Maybe - I should X-ray the lens but I do not want to damage its chips
Wasn't the old 24-70 2.8 EF lens designed that the lens hood worked perfect for all focal lengths
by shifting the (effective) principle plane for 24mm to the front? The effective aperture should be located
in or close to the principle plane.
The EF 24-70mm f/2.8 L is a retrofocus design from one end of the focal length range to the other. This means the rear parts of the lens are like a telephoto design turned around backwards.
The lens is fully extended at 24mm and fully retracted at 70mm. This is unlike many designs with focal length ranges that start out shorter than a mount's registration distance but go past the registration distance by the longest focal length in the lens' range (such as all of the 18-55mm variants that start out retrofocus at the wide end but move some groups in one direction while moving other groups in the opposite direction as the focal length is increased and are not retrofocus by the time they are at their longest focal lengths).
The hood attaches to the main barrel as the secondary barrel extends and retracts inside the stationary hood.
This means the front element is much nearer the front of the hood at 24mm and furthest from the front of the hood at 70mm.
I'm not sure if the hood works "perfectly" at every focal length, but it certainly provides much narrower coverage at 70mm than at 24mm.
I still use mine regularly.
As to the "effective aperture" (more properly called the entrance pupil), its location does not correspond to the physical location of the diaphragm any more than its size is the same as the physical diaphragm. The size and location of the entrance pupil are determined by the effect of all of the lens elements between the physical diaphragm and the front of the lens. The size and location of the exit pupil are determined by the effect of all of the lens elements between the physical diaphragm and the rear of the lens.