April 23
Anjum Shabbir
Anjum Shabbir
15th March 2021
External Relations & Trade Human Rights Institutional law

Op-Ed: “AG Hogan’s Opinion in Avis 1/19 regarding the Istanbul Convention” by Merijn Chamon

This Op-Ed will comment on Advocate General (AG) Hogan’s Opinion in Opinion procedure 1/19 regarding the Istanbul Convention. Of course, an Op-Ed does not allow for a fully-fledged analysis and will therefore limit itself to discussing the main points and flagging up the more remarkable findings in the AG’s Opinion. For the same reason, the political dimension to the ratification of the Istanbul Convention will not be covered either, and this Op-Ed will focus purely on the legal issues raised by the Convention.

Considering the constraints of space, the context of and background to Opinion procedure 1/19 will not be repeated here. Readers are instead directed to the earlier EU Law Live Insight published on the occasion of the oral hearing. Dispensing with this allows this Op-Ed to go straight to the discussion of AG Hogan’s Opinion, which will be structured around the three main legal questions at issue: the legal basis issue, the question of the splitting of decisions, and the question of the requirement of common accord. The AG’s analysis of the admissibility of the European Parliament’s request also raises interesting questions but to keep the focus exclusively on the questions which the Parliament puts to the Court of Justice, they will not be analysed here either.

Legal basis

To recall, the European 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 the signature, the Council of the European Union 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).

The analysis of the AG here is rich but quite complex, especially for those not fully familiar with the law of EU external relations. To properly understand the AG’s Opinion, two preliminary points need to be made:

First, as the AG rightly notes (points 66-85), the normal rules on ‘the choice of legal basis’ function a bit differently when it comes to choosing the legal basis of a decision on the signature or conclusion of a mixed agreement. Where for an internal legislative act (or an EU-only agreement), the assessment must be based on the objectives and content of the entire act, the test for a mixed agreement does not relate to that agreement in its entirety but only to those parts to which the EU will commit itself. The centre of gravity of the Council Decision concluding the Istanbul Convention may therefore be different from the centre of gravity of the Istanbul Convention itself.

To which parts of the Convention will the EU commit itself? The AG finds that for those parts coming under EU exclusive competence, in the sense of Article 3 TFEU, the issue is simple: since the Member States lack that competence, the EU must exercise its competence (point 91). For the parts coming under shared competence in the sense of Articles 4-6 TFEU, the AG puts forward that there is a (presumably unfettered) political choice to be made by the Council on whether it wishes the EU to exercise those competences. By taking that position (and by not identifying any legal check on this political choice, despite such checks being suggested by other AGs (see point 120 of AG Wahl in Opinion 3/15, and point 114 of AG Kokott’s Opinion in Joined Cases C‑626/15 and C‑659/16)) the AG essentially accepts that what the Court consistently refers to as ‘a test based on objective factors amenable to judicial review’ becomes subverted by the judicially unreviewable subjective will of the Council: the legal basis that will have to be relied on and which is proof of EU competence is determined by the extent to which the Council wishes the EU to exercise its competences. This premiss permeates the AG’s analysis of the legal basis-test and is consistently repeated throughout (see inter alia points 82, 84, 85, 87, 88, 89, 90, 91, 96, 112, 131, 135, 136, 159, 161, 164 and 166).

A second preliminary point is at the same time a point of critique. More than once (inter alia points 89 and 146) the AG suggests that his analysis is informed by the very specific case of the Istanbul Convention, namely that what ‘makes this case so specific [is] that the Union will not exercise all of the competences it shares with the Member States’. The AG is right that it might be unusual for the EU not to exercise the competences that accord to the agreement’s main objectives and provisions (and instead restricts itself to the competences according to an agreement’s subsidiary objectives or provisions). However, that the EU does not fully exercise its shared competences in its external relations is actually the rule which explains why many agreements are concluded as mixed rather than EU-only agreements.

In performing the legal basis-test in casu, the AG thus starts from the premiss that the EU will not fully exercise all its relevant shared competences and instead takes as a baseline the minimum common denominator on which all EU institutions agree that the EU should commit itself. This baseline effectively corresponds to the Council’s position, which is thus used to determine whether the Council rightly (intends to) alter the legal bases proposed by the Commission (point 90). The AG thereby notes that should the easy route be taken, and should the Council allow the EU to exercise its competences to the fullest, the relevant legal bases would simply be Articles 3(3) TEU and 19 TFEU (point 129).

The AG then proceeds to check three issues: (i) does the minimum common denominator properly take into account those parts of the Istanbul Convention that come under EU exclusive competence (and which therefore must be covered by the decision on conclusion); (ii) do these commitments coming under EU exclusive competence constitute the predominant aim and content of the decision on conclusion (thus requiring a separate legal basis); and (iii) are the other commitments which the EU intends to enter into predominant or merely ancillary? Evidently, in the latter case they would not require a separate legal basis.

As for the possible exclusive competences in play (i), the AG identifies a partially exclusive EU competence (as a result of the AETR doctrine) for judicial cooperation in civil matters and conditions of employment of EU officials and servants (points 143 and 145). However (ii), of those two, the former is ancillary (point 153) and therefore, only the latter requires a separate legal basis in the form of Article 336 TFEU (point 164). Here again the AG recalls that the Istanbul Convention’s predominant aim and content does not at all relate to how the EU treats its civil servants but that this is also not the relevant question to be answered: since the Council will make a choice to leave most of the competence for the Istanbul Convention to the Member States, the relevant question is what the predominant components are of what is left for the EU to agree to. And among the limited remaining aspects left to the EU, that of how it treats its own civil servants does constitute a main component. On the other commitments (coming under shared competence) (iii), the AG retains Article 82(2) TFEU as a legal basis, noting that the three political institutions were also all in agreement that at least this provision was one of the legal bases (point 152). The AG does not retain Article 83(1) TFEU as a legal basis since the substantive criminal law provisions in the Istanbul Convention only in some cases come under the heading of trafficking in human beings or of the sexual exploitation of women and children mentioned in Article 83(1) TFEU (point 155). The AG also retains Article 84 TFEU as a legal basis by reading that provision broadly and because the Istanbul Convention stresses the importance of the prevention of violence (points 156-159). The AG finally also retains Article 78(2) TFEU as a legal basis. Hold on, the reader familiar with the Council’s decisions on the signature of the Istanbul Convention will now say. Didn’t the Council in its decision on signature identify the Convention’s provisions on asylum as coming under EU exclusive competence? Indeed, the AG disagrees on this point with the Council and finds (applying Opinion 2/91) that there is no AETR effect, since both the common rules and the Convention’s provisions lay down minimum requirements (point 104). Since it has also been argued that despite these minimum rules, the Convention may still very well interfere with EU asylum law (thus resulting in an AETR effect), it will be interesting to see if and how the Court engages with the AG’s proposed solution. Coming back to the AG’s assessment of the EU’s shared competence on the Convention’s provisions related to asylum, he finds these to be substantively significant (again, especially given the limited commitments which will be left to the EU to commit to) since it requires parties to recognise violence against women as one of the forms of persecution that may give rise to refugee status (points 160-162).

All this means, according to the AG, that the proper legal bases are Articles 78(2), 82(2), 84 and 336 TFEU.


Splitting of decisions

A second issue on which the European Parliament queries 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 (notably Articles 60 and 61 of the Convention), just like it split the decision on signature. The argument of the Parliament here appears to be that since these relevant provisions of the Convention are already largely covered by EU common rules which Ireland follows internally, Ireland could no longer opt-out in the conclusion of the Convention (point 185). The AG refutes the Parliament’s reasoning on two accounts.

The AG first puts forward that it is difficult to see which requirement of EU law prohibits the splitting of decisions (point 172), let alone that such a requirement, if it were to be found, would constitute an essential procedural requirement (point 175). Under the Court’s well-established case law on Article 263 TFEU however, only the violation of essential procedural requirements will give rise to the annulment of the contested act. The AG thereby dismisses the idea that the Council splitting a decision could undermine the prerogatives of the Parliament or Commission. Yet, assuming there is no need to split (and this is a crucial assumption, see below), it is difficult to argue that those prerogatives are not affected: if there is no objective need to split, splitting unduly frustrates the exercise of Parliament’s scrutiny prerogatives under Article 218 TFEU and frustrates the Commission’s task to act on behalf of and watch over the general EU interest, reflected in Articles 17(2) TEU and 293 TFEU. This would be so even if the Commission has agreed to the splitting (as it did, see point 12) since an EU institution cannot, even voluntarily, dispense with its own prerogatives. Drawing on the MFA case, the AG suggests that Article 293 TFEU only applies to the legislative process (point 174) but that would be an awkward reading of MFA and Article 293 TFEU. The latter lays down a general rule and in MFA the Court merely noted that it applies especially in the legislative process, not only in the legislative process.

Of course, even if there is no requirement prohibiting splitting, the decision to split (pursuant to Protocol No 21) would still be unlawful if this violates Protocol No 21 itself. This is the second point on which the AG disagrees with the Parliament. The AG believes that splitting the decision would instead be required by Protocol No 21 which in its Article 4a provides that the Protocol also applies to measures proposed or adopted that amend existing measures which Ireland has agreed to participate in. AG Hogan thus reads Article 4a as also preventing an AETR effect from arising, otherwise, according to the AG, ‘Article 4a of Protocol No 21 would be devoid of any real meaning’ (point 188). However, this finding is rather puzzling and seems to disregard the nature of the AETR effect which is produced by the common rule (to which Ireland agreed to participate) and which cannot be equated with an amendment or modification since it is triggered solely by the risk that common rules may be affected (namely it is not triggered by the de facto modification or amendment of common rules). Unfortunately, the AG does not further explore the question whether Ireland could then at all be a party to the Istanbul Convention, since the AG accepts that even if not bound through the EU, Ireland may still not undermine the common rules to which it agreed under Protocol No 21 (point 190). It will be interesting to see if and how the Court addresses this point, since that would further clarify how AETR interacts with Protocol No 21 and perhaps the nature of the doctrine. Indeed, under the AG’s reasoning, the AETR’s pre-emptive effect (Ireland cannot act externally in a way that would undermine the common rules to which it agreed) is effectively dissociated from the ‘creation’ of an exclusive competence for the EU (the EU not having competence, according to the AG, to commit Ireland without its agreement), whereas as normally both go hand in hand.


Common accord

Finally, the European Parliament’s second question will for the first time allow the Court to address the legality of the contested practice of finding a common accord between Member States (by unanimity) on the conclusion of a mixed agreement, before the Council will allow the EU to conclude the agreement (even if the Council could do so with a qualified majority).

According to the Parliament, putting EU-decision-making on hold until all Member States agree goes beyond ‘sincere cooperation’ and instead de facto changes the voting rules within the Council, thereby resulting in hybrid decision making (point 196) which the Court prohibited in the Hybrid acts case.

On this the AG recalls that following the signature of an agreement (such as the Istanbul Convention), the only international law obligation on the EU results from Article 18 of the Vienna Convention (VCLT), meaning there is no obligation under international law on the Council to initiate the conclusion of the Istanbul Convention. On whether such an obligation exists under EU law, the AG is less clear. Above it was recalled that the AG suggested that the Council is entirely free to decide which shared competences of the EU it will elect to exercise (since only the exclusive competences must be exercised). When it comes to taking the decision to conclude however, the AG notes that the Council has a large margin of discretion (point 200) which suggests that it is still judicially reviewable and that hence there must be some relevant legal standard to test against. However, the AG does not identify that legal standard, since he finds that the Parliament’s argument is in any event incorrect: by waiting for all Member States to agree, the Council is not altering or ‘hybridising’ the voting procedure. Here the AG even adduces reasons justifying the Council’s approach, since if the EU would conclude in absence of a common accord and the mixed agreement would enter into force, the EU could be held jointly liable for Member States’ (in)action, even if those Member States act pursuant to retained national powers and are not themselves party to the agreement in question (point 205). While in general such problems may be pre-empted by the EU expressing a reservation, the Istanbul Convention does not permit reservations that could be used for this purpose (point 207). The Istanbul Convention, unlike a lot of other mixed multilateral agreements, does not contain a RIO-clause either. Because of this, the AG finds that the EU cannot adopt a declaration of competences, setting out the limited extent to which it commits itself and its Member States, since this would amount to a prohibited reservation (points 208-210). Of course, had there been a RIO-clause in the Istanbul Convention, the situation would have been entirely different, which underscores the need for the Commission to be involved, on behalf of the EU, in the negotiation of these agreements, which in this case it was not.

At the same time as finding that the Council does not manifestly err when waiting for the common accord of the Member States, the AG finds that neither is the Council required to wait for such a common accord (point 222). Incomplete mixed agreements, because of the EU acceding without all its Member States or because of Member States denouncing the agreement following its entry into force, result in practical problems but these problems do not legally prevent the EU from becoming (or remaining) a party to a mixed agreement (points 221-225).

The AG’s findings on the issue of common accord will undoubtedly disappoint the European Parliament (and the Commission) but in light of the Court’s post-Lisbon external relations case law it would not be surprising for the Court to follow the AG on this point. While the Court has been enthusiastic in finding EU exclusive competence wherever possible, its decisions in COTIF I (paragraph 68) and AMP Antarctique (paragraph 126) also make clear that it does not wish to intervene in the political process when the Council has to decide on the exercise of EU shared competences.


Merijn Chamon is an 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. This Op-Ed benefited from comments by Thomas Verellen, Bruno De Witte, and Peter Van Elsuwege. All errors or omissions remain those of the author.


Your privacy is important for us

We use cookies to improve the user experience. Please review privacy preferences.

Accept all Settings

Check our privacy policy and cookies policy.