It doesn't matter--the number of PhDs is an order of magnitude larger than the number of tenure positions opening.

MIT graduates about 400 PhDs per year and only has about 40-50 tenured positions open each year.

I suspect every other program is similar. There are simply too many PhDs being graduated, but the universities are dependent upon them for slave labor.

It sounds like his pay as a "professor, post-tenure" was reduced to 49% of what it was before, so it's almost certainly a money problem.

But these post-tenure 49% positions are the new thing to encourage people to go -- the previous option was emeritus, not only 0% but also unable to be the PI on grants etc. So they must have found a magic money tree somewhere to pay for them.

Not mentioned yet is the change from defined benefit pensions. If you had a final-salary pension, then there was no financial incentive not to go emeritus, in the good old days. Whereas on a defined-contribution scheme, retiring a year earlier does literally cost you (or your estate) a year's salary.

Large and famous universities sit on very large piles of money. They just choose to invest them differently — or maybe the monies are granted on particular conditions to spend them.

I mean a school like MIT must be able to afford opening more tenured spots, right?

You do get some growth by increasing the number of students you have and if research budgets expand but I'm guessing that is a function of inflation and population growth. Technology also allows you to teach more students with less faculty input so maybe you get productivity gains there.

Money.

Then they 'd have to enroll more students.

Or stop one-upping each other by building ever more expensive sports facilities?

You're referring to MIT's championship football team?

I'm referring to all the times I've heard on HN that US universities one-up each other in building sports facilities.

I heard sport facilities are actually a source of profit for the universities?

Schools that don’t do this still have the same money problems.