*Names in bold indicate Presenter
In a previous study, we addressed this potential source of bias by estimating the effects of state-level job losses on eighth grade math and reading scores. State-level job losses, as opposed to parental job losses, are an arguably exogenous source of variation in job loss with respect to individual youths within a community. Community-level analyses also capture spillover effects on youths whose parents maintain employment but who may be affected by friends’ or neighbors’ job losses or by concern about their own future employment prospects. In that work, we found that state job losses significantly decreased average math scores. This paper extends our previous work by applying this framework to study youths’ risk-taking behaviors. While it is plausible that youth health behaviors, like their test scores, are worsened by downturns, previous research (e.g. Ruhm 2000, Miller et al. 2009) has found that negative economic shocks actually improve adult health behaviors. It is unknown whether the health behaviors of youths, whose self-control and future orientation are still developing, respond in the same way as adults’ behaviors to economic stress.
The data for this analysis are derived from two main sources. Job loss data are from the Bureau of Labor Statistics’ (BLS) Mass Layoff Statistics, which report the number of workers in a year who are affected by mass layoffs that last longer than thirty days. Observations are for each state and the District of Columbia from 1995 to 2009. For the purposes of our analysis, we express job losses as a percentage of the working age population (age 25 – 64) in each state.
Youth risk behavior data are from the Centers for Disease Control’s Youth Risk Behavior Surveillance System (YRBSS), which has administered surveys to a nationally representative sample of high school students in roughly two year intervals since 1991. Analysis are focused on measures of controlled substance use, sexual activity and contraceptive use. We use a “differences-in-differences” regression framework, regressing our measure of job loss on student behaviors and adjusting for state and year fixed effects.
Our results indicate that job losses significantly reduce the use of cigarettes, alcohol, and marijuana among high school students. Losses significantly increase the probability that a teen is sexually active, but significantly decrease the average number of sex partners and increase the probability that a teen used contraception at last sexual intercourse. Results are consistent with previous studies finding improved adult health behaviors during downturns. However, the findings leave open questions about why students’ academic performance worsens at the same time that their health behaviors improve.