Thread by @NinaBuchmann6

WP with Doug Bernheim, @zdgroff and @seb_otero: The Effects of Large Group Meetings on the Spread of COVID-19: The Case of Trump Rallies

Matching and regression analysis suggest that the first 18 Trump rallies likely resulted in > 30,000 cases & 700 deaths. (1/N)

#EconTwitter

Here's the link to the paper: drive.google.com/file/d/1904CiW… (2/N) #EconTwitter #TrumpRallies #COVID19

In these control counties, we estimate the relationship between COVID-19 in the weeks after the rally and various conditions, such as pre-existing COVID-19 prevalence and pandemic related restrictions. (3/N)

We use the estimated model to predict post-event confirmed COVID-19 cases for the rally county. The difference between the actual and predicted incidence is an estimate of the treatment effect. We then calculate the average treatment effect across event counties. (4/N)

For most matching procedures, the average treatment effect implies that the 18 rallies subsequently increased confirmed cases by more than 250 per 100,000 residents. Extrapolating to the population, the rallies resulted in more than 30,000 incremental COVID-19 cases. (5/N)

Applying county-specific post-event death rates, we conclude that the rallies likely led to more than 700 deaths (not necessarily among attendees). (6/N)

To test for pre-trends and whether our method systematically mispredicts for event counties, we perform the same analysis for "placebo events" 10 weeks before the actual events. The pseudo-treatment effects are small, slightly negative, and statistically insignificant. (7/N)

Paper now on SSRN: papers.ssrn.com/sol3/papers.cf…. Looking forward to your feedback! (N/N)