There has been no surprise in the two judgments handed down by the Court of Justice of the European Union (CJEU) regarding the amnesty. There is no surprise because the Luxembourg court has adopted almost all the arguments that had already been advanced by the Court’s Advocate General in November 2025. And there is no surprise either because, in fact, the CJEU could not have done anything other than what it has done in applying European Union law, which leads us to the question of why these preliminary questions were formulated in the first place if the answer was going to be so clear and predictable. I will return to that later. The important thing is that they remove one of the few obstacles remaining for the full application of amnesty and, by doing so, they continue to evoke a certain national embarrassment for all Spaniards.
“The LOA was valid from the moment it was approved by the General Courts, the only body competent to enact it in a democratic state governed by the rule of law”
Almost all the media have highlighted that the CJEU has “upheld” the amnesty, that is, Organic Law 1/2024, of June 10 (hereinafter, LOA). I do not think this is the best formulation, because it gives the impression that the Luxembourg court had the power to evaluate the law in its entirety, and in reality it did not. In a country as uncertain as ours, we often seem to feel the need for external validation, when legally that is not the case. The LOA was valid from the moment it was approved by the General Courts, the only body competent to enact it in a democratic state governed by the rule of law. The other supplementary endorsement it needed was that of the Constitutional Court, responsible for verifying its compatibility with Spanish constitutional legality, which it did in STC 137/2025. It is not necessary to repeat all its arguments here, but that ruling already put to rest, once and for all, and with erga omnes effects for all public powers in Spain, the question of the amnesty’s validity. One may not like the LOA for political reasons, one may even question its full democratic legitimacy (despite it undoubtedly having the parliamentary support required), but its legal and constitutional validity cannot be questioned. Sure, one can discuss everything, but the arguments for doing so do not appear plausible. All constitutional democracies in the world have the institution of amnesty. It is so well known that, in fact, the CJEU’s judgments do not even address it, except as a basic premise of their reasoning, when, indeed, they acknowledge the sovereign power of member states to decide a legal framework for a given amnesty. The CJEU, in reality, could not enter into that question.
What has the CJEU decided about the amnesty law?
What exactly has the CJEU decided in these two judgments? Let us recall that the aim was to respond to eight prejudicial questions raised by the Court of Auditors and nine more raised by the National Court, in relation to two proceedings opened respectively in both bodies, one concerning the misappropriation of public funds established in the procès case with possible implications for the European Union’s financial interests, and the other concerning alleged terrorist actions carried out by some CDRs. What was asked of the CJEU was to clarify how EU law should be interpreted on two issues expressly opened by various articles of the LOA itself, which I will explain briefly below. In short, the CJEU had to judge the compatibility between the European legal framework and certain provisions of the LOA and its subsequent application. And the EU’s top court’s answer could not be clearer, more forceful, or more predictable: EU law does not oppose the LOA, except for one minor point, which I will specify later, nor to its full application to its direct beneficiaries, among whom are Carles Puigdemont, Oriol Junqueras, and a long list of Catalan separatists. What the judgments do, therefore, is remove an obstacle, more apparent than real, to the full application of the amnesty.
Let us look, one by one, at the four main substantive legal questions, before asking what should happen from now on.
“The response of the high European tribunal could not be clearer, more forceful, or more predictable: EU law does not oppose the LOA, except for one minor point”
The first concerns the first of the prejudicial questions raised by the Court of Accounts in the context of a repayment procedure initiated in 2022 by Sociedad Civil Catalana and the Public Prosecutor’s Office for those convicted of misappropriation by the Supreme Court’s October 14, 2019 ruling on special case 20907/2017 to repay the misappropriated funds, by virtue of their accounting liability. This repayment process, therefore, predates the LOA’s very approval on June 10, 2024. However, the lodging of the prejudicial question took place on July 29, 2024, a few weeks after that approval. And it makes sense, because it is article 2.e) of the LOA that, by excluding from the law’s application and therefore not granting amnesty to “acts penalized as crimes affecting the EU’s financial interests”, raises the issue of whether the misappropriation for which Junqueras, Turull, Romeva and Bassa were convicted, and for which those who were not tried, including Puigdemont, could be convicted, was a crime that could affect the EU’s financial interests. Seen this way, it would seem fully justified that the Court of Auditors raised this prejudicial question. But things are not that simple.
The order by which the Court of Auditors raises the prejudicial question to the CJEU on July 29 frames the question in terms that are absolutely aligned with the loopy and at times delirious reasoning of the Supreme Court when, in its July 1, 2024 order, later ratified in another order of September 30 of the same year, it refused to apply the amnesty benefit to the aforementioned independents on the basis that their misappropriation had indeed caused potential harm to the EU’s financial interests and thus could not be amnestied. Note that the Supreme Court did not deem it necessary to raise the prejudicial question to the CJEU, to which it certainly was not obliged, because it believed it could interpret EU law correctly on its own. The Supreme Court’s main argument, later adopted four weeks later by the Court of Accounts as the basis for its question, is that, by misappropriating Spanish public funds, in this case from the Generalitat de Catalunya’s budget, the convicted or fugitive separatists had attempted to achieve Catalan independence which, if accomplished, would have caused a necessary reduction in Spain’s contribution to the EU treasury and, consequently, real damage to them. The argument is so flimsy and so poorly constructed that it can only be understood as a desperate attempt by the Supreme Court to find a seemingly legal way to refuse to apply the amnesty to certain Catalan leaders, especially Puigdemont.
It is evident what the second paragraph of article 2.e) of the LOA was designed for: for the hypothetical case in which EU funds had been used to commit any crime within the procès framework, i.e., misappropriation of European funds. And the reason for excluding that case from the amnesty’s application is clear: the Spanish Parliament has competence to grant amnesty for breaches of Spanish law, but it is debatable whether it can grant amnesty for those infringements that violate European law. The aim was not to keep punishing any action that hypothetically and indirectly might affect the EU’s general interests, as would be the case, in some sense, with all actions undertaken to obtain Catalan independence, provided one assumes that the EU’s interests (financial and political) lie in the territorial integrity of the member states and in Catalonia’s continued membership in the Union. If that were the case, there would be no special element related to misappropriation that, as the Supreme Court itself declared proven in the 2019 ruling, affected only the Generalitat’s budget. In reality, there is no necessary causal link between that misappropriation and Catalan independence, and thus neither with the alleged harm to the EU’s financial interests. It would have been less crude to apply such argument to the commission of the offense of sedition, which the Supreme Court did not do.
“A crime that harms only the national budget does not injure the EU’s financial interests”
This ridiculous argument is precisely what the CJEU rejects outright, in paragraphs 95-100 of the judgment in case C-523/24, in a succinct but forceful manner. After admitting that EU financial interests can be harmed not only by completed acts that cause an actual reduction in resources but also by attempted acts that produce such a reduction only potentially, the CJEU warns that a crime that harms only the national budget does not harm the EU’s financial interests, even if one consequence of that crime could be the independence of a territory from a member state, with the consequent reduction of the national income and, afterwards, of that state’s contribution to the Union’s coffers (paras. 97-99). As that territory becomes independent from the member state and thus, automatically, from the European Union itself —the CJEU notes— its obligation to contribute to the Union would cease. It should be noted, incidentally and without giving it undue importance, that the CJEU has clarified one of the questions discussed in the years preceding the illegal 2017 referendum, about whether Catalunya’s possible independence would be compatible with its continued membership in the EU. Could the CJEU have said something different on this point? It is hard to imagine. An “extensive interpretation” of the notion of EU financial interests, such as the Court of Auditors asked for to subsume misappropriation within the exclusion of article 2.e) of the LOA and as the Supreme Court did, would have implied a colossal principle of EU interference in the sovereign competence of member states to grant amnesty for acts it deems appropriate if they are properly delimited and justified as exceptional. And it would imply that, for consistency, not only misappropriation but any criminal act undertaken with the aim of achieving territory independence would cease to be amnestied. The Supreme Court’s lack of consistency in its 2024 order does not imply that the CJEU will refrain from being consistent. And, as I have said, it was predictable that a serious and independent tribunal would respond as it did to a challenge that sprang from a ridiculous argument that only makes sense in a desperate attempt by some Spanish courts to refuse to amnesty those whom the amnesty law clearly aimed to forgive.
Does the amnesty law violate the rule of law?
The second substantive legal question concerns various prejudicial questions raised by the Court of Accounts on whether the LOA harms the European legal framework by violating the principle of effective judicial protection, the principle of legal certainty, the principle of loyal cooperation, the principle of equality before the law, and the principle of non-discrimination. Although the prejudicial questions could not lead the CJEU to rule on the legal validity of the LOA, as I have explained, this was the path chosen by the Court of Accounts to attempt an indirect ruling through showing that certain aspects of the LOA did oppose fundamental principles of the rule of law, not only of the member states but also of the EU itself. Remember that those who sought to challenge the LOA’s constitutionality, beginning with the Supreme Court itself in its 2024 orders, had argued precisely that the LOA was incompatible with all these essential principles of the rule of law. The CJEU’s answer could not be clearer. That democratic and rule-of-law states have the competence to enact amnesty laws is not in question, and by definition, an amnesty law implies an exception to the basic principles and rules of the rule of law, which is an obvious fact. What matters is that such exceptional effects do not have a systemic or permanent character. And in the current case, the CJEU seems to think it is obvious that this is not the case. Again, it was evident that this would be the Court’s response, and the Luxembourg court had no room to maneuver otherwise.
“The fact that the LOA extends the amnesty benefit to those terrorism offenses that do not amount to serious intentional human rights violations is not contrary to EU law”
The third substantive legal question concerns eight of the nine prejudicial questions raised by the National Court with regard to alleged terrorism offenses that some members of one of the CDRs could have committed while being prosecuted by this special Spanish court. Again, the origin of the question lies in one of the LOA’s formal exclusions, this time section 2.c), which explicitly refers to European Directive 2017/541 on the fight against terrorism. That article of the law excludes from amnesty coverage all terrorism offenses that have caused, intentionally, “grave human rights violations.” Of course, the CJEU is not competent to rule on whether those CDRs or any other person have committed such offenses or whether they severely violated human rights. The only aspect it has addressed in the judgment on case C-666/24 is whether the LOA’s extending the amnesty to those terrorism offenses that do not amount to grave intentional human rights violations is contrary to the Directive and to EU law. Again, the clear and forceful answer of the CJEU is no (paragraph 75). And it also finds no issue with the abstract terms in which the LOA expresses the requirement of grave violation, with regard to legal certainty, the principle of loyal cooperation, and the principles of equality and non-discrimination (paragraphs 76-87).
Finally, in the CJEU’s judgments this week only a possible incompatibility of a certain interpretation of the LOA with EU law is pointed out, specifically in case C-523/24. Once again, this pronouncement was entirely predictable and, in fact, the CJEU could not resolve it in any other way. It concerns whether the two-month deadline that LOA sets in article 10 for judicial, administrative, or accounting bodies to apply the amnesty in a fast-track and urgent process can be interpreted as an absolute obligation in which it is not possible to suspend the deadline on the grounds that a prejudicial question has been submitted to the CJEU. Although article 10 does not prevent the filing of appeals, it warns that such appeals cannot have suspensive effects on the obligation to apply the amnesty within that two-month period. That article says nothing about cases in which a prejudicial question is deemed appropriate to be submitted, as in the questions raised by the Court of Accounts and the National Court, or other courts that have presented such questions and are still awaiting resolution. Should it be understood that such prejudicial questions would have suspensive effects on the deadline? The CJEU does not pronounce on the LOA’s validity or compatibility with EU law in this regard. It merely notes that, if article 10 were interpreted in such a way as to prevent courts from presenting such prejudicial questions and awaiting their resolution before continuing with the amnesty’s application, then there would be a conflict with European law, as any jurist examining the case would readily concede. Precisely for this reason, it is also obvious that no one has interpreted it that way, because, in fact, these prejudicial questions were able to be presented and, indeed, the two-month deadline to apply the amnesty was suspended as evidenced by the fact that we are still dealing with this matter more than two years after the LOA was approved.
“It would seem that the aim was merely to delay the process, to stall the inevitable application of the LOA”
In light of all these considerations, one might ask why, if the CJEU’s answers were so clear and predictable, were these prejudicial questions presented by the Court of Accounts and the National Court. We cannot know for certain what objectives they pursued. One might think that, armed with the delusional arguments offered by the Supreme Court in its July 1, 2024 order, they sought to delay the inevitable application of the amnesty to certain independent leaders. Since it was predictable that the Constitutional Court would validate its overall constitutional fit, as it indeed did a year later, just as it is predictable that it will soon resolve the amparo petitions filed against the Supreme Court’s resolution and extend its application to pending cases, and as it was also predictable that the CJEU would respond as it has, it would seem that only a delay of the process could have been sought. But there is also another very different interpretation. Let us recall that the Supreme Court chose not to present a prejudicial question similar to the ones now resolved. It believed it did not need the CJEU’s help to interpret European law itself. Or perhaps it anticipated the obvious: that if it did, the CJEU would resolve as it has, and the Supreme Court would thus lose control over that interpretive issue. The Court of Auditors and the National Court could simply have embraced the Supreme Court’s absurd doctrine. But they chose not to, and they brought the issue to the European jurisdiction in the form of a prejudicial question. Even with the temporal delay that entailed, it was the only way to give the CJEU the opportunity to say what it has said with such clarity. So one could read it as an attempt to pit the Supreme Court against the weaknesses of its own reasoning. All of this is, of course, speculative and we will probably never resolve it.
What now for the amnesty and Carles Puigdemont?
What brings me to the final point of my analysis: what effects do these judgments have and what will happen from now on. The effects are, in one sense, immediate and erga omnes. All Spanish and European judicial, administrative and accounting bodies are obliged to interpret EU law as the CJEU has done, although it is not the European court that can and must apply the LOA to those who are still waiting to obtain that benefit. The matter, therefore, must return to the Court of Auditors and the National Court to apply the amnesty without further delays to those affected by their respective cases. Unfortunately, this will take some time yet.
With respect to those whom the Supreme Court refused to grant amnesty in 2024, such as Carles Puigdemont, it should be remembered that an arrest warrant still hangs over him. The first step is for the Constitutional Court to resolve the amparo petitions filed by those harmed by the Supreme Court’s order. It is predictable that the TC will rule in their favor in the coming months and will order the Supreme Court to issue a new order in which, now indeed, they will be granted the LOA’s protection. But new chapters may yet open in this long story. We cannot foresee how long it will take the Supreme Court to comply with such an instruction from the Constitutional Court or whether it will resort, for example, to presenting a new European prejudicial question about some new unresolved aspect that has not yet been settled by this week’s two judgments and that again disrupts deadlines.
It is very likely that all of this will not be resolved definitively until well into 2027. Unless something extraordinary occurs, such as Puigdemont entering Spanish territory with an ensuing arrest, which would hasten some of the pending decisions. But let us not dismiss, even in that scenario, the formidable capacity of some courts, particularly the Supreme Court, to complicate matters and resist a legislation they dislike, i.e., to fail to comply with their strict obligations within a state governed by the rule of law.
“It is very possible that all of this cannot be resolved definitively until well into 2027”
If we are where we are, without being able to settle this issue once and for all, it is precisely because some of those courts have overstepped the bounds of their jurisdiction, a point that has been implicitly noted by the CJEU’s own judgments when they remind us that the courts should limit themselves to applying the laws, including amnesty laws. Being reminded of this so clearly should give us all a certain national embarrassment.
