Replying to @realjoshkoenig @robkhenderson
Robust to many different synthetic controls but these are the main weights.
The results are pretty comparable (29 and 23.8% of OD 2021-2023 are attributed to regime change). And while both OR and WA decriminalized, they did so in very different ways.
1
17
12/ Bottom line: across Oregon and Washington, and across many robustness checks, decriminalization was followed by sustained increases in overdose mortality relative to matched counterfactuals.
Future work will look more directly at mechanisms, deflection, and related outcomes.
11/ Across these measures, the structural-break evidence is similar when allowing for one break. We also show results allowing for multiple breaks.
And importantly, the SCM often places weight on late-fentanyl-arrival states anyway (weights in blue rings).
10/ We examine some alternative measures, including:
a. Fentanyl reports per-capita
b. Fentanyl share of all opioid seizures
The second measure is useful because it may better capture the transition from heroin to fentanyl.
9/ Assessing fentanyl controls:
One issue is that fentanyl’s share of NFLIS drug reports can move for two reasons:
More fentanyl reports
Fewer non-fentanyl reports after enforcement drops
So the fentanyl-share measure may itself be affected by decriminalization.
6/ Importantly, we find that the overdose gap widens over time (particularly 2022 and 2023). This helps explain results from earlier studies that had limited post-periods.
5/ The main result: Relative to synthetic controls, overdose mortality increased by about 0.806 deaths per 100,000 per month in Oregon and 0.732 in Washington.
That implies roughly 1,186 excess deaths in Oregon and 1,895 in Washington from 2021–2023.
4/ The idea with the SCM is to build counterfactuals for OR and WA using a weighted combination of other states. Appendix Figure 3 shows the main donor weights.
The result is not unique to SCM; our results are robust to TWFE and a slew of alternative estimators.
2/ We first show/confirm that these policies did indeed reduce arrests. We find that WA was decriminalized in practice until the passage of SB 5536.
We then look to what happened next with regard to overdose mortality.
Going to be riding this high all the way through to finals week.
(re: Assigned reading/replication of the @JevayG and @berto_econ AEA P&P "Examining Medicaid Expansion and the Treatment of Substance Use Disorders" for my intro to econometrics course).
87
The anxiety wondering how these slides would go over for my discussion at the #appam2025 Drugs and Crime Student Session? Not great.
The response? Amazing.
1
2
106
Starting a new blog series that explored substance use disorder treatment center ownership. Made some pretty cool visualizations in my opinion, please read!
studhall.github.io/academics…
Attaching a nice example of the .gifs you can find!





















