Ground fault is typically resistive, but the inductive component comes into play, too. The NEC tables still apply- basically, you just calculate what current would cause 120V of voltage drop, and that's an upper limit for your fault current. The capacity of the source (transformer, generator, etc) comes into play, too, and there are some differential equations and all, but it's not hard to make conservative estimates. The important thing is that the ground wire be able to conduct current enough to trip the OCP before the wires explode, and that's what NEC is mandating.
250.122(B) doesn't seem to make that exception. Maybe it should? Or is there such an exception elsewhere?
250.122(B) only refers to the ungrounded conductors; the neutral is implicitely excepted. None of the other items in 250.122 appear to apply to an oversized neutral, either.