The primary rationale for tenure is academic freedom.  A researcher may want to pursue an agenda which is revolutionary or offensive to Deans, students, colleagues, the public at large etc.  However, the agenda may be valuable and in the end dramatically add to the stock of knowledge.  The paradigmic example is Galileo who was persecuted for his theory that the Sun is at the center of our planetary system and not the Earth.  Galileo spent the end of his life under house arrest.  Einstein considered Galileo the father of modern science.  Tenure would now grant Galileo the freedom to pursue his ideas without threat of persecution.

From the profound to the more prosaic: the economic approach to tenure.  For economists, tenure is simply another contract or institution and we may ask, when is tenure the optimal contract?  My favorite answer to this question is given by Lorne Carmichael’s “Incentives in Academics: Why Is There Tenure?” Journal of Political Economy (1996).

Suppose a university is a research university that maximizes the total quality of research.  Let’s compare it to a basketball team that wants to maximize the number of wins.  Universities want to hire top researchers and basketball teams want to hire great players.  Universities use tenure as their optimal contract but basketball teams do not.  Why the difference?

On the basketball side of things it’s pretty obvious.  Statistics can help to reveal the quality of a player and you can use the data to distinguish a good player from a bad player.  And this can inform your hiring and retention decisions.

On the research side, things are more complicated.  Statistics are harder to come by and interpret.   On Amazon, Britney Spears’ “The Singles Collection” is #923 in Music while Glenn Gould’s “A State of Wonder: the Complete Goldberg Variations” is #3417.   Even if we go down to subcategories, Britney is #11 in Teen Pop and Glenn is #56 in Classical.

So, is Britney’s stuff better than Bach, as interpreted by Glenn Gould?   I love “Oops..I did it Again”, but I am forced to admit that others may find Britney’s work to be facile while there is timeless depth to Bach that Britney can’t match.

I’ve tried to offer an example which is fun, but it is also a bit misleading as the analogy with scientific research is flawed.  First, music is for everyone, while scientific research is specialized.  Second, there is an experimental method in science so it is not purely subjective.  But the main point is there is a subjective component to evaluating research and hence job candidates  in science.  There is less of this in basketball.  Shaq is less elegant than Jordan but he gets the job done nonetheless.  The subjective component actually matters a lot in science because of the specialization.  Scientists are better placed to determine if an experiment or theory in their field is incorrect, original or important.  And they are better placed to make hiring decisions, when even noisy signals of publications and citations are not available.

Subjective evaluation is the starting point of Carmichael’s model of tenure.  If you are stuck with subjective evaluation, the people who know a hiring candidate’s quality best are people in the department that is hiring him.  If the evaluators are not tenured, they will compete with the new employee in the future.  If the evaluators hire who is higher quality  than they are themselves, they are more likely to get sacked than the person they hire.  In fact, the evaluators have the incentive to hire bad researchers so they are secure in their job.  This reduces the quality of research coming out of the university.  On the other hand, if the evaluator is tenured, their job is secure and this increases their incentive to be honest about candidate quality and leads to better hiring.  If there are objective signals as in sport, there is less need for subjective evaluation and hence no need for tenure.

This is the crux of the idea.  It is patronizing for anyone to impose their tastes of Britney vs Bach on others.  Everyone’s opinion is equally valid.  It is possible to say Scottie Pippin was a worse basketball player than Jordan – the data prove it.    Science is somewhere in between.  There is both an objective component and a subjective component.  We then have to rely on experts.  Then, the experts may have to be tenured.