Dick,
I agree with what you say regarding fully powered systems, accepting design complexities such as those described for the B737 et al.
I think that the way to consider what is happening with power assisted elevators is to start off considering manual elevators on variable incidence tailplanes. The zero hinge moment elevator angle, for a given angle of attack, will be a function of the downwash angle of the airflow off the tailplane. This will be affected in particular by the camber of the tailplane section but also by tailplane aspect ratio, taper ratio etc. Therefore, even at the zero lift angle of attack of the tailplane, the elevator angle may not be zero. This will also be influenced by the reduction in the influence of the wing downwash as the airflow travels rearwards.
Now, if we go to a power assisted system, we are somewhere between the two cases above, depending on the amount of power assistance and the artificial feel system.
Out of interest, the elevators on all marks of Hunter were actually power assisted as the stick was permanently connected to the elevators (and ailerons). However, the hydroboosters effectively provided all of the force to overcome the hinge moments. The spring feel bias had to be adjusted post airtest to minimise the trim change following reversion to manual, thus showing that the artificial feel influenced the elevator angle when in trim.
Rgds
L
PS. I remember some interesting discussions with you when I was doing my ATPL exams (Bristol, Dec '98)