Are Restaurants Really Supersizing America?I. Theoretical FrameworkII. Data and Descriptive StatisticsIII. Restaurant Proximity and Body MassA. First-Stage RelationB. Reduced-Form RelationC. Instrumental Variables ResultsIV. Analysis of Alternative InterpretationsA. Does Highway Proximity Increase Restaurant Consumption?B. Can Residential Sorting across Towns Explain the Results?V. Why Don’t Restaurants Affect Obesity?VI. ConclusionREFERENCES152American Economic Journal: Applied Economics 3 (January 2011): 152–188http://www.aeaweb.org/articles.php?doi=10.1257/app.3.1.152“There’s nobody at McDonald’s shoving fries in your mouth.”—Bonnie Modugno, McDonald’s Chief Nutritionist1Obesity rates in the United States have been growing rapidly in recent years. Whereas 15 percent of Americans were obese in 1980 (defined as having a body mass index of at least 30), 34 percent were obese in 2004 (Center for Disease Control 2007). The time series of obesity rates in the United States, plotted in Figure 1 (solid line), reveals that the rate of increase over the past quarter-century has been substan-tially greater than during the preceding two decades. Medical research has linked obe-sity to diabetes, heart disease, stroke, and certain cancers. Treating these diseases is expensive. Health care spending attributed to obesity reached $78.5 billion in 1998 and continues to grow (Eric A. Finkelstein, Ian C. Fiebelkorn, and Guijing Wang 2003). Although obesity is a serious and growing problem, its causes are not well understood.1 Kim Severson, “The Obesity Crisis,” San Francisco Chronicle, March 7, 2004.* Anderson: Department of Agricultural and Resource Economics, University of California, Berkeley, 207 Giannini Hall, MC 3310, Berkeley, CA 94720-3310 (e-mail: [email protected]); Matsa: Kellogg School of Management, Northwestern University, 2001 Sheridan Road, Evanston, IL 60208 (e-mail: [email protected]). We thank Patricia Anderson, Joshua Angrist, David Card, Carlos Dobkin, Michael Greenstone, Alex Mas, Amalia Miller, Enrico Moretti, Kathryn Shaw, May Wang, and seminar participants at Columbia University, Georgetown University, IUPUI, Northwestern University, Stanford University, Tel Aviv University, UC Berkeley, the 2009 AEA Annual Meeting, and the NBER Summer Institute for helpful comments. Ellen Kersten, David Reynoso, Tammie Vu, and Kristie Wood provided excellent research assistance. We appre-ciate the cooperation of the Departments of Health of Arkansas, Colorado, Iowa, Kansas, Maine, Missouri, North Dakota, Nebraska, Oklahoma, Utah, and Vermont in providing the confidential data used in this study. Their cooperation does not imply endorsement of the conclusions of this paper.† To comment on this article in the online discussion forum, or to view additional materials, visit the article page at http://www.aeaweb.org/articles.php?doi=10.1257/app.3.1.152.Are Restaurants Really Supersizing America?†By Michael L. Anderson and David A. Matsa*While many researchers and policymakers infer from correlations between eating out and body weight that restaurants are a leading cause of obesity, a basic identification problem challenges these con-clusions. We exploit the placement of Interstate Highways in rural areas to obtain exogenous variation in the effective price of restau-rants and examine the impact on body mass. We find no causal link between restaurant consumption and obesity. Analysis of food-intake micro-data suggests that consumers offset calories from restaurant meals by eating less at other times. We conclude that regulation tar-geting restaurants is unlikely to reduce obesity but could decrease consumer welfare. (JEL I12, I18, L51, L66)ContentsAre Restaurants Really Supersizing America?† 152I. Theoretical Framework 156II. Data and Descriptive Statistics 159III. Restaurant Proximity and Body Mass 161A. First-Stage Relation 162B. Reduced-Form Relation 165C. Instrumental Variables Results 168IV. Analysis of Alternative Interpretations 171A. Does Highway Proximity Increase Restaurant Consumption? 171B. Can Residential Sorting across Towns Explain the Results? 176V. Why Don’t Restaurants Affect Obesity? 180VI. Conclusion 184References 186VoL. 3 No. 1 153ANdErSoN ANd MAtSA: ArE rEStAurANtS rEALLy SupErSIzINg AMErICA?One popular idea among public health advocates is that eating restaurant food causes obesity.2 Restaurant food has high caloric content, and the portion sizes served are relatively large (Lisa R. Young and Marion Nestle 2002). Concerned policymakers are turning to new regulations on restaurants in efforts to fight obesity. For example, in response to high obesity rates in low-income neighborhoods, the Los Angeles City Council unanimously approved a law on July 29, 2008 banning the opening of new fast-food restaurants in a 32-square mile area containing 500,000 residents (Tami Abdollah 2007; Molly Hennessy-Fiske and David Zahniser 2008). State legislators in Mississippi, which is reported to have the highest obesity rate in the world, recently proposed legislation that would prohibit any restaurant from serving obese customers (the bill died in committee) (Bobby Harrison 2008; Nanci Hellmich 2008). If large portions and effective marketing (presentation and pricing) lead people to eat more when they go to restaurants than when they eat at home, then these regulations may be effective. But it is not obvious that the empirical link between eating at restaurants and obesity is causal. If consumers’ lifestyles are increasingly conducive to excess energy 2 Other hypotheses include changes in food prices, increasingly sedentary lifestyles, and technological change in food production (Darius Lakdawalla and Tomas Philipson 2002; David M. Cutler, Edward L. Glaeser, and Jesse M. Shapiro 2003; Lakdawalla, Philipson, and Jay Bhattacharya 2005; Sara Bleich et al. 2008).1253575100202530255051015Restaurants (per thousand square miles)Obesity rate (percent)Obesity rate001960 1965 1970 1975 1980 1985 1990 1995 2000 2005Restaurant densityFigure 1. Obesity Rates and Restaurant Density, 1960–2004Notes: This figure plots the obesity rate and restaurant density in the United States from 1960 through 2004. The obe-sity rates are age-adjusted estimates for the percent of adults aged 20–74 with a body mass index greater than or equal to