Joni Hersch of Vanderbilt has a new paper on this topic. Given the multiple dimensions of unobserved quality, I wonder if there is any method which can convince me on such questions. Still, I am glad to see someone putting the effort in. Here is what the author came up with:
Income disparities arise not only from differences in the level of education but also from differences in status associated with an individual’s degree-granting college or university. While higher ability among those who graduate from elite undergraduate institutions may account for much of the earnings premium associated with elite education, ability should be largely equalized among those who graduate from similarly selective graduate programs. Few graduates of nonselective institutions earn post-baccalaureate degrees from elite institutions, and even when they do, undergraduate institutional prestige continues to influence earnings overall and among those with law, medical, graduate business and doctoral degrees.