I would say that the total contribution of the momopole shift remains positive as long as the effect of epsilon remains small compared to the total nuclear charge. The model is used for this situation as electrons don’t spend all their time inside the nucleus.