An econometric analysis of self-assessed health: what does it mean and what is it hiding?

Self-assessed general health (SAH) is the most frequently employed health measure in economics research; however, it is poorly understood. In this paper we answer two questions: (i) what components of health does SAH measure? and (ii) does the use of SAH conceal important health effects? To answer the first question, we use a detailed health instrument and linear, logit and dynamic fixed-effects models to estimate the drivers of SAH. To answer the second question, we estimate the effects of income on disaggregated health measures using instrumental-variables fixed-effects models. We find that some health components – especially vitality – are very important to an individual when they assess their health, while others are inconsequential. We also find that this fact is partially responsible for why econometrically-sound studies find weak socioeconomic gradients in SAH. Regression results show that the effects of income on SAH are near-zero, even though income strongly affects several health components.