"Calorie Labeling and the Dunch Problem"
I wrote earlier today about the problem of large portion sizes in which many of us who might have one kind of desire to consume fewer calories nonetheless find ourselves drawn toward high-calorie orders in the moment.* Ezra Klein’s latest column in the print Post also tackles this subject and reports on the idea that mandatory nutritional labeling could make a surprisingly large difference:
We’re still waiting for the full data from New York’s experiment. But the researchers there shared unpublished numbers with the County of Los Angeles Public Health Department, which was preparing an analysis in case Los Angeles wanted to follow New York’s lead. Based on those numbers, Los Angeles researchers settled on a “conservative” estimate: 10 percent of chain restaurant patrons would order meals that were merely 100 calories lighter.
Surprisingly, that mild change in behavior has a huge and immediate effect: It would avert 38.9 percent of the county’s expected weight gain in the next year. If 20 percent of patrons order meals with 150 fewer calories, it would avert 116 percent of the expected weight gain, which is to say that the County of Los Angeles would actually lose weight.
Unhelpfully, the print column does not include this useful table which Ezra has previously blogged:
Now of course you’ll hear a libertarian argument to the effect of, “if people really wanted to know this stuff the market would respond automatically” which I think you’d have to say was naive at best. I do think that part of the key to making this have the desired effect is to be crude and obvious with the labels:
Chain restaurants will have to list caloric information on their menus and menu boards. Not behind the desk, or off to the side, or up on the ceiling. Where you can see it. New York, among other cities, has already instituted that policy. Every Starbucks in Manhattan now must post the calories in a MochaFrappaWhatsIt right next to the drink name.
What seems really wrongheaded about the NYC law is to limit its effect to chain restaurants. If the data on this kind of very soft paternalism looks promising, then I’d want to see its scope expanded.
* I was talking to Neil Sinhababu recently and he usefully drew a distinction between desires you would attribute to someone even if he was asleep at the time (“John wants to lose weight, find a better job, and pay down his student loans”) and desires you might attribute to someone in the moment (“John wants to hang out at the bar for another two hours and drink more beer”).