I just finished reading the press release of a paper published last Tuesday in the British Medical Journal advocating taxing junk food -- you know the stuff, sugary drinks, chips, candy bars, burgers -- by at least 20 per cent if we want to reduce obesity rates. The authors at the Department of Public Health at Oxford University, also advocate subsidizing the good stuff such as fruits and vegetables.

Not to disparage such a prestigious institution, much of the media have jumped on this study because the paper has supposedly found a causal link between taxing "junk food" and a reduction in obesity, a modest reduction mind you, but a reduction nonetheless. This same topic was covered in the Globe and Mail. News flash: this is really old news. A number of studies published over the last decade in the US have found the same results.

Here's another news item: A number of studies, of similar prestigious pedigree, have also found no link between taxing junk foods and obesity levels as measured by mass body index. Now let me ask you, do you think those studies would have enjoyed the same news coverage regardless of how well they were researched? News is only made when the results seem to fit the politically acceptable storyline. In the literature this is called 'white hat' bias, and it's common in health and environmental studies. The more politically cememted the research finding, the higher the chances these studies will be published and reported.

The problem with most studies that try to estimate the impact of price increases on the level of junk food we consume is that they're based on 'price elasticities' or how demand responds to changes in price. And that's what they are, estimates, so any results in a way are self-fulfilling and often based on simulated modeling. Increase the price of a product, demand falls and you come to the conclusion that less junk food intake leads to lower levels of consumption and hence, higher and more sustained obesity rates.

Another problem is that many of these studies make the false analogy with tobacco. They assume that high taxes on tobacco led to the decline in smoking. I won't deny taxes had an impact, but rates of cigarette smoking started declining in the mid-1960s long before taxes took hold. Besides, taxing one product such as tobacco is far different than taxing a limitless range of food products where people have almost an unlimited chance to substitute one product for another.

But even with sugary drinks people have been cutting back without taxation. On a per capita basis, Canadians consumed 82.14 litres of carbonated soft drinks in 2010. This represented a 30 per cent decline from 1999, when consumption was 117 litres per person, according to Statistics Canada. A lot of people are switching to water, bottled water mind you, but water nonetheless.

The BMJ study also gives the impression that economists are in agreement that taxes are an effective way to bring down obesity. Not so. In a sample survey of 120 economists who were members of the American Economic Association, 60.8 per cent disagreed with a tax on unhealthy foods (which included sugary beverages), 14.6 per cent were neutral and 24.6 per cent agreed. This skepticism probably arises in part because such a tax would be extremely difficult to craft. In that event, the unintended consequences may end up creating problems with a misallocation of resources.

Many studies underestimate the difficulty of using incentives such as taxation as a way to change behavior. There are millions of adults throughout North America who are on some sort of diet spending billions of dollars trying to lose weight. And most dieters are highly motivated. What makes one think that adding a few cents to a can of soda will be more effective?

The authors of the BMJ study hope that taxes raised on junk food will be allocated to anti-obesity programs, a wish shared by many groups advocating similar solutions. Governments, however, are loath to dedicate tax revenues to specific programs. This means, in the end, all taxes end up in general revenues.

But don't surveys tell us that people want to tax junk food as a way to encourage healthy eating? That's what people tell us when nothing is at stake. But try raising taxes and people have a way of changing their minds as they did in New York State recently when a modest tax on soft drinks was proposed and quickly rejected.

If we've learned anything about obesity over the last couple of decades it is that getting rid of weight and keeping it off is one of the hardest things most of us will ever do. We are just now getting a sense of how our hormones operate and often undermine our best intentions.

So what works? To be blunt, we don't have many answers. When it comes to kids there's no substitute for good parenting. Consider that one proven predictor of whether kids will become fat is tooth brushing. Not that clean teeth keep you thin, but it indicates that parents making good choices for their children lends the best hope for a healthy society.

We also know that simplistic top-down approaches such as taxation or public announcements telling us to exercise and eat our vegetables don't work. The obese and overweight get that way for many reasons and that's why more nuanced solutions are needed. What we don't need are studies that cover the same ground leading nowhere. Here's an idea. Rather than taxing junk food, we might want to consider taxing junk science.

Patrick Luciani is the co-author, with Neil Seeman, of "XXL: Obesity and the Limits of Shame," published by University of Toronto Press . The book was a finalist for the 2011 Donner Book Prize for the best book in public policy.