I'm past my edit time limit, but I should add context to the assertion that we can't estimate the value of a year of college. There are some good methods for doing so, but none that stand out as the correct way. It's a classic hard problem in econometrics.