« Back to Results

Economics of Marijuana and Opioids

Paper Session

Sunday, Jan. 3, 2021 12:15 PM - 2:15 PM (EST)

Hosted By: Health Economics Research Organization
  • Chair: David Bradford, University of Georgia

Marijuana liberalization policies and perinatal health

Angelica Meinhofer
,
Cornell University
Allison Witman
,
University of North Carolina-Wilmington
Jesse Hinde
,
RTI International
Kosali Simon
,
Indiana University

Abstract

We studied the effect of medical marijuana laws (MML) and recreational marijuana laws (RML) on maternal substance use disorders and newborn health. We used restricted hospital discharge and birth certificate data from 2007-2017 and a multiperiod difference-in-differences estimator that exploited cross-state variation in effective dates of marijuana liberalization policies. We found that maternal marijuana use disorder increased by 36% in the first three years following RML implementation, with larger effects in states authorizing commercial sales of marijuana. This growth was accompanied by a 2.5% increase in the proportion of births with low gestational age but no other newborn outcomes were affected. MML implementation had no significant effect on maternal substance use disorders or newborn health and even fairly small effects could be ruled out. In absolute numbers, our findings implied modest or no adverse effects of marijuana liberalization policies on the array of perinatal health outcomes considered.

The Labor Supply Consequences of the Opioid Crisis

David Powell
,
RAND Corporation

Abstract

In this paper, I focus on the effects of broad changes in opioid access on labor outcomes for 1981-2018. I examine changes in labor outcomes over time based on state-level variation in initial conditions which exposed some states to the opioid crisis more than others. Complementary work shows that the introduction of OxyContin in 1996 explains a large share of the national rise in overdose death rates (Alpert et al., 2019). My focus on OxyContin is due to its pivotal role in the opioid crisis. Within just a few years after its introduction, OxyContin was the most abused opioid, outpacing all other types of oxycodone combined. Evidence from Purdue Pharma’s launch plan for OxyContin suggests that they targeted less promotional activity to “triplicate states,” states with especially stringent prescription drug monitoring programs in 1996. Triplicate states were less exposed to the introduction of OxyContin as measured by morphine equivalent doses, prescriptions, and detailing. These differences persist to today.

I implement a difference-in-differences design, comparing labor outcomes in non-triplicate states to outcomes in triplicate states both before and after the introduction of OxyContin, while relying primarily on event studies to transparently trace the conditional trajectory of these outcomes over time. The traditional event study approach, implemented regularly in difference-in-differences designs, is problematic, however. Under reasonable conditions which likely hold regularly in empirical applications, event study estimates may be biased even when the "parallel trends" assumption holds. I provide theoretical and empirical evidence of this bias.

Hassle Costs versus Information: How Do Prescription Drug Monitoring Programs Reduce Opioid Prescribing?

Mireille Jacobson
,
University of Southern California
Abby E. Alpert
,
University of Pennsylvania
Sarah Dykstra
,
University of Pennsylvania

Abstract

Past work demonstrates that mandated (or “must access”) prescription drug monitoring programs (PDMPs) decrease opioid prescribing, but provides limited evidence on mechanisms. PDMPs provide physicians with information but also introduce a hurdle to writing an opioid prescription (a hassle cost). We analyze Kentucky’s landmark PDMP to understand how mandates reduce prescriptions, disentangling the role of information versus hassle costs. Focusing on the emergency department (ED), we show that opioid prescriptions declined sharply for both patients who are opioid-naïve and those who are non-naïve, implying an important role for hassle costs. The decline was largest for patients who exhibit “doctor shopping” behaviors and effectively zero for opioid naïve patients presenting with opioid-appropriate conditions, suggesting that doctors also use the information in the PDMP. On net, although the PDMP mandate clearly affected prescribing through the information provided, the mandate’s hassle cost explains the majority of the decline in prescribing.

The Hazards of Unwinding the Prescription Opioid Epidemic: Implications for Child Abuse and Neglect

Mary Evans
,
Claremont McKenna College
Matthew Harris
,
University of Tennessee
Lawrence Kessler
,
University of Tennessee

Abstract

We examine how two interventions designed to curtail prescription opioid misuse, the reformulation of OxyContin and the implementation of must‐access prescription drug monitoring programs (PDMPs), affected child abuse and neglect. Our results suggest that counties with greater initial rates of prescription opioid usage experienced relatively larger increases in substantiated child abuse and neglect subsequent to OxyContin’s reformulation. We also find larger increases in child abuse and neglect after must‐access PDMP implementation in counties with higher pre‐intervention exposure to opioids. Our results uncover unintended consequences of reducing the supply of an addictive good without adequate support (or alternatives) for dependent users.
Discussant(s)
Jevay Grooms
,
Howard University
David Beheshti
,
University of Texas-San Antonio
Anne Burton
,
Cornell University
Christine Durrance
,
University of Wisconsin
JEL Classifications
  • I1 - Health