Certainly overall salaries of teachers are held down by the failure to recognize that some teachers are more effective than others.
No playing with refined statistics can make flawed data less so.
VAM (value-added measurement) is new to the scene, but as you know has been debunked for its unreliability and probably lack of validity by many others.
While citing a litany of potential flaws in testing, the opponents of teacher and school accountability often create a straw man that bears little relationship to the real policy discussion.
Whether scores rise or fall wouldn't influence me because I think they are phony measures.