Postwar higher education has proven one of America's most effective subsidies, and it has paid for itself many times over.
Consider the phrase "many times over." To have paid for itself, it would have had to generate a present value of returns equal to the present value of costs. "Many" must mean at least three. So that would be a present value of returns equal to at least three times the present value of costs.
Is that plausible? I think not. Ignore, for a minute, which Tyler appears to do, the strong case made by Bryan Caplan [the links for Bryan's posts are too numerous: just do a search, within Econlog, on "signaling"] that much of higher education is signaling. Even if that were completely false and none of it is signaling, a huge part of the gain from education is a private good captured by the person who is educated. To make Tyler's case, one would have to make a case that a large part of the return from education is a public good. But he doesn't make that case or even link to a case.
Parenthetically, I found it ironic that one of Tyler's commenters, Gibbon, accused him of pretending that public goods don't matter and of denigrating "such key institutions" as state colleges. For Tyler's case to make sense he has to be arguing that there is a huge public good. Further, parenthetically, although I think the term "ad hominem" is thrown around way too frequently--most alleged claims of ad hominem arguments are actually not--in his last sentence, Gibbon gives a pretty clean example of an ad hominem argument when he writes, "I guess most libertarians are rich enough that this [eviscerating state colleges] doesn't bother you." Still further, the irony here is delicious: in a post in which Tyler is arguing for subsidies to state colleges, the obvious charge to make, if one were to make an ad hominem argument, would be to attack Tyler for arguing for subsidies for his employer, a state college. I hasten to add that that is not at all what I think motivates Tyler: I'm not making the ad hominem argument. My point is that Gibbon's ad hominem argument doesn't even work on its own terms: as an ad hominem argument.
Back to the issue. When one brings in Bryan Caplan's points about schooling as signaling, much of the subsidy to colleges is wasted and Tyler's case becomes even weaker.