Fifteen Years Is Enough. Europe’s Top Court on VAT Fraud, Slow Judges and Why Running a Criminal Group May Cost More Than the Crime

On 15 October 2026 the Court of Justice of the European Union ruled in Gidzhinov (C-321/25) that a fixed, fifteen-year limitation period for serious VAT fraud does not breach EU law, provided it does not create a systemic risk of impunity in the Member State concerned, and that a state may punish membership of a criminal organisation more harshly than the very fraud the organisation was set up to commit. The first answer draws a line for a state that ran out of time. The second shows that where a state wants to punish more, Luxembourg will draw no line at all. Between those two answers lies the whole modern argument about the rule of law and organised economic crime.

Sofia, three benches and two clocks

The indictment was drawn up in 2015. Five people, identified in the judgment only as A, B, V, G and D, were to answer for a criminal group said to have operated from March 2011 to 14 November 2012. Two were charged with leading it, three with belonging to it. Four were also charged with VAT fraud committed between June 2011 and October 2012, each count worth between a little over 600,000 and a little over 1.2 million euros, more than 4 million euros in all. These are figures from the indictment, not from a conviction; no verdict has ever been given in this case.

On 26 November 2015 the case was assigned to a bench of one professional judge and two lay judges. Over six years that bench held forty-eight hearings. Then the judge was appointed a European Delegated Prosecutor, which is to say he left to join the body the Union had created to prosecute fraud against its own finances, and Bulgarian procedure requires the same bench to hear a case from beginning to end. The VAT fraud trial therefore started again from scratch because its judge had gone off to prosecute VAT fraud. The second bench held eighteen hearings between 2022 and 2025, and the Court records drily that no concrete step was taken to gather evidence. On 14 March 2025 the second judge retired. The third bench, rather than begin a third time, asked Luxembourg.

The question arose from the arithmetic of two clocks. The Bulgarian Criminal Code ties limitation to the maximum penalty: ten years for offences punishable by more than three years’ imprisonment, fifteen for offences punishable by more than ten. Every procedural act interrupts the period, but once one and a half times the basic period has run, prosecution is barred absolutely, whatever the interruptions. VAT fraud (three to eight years) and membership of the group (three to ten) are thus barred after fifteen years: in 2026, in 2027 and on 14 November 2027. Leading the group, punishable by five to fifteen years, will not be barred until 14 May 2035, twenty-two and a half years after the group dissolved. The Sofia court said plainly that it could not get through three instances within the first deadline, but could within the second. And it asked whether EU law forbids a national statute that computes limitation solely from the maximum penalty, with no regard to the complexity of the case or to unforeseen events, when the result may be impunity.

It is worth pausing over who was asking. No defendant was pleading for protection. A court that had failed for nearly ten years to try a case was asking whether EU law would let it keep hold of the accused. The argument about the rule of law is rarely put so cleanly.

The first answer: the state had enough time

The Court first settled what the case was about. Acts committed in 2011 and 2012 still fall under the Convention on the protection of the European Communities’ financial interests of 26 July 1995 (OJ C 316, 27.11.1995, p. 48), which was replaced by Directive (EU) 2017/1371 of 5 July 2017 (OJ L 198, 28.7.2017, p. 29) only with effect from 6 July 2019. Fraud of more than 4 million euros is “serious” within the meaning of Article 325 TFEU and Article 2(1) of the Convention, and setting up a criminal organisation to commit VAT fraud has, since Taricco, counted in itself as serious fraud against the Union’s interests. At the time of the acts nobody in the Union had harmonised limitation periods; the Member States set their own, bound only by the duty not to let their rules generate a systemic risk of impunity.

The Court then did something the referring court probably did not expect: it praised the absolute limit. Such a limit, paragraph 46 reads, strengthens legal certainty by preventing limitation from being dragged out indefinitely through repeated interrupting acts. It need only be long enough not to make the protection of the Union’s finances impossible or excessively difficult, which is judged by looking at the limitation system as a whole and at the complexity of the proceedings it has to accommodate (paragraph 47). Fifteen years leaves the authorities, as a rule, a reasonable time to bring a case to a final judgment, even through three instances (paragraph 50). The referring court, for its part, had produced nothing to suggest that it is systemically and excessively difficult in Bulgaria to try serious fraud within fifteen years; changes of bench in individual cases are not enough to establish a systemic risk of impunity (paragraph 51).

Advocate General Manuel Campos Sánchez-Bordona, in his Opinion of 23 April 2026, said the same thing without the diplomacy: nothing in the file suggests that Bulgarian courts are missing deadlines wholesale, whereas the sequence of events described by the referring court points, in the Commission’s phrase, to the systemic inefficiency of the court itself (paragraphs 48 to 50 of the Opinion). In other words, the problem in Sofia was not the statute. It was the court.

That is the heart of the first answer. The Court moved the question from the level of the case to the level of the system. A state may lose an individual trial through its own inefficiency; what it may not do is build rules that lose trials as a matter of course. But the fact that it lost one is not something it may repair by helping itself to the defendant’s time. The citizen’s guarantee is not a reserve on which the state draws when it is late. That is a line Luxembourg drew for the state, and it is worth remembering, because in the second half of the judgment Luxembourg will show where it draws no lines at all.

Two caveats, so as not to read more into the judgment than is there. First, the Court did not hold that fifteen years is a minimum, or that shorter periods are contrary to EU law; it held that this period, in this system, in the absence of data about systemic impunity, is in order. Secondly, accepting a fifteen-year limitation period is not accepting a fifteen-year trial. Whether proceedings have lasted a reasonable time is assessed separately and case by case, having regard to complexity and to the conduct of the authorities and the parties, as the Court recalled in Kolev (judgment of 5 June 2018, C-612/15). The defendants in Sofia have been awaiting a verdict for more than ten years, and that is a separate problem which no limitation period solves.

Italy: champagne, a carousel and a limitation period that only pretended

To grasp the weight of this restraint one has to remember where the Court came from, because it came from a long way off. In 2014 the court in Cuneo, in Piedmont, was trying a group that between 2005 and 2009 had run a carousel on bottles of champagne: shell companies bought the goods free of VAT and issued fictitious invoices, a company called Planet deducted the tax on those invoices and sold the champagne below market price, and the shells filed no returns or paid nothing (paragraphs 18 and 19 of Taricco). Italian law computed limitation from the maximum penalty, but an interruption could extend the period by at most a quarter; that gave eight years and nine months for the organisers and seven and a half for the members, and every count would be time-barred by 8 February 2018 at the latest, before any final judgment could be given; one defendant had already left the case that way (paragraphs 20 to 22). The Cuneo court wrote the sentence that set the whole line of case law in motion: in tax fraud cases, impunity in Italy is the rule rather than the exception, and the revenue as a rule never recovers the tax (paragraph 24).

The Grand Chamber answered on 8 September 2015 in Taricco (C-105/14, ECLI:EU:C:2015:555) with a force Luxembourg rarely uses in criminal matters. Article 325(1) and (2) TFEU imposes on the Member States an unconditional obligation of result and has direct effect (paragraphs 51 and 52). If national rules on interrupting limitation mean that, in a significant number of cases, serious VAT fraud becomes time-barred before a final judgment, the national court must disapply them without waiting for the legislature (paragraphs 47 and 49). Equality was added for good measure: Italy had no absolute limit for organised tobacco smuggling but had one for VAT fraud, which is to say it protected its own excise duty better than the Union’s tax (paragraph 48). The defendants’ rights, the Court held, do not suffer, because extending a period that has not yet expired is not retroactive punishment; so says the Strasbourg case law since Coëme (paragraphs 55 to 57).

Italy did not take this calmly. By an order of 23 November 2016 the Corte costituzionale asked Luxembourg whether it meant this literally, and signalled that in the Italian order limitation belongs to substantive criminal law, covered by the constitutional principle of legality in Article 25 of the Constitution, and that the test of a “significant number of cases” is so vague that a criminal judge cannot tell when to apply it (paragraphs 13 to 17 of M.A.S.). In M.A.S. and M.B. of 5 December 2017 (C-42/17, ECLI:EU:C:2017:936), decided under the expedited procedure and known as Taricco II, the Grand Chamber stepped back without retracting anything in terms. The duty to disapply limitation rules remains, but it yields where carrying it out would breach the principle that offences and penalties must be defined by law: where the state of the law would become unforeseeable (paragraph 59), or where disapplication would reach acts committed before Taricco was delivered, since that would retroactively worsen the conditions of liability (paragraph 60). The Court added a sentence it would return to: it is for the national legislature, first of all, to build a limitation system that does not produce impunity (paragraphs 41 and 61).

Romania: four years in which nothing stopped the clock

The Romanian chapter began with a constitutional accident. In 2010 several people sold diesel bought under an excise suspension arrangement without recording the sales; the VAT and excise loss came to about 3.24 million euros, and convictions followed in 2018 and 2020, including for forming an organised criminal group (paragraphs 19 to 22 of Lin). Meanwhile, on 26 April 2018, the Romanian Constitutional Court held that a provision under which “any procedural act” interrupts limitation, including acts of which the suspect knows nothing, was unforeseeable and therefore unconstitutional (paragraphs 24 to 26). The legislature did nothing for nearly four years (paragraphs 27 and 94). On 26 May 2022 the same court found that throughout that time Romanian law had known no cause of interruption at all (paragraph 28); four days later the government closed the gap by emergency decree (paragraph 30). The Supreme Court then ruled that the rules on interruption belong to substantive law, so that the interruption-free state of affairs applies retroactively as the more lenient law and reopens even final convictions (paragraphs 31 and 32). The European Commission wrote in November 2022 that thousands of defendants might escape liability (paragraph 90).

Lin of 24 July 2023 (C-107/23 PPU, ECLI:EU:C:2023:606), handed down under the urgent procedure because two of the convicted men were in prison, cut this knot with precision. A systemic risk of impunity exists (paragraph 91). But the Constitutional Court’s judgments, resting on the foreseeability and precision of criminal law, need not be disapplied, exactly as in Taricco II; protection against arbitrariness is a component of the rule of law under Article 2 TEU (paragraphs 114 to 118). What the Romanian courts must disapply is the Supreme Court’s ruling, because it stretches the hole backwards over acts before 2018 by a further four years and more, and does so not in the name of foreseeability but of a national lex mitior standard which the Charter of Fundamental Rights does not cover: Article 49(1) of the Charter does not deal with limitation rules at all (paragraphs 108 and 122 to 124). The Court added that a judge who applies EU law may not be disciplined for it (paragraphs 136 and 137).

Romania answered as Italy had, only more openly. Two panels of the Supreme Court held in 2024 that Lin breached Article 7 of the European Convention on Human Rights and the constitutional ban on stitching a “third law” together from fragments of successive regimes, so that Lin applied only in part (paragraphs 40 to 45 of Lin II). The case that went back to Luxembourg was small: a director of two companies, twelve fictitious invoices from 2012 to 2014, a loss of about 59,000 euros, of which VAT about 36,000 (paragraph 30). In Lin II of 16 July 2026 (C-280/25, ECLI:EU:C:2026:586) the Grand Chamber first held that any fraud in which the total loss exceeds 50,000 euros is “serious”, whatever share of that loss falls on the Union budget (paragraphs 79 and 96). It then confirmed Lin in full: the Romanian courts must disapply the lex mitior ruling despite the ban on a “third law” (paragraph 163), despite the absence of national criteria for assessing systemic risk, since the Court had found that risk itself (paragraphs 168 and 175), and even where under the Romanian standard the period would have expired before Lin was delivered (paragraph 236). The one boundary is res judicata: a finding of limitation that has become final need not be disturbed (paragraphs 180 and 235). The Court reproached the Romanian Supreme Court for ruling without asking Luxembourg and thereby manufacturing legal uncertainty of its own (paragraph 234), and recalled in passing that professionals, VAT-registered traders among them, are expected to assess the legal risks of their business with due care, taking advice where necessary (paragraph 209). Romanian scholars replied that the state had shifted on to the individual the cost of its own four-year inaction, an argument set out in a polemic by Hotca and Neagu.

What Luxembourg will not tolerate, and what it praised

Set the three countries side by side and it becomes clear what the Court is really fighting. In Italy the period could formally be extended, but only by a quarter, so in complex cases the extension was a fiction. In Romania the interruption mechanism vanished for four years through a constitutional accident and a parliament’s inertia, and was then stretched backwards by a court. In both, the law pretended to allow prosecution while in practice preventing it. Bulgaria said it straight: fifteen years and not a day more. Luxembourg hunts down sham limitation; honest limitation it not only leaves alone but commends. That, in my view, is the key to Gidzhinov and at the same time the most practical lesson for any legislature: a long rigid period is better than a short leaky one.

The pendulum was pushed, each time, by a national court that wanted to convict and could not, and in Sofia the same kind of court was asking. The Court could have taken the Taricco road and said that a statute which refuses to extend prosecution in the face of unforeseen events threatens impunity. It did not. Instead of asking how to get around the deadline, it asked why fifteen years had not been enough for the state. The impunity EU law requires states to fight is impunity manufactured by rules, not impunity caused by a judge’s retirement; which is, incidentally, an answer to the Romanian objection to Lin II.

Two reservations follow from the same reading and cool the enthusiasm. First, the protection the Sofia defendants obtained is not a right. Article 49(1) of the Charter guarantees nobody that they will remain subject to the limitation rules in force on the day of the offence; it forbids only the resurrection of liability that has already expired (Lin, paragraph 108; Lin II, paragraphs 189 and 190). The Court did not hold that a citizen is entitled to fifteen years; it held that the state had failed to prove systemic impunity. That is an evidentiary ruling, not a guarantee, and had the Bulgarian government brought statistics of the Romanian kind, the answer might have been different. Secondly, the assessment of systemic risk no longer belongs to the national judge. In Taricco the Court told the Cuneo court to carry it out, in Taricco II it conceded that the test was too vague for a criminal judge, and in Lin II it solved the problem by making the finding itself, on the basis of Commission reports, and relieving national courts of the exercise (paragraphs 168 and 175). A Polish court that wished today to invoke systemic impunity, or its absence, could not do so on its own; it would have to ask Luxembourg and bring data. The same goes for a prosecutor.

There is something in this reading that sits uneasily with equality before the law. In Romania a defendant accused of fraud against the Union budget will not benefit from the more lenient law, while a defendant accused of fraud against the Romanian budget alone will; the Court found that difference rationally justified by the aim of protecting the Union’s finances (Lin II, paragraphs 159 and 160). Limitation in the Union now runs at two speeds, depending on whose budget was shortchanged, and the 50,000-euro threshold on total loss means that in VAT cases it is almost always the Union’s.

This line of case law deserves its due, because it holds strong cards and was not born of zeal. The Italian model did not produce the occasional accident; it produced impunity in series, and the Romanian hole threatened thousands of dismissals. EU law does not require the state to win; it requires it not to lose by its own design. Dzivev of 17 January 2019 (C-310/16) marked the boundary of that logic from the other side: even where wiretaps authorised by a court lacking jurisdiction are the only evidence of VAT fraud, EU law does not require their use. Gidzhinov completes the picture: neither unlawful evidence nor unlimited time is the price a citizen pays for the state’s effectiveness.

The second answer: the structure may cost more than the deed

The second question looked academic and is, for a Polish reader, the more important. Bulgaria punishes leading a group formed for gain with five to fifteen years’ imprisonment, membership with three to ten, and the VAT fraud itself, in its aggravated form, with three to eight. The Sofia court reasoned that an agreement to commit an offence is criminal solely by reason of its object, so the penalty for the agreement cannot exceed the penalty for the object; otherwise proportionality is breached. It asked about Article 3(1)(b) of Council Framework Decision 2008/841/JHA of 24 October 2008 on the fight against organised crime (OJ L 300, 11.11.2008, p. 42). That provision requires Member States to make such an agreement punishable by a maximum term of imprisonment that is “the same as that provided for the offence at which the agreement is aimed”, or else by “a maximum term of imprisonment of at least between two and five years”.

The phrase itself is the problem. The Advocate General wrote that a “maximum term of at least between two and five years” can hardly satisfy the rules of logic (paragraph 66 of the Opinion): a maximum that is “at least” is not a maximum. The Court waded through the language versions (the Italian speaks of a term “of between two and five years”, the Spanish, German, English, Polish and the rest of “at least” that range) and fell back on structure: the Framework Decision was adopted on a Treaty basis that permitted only minimum harmonisation, now Article 83(1) TFEU. The Council conclusions of 24 and 25 April 2002 on the approximation of penalties confirm it: they set four minimum thresholds for maximum penalties (one to three years, two to five, five to ten, at least ten) and expressly stated that nothing prevents Member States from setting higher ones (paragraphs 66 and 67 of the judgment, 61 to 64 of the Opinion). The range “two to five years” is therefore a floor, not a ceiling. A state may threaten the agreement with more than five years, and with more than the offence the agreement was aimed at.

Why may it? The Court gives two reasons. The first is textual: the Framework Decision defines a criminal organisation as a group aimed at offences punishable by at least four years, yet allows the agreement to be punished with up to five; the EU legislature itself therefore contemplated the agreement costing more than its object (paragraph 70). The second is purposive: an agreement in the sense of the Framework Decision is not an ordinary conspiracy of two, but conduct connected with a structured association of more than two persons established over time, and the Framework Decision was adopted to combat the specific dangers of such organisations and their spread (paragraphs 71 to 73). The Advocate General spells out what the judgment leaves implicit: an occasional agreement about a single act is not the same as an agreement to commit offences steadily and in an organised way within a structured group; whoever builds such a structure makes possible offences that could not be committed without it, and may be answerable for the acts of others (paragraphs 80 to 82 of the Opinion). Proportionality of penalties under Article 49(3) of the Charter has not been repealed, but the Bulgarian ranges do not “in themselves” appear to exceed what is appropriate and necessary (paragraph 77). The Advocate General was blunter still: the assessment of proportionality belongs to the national legislature and takes precedence unless it manifestly oversteps the bounds of reason (paragraph 83 of the Opinion).

The infrastructure argument deserves to be stated in its strongest form, because in VAT cases it is true. A carousel is not an act; it is a machine: the missing trader, the buffers, the broker, the bank accounts, the cash courier and the lawyer who knows how to service it. The person who never touches an invoice is often the most dangerous part of it, because that person supplies durability, division of labour and resilience to the loss of a single link. The harm of a structure is measured by its capacity to repeat, not by the size of one invoice. Europol’s SOCTA report of 2025 describes VAT fraud as the business of networks using buffer companies, captured legitimate firms and professional tax, financial and legal expertise. If the law wants to strike the machine rather than its cogs, it must have a provision that punishes the building of the machine.

And here begins the problem the judgment leaves untouched. The offence that may be punished more harshly than the fraud is also the offence easiest to allege. The Framework Decision defines a “structured association” as one whose members need not have formally defined roles, continuity of membership or a developed structure (Article 1(2)). Polish case law says the same: even the slightest degree of organisation, without fixed roles, does not preclude an organised criminal group (Court of Appeal in Białystok, judgment of 29 December 2015, II AKa 179/15). It happens that three companies in a supply chain and a shared bookkeeper suffice for the word “group” to appear in an indictment, and with it the harshest available penalty, the longest limitation period and the whole arsenal of measures that such a charge switches on. The ceiling on punishment is high; the threshold for the charge is low. Gidzhinov sanctions the first and says nothing about the second, because nobody asked. For an honest trader caught up in a supply chain, the difference between “knew or should have known” in the tax case law and intentional participation in an organisation with knowledge of its aim, which Article 2 of the Framework Decision requires, is in practice thinner than the textbooks claim; I have written about it in (VAT Carousel Fraud).

There is one more asymmetry, technical but far-reaching. The effectiveness obligation in Article 325 TFEU has direct effect: in Taricco it was able to order a national court to set aside a statute to the defendant’s detriment. The limits written into a Framework Decision have no direct effect; as the Advocate General recalled, citing Popławski (paragraph 73 of the Opinion), a national court is under no obligation to disapply a national provision that conflicts with a Framework Decision. Had the Court found fifteen years for leading a group excessive, the Bulgarian court would still have had to apply the provision. The only hard limit left is the Charter: legality, defended in Taricco II, and proportionality, which Gidzhinov tested with the words “in themselves”. EU law in this field is, to simplify only slightly, a mechanism with a one-way drive: it can force more repression and can rarely force less.

Poland: a ratchet that needs no help from Luxembourg

I described this mechanism in (The Boomerang: How Instruments of Repression Turn on Their Makers): instruments of repression do not go back in the box, and every failure of effectiveness becomes an argument for tightening them. Mises called it the spiral of interventionism; Robert Higgs, following Hayek, the ratchet effect. Gidzhinov adds to that picture the piece I was missing at the time. The European level is not a brake on the ratchet. It is its lower stop: it fixes how far down one may not go and leaves open how far up one may climb.

Poland has tested this experimentally. Since 1 October 2023, after the amendment of 7 July 2022 (Dz.U. 2022, item 2600), participation in an organised criminal group or association under Article 258 § 1 of the Polish Criminal Code of 6 June 1997 (Kodeks karny, Dz.U. 1997 No. 88, item 553, as amended) carries six months to eight years (previously three months to five), and founding or leading such a group under Article 258 § 3 carries two to fifteen years (previously one to ten). The ceiling for leading a group is thus the same as in Bulgaria, while the tax fraud itself is punished more leniently than there. Making false statements in a tax return (Article 56 § 1) and obtaining an undue refund (Article 76 § 1) of the Fiscal Penal Code of 10 September 1999 (Kodeks karny skarbowy, Dz.U. 1999 No. 83, item 930, as amended) carry imprisonment which, under that code, runs to at most five years (Article 27 § 1), and which after extraordinary aggravation, provided among other things for acting in an organised group (Article 37 § 1(5)), may not exceed ten years (Article 28 § 2). The organiser therefore risks fifteen years, the fraudster ten. Bulgaria asked Luxembourg whether it was allowed fifteen against eight. We have fifteen against ten and nobody asked anybody. On top of that, Article 65 § 1 of the Criminal Code treats an offender acting in a group like a multiple recidivist: a sentence above the statutory minimum, up to the maximum increased by half.

And where the Polish legislature did not wish to punish the structure more than the deed, it solved the problem differently: it raised the penalty for the deed. Since 2017, forging or falsely certifying invoices for a total exceeding PLN 10 million has been a felony under Article 277a § 1 of the Criminal Code, now punishable by five to twenty-five years. That is more than the maximum for causing grievous bodily harm and for armed robbery, which in both cases is twenty years. With a penalty like that for the deed, the question whether the group may be punished more harshly loses its point; in Poland the ratchet clicked past that tooth long ago.

Limitation tells a similar story, though here the picture is more interesting, because Poland formally has what the Court praised: fixed periods with no interruption mechanism. A misdemeanour punishable by more than five years is barred after fifteen (Article 101 § 1(2a) of the Criminal Code), and if proceedings against the person were instituted within that time, liability expires only after a further ten (Article 102, since 2 March 2016; previously five). Leading a group is thus barred in Poland after twenty-five years, the invoice felony after thirty, and fraud under Article 56 of the Fiscal Penal Code after ten years plus ten (Article 44 § 1(2) and § 5). Our absolute periods are longer than Bulgaria’s. Had a Polish court asked Luxembourg about Gidzhinov, the question would not have been “may we be so short” but “may we be this long”, and the answer would be the same: yes, because EU law sets only a floor.

The threshold from Lin II is worth keeping in mind here. Fraud is “serious” within the meaning of EU law once the total loss exceeds 50,000 euros, a little over PLN 200,000, counting the whole loss to the national budget and not only the Union’s share. In Polish conditions that means almost every carousel case that reaches a prosecutor falls under the effectiveness regime of Article 325 TFEU, with everything that follows: the duty to prosecute, the ban on rules that generate impunity, and the possibility of disapplying a national standard of protection unless it rests on the foreseeability of criminal law.

Poland’s real problem lies elsewhere: in the fact that formally rigid periods have in practice proved elastic. Two episodes. The first is Article 15zzr¹ of the COVID statute, which in criminal and fiscal criminal cases suspended limitation for the duration of the epidemic and the state of epidemic threat, with no end date and no cap, and left the decision as to when time would start running again in the hands of the executive. The Constitutional Tribunal, in its judgment of 12 December 2023, P 12/22, found the provision incompatible with Article 2 of the Constitution, and whatever one thinks of the disputes over the Tribunal’s status in that period, its reasoning is the same that Luxembourg set out three years later in paragraph 46 of Gidzhinov: a period without an end is not a period. The provision lapsed on 1 October 2023.

The second episode is ongoing. Article 70 § 6(1) of the Tax Ordinance of 29 August 1997 (Ordynacja podatkowa, Dz.U. 1997 No. 137, item 926, as amended) suspends the limitation of a tax liability the moment fiscal criminal proceedings are opened, and for years such proceedings were opened in the last weeks of December, not to prosecute anyone but to stop the tax from becoming time-barred. The Supreme Administrative Court, in its resolution of 24 May 2021, I FPS 1/21, allowed the administrative courts to examine whether such an opening was instrumental; I described the mechanism in (Tolling and Interruption of the Statute of Limitations on Tax Obligations: When the Fisc Buys Time). On 15 May 2026 the Sejm passed, by 241 votes to 199, a statute repealing that provision. But the same statute repealed Article 44 § 2 of the Fiscal Penal Code, under which liability for a fiscal offence also expires once the underlying tax claim is time-barred, and opened the way to recovering the equivalent of the tax in criminal proceedings after the tax itself had lapsed. On 11 June 2026 the President declined to sign the statute, reasoning that while it abolished suspension by the opening of proceedings, it simultaneously allowed the authorities to pursue tax claims without any time limit, which he found contrary to Articles 2 and 217 of the Constitution. Overriding a veto takes three fifths of the votes; as of the day I write, the Sejm has not overridden it. The law therefore stands as it did before.

I do not wish to judge who was right, because both sides were, and that is precisely the lesson. The legislature wanted to close the door through which the tax authority bought itself time, but attached to it a new way of reaching for money outside time. The veto blocked the new way and kept the old door. The citizen was left with the status quo. The ratchet not only declines to go back of its own accord; when it cannot go forward it stands still, and every attempt to turn it back comes home with a new tooth attached. The Court told the Bulgarian state that fifteen years is enough. In Warsaw in 2026 nobody managed to agree what enough is, so it stayed what it was.

Numbers that justify the seriousness but do not replace the evidence

An opponent of this diagnosis will answer with numbers, and is entitled to. The European Public Prosecutor’s Office reported in March 2026 that at the end of 2025 it had 3,602 active investigations with an estimated damage of 67.27 billion euros, of which 981 VAT and customs cases accounted for 45.01 billion, more than two thirds of the total. A year earlier the estimate was 24.8 billion euros, and a tripling within a year says, in my view, as much about better detection and a handful of very large cases as about the growth of crime. The European Commission estimates the VAT gap in the Union for 2023 at 128 billion euros, 9.5 per cent of the tax due, but the gap also covers insolvencies, errors and late payment, not only fraud. These numbers justify treating VAT fraud as organised crime and justify, in principle, punishing the structure separately. They do not justify prosecuting the honest link in the chain, or stopping the clock.

More instructive is what the numbers do not say. In the essay on the irreversibility of repression I cited figures from Poland’s Supreme Audit Office: between 2014 and 2018 about PLN 3 billion was secured against illicit gains estimated at PLN 200 to 500 billion. The European Public Prosecutor’s Office obtained freezing orders in 2025 worth 1.13 billion euros against 67 billion of estimated damage. Criminology has repeated for decades what the United States National Institute of Justice summarises: what deters is the certainty of detection and punishment, not its severity. The Sofia case is laboratory proof. Those accused of leading the group faced fifteen years in prison, and it had no bearing whatever on their calculation, because the calculation was set not by the size of the penalty but by the probability of a final judgment, which three benches failed to produce in more than ten years. The effectiveness of criminal law is a property of institutions, not of statutory maxima. A state that raises penalties instead of staffing its courts and gathering evidence buys itself a feeling of action with other people’s money.

Fifteen years as a European measure

One further circumstance gives the number fifteen a wider meaning. Directive (EU) 2024/1260 of the European Parliament and of the Council of 24 April 2024 on asset recovery and confiscation (OJ L, 2024/1260, 2.5.2024), which Member States must transpose by 23 November 2026, requires in Article 15 that confiscation be possible without a conviction where criminal proceedings were initiated but cannot continue, among other reasons because the limitation period under national law is shorter than fifteen years and expired after the proceedings began. Luxembourg says that fifteen years is enough to convict. Brussels says that whoever has less than fifteen must be able to take the property without convicting. The same number becomes a European measure of the state’s time: that is how long it has for a verdict, and below that it must keep a fallback. In the logic of the spiral I described, this is the next turn: when the criminal track runs out, the property track runs on. What the Directive changes in Polish law, and how one defends against such seizure, is set out in (Criminal Asset Forfeiture and Extended Confiscation). The Directive does not apply to acts from 2011 and 2012 and no retroactive basis for anything may be drawn from it; my point is about the direction, not this case.

What follows

Gidzhinov is neither lenient towards defendants nor severe. It is a judgment about separating three things that public debate on economic crime constantly runs together: the individual’s guarantees, which the state may not breach because it is in a hurry; rules that actually produce impunity, which EU law requires to be removed; and the ordinary inefficiency of the state, which may not be repaired at the defendant’s expense. The thesis I am putting forward is this: a state cannot choose between impunity and lawlessness; it has to be able to prosecute effectively within the law, and that last part is a task of organisation, not legislation. The Court’s first answer supports the thesis outright. The second shows its price: wherever a state chooses more repression, EU law will be silent, and the burden of keeping the measure falls on national constitutional culture, on constitutional courts, on courts reviewing instrumental proceedings and, as it turned out in June, on a veto.

For the defence the lesson is sober: an argument from delay is not an argument from systemic impunity and the two must not be confused, and a state that wants more time for itself must show a defect in the rules, not the history of one case, and must show it in Luxembourg, with data, because that, not a regional court, is where the finding on systemic risk is now made. It should also be remembered that EU law knows no right to limitation; it knows only a ban on resurrecting liability that has already expired. The rest of the guarantees are given, or not given, by the national constitution and statute. For an entrepreneur who receives a summons about a supply chain from years ago, the lesson is different: it is not the charge under Article 56 of the Fiscal Penal Code that changes his life but the word “group”, because behind it come fifteen years of exposure, twenty-five years of limitation, Article 65 of the Criminal Code, prejudgment seizure and extended confiscation. The dispute over whether the structure existed at all is, in such cases, a dispute over everything. For the legislature, finally: before reaching for a sixteenth year, it should ask why fifteen were not enough.

The sword I wrote about does not ask whom it is to strike. Gidzhinov adds the old truth that it does not ask, either, whether the state will make it in time. A state that cannot finish a trial in fifteen years does not need a sixteenth year. It needs a court.

Legal position as at 25 October 2026.

Robert Nogacki

Robert Nogacki is a Polish attorney at law (radca prawny), the founder and managing partner of Kancelaria Prawna Skarbiec (Skarbiec Law Firm), which has operated continuously since 2006.

The law is equal for everyone, but the parties rarely are: on one side stands an organization with time, money, and lawyers, on the other a person with one business, one nest egg, and one life.

Clients rarely come to him with a legal problem. They come with a problem that also has a legal side: an audit that began with a single invoice, money entrusted to someone who has disappeared, a company that has to be passed on before it is too late. Most such matters are decided long before the first letter is written, in decisions made without asking and in deadlines nobody remembered. So he begins by asking how the client got here, not what the client should have done.

He advises entrepreneurs and families from more than a dozen countries, including those whose accounts the tax office has just seized and who do not know what to do tomorrow morning. He defends them in tax audits, customs and fiscal inspections, disputes with the tax authorities, and criminal tax proceedings. He represents victims of investment fraud and Ponzi schemes. He helps families set up family foundations and plan succession, so that a life’s work outlasts a single generation.

Not every case can be won. Every case can be run so that the client knows where they stand. Since 2006 he has represented the victims in the WGI case (Warszawska Grupa Inwestycyjna, the Warsaw Investment Group), one of the longest criminal cases in the history of the Polish financial market, because some things must not be left half finished, even when they take two decades. In the case of the collapsed cryptocurrency exchange Zonda (Zondacrypto, operated by BB Trade Estonia OÜ), he represents several hundred victims in the criminal investigation conducted by Poland’s National Prosecutor’s Office and in the Estonian bankruptcy proceedings.

Kancelaria Prawna Skarbiec is listed in the rankings of Poland’s largest tax advisory firms published by Dziennik Gazeta Prawna and Rzeczpospolita, and it is a four-time recipient (2015 to 2018) of the European Medal awarded by the Business Centre Club and the European Economic and Social Committee. Robert Nogacki publishes regularly, in the press and on the firm’s website, for people who have a problem rather than a law degree, because a legal opinion the client cannot understand protects only the lawyer.

He believes that the best legal advice is the kind that means the client never has to appear in court.