20 May 2012

What College Bubble?

The number of PhD recipients on food stamps and other forms of welfare more than tripled between 2007 and 2010 to 33,655, according to an Urban Institute analysis cited by the Chronicle of Higher Education. The number of master's degree holders on food stamps and other forms of welfare nearly tripled during that same time period to 293,029, according to the same analysis. [Hat tip.]

There have been some that have proposed that the current surge in college costs is not proof of a bubble, but rather the natural byproduct of college’s sorting function.  (I think I first heard this proposal at Foseti’s.)  If that were the case, it doesn’t make sense that holders of advanced degrees are having difficulties getting good jobs, since the natural purpose of sorting is to take the best and brightest and put them in the best positions.

The theory of sorting makes its case on the grounds that colleges are largely meritocratic—a dubious claim at best, though true relative to colleges of, say, fifty years ago—and that they can be trusted to determine the best, brightest, and most dedicated.  Naturally, employers cannot perform direct testing for this, mostly because those sort of tests are racist, and so they need other proxies.  The meritocratic elite just so happen to provide those proxies.

Unfortunately, the sorting theory of higher education is untrue because the reality does not follow the model:  namely, those who have earned high educational honors and degrees aren’t more employable or working the better jobs.  Thus, if college is supposed to sort people, it has obviously failed, as evidenced by the fact that holders of advanced degrees are 300% more likely to receive food stamps now than three years ago, while US citizens in general are only 43% more likely (see linked article above.)

Funnily enough, there is a model that would generally predict this occurrence, and it is the bubble model, which posits that wages for holders of college degrees will decline as the supply of holders of college degrees increase, which is a direct result of government intervention into the market, particularly through the expansion of cheap credit and direct subsidy.  Low and behold, this has come to pass, mostly because the bubble model has better predictive power than the sorting model, and is thus more correct.

Since we’re on the subject of college degree holders, I’d like to point out as an aside that the idea that degrees aren’t real property because they aren’t transferrable is partially false.  It is true that one student can’t sell his credentials to another student, but it should also be noted that students aren’t the only ones who use the credentials they earn.  Employers also use student credentials by hiring employees who have certain credentials.  While they don’t “transfer” credentials per se, it is observably true that when someone switches jobs, the people employing their credentials also changes as well.  Given that there are signaling elements to college credentials (as evidenced by every guidance counselor that ever repeats the trope that college grads do better on the job market because they’re college grads), it should be plausible that there is a type of transference that exists with college credentials, except that is at the employer level, not the possessor level. Incidentally, this conceptual model reinforces the idea of a college bubble since it suggests that there can be diminishing marginal returns to adding one more college educated participant to the labor market, thus driving down wages.


  1. I've seen this stat in several places, but none break it down by degree, even in broad terms. My bet is that the humanities / liberal arts / women's studies crowd is hurting but folks with STEM degrees are doing just fine.

  2. I'd be interested to read a precise definition of the College Sorting Theory. On the one hand, colleges do obviously sort in at least two ways: 1) admission rates, which are a proxy for SAT 25/75 breakdown, which is a proxy for IQ; and 2) by major field (non-STEM vs. STEM), which also is a slightly weaker proxy for IQ.

    So the Columbia journalism grad and the Ga Tech ChemE grad have both been graded by college, but in different ways. The Columbia ANYTHING grad has been graded high IQ because he got into Columbia, duh. The Ga Tech guy had to major in a hard field to prove his bona fides. But Womens' Studies grad, from an otherwise non-selective school, hasn't proved much of anything even with a PhD (except may be she's above average in the self-delusion department).

  3. @southern Man- I've no doubt that the STEM majors are doing better than thier intellectually wimpier counterparts. My point was more that the sorting model is bunk, and the better predictive model is the bubble theory.

    @Steve- The sorting theory, in brief, is that colleges serve as an occupation sorter. This is mostly due to businesses not being able to sort directly because direct testing of cognitive abilities and work capabilities is racist. As you noted, colleges do have to have some sort of sorting mechanism (tests), though the push for affirmative action has diluted it. Thus, grades and actual graduation stand as the better proxy for sorting. In theory, it's supposed to be the case that having a college degree ensure that you get better jobs but as my post on Manufacturing jobs shows, this is not the case anymore. Basically, the best and brightest go to the best and brightest colleges and get the best and brightest jobs. Mids go to mid colleges and get mid jobs, low go to low-prestige colleges and get low prestige jobs, and those who don't go to college get shit jobs, and those who don't graduate from high school go on welfare. This is not a very accurate depiction of reality, though, which is one reason I dislike this theory. You can do some searching and link-chasing at Foseti's and probably find some academic stuff, if you want.

  4. But I don't see how the Sorting Model and the Bubble Model are necessarily opposed. In fact, they seem completely complementary. The Bubble Model is precisely what you get when you fail to understand the Sorting Model. When you fail to understand that college attendance and/or choice of major sort individuals into neurological categories, you will naturally assume that you can simply buy whatever credential those who get sorted into the "Make Pretty Good Money" category are getting. It's just a classic failure to appreciate that correlation does not imply causation, or more properly that correlation can imply causation exactly the other way around.

    The sorting being done by STEM degrees and/or Ivy (or near Ivy selectivity) admissions is as strong, if not stronger, than ever. Obviously AA screws with this, but the Ivies get their pick of the very best AA admissions (often not traditionally African American or traditionally Mestizo.) Never in history has the US Supreme Court been is little diverse as it is today in terms of almae matres. It's just that a whole lot more idiots are now getting business or liberal arts degrees from Ohio State because of the Bubble Model. But that doesn't mean that college don't sort. Only that they're not doing a good job below about +1 SD.