1. Introduction

The Wallonian incident[1] arguably represented the height of controversy concerning the introduction of an investor-State dispute settlement (ISDS)[2] mechanism within the Comprehensive Economic and Trade Agreement between Canada and the European Union (CETA).[3] A few days before the date agreed for the signature of the treaty, Belgium has created a standoff, after the Parliament of Wallonia threatened not to approve the signature of the agreement, if the ISDS mechanism was to remain therein. Notwithstanding the innovations introduced by CETA, States’ concern to see their policy space restricted and domestic adjudicatory jurisdiction over investment-related issues removed had notably permeated deliberation over the approval of the treaty. The exclusion of such investment dispute-related clauses from the provisional application[4] of CETA temporarily shelved the debate: Council Decision 2017/38/EU has sanctioned the provisional application of most of the treaty, leaving out (among other provisions) Chapter 8, Section F regulating the ISDS mechanism. However, renewed discussions are expected in conjunction with EU Member States’ domestic procedures for the ratification of CETA.[5] Such discussions are all the more likely to arise in light of the recent Opinion of the European Court of Justice (ECJ) on the Singapore-EU Free Trade Agreement (FTA) confirming the shared competence of the EU and its Member States over the investor-State dispute settlement mechanisms.[6] Amongst others, Belgium explicitly reiterated that it would not ratify CETA if the ISDS is to remain in text as currently devised.[7] This prediction is further evidenced by the recent claim brought by more than one hundred French Members of the Parliament to the French Conseil Constitutionnel on the compatibility of CETA (including the ISDS mechanism) with the Constitution, which has revived concerns on the eve of the provisional application of the treaty.[8] These recent developments give rise to the risk that EU Member States might effectively delay or close off ratification procedures,[9] eventually leading to a situation where CETA does not enter into force.[10]

This article explores whether the Law of Treaties offers tools to accommodate States’ trepidations on the ISDS mechanism, while preventing a stalemate. Should CETA be provisionally applied,[11] the most obvious choice would be to modify the treaty via the amendment procedure provided in the text of the treaty itself.[12] Even so, the resort to the formal amendment procedure raises at least two concerns. First, it is likely that such an amendment procedure will be politically controversial, especially in the countries that have already ratified CETA.[13] These countries may find it difficult to justify reopening the debate at the domestic level to discuss modification to certain provisions right after reaching an internal consensus on the ratification of the treaty. Additionally, the formal amendment procedure might also be perceived as procedurally inconvenient and cumbersome, in that it requires the fulfilment of additional respective internal requirements, which might include[14] doubling the number and timeframe of domestic processes and most probably new parliamentarian debates.

These considerations warrant an examination of alternative and less controversial instruments for the modification of CETA. Therefore, this contribution seeks to examine the processes by which revision of the CETA ISDS provisions[15] might take place. In short, it focuses on how the text of the treaty can be altered. As such, this study complements proposals for substantive reforms of the existing text of CETA (ie what should be revised), that have already been largely called for by various actors, including States, civil society and scholars.[16] To this effect, section two presents an overview of the most significant challenges identified in relation to the provisions of the CETA ISDS mechanism and the implications of potential reforms. Drawing from the Law of Treaties, section three will analyze those ‘modification’ tools States have at their disposal. Section four and five will combine the results of the previous parts, in order to identify what type of treaty law instrument in the specific context of CETA is best suited to realize each substantive suggestion for reform.[17]


2. Sources of discord and suggested fine-tuning

While a discussion over potential amendments of substantive standard of treatment provisions in Chapter 8 (Investment Protection) may be highly significant,[18] this article will only take into consideration the investor-State dispute settlement mechanism, mainly for two reasons.  First of all, the ISDS has represented the main bone of contention so far; furthermore, in the same vein as NAFTA, also in CETA substantive standard of treatment may be modified via binding interpretations issued by the Joint Committee,[19] thus reducing the urgency for modification.

The ISDS has been subject to thorough reconsideration in the last decade, calling for a substantive reform of the system as it stands in most international investment agreements.[20] The main argument against the ISDS mechanism relates to the risk of investors’ claims against host States. Arguably, such claims could lead to a ‘regulatory chill’, ie the shrinking of States’ regulatory autonomy. Other arguments relate to more institutional and structural flaws inherent to ISDS mechanisms. These include (1) the absence of an appeal mechanism, (2) the non-permanent role of arbitrators, and (3) the perceived lack of transparency surrounding international investment arbitration.[21] In response to these critiques, the EU and Canada have introduced in CETA a revised ISDS mechanism, namely the Investment Court System (ICS),[22] which endorses some of the suggested reforms, eg by introducing an appeal mechanism and expunging disputing parties’ choice of arbitrators.[23] While it is beyond the scope of this analysis to elaborate on the already thoroughly catalogued general critiques of ISDS and substantive reform conveyed by the ICS, this section will address some of those challenges that have not been addressed in CETA’s improved ISDS mechanism. More specifically it will mainly deal with (1) the potential eradication of the ICS mechanism; (2) the absence of a rule on the previous exhaustion of domestic legal remedies and (3) the borrowing of procedural arbitration rules from systems extrinsic to CETA.

The first and most radical claim for modification is the tout court removal of the ICS mechanism from the treaty. States, in particular, still remain apprehensive about the CETA’s revised investment-State dispute settlement mechanism.[24] Their paramount concern is that it still lacks a significant connection with the domestic legal orders. More specifically, the ICS (along with more classic ISDS mechanisms) removes national courts’ jurisdiction over investor-State disputes. This is cause for concern, because often sensitive matters are left to adjudication within a system which does not fully resemble a proper adjudicatory body, let alone a permanent court.[25] Eliminating the ICS would call for a systematic interpretation of the other provisions of CETA which might be affected by this (fictional) effacement. In fact, such an eradication will also have a more general spillover effect. The immediate consequence would be that domestic courts would regain direct competence over investment-related disputes. Nonetheless, this competence would be constrained (at least from the point of view of investors) due to the explicit prohibition in Article 30.6(1): ‘Nothing in this agreement shall be construed (…) as permitting this Agreement to be directly applicable in the domestic legal systems’.[26] Thus, investors would not have the possibility to claim the substantive standard of protection included in CETA before domestic courts, since no provisions of the treaty may exert direct effect in national legal systems. This notwithstanding, should Parties decide to go ahead without altering the provision on direct effect, national jurisdictions would still be able to adjudicate investment-related disputes on the basis of domestic law, which should, in principle and by way of compliance with international law, conform to the treaty standard.

An alternate option to removing ISDS is the introduction of an obligation to exhaust domestic legal remedies before turning to the ICS.[27] The absence of a previous exhaustion of local remedies rule has been at the centre of the debate around the ISDS, which has further contributed to the reticence of States. The critique is articulated mainly in two arguments. Firstly, when dealing with investor-State disputes, arbitrators take decisions about the appropriate balance between the right of host States to regulate in the public interest and the private interests of the investor. Thus, distinct from a purely commercial arbitration, in investor-State disputes arbitrators act more as ‘public adjudicators’.[28] It has been argued that domestic courts are better equipped to assess and strike the appropriate balance between public and private interests (especially in concreto). National judges usually adjudicate taking into account the domestic legal environment in its entirety, thus contextualizing decisions in light of the overall coherence of the national legal system where the investment is exerting its effects. Secondly, traditionally, the exclusion of domestic courts’ jurisdiction was justified because of the potential bias of the domestic adjudicating bodies toward the host State, especially in the absence of an admittedly strong rule of law environment and stringent legal institutions. Assuming for the sake of the argument that the aforementioned rationale holds true, this is certainly not the case for Canada and the EU, since both systems have a long tradition of judicial independence and solid institutional environment.[29] Thus, the introduction of a rule on the previous exhaustion of domestic remedies would be a desirable addition in the investor-State dispute settlement system introduced by CETA. However, there are two provisions in Chapter 8 which might represent a stumbling block in the establishment of such a rule. In order to submit a claim to the ICS, Article 8.22(4) of CETA obliges claimants to withdraw or discontinue (f) or waive (g) any proceeding (pending) before either domestic or international adjudicatory bodies. This wording seems to go exactly in the opposite direction to what is prescribed by the previous exhaustion of domestic legal remedies rule, which, instead, requires claimants to activate (and exhaust) national proceedings before resorting to the ICS.[30]

Another problematic element of the ICS that has been mainly highlighted by scholars and courts[31] relates to the significant referral the provisions of CETA make to already existing arbitration rules and procedures extrinsic to CETA. While claiming to represent a breakthrough in respect to the traditional ISDS system, the deference to existing rules regulating investor-State dispute settlement[32] casts a legitimate shadow of doubt on the extent of purported innovation, suggesting that the new system ‘cloaks arbitration in judicial robe’.[33] This absence of tailor-made procedural rules for the ICS adds yet another criticism to the functioning of the CETA investor-State dispute settlement. Article 8.23 of CETA allows for submission of claims under ‘a) ICSID Convention and Rules of Procedure for Arbitration Proceedings; (b) the ICSID Additional Facility Rules if the conditions for proceedings pursuant to paragraph (a) do not apply; (c) the UNCITRAL Arbitration Rules; or (d) any other rules on agreement of the disputing parties’. Notwithstanding the significant advancement in the sense that disputing parties cannot choose their arbitrators, a significant level of party autonomy is retained due to the fact that they can still select arbitration rules regulating the dispute or any other set of rules on the initiative of the claimant.[34] Further, one must assess the margin for potential development within the treaty, considering that the text does not allow for departing from existing rules and procedures. While Article 8.23(6) specifies that these rules (ie arbitration rules) apply ‘subject to the specific rules set out’ in Section F, Article 8.44(3)(b) clarifies that the Committee on Services and Investment can adopt or amend rules only supplementing those arbitration rules and procedures included in Article 8.23(2). This implies that the Committee cannot adopt ICS rules of procedures that do not conform to those listed therein. The same reasoning applies to the grounds of appeal delineated by CETA which operate a renvoi to both the ICSID and New York Convention. Thus, more than a borrowing of rules, it seems that these referrals establish an inter-treaty mechanism which makes ICSID rules ‘(in)directly applicable under ICSID’s own terms’, casting a shadow over the legal feasibility of some of these provisions.[35] The adoption of independent ICS arbitration rules would certainly ease many of the issues caused by this referral system. At the same time, such an adoption would prima facie be inconsistent with the text of the treaty allowing for supplementing additional rules only.[36]

This brief overview shows that while some of the substantive reforms mentioned above might qualify as ‘mere’ specifications of existing relevant CETA provisions, others go explicitly beyond the text of the treaty. Such circumstance calls for caution in the identification of the type of instruments needed to accommodate substantive suggestions.

The next section will introduce potentially relevant tools offered by the Law of Treaties to achieve this aim.


3. Modes of treaty ‘alteration’: The potential of ‘subsequent agreements’ under Article 31(3)(a) VCLT

It is well known that after the authentication – which has already happened for CETA – the text of the treaty becomes definitive and cannot be changed. This would mean that any substantive modification suggestions made by States would not be accommodated before States have expressed their consent to be bound by the treaty. Article 40 VCLT allows formal amendment procedures for multilateral agreements to be started among contracting parties,[37] ie those States that have expressed their consent to be bound by the treaty, independently from its entry into force. In the case of CETA, ratification is required for the valid expression of such consent. However, through their signature, States have agreed to allow CETA to apply provisionally. Thus, should provisional application start, States might initiate the amendment procedure as regulated by Article 30.2 CETA.[38]  This notwithstanding, as mentioned above, the recourse to such procedure, even if possible, will most probably not materialize. In light of this scenario, it seems useful to investigate other avenues of resolution for the accommodation of the substantive challenges highlighted above.

In this context, the wide category of subsequent agreements pursuant to Article 31(3)(a) VCLT regarding the interpretation and application of the original treaty might provide the appropriate level of flexibility (albeit within certain limits) to usefully ‘specify’ or ‘adjust’ the text of CETA. While still representing a significantly thorny issue, subsequent conduct of State parties has been thoroughly analyzed by scholarship.[39] Doctrine has shown the potential of subsequent conduct for the interpretation of the original treaty, but also its limits, especially when instruments included in the wording of Article 31(3) stretch interpretation significantly beyond the ordinary meaning of terms. The following paragraph will attempt to explore the extent to which subsequent agreements pursuant to Article 31(3)(a) may be used to achieve changes acceptable to all parties to CETA.

Albeit State and judicial practice are not entirely uniform as to the scope, extent and characteristics of subsequent agreements,[40] it is possible to identify some recurring elements that need to be taken into account in order to assess how far Article 31(3)(a) can be relevant for the aim of this analysis.

As to the ‘subsequent’ aspect, a combined reading of Article 31 paragraphs (2) and (3) suggests that these agreements refer to a moment in time after the ‘conclusion’ of the relevant treaty. A question remains as to what is meant by ‘conclusion’, since the term is neither defined in VCLT, nor necessarily univocal.[41] An analysis of judicial practice suggests to set the minimum threshold at the moment in time when the treaty text is crystallized. Entry into force does not equate to conclusion and is not necessary for the characterization of a posterior agreement as ‘subsequent’. This is the current situation of CETA.

With regard to the form of subsequent agreements, it is well established that these do not need to be concluded in treaty form. Besides the substantive relation between the original treaty and the subsequent agreement,[42] what needs to emerge is that the instrument intends to clarify the interpretation or the application of the relevant provisions.[43]  Subsequent agreements are not required to be legally binding per se,[44] in the sense of creating new or extended obligations for treaty parties beyond what is already provided in the original treaty. This notwithstanding, subsequent agreements still generate legal consequences, originating from the very same provisions as interpreted by the Parties.[45] An analysis of the case law shows that the relevant aspect is the fact of the agreement, not the compliance with any given formal requirement.[46] Thus, paragraph (3)(a) includes any single common act,[47] establishing the agreement of the parties over a specific meaning or application of the treaty itself.  In its Commentary to the 1966 draft of the VCLT, the ILC had clarified this notion, stipulating that an agreement of the parties in the sense of Article 31(3)(a) ‘represents an authentic interpretation by the parties, which must be read into the treaties for purposes of its interpretation’.[48] Yet a question remains as to the ‘supremacy’ or ‘authoritativeness’ of such an agreement, especially in light of the wording of Article 31(3) which requires the interpreter to ‘take into account’ the elements of its following paragraphs.[49]  While some scholars consider the Parties’ authentic interpretation by means of subsequent agreements to be ‘endowed with binding force’,[50] ILC’s work on subsequent practice and agreements is clear in specifying that ‘subsequent agreements and subsequent practice shall, after all, only “be taken into account” in the interpretation of a treaty, which consists of a “single combined operation” with no hierarchy among the means of interpretation which are referred to in article 31 (…) subsequent agreements and subsequent practice which establish the agreement of the parties regarding the interpretation of a treaty are not necessarily conclusive, or legally binding’.[51]

When it comes to non-binding subsequent agreements, there is no general presumption of ‘supremacy’. [52] While the interpretation conveyed by these instruments might still override other possible meanings of a certain provision, their interpretative value will be typically balanced with the other elements of the uniform rule of interpretation. Hence, their weight will depend on a interpretation of the non-binding subsequent agreement itself, both in terms of content and purpose displayed by the parties.[53]

The situation is different when States agree on binding subsequent instruments on the interpretation or application of a specific term or provision of the original treaty.[54] In this case, the interpretation agreed on will prevail over possible interpretations attained via other methods.[55] To do so, however, the agreement must display an explicit connection with the original treaty and its language must convey a ‘prevailing’ effect,[56] mostly worded in phrases such as ‘the agreement and the (original) treaty are to be interpreted as a single instrument’ or ‘in case of doubt as to the meaning or interpretation of provision XY, this agreement shall prevail’.[57]

The potential ‘authoritative’ effect of subsequent agreements leads to another issue – possibly the thorniest, albeit the most relevant to the present analysis – related to the potential amending effect of subsequent agreements pursuant to Article 31(3)(a). There is no specific rule[58] in VCLT on the effects that subsequent agreements might display beyond interpretation. Judicial practice seems equally ambiguous as to this aspect. In fact, interpretation and modification might be seen as intersecting and sharing a blurred line, which makes it complicated to distinguish between the two.

VCLT seems to address the issue of amendment separately from interpretation. Articles 39 through 41 are specifically on point with this assertion.[59] However, the fact that no specific, formal criteria are required by either Article 39 or Article 31(3)(b) has led some authors to suggest that subsequent agreements – in any form – might determine a modification of the treaty.[60]  This assertion is further evidenced by the fact that the contrary is true: agreements explicitly amending previous treaties may continue to illuminate the interpretation or application of certain treaty provisions. Their use in this context often originates from the need of the parties to solve disputes on interpretation.[61]

Hence, while an amending effect of subsequent agreements cannot be excluded,[62] some considerations are in order as to the potential limits to modification. These should prove useful as the groundwork for section four, which clarifies which instruments can be used to achieve modifications to CETA.

Foundationally, the mere existence of another relevant agreement covering the same substance of the original treaty cannot be considered sufficient evidence to infer the modification of provisions within the original treaty. In fact, a contrary presumption exists towards the stability of treaties, rather than the alteration of or derogation from existing rights and obligations encompassed therein.[63] This is particularly true in the case of non-binding agreements.[64] In this respect, ‘an agreement to modify a treaty is thus not excluded but also not to be presumed’.[65] As to binding agreements,[66] the distinction between Article 31 (3)(a) and Article 39 becomes more blurred. In this case, the will of States as to the purpose they aim to achieve seems to play a significant role in the characterization by judicial bodies and State parties of an instrument in regard to whether it merely interprets or more substantially modifies previous provisions. In fact, even where subsequent instruments seemingly change previous treaties, States are reluctant to confirm that an amendment has been made, both because they do not want to activate domestic procedures for approval of amendments and are hesitant to formally endorse new or modified legal obligations. Judicial bodies tend to accommodate the will of the Parties, respecting the effect or intention displayed in the subsequent agreement.[67] Thus States are accorded a wide margin in stretching beyond the ordinary meaning of terms, while formally staying within the confines of ‘authentic interpretation’. This deference is evidenced not only by judicial practice, but also by the increase in the numbers of subsequent agreements used to modify treaties while avoiding formal amendment procedures, as the new mechanisms endowed within multilateral treaties show.[68]

The presumption against modification via subsequent agreements is stronger where the treaty text provides for a specific amendment procedure.[69] If the subsequent agreement, be it binding or non-binding, does not comply with the requirements enumerated therein, amending language should be ‘interpreted narrowly as not to purporting to modify the treaty’.[70] The presumption in that case is that States would not desire to depart from the modalities of amendment agreed in the original treaty. This is particularly relevant for CETA, since Article 30.2 explicitly regulates amendment procedure. Yet again, in practice there is a plethora of examples to the contrary.[71]

Some authors have attached particular significance to the object and purpose of the treaty regarding ‘constraints’ imposed on the modifying extent of subsequent agreements.[72] As stated, Article 31 provides a ‘uniform’ rule of interpretation, hence the object and purpose of the treaty should represent the limits imposed on subsequent agreements related to interpretation and application of previous treaties. In this context, the object and purpose would operate in the same manner as they do for reservations. However, while the rule requires consistency in regard to the object, purpose, and plain meaning of terms, it does not rigidly bind States. This is the case for three reasons. Firstly, the object and purpose of the treaty may be sufficiently vague allowing for a wide margin for ‘dynamic’ or ‘evolutive’ interpretation even beyond the ordinary meaning of terms. Secondly, subsequent conduct might contribute to defining the object and purpose.[73] Thirdly, even if the presumption against modification is the strongest where the agreement impacts an essential element of the treaty, if the subsequent agreement is sufficiently clear, then the latter may modify the object and purpose of the former.[74] Therefore, although the presumption is to the contrary, the constraint of object and purpose is not absolute.

There are two final recurring elements in the debate over whether subsequent agreements carry interpretative or modifying effects. The first relates to considerations of compatibility with constitutional rules on treaty-making. Domestic courts deal regularly with the question of whether a specific subsequent agreement should be interpreted as clarifying relevant treaty provisions for the sake of interpretation or whether it is an outright modification.[75] The reason being that under domestic law, any modification – other than the addition of interpreting language – usually requires the involvement of domestic legislative bodies. Hence, considerations of legitimacy[76] of the treaty at home arguably require caution especially when subsequent amendments do not comply with the procedural requirements enumerated within the treaty, which sometimes envisage the fulfillment of previous domestic procedures for the entry into force of the amendment.

The final element, namely whether subsequent agreements might alter rights accorded to third parties, is particularly relevant for the present discussion. In the context of CETA this purportedly refers to investors. Investors under CETA enjoy procedural rights against State treaty Parties. However, per se, these third party rights of investors do not limit the rights of States to interpret and modify the provisions prescribing such third-party rights. Investors’ rights are embedded in the legal framework of the treaty.[77] The arguments made so far in this contribution demonstrate that States are not strictly limited in being ‘the masters of their own treaties’, via either formal amendment or subsequent agreements, and practice does not clearly distinguish between the effects of these two. Notwithstanding this consideration, some scholars[78] consider that a distinction may be drawn in terms of the timing of the effects of interpretation and modification respectively. In this regard, interpretation processes have retroactive effect, while amendments typically have prospective effect. Hence, some constraints could arise in instances where de facto modifications are enacted via a subsequent agreement that formally has been ascribed interpretative effect. This could operate to the detriment of investors’ rights where they have previously submitted a claim for ISDS. In the case of an amendment, regardless of its form, the ‘presumption of international law against retroactivity should prevail’,[79] and thus only exert a prospective effect.



  1. Application of the theoretical framework to CETA


The following sections will try and transpose the overarching legal framework on the scope of subsequent agreements into the specific context of CETA. As mentioned, these instruments seem to better respond to substantive challenges and reforms, to be introduced both before and after provisional application of CETA. The considerations below adopt a cautious approach, arguing for the adoption of either non-binding or binding agreements, within the margin of the presumptions discussed above. Section 4.1 will present how non-binding subsequent agreements could accomplish the desired treaty alterations referred to in section 2, while section 4.2 deals with binding agreements on the interpretation and application of CETA. It seems interesting to state here that both categories might well be included in those ‘implementation’ measures that the Parties have undertaken to adopt for the purposes of provisional application, before its official start.[80]


4.1. Non-binding agreements

The array of agreements that States may use to accomplish the objectives identified in section 2 is significantly varied, especially considering that – as mentioned – form does not feature as a necessary requirement. Depending mostly on what State parties want to convey, both substantively and politically, they might make recourse to (Joint) Declarations,[81] Memoranda of Understanding,[82] and any other type of informal agreement. The foundational aspect of any tool they decide to adopt is that it should display an apparent association with CETA, since it must express the agreement of parties regarding the interpretation or application of rights and obligations included in the treaty.

To begin with, subsequent agreements could be used to introduce a prior exhaustion of domestic legal remedies obligation for investors wishing to submit a claim to the ICS. Such a stipulation arguably falls within previously identified (albeit non-absolute) limitations regarding the scope of non-binding subsequent agreements. This is justified by a number of reasons. First, the introduction of a prior exhaustion of domestic legal remedies requirement does not alter the legal positions of State Parties to the treaty. In this scenario, investors would still retain the right to submit a claim against host States, but only after all domestic legal remedies have been exhausted. The said provision would introduce a requirement for proceeding, while leaving the investor-State dispute system intact. Even so, the ICS would not (and should not) function as a super appeal mechanism against domestic courts’ decisions. CETA does not allow the ICS to apply domestic law to investor-State disputes[83] and consequently it cannot review domestic courts’ decisions. Hence, the tribunal would still be required to review the compatibility of the host States’ conduct (that obviously includes the judiciary as an organ of the State), weighing it against rules and principles of international law pursuant to Article 8.31.[84] Such a practice is not so distinct from processes which occur at the European Court of Human Rights or before other international fora which require the exhaustion of domestic legal remedies.[85]

Inserting a requirement for exhaustion of domestic legal remedies would also be abreast with the procedure already existing within CETA[86] which allows the Committee on Services and Investment to adopt, with the agreement of the parties, rules ‘supplementing’ Section F, Chapter 8 on investor-State dispute settlement.[87] The adoption of a rule on the prior exhaustion of domestic remedies would manifest the understanding of the parties towards the specific application of the provision at Article 8.23 on the submission of claims, further indicating the conditions under which a claim can be submitted.

As mentioned, there are two provisions which might call for caution in drafting a Joint Declaration or a Memorandum of Understanding requiring the exhaustion of domestic legal remedies. Article 8.22(4) of CETA obliges claimants to withdraw or discontinue (f) or waive (g) any proceeding before either domestic or international adjudicatory bodies, before submitting a claim to the ICS. The purpose of these provisions is to avoid parallel proceedings.[88] Therefore, although prima facie it might seem so, these clauses do not contradict per se the ratio of the previous exhaustion of domestic remedies rule. In fact, while requiring the previous exhaustion of national proceedings, the latter still achieves exactly the same objective as the provisions on discontinuance. However, subsequent agreements should be worded in a way as to clarify the interplay between the previous exhaustion of domestic legal remedies requirement and the obligation to withdraw or waive any other proceedings. This could be easily achieved by specifying that Article 8.22(4) (f) and (g) refer to other international courts or tribunals only.

As to the third parties, ie potential investors, there seems to be no counter indication towards the adoption of a less formal agreement on the application of Chapter 8, Section F. There are two particulars to consider here. First, the treaty text already provides for a mechanism that allows for supplementing existing rules under the relevant section, making potential investors already well aware of the possibility for further specification of the rules pertaining to the dispute procedures and requirements. Additionally, and most importantly, the treaty itself is not in force and provisional application does not cover the ISDS. As such, investors have not yet been bestowed any procedural rights. This means that investors will decide which States to invest in based on provisions currently regulating substantive protection only, without any legitimate expectations as to how to access the ICS system. In fact, even entertaining such a perception of legitimate expectations at this point in time would be dubious and ill-advised.


4.2. Binding agreements

A slightly different reasoning applies to the introduction of procedural rules for tribunals. In fact, the treaty clearly refers to the possibility of supplementing the rules and procedures listed in Article 8.23 CETA. It seems, therefore, that simply a ‘specification’ in the application of the rules would not be sufficient to introduce new procedural rules for the tribunal. Rather, an objective interpretation of the text of the treaty would point towards excluding the possibility for a modification of the arbitration rules and procedures on a recommendation issued by the Committee on Services and Investment. Consequently, as demonstrated, an informal agreement regarding the introduction of new arbitration procedures would challenge the presumption against modification in those cases where (1) the treaty text clearly displays an indication to the contrary and (2) an amendment procedure is provided for in the treaty itself. Therefore, the adoption of binding agreements seems more justifiable, since it allows more leeway for (partial) adjustments of CETA. This is the case for two reasons. First, a binding agreement should leave little doubt as to the ‘supremacy’ effect mentioned above on the authoritativeness of subsequent agreements. Furthermore, the adoption of binding agreements, regardless of their characterization (as interpretative, amending or any other), would still be coherent with the procedural requirements for amendment included in CETA at Article 30.2. While the treaty is still not in force (at least until the moment the treaty is provisionally applied), a good faith obligation deriving from its signature would suggest to adopt agreements in line with those modification requirements specified therein.

In this case, Parties could resort to an implementation agreement.

The notion of implementation agreements is not clearly defined under international law.[89] Generally speaking, these agreements are considered to be instruments related to the relevant original treaty, which provide for measures that make specific treaty provisions operational.[90] Thus, the overarching function is to clarify or specify processes or legal obligations referred to in the original treaty. In this sense, implementation agreements should be included within ‘subsequent agreements regarding (…) the application’ of the original treaty under Article 31(3)(a), thus contributing to (rectius, providing for) the ‘correct’ interpretation of relevant treaty provisions.

Moreover, the amendment procedure under CETA is arguably not particularly stringent. Other than the written form, the provision stipulates that entry into force of the amendment occurs after ‘the exchange of written notifications certifying that they (ie the Parties) have completed their respective applicable internal requirements and procedures necessary for the entry into force of the amendment, or on the date agreed by the Parties.’ In this respect, parties are allowed a considerable margin of flexibility on the means of revising the treaty. A written and binding implementation agreement seemingly conforms to these conditions, provided that it characterizes clearly its relationship with the original treaty, ie that the former should supersede the latter in case of inconsistencies between the two. However, as to its appearance, an implementation agreement seems far less controversial than a formal amendment.

A prominent example in State practice is the 1994 Implementation Agreement on Part XI of UNCLOS. [91] It is widely known that the Agreement was introduced to further facilitate universality of the Law of the Sea Convention and avoid a fragmented regime of the oceans, with specific regard to the deep seabed mining activities. The granular drafting technique allowed for the introduction of an agreement de facto modifying some parts of the Law of the Sea Convention, even though the Agreement was designed in a way as not to appear as an amendment.[92] Politically, a formal modification would have been unacceptable to the States that had already ratified UNCLOS. Accordingly, other than introducing interpretative and implementing provisions, the Agreement declares inapplicable a certain number of provisions of Part XI, and introduces rules which shall be applicable instead, resulting in an Agreement on the nonapplication of the original treaty. Thus, at least formally, it could still qualify as an agreement on the application of the treaty, rather than an amendment. The technique designed for the expression of consent was equally sophisticated. The circumstances there called for a mechanism that would avoid ‘cumbersome and perhaps politically embarrassing domestic procedures’.[93] This resulted to the simplified procedure based on unpronounced acquiescence mechanism after signature (ie not requiring ratification for the consent to be bound) of the Agreement.

It should be noted that the experience of the Deep Seabed Mining Agreement might apply mutatis mutandis also to CETA. In fact, similar political circumstances are at play that disfavor a formal amendment of the treaty (be it in the form of a Protocol or a revised agreement in its entirety). The adoption of procedural rules appropriate to the Tribunal, for example, might well be established via an implementation agreement on the non-applicability of those CETA Articles referring to external arbitrational rules, while at the same time introducing new, alternative, rules to be applied.

Since it is plausible that an implementation agreement could be utilized to introduce a dispute settlement mechanism, such an agreement could be also concluded in order to regulate or eradicate an already existing dispute settlement mechanism. Thus, the outright eradication of the ICS or the disguised replacement of arbitration procedural rules in CETA could be arguably achieved via an implementation agreement. While directly replacing the applicable rules tout court would beget the same effect, an implementation agreement adopting new rules and rendering inapplicable the previous ones might be politically more acceptable. Furthermore, the introduction of procedural rules for the ICS does not seem to call necessarily for a domestic ratification scheme. A simplified procedure might well be sufficient, at least for many of the constitutional domestic systems of the Parties. This would also be in line with Article 30.2 CETA, which sanctions the entry into force of a potential amending agreement to a confirmation that ‘applicable’ domestic requirements have been fulfilled or to a date set by the Parties for this purpose. Thus it would seem that a realization of domestic parliamentary procedures is not required for the purposes of this provision either.[94]

A more drastic solution would be adopting an implementation agreement declaring the non-application tout court of the ICS mechanism. While seemingly more radical, the ISDS has been either abandoned or negotiated on an ad hoc basis in several instances.[95] Investors retain the possibility of bringing claims before domestic courts (see supra) and persistently urge their State of nationality to initiate a State-to-State dispute under Chapter Twenty-Nine of CETA.[96]

Based on this rationale, a prior involvement of domestic legislators is seemingly unnecessary. While implying an alteration of the legal position of Parties − in the sense that they would not be challenged as respondent anymore −, the non-application tout court of the ISDS admittedly decreases the burden over Parties, hence re-expanding the previous constraints levied over their ‘sovereignty’.

A different scenario applies if Parties want to remove the ICS, but also intend to revert the abovementioned provision on direct effect in its entirety. In fact, this would arguably alter the domestic legal system of the Parties, allowing for the internal enforceability of international standards before domestic courts. Such a circumstance would likely entail national legislative action.[97]


5. ‘Unilateral’ modifications to the applicability of the ICS provisions

While being a more peregrine option, it makes sense to briefly elaborate on the possibility to sanction the exclusion of the ICS through either (unilateral) reservations of Parties or the negotiation of an Opting-out Protocol. While being two significantly different instruments – the second requiring an agreement among the Parties – it makes sense to discuss about their potential effects together, since they both warrant a non-universal application of some provisions of the treaty. At the same time, should some Parties obstruct the entry into force of CETA in its entirety, these instruments would allow flexibility, modifying the treaty for some of the parties only, while guaranteeing wider participation.

An assessment of the two instruments requires an analysis under both EU and international law; furthermore, both an opting-out protocol and a reservation are strictly intertwined to what shall be called here, borrowing from WTO language, a kind of ‘variable geometry’ of CETA.[98]

First and foremost the question is whether EU Member States (MS) can put a reservation on provisions of the treaty falling under shared competence, such as the ISDS mechanism in CETA. The articulation of EU external competence has always been particularly convoluted. With regard to mixed agreements, the ECJ has been quite restrictive of the discretion MS have in their implementation.[99] In its judgments, the Court has heavily relied on the duty of loyalty that MS bear towards a unitary representation of the EU towards third subjects on the international scene.[100]

More particularly, in its case law the ECJ has clarified that the interest of the EU towards a uniform and effective application of mixed agreements over the entire territory of the Union shall be protected.[101] Hence, the duty of loyalty[102] would serve as the juridical basis to justify the existence of an (EU) obligation for MS to fully and correctly apply all parts of mixed agreements, regardless of the specific competence covering one or the other provision.[103] A correct application of the agreement can be fully guaranteed only when both the EU and MS concur to implement the treaty. The discretion of MS in the implementation of mixed agreements is hence significantly reduced to the point of including a duty for MS to abstain from actions which, albeit included in their sphere of competence, might still put at risk or have an impact on the correct application of the treaty on the part of the Union or the implementation of relevant secondary EU Law. Thus, under a more preventive dimension of the duty of loyalty, MS have an obligation to consult with and facilitate the EU in exercising its competences and achieve its objectives,[104] and avoid any risk of infringing already existing or pending adoption EU Law. Therefore, it seems unlikely that contracting States will put a reservation excluding outright the recourse of the ICS for investments installed in their territories. The inclusion of such a reservation by MS would also further complicate things. Should MS put such a reservation, a situation will be produced by which some EU companies residing in the States which accept ICS will benefit from the investor-State dispute mechanism, whereas others residing in reserving States would not have this possibility.[105] This outcome might be seen as contradicting the principle of non-discrimination based on nationality, by which EU natural and legal persons should receive equal treatment on the whole territory of the Union. These considerations make the prospect of an Opting-out Protocol even less likely.

Under international law, the situation is slightly different. The question is in primis whether such a reservation or opting-out mechanism would contradict the purpose and object of the treaty. While some commentators argue that an investor-State dispute mechanism represents an essential part of a FTA, this is not necessarily the case. If one looks at the preamble of CETA, the ISDS does not emerge as a fundamental feature to achieve the objectives and purposes of the treaty. Furthermore, claims might still be submitted to the State-to-State dispute settlement mechanism and investors are anyway left with the possibility of recurring to mediation and domestic courts.

There is, however, another obstacle deriving from the concurrent participation of both the EU and the MS to the treaty. In particular, Article 8.21 CETA allows investors to submit claims to the ICS both against the EU and a MS. The EU has 50 days to notify to the claimant the correct respondent. In a situation where a MS has put a reservation or opted out from the ICS mechanism, the investor will most likely submit a claim against the EU for infringements occurred in a (national) territory where the ICS does not apply. Article 8.21(4)(b) (Determination of the respondent for disputes with the European Union or its Member States) stipulates that ‘In the event that the investor has not been informed of the determination within 50 days of delivering its notice requesting such determination: (…) if the measures identified in the notice include measures of the European Union, the European Union shall be the respondent’; furthermore at paragraph 6 ‘If the European Union or a Member State of the European Union is the respondent, pursuant to paragraph 3 or 4, neither the European Union, nor the Member State of the European Union may assert the inadmissibility of the claim, lack of jurisdiction of the Tribunal or otherwise object to the claim or award on the ground that the respondent was not properly determined pursuant to paragraph 3 or identified on the basis of the application of paragraph 4.’ The fact that an investor might still submit a claim against the EU – even in the case where a MS has put a reservation to the ICS mechanism – will seemingly enhance the burden for the EU to justify that the challenged measure is exclusively a national one. This becomes more difficult when the State has enacted the challenged measure in application of an EU act. Besides the difficulty a (Canadian) investor might have in identifying whether a measure is exclusively national or not, the wording ‘include measures of the EU’ is extremely vague. One could say that measures adopted in implementation of EU acts – where States might still preserve a significant level of discretion – do ‘include’ a measure of the European Union. Hence, it would seem that for a large part of potential infringing actions, the EU would still legitimately qualify as the respondent. Additionally, prima facie, it seems that according to Article 8.21(6) the EU might not challenge the determination of the Respondent under paragraph (4). This puts the additional question of how the EU will (internally) reciprocate over the MS who has put the reservation on the ICS.

Thus, while being theoretically viable under international law, both an Opting-out protocol and the application of reservations do not satisfactorily address the flaws of the ICS. On the contrary, both solutions might risk creating a fragmented and less effective regime.


6. Concluding remarks

The transatlantic agreement between Canada and the EU has produced whirlwind controversies throughout. One of the most delicate aspects of the debate relates to the vexed question of the introduction of a (modernized) investor-State dispute settlement mechanism in the treaty. While several challenges have been addressed with the introduction of the ICS, States’ concerns have not been completely alleviated, as the occurrence of recent ‘incidents’ has shown. This contribution has attempted to give an overview of an array of potential instruments that the Parties might adopt before or after the advent of the provisional application of CETA in order to ‘alter’ the text of the treaty. These might timely answer those ISDS questions that are still open, notwithstanding the novelties brought by the ICS. Drawing from the Law of Treaties and examples of State practice, Memoranda of Understanding and Implementation Agreements seem to have the potential to usefully address substantive reforms, depending on the degree of textual ‘adjustment’ needed. While formal amendment procedures seem relatively unrealistic, these instruments prove quite flexible and could legitimately qualify as those implementation measures that are currently being arranged in preparation of the beginning of the provisional application of CETA.



*Assistant Professor in International Law, Erasmus School of Law, Erasmus University Rotterdam. The author would like to thank Alessandra Arcuri, Anastasia Karatzia, Federica Favuzza, Marco Longobardo, Jennifer Riter and the editors of the journal for their insightful comments on a previous version of this article. The usual disclaimer applies.

[1] See A Arcuri, ‘Who is afraid of CETA?’ The Department of International and European Union Law Blog (26 October 2016) <http://blog.eur.nl/esl/ieul/2016/10/ 26/who-is-afraid-of-ceta/>; R Repasi, ‘Wallonia Locuta, Causa Finita’ The Department of International and European Union Law Blog (26 October 2016) ˂http://blog.eur.nl/esl/ieul/2016/10/26/wallonia-locuta-causa-finita/˃.

[2] The literature on rationale and challenges of the ISDS mechanism is significantly rich, see R Howse, ‘International Investment Law and Arbitration: A Conceptual Framework’ (2017) Institute for International Law and Justice Working Paper 2017/1, <www.iilj.org/publications/international-investment-law-arbitration-conceptual-framework/> and the copious scholarly work cited therein.

[3] Comprehensive Economic and Trade Agreement (CETA) Between Canada, of the one part, and the European Union and its Member States, of the other part (30 October 2016) <http://data.consilium.europa.eu/doc/document/ST-10973-2016-INIT/en/pdf>.

[4] Council Decision 2017/38/EU of 28 October 2016 on the provisional application of the Comprehensive Economic and Trade Agreement (CETA) between Canada, of the one part, and the European Union and its Member States, on the other part [2017] OJ L11/1080. During the last G20, Canada and the EU have agreed September 21st as the date of start of provisional application of the treaty ˂http://europa.eu/rapid/press-release_STATEMENT-17-1959_en.htm˃.

[5] See for example demonstrations going on in Italy during discussions on ratification of CETA at the Foreign Affairs Committee of the Senate. D Giovinazzo, ‘CETA, il senato accellera sulla ratifica tra le proteste. ‘Mattarella: “Passo imporante”: La Commissione Esteri ha inviato il testo in aula respingendo una richiesta di sospensiva votata solo da sinistra italiana, M5S e Lega’ EU News, L’Europa in Italiano (Rome, 27 June 2017) ˂www.eunews.it/2017/06/27/ceta-il-senato-accelera-sulla-ratifica-tra-le-proteste-mattarella-passo-importante/88865˃. A general overview of domestic discontentment over CETA can be retrieved here: B Patterson, ‘CETA Ratification Faces a New Hurdle Following the Election in the Netherlands’ The Council of Canadians (16 March 2017) ˂https://canadians.org/blog/ceta-ratification-faces-new-hurdle-following-election-netherlands˃.

[6] Avis 2/15 Accord de Libre-Échange avec Singapour (CJEU, 16 May 2017) paras 285-293.

[7] See the compromise deal between the Belgian regional and community governments of last October. Part of the deal is that the Belgian government will seek the Opinion of the ECJ on the compatibility of the Investment Court System with EU Treaties. See ˂http://rf.llb.be/file/6f/5811e50fcd70fdfb1a589e6f.pdf˃.

[8] See ˂http://www.conseil-constitutionnel.fr/conseil-constitutionnel/emailing-utf-8/2017/saisine-2017-749-dc-[ceta].148634.html˃.

[9] Provided that it takes place later in the year as agreed, in a situation of radical disagreement, States might even decide to terminate provisional application in conformity with art 30.7 of CETA, See Pantaleo in this issue for a thorough discussion on provisional application.

[10] The Commission has treated CETA as a mixed agreement already from the outset. Thus the ratification of each single Member State is required for the treaty to enter into force.

[11] Provisional application had been further delayed because of an alleged misalignment on cheese tariff quota, see CBC News, ‘Canada, EU to Provisionally Apply CETA in September: New Date Set after Parties not Ready for Provisionally Application of Trade Agreement on Canada Day’ (8 July 2017) ˂www.cbc.ca/news/ politics/ceta-september-provisionally-1.4196210˃.

[12] CETA (n 3) art 30.2, see infra.

[13] Amongst others Canada, Spain, Denmark, Latvia and Croatia.

[14] Domestic (constitutional) courts are generally adamant that formal amendments of treaties require constitutional treaty-making procedures in order to safeguard domestic legitimacy. See the national case law cited in G Hafner, ‘Subsequent Agreements and Practice: Between Interpretation, Informal Modification, and Formal Amendment’ in G Nolte (ed), Treaties and Subsequent Practice (OUP 2013) 105, 116.

[15] While modification and revision are generally considered to be two different concepts, they will be treated as synonym for the purposes of this contribution.

[16] Besides academic work, scholars have been quite active in engaging in the public debate over CETA and MegaRegionals more in general. See amongst others the Legal Statement issued in October 2016 by a large number of international law scholars, STOP TTIP, ‘Legal Statement on Investment Protection and Investor-State Dispute Settlement Mechanisms in TTIP and CETA’ (October 2016) ˂https://stop-ttip.org/wp-content/uploads/2016/10/13.10.16-Legal-Statement-1.pdf˃.

[17] This paper endorses the ‘substance dictates form’ approach, adopted in the occasion of the drafting of the Agreement Relating to the Implementation of Part XI of the United Nations Convention on the Law of the Sea (adopted 28 July 1994, entered into force 28 July 1996) 1836 UNTS 3, extensively elaborated in T Treves, ‘Entry into Force of the United Nations Law of the Sea Convention: The Road Towards Universality’ in N Al-Naouimi, R Meese (eds), International Legal Issues Arising under the United Nations Decade of International Law (Martinus Nijhoff 1994) 443.

[18] In this direction, S Hindelang, ‘Study on Investor-State Dispute Settlement (ISDS) and Alternatives to Dispute Resolution in International Investment Law’ (2016) 13 Transnational Dispute Management 1; P Hainbach, ‘The EU’s Approach to Investor-State Arbitration in the Comprehensive Economic and Trade Agreement (CETA)’ (2016) 13 Transnational Dispute Management 1; C Titi, ‘The European Union’s Proposal for an International Investment Court: Significance, Innovations and Challenges Ahead’ (2017) 14 Transnational Dispute Management 1.

[19] CETA (n 3) art 8.44.

[20] For a thorough overview of the challenges relating to the ISDS more in general, see Hindelang (n 18).

[21] See Howse (n 2).

[22] Chapter 8, Section F.

[23] For an extensive analysis of the innovative elements and challenges of the reform, see Titi (n18).

[24] See supra in the Introduction.

[25] Amongst others, because of the absence of tenured track judges, the renvoi to already existing arbitral procedural rules, party autonomy in procedural rules’ choice, equivalence of final decisions with arbitral awards. See infra in this Section.

[26] See on this aspect D Gallo, ‘Portata, estensione e limiti del nuovo sistema di risoluzione delle controversie in materia di investimenti nei recenti Accordi sul libero scambio dell’Unione Europea’ (2016) 4 Diritto del Commercio Internazionale 827, 837, and M Distefano, ‘L’effetto diretto delle norme pattizie ovvero di “casi eccezionali”’ in M Distefano (ed) L’effetto diretto delle fonti dell’ordinamento giuridico dell’Unione Europea. Riflessione sui rapporti tra fonti dell’Unione Europea e fonti interne (Editoriale Scientifica 2017) 27, 42.

[27] See Hindelang (n 18) 46, 60.

[28] Howse (n 2) 57. On the ‘public’ nature of the jurisdiction exercised by international arbitrators see I Venzke, ‘Investor-State Dispute Settlement in TTIP from the Perspective of a Public Law Theory of International Adjudication’ Amsterdam Center for International Law Research Paper No 2016-6 <https://papers.ssrn.com/ sol3/papers.cfm?abstract_id=2742173>.

[29] Many studies have shown how the ISDS is actually exerting a detrimental effect on the rule of law, both domestically and internationally, for an extensive account see Howse (n 2) 33 and A Arcuri, ‘The Great Asymmetry’ paper presented at the 2017 WTO Conference, on file with the author.

[30] See infra section 4.1.

[31] See Titi (n 18) 22.

[32] Problems arise also in relation to the enforcement mechanism, which sanctions the equivalence of ICS awards adopted under CETA to ICSID awards. Clearly, this equivalency with ICSID awards for the purposes of enforcement can only be binding between the EU and Canada and cannot extend to the other States party to the ICSID Convention that are not bound by CETA. See on this Titi (n 18) 23.

[33] The now well-known quote is taken from LE Peterson, Europe’s latest TTIP investment proposal cloaks arbitration in judicial robe, tightens ethical screws (further), and thinks seriously about small claims (Investment Arbitration Reporter, 17 November 2015) <https://www.iareporter.com/articles/analysis-europes-latest-ttip-investment-proposal-cloaks-arbitration-in-judicial-robe-tightens-ethical-screws-further-and-thinks-seriously-about-small-claims/>, accessed 7 July 2017.

[34] Titi (n 18) 23; Hainbach (n 18) 33.

[35] Titi (n 18) 24. On the legal feasibility issues, see B McGarry, J Ostransky, ‘Modifying the ICSID Convention under the Law of Treaties’ EJIL:Talk! (11 May 2017)  <www.ejiltalk.org/author/bmcgarryjostransky/>.

[36] See infra section 4.2.

[37] Art 2(1)(f) Vienna Convention on the Law of Treaties (adopted 23 May 1969, entered into force 27 January 1980) 1155 UNTS 331 (VCLT).

[38] K Odendahl, ‘Amendment of multilateral treaties’, in O Dörr, K Schmalenbach (eds), Vienna Convention on the Law of Treaties: A Commentary (Springer 2012) 714, reasons that signature equates to the expression of consent to be bound by the treaty. However, his opinion cannot be shared, since this would be the case only for those treaties approved via simplified procedure, which is not the case of CETA.

[39] See amongst others the collected works of E Cannizzaro (ed), The Law of Treaties Beyond the Vienna Convention (OUP 2011); Nolte (ed) Treaties and Subsequent Practice (n 14).

[40] See extensively ILC, ‘Second Report on Subsequent Agreements and Subsequent Practice in Relation to the Interpretation of Treaties, by Georg Nolte, Special Rapporteur’ (26 March 2014) UN Doc A/CN.4/671.

[41] See R Gardiner, Treaty Interpretation (OUP 2015), 232

[42] ibid 243; Hafner (n 14) 114.

[43] See ‘Second Report on Subsequent Agreements’ (n 40) 4. Hence a mere overlap in terms of content is not sufficient to the purpose. On this, M Kohen, ‘Keeping Subsequent Agreements and Practice in Their Right Limits’ in G Nolte (ed), Treaties and Subsequent Practice (n 14) 34.

[44] O Dörr, ‘Art. 31 General Rule of Interpretation’ in O Dörr, K Schmalenbach (eds), Vienna Convention on the Law of Treaties: A Commentary (n 38) 553; ‘Second Report on Subsequent Agreements’ (n 40) 27; L Crema, ‘Subsequent Agreements and Subsequent Practice in Their Right Limits’ in G Nolte (ed), Treaties and Subsequent Practice (n 14) 13, 25.

[45]  ‘Second Report on Subsequent Agreements’ (n 40) 28; Gardiner (n 41) 246.

[46] See eg WTO, Panel Report, EC-IT Products, WT/DS375/R, WT/DS376/R, WT/DS377/R, 16 August 2010, para 7.558; Territorial Dispute (Libyan Arab Jamahiriya v Chad) (Judgment) [1994] ICJ Rep 6, para 60; Kasikili/Sedudu Islands (Botswana v Namibia) (Judgment) [1999] ICJ Rep 1045, para 63.

[47] ‘Second Report on Subsequent Agreements’ (n 40) 26.

[48] ‘Report of the International Law Commission on the work of its eighteenth session, Geneva, 4 May – 19 July 1966’ (1966) II YB of the Intl L Commission 221, para 14.

[49] See J Crawford, ‘Subsequent Agreements and Practice within the Vienna Convention’ in G Nolte (ed), Treaties and Subsequent Practice (n 14) 29.

[50] M Villiger, ‘The Rules on Interpretation – Misgivings, Misunderstandings, Miscarriage? The “Crucible” Intended by the International Law Commission’ in E Cannizzaro (ed) The Law of Treaties Beyond the Vienna Convention (n 39) 105, 111.

[51] ‘Report of the International Law Commission, Sixty-fifth session, 6 May-7 June and 8 July-9 August 2013’, General Assembly Official Records, Sixty-eight Session, Supplement No 10 (UN Doc A/68/10) 21.

[52] J Crawford (n 49) 31-32. To borrow the words of the author: ‘First of all, agreements can be informal and of course they require interpretation. There is no automaticity. Secondly, although the Vienna Convention does not pay much respect to the notion of a hierarchy of treaties, it seems reasonable to posit that there may be a hierarchy of treaties as between more formal and established expressions of state will and more ephemeral or temporary expressions. Not every coming together of states is Page Id: 31Vienna Convention on the Law of Treaties (United Nations [UN]) 1155 UNTS 331, UN Reg No I-18232 IC equal to every other coming together of states. Not every press release is equal to every other press release’.  See also Crema (n 44) 25.

[53] ibid 26.

[54] ‘Report of the International Law Commission’ (UN Doc A/68/10) (n 51) 21.

[55] ‘Second Report on Subsequent Agreements’ (n 40) 29; Gardiner (n 41) 226.

[56] Dörr (n 38) 553. See WTO, Panel Report, Chile-Price Band System and Safeguard Measures Relating to Certain Agricultural Products WT/DS207/R (3 May 2002) para 7.83 as quoted ibid, where the WTO Panel refused to consider a certain agreement between Chile and MERCOSUR as a subsequent agreement pursuant to art 31(3)(a), because the preamble of the instrument stipulated that the latter ‘shall adjust’ to the WTO Agreements.

[57] See eg the 1994 Implementation Agreement (n 17) art 2.

[58] J Alvarez, ‘Limits of Change by Way Subsequent Agreements and Practice’ in G Nolte (ed), Treaties and Subsequent Practice (n 14) 123; Hafner (n 14) 115.

[59] See WTO, Appellate Body Report, EC-Bananas III, WT/DS27/AB/RW2/ECU, WT/DS27/AB/RW/USA (26 November 2008) paras 391-393

[60] See A Aust, ‘Amendment of Treaties’ in A Orakhelashvili, S Williams (eds), 40 Years of the Vienna Convention on the Law of Treaties (British Institute of International and Comparative Law 2010) 41. See also J Klabbers, ‘Treaties, Amendment and Revision’ (2006) Max Planck Encyclopedia of Public Intl L <http://opil.ouplaw.com/view/10.1093/law:epil/ 9780199231690/law-9780199231690-e1419?prd=EPIL> and S Murphy, ‘The Relevance of Subsequent Agreements and Practice for Treaty Law to Date and in the Future’ in Georg Nolte (ed), Treaties and Subsequent Practice (n 14) 82, 88. Amid judicial practice, Territorial Dispute (n 46) para 60; See also Methanex v United States, Jurisdiction and Merits NAFTA Ch 11/UNCITRAL Arb Trib (3 August 2005) paras 20-21.

[61] Gardiner (n 41) 250.

[62] Klabbers (n 60); ‘Second Report on Subsequent Agreements’ (n 40) 62.

[63] See Pulp Mills on the River Uruguay (Argentina v Uruguay) (Judgment) [2010] ICJ Rep 14, para 131. On this Kohen (n 43) 39.

[64] Kohen (n 43) 44.

[65] ‘Second Report on Subsequent Agreements’ (n 40) 62.

[66] These include also ‘interpretative’ instruments adopted within the framework of a treaty. Both NAFTA art 1131(2) and CETA art 8.44 allow for bodies established via the treaties to issue interpretative provisions, which are binding upon adjudicatory bodies.

[67] See eg Pope & Talbot Inc (Claimant) v Government of Canada (Respondent), UNCITRAL Arbitration Under NAFTA Chapter Eleven, Award on Damages (31 May 2002) paras 46-47. On this extensively A Roberts, ‘Power and Persuasion in Investment Treaty Interpretation: The Dual Role of States’ (2010) 104 AJIL 179.

[68] Klabbers (n 60); Murphy (n 60) 88. These entail decisions of Conference(s) of the parties, tacit consent procedures, majority provisions, opting out procedures etc.

[69] ‘Second Report on Subsequent Agreements’ (n 40) 66.

[70] ibid 68.

[71] See in particular the case of a series of modifications introduced by the State Parties of both UNCLOS (n 17) and the Montreal Protocol on Substance that Deplete the Ozone Layer, 1987 outside the formal amendment procedures provided for in the respective treaties, both referred to in ‘Second Report on Subsequent Agreements’ (n 40) 64-65.

[72] See Alvarez (n 58) 126; Kohen (n 43) 44. Hafner (n 14) 119, explains that the vagueness and nature of the object and purpose of a treaty might determine the broadness of the margin of interpretation. Thus, some treaties might be subject to a stricter margin, e.g. the ECHR or WTO Agreements. However see for example Soering and Ocalan cited in Roberts (n 67) 206, where the European Court of Human Rights has arguably shown openness towards de facto amendments.

[73] Hafner (n 14) 115.

[74] ‘Second Report on Subsequent Agreements’ (n 40) 68; Crawford (n 49) 31.

[75] See amongst others S Kadelbach, ‘Domestic Constitutional Concerns with Respect to the Use of Subsequent Agreements and Practice at the International Level’ in G Nolte (ed), Treaties and Subsequent Practice (n 14) 145; Hafner (n 14) 115.

[76] Alvarez (n 58)130.

[77] On ‘non-treaty’ parties, see Roberts (n 67) 211. However, not all scholars agree that substantive standards of treatment might equate to ‘rights’ that investors enjoy. G Van Harten, Investment Treaty Arbitration and Public International Law (OUP 2007) 167. While this interpretation can be endorsed, investors still seem to enjoy at least a procedural right for the enforcement of these standards.

[78] Roberts (n 67) 212.

[79] ibid 213.

[80] See G20 declarations (n 4).

[81] Parties have already adopted a Joint Declaration prior to the signature of the treaty which can be characterized as an instrument pursuant to art 31(2)(a) VCLT, see A Arcuri in this issue.

[82] On the hybrid nature of this instrument see Crema (n 44) 25.

[83] See CETA (n 3) art 8.31.

[84] S Schill suggests that the ICS might oversee the first instances proceedings, S Schill, ‘Das TTIP-Gericht: Keimzelle oder Stolperstein für echte Multilaeralisierung des Internationalen Investitionsrechts?’ Verfassungsblog (25 November 2015) <http://verfassungsblog.de/das-ttip-gericht-keimzelle-oder-stolperstein-fuer-echte-multilateralisierung-des-internationalen-investitionsrechts/>.

[85] This notwithstanding, a tribunal might still benefit from the conclusions of national courts as to the correct application of domestic law, with the possibility of considering the latter as a matter of fact only. CETA (n 3) art 8.31(2) stipulates that ‘in determining the consistency of a measure with this Agreement, the Tribunal may consider, as appropriate, the domestic law of the disputing Party as a matter of fact. In doing so, the Tribunal shall follow the prevailing interpretation given to the domestic law by the courts or authorities of that Party and any meaning given to domestic law by the Tribunal shall not be binding upon the courts or the authorities of that Party’.

[86] The existence of specific procedures within a treaty are not exclusive of the adoption of subsequent agreements outside these procedures; these are considered to coexist and complement each other. See Hafner (n 14) 120.

[87] CETA (n 3) art 8.44.

[88] Hindelang (n 18) 65.

[89] C Kojima, V Vereshchetin, ‘Implementation Agreements’ (2013) Max Planck Encyclopedia of Public Intl L <http://opil.ouplaw.com/view/10.1093/law:epil/ 9780199231690/law-9780199231690-e1419?prd=EPIL>.

[90] See ibid the various functions that these implementation agreements might perform.

[91] 1994 Implementation Agreement (n 17), also known as Deep Seabed Mining Agreement. A similar practice is the adoption of the MARPOL 1974 Protocol.

[92] The Agreement includes only two articles which regulate the relationship with the Convention, whereas the substantive part is articulated in the Annexes of the Agreement.

[93] Treves (n 17) 471.

[94] As to the Appeals Tribunal, the Joint Committee ‘shall’ adopt a decision as to the administrative, organizational and procedural rules for the Appeals Tribunal to function, CETA (n 3) art 8.29. Hence, no prior implementation agreement seems to be needed.

[95] For an account of those States that have either renegotiated or abandoned the ISDS, see Hindelang (n 18) 11; A Roberts, ‘A Turning of the Tide against ISDS?’ EJIL:Talk! (19 May 2017) <www.ejiltalk.org/a-turning-of-the-tide-against-isds/>.

[96] Furthermore, albeit not uncontested, investors retain the possibility to include international arbitration in their investment contracts.

[97] E.g., while being quite obscure, art 80 of the Italian Constitution sanctioning the approval of the Parliament for those treaties implying ‘modification of the law’ would seem to include the hypothesis of direct effect of international substantive standard of treatment included in CETA.

[98] See Pantaleo in this issue on the multilaterality and mixity of the agreement. In several provisions, CETA refers to two parties only, see eg art 30.7 ‘A party shall notify.. the other Party’. Thus the question arise onto whether MS are contemplated at all in these provisions, in conformity with the list of Parties included in the preamble of the Treaty, or whether these provisions refer exclusively to the EU and Canada, as the text would explicitly suggest. While this question is out of the scope of this Article, the ambiguity as to who are the parties of the treaty produces some consequences in the application of CETA itself.

[99] See amongst others Case C-53/96­­, Hermès International (a partnership limited by shares) v FHT Marketing Choice BV [1998] ECR I-03603; Joined Cases C-392/98 and C-300/98, Parfums Christian Dior SA v TUK Consultancy BV [2000] ECR I-11307.

[100] See S Saluzzo, ‘Il diritto dell’Unione Europea e gli accordi internazionali degli Stati membri’ Doctoral Dissertation (Università di Palermo, 2016) 211 <https://iris.unipa.it/retrieve/handle/10447/163133/262594/Stefano%20Saluzzo%20-%20Tesi%20di%20dottorato.pdf>; E Neframi, ‘The Duty of Loyalty: Rethinking its Scope through its Application in the Field of EU External Relations’ (2010) 47 Common Market Law Review 323.

[101] Case C-104/81, Hauptzollamt Mainz v Kupferberg [1982] ECR 3641.

[102] According to S Saluzzo (n 100) 211, the duty of loyalty acquires an autonomous dimension in the context of shared competences.

[103] Hermès v FHT (n 99).

[104] Saluzzo (n 100) 213. Case C-459/03 Commission of the European Communities v Ireland [2006] ECR I-4657.

[105] Not to mention the risk of companies migrating to those MS allowing the ICS mechanism.