Morality frameworks defined as a single rigid rule will always encounter some "repugnant conclusion" given a sufficiently contrived situation, news at 11.
The "retributive morality" the author espouses doesn't actually dodge "justification of family-punishment" either. It relies on the same external axiom, that family should not be held culpable for their members' crimes (and are thus categorized as "innocents"). The same assumption should cause consequentialism to assign an extremely negative valence to family-punishment, because the consequence of family-punishment is that "everyone lives in a society where family members are held culpable for their members' crimes, and live/act in fear of such largely-uncontrollable punishment".
Easy -- No one is ever punished for anything and crime gets to the point that society crumbles, because as humans we cannot enact a justice system that does not occasionally punish innocent people.
Slightly less easy, but closer to current reality: For "don't knowingly punish innocent people" -- enact a justice system that is terrible at fact finding (or has gross bureaucratic and procedural inefficiencies), so that you punish a lot of innocent people but never do so knowingly.
Easy again: For "don't punish innocent people, knowingly or otherwise -- you have no criminal justice system that is workable within these constraints in the real world.
The "retributive morality" the author espouses doesn't actually dodge "justification of family-punishment" either. It relies on the same external axiom, that family should not be held culpable for their members' crimes (and are thus categorized as "innocents"). The same assumption should cause consequentialism to assign an extremely negative valence to family-punishment, because the consequence of family-punishment is that "everyone lives in a society where family members are held culpable for their members' crimes, and live/act in fear of such largely-uncontrollable punishment".