Comment by phreeza

I was simplifying, the point is that the model should always reflect the current state of the population. You shouldn't expect the kind of "catch up" effects that OP is referring to, unless you have static predictions that don't take into account actual deaths and births.