College Access and Adult Health

We investigate the relationship between college enrollment and health behaviors and outcomes later in life. Though a large prior literature attempts to isolate the causal effect of education on health, most studies use instruments that affect schooling behavior in childhood or adolescence, i.e. before the college enrollment decision. Some recent studies examine the effects of college on health, but ours is the first to use tuition prices across geography (states) and time as a source of plausibly exogenous variation in college outcomes that can be leveraged to isolate college enrollment and duration effects on health. 

Our contributions are important for at least three reasons. First, the instrument, though arguably less valid than some others, represents a policy-relevant mechanism for affecting education levels and, subsequently, health. This contrasts with some (natural) experiments, which use variation in education that is difficult to manipulate in practice, e.g. calendar-based primary and secondary school age eligibility thresholds, thus harming the external validity of their estimates. Second, our instrument likely has its strongest effect on a different population than are induced to change behavior in the face of other instruments, e.g. increases in mandatory education, meaning we may recover a different and separately informative local average treatment effect. Third, U.S. public higher education institutions, which enroll the vast majority of American young adults pursuing a college degree, have implemented dramatic tuition increases over the past decade following a major recession and significant state budget changes, so it is important to understand the effects of those changes on a variety of outcomes for young adults, including health. 

This paper marks an early attempt to understand the potentially adverse consequences of pricing some students out of higher education. We use 30 years of data from the Behavioral Risk Factor Surveillance System to match individuals to tuition prevailing in their current state of residence at the time they were 18 years old. We then estimate the effect of this tuition on 1) college attendance and completion and 2) various health behaviors and outcomes (including smoking, drinking, and BMI) controlling for state and year fixed effects, state-specific trends, and a variety of other state and individual-specific factors. Our estimates of the college enrollment price elasticity are in line with previous estimates. Furthermore, we estimate advantageous effects of college education on health behaviors and outcomes, meaning that increases in tuition not only reduce college enrollment but harm health.

Our findings inform a wide variety of ongoing policy debates about higher education, which is in a period of substantial change. Recently, prominent policy proposals included redirecting more students away from traditional degree programs to shorter, more job-oriented degrees or certifications; adjusting federal or state education financing support by changing student loan programs or directly adjusting tuition levels; or even increasing the age at which public education is offered universally, e.g. at the community college level. To the extent that the policies change the enrollment and duration of college education, our results inform these important discussions.