Right, but it seems more accurate to attribute the income discrepancy to the productivity of those people, rather than "oh you went to uni here's extra money". This means we need to decide if "people who are super-productive tend to go to uni at some point" or "uni produces super-productive people".