The latest issue of Nature had a lot of articles -- ranging from indifferent to bad -- about the PhD, Jonathan Mayhew has been talking about grad school/academia on his weird quasi-professional blog, and then there's the whole Peter Thiel-centered discourse on whether people should even go to college... research pressure prevents a long or esp. reasoned post on any of this so I'll just give a brief account of my own reasons for going to grad school, as I can't find anyone expressing the same view anywhere. (In general my views on higher education are "conservative" in the sense that I think radical changes are likely to destroy value, and "left-wing" in favoring high spending levels on education. But I've blogged about this in the past.)
I'm puzzled by the campaign to reduce the rate of production of science PhD's on the grounds that people are somehow being gypped. I am convinced that these hypothetical victims are mostly straw-men. I don't know of anyone who starts (e.g.) a physics PhD thinking it's an efficient way to increase their potential earnings -- or, for that matter, that a PhD guarantees a tenured job. Personally, I was intellectually adrift as a senior; I didn't want to take up a demanding job that was unrelated to my interests, and on the other hand I didn't trust myself, left to my own devices, to accomplish or learn anything, being indolent and dilettantish. Besides, there was much to be said for high-speed internet, insomnia-friendly hours, and access to university libraries. (I remember not applying to explicitly technical schools like MIT because I was worried they wouldn't have well-stocked libraries.) It seemed worthwhile to give grad school a shot. While some of this is idiosyncratic, a reasonable fraction -- between a third and a half -- of people seem to drift into grad school along vaguely similar paths.
(Why physics? (I was one of those multiple-majors.) One consideration, which should have been more important than it in fact was, was money. I didn't want to be dependent on my parents or on loans and this nixed law school. I briefly considered English grad school but was influenced by Brad Leithauser telling me that if I wanted to do English grad school it'd be wise to take a few years off after college; as I don't think far ahead his advice had the effect of just entirely killing the idea. I chose physics over math because the physics dept. at Amherst was friendlier and I had the amorphous sense that they'd do a better job of getting me into grad school somewhere.)
By and large, with a couple of minor sociological caveats, I think grad school served my purposes quite well. One gets paid enough to break even, without exceptional effort. The initial spurt of classes kept one engaged, challenged, etc. -- maybe pedagogically "worse" than college in general, but pedagogy is irrelevant beyond a certain level; what matters is whether the material is interesting -- and of course one can be essentially arbitrarily busy with research. (This is how one ought to do it but the flesh is weak.) There's more job security than in the average corporate workplace, esp. at a large public university; I think some people have been grad students for a decade or so. And I imagine that if I were to leave academia after my PhD and a postdoc, I'd find some manner of (semi-)quantitative work out there; while it's true that I'd earn much less than classmates who went straight to the real world after college, that is the nature of backup plans... admittedly the calculus is quite different if one (say) has kids to support, etc. but most people in the relevant situation do not.
(I suspect that the right kind of grad school would also be an ideal place to job-hunt from. Urbana isn't that, being just a little too far from Chicago. On the other hand, stipends do not scale with living expenses and one would be much more out-of-pocket in a major city.)