I, like him, was interested in data beyond anecdotes, since attrition rates, outcomes, and the "value" of a PhD from a non-Top 5 school seem to me to be very dependent on field. Looking in detail for my field (and with the caveat that there are definitely problems with the NRC dataset), I was very surprised by the data. Here is what I found (using the NRC S-ranking, and noting that academic plans means signed contract for postdoctoral fellowship or permanent academic job):
Top 10 schools: ~60% complete PhDs, ~20% with academic plans
Top 20 schools: ~50% complete PhDs, ~20% with academic plans
Bottom 10 schools: ~50% complete PhDs, ~20% with academic plans
Bottom 20 schools: ~50% complete PhDs, ~20% with academic plans
I will note that academic plans does not mean faculty positions (as implied by the Chronicle for Higher Ed flash tool, now pay to use)--the NRC definition is negotiating or signed a contract for an academic position including postdoctoral fellowships. In fact, I suspect most of these are postdocs. I think it would be really hard for schools to report on how many students have faculty positions for my field, given that a postdoc is more or less required, and most schools don't seem to follow up much after the initial placement (at least no one I know has been contacted as an alumni for that information). Still, the numbers are a lot smaller and a lot more consistent across the rankings than I would have expected, especially given the hysteria about overproduction of PhDs. I should also note that for a graduate of a US University, it isn't incredibly difficult to find a postdoctoral position in my field, though positions in top schools and/or top labs are very competitive.
The average number of PhDs graduated per year ranges from ~2 to ~30, with a mean of ~14. The net result is ~2500 PhDs produced per year (which is in line with the data from my field's society), of whom ~500 will go into academia (probably postdocs) immediately after their degree. This does not seem like a terrible oversupply of PhDs/postdocs even just considering those 500 per year, especially since anecdotally, some large percentage of them will not pursue faculty positions. Granted, not all of them can get positions at large research universities (which is what many people who want faculty positions profess to want), but I would guess that across academia at all levels, there are probably 200-300 faculty openings per year in my field. This is a much better match of supply with demand than I was expecting, even given my hand waving invocation of anecdata at the end.
UPDATE: Note, that I am NOT implying that getting a faculty job is easy, since I know it isn't. Nor I am implying that many openings don't have hundreds of applicants (I know we get that at ProdigalU). Just that supply of jobs across academia is probably a lot closer to demand than I would have guessed in the absence of data, especially if we lump all the different types of Universities together, which is obviously a gross simplification. I know that many of my research colleagues (me included!) would have gone to industry to do development rather than a community college to teach due to personal work preference.