https://www.theguardian.com/us-news/2019/dec/11/boeing-737-max-plane-faa-regulators-crash-risk
TLDR: FAA estimated one extra Max crash per 2 - 3 years due to MCAS when deciding how to proceed after the Lion Air crash, but decided to let Max continue flying (effectively leading to the Ethiopian crash).
Let's assume 1 extra crash per 2.5 years
c390 Maxs in service at grounding. Production rate of 52 per month. Average number of Maxes in service over next 2.5 years = 390+(2.5*12*52)/2 = c1070
For ROM purposes, let's assume those aircraft fly 12 hours a day, 330 days a year. Over 2.5 years, that's 2.5*330*12 = 9900 hours over the period.
Combining the two to get total Max flight hours per extra hull loss, that's 1070*9900 = 10,593,000, so call it 10.6m flight hours.
Permissible hull loss rate for certification = 1*10^-9 per flight hour, or 1 per 10^9 flight hours.
Dividing the two to get the increase in risk, 10^9/10.6m = 94.3.
So that FAA decision let Max fly with a known issue causing a hull loss rate almost 100 times that permissible for certification.
(NB, I'm just pulling the figures out of the Guardian report, the FAA report released by Congress seems to have calculated risk over the life of the aircraft from the phrasing in the article, so someone has converted that to the 1 per 2 - 3 years figure, and it's not clear if that's FAA, Congress, or the Guardian, the fleet size calculation could be off depending on how they've done this).