I also discussed this with some other professors in chemical engineering (a much smaller program, with very few TA's) and they told me that their dropout rate was essentially zero.
I was in math grad school for a year at UW-Madison.
The inelasticity seems to be on the labor side of the equation -- no matter how badly students and post-docs are treated, they keep coming back for more. My intuition is that stricter limits on fellowships (i.e. at most X years of government funding; more like the UK system), would provide less perverse incentives for everyone. But any politician to propose such a move would be pilloried for opposing science.
To the contrary, I would say that if we're going to have a serious discussion about the flaws in higher education system, we can correctly start from a premise that LEARNING is a virtue, a personal virtue that has positive externalities for society. But then we have to ask if institutions of higher education are effective at all, and economically efficient if shown to be effective, at providing learning. Learning is a wonderful thing, but much of learning occurs outside school walls, and some very socially useful learning doesn't gain the learner a credential such as a Ph.D.
Plus, not all learning (I would say the most important bits) have anything to do with economics or being socially useful, especially in the natural sciences.
Some education does, some doesn't. You can't conflate EE with Gender Consciousness in Mediaeval French Poetry, or whatever nonsense is being taught in the humanities these days. The flawed ideas is that all education is equally worthwhile and I'd argue that infact, that was never true.
Nor do I think people are bashing the wonderful and professional folks that are part of the collegiate system -- although some comments anecdotally move in that direction. Part of growing up is realizing that good people can be stuck in bad situations, i.e., nobody goes to work thinking "I'm sure glad I'm part of a system that is terribly awful!"
The rub of the matter is twofold: 1) Do the packages of learning and education put together by institutions actually hold together into something you might find cohesive and useful 30 years later? 2) Are the results of consuming these pre-packaged experiences worth the clearly staggering costs?
When I look at systems, whether political, computerized, social, or process, I'm looking for feedback mechanisms. How does the system self-adjust? In democratic politics, for instance, we have elections which can completely change policy and the direction of the country. Likewise in programming we have unit tests and frequent validation with users to make sure we're doing what we're supposed to.
I don't see that in higher education. When I look at the inflation-adjusted costs for a four-year degree in the U.S., for instance, it looks an awful lot to me like a system run amok. There's simply no way we're getting that much more value from a 4-year degree than we were 40 years ago. In fact, to be frank, it looks like public funding has created a system where colleges keep pushing the limits wondering where the ceiling is. If Joe Blow School of Fine Arts is charging $40K per year, surely Harvard is worth $80K, right? And whatever Harvard charges, guess what? Rich folks will pay. And because rich folks will pay, the rest of us will pony up the money so poor kids can go too. It just continues. It's a system without any feedback.
People go on at length about the socialization that happens in college, and I believe there are some valuable things there. But I spent time in the military and in college, and I got a lot of good socialization skills in the military too. Got to spend a lot of time outdoors, both learning and immediately applying knowledge in real-world situations. The Peace Corps, I imagine, is the same way.
So it's not a good or bad question. It's more of a question of "can't we get some of these same things somewhere else in a less expensive and easier-to-consume manner?" What are our options for training our future workforce? You can't have optimum performance in a system with only one option. It's a recipe for stagnation and laxness.
College used to open doors. Now it's essentially mandatory if you want even the right to apply for a decent white-collar job, and doesn't guarantee anything, because far too many people are going and there are too few good entry-level jobs out there.
The only way to save college is to cut it back to the status of something that's available to an elite 5-10%-- and I (along with most of us) would obviously prefer that this "elite" be chosen based purely on talent and not on means or family connections.
However, my point on the "idea [...] that schooling opens doors and improves one's life" was not that this idea is false. On the contrary, it's true of grade school, high school, and college. It's an idea that is dangerous on account of being true for a long time, then failing abruptly and catastrophically. (These are the sorts of ideas that form bubbles, e.g. tulip bulbs/tech stocks/houses are sure, good investments.)
With humanities PhD programs, the connection between education and social mobility semi-silently ceases to be true at that level, and there are a lot of students who aren't aware of the switch that has taken place until it's too late to turn back.
You went too far -- education does open doors, and if we're going to have any hope for the future, we need to continue to strive for a world where that's a fact. I know that it's fashionable to bash higher education in this forum, but an economy that's driven by technological progress takes more than high-school kids who know Python and HTML. At any rate, I sincerely hope that this isn't the first generation where education no longer correlates positively with economic success.
That said, I agree that Universities are taking advantage of naive young people -- especially when it comes to PhD programs. But if we're going to have a serious discussion about the flaws in higher education system, we have to start from the premise that higher education is itself a virtue.