AG bot in opinion 1/17. The autonomy of the EU legal order v. the reasons why the CETA ICS might be needed
The EU’s exercise of its post-Lisbon competences over foreign direct investment (FDI) has been anything but smooth. In Opinion 2/15 the CJEU clarified the EU and Member State competences over the EU’s new generation free trade and investment agreements, resulting in the splitting of the EU‑Singapore agreement into a separate trade and investment agreement. Then, in Achmea the Court found investor-state arbitration (ITA) clauses under intra-EU BITs to be incompatible with EU law, which will result in the termination of almost 200 intra-EU BITs and the non-enforcement of ITA awards rendered under them within the EU. Now, everyone is anxiously awaiting the outcome of Opinion 1/17 – requested by Belgium under the insistencies of Wallonia – and whether the Investment Court System (ICS) under CETA is compatible with EU law. This opinion will not only affect the entry into force and conclusion of the trade and investment agreements with Canada, Singapore, Vietnam and Mexico, but it will have broader implications for the multilateral ISDS reform process and the EU’s investment policy.
Therefore, Advocate General Bot’s extensive opinion delivered on 29 January 2019 (first commentaries here and here) in which it found the CETA ICS to be compatible with EU law deserves scrutiny. I will only focus on the AG’s arguments concerning the exclusive jurisdiction of the Court of Justice over the definitive interpretation of EU Law. In a separate post, Harm Schepel will focus on the AG’s arguments on non‑discrimination.
The main points of AG Bot’s arguments
The AG extensively uses arguments from the Court’s prior case law on the relationship between EU law and extra-EU dispute settlement mechanisms (DSMs), as well as some novel arguments based on reciprocity and the lack of direct effect of CETA to conclude that the ICS is compatible with EU law. According to the AG:
a) Pursuant to the principle of sincere cooperation, the courts of the Member States share the responsibility with the Court of Justice to ensure the full application and integrity of EU law, a corner stone of which is the preliminary reference procedure. However, the preservation of the autonomy of EU law ‘is not a synonym for autarchy’.
b) International agreements concluded by the EU are an integral part of the EU legal order, CETA was denied direct effect by its drafters, and CETA needs to be entirely compatible with the EU’s constitutional principles and the Treaties. In essence, an international agreement with its own DSM is not incompatible with EU law, provided the decisions of the said mechanism do not ‘have the effect of binding the EU and it institutions, in the exercise of their internal powers, to a particular interpretation of the rules of EU law’.
c) The requirement of reciprocity and the need to protect EU investors when investing abroad must be taken into account when examining whether the CETA ICS adversely affects the autonomy of the EU legal order. International investment agreements are needed because:
- it is not guaranteed that EU investors in the other contracting party benefit from the same set of local remedies as foreign investors in the EU
- reciprocity is one of the guiding principle of EU external relations
- the free movement of capital and payments between Member States and third countries needs to be fully secured.
- unlike in the case of the Member States, EU-third country relationships are not based on mutual trust. Thus, the parties must ensure that a neutral international forum exists to settle the disputes.
d) The ISDS mechanism under CETA is needed because the contracting parties decided that CETA will not have direct effect and investors cannot invoke it before domestic courts.
e) The Court’s arguments in the Achmea judgment cannot be transposed to the present case, as the former is premised on a different reasoning:
- the relationship between Canada and the EU is not based on mutual trust. Thus, CETA cannot affect the principle of mutual trust and sincere cooperation.
- the CETA Tribunal, unlike the one in Achmea, can only take into account EU law as a matter of fact.
f) The guarantees included in the ICS are sufficient to protect the Court’s exclusive jurisdiction to give definitive interpretations of EU law and to safeguard the dialogue between Member State courts and the Court of Justice. The CETA Tribunal:
- only has jurisdiction over investor claims for damages resulting from the breach of the standards found in the investment chapter
- has no jurisdiction to apply EU law/domestic law or to determine the legality of such rules
- can only consider domestic law as a matter of fact
- has to respect the parties’ right to regulate for the public interest
- must follow the prevailing interpretation of domestic law given by domestic courts
- is bound by an interpretation of EU law given by the Court of Justice
- will deliver awards that will be binding between the disputing parties ‘only in that particular case’. Even if it were to interpret EU law, the Court of Justice could always dismiss such an interpretation, without triggering a breach of the EU’s international obligations.
- has a narrow margin to interpret the treaty, because a treaty committee can hand down binding interpretations of the agreement
- can be corrected by the Appellate Tribunal in case of errors
- would not have the jurisdiction to rule on disputes internal to the EU, unlike the international courts in Opinion 1/09 and 2/13
- does not have the power to define whether the EU or the Member States are respondent, which is to be determined pursuant to Regulation 912/2014.
g) The ICS will not affect the role of national courts to ensure the effective application of EU law:
- it has no jurisdiction to interpret EU law
- it only has alternative jurisdiction, because investors can choose domestic courts
- even if investors choose domestic courts over ITA, they will not be able to invoke CETA directly and those domestic remedies will be different
- Member State courts will not have to apply CETA and they can refer preliminary rulings to the Court of Justice.
h) The establishment of a mechanism for the prior involvement of the Court and the full review of the ICS awards by domestic courts is not necessary:
- such mechanisms would go against the intention of the contracting parties to establish a DSM meant to stand outside their judicial systems
- review of the ITA awards is possible based on public policy grounds
- the ICS – in the spirit of reciprocity – would also need to make preliminary reference to the domestic courts of the other contracting parties, which would run counter to the purposes of this DSM.
Autonomy v. reciprocity
By now, it is not a novelty that the Court of Justice has an uneasy relationship with other international courts. The likelihood that the Court will find another international court to be compatible with EU law is quite low, if one is to consider the Court’s long standing case-law (Opinions 1/91, 1/92, 1/00, 1/09, 2/13 and Achmea). The accession to the European Court of Human Rights, investor-state tribunals under intra-EU BITs, the proposed European Patent Court and the proposed EEA Court have all fallen ‘victims’ to this case-law. Thus, it is quite remarkable that in light of these cases AG Bot concluded that the ICS is compatible with EU law. Nonetheless, parts of his reasoning are prone to criticism.
a) Reciprocity is an external feature of the international agreement not an internal EU requirement for the agreement’s compatibility with EU law
The AG sets the tone of his Opinion in para. 59, in which he states that the preservation of the autonomy of the EU legal order ‘is not a synonym to autarchy’. This statement is meant to be a criticism of the Court’s earlier case law. It is well known that the Court always affirms that – as a general rule -other international DSMs are not incompatible with EU law, only to conclude in the end that the DSM in question can adversely affect the autonomy of the EU legal order. It is also commendable that the AG spends time explaining the importance of the EU as an international actor to enter into international commitments, including subjecting itself to the rulings of international courts. However, the reasoning of the AG is difficult to square with the Court’s analysis in previous cases.
Achmea is a good example of how the Court goes about checking whether an international DSM is compatible with EU law. The Court first lays out the internal requirements of the EU legal order that an international DSM must meet, such as:
- the protection of the autonomy of the EU legal order;
- the primacy and direct effect of EU law;
- the preliminary reference procedure and its quintessential role in ensuring the effective and uniform application of EU law;
- the role of national courts in applying EU law;
- the principle of mutual trust and sincere cooperation;
- the need to respect the limits of Article 344 TFEU
- the Court’s exclusive jurisdiction to give binding interpretations of EU law
Against these internal benchmarks, the Court then assesses whether the international DSM satisfies them. The AG, however, instead of commencing his analysis with the internal requirements of the EU legal order, spends a considerable number of paragraphs on reciprocity and the necessity to have international investment agreements and the ICS.
First, it is puzzling why ‘reciprocity’ should be taken into account when examining the possible adverse effects an international DSM might have on the autonomy of the EU legal order. As mentioned, the first part of the analysis should map out the internal requirements of the EU legal order. In this case, however, reciprocity is a feature of the international DSM and of the international agreement, not an internal EU requirement. Truth be told, reciprocity was one of the main arguments used by the Court to deny direct effect to the WTO Agreement. However, that line of cases law concerned the effects of an international agreement in the EU legal order and not the issue of compatibility of an extra-EU DSM with EU law. Furthermore, the Court used the reciprocity argument when looking at the features of the international agreement and not the internal requirements of the EU legal order.
Second, the AG further conflates the internal requirements of the EU legal order with the possible need to have ICS in CETA. According to the AG, the ICS is needed as a neutral DSM because it is not guaranteed that EU investors abroad will benefit from the same local remedies as foreign investors in the EU. Furthermore, because the drafters of CETA deny it direct effect, investors will not be able to invoke it before domestic courts.
I find this part of the reasoning of the AG the most problematic. Just as with the reciprocity argument, these two remarks concern the international agreement and the ICS, and not the internal requirements of the EU legal order. Whether or not an international DSM is compatible with EU law is a different question than whether or not that DSM is needed. Furthermore, the need to have or not to have the ICS and investor-state arbitration is a political argument, not a legal one. As the AG rightly notes in the beginning of his Opinion, investor-state arbitration is facing multiple challenges, some of which the EU wants to tackle via its ICS. However, there is no consensus on whether the ICS is the right way of mending some of these challenges and whether investor-state arbitration is needed at all.
Concerning the direct effect argument, the AG is right to point out that CETA’s Article 30.6 denies it direct effect in the sense of prohibiting private parties to invoke it before domestic courts. As Semertzi noted in 2014, the denial of direct effect in the text of the agreement is a more recent phenomenon in EU economic agreements. Whether this is a welcome development, depends on the perspective one takes. However, the lack of direct effect of an international agreement is a feature of the international agreement and is not part of the internal EU requirements that an international DSM must meet.
Furthermore, whilst it is true that in a domestic court a foreign investor could for example not invoke directly the national treatment obligation of CETA’s investment chapter, it is also true that there is nothing to deny the agreement indirect effect (consistent interpretation). Moreover, as an ongoing project of mine – on what happens in domestic courts with foreign investors prior to the commencement of international arbitration – shows, investors in domestic courts overwhelmingly rely on domestic law and not on international investment law.
b) Achmea is more than just mutual trust
The AG also argues that Achmea is not applicable to the CETA ICS because Achmea is premised on the intra-EU nature of the BIT in question and on the need to protect the principle of mutual trust between the Member States.
As I have previously argued, the effects of Achmea might go beyond arbitration under intra-EU BITs, depending on whether one takes a narrower or a broader reading of it. One could argue, that the ICS – unlike the arbitration tribunal under the Dutch-Slovak BIT – would be common to a number of Member States (all of them technically) and it does not directly touch upon the principle of mutual trust. However, one should not overlook the overall broad and formalistic nature of the Court’s arguments in its established case law on the relationship between international DSMs and EU law. Even a hypothetical interference with the autonomy of the EU legal order is enough for the Court to find an incompatibility. As I argue in the next section, there are plenty of things that can go wrong: the lack of a preliminary reference mechanism, the limits set by article 344 TFEU, the location of the ICS ‘outside the EU legal system’, the possible de facto interpretation of EU law, etc.
c) To interpret or not to interpret? That is the question
The part of the AG’s Opinion that raises the issues, which – in my opinion – are the crux of Opinion 1/17, is summarized above in points (f) to (h).
The power of the Court to interpret EU law in a definitive matter is a tricky issue. It is commendable that the Commission tried following the Court’s prior case law and included some safeguards on the possible interpretation of EU law by the ICS. However, as I have previously argued, the Court also looks at the hypothetical possibility of a foreign tribunal to give a binding interpretation of EU law, and this is where the formalistic safeguards might not stop the ICS from de facto interpreting EU law.
First, even if the CETA Tribunal is bound to follow the Court’s interpretation of EU law, there will be provisions not yet (fully) interpreted by the Court. One such provision is Article 344 TFEU. The Court has yet to clarify whether it precludes recourse to investor-State arbitration in agreements with third states. Due to this uncertainty, in the original Achmea arbitration that preceded the case before the Court of Justice, the arbitral tribunal in essence interpreted Article 344 when it found that it did not preclude investor-State arbitration (here, para. 209). The Achmea tribunal also interpreted the scope of primary EU law provisions on the free movement of capital and establishment.
Such an interpretation – whether de jure or de facto – then becomes part of an arbitral award that under Article 8.41 CETA is binding on the disputing parties and that can be enforced in all EU Member States and beyond. If delivered under the ICSID Arbitration Rules, then it also cannot be challenged based on public policy grounds. I would argue that such a de facto interpretation of EU law is a binding interpretation of EU law (for a different opinion here) as the award is binding in its entirety and not just its operational part.
Such an interpretation also has practical consequences. For example, the domestic courts of one Member State could enforce an ICS award (just as Romania originally began doing after the Micula I award) that includes an interpretation of EU law not yet endorsed by the Court of Justice. However, the courts of another Member State – where enforcement is also sought – could deny the enforcement of the award because it contains an interpretation of EU law not yet endorsed by the Court. Such an outcome would affect the uniform application of EU law. Moreover, it could also affect the mutual trust between Member State courts, as some might give precedence to their ICSID obligations, especially if it is not yet confirmed by the Court that the award runs counter to EU law. As the AG notes, the Court of Justice could dismiss such an interpretation as incorrect. However, this would only happen if one of the national court sends a preliminary reference to the Court.
Second, the AG argues that the CETA Tribunal has a narrow margin to interpret the treaty, because various treaty committees have the right to give binding interpretations of the treaty (Article 8.32.3 CETA). As I have previously argued, in practice such committee interpretations will be very rare and due to the requirement of unanimity, any interpretation will need the approval of the other contracting party.
Third, the AG uses a puzzling argument according to which the ICS will not affect the role of domestic courts to ensure the effective application of EU law, because even if investors choose domestic courts, they cannot apply CETA directly. However, as mentioned, there is nothing to stop the indirect application of CETA, so domestic courts could still apply it. Furthermore, as explained, the enforcement of an award in which the ICS de facto interprets EU law can in itself affect the uniform application of EU law.
Fourth, the AG’s arguments that there is no need for the ICS to refer preliminary references to domestic courts is bizarre. According to him, in the spirit of reciprocity, a preliminary reference system would entail that the CETA Tribunal could refer a question to the tribunals of the courts of the other contracting party as well. However, this would run counter to the purpose of the ICS to provide a neutral DSM for foreign investors.
There are several problematic elements with this reasoning. First of all, the AG once again conflates the internal requirements of the EU legal order with the features of the ICS and the necessity to have it or not. The idea of a preliminary reference system is to make sure that if the CETA Tribunal is unsure about the meaning of EU law (or other domestic law) it gets a clarification from the Court of Justice (or other domestic courts). This ensures that the courts of the Member States will not enforce an ICS award that includes an interpretation of EU Law not yet offered by the Court. The lack of such a mechanism will make it impossible for the CETA Tribunal to know the proper reading/interpretation of an EU law provision that has yet to be interpreted by the Court of Justice.
Such a preliminary reference mechanism is not unusual. The DSM under the Draft Withdrawal Agreement of the UK in Article 174 (and also the Draft Framework Agreement between Switzerland and the EU) provide that the inter-State arbitral tribunal shall request the Court of Justice to give a ruling on a question that concerns the interpretation of EU law or the UK’s compliance with Court judgments prior to its exit. Whether such a mechanism would work for the ICS depends of course on whether the Court would allow the ICS to submit such a reference, as it is not a ‘court or tribunal of a Member State’ in the sense of Article 267 TFEU.
It is also problematic that the AG assumes that such a mechanism would affect the neutral nature of the ICS because the courts of the other party – that might not be impartial – could also receive such a request. The ICS already has to follow the prevailing interpretations of domestic law given by the courts of either contracting party, regardless of the nature of the domestic judiciary.
In conclusion, the main criticism of the AG’s Opinion is that it conflates the internal requirements of EU law concerning the effects of an international DSM in the EU legal order with the features of that DSM and the potential need to have it created.