Your brain is a supercomputer and your body is an adaptation machine. Which is to say I am not exactly sure what the brain and body experience and compute at different setbacks and crank lengths. But I trust in their ability to make good choices if a proper process is adhered to.

Both setback and crank length effect the acute hip angle, or thigh-torso angle, at the top of the stroke. Saddle forward or crank shorter have the same vector of effect, but different magnitudes. However, there are more factors at play when modifying setback, as it effects muscle group utilization, and this is probably a larger driver. I couldn't tell you what percentages of glute, hamstring, or quad to involve in your pedal stroke, but that is exactly what your supercomputer is doing, so I don't have to. Your body is capable of choosing a setback, and I think you should do that first.

You play with crank length after setback is determined, as your setback can perhaps effect your preferred crank length, but a change in crank length isn't likely to change your preferred setback.

The direct answer to your question is yes, if you move your saddle back and close the hip angle to a large enough degree, you may need a shorter crank to compensate. This assumes everything was optimized to begin with, in which case why are you changing setback anyway? Or everything wasn't optimized, so why didn't you find out your setback first?

Finally, and in a practical sense, I think you would have to make a fairly large change in setback to necessitate a change in crank length. I would not try to micro manage crank length for small changes in setback.