Three questions:
1-- Do you know which tolerancing standard would be governing the drawing? (If you're in the US, it is most likely ASME Y14.5.)
2-- Are you concerned about the two choices simply based on ease of inspection, or might there be a functional concern that "drives" the decision?
3-- From the functional point of view... is the roundness of the bores (form) just as important as the alignment? This would be the key question to swing the vote.
I can tell you that total runout is a more stringent control. But it is often the easier one to measure, using the dial indicator method. Concentricity is sometimes harder to measure, but it actually is more forgiving (because the bore could be oval-shaped yet still be concentric).
The fact that the part will be measured with a CMM shouldn't matter. I don't think the part literally has to be spun; the computer can do all the math as long as sufficient points are sampled.