Super spreader cohorts and covid-19

A simple two-cohort SIR like model can explain the qualitative behaviour of the logarithmic derivative estimations of the covid-19 epidemic evolution as observed in several countries. The model consists of a general population in which the R_0 value is slightly below 1, but in which a super-spreading small subgroup with high R_0, coupled to the general population, is contaminating a significant fraction of the population. The epidemic starts to slow down when herd immunity is reached in this subgroup. The dynamics of this system is quite robust against non-pharmaceutical measures.