It’s Not a Trap

Despite most countries having trouble getting rid of bribery in daily life, only few so far have dared integrity testing: sending out undercover testers disguised as ordinary citizens to contact the public administration and check which public employees ask for bribes. The main argument against such undercover tests has been that they constitute “entrapment”. However, in Cavca, the ECtHR finally dispels the myth that these tests in and of themselves equal entrapment. Anticorruption reformers have long awaited such a decision, not least since the Venice Commission had fuelled the entrapment argument for more than 10 years. Yet, the decision leaves one key question unaddressed: Just when does integrity testing become entrapment? This article shows that testers can have a more active role in disciplinary tests than in criminal sting operations. However, excessive bribe offers or emotional blackmail should be off-limits.

A tool too effective for corrupt governments

A culture of bribery in public administration is hard to break: Both sides involved usually stay silent about the bribe. The public official profits from it, as does the citizen who secures a public service otherwise not available. In a perverse way, both sides can trust each other. Professional integrity testing destroys this trust. Moldova is an impressive example. The moment integrity testing started (in 2014), the number of public officials reporting bribe offers jumped from almost zero to several dozen. Suddenly, any citizen coming to an office could be an undercover officer. Citizens, too, became alert as public officials stopped trusting them and reported bribe offers to law enforcement authorities. For the first time in Moldova, even judges reported bribe offers (and one can only imagine what they did until integrity testing started). “Who are the testers? Who will be tested next?” – social media and talk shows were humming with discussions. The deputy head of the anticorruption agency spearheading this reform coined the term of a “healthy hysteria”.

The success of this reform scared the then ruling elite, who profited from endemic corruption, but had initially supported integrity testing believing it would remain what other reforms were: a fig leaf, meant to appease the public and international observers. Only a few months later, using the argument of alleged entrapment, it was easy to set back the reform through a constitutional court captured by the ruling elite. The number of reported bribe offers went back to almost zero.

Integrity testing is not “entrapment”

About 10 years later, the ECtHR finally dethrones the entrapment argument in Cavca v. Moldova:

“The nature of such testing involves the authorities artificially creating situations which are like those that might occur in the context of the professional activity of the persons being tested, to see how they react. Therefore, […] such testing cannot be considered incitement to commit an offence” (§ 55).

One might wonder why a decision by the ECtHR was necessary to clarify this point. Only in the perverted logic of a corrupt public official could a basic request for public service be seen as a provocation (“The citizen applied for a passport – time for my bribe”).

The ECtHR lists the following four preconditions for tests to meet fair (civil) trial standards:

First, public officials failing the tests face only disciplinary liability (§ 55). Thus, only the civil limb of Art. 6 ECHR applies with its lower standard of fair trial compared to the criminal limb. The limitation to disciplinary liability was the ingenious global innovation by the Moldovan anticorruption agency to avoid the entrapment argument while still cleaning up the public sector.

Second, an individual suspicion is not needed (§ 59). Rather, “a risk of corruption […] identified” within a certain sector or agency is already sufficient. In the concrete case, the source of data was “previous reports, a decision by the Court of Auditors, reports by non‑governmental organisations, statistics and specific confirmed cases of corruption amongst […] [the agency’s] employees” (§ 5). A risk is not a fact, but only a probability. Therefore, in other cases than Cavca, one source of data could suffice if a judge confirms the risk as objectively identified (§ 60).

Third, prior “judicial authorisation and subsequent supervision” are required, “reviewing both the need for such testing and the specific manner in which it would be carried out” (§ 58).

Fourth, once disciplined, a public servant has, of course, the right to a court “hearing him and the tester” about alleged entrapment, as the ECtHR underlined (§ 63). The decision does not explain the possible motives of the Moldovan legislator violating the Convention in this regard (§ 64): First, audio and video evidence often speaks for itself. Second, guarding the identity of undercover agents is key for the testing scenarios to continue working in the future. These are highly trained experts and not disposable single-use extras. However, if needed, there are mechanisms for introducing undercover witnesses as evidence in criminal procedures, without divulging the identity of the tester (Van Mechelen, § 54). Civil procedures could probably deal with guarded identities of witnesses more flexibly than criminal procedures (Guide, § 422).

When does a test scenario become “entrapment”?

The test scenario used in the Cavca case shows how creative integrity testers need to be: an undercover agent had posed as a homeowner, illegally cutting down a tree. Confronted by a public official from the (corruption-prone) environmental agency, the tester “suggested […] that the issue could be resolved without there being an official procedure, and offered […] various items as a bribe” (§ 6). The public official initially refused, as he felt observed by other persons present. The tester repeated the offer later when the two were alone, and finally the public official asked the undercover agent to leave the chainsaw used in the tree-cutting in the public official’s private car. Did the repeated bribe offer amount to entrapment? Unfortunately, the ECtHR considered “that in the present case there is no need to decide on whether the applicant was indeed subjected to entrapment” because national courts had assessed the evidence without him (§ 61-63).

However, the Court underlines that in the disciplinary context, lower standards apply than in criminal cases: the “guarantees of a fair trial developed in its case-law under Article 6 § 1 in respect of entrapment in the context of criminal proceedings” serve only for “drawing inspiration, with appropriate adjustments” (§ 46). Lower standards mean: undercover testers are not confined to behaving in an “essentially passive manner” as in criminal investigations (Ramanauskas, § 55) but can play a more active role. It should be noted that even in criminal cases, an active role becomes justified where tested officials show an informal attitude towards the law (Rotaru, § 30):

“[T]he undercover agents offered bribes to the applicant only after the latter had shown that he was willing to ignore their infractions […]. There is therefore no indication that the investigators involved have exceeded the limits of a passive attitude”.

If in disciplinary (civil) cases limits are wider, undercover agents could even display their general willingness to bribe from the start. This may start with questions: What could the citizen do for the official to obtain leniency or speed up things? Could the citizen speak with the public official in private? Public officials must react to a first bribe offer with a clear and firm rebuke and with reporting the tester (Article 16(3) of Law 82/2017). Whilst they do not react fully legally, repeated bribe offers are appropriate, as in the Cavca case. By comparison, excessive bribe offers (in Malininas, § 17, about double the market price) or emotional blackmail on the tested official (cf. Vanyan, § 11, fear of suicide) would be off-limits.

Two Venice Commission opinions overruled

In contrast to the ECtHR’s nuanced approach in Cavca, the Venice Commission (VC) leaned against integrity testing about 10 years ago. In two opinions regarding Moldova (2014) and Ukraine (2015), it applied the stricter criminal limb of Article 6 to disciplinary integrity testing (§ 81), ignoring the earlier ECtHR’s clarification that such disciplinary proceedings fall under the civil limb (Volkov, §§ 93-95; in Cavca: § 36). The VC further diverged from the ECtHR by misinterpreting all integrity tests to be bribe offers and thus entrapment (§ 84), ignoring key case law (such as Rotaru). This stands out even more given that the Moldovan law defined a “professional integrity test” only as “simulated situations, similar to those in the work activity, […] to passively monitor and establish the reaction and conduct of the tested public agent” (Article 4 of Law 325/2013).

Furthermore, the VC’s draft opinion sympathised with “notoriously inferior – unaided – judge[s]” (§ 43) as the testing subjects and wandered off into concerns about judicial independence, although the Moldovan Law had excluded the judicial sector from its scope (Article 4 of Law 325/2013). In doing so, the VC walked into the trap set by the then untrustworthy Moldovan Constitutional Court, which misrepresented the integrity testing law as applying to the judicial sector.

Finally, the VC describes the Moldovan Law on integrity testing as “a highly disputable instrument, the suitability of which for the fight of wide-spread corruption phenomena in the public sector is questionable” (§ 93), while contradicting an earlier opinion sent by the Council of Europe’s Secretary General to the Moldovan government, supporting the (then draft) Law and recommending integrity testing “as a very effective tool against corruption” (ECCU-BO-MD-2/2012). Equally, the ECtHR underlines integrity testing’s “clear benefits for society” because “the possibility that any offer of a bribe could potentially be part of an integrity test clearly has a preventive effect” (Cavca, § 41).

Testing for administrative purposes

Cavca relates only to undercover tests directed at individual persons for their disciplinary liability. It should not be misread as applying to test purchases targeted at businesses for monitoring general compliance. For example, European Union law empowers authorities to conduct “test purchases […] under a cover identity” as a means for monitoring compliance with consumer protection rules (e.g. Art. 9 III d EU 2017/2394). This reflects undercover test purchases already conducted in EU countries to ensure that businesses comply with administrative or tax rules (for Germany see VGH Bavaria 22 CS 02.2687; VGH BW 6 S 1110/07). German courts do not even mention the need for prior judicial approval because they see a fundamental difference to undercover testing primarily aimed at sanctions. Furthermore, test purchases are about the compliance of businesses, not about behaviour of individuals (Hoppe, NVwZ 2022, 303). Consequently, the German legislator foresees administrative undercover testing without prior judicial involvement (§ 6(5) EU-VSchDG; BT-Drs. 19/16781, S. 30).

Looking ahead

The ECtHR has finally underpinned a wider use of integrity tests with legal certainty. Countries with high levels of daily corruption can now introduce this tool. However, clearing the legal hurdle is not even half the battle. As the explanatory notes on a model law for integrity testing state, “the testing unit has to be of highest integrity and confidentiality, as information of the targets and timings would be ‘commodities’ for which corrupt officials could be willing to pay a good price.” Notably, in countries with a high level of corruption it is often hard to find a suitable state body. Furthermore, scenarios must be realistic to work – this requires resources and training. Still, as the ECtHR pointed out, international organisations recommend these tests, and for good reasons so.

The author would like to thank Anton Marchuk, Sviatoslav Tkachuk, and David Wellstein for their comments on a draft of this blog post.