Job Market Paper
Adolescent Depression and Adult Earnings: the Roles of Direct and Indirect Effects
Abstract: Youth with depression earn less as adults, but we understand little about the pathways of effects that lead to this gap. I use data from a nationally representative survey of youth to estimate the direct and indirect effects of adolescent depression on adult earnings with a mediation analysis framework. I use several identification strategies to address issues of omitted variables bias. I find that the direct effect of adolescent depression on earnings is close to zero and that indirect pathways are most important. More than half of the total effect of adolescent depression on earnings is mediated by educational attainment, while the rest of the effect is mediated by adult depression. Results highlight the large economic benefits that could still be achieved by better treating adolescent depression or targeting gaps in mediating variables.
Check out a working draft of this paper here!
Works in Progress
Estimating the Cost Savings of Tiered Care Coordination for Youth in the Juvenile Justice System: Evidence from North Carolina
With Jeremy Bray
Abstract: In North Carolina, youth involved with Juvenile Justice (JJ) make up less than 1% of the adolescent population and account for 20-60% of expenditures on high-level behavioral health services (BHS). State agencies are faced with the task of providing BHS to JJ-involved youth without taking on overwhelming costs. Between 2017 and 2021, North Carolina funded Tiered Care Coordination (TCC) in five counties to better engage and serve youth involved with JJ in need of BHS, including mental health and substance use services. This paper estimates the effect of TCC pilots on the cost of care for participating youth. We use a novel data set containing individual-level data from North Carolina Medicaid, Division of Mental Health (DMH), and JJ for youth in five pilot counties and six comparison counties. We use a difference-in-differences approach and a poisson fixed effects model to estimate the average treatment effect on the treated (ATT) of TCC pilots on the average monthly cost of care following referral. Our approach explicitly models the non-negative nature of expenditure data, controls for time-invariant individual-level heterogeneities, and controls for the time-varying impacts of age, race, gender, and referral source on the cost of care. While Poisson fixed effects models can only identify proportional treatment effects, we combine it with an estimator of unobserved heterogeneity to consistently estimate the ATE in levels (i.e., dollars). Preliminary results from Medicaid and DMH data find that referral to TCC saves $2,092 (95% CI $685-4,870) per month per youth in the year following referral, which far outweighs the cost of implementing TCC pilots. These savings are primarily driven by drops in expenditures on psychiatric inpatient, other inpatient, and BHS. We find no significant change in residential or other medical expenditures. This paper presents the first causal estimates of the cost savings of TCC and High-Fidelity Wraparound pilots and finds that they lead to significant cost savings. This is a valuable insight for policymakers around the country who are looking for ways to address the increasing costs of providing BHS to JJ-involved youth.
Adolescent Mental Health and Adult Labor Market Outcomes: A Scoping Review
With Jeremy Bray
Background and Aims: Mental illness and substance use are frequently comorbid among adolescents and present long-term economic consequences. The aims of this scoping review are to summarize the peer-reviewed literature on the association between adolescent mental illness, substance use, and their comorbidity and adult labor market outcomes, and discuss ways future research can better accomplish policy aims. Methods: A systematic scoping review was conducted in accordance with the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) methods. Studies were included if they use individual-level data to conduct empirical analyses on the effects of adolescent mental illness, substance use, or their comorbidities on adult labor market outcomes. Study heterogeneity prevented meta-analysis, so a narrative review is given. Results: Thirty-two studies from seven countries met the inclusion criteria. No study evaluated the relationship between comorbid adolescent mental illness and substance use and adult labor market outcomes. Instead, studies focused on either adolescent mental illness or substance use. Adolescent depression and attention deficit hyperactivity disorder (ADHD) are strongly associated with lower average earnings and rates of employment in adulthood. There is strong evidence that adolescent marijuana use is associated with poor labor market outcomes and conflicting evidence on the effects of alcohol use. The mechanisms driving these relationships are contested and studies seldom estimate mediated effects. Conclusions: There is a surprising lack of research on the association between adolescent comorbidities and adult labor market outcomes. Addressing this gap in the literature may improve estimates of long-term economic costs and lead to more accurate policy recommendations.
The Cost Savings of High Fidelity Wraparound in North Carolina
With Jeremy Bray
The Cost Savings of Eviction Mediation: Evidence from Guilford County
With the Center for Housing and Community Studies
Treatment Effects in a Poisson Fixed Effect DID Model: a Statistical Package
Developing an Interactive Energy Dashboard for UNCG
With the UNCG Green Fund