The key part is the assumption that in between step 1 and step 2 that productivity and everything related to it remains stable. This is almost certainly never true.