THE INFLUENCE OF PHYSICAL ACTIVITY ON HEALTH

The primary goal of this study is to enhance the knowledge about the causal impact of physical activity on health. In fact, health can be quantified by numerous variables. Our specification for being healthy is having a healthy Body Mass Index (BMI). Although the positive association between physical activity and health outcomes is well acknowledged, the causality between them is quite doubtful. Namely, it is plausible that an extreme BMI value, i.e. obesity, harms the intensity or/and frequency of exercising. Further, it is likely that unobserved genetic factors, family background and risk aversion correlate with both BMI and physical activity. These elements illustrate the potential endogeneity of physical activity. In particular, we distinguish three physical activity categories: strenuous, moderate - and low intensity. Then, for our empirical analysis, we utilize the Longitudinal Internet Studies for the Social Sciences (LISS) panel, which consists of 4500 Dutch households. Further, we exploit both Ordinary Least Squares (OLS) – and fixed effect (FE) estimations to examine whether time-invariant unobservables lead to biased estimations. Also, we add lagged physical activities to control for the reverse impact from BMI on physical activity. Our results show that an OLS approach overestimates the beneficial impact of physical activity on BMI. Moreover, the overall FE estimation indicates that an additional day of strenuous physical activity throughout the week significantly lowers BMI with 0.029. In addition, it is surprising that the significant impact from low intensity physical activities on BMI disappears when we separate the regressions by gender. The causality of our findings is however debatable.