The impact of health on wages: Evidence from the British Household Panel Survey

Abstract. While income is generally considered an important determinant of health, little evidence has been offered on the reverse relationship, particularly for developed economies. This paper considers the effect of self-assessed general and psychological health on hourly wages using longitudinal data from six waves of the British Household Panel Survey. We employ single equation fixed effects and random effects instrumental variable estimators suggested by Hausman and Taylor (1981), Amemiya and MaCurdy (1986), and Breusch, Mizon and Schmidt (1989). Our results show that reduced psychological health reduces the hourly wage for males, while excellent self-assessed health increases the hourly wage for females. We also find the health variables to be positively correlated with the time-invariant individual effect. Further, we confirm the findings of previous work which suggested that the majority of the efficiency gains from the use of the instrumental variables estimators fall on the time-invariant endogenous variables, in our case academic attainment, and add further support to the hypothesis of a negative correlation between educational attainment and individual characteristics which affect wages.