The Peter Principle

Do firms systematically mispromote workers?

The Peter Principle is the idea that everyone will eventually be promoted to their level of incompetence. If multiple tiers of jobs exist, and wages within a tier are relatively fixed, then companies will offer promotion to the next tier as a way to incentivize additional effort. If the jobs are substantially different in their skill requirements, then people will be promoted until they reach the job where they are no longer capable, where they stagnate.

Some economists are very skeptical of the idea. Firms will anticipate that this will happen, and adjust their behavior such that there’s no effect. Skills at different jobs might all be positively correlated, even if seemingly unrelated, if job performance is due to some underlying variable like intelligence. Edward Lazear (2004) launched a subtle attack on it, describing it entirely as statistical error. Suppose that ability is realized with some randomness — somebody might do an exceptional job, but they just got lucky. Even if skills from job to job are identical, people will regress toward their average ability. This is the same reason why sequels will tend to be worse than the original. The first was somewhat lucky to have a great idea, so we can’t expect the second to be as good.

There are good theoretical reasons to think that the Peter Principle is in fact true, though. Promotions are partly a way to sort people into jobs, and partly an incentive scheme. Fairburn and Malcolmson (2001) suggest that the preponderance of promotion incentives, rather than bonuses, is due to the misalignment of manager incentives. Bonuses are given at the discretion of those best able to evaluate them. If managers favor some people over others for reasons unrelated to things which benefit the company — for instance, if they are having an affair with them — giving them a promotion requires them to leave the department.

Perhaps most importantly, there is strong evidence that it happens. A recent article in the QJE, “Promotions and the Peter Principle”, by Benson, Li, and Shue, is, to my knowledge, the only paper credibly evaluating the hypothesis. They examine salespeople across 131 firms, and isolate the value add of a manager by controlling for their subordinates, month, company, and year. This is negatively correlated with prior sales performance, as expected. To show that this is due to discrimination in favor of workers with better sales, they use a “Becker outcomes test”. They isolate only the “marginal” promotions by comparing differences in promotions between high-demand and low-demand periods for that particular firm. You can show that they’re applying lower standards for workers with good sales performance. Note that they cannot show that this is not optimal — simply that there is discrimination.

The inefficiency of promotion-based incentives, as compared to an ideal world, suggests a mechanism through which pay transparency is harmful. There are strong norms against paying different people in the same job different amounts, even if they are more productive. If a firm cannot back up its pay differences with objective criteria, it leaves itself at risk of expensive litigation if the pay disadvantages a protected group. No such stigma against paying different jobs different wages exists, so a firm looking to incentivize work through pay will promote people, even when the social optimum would simply be an increase in pay. Cullen and Perez-Truglia (2022) conduct a field study on a large company in Southeast Asia, and find that learning about managers making more than they expected increased effort and productivity, while learning about people at the same level making more actually decreased effort. Zoë Cullen followed up that with an article arguing against pay transparency generally — it makes pay more equal through reducing wages overall. The desire for workers to pull each other down to the same level, like crabs in a bucket, is stronger than any change in market power. Far better to have pay transparency across firms, than within firms. Promotions over bonuses are also incentivized by statute. It is simply easier to win a suit alleging discrimination by race or by gender when the suit details unequal pay for the same work.

The Peter Principle — that people are promoted for their prior success, not their future prospects — is definitely real. Business owners should take heed. What policymakers can do, however, is not as clear. This may be the best we can do, in an imperfect world.