How much people earn is not the only test of the quality of their education.
The long term value production of the average prostitute strikes me as damn near zero, unless the outlet prevents the john from sexually assaulting someone else. I agree, though, that the short term value can be quite high, and I think that is what people pay for.
It seems to me that value delivered to a sufficiently large audience would have to be of the long-term variety, because otherwise it wouldn't last long enough to get sufficiently dispersed. It's possible something might contradict that assertion, but I can't think of any examples right now.
Here's why: The value derived from an individual from a restaurant is food, and perhaps being in a pleasing ambient environment for a short while. They derive value from the environment only while they are there, and from the food for the next day or so. After that, no more value unless you think he might have died otherwise, in which case he gained long term value in still being alive.
The deal with restaurants is that they continuously deliver short term value to a stream of people who, hopefully, come back again and again. The value of an individual meal is fleeting, but they're willing to provide that value again and again.
My point is not to confuse a steady stream of short term value with a discrete instance of long-term value. I can use a well-built chainsaw for far longer than I can use a New York strip steak from a high class restaurant. Maybe you and I are discussing value from different perspectives though, but I'm not arguing that short-term value is a lesser degree than long term value. I'm just arguing that they aren't the same.
For a country, most economists would consider GDP. For me, I would consider sustainability and happiness.
PG made a point (I think in one of his essays that I can't remember - correct me if I am wrong) that I've found to be fairly true (though of course not always) - that often, jobs for a given level of qualification have a set wealth they will pay, and this payout can come in the form of a more interesting job or more money. This generalization does have a subtlety though - when you are a higher valued candidate for work, you are not only offered more wealth but a greater diversity of choices of work that offer different proportions of interest vs. money. Essentially what I mean here is that it is much easier, coming out of Harvard, to get a high paying job than it is to get a really interesting job. When you are in the middle rungs of math at Harvard, it is not that hard to get a job as an I-banker (lots of money, but two years of very boring work), but you can't just trade off some of that money for a more interesting job. It is only when you are smarter that you get the option of having either the high paying I-banking job or a lower paying job that is more interesting (i.e. cutting edge grad school work at MIT's AI lab). A lot of the smartest people want this more interesting job not only because they are driven more by their interests than money, but also because these jobs are more prestigious since only the smartest people can have them.
On top of this, I knew a lot more people from Harvard who were willing to forgo making much money in their earlier life by taking travel grants or just going off to some country to make some documentary or doing a startup (something that I think has a much lower total expected outcome of income than a finance job). A lot of the smarter people were willing to sacrifice money for freedom because they were sure they would probably eventually make enough money to survive (and the degree helped this), and freedom was just more important to them.
Given all this, measuring money in these studies doesn't even seem good enough to me to be a heuristic for success. People at the most prestigious universities seem to often have the freedom and luxury to have different priorities and to simply not care as much about money. It seems like even a study asking people about their job satisfaction some number of years later would be better.