Perhaps the fellowship for PhDs should only be two years, and a faculty review at the end of the two years cuts out the majority of students, so many more can b

Perhaps the fellowship for PhDs should only be two years, and a faculty review at the end of the two years cuts out the majority of students, so many more can be admitted and evaluated than actually graduate? And maybe another round after the 4th year? That would solve both problems: most people going into the programs wouldn't graduate, but many more people who wanted to try would be able to at least get in, and most of the students who did graduate would be able to get jobs (maybe, out of a 30 person starting cohort, 2 graduate with a PhD). It could vary year by year, so a year with a really great cohort, maybe the majority would come through, whereas with a bad cohort almost all of them wouldn't make it.