Well, in your parent comment you say "The problem is that we assume we need to test students, and not teachers." So, that does seem to be what you were saying.

Of course a teacher's supervisor should be in charge of evaluating them. But they must be competent to do so and they should do so through evaluation of individual lesson plans and classroom observation, not the performance of the teacher's students on a standardized test. That's not a straw man by the way, there are a lot of people in this debate who think that's an acceptable metric for making decisions on compensation and hiring and firing [1]. Whether you are one of them, I have no idea. But the argument you make above is one I have encountered many times, and typically its adherents are quite myopic about the culpability of everyone else in the system.

Of course there should be a way to evaluate teachers and hold them accountable. And there should be a way fire them if it's called for. But it has to be equitable. And it shouldn't hold teachers hostage to factors they don't control. But even granting you that, I haven't seen any evidence that this is going to lead to much an improvement in the educational system. There are much bigger problems here than "bad teachers".

[1] http://www.epi.org/publication/bp278/