Health and the Wage Rate: Cause, Effect, Both, or Neither? New Evidence on an Old Question

In this paper we investigate two-way causality between health and the hourly wage. We employ insights from the human capital and compensating wage differential models, a panel formed from the National Longitudinal Survey of Youth 1997, and dynamic panel estimation methods (for example, Arellano 2003, Baltagi 2008) in this investigation. We adopt plausible specifications in which a change in health induces a change in the wage with a lag and in which a change in the wage induces a change in health, also with a lag. We uncover a causal relationship between two of five measures of health and the wage in which a reduction in health leads to an increase in the wage rate in a panel of U.S. young adults who had completed their formal schooling by 2006 and were continuously employed from that year through 2011. There is no evidence of a causal relationship running from the wage rate to health in this panel. The former result highlights the multidimensional nature of health. It is consistent with an extension of the compensating wage differential model in which a large amount of effort in one period is required to obtain promotions and the wage increases that accompany them in subsequent periods. That effort may cause reductions in health and to a negative effect of health in the previous period on the current period wage. In this framework, employees have imperfect information about the effort requirements of a particular job when they are hired, and employers have imperfect information about the amount of effort new hirers are willing to exert. The result also is consistent with a model in which investments in career advancement compete with investments in health for time—the ultimate scarce resource. The lack of a causal effect of the wage on health may suggest that forces that go in opposite directions in the human capital and compensating wage differential models offset each other.