pkjo: ah, that's a *workflow* question.
(incidentally, it's completely orthogonal to the question of whether you work with 1 layer/frame or 1 image / frame.)
What you describe is a very common process among animators, because that is a pretty effective method.
I used it to make this yesterday:

however, it has larger stages. As a rule, first the key frames are done, then the same process is used to fill in in-between frames, and once the whole animation is looking good, it's rendered (previously being just outline, thumbnails, silhouette or flat colors)
If you're talking about multiple independently-animating objects, these are usually individually animated then inserted into the animation as an animation (on their own layer or similar.). I definitely recommend something advanced like GIMP-GAP for such an application. Layers and comfortable support for 'sub-animations' are a must for such projects.