It's an interesting article, but how exactly is "skill" measured for things like [1]? I couldn't find a description of this in the article or the linked paper (that I skimmed).

Furthermore, how does this account for personal preference for less dangerous jobs for example? I'd much rather be a waiter than a coal miner, even though you could say both require the same/no skill.

Pay is only one of many different aspects of compensation (a more dangerous/stressful job will likely pay more even if it requires the same skill; many teaching jobs have a lot of vacation which you might value more than a pay increase, and so on).

[1]: http://www.brookings.edu/~/media/Blogs/social-mobility-memos...