Navigating the Labyrinth of Jurisdiction in Restrictive Measures Cases: Neves 77 Solutions (C-351/22)

Table of Contents: 1. Introduction – 2. Facts of the case and questions referred – 3. Opinion of Advocate General Ćapeta – 4. Judgment of the Court – 5. Analysis of selected issues concerning the Court’s jurisdiction – 5.1. In the shadow: fundamental rights protection as a question of competence – 5.2. Categorization of restrictive measures: binary constitutional choices meet complex political realities – 5.3 The comeback of Article 40 TEU: a missed opportunity? – 6. Concluding remarks

Abstract: On 10 September 2024, the ECJ delivered its judgment in Case C-351/22 (Neves 77 Solutions). In this case, the Court was confronted with the question whether it has jurisdiction to hear a preliminary reference on the interpretation of a CFSP Decision concerning restrictive measures. While the judgment constitutes an important clarification with regard to the ECJ’s jurisdiction in the area of CFSP matters, it appears that the relevance of the ruling was largely underappreciated when compared to the judgment in Joined Cases C-29/22 P and C-44/22 P (KS and KD v Council and Others), which was delivered on the same day. In this annotation, I explore the complex legal questions concerning the Court’s jurisdiction underlying the Neves 77 Solutions ruling, discussing both merits and flaws of the judgment. While at first sight Neves 77 Solutions might seem relatively unspectacular, I argue that the implications of the judgment are highly significant, particularly in view of the EU’s envisaged accession to the ECHR.

Keywords: CFSP – jurisdiction – carve-out – claw-back – restrictive measures – mutual non-affection clause.

1.   Introduction 

Does the European Court of Justice (ECJ) have jurisdiction to answer a preliminary reference from a national court asking for an interpretation of a Common Foreign and Security Policy (CFSP) Decision on restrictive measures? As is well-known, the jurisdiction of the Union courts in the area of the CFSP is subject to certain limitations. And while studying the text of the Treaties might give the impression to the uninitiated that there is indeed very little role for the Union courts in this policy field, the post-Lisbon case law has significantly narrowed the jurisdictional carve-out for the CFSP and at the same time widened the jurisdictional claw-back in the area of restrictive measures.[1]

As far as the latter is concerned, the Court in Rosneft ruled that its jurisdiction to rule on restrictive measures against natural or legal persons would not only encompass annulment actions brought in accordance with Article 263(4) TFEU, but also extend to preliminary references on the validity of such restrictive measures.[2] Only a few years later, the Court in Bank Refah Kargarandecided that actions for damages would also be covered by the claw-back provision.[3] The question of preliminary rulings on the interpretation of restrictive measures, however, constituted an entirely novel issue that had not arisen until the Regional court of Bucharest issued a preliminary reference of this kind to the ECJ in the case of Neves 77 Solutions.[4]

In this case note, I will explore the intricate jurisdictional issues that underlie this seemingly unspectacular and little appreciated case, arguing that there is far more than meets the eye in view of the broader implications of the judgment. 

2.   Facts of the case and questions referred 

In 2019, the Romanian company Neves 77 Solutions acted as an intermediary for a transaction between a Ukrainian and an Indian company involving the delivery of 32 radio sets originally manufactured in Russia.[5] The Department of Export Control of the Romanian Ministry of Foreign Affairs subsequently informed Neves 77 Solutions that the brokering transaction came within the scope of CFSP Decision 2014/512[6] concerning restrictive measures in view of Russia's actions destabilising the situation in Ukraine.[7] Although Neves 77 Solutions disputed these allegations on legal grounds, the Tax Fraud Department of the Romanian National Tax Administration Agency imposed an administrative fine on Neves 77 Solutions (equal to around 6000 euros) and ordered the full confiscation of the proceeds resulting from the contract with the Ukrainian company (equal to almost three million euros) for infringing Article 2(a) of CFSP Decision 2014/512 (prohibiting the provision of brokering services in relation to military equipment) as well as several provisions of the domestic Decree-Law on the implementation of international sanctions.[8] Neves 77 Solutions challenged this decision before Romanian courts, arguing that its conduct did not in fact constitute an administrative offence as alleged by the Tax Fraud Department and that, in the alternative, a full confiscation measure would be disproportionate and be in breach of its right to property under Article 1 of Protocol No. 1 to the European Convention on Human Rights (ECHR).[9]While the first instance court dismissed the action, the Regional Court of Bucharest decided to refer three questions to the European Court of Justice: 

(1) Can [Decision 2014/512], in particular Articles 5 and 7 thereof, in the light of the [principle] of legal certainty and [the principle that penalties must be defined in law (nulla poena sine lege)], be interpreted as permitting (by way of a civil penalty) a national measure authorising the confiscation of the entire proceeds of a transaction, such as the one referred to in Article 2(2)(a) of [Decision 2014/512], in the event that an act categorised by domestic law as a summary offence is found to have been committed?

(2) Is Article 5 of [Decision 2014/512] to be interpreted as allowing Member States to adopt national measures providing for the automatic confiscation of any proceeds resulting from a breach of the obligation to notify a transaction falling within the scope of Article 2(2)(a) of [Decision 2014/512]?

(3) Is the prohibition laid down in Article 2(2)(a) of [Decision 2014/512] applicable where goods constituting military equipment, which were the subject of brokering transactions, were never physically imported into the territory of the Member State?[10]

3.   Opinion of Advocate General Ćapeta

It is worth noting that all post-Lisbon cases concerning the Union courts’ jurisdiction in the CFSP so far have been dealt with by different Advocate Generals (AGs). The views expressed in the respective Opinions of the AGs differ greatly, ranging from rather conservative (AG Kokott in Opinion 2/13[11] and AG Wahl in H v Council and Commission[12]) to fairly nuanced (AG Jääskinen in Elitaliana v Eulex Kosovo,[13] AG Wathelet in Rosneft,[14] and AG Bobek in CSUE v KF[15]) and slightly progressive (AG Hogan in Bank Refah Kargaran[16]).

AG Ćapeta, who was not only assigned as the AG in Neves 77 Solutions, but also in KS and KD v Council and Others has advanced the so far most progressive (i.e. least restrictive) interpretation of the carve-out provision enshrined in Article 24 TEU and Article 275(1) TFEU. In her Opinion in KS and KD v Council and Others delivered on the same day as in Neves 77 Solutions, AG Ćapetasubmitted that the carve-out provision could not be interpreted as excluding review of CFSP measures in light of fundamental rights.[17] In her view, only the review of CFSP measures in light of the CFSP provisions enshrined in primary law and the interpretation of CFSP provisions under both primary and secondary law would be excluded from the Union courts’ jurisdiction.[18] AG Ćapeta’s Opinion in Neves 77 Solutions must be read in light of these considerations. 

Before taking on the questions referred, the AG thus considered it necessary to reformulate the first and second question as to relate to the interpretation of EU fundamental rights instead of an interpretation of Decision 2014/512. Accordingly, AG Ćapeta reframed the first and second question as asking in essence whether the Romanian confiscation measures were unlawful in light of the principle of legal certainty, the principle of nulla poena sine lege, and the right to property.[19] Having found the first two principles not to be relevant to the case, the AG then examined whether Article 17(1) CFR precluded the Romanian confiscation measures in question. She concluded that the confiscation of the proceeds, albeit a limitation of Article 17(1) CFR, was proportionate and thus in accordance with the right to property as enshrined in the Charter.[20]

Turning to the third question, the AG opined that unlike the first and second question, the third question did not concern an interpretation of the Charter, but was about giving an interpretation to a general provision of a CFSP Decision.[21] According to the AG, the Court lacks jurisdiction to do so, as the drafters of the Treaties were not willing to give the Court a role in influencing policy decisions in this area.[22] While the Court could assess the lawfulness of a CFSP measure as understood by the Council, it was precluded from determining that one of several possible meanings was the correct one. Thus, the notion of ‘brokering services’ pursuant to Article 2(2)(a) of Decision 2014/512 could be given different meanings by different domestic courts.[23]

4.   Judgment of the Court 

In its judgment, the Court recalled that although it did as a general matter not have jurisdiction in the area of the CFSP, pursuant to Article 24(1) TEU and Article 275(2) TFEU there were two exceptions to this limitation: monitoring compliance with Article 40 TEU and reviewing the lawfulness of restrictive measures against natural or legal persons.[24] However, given that the latter would only apply in instances of restrictive measures of an individual nature and because Article 2(2)(a) of Decision 2014/512 constituted a general measure, the Court could not rely on the second counter-exception to establish its jurisdiction in this case.[25] The Court further recalled that a regulation adopted under Article 215 TFEU giving effect to a CFSP Decision on restrictive measures would come under the full jurisdiction of the Court. In the present case, however, the Council had not implemented the prohibition to provide brokering services laid down in 
Article 2(2)(a) of Decision 2014/512 by way of a regulation.[26]

Notwithstanding this fact, the Court stressed that it had full jurisdiction to monitor compliance with Article 40 TEU as the first counter-exception does not require any further conditions to be fulfilled.[27] Highlighting the difference in wording between Article 215(1) and (2) TFEU, which requires the adoption of a regulation in cases of general measures, while being optional in the case of individual measures, the Court considered that the Council’s powers under paragraph 1 were ‘circumscribed’. As the Council had failed to implement Article 2(2)(a) of Decision 2014/512 by way of a regulation although it was obligated to do so, the Court considered that in light of Article 40 TEU it would enjoy jurisdiction to give a preliminary ruling for every provision of a CFSP measure that should have been given effect to by a regulation adopted under Article 215(1) TFEU.[28]

Given that Article 2(2)(a) of Decision 2014/512 constituted a measure aimed at ‘the interruption or reduction, in part or completely, of economic and financial relations with a third country’ and that the goods referred to therein fell within the competences of the TFEU, the Court affirmed its jurisdiction to give an answer to the questions referred by the Romanian court.[29]

Answering the questions referred, starting with the last one, the Court confirmed that the prohibition on providing brokering services pursuant to Article 2(2)(a) of Decision 2014/512 also applied to brokering transactions in which the military equipment concerned did not enter the territory of a Member State.[30] Answering the first and second question, the Court found the confiscation of the proceeds not to have violated the right to property pursuant to Article 17(1) CFR nor the principle of legal certainty and the principle that penalties must be defined by law, provided that the provisions under Romanian law were in force at the material time in the main proceedings and that they met the requirements of clarity and foreseeability.[31]

5.   Analysis of selected issues concerning the Court’s jurisdiction 

5.1.  In the shadow: fundamental rights protection as a question of competence

Despite their highly different factual and legal backgrounds, the case of Neves 77 Solutions is closely connected to KS and KD v Council and Others.[32] Both cases were assigned to the Court’s Grand Chamber sitting in the same composition (albeit with different judge rapporteurs) and to the same AG, the oral hearings took place on the same day, and both the AG’s Opinions as well as the Court’s judgments in both cases were delivered on the same day respectively. It is of course not hard to see why the Court chose to deal with these two cases together. With both cases involving questions of EU fundamental rights law and raising novel questions in relation to the jurisdiction in the area of the CFSP, it was virtually impossible to ignore the elephant in the (Court-)room, namely the implications of the two judgments for the EU’s accession to the ECHR.[33]

What is striking, however, is that while the judgment in KS and KD v Council and Others has attracted much scholarly commentary,[34] academic interest in the Neves 77 Solutions judgment at least so far seems to be largely non-existent.[35] Leaving aside the factual aspect that sympathy for a company that has lost its proceeds as a result of having been involved in a forbidden transaction under EU sanctions law is destined to be less pronounced than for individuals who have lost their loved ones to war crimes, I would assume that from a legal perspective the main reason for this relative lack of scholarly interest is that unlike KS and KD v Council and Others, Neves 77 Solutions only dealt with the interpretation of the claw-back provision, which many scholars presumably regard only as a side issue in the greater scheme of things relating to the Court’s jurisdiction in CFSP matters.[36]

Without a doubt, KS and KD v Council and Others is a significant case. The judgment answers some questions and raises a couple of new ones. It does not, however, give a definitive answer to the question whether EU accession to the ECHR will be possible without amending the Treaties, neither in the positive nor in the negative.[37] Neither does the judgment in Neves 77 Solutions, and yet the outcome of that case – particularly viewed against the Opinion of the AG – is of tremendous importance for keeping the prospects of accession to the ECHR alive.

As summarized above, in her Opinion AG Ćapeta suggested a rather unorthodox solution to the problem of jurisdiction. In her view, the Court would be competent to give a preliminary ruling only in relation to the rights and freedoms enshrined in the Charter of Fundamental Rights, whereas the interpretation of Decision 2014/512 would be off-limits for the Court. What the AG proposed was in essence a decoupling of the interpretation of the Charter rights in question (Questions 1 and 2) from the interpretation of the term ‘prohibition on providing brokering services’ (Question 3) and only answering the former, while leaving the latter to the domestic courts. AG Ćapeta’s Opinion – just like her Opinion in KS and KD v Council and Others – appears to be strongly motivated by the intention to pave the way to EU accession to the ECHR.[38] Paradoxically, however, it most likely would have done more harm than good had the Court followed her Opinion. 

The crucial difference between KS and KD v Council and Others and Neves 77 Solutions is that in the former case it is the Union institutions which by implementing a CFSP measure were responsible for an alleged infringement of fundamental rights (direct enforcement of EU law), whereas in the latter case the acts in question resulted from actions on the part of a Member State implementing a CFSP measure (indirect enforcement of EU law). While the applicability of the Charter is beyond dispute in a situation like in KS and KD v Council and Others, Member States are bound by the Charter only ‘when implementing Union law’ pursuant to Article 51(1) of the Charter. Lenaerts and Gutiérrez-Fons have popularized the metaphor of the Charter as the ‘shadow’ of EU law to describe this relationship.[39]

But how do we know whether a Member State is implementing EU law in a given situation? We only know if we interpret the legal act that is supposed to trigger the application of the Charter. And this is precisely the reason why it would have been problematic had the Court followed AG Ćapeta’s proposed solution. In the domestic proceedings before the Romanian courts, which subsequently gave rise to the preliminary ruling, it was called into question whether Neves 77 Solutions’ conduct actually qualified as a violation of the prohibition to provide brokering services enshrined in Article 2(2)(a) of Decision 2014/512. The alleged non-conformity with EU fundamental rights law of the confiscation of the proceeds of that transaction merely constituted an alternative argument. Importantly, however, that alternative argument could only be examined in light of the Charter if Neves 77 Solutions’ conduct was actually covered by the prohibition enshrined in Article 2(2)(a) of Decision 2014/512 as otherwise that conduct and consequently also the sanctioning thereof by the Romanian authorities would not have occurred within the scope of EU law. This is also the reason why the Court considered it ‘appropriate to examine in the first place’[40] the third question referred before the first and second, as the first and second question were premised on the assumption that the third question was to be answered in the affirmative. Had the Court answered the third question in the negative, then the first and second question would have become obsolete for the simple reason that the transaction in question was outside of Article 2(2)(a) of Decision 2014/512, and the sanctioning of that transaction under domestic law at best would have constituted a form of ‘gold plating’ of Decision 2014/512,[41] which, however, does not trigger Article 51(1) of the Charter.[42] The only applicable fundamental rights standard in that hypothetical alternative scenario would then have been national constitutional law and the ECHR. 

Of course, one could argue – in line with what AG Ćapeta suggested – that the drafters of the Treaties apparently were willing to accept that one and the same provision of a CFSP measure could be interpreted in different ways by different courts.  This is certainly a legitimate way of looking at the ambiguous wording of the carve-out provision enshrined in Article 24(1) TEU and Article 275(1) TFEU. The problem, however, is that this Humpty-Dumpty-approach to the interpretation of a CFSP measure is hard to square with Article 6(1) TEU and Article 51(2) of the Charter, both of which clarify that the Charter must not extend beyond the scope of EU law.[43] In other words, Article 51(1) of the Charter requires an objective assessment whether a situation falls within or outside the scope of Union law.  Thus, while one might accept the argument that the drafters of the Treaties might have been willing to forego a centralized and uniform interpretation of CFSP measures, they clearly did not envisage Article 51(1) of the Charter to serve as an on-and-off switch that domestic courts could trigger as they see fit. A decoupling of the interpretation of the Charter from the interpretation of the underlying provision of a CFSP measure would therefore have created serious tensions with the principle of conferral.[44]

Now what does all of this have to do with the prospects of an EU accession to the ECHR? 

Right now, the EU is not (yet) a party to the ECHR. A violation of a Convention right is currently only attributable to the Member States themselves and not to the Union, even in situations in which the Member States are implementing EU law.[45] After an EU accession to the ECHR, however, the question will arise who incurs responsibility for an alleged violation in situation of indirect enforcement of EU law. To that effect, Article 1(b) of Protocol No. 8 requires that the accession instruments must provide for mechanisms that would guarantee that ‘applications [under the ECHR] are correctly addressed to Member States and/or the Union as appropriate.’ In accordance with this requirement, the Draft Accession Agreement of 2013 (DAA 2013) therefore included two important mechanisms to address the issue of responsibility. 

First, there is the co-respondent mechanism, which allows for the EU to join proceedings as a party to the case in applications against one or more Member States before the ECtHR if the compatibility of EU law with the ECHR is called into question, ‘notably where [a violation of the Convention] could have been avoided only by disregarding an obligation under European Union law’.[46]

Second, and closely connected to the co-respondent mechanism, there is the prior-involvement mechanism. The prior-involvement mechanism creates a judicial link between Luxembourg and Strasbourg in proceeding to which the EU is a co-respondent in order to allow the ECJ to assess the compatibility of Union law with the Convention in instances where it did not yet have the opportunity to do so.[47]

Both mechanisms still form part of the revised Draft Accession Agreement of 2023 (DAA 2023), albeit with some modifications that were made in order to take account of the incompatibilities with the autonomy of EU law that the Court identified in Opinion 2/13with regard to the DAA 2013.[48] In Opinion 2/13, the Court inter alia opined that any review with regard to the plausibility of a request to become a co-respondent on the part of the ECtHR as well as the possibility for the ECtHR to depart from the general rule of joint liability for the respondent and the co-respondent would likely affect the division of powers between the EU and its Member States.[49] Moreover, the Court also opined that the prior involvement mechanism must allow for the interpretation and not just a review of the validity of EU secondary law.[50] In brief, the way in which these mechanisms were designed in the DAA 2013 would have undermined the Court’s jurisdictional and interpretative monopoly over EU law. 

Against this backdrop, it is not hard to see why it would have proven rather disastrous for the prospects of EU accession to the ECHR had the Court in Neves 77 Solutions declined its jurisdiction for the interpretation of Decision 2014/512. If a restrictive measures case like Neves 77 Solutions were to come before the Strasbourg court, it would be necessary to examine whether the conditions for the EU to become a co-respondent to the proceedings are fulfilled. Such a decision, as the ECJ noted in Opinion 2/13, ‘necessarily presupposes an assessment of EU law.’[51]  If the conditions are fulfilled, the EU side will then have to examine whether it is necessary to activate the prior-involvement mechanism,  and if that is also the case, then the ECJ will be called upon to rule on the interpretation and/or validity of the provision of EU law that is subject to the proceedings before the ECtHR. Such a ruling allows among other things for the correct interpretation of EU law and as a consequence for a correct attribution of a potential violation of the Convention. If, for example, the ECJ were to find that the EU’s decision to join proceedings against a Member State was based on an erroneous interpretation of EU law, it will become necessary to terminate the co-respondent mechanism in light of this finding. If the ECJ was barred from making such an interpretation in relation to a CFSP Decision imposing restrictive measures, then it would be exclusively for the ECtHR to interpret such a CFSP measure, which obviously would run counter to the autonomy of EU law. 

Some might object that such concerns seem overblown in the context of a case like Neves 77 Solutions. After all, while the interpretation of ‘providing brokering services’ indeed requires an interpretation of Article 2(2)(a) of Decision 2014/512, it was undisputed that the policy of confiscating the proceeds of a prohibited transaction did not result from Decision 2014/512, but from the specific implementing measures Romania undertook under its domestic law to give effect to the prohibitions laid down in that Decision.[52] While I am willing to concede that in Neves 77 Solutions it was easy to separate the normative content of the EU legal act in question from the conduct of the Member State acting as a decentralized enforcer of that legal act, this does not alter anything with regard to the analysis made above. First, not all cases will be that clear-cut when it comes to determining whether a potential violation of a Convention right results directly from the CFSP Decision itself or is the consequence of the way the measure is implemented by a Member State. Second, the wording of Article 3(2) of the DAA 2013/DAA 2023 suggests that the EU would become a co-respondent to the proceedings against a Member State ‘notably where that violation could have been avoided only by disregarding an obligation under European Union law.’[53] Although subject to considerable debate among scholars,[54] the way I read this provision is that the co-respondent mechanism may also apply outside the narrow situation in which a Member State has virtually no discretion whatsoever when implementing an act of EU secondary law. While unfortunately neither the Draft Accession Agreement itself nor the Explanatory Report enlighten us as to the situations in which it may occur that despite room for discretion recourse to the co-respondent mechanism would be conceivable, it is certainly within the realms of possibility that a case like Neves 77 Solutions would raise an issue that could trigger the co-respondent mechanism as well as the prior-involvement mechanism. 

By deciding Neves 77 Solutions the way the Court did, the gaps in the protection of fundamental rights as far as the issue of restrictive measures is concerned have been drastically reduced, if not closed altogether. Restrictive measures of general application are subject to the full jurisdiction of the Union Courts with regard to the regulation adopted on the basis of Article 215(1) TFEU, or – if the Council has failed to adopt such a regulation – based on Article 275(2) TFEU in conjunction with Article 40 TEU with regard to the CFSP Decision itself to the extent that a provision falls within the scope of the TFEU. Restrictive measures of an individual nature are subject to the jurisdiction of the Union Courts where the Council has chosen to implement them by way of a facultative regulation based on Article 215(2) TFEU, or with regard to the CFSP Decision itself in the context of annulment actions, action for damages, and preliminary rulings on the validity of such restrictive measures. 

Nonetheless, some questions remain. 

For one thing, the Union courts’ jurisdiction for restrictive measures of general application only covers those measures that substantively fall within the scope of the TFEU. But what if a measure of general application cannot be transposed by way of a regulation because it contains elements that fall outside of the scope of the TFEU and therefore can only be given effect through measures adopted under the domestic law of the Member States? From a practical point of view, such a scenario is very unlikely to occur. Prior to the judgment in Neves 77 Solutions, the prime example of restrictive measures of general application that were not implemented by a regulation were arms embargos. The mistaken assumption that such measures would fall outside the TFEU, however, was corrected by the Court in Neves 77 Solutions.  It thus seems that as a matter of fact there are no elements left in the context of restrictive measures of general application which would not have to be transposed by way of a regulation.[55]Nonetheless, a hypothetical risk that such a situation may arise persists. 

The other blind spot that is still left is whether the claw-back provision also grants the Court jurisdiction to decide on a preliminary ruling on the interpretation of restrictive measures which are of an individual nature and which are contained in a CFSP Decision itself. I think, however, it is safe to assume that in this case we are dealing with a ‘known unknown’ and that the Court would also have jurisdiction to hear such a case. Although there is no definitive statement from the Court in this regard, AG Wathelet has already addressed this question in his Opinion in Rosneft, arguing that if the Court was competent to decide on the broader question of validity, it must all the more be competent to give a preliminary ruling on the interpretation of a CFSP Decision.[56] By way of a constitutionally conforming interpretation the annulment or invalidation of a CFSP Decision could thus be avoided if that Decision was given a meaning that was in conformity with the rights enshrined in the Charter. AG Wathelet’s argument is persuasive and it is fair to assume that the Court would ultimately follow this line of reasoning.[57]

5.2.  Categorization of restrictive measures: binary constitutional choices meet complex political realities

The Union’s constitutional framework distinguishes between two types of restrictive measures: restrictive measures of general application on the one hand, and restrictive measures of an individual nature on the other hand. That distinction bears legal significance for a number of reasons. First, for restrictive measures of general application, the adoption of a Council regulation is obligatory, whereas in the case of restrictive measures of an individual nature, the adoption of such a regulation is merely optional.[58] Second, the question whether restrictive measures are of a general or an individual nature affects the standard of judicial review. While the Council is given a larger degree of discretion in cases concerning restrictive measures of general application, the level of scrutiny is stricter when it comes to restrictive measures targeting natural or legal persons.[59] Third, and most relevantly in the context of Neves 77 Solutions, the qualification of restrictive measures as either being of a general or an individual character affects the jurisdiction of the Union courts as far as the application of the carve-out provision pursuant to Article 275(2) TFEU is concerned. While the review of compliance with Article 40 TEU is not subject to any specific conditions, the review of CFSP decisions on restrictive measures is limited to those of an individual nature. Restrictive measures of a general nature are, by contrast, excluded from the jurisdiction of the Union courts.[60]

As the Court found Article 2(2)(a) of Decision 2014/512 to be of general application, its jurisdiction in any event (i.e. regardless of whether preliminary rulings on interpretation are permissible under that provision) could not be based on the second counter-exception to the carve-out provision of Article 275(1) TFEU. Nonetheless, I still consider it worthwhile to reflect on the question whether the conclusion that the prohibition on providing brokering services as laid down in Article 2(2)(a) of Decision 2014/512 constituted restrictive measures of general application was indeed inevitable.  

In its judgment, the Court did not spend many words on the issue. It merely stated in a single paragraph that ‘the scope of [Article 2(2)(a) of Decision 2014/512] is defined by reference to objective criteria, not to identified natural or legal persons’ and that therefore the prohibition on providing brokering services ‘is, in any event, a measure of general scope that does not come within the scope of a restrictive measure under the second paragraph of Article 275 TFEU’.[61] It is thus the definition by reference to ‘objective criteria’ which renders a provision of a CFSP decision on restrictive measures to be of general application and as a consequence outside the jurisdiction of the Court. 

The way in which drafters of the Treaties framed the difference between restrictive measures of general application and those of an individual nature is that of a binary logic, a choice of either/or: On the one hand, there are – in the words of Article 215(1) TFEU – CFSP decisions which provide ‘for the interruption or reduction, in part or completely, of economic and financial relations with one or more third countries’. One other hand, there are CFSP decisions imposing restrictive measures ‘against natural or legal persons and groups or non-State entities’ (Article 215(2) TFEU). In practice, however, matters are more complicated than that. The way in which restrictive measures are designed often times fails to reflect the clear-cut differentiation the drafters apparently had in mind.[62]

Decision 2014/512 illustrates this complexity quite well. Decision 2014/512 is not as such aimed at the interruption or reduction of economic and financial relations with Russia, nor does it exclusively target natural or legal persons. This ambiguity is reflected in the legal basis of Regulation 833/2014[63] (giving effect to Decision 2014/512 within the framework of the TFEU), which does not clearly attribute Regulation 833/2014 to either paragraph 1 or 2 of Article 215 TFEU, but instead refers generically to Article 215 TFEU without further distinction.[64]

Even if one looks at each provision of Decision 2014/512 independently, it appears challenging to determine whether a certain measure is of a general or of an individual character. The reason for this lack of clarity is that not only do CFSP Decisions as a whole escape a clear categorization in line with the structure of Article 215 TFEU, but also the measures themselves are not drafted in a way that strictly follows the predefined constitutional dichotomy of embargo measures versus measures against natural or legal persons. Various measures imposed by Decision 2014/512 are of a ‘hybrid’ nature, combining sector-specific restrictive measures relating to goods or services with a particular personal(-ized) element.[65]

The assessment whether a provision is of individual or of general application must therefore always be conducted on a case-by-case basis. But based on what criteria? 

Based on an analysis of the case law until 2018, Heliskoski concluded that ‘the notion of restrictive measures against natural or legal persons within the meaning of the second paragraph of Article 275 TFEU is to be understood as covering measures targeting identified natural or legal persons, in practice, entities individually named in the provisions of the act concerned.’[66] A contrario it would follow that all other restrictive measures are of general application. This conception, which in the ECJ’s case law finds particular support in the Rosneft judgment,[67] certainly explains why the Court in Neves 77 Solutions found Article 2(2)(a) of Decision 2014/512 not to be of an individual character. 

The case law of the General Court (EGC) on the differentiation between restrictive measures of an individual and of a general nature is more extensive than the ECJ’s case law on that matter. And while the EGC’s case law falls mostly in line with the ECJ’s approach adopted in Rosneft,[68] there are some notable and also some rather confusing elements in the former’s case law. 

For one thing, in a number of cases the EGC has ruled that a provision was not of an individual nature despite the fact that it referred to natural or legal persons listed in an annex to a particular CFSP Decision.[69] The reason for this categorization as measures of general application was that even though certain persons were identifiable through their listing in the annex, the listing itself related to ‘a category of persons envisaged in a general and abstract manner’[70]. Moreover, in two relatively recent cases, the EGC not only made the question of general-versus-individual dependent on the question whether a particular natural or legal person could be identified, but differentiated – within one and the same provision – based on the identity of the applicant in the respective context.[71]

I would respectfully submit that the EGC’s approach is not particularly convincing and is making matters unnecessarily complicated. In particular, the approach taken in the latter two cases mentioned above seems to be at odds with the wording of Article 275(2) TFEU, which only speaks of ‘decisions providing for restrictive measures against natural or legal person’. The fact that Article 275(2) TFEU makes no reference to the identity of an applicant indicates that the provision does not entail a subjective element, but that the notion of ‘natural or legal persons’ is to be understood in strictly objective terms. That Article 275(2) TFEU requires that proceedings be ‘brought in accordance with the conditions laid down in the fourth paragraph of Article 263 of this Treaty’ only alludes to the fact that the second counter-exception enshrined in Article 275(2) TFEU serves to protect the rights of individuals. Moreover, in Rosneft, the ECJ rejected the argument submitted by the Commission that the Court would only have jurisdiction to give a preliminary ruling on the validity of a CFSP Decision imposing restrictive measures if the applicant in the main proceedings fulfilled the conditions laid down in Article 263(4) TFEU.[72] If, however, Article 275(2) TFEU were to be interpreted so narrow as to mean that the Court would only have jurisdiction to hear a case where the applicants themselves are identified in the text of a provision or an annex, then this would imply the introduction of the Plaumann test through the backdoor of jurisdiction. It appears that the approach adopted by the EGC leads to a problematic conflation of questions relating to the jurisdiction and questions dealing with the admissibility of an action. While questions of direct and individual concern play a role in the context of admissibility for actions for annulment, they should be irrelevant for the issue jurisdiction, irrespective of the type of proceedings. 

Yet, even if one accepts the argument that this overly narrow subjective differentiation between measures of individual and general application is incorrect, we should still be critical and ask whether the ‘conventional’ approach endorsed by the ECJ, namely distinguishing between measures of general and individual application based on the identification of a particular natural or legal person, might not be too restrictive either. If one considers the wording of Article 275(2) TFEU, it is not self-explanatory at all that ‘restrictive measures against natural or legal persons’ necessarily would have to be understood in the sense that an individual natural or legal person is identifiable. Instead, one could also read ‘restrictive measures against natural or legal persons’ as restrictive measures being intended to have an adverse legal effect on natural or legal persons, which in the context of Neves 77 Solutionsarguably would have brought Article 2(2)(a) of Decision 2014/512 within the meaning of a restrictive measure of an individual application. 

One may of course ask why these considerations should matter that much. After all, from the perspective of effective judicial protection it should be irrelevant whether we are dealing with a measure of general application, which comes within the jurisdiction of the Union courts by way of a regulation adopted under Article 215(1) TFEU (or, absent of which, based on Article 40 TEU) or a measure of an individual nature, for which the second counter-exception permits access to the Union courts under certain procedural conditions. However, restrictive measures of general application can only be adopted by way of a regulation and thus will only come within the jurisdiction of the Union courts under the condition that they substantively fall within the scope of the TFEU.[73] As already elaborated above, although highly unlikely it cannot be categorically ruled out that some restrictive measures of general application contained in a CFSP Decision fall outside the scope the TFEU and thus can only be implemented by the Member States. 

I want to make clear that I am not suggesting that the conclusion at which the Court arrived in Neves 77 Solutions by categorizing Article 2(2)(a) of Decision 2014/512 as a measure of general application is wrong. Albeit accompanied by relatively little additional reasoning, that conclusion is rooted in precedent from previous case law. But if we take the promise of effective judicial protection against restrictive measures seriously, then perhaps it is warranted to adopt a slightly broader understanding of what may constitute ‘restrictive measures against natural or legal persons’ to prevent any risk of gaps in the system of judicial protection from arising.[74] In any event,  more clarity (i.e. a more refined methodology) when it comes to the criteria for differentiating between measures of general application and measures of an individual nature would be more than welcome, particularly in cases where that distinction may have a decisive impact on the question of jurisdiction.   

5.3.  The comeback of Article 40 TEU: a missed opportunity? 

Article 40 TEU, referred to as the ‘mutual non-affection clause’,[75] requires a strict separation of Union competences exercised under the TFEU, and the implementation of the CFSP. One could be inclined to assume that Article 40 TEU should play a central role in the EU’s post-Lisbon constitutional edifice. However, the actual track-record of the provision in the Court’s case law looks rather bleak. 

In a number of interinstitutional litigations involving questions relating to the correct choice of legal bases and the demarcation of the CFSP and non-CFSP policies, the Court either chose to ignore Article 40 TEU altogether,[76] or, where it did refer to Article 40 TEU, only did so in a way which ‘gave the impression of being mere lip service’.[77]

In Rosneft, the ECJ was confronted with the question whether the EU position on restrictive measures laid down in Decision 512/2014 was defined ‘in excessive detail’ and thus encroached ‘on the joint power of proposal of the High Representative […] and the Commission.’[78] The Court dismissed this argument, reasoning that although a CFSP Decision adopted pursuant to Article 29 TEU and a regulation adopted on the basis of Article 215 TFEU had different functions, a high level of detailedness of the CFSP Decision could not be regarded as an intrusion into to the non-CFSP procedure under Article 215 TFEU ‘when the measures relate to a field where there is a degree of technicality’[79]. The Court also rejected the argument that the measures laid down in Decision 512/2014 would in fact constitute ‘legislative acts’ and therefore infringe upon Article 40 TEU.[80]

In addition to the litigation brought before the ECJ, Article 40 TEU was also invoked in a number cases before the General Court, however, to no avail.[81]

Remarkably, Neves 77 Solutions is thus the first post-Lisbon case in which the mutual non-affection clause actually proved decisive for the outcome of a case. And yet, the way in which the Court relied on Article 40 TEU – based on a purely formalistic instead of a substantive line of reasoning – is rather disappointing and in my view a missed opportunity for clarifying the boundaries between the CFSP Decision pursuant to Article 29 TEU, and the regulation adopted on the basis of Article 215 TFEU. 

To illustrate what I mean by this, it is helpful to reconstruct the reasoning which the ECJ employed to justify its jurisdiction to interpret Article 2(2)(a) of Decision 2014/512. The Court’s reasoning essentially rests on three premises: (1) that unlike in the second alternative of the claw-back provision, there are no specific procedural conditions that need to be fulfilled when reviewing compliance with Article 40 TEU,[82] (2) that a regulation under Article 215(1) TFEU, whose adoption by the Council is obligatory, is subject to the full jurisdiction of the Court,[83] and (3) that the Court’s jurisdiction to conduct judicial review cannot be negatively affected by the classification, nature, or form of a legal act.[84]  

Based on these premises, the ECJ developed the following rationalization to justify the interpretation of a restrictive measure of a general nature enshrined in a CFSP decision: The Court is generally barred from giving an interpretation to Article 2(2)(a) of Decision 2014/512 because the provision does not constitute a restrictive measure against natural or legal persons pursuant to Article 275(2) TFEU. By contrast, the Court has full jurisdiction to interpret a regulation adopted on the basis of Article 215(1) TFEU which seeks to give effect to these measures. Because the Council failed to implement the prohibition to provide brokering services enshrined in Article 2(2)(a) of Decision 2014/512 by way of a regulation adopted on the basis of Article 215(1) TFEU even though it would have been required to do so, the Court must treat Article 2(2)(a) of Decision 2014/512 as if it was a regulation adopted under Article 215(1) TFEU for which it enjoys full jurisdiction and not as a provision adopted under the CFSP. 

The first and the second premise are fairly uncontroversial. Both the lack of procedural conditions for exercising jurisdiction to review compliance with Article 40 TEU as well as the Court’s full jurisdiction to review regulations adopted on the basis of Article 215 TFEU have been confirmed by the Court in Rosneft.[85] By contrast, the premise that ‘the power of judicial review […] is not limited by the classification, nature or form’[86] of an act intended to produce legal effects on third parties requires a more in-depth analysis. Although that premise is also rooted in precedent – the Court cites ERTA[87] and the Seat of the European Labour Authority[88] case in this regard –, I would carefully suggest that reference to these judgments is somewhat misplaced in this particular context. Cases like ERTA and Seat of the European Labour Authority were concerned with situations in which a decision taken by the Member States should in fact have been taken by the Community/Union institutions. This is clearly different from the situation which gave rise to the judgment in Neves 77 Solutions, in which in a two-step procedure on the EU level the final step was simply not completed – intentionally, as we know from the oral hearing.[89] The real issue here is therefore not so much the correct authorship of Decision 2014/512, which is undisputed, but rather the question about the relationship between a CFSP Decision pursuant to Article 29 TEU and a regulation adopted on the basis of Article 215(1) TFEU and the consequences of a failure to adopt a regulation pursuant to Article 215(1) TFEU. 

When considering potential consequences of such failure, the first and most obvious one which comes to my mind is that direct effect would be precluded. While a regulation is – in the words of Article 288 TFEU – ‘binding in its entirety and directly applicable in all Member States’, a CFSP Decision imposing restrictive measures not directed against natural or legal persons must be implemented by way of a regulation pursuant to Article 215(1) TFEU. As the doctrine of direct effect presupposes that a norm must be unconditional as well as sufficiently precise to be relied upon before national courts, failure to transpose restrictive measures of a general nature by way of a regulation would in principle stand in the way of such unconditionality.[90]

While in Neves 77 Solutions, the question of direct effect ultimately was not relevant because Romania undertook implementing measures under its domestic law to give effect to Decision 512/2014, it would be misguided to presume that this would make reflections about the relationship between the CFSP Decision based on Article 29 TEU and the regulation adopted under Article 215(1) TFEU obsolete. 

One way of looking at the relationship would be to regard it as a two-step composite procedure consisting of a CFSP and a non-CFSP part. The adoption of a regulation under Article 215(1) TFEU, ‘which serves as a bridge between the objectives of the [TEU] in matters of the CFSP and the actions of the Union involving economic measures falling within the scope of the [TFEU]’[91] would thus constitute the second and final step in the procedure for the adoption of restrictive measures of general application at the EU level. As an ‘umbrella competence’ comparable e.g. to Article 217 TFEU,[92] the regulation would cover every aspect of the CFSP Decision which falls within a competence conferred upon the Union by the TFEU. Importantly, under this conception both steps would constitute essential parts of the law-making procedure at the Union level, while certain implementing measures would still be for the Member States to take. Such a conceptualization, however, raises a number of questions when applied to the context of Neves 77 Solutions. 

In particular, if the procedure for the adoption of restrictive measures of a general nature is obligatorily to be completed on the Union level, on what basis did Romania transpose the prohibition to provide brokering services into its domestic law? This question is especially relevant given that the Court held that ‘trade in […] weapons, equipment and services [referred to in Article 2(2)(a) of Decision 2014/512] falls within the remit of the European Union under Articles 114 and 207 TFEU.’^[93] If, however, the transaction in question at least in part falls within the common commercial policy, how could Romania then adopt legally binding acts in this field without infringing the EU’s exclusive competence pursuant to Article 3(1)(e) TFEU?[94] If we consider the adoption of a regulation under Article 215(1) TFEU as a form of law-making, then the only possible way in which Romania could do so would be as a trustee of the Union interest.[95] However, it is highly dubious whether in the post-Lisbon Treaty framework such trusteeship could be assumed on a unilateral basis given that Article 2(1) TFEU requires that the Member States must be ‘empowered by the Union’ if they want to adopt legally binding acts in a field of exclusive Union competences.[96] But even if one were to accept that under exceptional circumstances this was still a possibility for the Member States also in the absence of an express authorization, it is obvious that in Neves 77 Solutions the necessary conditions for such unilateral conduct were not fulfilled. [97]  First, the Council’s inactivity was not caused by an inability to act but was a deliberate choice. Second, Article 215(1) TFEU does not provide for a fixed time limit within which the Council would have to adopt its regulation. Third, there are no indications that Romania would have informed the Commission of its intention to adopt legally binding acts in the field of the common commercial policy. 

Another – different – way of looking at the relationship, however, would be to treat the adoption of the regulation pursuant to Article 215(1) TFEU itself as an implementing measure. [98] While it is as a general rule for the Member States to implement binding acts of EU law in accordance with Article 291(1) TFEU, Article 291(2) TFEU recognizes that ‘[w]here uniform conditions for implementing legally binding Union acts are needed’, such implementing measures could be adopted by the institutions. Thus, one could argue that Article 215 TFEU serves as a lex specialis to Article 291(2) and (3) TFEU in the context of restrictive measures of general application. The way in which the Court characterizes Article 215 TFEU seems to lend support to this conceptualization, as it has described the function of Article 215 TFEU as ‘to give effect to restrictive measures where such measures fall within the scope of the [TFEU], and, in particular, to ensure their uniform application in all the Member States.’[99] But then again, one could ask how a Member State such as Romania could lawfully take implementing measures with regard to restrictive measures where the Council has failed to adopt a regulation which is supposed ‘to ensure their uniform application’. While Article 2(1) TFEU expressly allows for the adoption of legally binding acts by the Member States ‘for the implementation of Union acts’ also in an area of exclusive Union competence, the fact that the constitutional framework mandates the adoption of a regulation pursuant to Article 215(1) TFEU to guarantee uniform conditions for the application of restrictive measures of a general nature puts a huge question mark on the permissibility of any individual implementing measures taken by the Member States in the absence of such a Council regulation precisely because such measures would fail to provide the uniformity that a regulation is supposed to ensure.    

Against this backdrop, I argue that the approach taken by the Court with regard to the relationship between the CFSP Decision pursuant to Article 29 TEU and the Council regulation to be adopted under Article 215(1) TFEU as well as the role ascribed to Article 40 TEU is not particularly convincing. What would have been more compelling is in my view a clear substantive demarcation between CFSP and non-CFSP policies instead of a purely formalistic argument to justify the ECJ’s jurisdiction. 

Restrictive measures of an individual nature (i.e. listings of natural or legal persons) are by their very nature of such specificity to be inclined to produce legal effects vis-à-vis the listed individuals, which may also help explain why the adoption of a regulation under Article 215(2) TFEU is a purely facultative measure and why the second counter-exception under Article 275(2) TFEU grants the Union courts jurisdiction only for this particular type of restrictive measures.[100] By contrast, it can be argued that the rationale behind the obligatory adoption of a regulation on the basis of Article 215(1) TFEU for restrictive measures of general application is that the general definition of an ‘approach of the Union to a particular matter of a geographical or thematic nature’ does not necessarily imply that the specifics under which such restrictive measures may take effect vis-à-vis natural or legal persons are laid down in sufficient detailedness and therefore will most likely require further steps on the Union level by way of a regulation, which then as a result also come under the full jurisdiction of the Union courts. Although the Court in Rosneft has accepted a high degree of detailedness of the CFSP Decision adopted on the basis of Article 29 TEU itself, it has also indicated that the CFSP Decision and the Council regulation fulfil two different functions,[101] and in Bank Refah Kargaran the Court has confirmed that CFSP Decisions and regulations under Article 215 TFEU ‘may not be substantively identical’[102]. The Court thus could have built on these arguments and set out in greater detail what the content of the CFSP Decision imposing restrictive measures of general application is supposed to be and what is to be laid down in the regulation to be adopted on the basis of Article 215(1) TFEU instead of rubberstamping the currently prevailing ‘copy-and-paste’ approach in the field of restrictive measures.[103] After all, the ‘bridge metaphor’ that the Court has used to describe the role and function of Article 215 TFEU only makes sense if the two ends that need to be connected are actually located on different sides – not only formally, but also in substance. Failure to respect this functional demarcation should consequently be regarded as a substantive infringement of Article 40 TEU.

6.   Concluding remarks 

Neves 77 Solutions is the type of case which at first sight appears to be rather inconspicuous. The facts of the case seem pretty dull, and the legal questions involved rather technical. Yet, upon closer inspection, Neves 77 Solutions is an incredibly complex and fascinating case, full of legal puzzles in relation to the Court’s jurisdiction. And although in my view certain elements of the judgment exhibit some flaws, the overall line of reasoning is to be welcomed. And in view of the broader implications of the case, we might perhaps at some point in the future look back at the judgment as the case that saved the EU’s accession to the ECHR.