Must-Access Prescription Drug Monitoring Program Mandates and State Employment: A Staggered Difference-in-Differences Analysis

Abstract

This paper estimates the causal effect of state-level must-access Prescription Drug Monitoring Program (PDMP) mandates on aggregate employment outcomes using a staggered difference-in-differences design. Between 2013 and 2021 (full-exposure years), 46 U.S.\ states adopted laws requiring prescribers to query their state's PDMP before issuing controlled substance prescriptions; all 46 of which are included in the estimation sample. I exploit this staggered rollout using the estimator with not-yet-treated states as the primary comparison group and Bureau of Labor Statistics Local Area Unemployment Statistics data spanning 50 states from 2007–2023. The results constitute an informative null: the estimated average treatment effect on the treated (ATT) for log employment is $+0.0036$ (SE $= 0.0079$, $p = 0.647$), and the effect on the unemployment rate is $-0.242$ percentage points (SE $= 0.293$, $p = 0.407$). The not-yet-treated comparison group is preferred because only four states (Kansas, Missouri, Nebraska, South Dakota) never adopted universal must-access mandates, and this thin control group produces spurious pre-trend violations in event-study analysis. Using the never-treated control group as a sensitivity check, the ATT for log employment is $+0.0100$ (SE $= 0.0078$, $p = 0.203$)—also statistically insignificant. Group-level estimates reveal heterogeneity across adoption cohorts, but the overall pattern across all specifications is consistent with negligible aggregate employment effects. These findings suggest that must-access PDMP mandates—while effective at reducing opioid prescribing—do not produce detectable changes in state-level employment aggregates over the medium run.