Dietonomics

My oped today was on the economics of dieting. Full text over the fold.


The Economics of Obesity, Australian Financial Review, 28 July 2009

Ask an economist why obesity rates are rising across the rich world, and you’ll likely get two answers. First, the shift away from occupations requiring serious amounts of physical labour means that the typical worker burns fewer kilojoules on the job than in the past. Second, advances in agriculture and food technology have made kilojoules cheaper and yummier than ever before. (Empirically, it turns out that the former explained weight gain until about 1970, and the latter is the key factor since then.)

But would you trust an economist to tell you how to lose weight? Strikingly, part of the growth of behavioural economics has involved the dismal science turning its analytic toolkit to understanding why many of us overeat, and how we can reduce our food consumption.

The economic approach to estimating which kinds of dieting strategies succeed involves identifying cognitive biases, which are then tested in randomised experiments in the laboratory or real world. Coupled with this is the evaluation of policy changes, to see whether initiatives such as kilojoule labelling help or hurt consumers.

A trio of papers in the May 2009 issue of the American Economic Review summarises some of the new research in this field. Analysing eating patterns across the day, Marianne Bertrand and Diane Schanzenbach find that while Americans spend less time eating as their primary activity, it has been increasingly common to eat while doing something else (I hope you’re enjoying that sandwich, by the way). About a tenth of US kilojoules are consumed while watching TV. And while people tend to eat less when they have consumed more in the previous six hours, this rule doesn’t apply to television eating: people eat just as much in front of the television regardless of what they have recently consumed.

So could just turning off the television make a difference? To test this, Brian Wansink and co-authors ran a horse race between ten dieting strategies. Randomly assigning 1000 participants from a weight-loss website to one of the strategies, they were able to monitor compliance and weight loss.

The upfront finding – not surprising to anyone who has ever gone on a diet – is that very few people stick to the plan. No dieting strategy garners more than a 40 percent compliance rate.

Yet the successful strategies are not always those that are easiest to stick to. The best ways to lose weight are to use 25 cm plates (an average 0.9kg weight loss in a month), avoid eating while watching television (0.7kg weight loss), and eat fruit before snacking (0.5kg weight loss). If you’re munching through The Biggest Loser, you may be missing the point.

These results are consistent with the notion that we eat more with our eyes than our stomachs. Indeed, other evidence shows that those who decide when to stop eating based on external cues (of the kind ‘I stop eating when my drink runs out’) are more likely to be overweight than people who say they stop eating when they feel full.

Can public policy help too? One popular solution among politicians is to require restaurants to label the kilojoule content of each dish. In principle, one might think that more information can only benefit consumers. But by surveying customers before and after New York City’s 2007 adoption of such a law, Julie Downs and co-authors conclude that the rule did not reduce food consumption. Indeed, the authors find some evidence that dieters consumed more kilojoules after the labelling law came into effect, which they attribute to the fact that dieters tend to overestimate the amount of energy in a meal. If the legislature’s intention was just to increase information, then the law succeeded; but if they had hoped that it would help people lose weight, it was an abject failure.

Although behavioural economics is still a relatively young field, research such as this may be helping to bring a little more rigour into the notoriously unscientific diet industry. With obesity being one of the leading health concerns in developed nations today, even modest advances in our knowledge can have a substantial payoff. Who knows: perhaps the ‘behavioural economics diet’ will be coming soon to a bookshop near you?

Andrew Leigh is an economist in the Research School of Social Sciences at the Australian National University.

I was tempted to take a swipe at the end of the piece at the Dieticians’ Association of Australia for being unwilling to follow the lead of their US counterpart and publish a science-based critique of major diets. You can see why a peak body doesn’t like doing this kind of thing: all the quack members drop out of the organisation. On reflection, I probably should have put the boot in.

About these ads
This entry was posted in Health economics. Bookmark the permalink.

5 Responses to Dietonomics

  1. Michael W says:

    I am not sure whether you are aware of the science-based critique of reliance solely on randomised controlled trials.

    This is a critique from within scientific medicine of the reliance by the “evidence-based medicine” movement (exemplified by the Cochrane collaboration) for its reliance on published randomised controlled trials to the exclusion of basic science when examining medically implausible treatments.

    See

    http://www.sciencebasedmedicine.org/?p=553

    http://www.sciencebasedmedicine.org/?p=569

  2. Pedro X says:

    Would the idea of giving money away if you fail to give up weight qualify as a behavioural economics diet? This is what http://www.stickk.com/ seems to be.

    Any idea if there is anything solid that looks at that method?

  3. conrad says:

    I’m surprised how little economists care about other work. If price was the real reason, or even if its effect size was big, we should have seen obesity rise steadily everywhere food got cheaper vs real incomes, which was essentially everywhere that was rich. But we didn’t see that. There are massive differences across countries — Mexicans got fat, despite being far poorer than Australians, and the Japanese and Koreans somehow remained thin. So you are left with a hypothesis that is country specific and essentially untestable. You also have data going against the hypothesis within country (it would make a nice multi-level modeling example), because the less money people have in places like Australia, the fatter they are.

  4. Pingback: what we don’t know may make us fatter, depending on whether we’re in a sandwich shop or a hamburger restaurant « siphoning off a few thoughts

  5. samuel says:

    I would have thought the recent almost complete collapse of the developed world’s banking system, saved only by massive government intervention, would have for once and for all put to rest any arguments about the economic’s profession and it’s pseudo scientific basis ability to provide the answers to anything.
    Might I suggest you all go and get real jobs.

Comments are closed.