it was supposed to work on its own. The nanny state lobby assured us that it was "evidence-based". Backed up by public health academics, the government's Impact Assessment assumed that it would lead to people consuming 41 fewer calories from out-of-home meals. This, in turn, would lead to £5.6 billion worth of savings and benefits over 25 years. Since the policy would "only" cost businesses half a million pounds a year to implement, this made it highly cost-effective.

This was sheer fantasy. Even the study the government cited to support its claim about the 41 calories concluded that "labelling of menus with calories alone had no effect on calories selected or consumed". By the time the Impact Assessment was published in 2020, two randomised controlled trials had been conducted in the UK to study the effect on calorie labelling in restaurants. The first found "no overall effect". The second found "no evidence that prominent calorie labelling changed daily energy purchased".

The policy went ahead anyway because politicians wanted to be seen to be doing something. Sound familiar? Calorie labelling is the smallest of examples, but consider hundreds of similar cases in every sector of the economy and you start to see how the government is slowly throttling business. The £500,000 annual cost of calorie labelling is relatively trivial, but the Impact Assessment for the ban on HFSS (high in fat, sugar and salt) food advertising, due to be implemented next October, uses the same fantasy modelling to claim that the policy will reduce energy consumption by 3.3 calories per person per day at a cost to business of £199 million per year. The costs are real. The benefits, I suspect, will once again fail to materialise.

Dr Christopher Snowdon is head of lifestyle economics at the Institute of Economic Affairs