A decision boundary named once feels solved. It is not. What actually holds it is relitigating it visibly, with the team watching, until they can run the test themselves.
A rating is supposed to describe what a person did. Increasingly, it describes the budget, and gets handed to the person as if it were about them.
New hires aren't pricing a cut onboarding budget. They're pricing what their manager does with what's actually in the manager's control, and they read it correctly, fast.
The decision everyone treats as settled is not safe just because nobody questions it. If nobody left on the team can explain why it was made, it is already a live risk.