On 2 December 2025, I sat in the Grande Salle of the Court of Justice as four associations of European judges argued, through counsel, that the EU’s procedural architecture had left them without a route to challenge an act they believed entrenched systemic rule of law backsliding in Poland. The case, Medel and Others v Council (C-555/24 P), asked a deceptively simple question: can associations representing judges bring an annulment action under Article 263(4) TFEU to challenge a Council decision approving recovery and resilience milestones that they argue conflict with or fall short of the Court of Justice’s own case law on judicial independence?

On 16 April 2026, Advocate General Tamara Ćapeta delivered her Opinion (ECLI:EU:C:2026:304). She proposes that the Court set aside the General Court’s order dismissing the actions as inadmissible, find them admissible, and refer the case back for a decision on the merits. This piece examines what she decided, why it matters doctrinally, and what it reveals about the limits of the EU’s toolkit for addressing democratic backsliding when viewed from the inside.

1. How we got here

To understand why this case exists, it is necessary to ask what the existing tools had already produced. The rule of law saga is considerably more extensive than what is recalled here, but some of its key moments are worth revisiting.

Between 2019 and 2023, the Court of Justice issued a series of judgments in infringement proceedings finding that Poland had violated Article 19(1) TEU through its judicial reform legislation. In Commission v Poland (C-791/19), the Court found that the disciplinary regime applicable to judges, including provisions under which the submission of a preliminary reference could trigger disciplinary proceedings, infringed Article 19(1) TEU, Article 47 of the Charter, and Article 267 TFEU. In Commission v Poland (C-204/21), it further found that the Muzzle Law, which prohibited national courts from reviewing compliance with EU requirements for an independent tribunal and treated such review as a disciplinary offence, infringed the same provisions. Yet Poland’s response remained cosmetic. Rather than abolishing the Disciplinary Chamber, Poland renamed it the Chamber of Professional Liability while leaving its structural features largely intact. The underlying problems were reorganised rather than resolved. The Article 7(1) TEU procedure, initiated by the Commission in 2017, had produced no binding outcome and was formally discontinued on 29 May 2024, when the Commission withdrew its reasoned proposal.

The preliminary reference procedure was used extensively during this period. Polish courts made 39 references in 2022, 48 in 2023, and 47 in 2024 (Opinion, footnote 88). Some of the most significant judgments on judicial independence were built precisely on questions referred by individual Polish judges. But this route was not free. As the Court found in Commission v Poland (C-791/19), disciplinary proceedings were initiated against judges for making preliminary references. Milestone F1G of Poland’s recovery and resilience plan was supposed to address this directly, requiring Poland to ensure that preliminary references could not be used as grounds for disciplinary action. Whether it did so adequately was precisely what the associations disputed.

Into this landscape came the Recovery and Resilience Facility, established by Regulation (EU) 2021/241, which conditions disbursement on the achievement of milestones proposed by Member States and approved by the Council. On 17 June 2022, the Council adopted an implementing decision approving Poland’s plan, including three judicial independence milestones designated as super milestones, meaning preconditions for any disbursement. Milestone F1G required reforms to strengthen the independence and impartiality of courts. Milestones F2G and F3G required that judges sanctioned by the Disciplinary Chamber have access to review proceedings before a court meeting the requirements of Article 19(1) TEU, and that those proceedings be concluded within defined timeframes.

The associations brought two distinct types of claim, a distinction the AG identifies as analytically decisive (Opinion, paragraphs 67 to 70). On milestones F2G and F3G, the claim was one of conflict: the associations argued that the Court’s case law on the Disciplinary Chamber, including A.K. and Others (C-585/18) and W.Ż. (C-487/19), had the effect of rendering Disciplinary Chamber decisions non-existent, requiring immediate reinstatement without further procedure. Requiring review proceedings of any kind would worsen, not remedy, the situation of affected judges. On milestone F1G, the claim was one of insufficiency: the milestone failed to go far enough to restore effective judicial protection, in breach of Articles 20(5)(e) and 22 of the RRF Regulation, which require Member States to maintain effective internal control systems protecting EU financial interests.

In August 2022, the four associations, Medel, the International Association of Judges, the Association of European Administrative Judges, and Rechters voor Rechters, brought actions for annulment.

2. The gap the general court made visible

The General Court, sitting as Grand Chamber, dismissed the actions as inadmissible in June 2024 (Medel and Others v Council, T-530/22 to T-533/22). Its reasoning applied the settled three-situation framework for associational standing (paragraph 40 of the order under appeal): no procedural rights were granted by the RRF Regulation; the milestones did not directly alter the legal position of judges; and the associations had not established that their own interests as associations were affected. On all three grounds, standing was denied.

The General Court’s reasoning was internally consistent. By applying the Plaumann criteria with their habitual rigour, it made visible a structural gap that had always been there. All of the tools the EU had deployed, infringement proceedings, Article 7 TEU, RRF conditionality, were addressed to Poland as a state. None of them gave the judges most directly affected a route to challenge the validity of EU acts before the EU Courts. The preliminary reference procedure offered an indirect path, but it was not one the affected judges could initiate themselves. It depended on a live national case and could address only the question arising in that specific dispute. What was absent was any route by which those most directly affected could bring such a challenge on their own terms.

At the December hearing, Judge Lycourgos, the reporting judge, pressed the institutions on whether any deterioration in judicial independence, however extreme, could justify a more flexible reading of Article 263. The Council and Commission held their position: the existing toolbox was the proper channel. The General Court’s order reflected that institutional position exactly.

3. What the AG did

AG Ćapeta’s Opinion engages seriously with that gap. Her analysis proceeds in two distinct moves: one addressing the associations’ standing on behalf of sanctioned judges in relation to milestones F2G and F3G, and one addressing their standing in their own name in relation to milestone F1G.

On milestones F2G and F3G, the AG finds that judges actually sanctioned by the Disciplinary Chamber are directly concerned (Opinion, paragraphs 86 to 112). Since the associations are acting on behalf of their members, the threshold question is whether those members are directly and individually concerned by the milestones. The analysis turns on the two-condition test for direct concern established in Nord Stream 2 (C-348/20 P) and recently confirmed in WhatsApp Ireland v EDPB (C-97/23 P): the challenged act must be a direct source of a distinct change in the applicant’s legal position, and no genuine discretion must be left to the implementing addressee.

On the first condition, the AG parts ways with the General Court. The General Court had relied on Nord Stream 2 to conclude that, since the situation of the judges remained governed by Polish law even after the milestone was adopted, there was no direct link between milestone F2G and their legal position. The AG inverts this. Milestone F2G imposes an obligation of result analogous to a directive: it is the obligation to introduce review proceedings, whatever their form, that directly changes the legal position of affected judges (paragraph 86). The fact that their situation continues to be governed by Polish law does not break this link.

On the second condition, the AG finds that Poland’s theoretical discretion not to request RRF funds, and thus not to implement the milestone, is precisely that: theoretical. Poland itself proposed the milestone. It had already adopted the relevant reform through the Law of 9 June 2022, days before the Council formally approved the plan. Endorsing Poland’s own choice while treating that choice as non-binding would, the AG says borrowing from AG Dutheillet de Lamothe in Bock v Commission (62/70), represent excessive formalism (paragraph 109). Those judges also satisfy individual concern under Plaumann (25/62): they form a closed and identifiable group that the milestones specifically targeted (paragraph 220).

The AG’s conclusion on direct concern is, however, limited to judges who were actually subject to Disciplinary Chamber decisions. The appellants had argued that judges not subject to Disciplinary Chamber decisions were also directly concerned, pointing to chilling effects on judicial decision-making and increased workload caused by the absence of sanctioned colleagues. The AG rejects this. Neither the existence of chilling effects nor an increased workload, she finds, is sufficient to demonstrate a direct change in the legal position of those judges within the meaning of the case law (paragraphs 116 to 118).

Turning to milestone F1G, the AG also rejects the claim that judges of other Member States and of the EEA are directly concerned. The appellants contended that the insufficiency of milestone F1G directly affected cross-border judicial cooperation and produced spillover effects across legal orders. The AG acknowledges that rule-of-law deterioration in one Member State can have consequences for others, but finds that this does not establish the direct link required between the milestone’s alleged insufficiency and the rights and obligations of judges elsewhere in the EU and EEA (paragraphs 129 to 131). From a doctrinal perspective, these conclusions are reached relatively briefly. The cross-border dimension in particular deserved more sustained engagement: the functioning of mutual recognition and the preliminary reference mechanism depends on a baseline of shared judicial independence, and whether that observation could ground direct concern for judges in other Member States is a question the Opinion closes without fully exploring.

Before turning to the outcome on milestone F1G for members, the AG makes a prior finding of independent significance: the General Court simply failed to consider how direct and individual concern should be interpreted when applied to associations acting in their own name, referring back to its findings on members without any separate analysis. That absence of reasoning was, in the AG’s view, sufficient on its own to annul the order under appeal (paragraph 145). On milestone F1G, no individual category of judges is directly concerned in the conventional sense: the insufficiency claim does not establish a direct link to specific changes in identifiable legal positions (paragraphs 125 to 131).

Going further, the AG develops a framework for associational standing that the Court has not previously articulated with this clarity. She draws on Carvalho (C-565/19 P), where the ‘collective defending a collective good’ argument was raised but could not be considered (paragraph 188), and on AG Emiliou’s Opinion in Nicoventures (C-731/23 P), which proposed standing where a measure significantly impacts an association’s core statutory activity (paragraph 196).

The AG distinguishes two kinds of associations (paragraphs 189 to 191). The first aggregates members’ interests, and its own interests are the sum of those interests. The second exists to defend a collective societal interest that transcends its membership: judicial independence as a structural condition of the rule of law belongs here. Such associations have genuine own interests that constitute their identity. The AG notes that the three situations recognised in the existing case law are not a closed list (paragraph 139). The framework she proposes is a clarification of Article 263(4) TFEU, not a departure from it.

For direct concern, the test for associations acting in their own name is adapted: not whether the challenged act altered the association’s legal position as such, but whether it is a direct source of effects on the collective interest the association exists to defend (paragraph 207). For individual concern, the test is whether the association’s core interests differentiate it from other associations in relation to the challenged measure (paragraph 209). The AG also situates associational access to the EU Courts within the framework of participatory democracy under Article 11(1) TEU, while being careful to note that access to courts must rest on legal rather than political arguments (paragraphs 164 to 165).

Applied to the present case: milestone F1G concerns judicial independence in Poland, which is precisely the collective interest constituting the identity of all four associations. Poland’s discretion is purely theoretical for the same reasons as with F2G and F3G. Their core interests differentiate them from associations with a more general rule-of-law mandate not specific to Poland (paragraph 234).

4. What the opinion does not do

The associations argued that the dysfunction of Poland’s judicial system justified a broader reading of Article 263(4): if national remedies are unreliable, the EU Courts must compensate. The AG rejects this on both legal and factual grounds (paragraphs 151 to 154). The legal ground is that direct and individual concern are Treaty-level requirements the Court cannot modify, even for compelling reasons, a position anchored in UPA (C-50/00 P) and confirmed in Carvalho. The factual ground is that Poland’s judiciary, despite its problems, is not so broken that relevant cases could not reach Luxembourg by preliminary reference.

The AG similarly declines to follow the ECtHR’s KlimaSeniorinnen judgment (App. no. 53600/20) as direct precedent, finding that victim status under Article 34 ECHR and standing under Article 263(4) TFEU serve different functions (paragraphs 179 to 181). But she extracts its underlying logic: where individual standing is practically foreclosed by the difficulty of satisfying individual concern, associational standing prevents the courts from becoming inaccessible precisely where the stakes are highest (paragraph 184). An exception-based approach, under which standing is available because the situation is extreme, would be impossible to generalise. The identity interest framework is durable, principled, and applicable well beyond Poland and well beyond rule of law cases.

Taken together, these two rejections reveal a tension that runs through the Opinion. As Guillermo Íñiguez has noted in his own analysis of the Opinion, the line between reinterpreting Article 263(4) and modifying it is more blurred than the Court of Justice has traditionally made it out to be. The AG’s subsequent reasoning, which adapts the standing requirements significantly through the identity interest framework, does not fully resolve that tension so much as navigate around it.

5. Conclusion

The Opinion proposes that the case be referred back to the General Court for a decision on the merits, but the Court of Justice must first decide whether to follow it. If it does, the harder questions begin. The core claim about milestone F1G is that it was insufficient, that the Council failed to discharge its obligations under Articles 20(5)(e) and 22 of the RRF Regulation. A claim of insufficiency is only legally cognisable if the law imposed a clear enough obligation to do more. Whether a Member State can negotiate its own rule-of-law commitments and have them endorsed by EU institutions in a way that is insulated from legal challenge remains genuinely open.

What is already clear is that the Opinion proposes something more significant than a procedural win. If adopted, the identity interest framework would mean that associations defending collective constitutional values have a legally cognisable stake in the EU acts that shape those values, in part drawing on Article 11(1) TEU and the Treaties’ own recognition of participatory democracy, though the AG is careful to treat participatory democracy as a contextual consideration rather than a freestanding basis for standing. That would be the first serious attempt to give that kind of actor a home inside Article 263 TFEU. And crucially, it would do so on legal rather than political terms. As the AG observes, bringing a case before the EU Courts to protect collective interests should not be a continuation of a political battle by legal means – courts are not legitimised for political decision-making. The identity interest framework is valuable precisely because it provides a legal route that does not depend on the political circumstances of any particular moment.

That matters more than it might seem. The structural damage to Poland’s judiciary has not been undone by a change of government. A significant number of judges were appointed through procedures whose compatibility with EU law the Court of Justice has called into question. The Article 7(1) TEU procedure was discontinued in May 2024. Parliamentary elections are scheduled for 2027, and the restoration of judicial independence remains ongoing, incomplete, and politically contingent. This case should not be read as less urgent because Poland is moving in a ‘better’ direction. The value of a durable legal framework for associational standing is precisely that it does not depend on the outcome of elections, or on the continued goodwill of any particular government. What the EU’s legal architecture can do when those conditions sharpen again, or when they emerge somewhere else, is exactly what this case is in the process of determining.

The key is in the door. Whether it turns is another question entirely.