I would go with a more objective criterion: how many graduate and get a job in their field.
It would not be complicated to send the IRS a list of social security numbers of people who graduated five years prior, along with their classes and professors, have that matched up with gross income now, and after matchup strip the identifying social security numbers so as to anonymize the data.
You would now have the data needed to correlate professors and courses with after-graduation economic viability. Fire the professors whose students can't seem to get a job that's worth what they paid in tuition.
Another suggestion: mandate the professors pension funds invest in student loans. If the student loans get paid back, great. If the loans don't get paid back, the professors' pension funds take a hit.
Various measures of post-graduate success are available and are used in evaluating schools and majors. These are incorporated into U.S. News rankings of schools and Payscales.com offers first year and fifth year earnings of graduates by school and by major. This is good information for students.
However, college administrators are happy to take tuition dollars from all students regardless of whether the students care about whether their college degree will actually be worth something. The overwhelming majority of students nowadays are more interested in winning sports teams and student unions than in challenging programs involving math and science, no matter what the post-graduate outcome.
The data required to evaluate individual faculty based on post-graduate success of their students would be enormous. Assuming a student takes forty different 3-credit courses, this means 2.5 percent of a student’s post-graduate success might be attributed to any one of his professors. Furthermore, are we talking about the first-year salary or the fifth-year salary? First-year (starting) salaries are all about the reputation of the school, not about any particular student’s ability. To wait for fifth-year salaries means a very long delay between a particular professor’s work and the evaluation of it.
The point of this thread is tenure. More specifically, how tenured professors might go on “teaching” even though they could retire, when they might not be productive in the classroom. Tenure has its problems. But, ending tenure introduces other problems. Specifically, if colleges base keeping a job solely on student evaluations, the consequence will be a watering down of academic rigor and grade inflation. Post-graduate performance is impractical. I suggest student performance on a standardized exam.
I would like to conclude by thanking you for your confidence in professors. You think professors can overcome public policy, college administrators and unengaged students and turn around the deterioration of standards in our universities, if only professors were sufficiently motivated such as by having their pensions invested in student loans. While I think you’re mistaken about our ability, I accept the compliment.