That would be for a homogeneous spread inside a country. However if the epidemic spreads from region to region, you could reach a local herd immunity in the fi

That would be for a homogeneous spread inside a country. However if the epidemic spreads from region to region, you could reach a local herd immunity in the first regions infected way before seing the country-wide curve becoming linear. But thanks, trying to find a counter-argument to your post made me realize a single-curve model probably doesn't make a lot of sense to try and understand accurately the spread of an epidemic.