Grant more PhDs per year, while reducing the cost of education to the student. Grad student stipends are often meager, assuming you even get funding.

That's not more productivity, that's more effort.

Well, it's the same amount of "effort" -- you supervise the same number of students, they just graduate faster. And most department metrics Ive seen measure PhDs granted as an output.

Ah yes, in terms of number of PhDs you can chuck out, fair enough.