The Effect of Participating in Physical Activity on Health for Americans: Evidence from 1990-2010

Abstract

Using cross-sectional drawn data from the Behavioral Risk Factor Surveillance System (BRFSS) for the years 1990-2010 in the United States, I examine the causal impact of participating in physical activity on body mass index. I employ an Instrumental Variable framework as a solution to the endogeneity of physical activity, in which I employ the respondents’ answers to the survey question: “Are you using physical activity or exercise to lose weight or maintain your current weight?”, as the instrument. Consistent with the findings of previous studies, the OLS results suggest that participation in any level of physical activity is negatively associated with the body mass indices of individuals. Holding all other factors constant, participating in physical activity is associated with a 5.0% decrease an individuals’ body mass index, while engaging in irregular, regular, and active levels of physical activity is associated with decrease of 3.6%, 5.5%, and 8.7% respectively. Therefore, their statistical effects increase in magnitude as the physical activity levels increase. The attributes of being a female or a smoker are associated with decreases in BMI, while age is positively associated with BMI. Unfortunately, I was unable to obtain causal effects due to employing a weak instrument in the instrumental variables regression. Finally, I conclude that further research is needed to find additional strong and relevant instrument and to acquire a more consistent dataset with well-defined variables in order to obtain more reliable and complete results.