While I do think that 20% infection rate PLUS other mitigation measures (whether government enforced or not) may be enough to have sustained R_t < 1, it is worth noting that there ARE areas that had >20% infection rate.
In Bergamo, Italy -- one of the first hot spots that was late to shut down after experiencing their outbreak, had 57% of their residents get coronavirus.
https://www.reuters.com/article/us-health-coronavirus-italy-antibodies/over-half-of-people-tested-in-italys-bergamo-have-covid-19-antibodies-idUSKBN23F2JVhttps://www.medrxiv.org/content/10.1101/2020.04.15.20067074v3The death rate of the entire population was 0.58%, which means that if 57% of the entire population was infected, then the infection fatality rate is about 1%, which lines up with the calculated IFRs of similar early-breakout locations when treatments were still in their infancy.
That doesn't suggest to me that something like that could happen in the US now, by and large their outbreak happened before people took the virus seriously and they were extremely late to react. They also had the issue of multi generational housing and an older population. They were probably undergoing something close to the true R0 of 2.5 during their outbreak, whereas in the US the places that are "outbreaking" are peaking at R_t of 1.1-1.3.
But, to me it is reasonable evidence that there is not some magical barrier at 20% infection where the virus dies out regardless of what behaviors we pursue. I think we are just seeing that 20% infection of the population, in combination with other behavior changes, testing, tracing, etc. is probably enough to have sustained R_t <1.