Calories on Menus: Nationwide Experiment Into Human Behavior

Sabrina Tavernise
NOV. 26, 2014
The New York Times
Now it’s official. Starting next November, menus in many places where Americans eat — like chain restaurants and some movie theaters, convenience stores and amusement parks — will have to list calories.
Consumer health advocates were jubilant when the Food and Drug Administration announced the new policy on Tuesday. Many had fought for the rule for more than a decade, believing it would be a major weapon in the fight against obesity.
But will it?
The evidence on whether menu labeling works — either to move the national needle on obesity, or to reduce the number of calories an individual consumes after looking at a menu — is pretty skimpy, in part because the practice hasn’t been around that long.
In the few places where menu labeling exists, like New York and Philadelphia, most studies have observed a few thousand people over just a few weeks and months — too small a group and too short a time to detect the subtle changes that economists expect the policy will prompt.
Brian Elbel, associate professor of population health at New York University’s School of Medicine, has spent weeks outside fast food restaurants talking to customers and collecting their receipts.
The findings have been uninspiring so far. In a study he did in 2008 in New York City, only slightly more than half of consumers even saw the posted calories, and of those, a little over a quarter (around 15 percent of the total) said the information changed what they ordered. He conducted a larger study in 2010 in Philadelphia after that city started requiring chain restaurants to post calories, and the results were similar.
When he delved deeper, he found that those who changed their ordering behavior tended to be the more educated consumers — in other words, not the target population. Americans with more education tend to be less likely to be obese than those with less, though there are exceptions.
The researchers could not tell whether this subgroup ordered less because of the calorie posting, or because they would have behaved differently anyway. Either way, their change was not enough to make the average number of calories everybody was buying go down.
“It doesn’t appear to be changing what people order for fast food at a population level,” Mr. Elbel said.
Until now, most studies exploring menu labeling have been of fast food restaurants, Mr. Elbel said, and are missing the growing slice of the population going to sit-down restaurants, such as Applebee’s and the Cheesecake Factory, where meals are believed to be even more caloric. The new policy has excited researchers because it extends their laboratory to the entire nation.
One of the largest studies to date, involving hundreds of millions of transactions at Starbucks, found a small but real decline in the number of calories consumers bought.
Starbucks gave a group of Stanford University researchers access to transaction data in New York, Boston and Philadelphia from January 2008 to February 2009. The company began posting calories in its New York stores in April 2008, and researchers looked at consumer behavior there before and after posting and compared it with the other cities where calories had not been posted. Average calories per transaction from food fell by 14 percent in New York, a calorie reduction that lasted for at least 10 months after the calorie counts were first posted. Average calories per transaction for drinks did not change much.
Bryan Bollinger, one of the authors, who is now an assistant professor of marketing at the Fuqua School of Business at Duke University, said his study’s large size gave researchers confidence in their conclusions. He says that when small studies find no effect, it is impossible to tell whether there really wasn’t any change, or if the sample was simply too small to detect it.
“We found strong evidence that calorie labeling does change consumer behavior,” he said.
But George Loewenstein, an economist at Carnegie Mellon University who has researched how behaviors affect obesity, was skeptical that menu labeling would have much effect on consumer habits.
“There are very few cases where social scientists have documented that giving people information has changed their behavior very much,” he said. “Changing prices and changing convenience have big impact. Providing information doesn’t.”
He gave a few examples: nutrition information on packaged food; warning labels on drugs; terrorism alerts; airplane safety cards; and the government’s pyramid symbol (recently replaced by a plate) that is supposed to show people, in simple terms, how to eat a balanced diet.
“They made a big deal about the plate and how it was so superior to the pyramid, but I just wonder if there is a single person in the United States whose behavior was changed,” he said.
He pointed out that the upper-middle-class people making menu-labeling policy often had little insight into how the lower-middle-class people whom the policy was aimed at would use the information. Just getting consumers to understand what the numbers mean is hard. Once they do, they might use them to maximize their calories per dollar, instead of reducing them.
“The people who most need the information don’t know how to use it,” he said.
The main benefit of the new menu labeling policy, Mr. Loewenstein said, would probably not be in changing the behavior of consumers, but in changing the behavior of companies that want to sell them things.
“To the extent that calorie labeling is beneficial, by far the most likely mechanism is that restaurants end up changing their offerings, for example reducing calories in the food they offer,” he said.
But health experts cautioned against writing off menu labeling as a policy tool.
Marion Nestle, a professor in the department of nutrition, food studies and public health at New York University, calls menu labeling “nutrition education for the public big time,” and said it would take a while for consumers to learn how to use the information.
“When it first starts, people will be shocked, like we were in New York,” she said. “Even I couldn’t believe it. A smoothie had 1,000 calories in it. That’s half the calories someone needs for a whole day.”
Professor Elbel said it was unrealistic to expect that any single policy would fix the nation’s obesity problem. An easing in the obesity rate will most likely come from a combination of different actions applied systematically over time.
“No single policy is going to do it,” he said.