Speech is rated as heavily accented and transcribed with complete accuracy. Why is that not a contradiction on Munro and Derwing's framework?

B. Intelligibility and accentedness are partly independent, so a strong accent need not impede understanding

That the dimensions are partly independent. Perfect transcription means intelligibility was complete; a high accent rating means the speech was audibly non-local. Both can be true at once, and separating them is what allows a learner to aim at being understood rather than at sounding native.