226 – In praise of random promotions

Posted on March 28, 2013 by


“Past performance is no indicator of future success” is a warning attached to many financial products that are hawked in the Street. Few people pay any attention to the warning. In fact, “past performance” is the basis of meritocracy. The (predictable) outcome is the Peter Principle. The Peter Principle is a proposition stating that the members of an organization where promotion rests on achievement, success, and merit will eventually be promoted beyond their level of ability. The principle is commonly phrased, “Employees tend to rise to their level of incompetence. (…)” in time, every post tends to be occupied by an employee who is incompetent to carry out its duties” and adds: “work is accomplished by those employees who have not yet reached their level of incompetence.”[1]

We have a case of substitution (see my 225). The difficult question is to determine whether the person is fit for the job. The simpler question is to verify that he was good in his last job. Presto, switch past performance for future success as the selection criterion: alas, in the long run incompetent managers take over. This applies also to politicians, of course.

What is one to do? Laurence J. Peter, after whom the principle is named, proposed establishing an elite fast track as “second best” solution.[2] This is voodoo management to me: why should restricting the pool of candidates yield better outcomes? How is one to determine “brilliance” – at a young age? Why should career security be reward for “brilliance”? In fact, the effect is likely to be perverse: the caste will have an easier time protecting its collective incompetence.

Italian researchers have taken a different tack. Their solution is to expand the candidate pool: they propose promoting by lot. They used an agent-based modeling approach to simulate the promotion of employees and tested alternative strategies (please note their search for evidence rather than plausibility). Although counter-intuitive, they found that the best way to improve efficiency in an enterprise is to promote people randomly, or to shortlist the best and the worst performer in a given group, from which the person to be promoted is then selected randomly. This work won the 2010 Ig Nobel Prize in management science.

In a recent article,  they expanded the concept to Parliamentary elections.[3] Party politics favors partisan law-making at the expense of the “common good”. The “winner-take-all” attitude we encounter in politics today reflects this attitude.

The authors propose introducing “independent parliamentarians” drawn by lot. They do not have to be angels and vote only for laws that enhance the common weal. We want their independence to destroy the effect of party politics, where “right or wrong my party” prevails. How many such parliamentarians are required? The larger is the winning party, the larger must the number of independent agents be. What is interesting, however, is that the effect is non-linear and depends on circumstances. A few independent parliamentarians suffice to make a big difference, provided the majority party has no stronghold on Parliament.

The ancient Greeks and Romans understood the constructive role of chance. We still find traces of it in the jury system. We should embrace it too. We have one advantage over our forefathers: we can simulate the outcome.

[1]           http://en.wikipedia.org/wiki/Peter_Principle (also for other quotes in this blog)

[2]           “The prospect of starting near the top of the pyramid will attract to the hierarchy a group of brilliant [higher class] employees who would never have come there at all if they had been forced to start at the bottom.” Thus he concludes that the hierarchies “are more efficient than those of a classless or egalitarian society.”

[3]           Alessandro PLUCHINO . Andrea RAPISARDA – Cesare GAROFALO, Salvatore SPAGANO, Maurizio CASERTA (2013): L’efficenza del caso. Le Scienze, January 2013.