They are testing who HAS the virus. So even a korea is missing those who have HAD it. They’re only seeing a snapshot of current infections. Now that is without context of where you are in the progress of the epidemic.
The paper that says half may have had it relies on the assumption that 0.1% of cases have severe illness and I think everything points to that being far too low.
All the models are guesses until we have a widespread testing of a sample of the population for both current and past infection, and repeat that a few times.
DOES it though? Coronavirus has been spreading for months. Here's what we do know:
We know Coronavirus has been spreading since at least the middle of november in china.
We know that China did fuck all for a month to contain the spread, even then they let people leave wuhan and fly around the world for weeks before locking down.
We know this disease can survive on hard surfaces, can spread through the air, and is more infectious then typical flus.
We know that testing has been sporadic at best and atrocious at worse.
Now, while we dont know how far Corona as spread, we DO know how well other coronaviruses spread, and how other flus spread. We can make accurate estimates of infection numbers because we have been measuring them for decades. If we calculate the current spread of coronavirus, not hospitalizations, just confirmed cases, using other coronavirus examples, that leads to massive swaths of the population infected in order to spread the disease so far. A disease with 90K cases int he US isnt going to have a smattering across most counties.
And we know Corona is more infectious then other coronaviruses or flus. So why would we assume the case numbers were mysteriously lower? Past coronavirus/flu spread+asymptomatic transmission for several weeks+more aggressive spread=widespread infection. The large number of hospitalizations would be indicative of a disease that is running roughshot over the population, a pandemic already at its height, currently infecting tens of millions. Even in hard hit italy, its largely the hospitals in lombardy that are utterly overrun, the rest of the country has heavy case loads but not as bad as up north. If the hospitalization rate really was 20% and the fatality rate 1.2%, we would be seeing more evidence of it in countries like italy and s. korea. We would have seen 8x the number of deaths, easily, int he US by this point. And we would have, on our hands, a virus that can randomly skip all around the country with seemingly no line of transmission, making quarantine utterly useless. Hand washing and social distancing work because we know this thing transmits the same way other coronaviruses do, and thus we can use past coronavirus examples to model this new strain.
It is far more likely that the coronavirus has already infected hundreds of millions around the globe, not the 350,000 or whatever that is currently being reported, and that its death rate is nowhere near what people are speculating based on tests. I wouldnt be surprised if half the US wasnt already exposed to the disease.