The income-health gradient: Evidence from self-reported health and biomarkers using longitudinal data on income

This paper adds to the literature on the income-health gradient by exploring the association of short- and long-term income with a wide set of self-reported health measures and objective nurse-administered and blood-based biomarkers as well as employing estimation techniques that allow for analysis 'beyond the mean' and accounting for unobserved heterogeneity. The income-health gradients are greater in magnitude in case of long-run rather than cross- sectional income measures. Unconditional quantile regressions reveal that the differences between the long-run and the short-run income gradients are more evident towards the right tails of the distributions, where both higher risk of illnesses and steeper income gradients are observed. A two-step estimator, involving a fixed-effects income model at the first stage, shows that the individual-specific selection effects have a systematic impact in the long-run income gradients in self-reported health but not in biomarkers, highlighting the importance of reporting error in self-reported health.