After reading the article, I posted this response on The Economist's website:
I speak with some knowledge here: I hold a Ph.D. from the University of California, Berkeley specializing in evaluation of education and subsequently have been on the faculty of four graduate schools, including Berkeley.
In my view, the key to abbreviating the length of a Ph.D.and post-doc program is to eliminate the arcana, of which there are mountains. Primarily teach the tools required for excellent research: e.g., research design, how to critique research, and yes, expose the students to core cutting-edge areas in a discipline to help them identify their desired research focus. Such training could, in my judgment, most time-effectively be learned if the program were just two years long.
As important, I'd cut by 80% the number of slots in Ph.D. programs, referring the other applicants to equally abbreviated, practically-oriented doctorates (e.g., Psy.D, Ed.D.) taught not by Ph.D.s (disproportionately esoterica-focused theoreticians) but by master practitioners with the rare ability to convey their mastery to students.
So much university research is known, apriori, to be of trivial value, certainly known upfront to be cost-ineffective expenditures of taxpayer dollars--and much of such research indeed is funded by the taxpayer. Especially in these tough economic times, it would be wiser to allow taxpayers to retain their money than to fund yet another study on, for example, the deconstruction of the use of the doppelganger in 19th century literature.
UPDATE: Upon reading this post, a reader forwarded me this even more powerful caution against pursuing a Ph.D.
And I'm reminded of an article on the topic that now, five years later, still remains top-of-mind for me.