Case C-244/17 – Commission v Council: the centre of gravity test revisited in the context of Article 218 (9) TFEU

By Pieter Jan Kuijper


Case C-244/17 – Commission v Council (PCA with Kazakhstan) is one of the most recent cases in the long list of external relations cases and Opinions decided by the Court (in most cases in its Grand Chamber composition) since the entry into force of the Lisbon Treaty (beginning with Dai-Ichi Sankyo, Case C – 414/11).  These cases have covered many aspects of the horizontal balance of competences between the political institutions of the Union, as well as the vertical distribution of powers between the Union and its Member States, in the field of the external relations of the Union.

Inevitably sometimes both aspects are touched upon, as in the present case. On the one hand, there is the question of which institutions play, or should play, a role in the decision-making under Article 218(9); on the other hand questions arise which methods of decision-making should be followed, unanimity or qualified majority voting; whether this should be determined by which legal bases such decisions should be taken and which method should be used to select such legal bases. The first question seems – and is – simple at first sight, but raises an important question about democratic legitimacy. The second question seems very complicated, but – after reflection – can be easily decided on the basis of existing precedents.


It should be pointed out at the outset that this case, contrary to many other recent external relations cases, did not elicit great interest from the side of the Member States. In other such cases normally between half a dozen to 25 Member States (in Opinion 2/15) intervened, usually opting for the most restrictive interpretation of the new Lisbon provisions on external relations, thus creating the impression that the Member States did not really want the new Lisbon provisions on external relations any longer.[1] Not so in this case: there was no intervention by any Member State. Also the Parliament, as the institution excluded from the decision-making in Article 218(9), did not show any interest, although as co-legislator of Union legislative acts that in principle might be affected by binding decisions of treaty bodies it could have made some points of principle before the Court.

The subject of the case

The Enhanced Partnership and Cooperation Agreement is a mixed agreement between the EU and its Member States of the one part and the Republic of Kazakhstan of the other part. Its signature and provisional application were decided by a Council decision based on Articles 31(1) and 37 TEU (CFSP), Articles 91 and 100(2) TFEU (transport), and Articles 207 and 209 TFEU (trade and development cooperation) and the Agreement entered into force provisionally on 1 May 2016.  Like most such agreements the EU-Kazakhstan agreement was provided with a Cooperation Council at ministerial level, assisted by a Cooperation Committee at ambassadorial level and a number of Sub-committees, which each had the power to take decisions. The agreement charges the Cooperation Council with laying down the rules, in accordance with which itself, the Cooperation Committee, and its Sub-committees will function and what the division of tasks between these committees will be.

In order to fix the EU position in the Cooperation Council on this decision, the EU Council was to take a decision that, according to the joint proposal of the Commission and the High Representative, should be based on the procedural basis of Article 218(9) TFEU, in conjunction with Article 37 TEU, and the substantive bases of Articles 207 and 209 TFEU. This implied that the CFSP (Article 31(1) TEU) was omitted as substantive legal basis of the decision, whereas it had been included in the decision approving the provisional application of the Agreement with Kazakhstan.

The Council, however, modified the proposal and took all the legal bases that were mentioned in the Council Decision on signature and provisional application and adopted the position of the EU under Article 218(9) on that basis. That was the decision that was attacked before the Court by the Commission in order to have it annulled.

The interpretation of Article 218(9) TFEU

The first question that confronted Advocate-General Kokott and the Court was the interpretation of Article 218(9) and in particular what context was relevant for this interpretation. In this respect, it must be said that the Advocate-General (A-G) was clearer and more convincing in her reasoning than the Court, although their respective lines of argument ultimately led to the same result. The A-G emphatically rejects the Commission’s argument that all decisions of treaty bodies with legal effect, except those that affect the institutional structure of the treaty, have to be decided by the qualified majority of Article 218(8), first sentence (paras. 38-44). She advances the simple and decisive argument that the procedure mentioned in that sentence obviously refers to the procedure laid down in Article 218 preceding that sentence, i.e. the procedure of  opening negotiations, the treaty negotiation itself and conclusion of the agreements reached.

Therefore, the relevant context for the interpretation of Article 218(9) has to be found elsewhere, namely in Article 16(3) and 31(1) TEU on decision-making by the Council respectively in the TFEU and the CFSP context. This leads her to the conclusion that “normally a qualified majority is sufficient for decisions taken by the Council in the context of the communitised (sic) policy areas, whereas with respect to the CFSP as a rule the unanimity principle continues to apply”( paras. 45-48). If a decision by one of the treaty bodies of an agreement contains elements of both policy areas, a solution should be found by using the usual centre of gravity approach employed by the Court in order to determine the ultimate legal basis for such a decision (para. 58).

The Court reaches broadly the same conclusion but restricts itself very much to the four corners of Article 218 TFEU and its own case law. This is justified by recalling that Article 218 itself, by differentiating the decision-making procedures, takes account of the specific features of the different fields of EU activity, notably the CFSP, and thus already reflects the institutional balance established for each of those fields established by the Treaties (paras. 20-24). It recalls its judgment in Case C-73/14, which characterized Article 218(9) as setting up a simplified procedure, compared to the treaty-making procedure (para. 65 of that judgment). And the Court adds that in the light of paragraphs 6, 9 and 10 of Article 218 read together, this simplification is restricted to a limitation of the Parliament’s participation (para. 26).

The voting rule for the positions to be taken under Article 218(9) is simply absent and the applicable voting rule must be found by reference to Article 218(8). Where this paragraph 8 refers to the unanimity required for all agreements that cover a field for which unanimity is required for the adoption of an EU act, the Court concludes that this must be taken also to refer to CFSP acts. Thus the Court establishes a link between the voting procedure and the substantive legal basis of decisions taken under article 218(9), which also preserves the symmetry between the procedures related to the EU’s external policies and the procedures related to its corresponding internal policies, which is a principle also inherent in the text of Article 218(8) (paras. 27-34).

Accordingly the Court arrives at the same conclusion as the A-G. The voting rule under Article 218(9) has to be found by reference to the substantive legal basis. And hence both return to the rules and principles that the Court normally applies in deciding the single legal basis of any EU act in cases where prima facie there might be more than one legal basis, including the centre of gravity test (para. 37 and Opinion of the A-G, para. 58).

The centre of gravity test in theory and practice

From this point onwards the opinion of A-G Kokott and the judgment of the Court run in parallel, that is to say that the Court follows Ms Kokott, whose opinion is more elaborate than the judgment of the Court, which especially in the practical application of the centre of gravity test to the facts of the case relies almost fully on her opinion and thus can be much more succinct.

Referring to both older and newer case law of the Court the A-G concludes that the institutional provisions of an international agreement are an ancillary aspect of such agreements that should reflect the overall objective of the agreement and the legal basis that should be based on this overall objective and not vary from segment to segment of the agreement. Hence the rules governing the decision-making in such institutions of the agreement should also be governed by this overall objective that must be determined on the basis of the centre of gravity test. The Court follows her completely in her approach.

However, the Court leaves out a caveat made by the A-G to the effect that the centre of gravity approach should always be restricted to the horizontal division of powers and should never interfere with the vertical distribution of powers between the Union and its Member States (Opinion of the A-G, para. 59).

Where the application of the centre of gravity approach is concerned, A-G Kokott gives the impression of bending over backwards to do justice to the sensibilities of the Council by describing in extenso all the points in the preamble and in the first dozen articles of the Kazakhstan Agreement, which deal with the CFSP (para. 42). Only to conclude immediately afterwards that the aims and content of the Partnership Agreement as whole bear out that out that it is principally pre-occupied with trade and business and other subjects that belong to the TFEU, such as cooperation in many different areas.

Moreover a quantitative analysis of the provisions of the Agreement demonstrates that the large majority of the articles treat trade and business issues, and not just the traditional trade problems, but also newer questions such as trade and sustainable development and other subjects of development cooperation and other types of cooperation, notably in the area of freedom, security and justice. All of these are TFEU governed subjects and not part of the CFSP (para. 43).

Since, therefore, the whole Enhanced Partnership and Cooperation Agreement with Kazakhstan falls preponderantly within the domain of the TFEU, most notably the trade and development provisions of that Treaty (in spite of the legal bases, including the CFSP provisions of Articles 31 and 37 TEU, which the Council had recourse to when adopting the provisional application of the Agreement in May 2016), the institutional decisions of its treaty bodies should be based on these same substantive provisions, namely articles 207 and 209 TFEU (paras. 75-78).

As already mentioned above, the Court shadows closely, but very succinctly these logical steps set out in extenso by the A-G.


Strictly speaking, this case is merely about the legal basis for the positions to be taken by the Union in respect of decisions of a general nature of association or cooperation councils and committees of international agreements (Article 218(9) TFEU). It is obvious that the AG and the Court are of the view that if such decisions relate preponderantly to the CFSP or to TFEU external relations, the legal basis for such decisions should be found respectively in the TEU or the TFEU.

However, since the legal basis of such decisions, if they are of a general nature, should be found in the legal basis of the agreement as a whole according to the Court, this case indirectly decided the question whether in May 2016, by including CFSP provisions in the legal basis for the provisional application of the agreement, the Council had acted in error. The implicit affirmative response to this question seems to have shocked quite some Member States[2], which is somewhat incongruous with their original lack of interest to intervene in this case.

 However, this judgment should not have come as a surprise. Anyone who had analysed older case law carefully (Cases such as the old ECOWAS case from the time before the Lisbon Treaty and the newer cases relating to the agreements bringing Somali pirates before courts in Mauritius and Tanzania) should have realized that the Court would apply the centre of gravity test to situations and legal acts on the borderline between the CFSP and TFEU external relations. Moreover, it was not unpredictable that, in the light of the sheer lack of balance between the numbers of “political dialogue” provisions and trade and development provisions in such agreements, including the agreement with Kazakhstan, the result of a centre of gravity test might be unfavourable to the Member States’ traditional position.

Before Lisbon’s entry into force that traditional position had been that even a few provisions on political dialogue should be considered qualitatively so important that they could be used as “poison pills” – to put it in rather graphic terms – in order to lift what were essentially trade agreements out of the exclusive competence of the European Union and turn them into mixed agreements, in the form of either association or cooperation agreements. In those days, the problem of horizontal distribution of competences was different as the CFSP was formally outside the European Community, even if it used its institutions.

In addition, the common commercial policy involved only the Council and the Commission and not yet the Parliament, as the latter did not yet have the power of consent of trade agreements. Moreover, the so-called “pastis theory” of mixed agreements, according to which one drop of Member State competence (even when it was merely in the form of shared competence) would make a whole (trade) agreement mixed just as one drop of pastis makes a glass of water completely milky, was still widely accepted.[3]  Recalling this slice of pre-Lisbon history also serves to remind us that in spite of A-G Kokott’s theoretically justified warning that the centre of gravity test should be applied only to the horizontal division of power between the institutions and never to the vertical division of competences between the Member States and the Union, these questions are in practice closely linked, especially when one of the powers at issue in the horizontal division is an exclusive power, such as trade policy. These traditional practices followed in the Council by the Member States before the entry into force of the Lisbon Treaty have continued under different guises for some time after Lisbon as part of what was characterized above as the rejection of that Treaty by the Member States. They included extremely restrictive interpretations of new TFEU provisions, such as Article 207 in respect of intellectual property and services, and of Article 3(2) TFEU and the ERTA precedents, attempts to turn back the powers of the Commission as negotiator and external legislative representative of the Union, the attempt to create mixed agreements through hybrid legal acts, one half being an act of the Member States meeting in the Council, and the other half a Union act, and many other well-documented examples.

The Court’s analysis of the EU-Kazakhstan Enhanced Partnership Agreement, which is both mixed as between CFSP and TFEU (as is now permitted by Article 218 TFEU) and as between the Union and the Member States (as usually was the practice with pre-Lisbon Association broad Co-operation Agreements), makes clear that those days are over.  However, to what extent they will be over remains to be seen.

After this case, it is clear (if there was still any doubt) that the conclusion of EU international agreements covering both CFSP/CSDP and TFEU matters has to respond to the same criteria for deciding their legal basis as all Union acts. That is to say that they must preferably rest on a single legal basis that is selected by applying objective and judicially verifiable elements and that is decided, if necessary on the basis of the centre of gravity test. If an EU act, in this case an international agreement as the one with Kazakhstan, has distinct components, such as a CFSP component and a TFEU component, that in itself is not sufficient to decide on multiple legal bases, especially if one component is clearly dominant or preponderant and the other is incidental to the dominant one. As appears from our case, preponderance has a quantitative aspect (number of articles devoted to the one or the other component) and a qualitative aspect (weight or importance of the component). However, when such agreement has to be implemented by EU legislative or executive acts or by binding decisions of treaty bodies of such an agreement on which the EU must take a position, those acts or that position have to be taken on the same basis as the agreement concerned, if their scope is as broad as the original agreement. If, on the contrary, the scope of such acts or positions is preponderantly restricted to either the CFSP or the TFEU aspects of the agreement, the corresponding legal basis has to be selected from either of the TEU or the TFEU.

The recent Antarctic Protected Marine Areas and the notion of false mixity

However, whether Member States can continue their practice of voluntary mixity, when there is no absolute need to do so, is another and much more difficult question. Possibly not entirely by coincidence A-G Kokott has also dealt with this question in her Opinion in two joined cases about the creation of two Marine Protected Areas (MPA’S) in the Antarctic, in which she seeks to restrict what Allan Rosas has called “false mixity” and what she, following her colleague Szpunar, characterizes as an erroneous confusion between “shared powers” and “mixed external action.” She thus is of the view that there was no legal necessity for joint action by the EU and its Member States (paras. 104-110). However, with respect to the question whether Member States that were Members of the Commission for the Conservation of Antarctic Marine Living Resources (CCAMLR) had the right to voluntary participation in that treaty body because of their special position there her reasoning was rather nuanced. Her final conclusion, however, was in favour of the annulment of the prohibition to take a position on behalf of the Union alone (paras. 111-124).

However, just a few days ago the Court has rendered its judgment in the Antarctic case.  The problems brought up by A-G Kokott have not been fully taken up by the Court, because they were not discussed during the hearing and the Court, after consulting the A-G, implicitly decided not to reopen the hearing (paras. 52-55). For the rest this judgment is a textbook application of the centre of gravity test, as applied to positions of the Council to be taken in treaty bodies, in this case the CCAMLR. The application of this test in this case yielded the result that the creation of the Marine Protected Areas in the Wedell and Ross Seas off the Antarctic continent fell preponderantly in the domain of environmental protection (of shared competence) rather than in that of the common fisheries policy (exclusive competence). Hence the proposals concerning the MPA’s could be made by the Union and the Member States together.[4]

As was predicted at the beginning of this blogpost, the question of how to select the legal basis for positions of the Union to be taken in treaty bodies of international agreements, though complicated, could in the end be solved by simply having recourse to the standard methods of choosing a legal basis for Union acts, including the centre of gravity test. On the other hand the question of the influence of Parliament on such positions, though easily answered in the negative on the basis of the text of Article 218(9), raises questions of principle whether this text is adequate, given the important consequences that such positions and the decision resulting from them may have on international agreements which Parliament has consented to.

[1] See “From the Board: “Litigation on External Relations Powers After Lisbon: The Member States Reject Their Own Treaty”, Legal Issues of Economic Integration, 43, no. 1 (2016); 1-14.

[2] Personal information obtained by the writer.

[3] See on the “pastis theory” and its doubtful validity in the post-Lisbon era: Pieter Jan Kuijper, “Post-CETA: How we got there and how to go on”, RTDEur avril-juin 2017, pp. 181-187, at 184.

[4] Ibid., paras. 75-103. The case contains a largely novel part (paras. 125-133) on the consequences for the exercise of mixed competence by the Union alone, which pays close attention to the position of the Union and its Member States in the overall Antarctic treaty system, of which the CCAMLR is a part. This part deserves a separate analysis at a later time.


  1. Bashar H. Malkawi

    The issue of mixed agreements are very complex. Every trade agreement includes both trade and non-trade provisions. One single voting procedure should be followed. Dividing the agreement into several parts and follow different procedures for different provisions add complexity and indeed discourage other countries to enter into negotiations with the EU. The dominant rule should be followed. This rule will help achieve the objectives of the agreements in question. Bashar H. Malkawi

  2. Ramses A. Wessel

    Just to add one sentence to this excellent analysis by PJK: I feel that perhaps another interesting (innovative?) element in the Antarctic cases is that the Court uses international law arguments to point to the need to keep the Member States involved. In fact, in this case international rather than EU law considerations seem to be conclusive.

  3. Szilárd Gáspár-Szilágyi

    So I checked the Antarctic Treaty case mentioned by Profs. Kuijper and Wessel – and the one that will be commented on by Christina. For me paragraphs 126-132 are more than interesting. The CJEU starts by reiterating its original stumble in Op 2/15, corrected in the OTIF case (the one on which we commented on here Thus, the Council can still decide that in an area of shared competence only the Union might exercise the external competences (facultative mixity is still alive), BUT a limit can be set to facultative mixity by the need to be in ‘observance of international law’ (para. 127). In other words, if the EU does not have a fully autonomous status in an international body – but this status depends on the membership of the Member States – then facultative mixity can be curtailed. Or am I reading it wrong?

    • Bernardo Carvalho

      I am studying the External Relations of the EU Law, still trying to figure out what does “facultative mixity”, “center gravity test” “false mixity” means, can someone clarify? Thank you so much

  4. Pieter Jan Kuijper

    “Facultative mixity means that Member States in the Council, when concluding an international agreement, have a possibility to choose a mixed agreement or not. That is to say that a future agreement is not at all or not fully covered by the provisions of Article 3(1) and 3(2) TFEU (i.e. no exclusive competence) and, moreover, that there is no part of the agreement that falls under exclusive Member State competence or that there are no legal factors that compel Member States in another way to be a party to the agreement (there is some speculation as to what these other legal factors might be, see Szilard Gaspar-Szilagyi above). In the last two mentioned cases, there is no doubt that there is a legal requirement to conclude a mixed agreement.

    “False mixity” for some is identical to “facultative mixity”, since there is no legal obligation to have recourse to a mixed agreement. For others – and this is the more common view – “false mixity” occurs, if the Member States in the Council actively seek to introduce provisions in the agreement that are of shared or Member State competence with a view to make the agreement mixed, which it otherwise would not have been.

Leave a Reply