Crooked Timber reviews a paper on how the job market in social science and humanities faculties have a great deal of hierarchical sorting. In the comments, Brad DeLong notes that there is a market failure for assistant professors in economics that are "underplaced". I'm not arguing -- our own department has benefited greatly from that failure. But what I wonder is why this continues to happen, particularly when places like Berkeley are supposedly trading in that market? Kieran refers to another paper that uses an anthropological explanation:

departments are tribes, graduate students are women to be married off, and areas of specialization are clan-markers that help define which exchanges are appropriate and which are taboo.

I bet Invisible Adjunct (link fixed, thanks Eric!) has a field day with that metaphor.

UPDATE (11/12): She found it.