read the paper maybe before you comment perhaps
https://www.nber.org/papers/w27408.pdf
So I read far enough and skimmed the rest. They're basing most of their hypothesis on SafeGraph phone tracking data, and the assumption that since there was a net increase in social distancing during the protests (folks stayed home to stay out of the chaos), that it will equate to no increase.
First, there's simply not enough time to compare that data given incubation periods and downstream impacts to actuals that are still coming in, and future actuals. The health data they used ended on June 20th, while if you follow their charts, protests didn't reach their heights until a week after the murder and later. The chart (page 33 panel D) is obviously not reflective of what's going on this week as we're seeing some of our highest numbers ever.
Second, it doesn't address that post protest, those that did protest went home or back to their communities. They mention persistent protests, but not sure they addressed the impact of those who persistently protested going back and forth (protest, then home, then...).
Third, it doesn't address that many that protested, were African American, who we know are hit harder in terms of %s (and who return to their communities which we know have been less adherent to social distancing policy, and often live in multi family housing at a higher %).
Forth, I didn't see them mention that the median age group for protesters is likely less than that of the general population. Given the demographics, more likely to be asymptomatic, and also less likely to seek early medical treatment for mild symptoms.
I can keep going and going. It also admits that there could be a problem with their data if protesters came from outside of the city in question.
In short, way too early, way too many assumptions, and doesn't take into account a lot of basic factors. Too much stat head, too little common sense.
