Key features of the Draft Accession Agreement

Before addressing the Opinion and its consequences in more detail, it is necessary to briefly introduce key aspects of the Draft Agreement. The Draft Agreement takes account of the EU’s executive federalism, i.e. the role of the member states as the entities mainly responsible for implementing EU law, and the fact that after accession, both the EU and its member states would be parties to the European Convention. For an applicant before the European Court of Human Rights this could result in considerable difficulty to decide which entity – EU or member state – was responsible for a violation of the Convention, as it cannot always be easily determined whether the exact legal basis for member state action is a rule of domestic law or of EU law.

In order to avoid these problems the Draft Agreement contains a rule of attribution in Article 1(4) according to which an ‘act, measure or omission of organs of a member State of the European Union […] shall be attributed to that State, even if such act, measure or omission occurs when the State implements the law of the European Union, including decisions taken under the [Treaty on European Union] and under the [Treaty on the Functioning of the European Union]’. Conduct is only attributed to the EU where there was no member state involvement. In order to include the EU or the member states in EU-related proceedings the drafters devised the status of co-respondent.Footnote ⁵ The co-respondent is a party to the case, so that it would be bound by the decision of the Court and enjoys the same procedural position as the respondent.Footnote ⁶ In light of the rule on attribution, most cases would initially be directed against a member state. Where this is so, Article 3(2) of the Draft Agreement provides that:

The Draft Accession Agreement provides further for the involvement of co-respondent(s) if the initial application is directed against the EU where an alleged violation ‘calls into question the compatibility with the [Convention] rights at issue [...] of a provision of the [Treaties]’.Footnote ⁷ The decision to become a co-respondent would be voluntary. A potential co-respondent either accepts an invitation from the European Court of Human Rights or makes a request to be joined to the proceedings as co-respondent. In the latter case the Draft Agreement provides for the Court of Human Rights to assess whether ‘it is plausible that the conditions [for the co-respondent status] are met.’Footnote ⁸

If a violation is found the respondent and co-respondent(s) would normally be jointly responsible.Footnote ⁹ The drafters considered this solution capable of avoiding an infringement of the autonomy of EU law given that the co-respondent mechanism could be seen as preventing the Court of Human Rights from making internally binding pronouncements on the division of competences between the Union and the member states.Footnote ¹⁰ As the following discussion of Opinion 2/13 will show, the Court did not consider this to be the case.

The rule on attribution would result, in most cases, in the need to exhaust the domestic remedies in the member state to which the conduct is attributed.Footnote ¹¹ These proceedings may result in a request for a preliminary reference to the Court of Justice in cases in which the interpretation or validity of EU law is at issue.Footnote ¹² As the parties to the proceedings before a national court cannot force it to request a reference, there is a potential gap in the involvement of the Court of Justice even where a preliminary reference would be mandatory. The Court of Justice considered this to be problematic with regard to the subsidiarity of the European Court of Human Rights’s jurisdiction and demanded that ‘external review by the Convention institutions can be preceded by effective internal review by the courts of the Member States and/or of the Union’.Footnote ¹³ In order to involve the Court of Justice in such cases, Article 3(6) of the Draft Agreement provides:

The exact ramifications of this involvement at EU level were to be laid down in EU-internal rules.

The extension of autonomy in Opinion 2/13

The key theme in Opinion 2/13 is the autonomy of EU law. In the context of external relations, a summary of the Court’s classic position on the autonomy principle can be found in Opinion 1/00:

It follows from this that an international court must not interpret EU law if such interpretation would carry binding effects. This could be seen clearly in Reynolds, where the Court held that the power of a United States District Court to decide whether under EU law the Commission was capable of bringing proceedings against tobacco producers in a United States court did not constitute an infringement of the autonomy of EU law because ‘a decision by a United States court as to the Commission’s power to bring legal proceedings before it is not capable of binding the Community and its institutions to a particular interpretation of the rules of Community law in the exercise of their internal powers. […] such a decision would be binding only in relation to the specific proceedings.’Footnote ¹⁵ This was the received wisdom at the time the Accession Agreement was drafted.Footnote ¹⁶

Moreover, the Court held in Kadi that an international agreement cannot have the effect of prejudicing the constitutional principles of the Treaties.Footnote ¹⁷ It further added in Opinion 1/09 that it was impossible for a Union agreement to replace the jurisdiction of member state courts as ‘ordinary courts’ within the EU legal order with that of an international court.Footnote ¹⁸

A narrow conception of autonomy, such as this, is appropriate as it serves the legitimate purpose of protecting the integrity of the EU law while retaining the EU’s capacity as an external actor. The Draft Agreement is the product of an effort to create a functioning system of judicial review by the European Court of Human Rights of alleged violations of the European Convention by the EU while ensuring that it stayed within the limits defined by the Court of Justice.Footnote ¹⁹ However, in Opinion 2/13, the Court of Justice moved the goalposts and, without expressly admitting it, extended the meaning of the autonomy of EU law by first adopting an extremely narrow view of the meaning of ‘interpretation of EU law’; second, by elevating mutual recognition to the status of a constitutional principle; and third, by largely ignoring the express duty for the Union to accede laid down in Article 6(2) TEU.

In this regard the Opinion is not devoid of a degree of irony, given that the Court of Justice expressly points out the truism that the EU is precluded from being considered a state under international law.Footnote ²⁰ While this would suggest that in contrast to some states, which robustly defend the idea of their own sovereignty, the Union would be more open to integration into an international human rights mechanism, the Court has used this argument to achieve the exact opposite.

The Court of Justice identified seven distinct shortcomings of the Draft Agreement, some of which consist of several subparts, which rendered the Draft Agreement incompatible with the Treaties. They are analysed in detail in the section below titled ‘Is Accession Still Possible?’ together with possible solutions that the member states can adopt in order to address them. It will be demonstrated that some of these shortcomings, such as the exact role of the European Court of Human Rights in the co-respondent and prior involvement procedures as well as the insufficient protection of Article 344 TFEU relate to details in the arrangements already laid down in the Draft Agreement. Others, such as the failure to adequately take account of Article 53 of the EU Charter of Fundamental Rights, of the principle of mutual trust, of the Common Foreign and Security Policy, and of the dangers posed by Protocol No. 16 to the Convention deal with issues outside the Draft Agreement, which, save for the Common Foreign and Security Policy issue, the negotiators do not seem to have discussed. Ultimately, these latter issues are more difficult to overcome.

The co-respondent mechanism: Procedure for the involvement of the co-respondent

The Court of Justice found fault with two aspects of the co-respondent mechanism: the procedure for the involvement of the co-respondent and the allocation of responsibility under Article 3(7) of the Draft Agreement.

The Draft Agreement foresaw two possibilities of involving a co-respondent: either the European Court of Human Rights would invite the co-respondent to join; or the co-respondent would request its involvement. In the latter case the Court of Human Rights ‘shall assess whether, in light of the reasons given by the [contracting party] concerned, it is plausible that the conditions in [Article 3(2) or (3) of the Draft] are met.’Footnote ³³ Hence a decision on the plausibility of whether ‘it appears that such allegation calls into question the compatibility with the [Convention] rights at issue [...] of a provision of European Union law’ would be required. The Court of Justice considered this to be liable to interfere with the division of powers between the EU and its member states given that the Court of Human Rights ‘would be required to assess the rules of EU law governing the division of powers between the EU and the Member States as well as the criteria for the attribution of their acts or omissions.’Footnote ³⁴

This strict reasoning by the Court exemplifies the extension of the autonomy principle. The plausibility review was adopted as a deliberately superficial standard of review designed to avoid a situation in which the European Court of Human Rights would be forced to make a binding assessment on the division of powers between the Union and the member states, which would potentially be contrary to the autonomy of EU law.Footnote ³⁵ Moreover, the Draft Agreement refers to the ‘calling into question’ of a provision of EU law, which would suggest that the Court of Human Rights would carry out its assessment on the basis of the application before it and the reasons given by the prospective co-respondent and would not conduct a final determination on the basis of the EU Treaties.

As far as the way forward is concerned, it is difficult to see how unilateral action, e.g. an interpretative declaration, on the part of the EU and its member states, could achieve the desired outcome given that the wording of the Draft Agreement would suggest the exact opposite and that it would not be binding on the European Court of Human Rights. Thus the most appropriate solution would be to amend the Draft Agreement by removing the power of the Court of Human Rights to review the plausibility of a co-respondent request. This amendment could provide that the prospective co-respondent would have to be granted this status without any review, i.e. purely at its request.

The co-respondent mechanism: allocation of responsibility under Article 3(7) of the Draft Agreement

As already mentioned, the Draft Agreement provides that the respondent and co-respondent would be jointly responsible if the European Court of Human Rights finds a violation of the Convention. Article 3(7) of the Draft, however, stipulates an exception by giving the Court of Human Rights jurisdiction to decide that ‘on the basis of the reasons given by the respondent and the co-respondent, and having sought the views of the applicant […] only one of them be held responsible’. It is clear that the autonomy of EU law stands in the way of a provision that gives the Court of Human Rights jurisdiction to make a final decision on the internal division of responsibility between the EU and its member states as such a determination would have to be made on the basis of the EU Treaties.Footnote ³⁶

A binding declaration on the part of the EU and its member states to never request that the European Court of Human Rights allocate responsibility might provide a solution to this problem. One can, however, harbour doubts as to whether the Court of Justice would accept this as sufficient. Throughout the Opinion it has become obvious that the Court of Justice does not trust the workings of its own legal order, e.g. with regard to Protocol No. 16 or Article 344 TFEU.Footnote ³⁷ It is not likely that it would accept an internally binding measure when the Court of Human Rights would still have jurisdiction to allocate responsibility under Article 3(7) of the Draft Agreement.

Hence an amendment of the Draft would be the safer option. One could simply remove the critical half sentence at the end of Article 3(7) of the Draft and thereby deprive the Court of Human Rights of its right to allocate responsibility in the circumstances set out therein. Halberstam seems to propose a different solution by suggesting that the Human Rights Court should be deprived of the power to second-guess the EU’s view on joint and sole responsibility.Footnote ³⁸ This would presumably mean that the Draft Agreement would be amended in such a way that the respondent and the co-respondent could inform the Court of Human Rights that the responsibility for the violation should be given to only one of them. Although this solution would certainly accord with the requirements of the autonomy of EU law, it would be contrary to the Convention system as a whole. After all, the applicant has brought a case against the respondent, which was then joined by the co-respondent, but could then be deprived of either the respondent or the co-respondent. The whole point of the compulsory jurisdiction of the Court of Human Rights seems to be that a respondent (and a co-respondent after they have agreed to join) cannot escape responsibility without the agreement of the applicant. A workable compromise between these two positions could consist of requiring the applicant to give permission to a request by the respondent and co-respondent that responsibility should be allocated to only one of them. It would have to make clear that the Court of Human Rights does not have jurisdiction to question this request.

The Court and the Advocate General identified the further problem that Article 3(7) of the Draft does not account for situations in which member states have made a reservation. In their view it could thus happen that a member state is held responsible despite a reservation. This they consider at odds with Article 2 of Protocol No. 8, which provides that accession must not affect ‘the situation of Member States in relation to the European Convention.’Footnote ³⁹ In order to be able to propose a solution, it is necessary to briefly outline in which situations the problem of reservations can materialise. Where a member state authority has acted, the alleged violation would be attributed to the member state under Article 1(4) of the Draft Agreement. If the member state had made a reservation in respect of the alleged violation, it cannot be a respondent and its position is not compromised. Hence the issue of reservations only materialises where the member state acts as a co-respondent, i.e. in a situation where the EU is respondent. The member state may decide to join the Union as co-respondent. If the Court were to find a violation of the European Convention, then, according to the Court, the member state would be jointly responsible for the violation despite its reservation. Given that the co-respondent status is voluntary, this is likely to be a very rare problem. Realistically it could only materialise if the proceedings are not limited to violations in respect of which the member state has made its reservationFootnote ⁴⁰ so that it may have an interest in being involved regardless.

Admittedly, the Draft Agreement is silent as to the effect of reservations on the co-respondent. But instead of assuming that the member state would be responsible despite its reservation, it would have been more convincing to hold that where the Draft Agreement is silent, the standard rules of the European Convention would be applied, which means that the reservation would be given effect.Footnote ⁴¹ Moreover, the question remains whether in light of the voluntariness of the mechanism one can truly say that the Draft Agreement itself affects the situation of the member states with regard to the Convention.

One option for resolving this would be an interpretative declaration to the effect that reservations made by the member states of the EU must be respected. Yet again the Court of Justice might not consider this sufficient, as the Court of Human Rights might still find a responsibility as co-respondent unless the Draft Agreement expressly excluded this scenario. Hence the safest solution would be a clarifying sentence in the Draft Agreement stating that a member state cannot be held responsible either as respondent or as co-respondent in so far as it has made a reservation.

The prior involvement mechanism

Perhaps surprisingly, the Court of Justice considered that the very mechanism it had demanded be included was incompatible with the Treaties.Footnote ⁴² Having confirmed that the introduction of a prior involvement procedure was constitutionally required, the Court of Justice held that the solution found in the Draft Agreement was insufficient. First, it considered that it was not for the Court of Human Rights to decide whether a prior involvement should take place, but for the competent EU institution;Footnote ⁴³ and second, it found that the prior involvement procedure was unduly restrictive by allowing the involvement only to take place where the compatibility, viz. validity, of a provision of EU law was at issue and not where mere questions of interpretation of EU law arose.Footnote ⁴⁴

Concerning the first point, the Court of Justice’s key argument is that the Draft Agreement would confer jurisdiction on the Court of Human Rights to interpret the case law of the Court of Justice.Footnote ⁴⁵ This is not convincing. It is important to recall the wording of Article 3(6) of the Draft Agreement concerning the prior involvement: all the Court of Human Rights is asked to do is to give the Court of Justice sufficient time to make an assessment as to whether the provision of EU law concerned is compatible with human rights. This is preceded by the condition that the Court of Justice must not yet have done so. On the basis of the wording of the provision, there is not much to suggest that the Court of Human Rights would carry out an interpretation of the case law of the Court of Justice. It would rather be a superficial glance at the domestic proceedings, and in a case where a reference request had been made, whether the Court of Justice’s answer engaged with the fundamental rights question before the Court of Human Rights. In this sense the Court of Human Rights’ ‘review’ would at most resemble an examination as to whether an applicant has exhausted domestic remedies. Admittedly, the wording of Article 3(6) Draft Agreement could be construed in a broader manner so that the Court of Human Rights would be required to assess whether the Court of Justice has ever considered the compatibility of the provision of EU law at issue, i.e. including in proceedings unconnected to those before the Court of Human Rights. But this would first require an interpretational stretch, as Article 3(6) Draft Agreement expressly refers to ‘proceedings to which the [EU] is a co-respondent’, which suggests that the proceedings in which the assessment must have taken place are the very proceedings pending before the Court of Human Rights. And second, one should bear in mind that the consequence of such an assessment by the Court of Human Rights is purely procedural: it must afford sufficient time to the Court of Justice to make such an assessment. This confirms that the prior involvement procedure is entirely a procedure under EU law, which the Court of Human Rights must allow to happen. The Court of Justice has thus revealed a blind spot for the practicalities of the procedure before the Court of Human Rights and reinforced its broad notion of autonomy including non-binding assessments of EU law by another court.

The argument concerning the second point rests on the premise that the prior involvement is necessary in order to ensure the autonomy of the EU legal order. The Court feared that after accession, the Court of Human Rights may be asked to interpret EU law and would be in a position to choose between different plausible interpretations.Footnote ⁴⁶ This argument is not conclusive, as it seems to misunderstand the way in which such cases would reach the Court of Human Rights. The Court of Human Rights would be seised only after a case had been decided by a domestic court, which may have interpreted EU law if that was necessary for its decision. The alleged violation would consist of the domestic court’s decision, i.e. the interpretation of EU law adopted by it. The Court of Human Rights would accept this interpretation as part of the facts of the case, on which it would base its assessment. Hence, the question is whether there can be a situation in which no reference has been made in a case that would later involve the EU as a co-respondent but which would nonetheless raise a mere question of interpretation without calling into question the compatibility of the provision with the Convention. This is difficult to imagine.

Nonetheless, the Opinion demands a reaction on the part of the EU and its member states. It would first have to be made clear that the decision to involve the Court of Justice is entirely one for the EU to take. Again, a mere interpretative declaration or other unilateral action might not satisfy the Court of Justice. Instead, an amendment of the Draft Agreement with a rule to the effect that the Court of Human Rights ‘must suspend proceedings if the EU as a co-respondent informs it that the Court of Justice would be involved’, might suffice. Moreover, the term ‘compatibility’ would have to be either re-defined as including interpretations or would have to be replaced by a sentence to the effect that in cases where the Court of Justice has not yet interpreted the EU law at issue, a prior involvement could take place. The suggested amendment might resolve this automatically, given that it would be entirely up to the competent EU institution, presumably the EU Commission, to decide whether a prior involvement should take place so that the Court of Human Rights would never be in a situation of having to decide about the admissibility of a prior involvement. In this case an additional clarifying unilateral declaration to the effect that ‘conformity’ is to be understood to include ‘interpretation’ might be sufficient.

The Court of Justice demanded in the subsequent paragraph that the ‘prior involvement procedure should be set up in such as way as to ensure that, in any case pending before the Court of Human Rights, the EU is fully […] informed’ so that the competent institution can decide whether a prior involvement should be initiated. This ignores the fact that the prior involvement question only arises where the EU has already decided to become co-respondent so that the Union would certainly be aware of these proceedings and could initiate the prior involvement where needed. The Court of Justice seems to have misunderstood the procedure before the Court of Human Rights in this regard. It would have been more convincing if the Court had taken on board the Advocate General’s concerns regarding the rights of defence in respect of the co-respondent mechanism.Footnote ⁴⁷ Protocol No. 8 expressly requires that arrangements be made that ‘individual applications are correctly addressed to Member States and/or the Union as appropriate.’ This presupposes that potential co-respondents are informed if applications with an EU law dimension are pending before the Court of Human Rights. Otherwise potential co-respondents may never find out about these cases, as reliance on the Court of Human Rights taking the initiative and spelling out an invitation cannot be considered sufficient. She therefore demanded the Draft Agreement should contain a requirement that a full and systematic information of a potential co-respondent occurs.

In reaction to this, the Draft Agreement would need to be amended with a provision obliging the Court of Human Rights to ‘ensure that in any case pending before it the EU is fully and systematically informed so that the competent EU institution is able to assess whether a prior involvement is necessary’. In light of the Advocate General’s more convincing criticism, it would make sense to extend this duty to inform the EU to all cases in which the EU could potentially be a co-respondent. Of course, this could not be easily achieved in practice given the sheer number of applications against EU member states that reach the Court of Human Rights every day, not all of which are likely to specify that there may be a connection between EU law and the violation. One option would be for the Court of Human Rights to communicate all cases brought against EU member states also to the EU, but this might be considered inappropriate given that the vast number of cases brought before the Court of Human Rights probably have nothing to do with EU law. A better solution would be to oblige the member states to ensure that all potential co-respondent cases are brought to the attention of the EU as soon as they have been communicated. This already flows from the duty of loyal cooperation under Article 4(3) TEU, but could be reiterated in the Draft Agreement itself to satisfy the Court.

The Court of Justice’s exclusive jurisdiction under Article 344 TFEU

The Court of Justice further held that the accession agreement only insufficiently protected its exclusive jurisdiction over inter-party disputes between member states and between the Union and member states. This exclusive jurisdiction flows from the autonomy of the EU legal order and is confirmed by Article 344 TFEU.Footnote ⁴⁸ It provides that ‘Member States undertake not to submit a dispute concerning the interpretation or application of the Treaties to any method of settlement other than those provided for therein’. The Court rightly considered that after accession the Convention, like any other Union agreement, would become an integral part of EU law as far as it came within the scope of EU law so that the Court of Justice would acquire jurisdiction to interpret it.Footnote ⁴⁹ Cases between member states could be brought before it under Article 259 TFEU.Footnote ⁵⁰ Article 55 of the Convention equally provides for exclusive jurisdiction of the Court of Human Rights over inter-party disputes. The drafters of the Draft Agreement realised that in case of accession, there could be a conflict of jurisdiction and therefore added Article 5 of the Draft, which states that proceedings before the Court of Justice ‘shall [not] be understood as constituting [a] means of dispute settlement within the meaning of Article 55 of the Convention’. By thus removing the exclusive jurisdiction of the Court of Human Rights over such cases, the drafters aimed to avoid a conflict between the two exclusive jurisdiction clauses. Member states would be able to bring their disputes over the Convention before the Court of Justice without infringing the (hitherto) exclusive jurisdiction of the Court of Human Rights, but if they chose to bring them before that Court, they would be in violation of Article 344 TFEU.

The Court of Justice did not share this conclusion and confirmed its lack of trust in the EU’s own legal order, which shows again how the Court has ratcheted up the requirements for compliance with the autonomy of EU law. It ignored the Advocate General, who had pointed out that the tools available under EU law to prevent or pursue a violation of Article 344 TFEU, such as infringement proceedings or interim measures, were adequate.Footnote ⁵¹ It held instead that the mere possibility of an inter-party dispute between two member states or between the Union undermines Article 344 TFEU.Footnote ⁵² Hence the Court demanded that the jurisdiction of the Court of Human Rights be expressly excluded in relation to the application of the Convention within the scope of EU law.

The Court of Justice’s view was defended by Halberstam, who pointed out that the Court added that ‘if the EU or Member States did in fact have to bring a dispute between them before the ECtHR, the latter would, pursuant to Article 33 ECHR find itself seised of such a dispute.’Footnote ⁵³ Halberstam contends that the compulsory jurisdiction of the Court of Human Rights in such a case would be a problem and that the Court of Justice could not be asked ‘to approve of an agreement extending the reach of a provision that, under certain factual circumstances mandates a violation of EU law.’Footnote ⁵⁴ This defence of the Court of Justice is unconvincing. It merely mirrors the hyperbolic character of the Court of Justice’s reaction to the accession agreement, which essentially finds fault with the agreement because it opens up an opportunity for the member states to breach the Treaties.

What this part of the Opinion mainly demonstrates is how the Court has raised the requirements for compliance with the autonomy of EU law. In particular its reliance on the Mox Plant case, in which Ireland had been found in breach of Article 344 TFEU, is problematic.Footnote ⁵⁵ As Johansen correctly argues, the agreement in that case equally provided for the jurisdiction of an international court over a mixed agreement.Footnote ⁵⁶ In stark contrast to Opinion 2/13, the Court in Mox deemed it sufficient that the agreement contained the possibility that such a conflict be avoided by giving the international court a means to divest itself of the dispute.Footnote ⁵⁷ This stance again reveals the Court of Justice’s lack of trust in the EU’s own legal order. The consequence of this is that the EU is becoming an even more awkward partner on the international plane. Requiring the protection of the autonomy of EU law in a watertight manner requires an externalisation of internally resolvable issues, which is new and worrying because it makes the EU a difficult partner to deal with.

The solution to this is not as straightforward as it may seem given that one must find a formulation that does not violate the autonomy of the EU legal order. The Opinion shows that an internal solution would not satisfy the Court of Justice. After all, such a solution already exists in the guise of Article 344 TFEU and was deemed insufficient. This suggests that any other exclusion of the jurisdiction of the Court of Human Rights agreed between the Union and the member states only would not be deemed sufficient either.

Hence the most obvious way forward would be an amendment of the Draft Agreement. As will be shown this may necessitate a broader exclusion of disputes as actually demanded by the Court, which insisted on an express exclusion of the Court of Human Rights’ jurisdiction ‘over disputes between Member States or between Member States and the EU in relation to the application of the Convention within the scope ratione materiae of EU law.’Footnote ⁵⁸ If the Draft Accession Agreement were amended to include a clause reflecting this quote, the Court of Human Rights would be required to decide in a hypothetical case between two member states whether the Convention was applicable as an integral part of EU law or not. This, however, would be contrary to the autonomy of EU law and it might thus be necessary to exclude the jurisdiction of the Court of Human Rights over all disputes between member states and between member states and the EU in a blanket manner.Footnote ⁵⁹ This would also result in the exclusion of disputes that have nothing to do with EU law. A possibility of avoiding this might be the introduction of a pre-clearing mechanism similar to the prior involvement procedure, which might allow the Court of Justice to make this assessment before the Court of Human Rights continues its proceedings. Yet such a pre-clearing mechanism would complicate such proceedings considerably and might fall at the same hurdles as the prior involvement mechanism.

Article 53 of the Charter of Fundamental Rights

A further concern addressed by the Court, but not the Advocate General, is the compatibility of the Draft Agreement with Article 53 of the Charter of Fundamental Rights, which provides:

In the Melloni decision, the Court of Justice had decided that this provision could not be interpreted in a way that would undermine the ‘primacy, unity and effectiveness of EU law’.Footnote ⁶⁰ As the Charter binds the member states only when they are acting within the scope of EU lawFootnote ⁶¹ this means that they are only allowed to apply a higher standard of protection than required by the Charter if this is compatible with their EU law obligations, which continue to enjoy primacy over both national constitutional law and over international law. In Opinion 2/13 the Court suggested that there could be a conflict between this reading of Article 53 of the Charter and Article 53 of the Convention, which allows parties to the Convention to lay down higher standards of human rights protection than what is strictly required by it. The Court considered this to potentially empower the member states to circumvent the limitations resulting from the primacy of EU law. It therefore required that there be coordination between the two Articles 53.Footnote ⁶²

This is not convincing and demonstrates again that the Court of Justice has no confidence in the workings of the EU legal order. From the point of view of EU law the legal situation in this respect would not change with accession. Member states would still be bound to comply with the primacy of EU law. The fact that the Convention, and thus its Article 53, would have to be considered an integral part of EU law would not make a difference for two reasons. First, international treaties concluded by the EU must still comply with EU primary law so that Article 53 of the Charter as interpreted by the Court of Justice would prevail in case of conflict.Footnote ⁶³ Second, Article 53 of the Convention merely provides that the Convention cannot be construed as limiting other fundamental rights contained in domestic legal orders. But this does not mean that it places parties to the Convention under an obligation to apply higher standards than what is strictly required by the Convention. As Halberstam has pointed out, Article 53 of the Convention cannot create a power that did not previously exist.Footnote ⁶⁴ There is no potential for conflict, so the problem identified by the Court of Justice does not exist. Article 53 of the Convention already enables member states to apply higher standards than those strictly required by the Convention. Moreover, Article 52(3) of the Charter provides that rights in the Charter corresponding to those in the Convention shall have the same meaning and scope. This includes the case law of the Court of Human Rights.Footnote ⁶⁵ It is therefore inconceivable that the Convention would provide a higher standard than the Charter.

As with Article 344 TFEU, this is an issue for which EU law as it currently stands provides an internally binding solution. An interpretative declaration by the EU to the effect that ‘the primacy, unity and effectiveness of EU law would be unaffected’ by accession might thus not satisfy the Court as it would not guarantee that the Court of Human Rights would respect this. Hence the safest option would be to add a clarifying provision into the Draft Agreement. The drafters would again need to be conscious of not accidentally allowing the Court of Human Rights to interpret EU law, and in particular Article 53 of the Charter. A possible solution could be a clause along these lines: ‘Article 53 of the Convention shall not be interpreted as requiring the EU’s Member States to provide a higher domestic standard of human rights protection than that provided for by the Convention’.

Mutual trust and mutual recognition

The Court raised a more fundamental issue with regard to the principle of mutual trust, on which the principle of mutual recognition, primarily of relevance in the Area of Freedom, Security and Justice, is predicated. Mutual trust is the basis on which a member state may be allowed to presume that fundamental rights have been observed by other member states. This prevents, for instance, a member state that has been requested to execute a European Arrest Warrant from subjecting this request to fundamental rights scrutiny. It can neither review the fundamental rights compliance on the basis of its own constitutional guarantees nor on the basis of the rights guaranteed by EU law.Footnote ⁶⁶ Instead, member states recognise each other’s human rights standards. The Court considered the Draft Agreement to endanger this very concept by opening up the possibility of requiring one ‘Member State to check that another Member State has observed fundamental rights’ as this would upset the underlying balance of the EU and undermine the autonomy of EU law.Footnote ⁶⁷ According to the Court of Justice, the key flaw of the Draft Agreement in this respect was that it treated the EU like a state by giving it a role identical to that of every other contracting party, which disregarded its intrinsic nature.Footnote ⁶⁸

It is necessary to deconstruct the Court’s argument in order to fully appreciate its significance. First, the Court considers mutual trust to be a foundational value of the EU. It locates it in Article 2 TEU, which names the common values on which the Union is founded. Given that all member states share these values, the Court considers that this implies and justifies the existence of mutual trust.Footnote ⁶⁹ This way the Court elevated the principle of mutual trust to the status of a core constitutional concept and equated it with direct effect and primacy, which are concepts long held to characterise the supranational nature of the EU legal order.Footnote ⁷⁰ Although they are not expressly laid down in the Treaties, they operate across the Treaties and are thus truly foundational. By contrast, mutual trust as the basis for mutual recognition is relevant in free movement lawFootnote ⁷¹ and in the Area of Freedom, Security and Justice. In particular in the latter, mutual recognition is mentioned in a number of Treaty articles, but mainly as an instruction to the legislator that it ought to be the basis for legislation in this area. One can thus question the doctrinal soundness of the Court’s argument.

Second, the Court additionally considered that by treating the EU like a state and ignoring its intrinsic nature, the Draft Agreement failed to appreciate that the member states have accepted that the relations between them ‘are governed by EU law to the exclusion […] of any other law’. It thereby tried to seal off the relations between member states against any review by the Court of Human Rights. With the traditional autonomy principle in mind, this is not surprising as far as the right to interpret the rules governing the relations between the member states is concerned. However, it is remarkable that this should exclude a review of the human rights compatibility of this set-up.

The Court’s strict stance suggests deep concerns about the Court of Human Rights’ case law touching on mutual recognition in general, and the Dublin system on asylum in particular. Under the Dublin Regulation, an asylum request is processed by the member state through which the asylum seeker entered the EU.Footnote ⁷² If an application for asylum is made in another member state, the asylum seeker is usually sent back to the state of entry. The Court of Human Rights’ basic doctrinal framework for extradition cases was developed in Soering and Chahal.Footnote ⁷³ An extradition must be stopped where the victim is facing a ‘real risk’ of having his human rights violated in the receiving state. These cases are often concerned with Article 3 of the Convention,Footnote ⁷⁴ but may also concern other violations, such as flagrant denials of justice contrary to Article 6 of the Convention.Footnote ⁷⁵ It is recalled that in M.S.S. the Court of Human Rights held that an EU member state violates Article 3 of the Convention if, on the basis of mutual recognition, it sends an asylum seeker back to a member state where that person would face a real risk of being subjected to treatment contrary to Article 3 in the receiving country.Footnote ⁷⁶ The respondent member state was thus forced to process the asylum application by making use of Article 3(2) of the Dublin Regulation, which constitutes an exception to the general rule that the first member state entered by the asylum seeker is responsible for processing that person’s asylum application. This was accepted by the Court of Justice in N.S. Footnote ⁷⁷ However, in a later case the Court of Justice held that situations such as in M.S.S. where there were ‘systemic deficiencies in the asylum procedure’ were the only ones in which the obligation to use Article 3(2) of the Dublin Regulation existed.Footnote ⁷⁸ This was a clear attempt to confine the fallout of the M.S.S. decision to extreme cases. In Tarakhel v Switzerland, however, the Court of Human Rights made it obvious that it disagreed with this strict reading of M.S.S. and held that even though there had not been a systemic breakdown of the asylum procedure in this case, the Swiss authorities, who wanted to send the applicant back to Italy, were under an obligation to take into account the applicant’s individual situation and, if necessary, to obtain assurances from Italy that the conditions for the applicant would be compliant with Article 3 of the Convention.Footnote ⁷⁹

This example highlights the potential of accession to undermine mutual recognition in EU law. So far the Dublin system seems to be the only area where this has materialised, but there are no guarantees that this would continue to be the case after accession. For instance, currently mutual recognition in Brussels Regulation cases, such as cases concerned with child abduction,Footnote ⁸⁰ escapes the Court of Human Rights’ scrutiny as it benefits from the Bosphorus presumption.Footnote ⁸¹ After accession, however, it could well be the case that the Court of Human Rights would give up the Bosphorus presumption and intervene more strongly. As far as the European Arrest Warrant is concerned, the Court of Human Rights has not yet interfered with the principle of mutual recognition, but applications such as Stapleton v Ireland, which was declared manifestly ill-founded, suggest a general willingness on the part of the Court of Human Rights to subject the execution of an European Arrest Warrant to human rights review.Footnote ⁸² This is best demonstrated by the Court of Human Rights’ remark in Ignaoua that ‘the mutual trust and confidence underpinning measures of police and judicial cooperation among EU member States must be accorded some weight’, which falls remarkably short of according it immunity from review.Footnote ⁸³

In order to find a viable solution to protect the principle of mutual recognition it is worth recalling the Court’s very words:

Hence a solution would have to prevent such cases from being adjudicated in Strasbourg. Reservations would not be a viable possibilityFootnote ⁸⁵ as they would have to be made in respect of many Convention rights and name a large amount of EU legislation, which might easily make them too general to be acceptable. Moreover, they cannot be updated, which would be necessary where such a fast-developing area as the Area of Freedom, Security and Justice is concerned.Footnote ⁸⁶

It would therefore be best to include a provision in the Draft Agreement itself. As the Opinion highlights, even the slightest hint of giving the Court of Human Rights the power to interpret EU law is likely to fall foul of the autonomy principle. A formulation mirroring the above wording of the Opinion might best reflect the Court’s demands and could be phrased like this: ‘A Member State of the EU cannot be requested to check that another MS has observed fundamental rights where EU law imposes an obligation of mutual trust.’ However, this would prompt an investigation by the Court of Human Rights whether an obligation of mutual trust existed or not and is thus not an option. The same would be true for any draft referring to ‘a Member State’s obligations in the Area of Freedom, Security and Justice’ or ‘where Member States act within the scope of EU law’, and so on. Hence one needs to apply a broader brush in order to achieve the desired result.

A safer option might be this: ‘Member States of the EU cannot be held responsible under the Convention for failing to carry out a review of another Member State’s compliance with Convention rights’. Crucially the Court of Human Rights would not be required to interpret EU law. All it would need to do is verify whether the case deals with two member states, which even the Court of Justice is unlikely to consider a violation of the autonomy of EU law. The drawback of this option is, of course, that it would also cover non-EU related cases, e.g. judicial cooperation based on bilateral treaties.

Protocol No 16

The Court further criticised that the Draft Agreement did not make provision to ensure that Protocol No. 16 to the Convention could not interfere with the autonomy of the EU legal order and the effectiveness of the preliminary ruling procedure.Footnote ⁸⁷ Protocol No. 16 allows the highest courts of the parties that have signed up to it to request an advisory opinion from the Court of Human Rights on ‘questions of principle relating to the interpretation or application’ of the Convention.Footnote ⁸⁸ When Opinion 2/13 was handed down on 18 December 2014, not a single member state had ratified the Protocol and only nine had signed it.Footnote ⁸⁹ The Protocol had not entered into force. The Draft Agreement does not foresee the accession of the EU to the Protocol. In fact, it predates the Protocol.

The Court identified two potential issues with Protocol No. 16, neither of which can convince in its entirety. First, the Court feared a circumvention of the preliminary reference procedure if national courts of last instance that were seised with a dispute concerning the compatibility of EU-related member state action with fundamental rights asked the Court of Human Rights for an advisory opinion instead of referring the case to the Court of Justice. Without expressly saying so, the Court seemed to fear that if the Court of Human Rights were to respond that there was a violation of the Convention in such a scenario, the national court would no longer request a reference from the Court of Justice to confirm this and (possibly) declare EU law invalid. Rather surprisingly, the Court failed to appreciate two important facts: the advisory opinion would not be binding on the national court;Footnote ⁹⁰ and under EU law, the national court would undoubtedly remain under a duty to refer the case to the Court of Justice.

The Court did not expressly spell out a potentially more serious concern, which is that after accession, the Convention would become an integral part of EU law as far as member states were acting within its scope.Footnote ⁹¹ Thus from the point of view of EU law, a request for an advisory opinion under Protocol No. 16 could be regarded as a request to interpret EU law, which would potentially fall foul of the autonomy of EU law.Footnote ⁹² As pointed out above, up until Opinion 2/13 the autonomy of the EU legal order had only been held to be infringed where another court was given jurisdiction to make a binding interpretation of EU law, which would not be the case with Protocol No. 16. This has clearly changed with Opinion 2/13.

Second, the Court saw a danger that the advisory opinion procedure could trigger the prior involvement of the Court of Justice and thus lead to a circumvention of the ‘proper’ route for involving it, which is Article 267 TFEU. This concern is, however, further evidence for the misunderstanding of the prior involvement on the part of the Court. Even a superficial reading of Article 3(6) of the Draft Agreement shows that the prior involvement is only supposed to be possible where the EU is a co-respondent. The co-respondent mechanism is designed for contentious cases. This is made clear by the wording of Article 3(2) and (3) of the Draft, which expressly state that the application must be directed against either a member state or the EU. In case of an advisory opinion, there is no respondent. It is a procedure between the national court and the Court of Human Rights, which does not decide the case but gives an opinion on the interpretation or application of a specific provision of the Convention. Moreover, the advisory opinion procedure is only open where fundamental questions are at stake, i.e. questions hitherto not decided by the Court of Human Rights. Even if a safeguard were built into the accession agreement, these questions would be likely to end up in Strasbourg anyway.

Requiring an unspecified safeguard against Protocol No. 16 to be included in the Draft Agreement thus seems to be an overreaction on the part of the Court of Justice. Even if one accepted the Court’s concerns as valid, it could have confined itself to making it clear that it did not think that after accession a member state could be party to Protocol No. 16. Thus member states would have been under an obligation to refrain from ratifying it. As no member state had done so, this would not have been problematic to achieve in practice. This suggests that, in parallel to the situation regarding Article 344 TFEU, the Court’s concerns would not be allayed by a unilateral commitment on the part of the member states not to sign up to Protocol No. 16. The safe option would thus be to amend the Draft Agreement.

One possibility would be to include a provision in the Draft Agreement that member state courts did not have the right to request an advisory opinion in cases in which they are interpreting or applying EU law, i.e. cases in which they would be entitled or indeed required to request a preliminary reference.Footnote ⁹³ However, such a solution might get caught in the autonomy trap, as the Court of Human Rights would then be empowered to review whether in any given request by a member state court this condition is satisfied, which would of course require an interpretation of EU law. Thus the safe solution would be an express and binding undertaking by the EU’s member states not to sign up or to Protocol No. 16 and for those that have already done so not to ratify the Protocol. This should, in theory, assuage the Court’s concerns.

The Common Foreign and Security Policy

The final objection concerned the fact that after accession, the Court of Human Rights would be able to review Common Foreign and Security Policy measures as to their compatibility with the Convention, whereas the Court of Justice’s jurisdiction over the Common Foreign and Security Policy would remain limited. Its jurisdiction concerning Common Foreign and Security Policy measures is generally excluded save for the review of restrictive measures and for policing the boundaries between the Common Foreign and Security Policy and all other Union competences under Article 40 TEU.Footnote ⁹⁴ While most measures that could raise human rights concerns can probably be classified as restrictive measures, such as sanctions against individuals, there may be situations in which the Court’s jurisdiction may be asymmetrical to that of the Court of Human Rights.Footnote ⁹⁵

The Court proceeded in two short steps. Having established that situations were conceivable in which it had no jurisdiction but which would nonetheless be reviewable by the Court of Human Rights,Footnote ⁹⁶ it held that this ‘would effectively entrust the judicial review of those acts, actions or omissions on the part of the EU exclusively to a non-EU body’.Footnote ⁹⁷ This would contradict what it had held in Opinion 1/09.Footnote ⁹⁸

This reasoning is unconvincing for two reasons. The first relates to the type of review carried out by the Court of Human Rights. As demonstrated in the above quote, the Court of Justice equated the judicial review that would be carried out by the Court of Human Rights in case of accession with the review that it would have to carry out if it had jurisdiction over Common Foreign and Security Policy measures. Admittedly, the Court added the caveat that ‘any such review would be limited to compliance with the rights guaranteed by the ECHR’, but did not consider this to be material.Footnote ⁹⁹ This is a relevant difference, however: the Court of Human Rights’ review is not based on EU law, but on the Convention. The fact that the Convention would become an integral part of EU law cannot make a difference in this respect, given that external accountability is the very point of the Convention so that Article 6(2) TEU, which explicitly provides for accession, must be seen as authorising such review to take place.Footnote ¹⁰⁰ As the Advocate General convincingly argued, there is no conflict with the autonomy of the EU legal order here because the unique supranational structure of Union law would not be affected by the Court of Human Rights’ review.Footnote ¹⁰¹ It is hard to disagree with her conclusion that the ‘absence of sufficient arrangements within the EU, by which the autonomy of EU law alone can be protected, can hardly be seen as an argument against recognition of the jurisdiction of the judicial body of an international organisation.’Footnote ¹⁰² Moreover, the Court of Justice conveniently ignores the fact that EU accession would give the Court of Human Rights jurisdiction to review the compatibility of EU primary law with the Convention. Hence the Common Foreign and Security Policy would not be the only area over which the jurisdiction of the Court of Justice and the Court of Human Rights is asymmetrical.

The second reason relates to the neglect of the potential role of national courts in these cases. The Lisbon Treaty places member states under an express duty to ‘provide remedies sufficient to ensure effective legal protection in the fields covered by Union law’.Footnote ¹⁰³ This confirms their position as ‘ordinary courts’ of the European legal order.Footnote ¹⁰⁴ As convincingly argued by Hillion, they can therefore ‘step in if and when the Court of Justice’s jurisdiction is restricted or non-existent’.Footnote ¹⁰⁵ As the Common Foreign and Security Policy is a field covered by Union law, the member state courts can therefore be deemed to have jurisdiction even in the absence of the Court’s jurisdiction.Footnote ¹⁰⁶ The fact that the Court of Justice is not competent to answer requests for a preliminary ruling in such situations therefore does not deprive individuals of remedies in the EU legal order. Thus the Court of Justice’s conclusion that the Draft Agreement would entrust the review of EU acts exclusively to an international court is not convincing. It is not without irony that in order to support this argument, the Court of Justice points to Opinion 1/09, where it had expressly held that the EU’s judicial structure is not only based on the Court of Justice but also on the courts of the member states.Footnote ¹⁰⁷

The main weakness in the Court’s argument stems from its neglect of Article 6(2) TEU. In contrast to other provisions conferring external competence on the EU, this one specifically relates to the Convention, which suggests that the member states, when including it in the Lisbon Treaty, had the basic features of the Convention system in mind and nonetheless wanted the Union to join. Of course, Article 6(2) TEU cannot be read as authorising accession under all circumstances. This is made clear by Protocol No. 8, which demands that the specificities of Union law be provided for. It is, however, suggested that Article 6(2) TEU should nonetheless be understood to have some substantive content in the sense that it equally demands that the basic workings of the Convention system of judicial review are deemed compatible with the Treaties.

The exclusion of the Court of Human Rights’ jurisdiction concerning Common Foreign and Security Policy measures in the Draft Agreement would create considerable drafting difficulties. As with mutual recognition, a violation of the autonomy of the EU legal order must be avoided, i.e. the Court of Human Rights must not be given jurisdiction to interpret EU law. This constraint means that it would be difficult to adopt a draft along the following lines: ‘The European Court of Human Rights has no jurisdiction to review acts of the Union or of the Member States adopted on under the Common Foreign and Security Policy.’ It is difficult to conceive of a formulation that would limit the Court of Human Rights’ jurisdiction as required by the Court of Justice and that would at the same time not empower the Court of Human Rights to interpret the Treaties at least in a superficial fashion.

As an alternative, one could envisage a variation of the prior involvement mechanism. It is obvious that the Court of Human Rights could review cases in which the Court of Justice has been involved. Where a member state is taken to that Court after domestic proceedings have ended and no reference has been made, the Court of Human Rights could be required to first ask the Court of Justice whether the complaint concerns an area over which the Court of Justice would have had jurisdiction. This might reconcile the requirement that the Court of Human Rights not interpret EU law and the requirement that it not be given jurisdiction over matters, which the Court of Justice is currently not competent to adjudicate. However, it would be very difficult to formulate this in practice without creating a monster, given that the vast majority of cases brought against EU member states have no EU law element whatsoever and that the Court of Human Rights must not be given jurisdiction to even superficially look into the EU Treaties.

This shows that with its objections to the Court of Human Rights’ jurisdiction concerning Common Foreign and Security Policy matters, the Court of Justice has managed to create a veritable Catch-22. On the one hand accession cannot go ahead without certain Common Foreign and Security Policy measures being excluded from the jurisdiction of the Court of Human Rights, as otherwise the specific characteristics of EU law would not be respected. On the other hand, any attempt to formulate an exclusion of Common Foreign and Security Policy measures would violate the autonomy of EU law and thus disrespect its specific characteristics.Footnote ¹⁰⁸ The only viable solution would thus be Treaty change, for which there are two options. One could either give the Court of Justice jurisdiction over all Common Foreign and Security Policy measures;Footnote ¹⁰⁹ or the member states could adopt a slimmed-down version of the ‘notwithstanding Protocol’ discussed above and clarify Article 6(2) TEU by stating that the Union shall accede to the Convention notwithstanding the limited jurisdiction of the Court of Justice over the Common Foreign and Security Policy.Footnote ¹¹⁰