Trends in Engagement with Opioid Use Disorder Treatment Among Medicaid Beneficiaries During the COVID-19 Pandemic (with Alyssa Tilhou, Laura Dague, Brendan Saloner, and Marguerite Burns)

Abstract: This paper examines trends in buprenorphine possession during the COVID-19 public health emergency in comparison with on-site services including urine drug testing and opioid treatment programs. We assess these outcomes using Wisconsin Medicaid administrative data containing detailed diagnostic information on all Medicaid-funded healthcare utilization. We find that among adult beneficiaries continuously enrolled from 12/2018 to 9/2020 with prior history of Opioid Use Disorder diagnosis, buprenorphine possession remained stable at the onset and for the first six months of the public health emergency. In contrast, completion of urine drug tests and receipt of opioid treatment program services declined with the start of the pandemic and recovered partially six months into the public health emergency. These findings suggest that patients in office-based settings were able to retain access to buprenorphine despite fewer on-site services like urine drug tests, while patients at opioid treatment programs experienced greater disruption in care.

Working Papers

Job Search, Wage Dispersion, and Universal Healthcare

Abstract: I develop and estimate a labor market search model with employer sponsored health insurance (ESHI), worker and firm heterogeneity, and wage dispersion arising from firm market power and job transition frictions. I estimate this model and use it to examine the impact of ESHI on the wage distribution, and to counterfactually predict the effect of removing health insurance from the labor market via the provision of free public insurance. I consider two alternative policies where universal healthcare is funded external to the model, or via a new corporate tax on revenue. In the first, I find a considerable degree of passthrough to wages, roughly 76%, of what is effectively a subsidy to firms that were previously paying insurance premiums. However, it takes almost ten years for these wage gains to fully accrue to workers. In the second policy, average wages are virtually unaffected, but in addition to providing insurance coverage to all individuals, the tax acts as a transfer of wealth from the highest to the lowest earners, and these distributional effects are realized much more rapidly. In both counterfactual regimes, wage inequality decreases by a little more than 2 percentage points, but unemployment, job mobility, and joint productivity are not significantly impacted by universal healthcare.

Side Hustles: Non-Wage Job Characteristics and Dual Job Holding

Abstract: This paper develops a simple structural model of the choice to work a second job. I examine the effects of non-wage job characteristics on this decision making in order to determine the extent to which individuals hold multiple jobs as a source of enjoyment, versus as a means of overcoming hours constraints in the primary job. To fit this model, I estimate a distribution of enjoyment parameters for individuals holding more than one job, and find that on average, individuals dislike their secondary jobs about 13.5% more than their primary jobs, but roughly 35% of these individuals enjoy their secondary jobs. Though this supports findings of hours constraints as the primary motivator of dual job holding, these results provide a framework for further study of the substantial portion of dual job holders that do prefer their second jobs.

Employment Shocks and Mental Healthcare Utilization During the Covid-19 Pandemic

Abstract: I examine the impact of the Covid-19 pandemic and its related income and employment shocks on the use of mental health resources in the Wisconsin Medicaid population. Using administrative Medicaid claims data from the Wisconsin Department of Health Services, I find a reduction in mental healthcare utilization during the public health emergency (PHE) that is similar to but significantly smaller than observed trends in overall outpatient visits. However, making use of the PHE declaration as an exogenous shock to employment, I find that this decline was 0.45 percentage points smaller for individuals who experienced a decrease in wages of 50% or more. This is largely driven by the subset of individuals with a pre-existing mental health diagnosis, as for this group I find the effect of an employment shock was a 1.3 percentage point smaller drop in mental health visit probability, a 4.35% difference relative to individuals who did not experience a reduction in wages. This suggests that individuals with pre-existing mental health conditions were more likely to continue care during the initial months of the pandemic if they were subjected to some form of job displacement.