Long-Term Effects of Income Received at Different Stages in the Life Cycle on Health

This paper examines the long-term effects of income received at different stages in the life cycle on health and health behaviors. The primary data comes from the longest running panel survey, The Panel Survey of Income Dynamics (PSID), which follows sample members and their descendants since 1968. I implement an instrumental variables approach to overcome concerns that income is likely correlated with unobserved factors which also affect health. The instruments leverage plausibly exogenous variation coming from changes made to the Earned Income Tax Credit over time at both the state and federal levels using data from the Current Population Survey and the NBER TAXSIM Software. I use multiple imputation to address concerns that income is missing in some years since the PSID became a biennial survey in 1997. The primary contribution of this paper is that I construct an instrument for family income which does not rely on one off shocks to family income but instead, exploits exogenous changes made to one of the largest safety net programs which has been around since 1975. The second major contribution of this paper lies in the fact that I observe the individuals in my sample from birth until the age at which the health outcomes are considered. Hence, I am able to account for income at all stages, unlike previous studies that consider only shocks encountered at a single point in time or over a short period of an individual’s life. I find that ordinary least squares estimates are biased toward zero and at best, income has only modest effects. This finding aligns with previous studies on the causal effect of income on health but contrasts findings from studies examining the long-term effects of in utero and early childhood environment. Results from a model which tries to separate the effects of permanent versus transitory income are forthcoming.