[Pyramus]

Here is a good estimation based on seroprevalence data from different countries [1]

> Taking France as a reference population, the ensemble model estimates a population IFR of 0.79% (95% credible interval, 0.68–0.92%).

Crucially, IFR correlates with the capacity to treat the infection.

[1] https://www.nature.com/articles/s41586-020-2918-0/figures/2

[_cs2017_]

Except that seroprevalence itself likely underestimates total infections: https://www.bmj.com/content/370/bmj.m3364.

Also, it's quite suspicious that Japan has the highest IFR in the world, with the gap over the average much larger than can be explained by population age. Especially strange since Japan has some of the best medical systems in the world, and has not (yet) exceeded the capacity of its hospital system.