If you take the stance that in-person work is more effective (an "uncomfortable truth" as you call it), then it would seem to reason that engineers who transferred to remote work would see a decrease in performance, but that is not the case as you also mention.

It seems like the real issue may be the hiring process between remote and in-person engineers, or biases in evaluating in-person vs. remote engineers.