Deans serve five year terms and often seek another 5 year term.  In this age of performance criteria, how are Deans evaluated?  Since they were originally chosen by the Provost, the Provost has a slight conflict of interest in determining whether the Dean has done a bad job during his/her 1st term.  Departments who are part of the Dean's Empire do not have the right incentives to tell the truth about the Dean's performance. If they anticipate that the Dean will be re-appointed and their remarks will become known to the Dean, then the Dean may seek revenge.   With faculty, external letters play a key role in determining evaluation but with Deans there is no role for such letters because all of the Dean's effort is unobserved by outsiders.  If the Deans are evaluated on money raised, then this raises the counter-factual of how much $ would another Dean have raised?  Are Deans at a university benchmarked relative to each other? But, how can the Dean of Social Sciences be compared to the Dean of the Medical School at the same University?

So, what is the "efficient way" to evaluate a Dean's performance? If we are unable to design such a performance metric and do not promote Deans based on this metric then what does this mean about the future of our universities?   You might say that the same issue arises with respect to managers of Fortune 500 firms but is this true? The external market for talent and poaching by other firms provides a market test of the quality of internal talent.  

With faculty members, we are one person firms and there is a clear link between our reputation and our teaching and research work which can all be quantified using various metrics. Deans engage in Team Production. If the Law School makes great strides during Dean X's reign, how much credit does the Dean receive and deserve?  Did the previous Dean make the hires and these people reached their peak production during Dean X's reign?  Did the economy boom when Dean X was in charge and the donations flowed in?  Note that this point is similar to the CEO pay for luck literature.

My big question here is how does a non-profit organization know when it does and when it doesn't have excellent leadership and what incentives are in place to encourage such efficient sorting?