How do early rollbacks play out?

Early rollbacks will result in higher rates of infection and death.

But thanks to persistent stalling by Trump and his eventual complete failure to come through with the needed supply of tests, public horror resulting in damage to the GOP at the polls and strong public pressure to return to lockdown will be headed off or at least much diminished, even if the bug is as deadly as has been feared.

So likely there will be no significant national return to lockdown until and unless we are on the way to that dreaded result, something very close to universal infection in the USA and deaths in the millions.

The GOP is charging ahead toward perhaps that result.

All of us have to hope the true lethality of this bug has been much overestimated because the lack of contact tracing and the dearth of testing have hidden the true number of Americans infected, perhaps 50 to 85 times as high as the number identified by individual testing.

[Update, 4/23. Subsequent studies in other parts of the country indicate the undercount of cases, while still large, is very, very much less than the Stanford study indicated, sometimes suggesting a true number of infected not more than twice the number found by individual testing.]

If that is not right we will lose maybe 2 to 5 % of those infected, which could be some 6 to 16 million dead Americans, assuming 320 million actually infected.

Or maybe a little less, or maybe a little more.

But of course if that Stanford result is right and at any rate we've got the number killed by the virus right (admittedly questionable if we've totally missed so many of the infected) then the true lethality of this bug is maybe 1/70th part of the rate based on infections identified by individual positive tests, the rate currently feared to be right or close to right.

[ 4/23. Or maybe it's 1/2 the currently feared rate.]

It could be 0.029 % to  0.071 %, far less than the usual annual flu, resulting in maybe 93,000 to 227,000 dead, again assuming 320 million actually infected.

[4/23. Or maybe the true lethality is about half what's now thought, and so maybe from 1 to 2.5 % yielding maybe 3 to 8 million dead nationally, on the assumption infection in the end is arbitrarily close to 100 % of the population.]

And that's plenty low enough for the GOP/Trump line that the threat has been wildly exaggerated by global panic and in fake news from the Dems and their media sympathizers to be vindicated in public opinion.

Right along with the most important and persistent of Trump's hunches, feelings, and intuitions.

And that will have exactly the dreadful political consequences for coming decades we Democrats fear.

Including even less pandemic preparedness.

None at all for the virus that spreads like this one but kills like Ebola.

[But lethality of 1 to 2.5 % and 3 to 8 million dead does not, I think, vindicate the Boob in the White House or those who agree with him.]