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.
The real issue may be one of the ones you mentioned. It could also be that employees who spend at least some significant time in person are more effective than those who are and always were fully remote.
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.