Op-Ed: “The Court’s Opinion in Avis 1/19 regarding the Istanbul Convention” by Merijn Chamon
On 6 October 2021 the Court of Justice (EU) ruled that for the conclusion of the Istanbul Convention, the EU is not required to wait for the completion of the ratification by the Member States but neither is the EU required to go ahead and complete its own ratification of the Convention before the Member States have done so. While this may be the main take home message from the Opinion, the Court also dealt with finer points of the law governing the EU’s external relations and the procedure before the Court.
This Op-Ed will build on earlier publications on EU Law Live, notably the Insight published on the oral hearing and the Op-Ed on AG Hogan’s Opinion. Just like the latter Op-Ed, this one will be structured around the three main legal questions at issue: (i) what should be the legal bases used to accede to the Convention; (ii) should the signature and concluding decisions be split to take into account the variable geometry of Ireland’s participation to measures in the Area of Freedom, Security and Justice (pursuant to Protocol No 21); (iii) and can the Council wait for a consensus between Member States on their individual accession to an international agreement before deciding by qualified majority to let the EU accede to that agreement. In addition, this Op-Ed will also look into the partial inadmissibility of the Parliament’s questions.
Before turning to these issues a remarkable and very welcome aspect of this Opinion should be stressed: in other recent Opinions, the observations submitted to the Court are usually summarized very briefly. In quantitative terms they take up around 10% of the entire Opinion (hitting a low of only 3% in Opinion 2/15). In Opinion 1/19 however, the summary of the observations takes up more than a third of the entire Opinion which effectively adds greater transparency to the Court’s rulings since neither the hearings nor the written submissions of the EU institutions and especially those of the Member States are pro-actively made public.
The Court first had to deal with the issue of admissibility of the questions since the Parliament had asked about the proper legal basis to sign and conclude the Istanbul Convention, even though the Council had already adopted its decision on signature. Was it possible to use the Opinion procedure to call into question the legality of an act that has already been adopted by the Council? This recalls earlier inter-institutional allegations of a possible abuse of the more flexible Opinion procedure to achieve outcomes that can only be secured through an infringement action or an action for annulment (see Opinion 1/13, para. 35 and for a reverse argument Case C-29/99, para. 52). In this case it could be argued that the Parliament should have brought an action for annulment (against the decisions on the signature of the Convention). On the latter the AG applied Textilwerke Deggendorf (TWD) by analogy (para. 40), even if so far,TWD only applied to non-privileged parties or to Member States that have omitted to challenge decisions addressed to them. The Court itself did not pursue the TWD route. It extensively recalled its settled case law and the purpose and possibilities of the Opinion procedure (paras 192-205). Subsequently it noted that the preventive purpose of the Opinion procedure could not be fulfilled anymore in relation to the signature decisions (para. 218), further noting that the Parliament had indeed been in a position to bring an action for annulment against those decisions (para. 219). Since this inadmissibility issue did not play for the future Council decisions on the conclusion of the Convention, the Court was still in a position to answer the other questions of the Parliament and to address the three substantive issues noted above.
To recall, the Commission had proposed to sign and conclude the Istanbul Convention based on Articles 82(2) TFEU (judicial cooperation in criminal matters) and 84 TFEU (supporting measures in crime prevention). In its decision on signature, the Council dropped the latter provision and instead added Articles 78(2) TFEU (asylum) and 83(1) TFEU (definition of crimes and sanctions for serious crimes such as human trafficking and sexual exploitation). Behind this disagreement on legal bases lies a disagreement on the extent to which the EU should become a party to the Convention and (conversely) the extent to which the EU should leave a legal space to the Member States to be parties in their own right to the Convention. In this case, like in most cases where mixed agreements, i.e. international agreements concluded with third countries whereby both Member States and the EU are formal parties on the EU side, are at issue, the Member States in Council prefer to restrict the EU participation to what is strictly legally necessary. This means that the EU would only exercise its exclusive competences.
On this the AG thereby noted that if the Council would simply allow the EU to exercise all its relevant shared and exclusive competences, the legal bases would be Articles 3(3) TEU and 19 TFEU (para. 129). Accepting that the Council can elect not to exercise the EU’s shared competences (but must of course exercise any exclusive competences), he suggested that the legal bases should be Articles 78(2), 82(2), 84, and 336 TFEU.
The Court settles the legal basis issue by first stressing that the correct legal bases will depend on the ‘agreement envisaged’ which is not the Istanbul Convention as such but only the parts to which the EU will commit itself (para. 278), following the suggestion of the AG (paras 66-85). The problem here of course is that there is a (largely political) disagreement on the extent to which the EU should become a party to the Convention (see above). The Council wants to limit the EU’s accession to what is strictly legally necessary (‘narrow’ EU accession). The Parliament prefers that the EU would exercise its competences (regardless of their nature) to the fullest extent (‘broad’ EU accession).
In this respect the Court first suggested that, differently from the AG, it would not simply yield to the viewpoint of the Council, and noted that it would approach the issue in light of the decisions on signature (the Council’s position) but also in light of the actual question put to it by the Parliament (para. 282). It recalled its well-known established case law on choice of legal basis (paras 284-288) and applied it to the case at hand. According to the Court both the context (para. 289) and the content (para. 290) clearly indicate that the ‘envisaged agreement’ is about the protection of women against violence, while the dual aim is to further the equality between men and women and to combat violence against women (para. 291). Having looked into these objective factors, the Court, just like the AG, proceeds by accepting that the assessment must also be informed by the subjective preferences of the Council. The Court thereby explicitly notes that the “limited purpose of the act concluding the envisaged agreement is confirmed by the substantive legal basis referred to [in both] signature decisions” (emphasis added) (para. 293) and this even though the Court earlier explicitly recalled that a decision on conclusion is “in no way a confirmation of the [decision on signature]” (para. 201). Clearly this does not fully hold true, since the Court accepts that the legal bases for signature will pre-determine those for conclusion and this because the choice of legal basis test is, in reality, not determined by objective factors but by the subjective will of the Council.
The Court’s premise thus boils down to the view of the Council as expressed in the signature decisions, and this premise is not merely a (rebuttable) presumption. This is remarkable given that the agreement is still only envisaged, and the proper approach would be to conditionally identify the legal bases subject to the caveat that the political preferences of the Council and Parliament could still change (just like AG Sharpston in Neighbouring Rightsconditionally identified the nature of the EU ‘s competence in relation to the to be negotiated Convention on the protection of the rights of broadcasting organisations, see para. 166) The outcome of the Court’s assessment is then the same as that of the AG, finding that that the legal bases should be Articles 78(2), 82(2), 84 and 336 TFEU. The Court thereby held that some provisions of the Convention were merely incidental: in line with established case law (see e.g. para. 70 & para. 276), the institutional provisions of the Istanbul Convention were held to be ancillary to the substantive provisions which they accompany (para. 309). Similarly, the Convention’s provisions dealing with the particularly serious crimes that come under Article 83(1) TFEU are so limited as to be incidental (para. 301). By contrast, even if the Convention’s provisions on asylum are quantitatively very limited (which since the Kazakhstanand Armenia cases is an important element to determine the ancillary nature of an agreement’s provisions), they form a separate chapter of the Convention and impose important substantive obligations precluding them from being considered incidental and requiring the addition of Article 78(2) TFEU (para. 304).
Splitting of decisions
A second issue on which the Parliament queried the Court is the Council’s intention to split the decision on conclusion to allow Ireland, pursuant to Protocol No 21, not to be committed through the EU to the Convention’s provisions on asylum since Ireland has not opted in to the Directives for which an ERTA effect results in exclusive EU competence.
Unlike the AG, the Court does not engage with the possible consequences of an ERTA effect arising from common rules that come within the scope of Protocol No 21. Instead it followed a systematic reading of the entire Protocol (para. 322) and reiterated (as per established case law) that the applicability of the Protocol depends on the legal basis relied on (para. 325). The Court’s reiteration of this principle is also important for the pending case asking about the continuance of the European Arrest Warrant in relation to the UK after Brexit. Since the Withdrawal Agreement and the Trade and Cooperation Agreement are not based on a provision coming under Title V of Part three of the TFEU, Ireland arguably did not need to explicitly opt-in to allow for this continuation.
Since (all but one of) the legal bases identified by the Court for the conclusion of the Istanbul Convention come under Title V of Part three of the TFEU, the Protocol also applies to the adoption of the decision on concluding (and signing) the Istanbul Convention. However, given the Court’s systemic reading of Protocol No 21 and the fact that Article 3 of the Protocol allows for Irish participation in ‘the proposed measure’, the Court adds that splitting decisions is in principle impossible and opting in or staying out is only possible for the measure as a whole. Without saying so explicitly, this suggests that the Council’s decisions on signature are invalid (para. 328). It should be clear that this solution is to be preferred over the one proposed by the AG who suggested that splitting should principally be allowed (para. 172). Indeed, the Court turns this suggestion upside down and only exceptionally allows the splitting of decisions. But this only so if an objective need requiring such splitting can be shown. According to the Court, this would be the case when the measure envisaged (here the Istanbul Convention) only partially comes under the scope of Protocol No 21 or Protocol No 22. As the Court notes, such a situation presents itself in casu, since the Court found that Article 336 TFEU (dealing with the EU’s own staff) should also be a legal basis.
Splitting the conclusion decision will therefore be possible but not in the way the Council envisaged. As noted above, the Court did not go into the repercussions for the ERTA doctrine. It should be recalled here that some common rules (for which Ireland used its opt-in under Protocol No 21) are affected by the Istanbul Convention, resulting in an ERTAeffect and therefore a need for the EU to exercise this exclusive competence. The Court’s solution suggests however that this ERTA effect only applies for the other Member States that are bound by the common rules and not for Ireland. Instead, for Ireland, the future decision concluding the Istanbul Convention would amount to an act amending those common rules to which Ireland again has the option to opt-in (or stay out) under Article 4a of Protocol No 21. While this is in no way explicit in Article 4a, that Article would thus constitute an empowerment in the sense of Article 2(1) TFEU: even if the common rules that are binding on Ireland result in an ERTA effect and therefore in an exclusive EU competence, Article 4a of Protocol No 21 authorizes Ireland to still exercise a competence which has become an EU exclusive competence.
The issue of common accord
Although the issue of the practice of relying on a common accord was the subject of the Parliament’s second (and last question), the Court found it appropriate to examine it first (without explaining why) (para. 229). As one commentator(pp. 233-234) put it “[t]he Council practice for agreements which do not entirely fall within the exclusive competence of the Union […] is not to vote in accordance with the majorities provided for in Article 218 TFEU, but to apply in practice consensus by following a procedure which […] consists in ‘verifying’ first whether there is ‘common accord’, and only then, once there is ‘common accord’, voting in accordance with the procedure provided for in the Treaties (qualified majority voting, if applicable). In practice, the Council considers that Member States are to some extent free to decide on the practicalities of determining that such ‘consent’ or ‘accord’ exists. This way of proceeding of ‘waiting’ for the ‘common accord’ is often presented by the Council as being in accordance with the principle of sincere cooperation, as the Union should not act before having checked whether its Member States have a problem or disagree, as the agreement also engages competences of the Member States.”
The Council resorted to this ‘sequenced’ common accord approach after the Court prohibited in Hybrid acts, the practice whereby the Council and the Member States (as sovereign actors under international law) jointly adopt a single decision to conclude mixed agreements on behalf of the EU and its Member States. This was not permissible, according to the Court, since it confuses the decision-making under the EU Treaties with intergovernmental decision-making. As was envisaged already at the time of the Hybrid acts case however, this prohibition might merely result in the supranational decision-making being held hostage by intergovernmental blockages, as the ‘common accord’ approach illustrates. Could the Council then be forced to break free and go ahead despite their not being a common accord at the intergovernmental level?
The Court takes the same vantage point as in Hybrid acts and recalls that the EU Treaties established a new legal order (para. 230), whereby the procedures for decision-making are not at the disposal of the Member States (or the institutions). In this regard, also the principle of sincere cooperation cannot be relied upon to alter the decision-making procedures prescribed by the Treaties (para. 242). In this given case, the Court stresses that it is commonly agreed that the Istanbul Convention will be a mixed agreement, that the Parliament’s consent will be required for conclusion and that the Council will have to decide by qualified majority (para. 239). The Council can therefore not invoke the requirement of a common accord between the Member States on the conclusion of a mixed agreement as a condition before it decides on the conclusion of that mixed agreement by the EU, since this would alter the decision-making procedure prescribed by Article 218 TFEU (para. 245). This important statement of principle is of course still relative, since it merely means that the Council has to avoid referring to a lack of common accord among Member States when it is asked to explain why it does not conclude a mixed agreement. The Court is lucid enough to also recognize this, since it notes that the Council still has to muster the required majority when concluding an agreement (para. 250). In very clear terms the Court goes on to state that “both the decision whether or not to act on the proposal to conclude an international agreement, and, if so, to what extent, and the choice of the appropriate time to adopt such a decision fall within the Council’s political discretion” (para. 252). The prospect of identifying useful legal limits that constrain this discretion (see e.g. here and here) has therefore dimmed considerably. The Council can legitimately decide to continue discussions in order to build a bigger consensus between the Member States (paras 253-254). At the same time the Court points out that a vote may be ‘forced’, since according to Article 11 of the Council’s Rules of Procedure, the Council Presidency must hold a vote on an issue if one Member State or the Commission so requests and a majority of the Member States agree to holding a vote (para. 255).
Having confirmed the Council’s political discretion, the Court also rejects the arguments by some of the Member States, and even the Council itself, suggesting that pushing ahead with the conclusion in absence of a common accord would be illegal under EU law or result in the liability of the EU under international law. Indeed, when the EU would accede without all its Member States, the EU cannot exercises the competences of those Member States (para. 264). As noted by Kübek in an earlier Analysis of the Opinion on EU Law Live, the Court’s reason why under international law, the EU’s liability would not arise is far from convincing.
A further puzzling part of the Court’s Opinion on this point (and also on the issue of the legal basis) is how it engages with its own statement in the AMP Antarctique case that the EU needs to exercise its powers ‘in observance of international law’. The Commission relied on this to argue that the EU should go for a ‘broad accession’ (see above) in order not to frustrate the objectives of the Istanbul Convention (para. 283), whereas the Council relied on it to argue in favour of waiting for a common accord, since without all the Member States being parties to the Convention the EU would not be able to ensure a proper implementation of the provisions to which it would commit itself (para. 270). The Court dismissed both sets of arguments (paras 272 and 283) by noting that the purpose of the Opinion procedure is to assess the compatibility an international agreement with EU law, not the other way around. Evidently, the Opinion procedure should not be used to assess hypothetical future (in)action of the EU that could lead to its liability under international law but when that future (in)action is pre-ordained by the way in which the EU accedes to an agreement it would run counter to the telos of the Opinion procedure were the Court not to look into the issue. In addition, the Court seems to view the requirement to respect international law as an external obligation imposed on the EU by international law, disregarding Article 3(5) TEU which makes clear that the strict observance of international law is an obligation that is internalized in the EU’s own constitutional charter.
Specifically how the Court dismissed the Council’s reliance on AMP Antarctique highlights the confused logic inherent in AMP Antarctique. To recall, in that ruling the Court held that international law may prevent the EU to exercise its shared competences on its own and instead require it to act jointly with the Member States. Arguably, that logic only applies to the EU’s shared competences and depends on the specific legal framework in international law. It is remarkable then that while the Court in Opinion 1/19 starts from the premise that the EU will limits itself to exercising its exclusive competences when acceding to the Istanbul Convention (see above), it does not at the outset reject the Council’s AMP Antarctique argument on the basis that the AMP Antarctique doctrine arguably only applies to the EU’s shared competences.
A way forward
The Parliament returns rather empty handed from its trip to Luxembourg. The Court confirms that the extent to which the EU accedes to the Convention is decided by the Council (which the Parliament can of course reject, but then the EU does not accede at all). While incomplete mixity may be a sign of the absurdity of mixity, the Court accepts the possibility of the EU concluding a mixed agreement to which not all its Member States are parties. The Court even identifies a way forward (a simple majority of Member States forcing a vote in the Council) but this still requires a qualified majority of Member States to disregard diplomatic conventions within the Council. Will recent initiatives in a number of Eastern Member States create the momentum to do so? In 2018, the Bulgarian Constitutional Court held that the Convention goes against the Bulgarian Constitution. In 2019, the Slovak Parliament called on the Government to terminate the ratification procedure. In 2020 the Hungarian Parliament did the same. And in 2021 a proposal was put forward in the Polish Parliament to make Poland withdraw from the Convention. The (other) Member States that have ratified the Istanbul Convention muster the necessary qualified majority to proceed and allow the EU to accede to the Convention as well and if the Commission alters its proposal, they could opt to go for the broadest possible accession of the EU to the Istanbul Convention. As noted by the AG, this would result in Articles 3(3) TEU and 19 TFEU as legal bases and by virtue of Article 216 TFEU, this would bind all Member States, including the recalcitrant ones.
Merijn Chamon is Assistant Professor of EU Law at Maastricht University, Visiting Professor at the College of Europe (Bruges) and voluntary collaborator at the Ghent European Law Institute. I would like to thank Graham Butler, Thomas Verellen, Bruno de Witte and Gesa Kübek for their comments on an earlier draft. All errors or omissions remain mine.