Do Fast Food Restaurants Contribute to Obesity?

Featured in print Bulletin on Aging & Health

Over the past thirty years, the prevalence of obesity and obesity-related diseases in the U.S. has risen sharply. Since the early 1970s, the share of children age 6 to 19 classified as overweight has more than tripled, from 5 percent to 17 percent, while the share of adults classified as overweight or obese rose from half to two-thirds of the population. Over this same period, the number of fast food restaurants more than doubled. Exposes such as "Supersize Me" and "Fast Food Nation" as well as reports in the popular press have frequently suggested that fast food is at least partly to blame for the U.S.'s rising obesity rates.

Despite the popularity of this view, it has been difficult to empirically establish a causal link between fast food and obesity. The simple fact that fast food restaurants and obesity have both increased over time is insufficient proof of this link, as are studies that rely on differences in fast food consumption across individuals, since people who eat more fast food may be prone to other behaviors that affect obesity.

In The Effect of Fast Food Restaurants on Obesity (NBER Working Paper 14721), researchers Janet Currie, Stefano DellaVigna, Enrico Moretti, and Vikram Pathania undertake a careful study of the effect of fast food on obesity using the exact geographic location of fast food restaurants. Specifically, the authors examine whether proximity to a fast food restaurant affects the obesity rates of 3 million school children and the weight gain of over 1 million pregnant women.

The authors have several strategies to overcome the concern that children whose school is close to a fast food restaurant may be more prone to obesity for other reasons. First, the authors compute whether the fast food restaurant is within 0.10, 0.25, or 0.50 miles of the school and test whether there is a differential effect by distance. Given that there is a fast food restaurant in the general area, whether the restaurant happens to be very close to the school is arguably random and thus a good way to identify the effect of fast food access on obesity. Second, the authors control for a rich set of school and neighborhood characteristics in their analysis and allow these to vary over time. Finally, the authors estimate models relating changes in obesity to changes in access to fast food restaurants, relying on schools that gain or lose a nearby fast food restaurant during the sample period to identify the effect of fast food access on obesity. The authors estimate their models using data on all California 9th graders for the years 1999 and 2001-2007.

In their analysis of pregnant women, the empirical approach is similar - the authors include women for whom they observe at least two births in their sample, so that they can examine whether changes in fast food access between one pregnancy and the next are associated with changes in the probability of excessive weight gain. The authors use Vital Statistics data from 1989 to 2003 for Michigan, New Jersey, and Texas, the states for which they were able to obtain confidential data with mothers' names and addresses.

Turning to the results, the authors find that proximity to a fast food restaurant significantly increases the risk of obesity. For children, having a fast food restaurant within 0.10 miles of school increases the probability of obesity by 1.7 percentage points, or 5.2 percent. Interestingly, there is no significant effect of having a restaurant 0.25 or 0.50 miles from the school. The effects of fast food access are larger for girls.

Is the magnitude of this finding reasonable? To probe this, the authors estimate how many additional calories would be needed to push enough students over the overweight threshold (85th percentile of the historical BMI, or body mass index, distribution) to generate the estimated 5.2 percent increase in obesity. They estimate this to be 30-100 calories per day (per student). Given the large number of calories in typical items from a fast food restaurant, the magnitudes seem plausible.

For pregnant women, having a restaurant within 0.10 miles increases the probability of gaining over 20 kilos during pregnancy by 0.2 percentage points, or 2.5 percent. Unlike for children, the effects are still discernable at 0.25 and 0.50 miles. The effects are largest for African American mothers and mothers with a high school education or less. For pregnant women, the estimated increase in calories needed to generate this effect is much smaller, only 1-4 additional calories per day.

The authors conduct several other interesting analyses. First, they explore the effect of other types of restaurants and find that only fast food restaurants affect obesity. Second, they ask whether the future presence of a fast-food restaurant is associated with greater obesity today. If that were the case, it would suggest that fast food restaurants locate in areas where obesity is trending up, providing an alternative explanation for the paper's findings. However, the authors find that only current fast food locations matter.

Overall, the study suggests that proximity to fast food significantly increases the risk of obesity. One possible explanation is that proximity to fast food reduces the costs of fast food, principally travel costs. A second explanation is that easier access to fast food tempts consumers who have self-control problems. The fact that the effects are larger for teens and fall off more quickly with distance suggests that travel costs are more important for teens than for pregnant women.

The policy implications of these findings are potentially important. As the authors conclude, "policies restricting access to fast food near schools could have significant effects on obesity among school children, but similar policies restricting the availability of fast food in residential areas are unlikely to have large effects on adults."