The relation between obesity and fast food

            Rising prevalence of obesity is often referred to as an epidemic, although it has apparently leveled off since 1999. Obesity is defined as a body mass index (BMI) of 30 or higher. Obesity has been associated with many health problems. Many causes for weight gain have been suggested. These include: increased consumption of sugar-sweetened beverages, reduction in real prices of food, urban sprawl reduced cigarette smoking, less time spent preparing healthy meals in the home, eating more food from restaurants, rising numbers of food stamp recipients, technological change leading to a more sedentary lifestyle, increased availability of restaurants, a growing lack of grocery stores selling healthy foods, and agricultural policies that encourage production of "excess calories".

            This paper focuses on the hypothesized link between obesity and fast food using panel data on all states. The paper attempts to determine whether variation in obesity prevalence across states, and over time, is explained by the associated variation in fast food restaurants. Data on numbers of fast food restaurants and their employees is available from 2001-2009 within the Quarterly Census of Employment and Wages (NAICS 722211 Limited-service restaurants). Other variables will be collected that control for other possible factors beside fast food that influence obesity prevalence. Such variables are suggested from the above-discussed studies that suggested demand and supply factors. These include age, gender, ethnic, education, income, exercise and diet characteristics at the level of states and over 2001-2009.

            It is important to deal with a possible endogeneity problem. The problem is that, while states with many fast food restaurants may also experience relatively high obesity prevalence, it is possible that this connection is simply that fast food restaurants are more frequently located in areas that have high demand for fast food and thus variation in fast food presence is explained by heavier people enjoying fast food more rather than the restaurants themselves causing obesity. In effect, if there is a link between obesity and fast food, it might simply stem from heavier people enjoying fast food more than other people. If true, then restrictions on fast food, or higher taxation, is unlikely to curb obesity. My examination will thus carefully sort out this endogeneity issue by using two stage least squares to estimate panel regressions. Briefly, this technique creates an instrumental variable that is correlated with fast food presence but by itself is not correlated with obesity prevalence. An appropriate instrument will be determined that deals with this endogeneity issue.