The Enduring but Unwelcome Role of Party Intent in Treaty Interpretation

Party “intent” is not one of the tools that the Vienna Convention on the Law of Treaties (VCLT) gives to treaty interpreters.¹ To be sure, party intent is presumably reflected in the “object and purpose” of the treaty, but it is not a separate criterion; in fact, the VCLT implicitly excludes party intent from playing an interpretive role. Yet many decision-makers, counsel, and academics persistently look to party intent for guidance when interpreting treaties. The most favored nation (MFN) debate illustrates why party intent endures as an interpretive touchstone: treaty language, even when analyzed in context and in light of the convention's object and purpose, does not always lead to clear answers. Both Simon Batifort and J. Benton Heath ² and Stephan Schill,³ in their different ways, depart from traditional VCLT analysis and hark to party intent as a reason to endorse a modified approach to treaty interpretation. Yet they also illustrate why party intent is an imperfect tool: party intent is too malleable to be a conclusive guide to treaty meaning.

Batifort and Heath describe the İçkale v. Turkmenistan ⁴ tribunal's approach as a welcome instance of “bottom-up” reasoning that pays close attention to differences in treaty text, contrasting it favorably with other tribunals’ propensity to engage in what they term “top-down” reasoning.⁵ In response, Schill challenges their assessment that a “top-down” approach exists.⁶ While Schill does not dispute that MFN clauses must be interpreted individually according to their specific terms, he also highlights the importance of the international law background against which the MFN principle has developed, which includes an emphasis on the multilateralizing function of MFN clauses.

Whither the Vienna Convention on the Law of Treaties?

Dissatisfaction with the VCLT seems to be a common feature of both articles. Neither dismisses the Convention,⁷ yet neither suggests that proper VCLT application will provide the solution the authors seek.

The VCLT is, to be sure, an imperfect instrument, and its application in practice is fraught with potential missteps. The interpreter is supposed to analyze in good faith the ordinary meaning of the treaty's language in context and in light of the treaty's object and purpose as if the process were occurring in a crucible—all three are to be considered simultaneously, and no one of them is to be given precedence over the others.⁸ Furthermore, the experience and expertise of the interpreters is almost certain to have an effect on interpretation. I have elsewhere discussed the effect of a decision-maker's interpretive community (or communities) on her approach to treaty interpretation.⁹ As Michael Waibel has similarly concluded, “The meaning of international law norms hinges on background principles shared by interpreters who form part of one or several interpretive communities.”¹⁰ Nevertheless, debating treaty interpretation without attention to the VCLT's strictures unmoors the authors’ methodologies from applicable law, leaving them drifting (problematically) toward party intent.

It is not altogether clear that Batifort and Heath's “bottom-up” and “top-down” nomenclature comports with ordinary treaty analysis. To put those categorizations in standard VCLT language, “bottom-up” reasoning appears roughly to refer to the ordinary meaning of language in the treaty in question, while “top-down” reasoning appears to refer to the presumptive “object and purpose” of the treaty provision in question, although not of the treaty itself.¹¹ Yet neither the “bottom-up” nor the “top-down” approach takes into account the VCLT requirement that interpretation be done in context, with context extending to the text of the treaty, “including its preamble and annexes.”¹² In other words, if the “top-down” approach risks losing the trees for the forest, the “bottom-up” approach risks losing the forest for the trees.

To Schill, on the other hand, “interpretation of MFN clauses is based on multilateral rules and includes multilaterally accepted interpretative presumptions.”¹³ The VCLT, however, does not discuss “multilaterally accepted interpretative presumptions.” This omission does not mean they are not important or that they should be discounted. They are, arguably, an indelible feature of the interpretive process. Yet employing these presumptions arguably extends “context” much further than is compatible with the VCLT.

Schill also confidently argues that MFN clauses, even in a bilateral context, are intended to perform a multilateralizing function: “[T]his commitment in bilateral treaties to multilateralism is what states intended to make by including MFN clauses in their investment treaty relations.”¹⁴ Again, the VCLT does not include a focus on party “intent.” This was a deliberate omission in the VCLT, which sought to eliminate the subjective intent of the parties to the treaty in the context of treaty interpretation and to focus on the language actually in the treaty, as assessed against the object and purpose of the treaty (again as expressed in the treaty itself).¹⁵ Party intent only plays a role if it can be established that the parties meant a particular meaning be given to a term, which presumably would otherwise be interpreted differently because of the “ordinary meaning” typically attributed to it.¹⁶

This does not mean that the VCLT has been successful in eliminating a focus on party intent. As Abby Cohen Smutny and Lee Steven have written in the context of investors’ use of MFN clauses to avail themselves of more favorable procedural provisions, “the common touchstone in all of the cases discussed here is the premise that the intent of the Contracting Parties to the treaty containing the MFN clause is paramount in determining whether an investor can rely on the clause to benefit from the dispute settlement provisions of other treaties.”¹⁷ But it is tricky to ascertain the intent of the treaty partners, particularly when one of them is acting in the guise of respondent and might have strategic reasons for asserting its intent in treaty-drafting. As Sir Frank Berman has observed,

[The tribunal] can clearly not discount assertions put forward in argument by the Respondent as to the intentions behind the BIT [Bilateral Investment Treaty] and its negotiation (since that is authentic information which may be of importance), but it must at the same time treat them with all due caution in the interests of its overriding duty to treat the parties to the arbitration on a basis of complete equality (since it is also possible that assertions by the Respondent may be incomplete, misleading or even self-serving). In other words, it must be very rarely indeed that an [International Centre for Settlement of Investment Disputes] Tribunal, confronted with a disputed issue of interpretation of a BIT, will accept at face value the assertions of the Respondent as to its meaning without some sufficient objective evidence to back them up.¹⁸

Whether MFN clauses are in fact intended to serve a multilateralizing effect might or might not be true in any given case. It is also plausible to argue that MFN clauses serve the more limited purpose of protecting the concessions gained by one treaty partner in negotiations with another. If A offers B access to its steel sector, but then offers to C access to both steel and iron ore, B's deal is suddenly less valuable. An MFN commitment offered to B ensures that B gets at least what C gets. In this scenario, multilateralization is a result, but not necessarily a goal, of an MFN clause. In other words, B would likely prefer to remain on bilateral terms with A while C gets nothing. B does not want to multilateralize but accepts that multilateralization is the price it pays to prevent itself from being undercut. Perhaps this caveat makes no difference in outcome; the MFN clause still performs a multilateralizing function. Yet that outcome might more persuasively be based on treaty text rather than on party intent.

The malleability—and unreliability—of intent as an interpretive touchstone is illustrated by the Free Trade Commission (FTC) Note of Interpretation negotiated by Mexico, Canada, and the United States. That Note expressly limited Article 1105 of the North American Free Trade Agreement (NAFTA) to providing the international minimum standard of treatment under customary international law, and expressly categorized fair and equitable treatment and full protection and security as subsets of that minimum standard. Did they intend this carefully negotiated language to be displaced by use of NAFTA's MFN clause (found in Article 1103)? A reasonable answer would be “no.” No doubt it would have been desirable for the NAFTA parties to have made clear the relationship between the Note of Interpretation and the MFN clause, but it is at least arguable that they did not believe that they needed to do so. Why would they have carefully negotiated an agreement clarifying (and limiting) the meaning of Article 1105 if Article 1103 could be used to extend the language in that provision? On the other hand, they are sophisticated negotiators who surely knew, or should have known, of the potential effect of Article 1103. Had they intended to exclude the possibility, why did they not do so?

For more than fifteen years, NAFTA tribunals have avoided deciding these conflicting views of party intent by evading whether NAFTA's MFN provision permits investors to rely on an unfettered fair and equitable treatment standard.¹⁹ This example makes clear that references to party intent do not provide additional interpretive purchase and may instead raise new questions.

Calcification of Treaty Language

The focus on intent recognizes the centrality of state power over treaty formation, but also illustrates, albeit indirectly, the paucity of tools that treaty interpreters have to ascertain a single, “true” meaning to attribute to a treaty term. It is less easy for states to amend or “clarify” existing treaty language than Schill suggests.²⁰ Again the NAFTA parties’ experience is illustrative. The NAFTA FTC, comprised of the trade ministers of the three NAFTA parties, has the power to issue notes of interpretation of the agreement. First, this is a political process. The trade ministers are not civil servants; rather, they are political appointees with a multifarious portfolio. One challenge, therefore, is to reach agreement about putting a potential interpretation on their agenda amid the multiple demands on their time. Further, all three countries would need to see this as a priority simultaneously. Absent an impending crisis, this is difficult. For more than fifteen years investors have sought to use the MFN clause to displace the Note of Interpretation relating to Article 1105, and each time they have failed. It is thus hard to argue that an incorrect use of Article 1103 (assuming it is incorrect) is threatening the proper functioning of NAFTA Chapter 11 or otherwise urgently requires the attention of the NAFTA trade ministers.

Moreover, the political nature of the process means that other priorities or desires are likely to be introduced alongside the potential change to treaty language. In the NAFTA context alone, perennial problems such as immigration through Mexico are likely to play a role in trade negotiations. Some trade agreements are meant to bolster economic influence in strategically sensitive regions; small disagreements about treaty text are unlikely to be addressed if by doing so the states parties would derail larger security goals. Multiple other political problems or timing issues could derail an interpretation initiative in a similar manner.

Second, note the limited power given to the FTC—it can issue a note interpreting the agreement, but not amending the agreement. Given the language of Article 1103 and the possibility to argue that in context it should be limited to internal measures rather than extended to dispute settlement or substantive NAFTA provisions, perhaps an interpretation would be in order. But scholars roundly criticized the NAFTA parties for their interpretation of Article 1105 on the grounds that it was not an interpretation but instead constituted an amendment of the agreement.²¹ The FTC is understandably loath to incur more opprobrium.

While limited in scope, this particular debate draws attention to problems currently vexing treaty negotiators, who are operating against a backdrop of thousands of similar treaties, including sometimes dozens of treaties concluded by the same state, which might or might not follow the text of a model treaty.²² What conclusions should one draw when states make explicit that which was (arguably) implicit, such as the addition of “like circumstances” to a hitherto unadorned MFN clause? Does the inclusion of “in like circumstances” really change the analysis and even the scope of the provision? Should it? Should we place the change against the backdrop of the practice of the states parties to the treaty? If A's international investment agreements had all featured “in like circumstances” language before, but B's had not, does the inclusion of “in like circumstances” in the new A-B treaty mean that B has changed its practice? Does it simply mean that A's treaty language ended up “winning” that particular battle? The NAFTA parties’ approach of generally retaining existing language but adding explanatory provisions or annexes “for greater clarity” risks leading to investment agreements the length of War and Peace. Yet alternatives are hard to find, given the assiduity with which counsel and arbitrators compare and contrast treaty language to discern party intent.

Conclusion

Batifort and Heath's and Schill's contributions show that we need an interpretive approach that takes into account individually negotiated treaty language but does not have a stultifying effect on treaties by overemphasizing minute differences in text. The existence of interpretive communities and the continued focus on intent illustrate the difficulty of achieving objective treaty interpretation reliant solely on language found in the treaty and other agreements. Treaty language and even subsequent agreements or state practice do not always enable us to capture accurately the meaning of the treaty, and state practice in amending treaty language is a slim reed on which to lean. But neither can party intent solve these interpretive challenges.