The Oxford Guide to Treaties Symposium: Can Severability Be Made Less Severe?

[Ed Swaine is Professor of Law at George Washington University Law School]

I’m pleased that the subject of reservations, which is near and dear to my heart, is attracting the attention of such esteemed commentators. The illuminating comments by Professor Stewart and Dean Koh, which I had the chance to read while preparing this post, focused in part on the severability solution, according to which impermissible reservations can at least sometimes be disregarded so as to leave a reserving state as party to the entire agreement.  Their comments, which query the compatibility of this approach with state consent and with securing Senate consent to future agreements, may be misperceived as resulting from parochial concerns – particularly in light of objections to the U.S. reservation to the Convention on Certain Conventional Weapons, in which some states likewise asserted the capacity to disregard the U.S. reservation and hold it to the “entirety” of the Convention.

I think the objection to the solution goes deeper, for reasons I could only touch on in my chapter.  ILC Guideline 4.5.3 would create a presumption that impermissible reservations are severable unless it is shown that the reserving state viewed the reservation as a sine qua non of its consent (as opposed to, say, a half-hearted ask or an attempt to yank another state’s chain).  Unless the matter rises to the legitimate attention of a treaty monitoring body or an international tribunal, this puts two inquiries to states trying to evaluate another state’s obligations – first, whether the reservation is impermissible because it violates the treaty’s object and purpose; and second, the psychological conditions of another state’s consent – which must also be anticipated by the reserving state in its turn.

This does not seem very workable, and the reason is in part because of the (ultimate) indulgence of the consent principle.  One might defend the first inquiry as potentially objective, though states are bound to differ on a treaty’s core values; the VCLT wisely pretermits the inquiry’s relevance to the extent it opens all reservations to objection on any grounds. The second inquiry, though, is not only subjective, but foreign to the responding state.  Still, it would have been open to the ILC to defend this presumption of severability as an information-forcing default rule – one that makes the reserving state disclose, via the reservation text or otherwise, its intent to insist on the reservation as a condition of adherence.

It is at that point, though, that the Guidelines seem to run out.  The opposability construct of the VCLT, to which the Guidelines are somewhat hostile, works after a fashion only because the call-and-response of reservations and objections establishes a shared understanding through what look like unilateral vetoes: first a veto by the reserving state of the agreed treaty content, then a veto by any objecting state to the reservation’s effect against it (and, unfortunately, either to the reserved-to provision or the reserving state’s consent, which is not an appealing choice).  Guideline 4.5.3 instead supposes that states will mutually intuit an offense against a treaty’s object and purpose; then develop a shared view of the reserving state’s “reservation price” based on the totality of circumstances relating to its intentions; and last but not least, as a belated concession to the consent principle and to its relative expertise, permits the reserving state to “express at any time its intention not to be bound by the treaty without the benefit of the reservation.”

The ILC is to be commended for taking seriously the problem of reservations that are inimical to a treaty’s mission and for addressing the limitations of the objections scheme. And it is notable that it did so while (at least nominally) resisting the position that each state can decide for itself whether a reserving state remains a party to the entire agreement, having recognized the importance of a state’s consent to the initial set of terms it assumed. But permitting a state to end its obligations at any time, with potentially retroactive effect, on the pretext that its reservation is impermissible – well, that converts a problem of uncited violations into one in which a team can blow the whistle on itself or, in any event, pick up the ball and go home when it chooses.  The Guidelines probably anticipate more refereeing by treaty monitoring bodies, but that will elicit its own objections. In the meantime, it seems likely that the proposed compromise on severability will gain adherents only slowly, if at all, and in the meantime make the players even more uncertain of the rules.