[astigsen]

> The main thing that gives me some doubt is it seems plausible that we'll retroactively find that the deaths directly attributable to COVID-19 were somewhat overstated, in the sense that my understanding is a presumed COVID-19 diagnosis is accepted in lieu of having a positive test.

Looking at the total number of excess deaths (for all causes, ignoring if it is COVID-19 or not), I would say that it looks more likely that it is severely undercounted.

You can find the graphs here: https://www.ft.com/coronavirus-latest (look for excess mortality rates for New York City).

I guess you could argue that the huge increase in deaths was caused by side-effects of the lockdown, but that does not really jive with the results from countries like Denmark, which did an early and hard lockdown, and as result show almost no extra mortality.