It isn't, and the same phenomenon has been observed in other cases. The 2009 swine flu for example had R0 estimated between 1.4-1.6, yielding a theoretical herd immunity threshold of 29%-33%, but estimates of how many people were actually infected range from 11%-21%.