On Monday, on ABC radio's The World Today, the Australian Association of National Advertisers let parents in on a secret: ads for food directed at children don't succeed in persuading them to eat more.
As the association's executive director, Collin Segelov, put it: "If consumption was as easy as advertising, then, my golly, everything would be easy. It doesn't work that way."
That surely can't be what advertising agencies tell their clients.
I can't imagine them saying: "Listen, this advertising campaign isn't going to grow the market, but if you merely want to fight over market share, who am I to stop you?"
Most of the food advertising that you see is where you've got companies competing with products against one another. So they're trying to get their brand into the equation. They're trying to get people to look at their brand as against someone else's brand.
It's a familiar argument. In the early 1970s, when there were moves to ban cigarette advertising in Australia, the manufacturers insisted ads didn't encourage smoking; they merely encouraged brand switching. As it happened, once the ads were taken off TV, smoking rates began to slide.
Cigarette companies also claimed that without their ads and sponsorship Australian television and Australian sport might go broke.
It's an argument Segelov has echoed this week in his campaign against moves to stop food companies advertising to children and sponsoring junior sport. He told the Herald: "If someone pulls the plug, then the sport could disappear"...
Perhaps fortunately for the advertising industry, the most comprehensive study of the evidence to date suggests that promoting food to children does encourage them to eat more of it. The University of Strathclyde in Glasgow concluded, after examining more than 100 studies at the behest of the British Food Standards Agency in 2003, that there was "sufficient evidence" to suggest food promotion encouraged consumption.
One of the studies examined obesity in Quebec, which banned television advertising directed at children in 1978. It enjoys the lowest obesity rate of any Canadian state.
Another used detailed diaries to record children's TV viewing habits. It found that the more food advertisements they saw, the more snacks and calories they consumed.
Late last year a study conducted for the US National Bureau of Economic Research concluded that a ban on advertising fast food restaurants to young people would cut the number of overweight children by 10 per cent and the number of overweight teenagers by 12 per cent. It found that even the more modest step of removing tax deductions for such advertising would most likely cut obesity among children by between 3 and 5 per cent.
They are gains worth having, and there is a precedent for them in Australia. We ban the alcohol ads until after 9pm. But advertising bans by themselves would make only a dent in childhood obesity.
The latest economic research on the growth in obesity in Western nations suggests that its causes are complex and not easily reversed. Writing in the Journal of Economic Perspectives, David Cutler and two colleagues from Harvard University put a lot of the blame on technological progress. Forty years ago most of our food was prepared laboriously at home. In 1965 it took a married woman who was not in paid employment two hours a day to cook, and then clean up after, a family meal. After decades of innovations, including vacuum packaging, deep-freezing and microwave cooking, it now takes half the time.
And the food is more processed.
As an example, Cutler says that Americans ate large quantities of potatoes before World War II, but the potatoes were usually baked, boiled or mashed. Chips were too hard to make, even for most restaurants. Now french fries can be peeled, cut, cooked and frozen in a few central locations using sophisticated technologies. Today the french fry is the dominant form of potato and America's favourite vegetable. Since 1977 potato consumption in the US has climbed 30 per cent, almost exclusively in the form of potato chips.
Cutler reasons that if it takes less effort to do something we like, we will do more of it. Only about a third of Americans reported eating two or more snacks a day in 1977. By 1996 it was almost a half.
And food is a lot cheaper, relative to earnings, than it used to be. Cutler uses the Economist magazine's so-called Big Mac index to show that the more affordable a country's Big Macs are (and, by implication, the cheaper its processed foods, generally), the more obese are its citizens.
Other economists lay blame at the feet of working women. Kristin Butcher and a team from the Federal Reserve Bank of Chicago conclude that the more hours a week a mother works, the more likely are her children to be obese. A mother who works an extra 10 hours a week appears to increase her likelihood of having an obese child by 3 to 4 per cent.
Curiously, this appears to be the case only for well-off families. The number of hours worked a week makes little difference to obesity rates among low-income families, which are, in any event, far more likely to raise obese children than are high-income families.
Depressingly, the economic research offers little usable advice on how to reverse the sudden growth in obesity. Redistributing income and unwinding technological change hardly seem practical.
But banning the advertising of junk food would be. It would be a start.