From a 1998 paper by Charles A.E. Goodhart (a great read, by the way):
One possible rationale is that the [Mengerian-form] theory was never meant to be a positive, explanatory theory, but instead a normative theory, of what should be. As one referee commented: ‘‘OCA theory is a normative, not a positive theory.’’ A weak form of this would be to recognize that, in practice, monetary institutions are inherently and au fond associated with considerations of political sovereignty, but that the subsidiary function of [Mengerian-form] OCA theory is to assess the balance of purely economic benefits and costs that this may generate. The problem with this is that the historical record of the association of money creation with the establishment and maintenance of a stable sovereign power is so overwhelming (apart from the case of tiny, and by the same token politically weak, states) that the balance of purely economic benefits and costs entailed by OCA must presumably be of second order importance.