That's debatable for the vast majority of them. "Clearly" is a strong word here. And certainly, regardless of how "good" they are, almost none are worth the hit in performance in the most basic of tasks for me.
Emacs Lisp is much less performant than a modern Javascript implementation. I don't see how choosing JS as the extension language raises any serious performance issues.
Extension quality is a bit subjective, sure. But if you are going to claim that emacs extensions are of higher quality because they're written in a more obscure language, then you should give evidence. That is not my experience.
If you weren't referring to the performance of plugins when you talked about the "hit in performance", then you are changing the subject from the quality of plugins to overall editor performance. I didn't say anything about that.
That was my point all along. The people that write extensions in this case are also the people "hacking" on the editor itself, and saying they (or web "devs" in general) are not performance conscious is an understatement to say the least.