That doesn't make sense to me. I would think you'd want to get to the point where you've reached the carrying capacity, so that "overshoot" of the equilibrium doesn't result.
See my other reply below. You can't overshoot 100%, and long before we get there herd immunity takes hold, and there's no question of the disease being kept from working its way through -- only a question of managing the rate at which it does. All in all the economic damage of continued shutdown does not make sense at this point (at 21% of the population exposed). We should stop now.
Sorry, but that's dumb. If 1/5th of the population has been exposed, then you know the total strain on the healthcare system is extremely unlikely to top an additional 4x what it has been, but not only it has 1/5th the population being exposed not in fact been enough to overwhelm the healthcare system, as more of the population is exposed you begin to get herd immunity, so in fact the total strain on the healthcare system is extremely unlikely to reach a total of 5x what it has been, very likely not even half that.
Give all of that, it makes no sense to continue the shutdown considering how costly the shutdown is.
To those who think this is a question of saving lives, that's not so. At this point the disease will work its way through the population because there's no time to finish testing, producing, and distributing a vaccine, which means the only question is of managing the maximum load on the healthcare system. Moreover, many of the most vulnerable have been hit already, and those who haven't can be protected by measures far short of a shutdown.
No, at 21% it is most definitely time to stop the shutdown. (Note: not all the measures, just the shutdown.)
We don't know that actually 21% have been exposed - hence the use of "preliminary" and "suggest". We need actual positive results that aren't based upon hiding behind the repeated use of "extremely unlikely". It seems like a bad idea to gamble based upon an initial suggestion.
The only actual good indicator we have is the rate of change of deaths and hospitalizations.
Weeks ago there was a report that 13.7% of pregnant women admitted during labor in NYC tested positive for covid-19. That was using a PCR test, not an antibody test, IIRC. That surely implies that weeks ago the rate of infection in the city was at least 13.7% minus false positives plus what would have been antibody test positive patients who were not producing the virus any longer. To go from ~13% to 21% in a few weeks seems perfectly reasonable. Moreover, "suggest" is a bit of a weasel word: there's no proof, only statistical approximation. So, yes, I suspect the city really is at 21% exposed.
> The only actual good indicator we have is the rate of change of deaths and hospitalizations.
Nonsense. The right thing to do is to perform randomized antibody testing of the population. The sample has to be random. The city (and other cities, and States) should pick an appropriate number of household addresses randomly, send techs and police to each one, pick a random person at each address, and test them. Given a sufficiently random and representative sample, you don't need to test so very many people to determine with a very high degree of certainty just how much of the population has been exposed.
Only then can hospitalizations and deaths be put in perspective. If SARS-cov2 spreads through enough of the population to reach "herd immunity" and its ultimate impact is, as I expect, much less than a bad normal-flu season, then the shutdown was not necessary and should stop (not necessarily other protective measures).
On the flip side, if very little of the population has been exposed to covid-19, and we were to stop the shutdown due to declining death and hospitalization rates, then the risk would be very high of a second wave.
Randomized antibody testing is absolutely essential here. We need to know how it progresses through the population, and we need to know ASAP.
EDIT: Read the comments below.