We analyze the effect of obesity on the incidence of hypertension, hyperlipidemia and diabetes in USA using a health production theoretical framework along with a bivariate flexible semi-parametric recursive copula model that account for endogeneity. In this approach, the effects of control variables are flexibly determined using additive predictors that allow for a variety of effects. Our findings suggest that there exist a positive and significant effect of obesity on the prevalence of all chronic diseases examined. In particular, after endogeneity is accounted for, the probability of having hypertension, hyperlipidemia and diabetes for obese individuals are, respectively, 35%, 28% and 11% higher than those under the obesity threshold. These findings suggest that lowering obesity rates could lead to significant reductions in the morbidity and mortality associated with these diseases.