Thanks for the vids. I'm also involved in some cardboard aided design in my build thread. I have no special training. I just hang around this forum. And it looks to me like there is a larger than stock recirculation zone on the glass in clip #1 and that that zone grows very significantly in clips 2 and 3 with the kamm. I would think these all indicate down stream turbulence, but then the flow seems attached at the hitch box. Still couldn't that attached flow be just the air getting forced around the box, with turbulent wake above it and downstream of the hitch box?
Can you bring the hitch box closer to the rear bumper and close the gap, especially at the top? I bet their is a sort of bow wave at the front of the hitch box caused by the air rushing down from the roof and in from the sides down into that gap at the tail for the trunk.
EDIT: It occurs to me that if all you end up with is a cargo box that adds no big drag but does not reduce it, either, then you are still ahead of the roof mounted units.]