Why are you assuming that epidemics only stop growing exponentially when they reach population saturation? That is not the case for all the recent respiratory epidemics.

Even if they do, why is 20 a "low" number? They are talking about estimating the current real figure, from the current laboratory-confirmed figure. 20 seems pretty reasonable if a little high, assuming the epidemic is indeed still in the exponential phase.