Difference-in-Differences with Variation in Treatment Timing

Goodman-Bacon, Andrew

doi:10.3386/w25018

Cited by 648 publications

(438 citation statements)

References 39 publications

(56 reference statements)

Supporting

Mentioning

411

Contrasting

Unclassified

Order By: Relevance

“…Models with variation in treatment timing are known to exhibit bias if the treatment effect is heterogeneous over time. 22 We test for temporal treatment heterogeneity bias by performing a Bacon decomposition. 23 Results from this analysis indicate that this source of bias is moderate (weight of Later Treatment vs.…”

Section: Estimationmentioning

confidence: 99%

The Effect of Stay-at-Home Orders on COVID-19 Cases and Fatalities in the United States

Fowler

Obradovich

et al. 2020

Preprint

117

105

View full text Add to dashboard Cite

Background In March and April 2020, public health authorities in the United States acted to mitigate transmission of and hospitalizations from the severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2), which causes coronavirus disease 2019 . These actions were not coordinated at the national level, which raises the question of what might have happened if they were. It also creates an opportunity to use spatial and temporal variation to measure their effect with greater accuracy. MethodsWe combine publicly available data sources on the timing of stay-at-home orders and daily confirmed COVID-19 cases at the county level in the United States ( N = 124,027). We then derive from the classic SIR model a two-way fixed-effects model and apply it to the data with controls for unmeasured differences between counties and over time. This enables us to estimate the effect of stay-at-home orders while accounting for local variation in factors like health systems and demographics, and temporal variation in national mitigation actions, access to tests, or exposure to media reports that could influence the course of the disease.Findings Mean county-level daily growth in COVID-19 infections peaked at 17.2% just before stay-at-home orders were issued. Two way fixed-effects regression estimates suggest that orders were associated with a 3.9 percentage point (95% CI 1.2 to 6.6) reduction in the growth rate after one week and a 6.9 percentage point (2.4 to 11.5) reduction after two weeks. By day 27 the reduction (22.6 percentage points, 14.8 to 30.5) had surpassed the growth at the peak, indicating that growth had turned negative and the number of new daily infections was beginning to decline. A hypothetical national stay-at-home order issued on March 13, 2020 when a national emergency was declared might have reduced cumulative infections by 63.3%, and might have helped to reverse exponential growth in the disease by April 10.Interpretation Although stay-at-home orders impose great costs to society, delayed responses and piecemeal application of these orders generate similar costs without obtaining the full potential benefits suggested by this analysis. The results here suggest that a coordinated 1 . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) : medRxiv preprint nationwide stay-at-home order might have reduced by hundreds of thousands the current number of infections and by tens of thousands the total number of deaths from COVID-19. Future efforts in the United States and elsewhere to control pandemics should coordinate stay-at-home orders at the national level, especially for diseases for which local spread has already occurred and testing availability is delayed. Since stay-at-home orders reduce infection growth rates, early implementation when infection counts are still low would be most beneficial.Funding None.

show abstract

Section: Estimationmentioning

confidence: 99%

The Effect of Stay-at-Home Orders on COVID-19 Cases and Fatalities in the United States

Fowler

Obradovich

et al. 2020

Preprint

117

105

View full text Add to dashboard Cite

show abstract

“…6 Second, linear state-trends do not fully remove bias from nonparallel trends in the situation of time-varying treatment effects. Directly controlling for statespecific trends in only the pre-period is a straightforward strategy to address both of these concerns (Goodman-Bacon, 2018). 7 Economic theory provides clear predictions about the unintended consequences from the zero cost-sharing mandate.…”

Section: Introductionmentioning

confidence: 99%

“…Directly controlling for statespecific trends in only the pre-period is a straightforward strategy to address both of these concerns (Goodman-Bacon, 2018). I control for state-specific pre-trends rather than linear state-trends for two reasons.…”

mentioning

confidence: 99%

Unintended consequences of health insurance: Affordable Care Act's free contraception mandate and risky sex

Willage

2019

Health Economics

View full text Add to dashboard Cite

Health insurance is a primary driver of rising medical expenditures. Economic theory suggests that insurance induces an increase in risky behaviors, but previous empirical evidence is mixed. I use a mandate in the Affordable Care Act in which contraceptives were covered at zero cost to consumers to test for unintended effects of insurance on risky sex. Leveraging mandated zero cost-sharing for contraception and pre-policy insured rates as a measure of treatment intensity, I provide evidence that this 2012 policy reduced fertility but caused unintended consequences: a decline in condom use and a subsequent increase in sexually transmitted infections (STIs). I discuss shortcomings of controlling for nonparallel pre-trends using state-trends, and I suggest an alternative to control for pre-trends directly in the context of dose-response difference-indifferences. Finally, estimates based on the 2010 dependent coverage mandate indicate health insurance provides an overall net positive effect on insurance and STI prevention.

show abstract

“…11 State specific trends do not appear to be a significant source of bias (Goodman-Bacon, 2018;Wolfers, 2006). When excluding the state specific trends controls, the estimates are somewhat larger but still not significant with the exception of the over 45 age group.T ABLE 1 Summary statistics by nurse practitioner independent schedule II prescribing authority status allowed NPs to prescribe opioids independently and have operational mandatory PDMPs experience a significant increase in treatment admissions for opioid misuse mainly driven by significant changes among the 18-24 and 25-44 years old population.…”

Section: Resultsmentioning

confidence: 90%