Economists have measured exactly that.
In 2019, Benson, Li and Shue ran the first large-scale test of it, which was published in the Quarterly Journal of Economics. They tracked sales workers across 131 companies. Strong selling predicted who got promoted. And the stronger the selling beforehand, the worse the team did afterwards.
They were testing the Peter Principle. Laurence Peter and Raymond Hull named it in 1969: people rise on performance in their current role until they reach one they cannot do. Peter meant it as satire. The data turned the joke into a finding.
The cause is structural. One promotion is asked to do two unrelated jobs: it rewards excellent work, and it fills a management seat. Those need different skills. The numbers a company already has answer the first and say little about the second.
Benson, Li and Shue also put a figure on the cost. Promote based on managerial potential rather than past sales, and team output could rise by nearly 30%.
Google built its answer into its structure. It runs two parallel ladders, one for managers and one for engineers who would rather build than manage. The technical ladder climbs as high as the management one. Jeff Dean reached Google's top technical rank without ever running a large organisation. At most firms, the only way up is to manage people.
One honest caution. Promotion is also the carrot that makes people stretch in the role they hold now. Remove it, and the pull goes too. The work is to build a second way up, so a specialist is not made to choose between craft and career.
It reaches well beyond software. Promote the strongest buyer to lead the procurement team, and the team loses its strongest buyer.
A promotion answers two questions at once. Who did the best work, and who should lead? Treat them as one, and you lose your best people twice over. Separate them before the next decision.
