The Effect of Income on Health: Evidence from New Health Measures in the NLSY

Abstract

While it is well known that income and health are positively associated, the critical question of whether this relationship is causal remains open. Income may cause better health, but causality could also run from health to income, or both income and health could be correlated with one or more unobserved variables. This paper attempts to overcome these problems by systematically exploiting features of the National Longitudinal Survey of Youth, which recently added several important health measures. To account for reverse-causality, I reason that health effects income mainly by limiting labor force activities, and focus on a subpopulation that has never reported a health related work limitation. To account for omitted variables, I estimate sibling fixed-effects models to control for childhood conditions and genetic background, and include direct controls for discount rates, risk aversion and cognitive ability. I find that income has a significant causal effect on physical health and obesity, but not on mental health, smoking or heavy drinking. My preferred models predict that an individual with a permanent income of $70,000 is approximately 10% less likely to be obese than if they had a permanent income of $30,000, and reports physical health that is .25 standard deviations more favorable.