Interesting. I can see it both ways - just a moment around an axle centerline......but also the same vector at different distances from a trailing radius arm pivot point. Hhhhhmmmmm.....still confused.
Mine comes from a static analysis of the forces. Whatever force the caliper absorbs is translated into a torque (between the point of pressure of the pad on the rotor) and the axis of rotation (center-line of the wheel). That torque is balanced by the opposite reactions at the upper and lower hub connections. The braking torque is the same (direction and magnitude) whether the caliper is mounted on the front or back.
I just looked in my several books on suspension design, hoping to find a nice quote. None of them even mentioned any preferred location of the caliper.
I think most cars have been configured for trailing calipers in order to more readily allow the cooling of the discs.
In regard to live axles - General motors seems to also think that torque reaction as well as braking forces react differently dependant upon the side of the car. Noting that many have shock absorbers staggered with one side in a leading configuration and the other side trailing.
While another kettle of fish -
I’ve worked on stock cars (hardly “stock”) which had the rear brake calipers mounted on floaters with links to the chassis. This would load the chassis with a verying degree of weight in each corner and could be tuned.