| During the past three decades, the United States and most of the rest of the developed world have experienced a rapid and sustained rise in the obesity rate. This trend has stimulated a considerable amount of research by economists and other social scientists dealing with its causes and with policies to combat it. To date, the focus has been on obesity defined by a body mass index (BMI, weight in kilograms divided by height in meters squared) greater than or equal to 30. This measure has been criticized because it fails to distinguish body fat from lean body mass. It is the former that is responsible for the detrimental health effects of obesity. Therefore, in my dissertation I introduce the percentage body fat (PBF, the ratio of body fat to total weight multiplied by 100) and an obesity indicator based on PBF as alternative measures of body composition. I generate equations by gender and race to predict these measures from height, weight, and age in the Third National Health and Nutrition Examination Survey and use the estimated coefficients to obtain PBF and obesity based on PBF in the Behavioral Risk Factor Surveillance System for the period from 1984 through 2009. I then examine the effects of socioeconomic characteristics and state-level measures pertaining the per capita number of restaurants, the prices of a meal in fast-food and full-service restaurants, the price of food consumed at home, the price of cigarettes, and clean indoor air laws on BMI, PBF, BMI-defined obesity, and PBF-defined obesity. My results suggest that most of the determinants at issue have similar qualitative and quantitative effects on the outcomes at issue. Finally, I assume that PBF-defined obesity correctly identifies obese and non-obese individuals, but BMI-defined obesity results in error. I use these assumptions to estimate the parameters of a binary choice model by nonlinear least squares. My results show that this procedure successfully corrects the downward bias in the marginal effects of a probit model for BMI-based obesity. |
