Phase III trials don't run with enough participants to detect 0.01% effects with any confidence. These are generally found postmarketing, like Vioxx. In terms of vaccine harm, a badly behaved swine flu vaccine caused 3 cases of narcolepsy per 100k juveniles vaccinated. This effect simply would never be seen in Phase III trials.
Finally, beyond the raw CFR we need to consider other adverse outcomes from covid: cognitive impairment following mechanical ventilation and cardiac damage are two that really worry me.
I've amended the post's math - it's definitely not as hard a point, but the secondary consideration I didn't include was that it's also not an all or nothing proposition - a fatality rate from a vaccine is one thing, but even COVID-19 doesn't just have a fatality rate - it's got a much much higher "severe long term complications" rate. Which is also very much what you're looking for with trials.
Having things look good after 3 months in a vaccine trial doesn't tell you if they look good after a year necessarily - and all of this is really weighted against "a sensible response to limit the spread of the virus is possible, the US is just not doing it".
EDIT: The other possibility is that a vaccine that just straight up doesn't work very well would also make the situation worse - churn out millions of doses, get a limited amount of immunity in the recipient population, government is compelled to wind back "mechanical" spread prevention measures - and the epidemic cinders for a while and then rips through the population because everything thinks they're immune.
Because if it doesn't work, you haven't had a doctor give someone something with unforseen side-effects: they injected them with a deadly virus and killed them (in the worst case).
Now we have treatments like Regeneron which the President received, which most likely made a very big difference in his fairly rapid convalescence - doing it when you have more plausible interventions if it doesn't work is quite different.
I mean all of this stuff is, relatively, going at lightning pace - if we have a vaccine early next year, that's what, 1.5 years between initial detection and immunization for a novel virus? That is breakneck pace.
I believe the silver bullet for future pandemics will be pre-approved mRNA vaccine pathways with the only change to the specific vaccine being the mRNA payload that targets the pathogen. Then testing could be a lot more straightforward and quicker which could end future pandemics.
Might wanna check that math, dude.
Even with your preposterously deadly made-up vaccine, you’re only hitting about one month’s worth of COVID deaths.
[1] https://timesofindia.indiatimes.com/life-style/health-fitnes...
So whereas the US has currently had a total of about ~10 million cases to date, this represents about 3.3% of the total population. There have been about 200,000 deaths so far, so the US case-fatality rate is about 2% - note, the mortality rate is most likely quite a bit lower overall since it's spreading so much. This is simply deaths vs. observed cases.
EDITED(got the math off by a factor of 10 here previously - at 0.01% fatality it's not 360,000 - that said, see comments below, actually establishing that in clinical trials is not easy): Let's suppose we rush through a vaccine that turns out to cause fatal complications in 0.01% of the population, and we give it 90% of the population. Straight up - we've just killed 36,000 people. But that's assuming that the complications are binary - but COVID-19's complications aren't binary, and neither would negative side-effects be. So we might kill a bunch of people, and injure tens of thousands more seriously. Or we might have inadequately diverse trials and miss a whole population of potential negative side-effects (a classic problem in a lot of trials: university aged students with the distribution of whichever university developed the product are over-represented in samples).
You cannot rush clinical trials. To even stand a hope of getting to 0.01% risks you need a population of at least several thousand in your trial. You need to monitor them aggressively because hopefully you have good data that whatever can go wrong will be slow enough you can intervene, but...there are always risks. Clinical trials can and do go badly wrong and people involved in them are taking a selfless risk for the betterment of the rest of us - and that's a very different proposition to declaring "this'll probably be fine" and not collecting the data.