Estimation of a dynamic model of weight

The ongoing debate about the economic causes of obesity has focused on the changing relative prices of diet and exercise. This paper uses a model that explicitly includes time and spatially varying community-level urbanicity and price measures as instruments to obtain estimates of the effects of diet, physical activity, drinking, and smoking on weight. The instruments control for bias due to unobservables that may be common determinants of weight and these explanatory variables. We apply a dynamic panel system GMM estimation model to longitudinal (1991-2006) data from China to model weight and find that among adult men in China, about 5.4% of weight gain was due to declines in physical activity and 2.8-3.1% was due to dietary changes over this period. Combined, changes in physical activity and diet only explain around 8% of the short run gain in weight and about 11% of the long-run gain in weight among Chinese men.