professors - Do universities hire graduates from lesser universities

Recently, someone analyzed computer science professors at top universities and found that over half of the professors at the top 51 universities graduated from a top 10 university. Others have also brought this up. From my personal observations, most schools do hire graduates from better ranked schools.

1. Is this because of the competitive job market? We have so many good applicants, we have to narrow it down some how!


2. Or is it simply that these schools produce the most PhDs?


3. Has this always been the case?


4. How rare are exceptions to this? I know of a few people who graduated from a top 75 school and got hired at a top 50 school. But what about bigger gaps? The top 10 schools seem to just swap graduates, do they ever hire from a 50+ ranked school?

Update 2018: I have accepted a tenure-track position at a top 75 department at an R1 university immediately after graduating from an unranked department at an R2 university. It does happen!

_{This may or may not generalize to other fields and countries.}

Answer

People are occasionally hired by far more prestigious universities than the ones they studied at. For example, there's a tenured professor in the Princeton math department (unambiguously among the top 5 departments in the U.S.) who received his Ph.D. in 1999 from Kansas State (which wouldn't necessarily make the top 75). Where your degree is from is a negligible factor in hiring decisions compared with how outstanding your research is.

On the other hand, research excellence is highly correlated with which doctoral program you attend. The top programs tend to get the students with the most talent, determination, and preparation, and they usually provide the most support for these students to succeed. Of course this is just a statistical assertion, not an absolute law. However, in mathematics in the U.S., the number of students graduating each year from rank 50-75 universities whose job applications are as impressive in research as those of the average top-5 graduate is tiny. If you're hiring based on research promise, then even the most unbiased search should lead to hiring mainly people from higher-end schools. Of course there's presumably some prejudice as well, but I don't think it's a substantial factor at research universities. (I have no first-hand experience with hiring at teaching-oriented schools. In particular, I don't know how overrepresented graduates of prestigious universities are or which factors are responsible for it.)