If it wasn’t for Jimmy Savile I’d be dead. I suppose it was inevitable that one day I would run my car smack into the side of a bus and that it would happen because I was discussing the centrality of regional government to the German social market model when I should have been watching the road. But when the moment came I didn’t die.

You see, I had followed the old DJ’s advice from the adverts of my youth — I had clunk clicked. And my seatbelt saved my life. So you may think what I am about to say perverse. I don’t think making people wear seatbelts saves lives.

In his book Risk, John Adams, of University College London, explains it all very simply. Yes, if you are in a crash your seatbelt may well save your life. Yet you are more likely to be in a crash if you wear a seatbelt. Why? Because with the belt on you feel safer, and may drive more carelessly. So making the wearing of seatbelts compulsory does not reduce the number of fatal accidents, it redistributes them from the driver to the people they crash into.

The data produced by the professor to support his view on seatbelts are very convincing. Yet the detailed dissection of car accident data is less important than the more general point that arises out of his work. People don’t just sit still and allow themselves to be regulated by government, they change their behaviour in response to the law. And they may do so in ways that render the regulation worse than useless.

My car accident happened 15 years ago and Risk was published in 1995. So why am I telling you all this now? Because of David Cameron and the speeches he’s been making about absent fathers.

It was in his party conference speech last year, that Mr Cameron first started talking about his big theme: social responsibility. What this meant, it turned out, was that he intended to encourage people — retailers, families, faith schools — to behave responsibly, but that he wouldn’t necessarily do that through new laws. And the response was predictable: what does all this amount to? Where are his policies? Where’s the beef? Fine, so David Cameron doesn’t like absent fathers, thinks they are as bad as drunk drivers, but what is he actually going to do?

I find myself almost alone as a social responsibility militant. I don’t just believe that government can change behaviour through persuasion as well as through law, I believe that persuasion is often much more effective than law. In fact, I believe that laws won’t work without persuasion.

Some new research may help you to understand why. Last year Parliament agreed a ban on smoking in public places. Medical evidence of the harmful nature of passive smoking was critical to this decision. But strangely, studies that measure the impact of smoking bans on nonsmokers have been remarkably thin on the ground.

Now Jérôme Adda and Francesca Cornaglia have attempted to fill this gap with a paper published by Bonn’s Institute for the Study of Labour. The authors have used data on cotinine concentration in body fluids, a measure of the impact of passive smoking that is routinely used in medical literature. They then looked at how the cotinine data changed as policy changed in different parts of the United States. They also broke the data down so that the impact of policy on different age groups and classes could be analysed.

Their results are striking and show just how impoverished was the debate we had before the smoking ban was agreed. Bans do not reduce passive smoking — they redistribute it. Better-off adults show less signs of exposure, while poorer children are more exposed. The public health lobby promoted the smoking ban partly because of their intense concern about health inequality. The bad news from the new study is that smoking bans increase health inequality.

If you ban smoking in restaurants and bars, some users — those who are mainly social smokers — may indeed give up. But others simply go home and smoke. In winter they stay indoors, close the windows and allow their children’s cotinine concentration to soar.

Taxes have a different impact. They reduce the impact of passive smoking among children, but not among adults. Tax increases reduce the number of cigarettes that people smoke, but there is some evidence that they actually smoke more of each fag. Tax also appears to increase the proportion of cigarettes smoked at home rather than socially.

The new paper, then, echoes Adams’s seatbelt findings. People don’t just sit still and let government act on them.

One idea, of course, would be to ban smoking in the home as well. The problem is that you would have to enforce such a ban. Quite apart from the civil liberties questions, how would you do that? How would you find the resources for a start?

So what would work? The answer lies in Scandinavian drink-drive statistics, as answers often do. In the 1970s it was believed that Scandinavia owed its low rate of alcohol-induced accidents to its draconian legislation. Then H. Laurence Ross produced his study, The Scandinavian Myth. Ross’s data suggested that drink-drive laws only work when they ratify strong and well-established public opinion. The law follows changing behaviour rather than the other way round. A ban won’t stop people smoking at home, but social pressure might. Laws won’t make absent fathers look after their children, but social disapproval might.

That’s why I am a social responsibility militant. All talk and no action, that’s my slogan. Where’s the beef? Who cares.

Times Recommends

• We have kit and leadership. Let’s go
• Behind the scenes at the biggest deal
• Conferences — what a waste of money