Are Europe’s Competition Authorities Becoming the Market Police?

Unannounced inspections, searches of private homes, sealed offices, forensic copies of entire mailboxes, police officers at the door – and, if a current Polish bill is enacted, an official who buys a subscription under an assumed name to see whether it can be cancelled. EU enforcement law has steadily equipped administrative authorities with methods long associated with criminal investigations. Are national competition authorities (NCAs) becoming a form of market police?

The driving force is competition law. The ECN+ Directive 2019/1 completed what Regulation 1/2003 had begun: it obliged Member States to give every NCA a common core of coercive powers modelled on those of the Commission. The same model has since migrated into newer EU instruments, from the Digital Markets Act and the Digital Services Act to the Foreign Subsidies Regulation, which leads to observe that, when proposing new legislation, the Commission routinely suggests vesting supervisory authorities with far-reaching investigatory powers. Consumer law follows the same path. The CPC Regulation, as an earlier post on this Blog pointed out, harmonises minimum powers for consumer authorities much as ECN+ did for antitrust authorities, dawn raids included. Poland’s UC142 bill, which would implement the CPC Regulation, shows where this migration may lead: certain covert test purchases would be carried out by the Police at the request of the competition and consumer authority.

The benchmark for this development is set as much in Strasbourg as in Brussels or Luxembourg. Since Menarini, the Strasbourg Court has treated antitrust fines as ‘criminal’ for the purposes of Article 6 ECHR; since Société Colas Est, competition inspections have been measured against Article 8. In an earlier post on this Blog I traced how the European Court of Human Rights built from these cases a standard that is substantive rather than formal: what counts is whether review works in practice, not whether a judge signed something at the outset. That standard yields three questions for any new instrument – whether it is reviewable, whether review arrives in time to matter, and whether unlawfully obtained evidence can actually be excluded. Covert purchases add a fourth, which Strasbourg has so far asked mainly in criminal cases: where legitimate testing of a trader’s conduct ends and incitement begins.

My answer to the opening question is a qualified yes, and it rests on two propositions. First, as regards investigative methods, the Polish system already comes close to criminal procedure, and it has done so chiefly through competition law; UC142 extends that toolkit to consumer enforcement rather than inventing it. Second, the difficulty lies less in any single power than in remedies that depend on the route chosen. A covert purchase requiring identity-concealing documents, made by the Police, would carry an appeal against the court’s consent and, if the appeal succeeds, the evidence-exclusion rule borrowed from competition searches; a purchase under assumed personal details, made by UOKiK’s own staff, would remain subject to the older rule excluding an appeal against consent. Whether safeguards keep pace with powers is therefore a question of remedial design.

The Competition Template: ECN+ and Its Polish Transposition

ECN+ Article 6 requires powers to inspect business premises, copy records, seal business premises and obtain police assistance. Article 7 extends inspections to other premises, including homes, on reasonable suspicion that relevant records are kept there and with prior authorisation by a national judicial authority. Article 6(3) preserves national authorisation requirements for business inspections; Article 3 requires procedural safeguards. Articles 8 and 9 add requests for information and the power to summon representatives for interview, while Article 15 requires a maximum fine of not less than 10% of worldwide turnover. These techniques have precedents in Commission inspections under Regulation 1/2003, Articles 20–21. What ECN+ changed was their status. A national option became a minimum standard binding on every Member State.

Poland was a late transposer. Well after the February 2021 deadline the draft bill had still not reached Parliament, as the 2021 Polish country report on this Blog recorded. The President of the Office of Competition and Consumer Protection (UOKiK) was hardly toothless in the meantime: the same report notes fines of roughly €175,000 imposed on six individual managers for infringing Polish and EU competition law. The Act of 9 March 2023, in force since 20 May 2023, illustrates both sides of the eventual transposition. It expanded the inspection framework, including sealing and continued examination off-site, and regulated police assistance in greater detail. It also introduced express protection for certain lawyer–client communications and strengthened judicial remedies (Competition and Consumer Protection Act, Articles 105b, 105ca, 105da and 105n–105nb). Describing ECN+ solely as an expansion of coercion would therefore miss its second movement. The punitive side of the reform was developed further in UOKiK’s new fining guidelines, discussed on this Blog as one of the more tangible consequences of the Directive.

For competition practitioners, the Polish toolkit is already closer to criminal procedure than its administrative label suggests. A search requires authorisation by the Court of Competition and Consumer Protection (SOKiK), may extend to residential premises (Articles 105n and 91) and is carried out with means – copying of records and data, sealing, police assistance – that any criminal investigator would recognise. None of this is unlawful in itself: ECN+ requires most of it, and Strasbourg accepts such powers where they are effectively controlled. Its significance lies elsewhere. These competition-law devices, and the remedies built around them, are becoming the model for consumer enforcement as well.

Consumer Enforcement Borrows the Toolkit

The chronology needs care. The Consumer Protection Cooperation (CPC) Regulation 2017/2394 predates ECN+. The relevant Polish sequence, however, is ECN+ transposition in 2023, followed by UC142’s proposed CPC implementation. Nor does CPC empower every NCA: it addresses designated consumer-enforcement authorities. Because the President of UOKiK combines both functions, the accumulation of powers is particularly visible in Poland, and tools refined in one field are easily read across into the other.

CPC Article 9(3)(c)–(d) requires inspection and test-purchase powers, including cover identities where necessary. Article 10(2) requires proportionality and procedural safeguards. The Regulation’s mandate concerns the cross-border infringements within its scope; applying similar instruments more broadly to domestic consumer enforcement is a national legislative choice. Articles 9(2) and 10(1) allow powers to be distributed across authorities and exercised through other public bodies or courts. In other words, CPC does not dictate Poland’s institutional design, and the choices discussed below are Warsaw’s, not Brussels’.

From Mystery Shopping to Police Purchases

UC142 does not invent Polish mystery shopping. Existing Article 105ia already permits, with SOKiK’s prior consent, steps aimed at purchasing and their covert recording, but not the conclusion of the contract itself. The UC142 draft dated 20 August 2026 would allow the inspecting official to purchase goods or services and to check how the right of withdrawal and other consumer rights are honoured after the purchase (proposed Article 105ia(1), point 2). For distance sales, the inspection may be conducted remotely, by post or electronic means (paragraph 1a), and the official is to act under a name, address and e-mail address other than his or her own, drawn from a database kept by the President of UOKiK (paragraphs 1b–1c). These are consumer-protection powers, not a new undercover cartel-investigation regime.

Two questions should be kept apart here: whether the instrument works, and what it costs in terms of the values that procedural safeguards protect. On the first, the enforcement case is compelling. An advertised cancellation right may be honoured on paper but frustrated after payment, and an authority unable to complete a transaction cannot reliably reproduce that experience. A test subscription followed by an attempted withdrawal could reveal conduct that reviewing contractual terms alone would miss. Effectiveness, however, does not settle the second question. The relevant line runs between revealing a trader’s usual conduct and inducing the infringement that the authority later seeks to punish, and only a reliable record of the transaction allows a court to tell the two apart.

Proposed Article 105ib draws the institutional boundary. Where the purchase requires documents that prevent the persons carrying it out from being identified, the Police make it at the request of the President of UOKiK, and the operation requires SOKiK’s consent (Article 105ib(1)–(2)). The explanatory memorandum explains the choice. The basic variant, with assumed details taken from UOKiK’s database, does not suffice where the whole sales process must be tested, in particular for complex and risky financial services: the contract must then be concluded with a document concealing the inspector’s identity and bearing a PESEL number, and for more complex products with a matching transaction or credit history. The memorandum classifies such purchases as operational activity, which the President of UOKiK does not conduct, and considers that combining administrative and operational functions in one authority would not serve the systemic coherence of the legal order. The distinction prevents a direct institutional merger, while connecting consumer enforcement to police methods.

The competition-law lineage of the police route is visible on its face. Instead of designing remedies of its own, proposed Article 105ib(2) borrows those built for competition searches, discussed in the next section. Competition lawyers should also watch the boundary between the two fields. A test purchase can reveal more than a consumer-law breach: a checkout page may disclose a resale price maintained across a distribution network, or the tying of products by a dominant platform. Whether material gathered through a consumer test purchase may later feed a competition investigation deserves an answer on the face of the statute, because the safeguards attached to the two routes are not the same.

UC142 would also extend searches to proceedings concerning collective consumer interests and specified preliminary inquiries (proposed Article 105n(1)(3)). According to the explanatory memorandum, such searches are to follow the same procedural provisions as searches in antitrust cases; the competition search regime is thus carried over into consumer enforcement. Business concerns are concrete. In its October 2026 consultation response, Confederation Lewiatan questions the necessity of consumer searches and warns that invented identities could trigger legitimate antifraud systems, potentially being mistaken for obstruction. This is a warning about the draft’s operation, not evidence that such penalties have already occurred.

What Does SOKiK’s Approval Actually Guarantee?

SOKiK is central, but ‘judicial control’ covers different mechanisms. Ordinary mystery shopping requires prior consent under existing Article 105ia(4)–(5), which excludes an appeal against that consent. UC142 extends that consent requirement to actual purchases but leaves the exclusion of an appeal untouched. The police route is different. Through Article 105ib(2), it imports the search-authorisation procedure of Articles 105n(4)–(9), 105na and 105nb(1): a decision within 72 hours, reasons, an appeal against the consent and, where the appeal succeeds, a bar on using the evidence obtained (Article 105na(4)). Article 105ib(2) does not, however, refer to Article 105p, which allows the conduct of a search to be challenged before SOKiK. The stronger remedy thus travels with the police, not with the covert purchase as such, and it concerns the authorisation of the purchase rather than the way it was carried out.

Existing protection also matters. Article 59 of the Entrepreneurs’ Law permits objections for specified inspection breaches, with SOKiK review of UOKiK’s ruling under Article 479²⁸ § 1(2a) of the Code of Civil Procedure. Proposed Article 105ie disapplies, for all test purchases, several provisions of the Entrepreneurs’ Law, including the requirement to show the service card and serve the authorisation before inspection activities begin (Article 49(1)–(3), (7), (8) and (10)); the memorandum explains that the authorisation, the service card and the court’s consent are produced only once the purchase is completed, since the trader does not know of the activities until then. Different intrusions can justify different safeguards. In my view, however, the expanded ordinary purchase warrants reconsideration of the older no-appeal rule, and the remaining objection routes should be tested against transactions of which the trader learns only after the event.

In P 19/17, decided on 16 January 2019, Poland’s Constitutional Tribunal held unconstitutional the former prohibition on appealing search authorisations (Article 105n(4), second sentence). It did not decide the constitutionality of Article 105ia(5). Its reasoning nevertheless supports asking whether each expanded instrument has a remedy proportionate to the interference, beyond the initial judicial signature. That is, in substance, the Strasbourg question.

Strasbourg: Effective Review, Not Administrative Labels

The European Court of Human Rights distinguishes sanctions from investigative measures. In A. Menarini Diagnostics v Italy (2011, §§ 38–44, 59, 63–67), a punitive antitrust fine attracted Article 6’s criminal guarantees despite its administrative classification. Such a sanction may be imposed by an administrative authority, provided that its decision is subject to review by a judicial body with full jurisdiction. There was no violation, because the domestic courts actually examined the factual and legal issues, including proportionality, and could vary the penalty – a point this Blog drew out at the time in a comparison with Luxembourg’s approach to judicial review. Applying the Engel criteria, I would expect similar reasoning to reach Polish consumer fines, which can amount to 10% of turnover for practices infringing collective consumer interests. That expectation does not make every consumer investigation a criminal proceeding.

Article 8 also protects business premises and correspondence, as the Court confirmed for competition inspections in Société Colas Est v France (2002). Delta Pekárny v Czech Republic (2014, §§ 83, 87, 91–94) recognised that effective subsequent review may compensate for the absence of prior authorisation, but found the safeguards insufficient, since the courts had not examined the factual grounds, necessity, duration and scope of the inspection. The judgment’s practical aftermath for the Czech authority’s dawn raids was discussed on this Blog. In UAB Kesko Senukai Lithuania v Lithuania (2023, §§ 114–127), prior authorisation did not cure the refusal to review the inspection’s execution, and subsequent discontinuance removed the opportunity to raise those complaints in proceedings concerning a final infringement decision. The Court nevertheless expressly declined to require ex post judicial review in every business-search case (§ 117). As a post on this Blog observed, Kesko appears to set the bar higher than the Court of Justice did weeks earlier in Les Mousquetaires and Casino. The moment at which review becomes available matters as much as the authorisation itself.

Vinci Construction and GTM Génie Civil et Services v France (2015, §§ 78–79) requires concrete judicial examination of reasoned objections to specifically identified documents allegedly privileged or unrelated to the investigation, with return or deletion where appropriate – a framework that national courts have since applied to bulk copying, as in the Irish ruling discussed in To Seize but Not Search. The counterpoint is BRD–Groupe Société Générale v Romania (2025, §§ 104, 107–117): in the competition-inspection limb, the Court found no Article 8 violation despite the absence of prior judicial authorisation, given limited seizure and effective subsequent scrutiny. Safeguards must work in the particular procedure; Strasbourg does not impose an identical warrant model for every investigation.

Covert purchases engage a different strand of the case law. In Teixeira de Castro v Portugal (1998) and, authoritatively, in Ramanauskas v Lithuania ([GC], 2008, § 55), the Court separated investigation conducted in an essentially passive manner from incitement, in which officials exert such influence that they provoke an offence that would otherwise not have been committed. Evidence obtained through incitement renders the proceedings unfair under Article 6. That case law concerns criminal charges and does not automatically govern consumer testing. In my view, its logic nevertheless corresponds to the line drawn above between revealing a trader’s usual conduct and manufacturing the infringement. If consumer fines fall within Article 6’s criminal limb, a test purchase that steers the trader towards a breach could expose the resulting decision to the same objection, and the safeguard Strasbourg expects – a record allowing a court to examine how the transaction actually unfolded – is precisely what UC142 should guarantee.

The inspection cases themselves concern searches and seizures; they do not establish that every ordinary test purchase engages Article 8. What they offer is a method: assess the actual interference and the practical remedy.

The CJEU’s Grand Chamber judgment of 16 July 2026 in IMI, Joined Cases C-258/23–C-260/23 applies the same method. It accepted the seizure at business premises, without prior court authorisation, of business emails relating to the subject matter of the inspection, provided the law lays down a strict framework and full ex post judicial review guards against abuse (paras 108–111). It added a specific reservation for devices belonging to managers or employees rather than to the undertaking: because such devices may also serve private purposes, access to their data must first be reviewed by a court or an independent administrative body, where appropriate after the devices have been sealed (paras 112–118, especially para 114). Luxembourg and Strasbourg thus proceed from the same premise: the intensity of the interference, not the label of the procedure, determines how much control is due and when.

The Strongest Objection: Lighter Intrusion, Lighter Control

The strongest objection to this analysis can be put as follows. A test purchase is far less intrusive than a dawn raid. The official does what any consumer could do: no premises are entered, no documents are seized, and the trader’s own commercial offer is simply taken up. On this view, a further appeal against SOKiK’s consent would add delay without adding much protection, while the Court of Justice in Les Mousquetaires and Casino accepted that a combination of remedies – including review of the final decision – can together satisfy the right to an effective remedy.

I accept the premise for a purchase that merely takes up the trader’s offer, as any consumer could. It is the reason why different intrusions may carry different safeguards, and why prior appeal is not the only acceptable model. The objection weakens, however, once the official, under an assumed identity, carries the transaction into the post-contractual phase to test withdrawal and other rights. The official’s own conduct then shapes the very interaction that becomes the evidence of the infringement, which is what brings the Ramanauskas line into view; and the trader learns of the measure only afterwards. Nor does the delay argument hold: review timed to the disclosure of the covert act does not slow the operation itself. The combination-of-remedies argument is also weaker than it appears. A covert purchase that reveals nothing will usually lead to no infringement decision, and reliance on review of a final decision becomes problematic where the investigation is discontinued. The relevant question is then whether another effective remedy remains available. Kesko illustrates that risk in a business-search case; its application to covert purchases requires a separate assessment, which in my view the statute itself should make unnecessary.

Powers and Remedies Must Develop Together

The answer to the opening question is therefore a qualified yes. EU legislation consolidates and extends methods familiar from criminal enforcement, and competition law is where that extension began. Poland’s proposal would assign certain covert consumer test purchases to the Police, while the President of UOKiK retains a specialised administrative mandate. The concern lies in the cumulative reach of distinct instruments and the uneven remedies accompanying them. For competition lawyers, the lesson is that the ECN+ toolkit has become the backbone of UOKiK’s enforcement as a whole, and every extension into consumer law imports both its strengths and its remedial gaps. The revision of Regulation 1/2003 is still at the preparatory stage: after the 2025 consultation, the Commission plans to adopt its proposal in the second quarter of 2027. Before that text is settled, the Polish case shows the Union legislature what happens when powers travel faster than remedies.

The argument has limits that should be stated. It does not show that UC142 is incompatible with the Convention: the bill is a consultation draft, the Strasbourg inspection case law does not transfer automatically to test purchases, and much will depend on how SOKiK applies the remedies in practice. Nor does it suggest that UOKiK is turning into a police body; the bill itself keeps operational activity with the Police. Its claim is narrower. Where the legislator adds a covert power, it should add a remedy that works for that power.

As a matter of legislative design, UC142 should therefore adapt existing remedies to its new instruments: clarify their scope, align deadlines with disclosure of covert acts, and ensure a record sufficient for judicial scrutiny. Authorisations should explain why the particular test and its concealment are necessary. Legitimate fraud prevention should not become obstruction merely because it detects an official’s assumed identity. The framework should secure practical review of challenged conduct even without a final infringement decision, and attach effective consequences to evidence obtained through incitement. The first postulate responds to the risk illustrated in Kesko; the second follows the logic of Ramanauskas. These guarantees make investigative powers legally sustainable and their limits predictable.

The author writes in a personal capacity. Views expressed do not represent those of any institution.

For further reading on this Blog: Strasbourg’s Long Shadow: The European Court of Human Rights and the Standard of Antitrust Inspections · Dawn Raids and Human Rights: Where Are We Now? · Two Courts – Two Views on the Procedural Safeguards to Surround Dawn Raids · Dawn Raids: The ‘New Normal’ Since the Pandemic · The CPC Network – Consumer Protection, EU Style