What you can say is that if you have the THEORETICAL case with the surface having no fuse, prop, sideslip effects and balanced about the 1/4 chord point where theoretically the moments from deflection are zero then it would take a very small servo to move the surface a few degrees about the zero deflection setting.

Move to the real world with wind downwash, fuse effects, sideslips effects, etc. then you have to account for a lot of variable airloads on the surface which would require a larger servo.

The torque required in most model applications is really unknown, we just throw a large servo at it and see if it works:-) I doubt anyone has sat down and calculated the exact servo torque requirements on any airplane. We try someting, it works and so we make that the requirements in the kit. Certainly you could progressively try smaller servos until you found the limit of control effectiveness but that is a lot of work.