This paper presents a life cycle model for the demand for health, and derives empirical specifications that distinguish between permanent and transitory wage responses. Using panel data, we estimate dynamic health and health input demand equations. We find evidence of negative transitory wage effects, and positive permanent effects. Estimation results based on our life cycle framework lead to very different conclusions than those based on static cross section analyses that are common in the literature. The analysis emphasises the importance to analyse health related behaviour in a dynamic life cycle context.