Response to consultation on methodology for setting fines under MiCA
Q1. Do you have any comments on the proposed classification of infringements based on their severity, as well as on their distribution across the severity scale? Please justify your answers.
I do not object to a three-tier classification, and the ordering — organisational and operational requirements and disclosure above conflicts of interest, and both above obstacles to supervisory activities — is defensible and consistent with the sectoral precedents the paper cites.
One point in the methodology’s favour, since consultation responses rarely record these. I checked the allocation at section 4, paragraph 2: Annex V points 1 to 87 and Annex VI points 1 to 46 are each allocated to exactly one band, with no point omitted and no point appearing twice. The table is complete.
On the inclusion of infringements outside the Annexes, the EBA’s footnote 10 is well founded as to the power, and I say so because the point has been queried elsewhere. Article 122(2)(f) and (3)(f) each require the EBA to “indicate the fine provided for in Article 131, where the answers to questions asked are incorrect or misleading”, and Article 123(2) refers in terms to “the fines provided for in Article 131, where the answers to questions posed to issuers of significant asset-referenced tokens or issuers of significant e-money tokens are incorrect or misleading”. Those provisions engage Article 131 independently of Annexes V and VI, as the footnote says.
What does not follow is that the maxima apply. Article 131(3) caps “the maximum amount of the fine referred to in paragraph 1”, and Article 131(1) is confined to an infringement “as listed in Annex V” or “as listed in Annex VI”, found “in accordance with Article 134(8)” — which is itself a decision on “whether an infringement as listed in Annex V or VI has been committed”. A fine for incorrect or misleading answers is a fine provided for in Article 131 but is not a paragraph 1 fine. Whether the 12.5% and 10% ceilings govern it is therefore not settled by the words. The methodology assumes that they do — it places those infringements in the 2% band and then applies the maxima at section 4, paragraphs 9 and 10 — and that is the protective answer and, I would expect, the right one. It should be reasoned rather than assumed.
Recommendation 1. State in the methodology the basis on which the Article 131(3) and (4) maxima are applied to fines imposed for infringements of Articles 122 and 123. An issuer facing such a fine should not have to construe paragraph 1 and Article 134(8) to discover what its exposure is capped at.
Q2. Do you have comments on the level of the proposed percentages for each category of infringements? Please justify your answers.
My comment is not on the level of 5%, 3% and 2% but on the derivation offered for them. Section 3, paragraph 15 states that the percentages “have been inspired by the percentages of annual turnover identified in Article 111 of MiCA regarding administrative penalties imposed by national competent authorities”, and that they “already provide an indication of the severity of the different types of infringements on behalf of the co-legislators”.
The percentages in Article 111 are maxima. They are the ceilings Member States must enable competent authorities to impose, expressed as proportions of total annual turnover, and they run from 2% to 15% across the infringement families. The percentages in this methodology are basic amounts, to which multiplying coefficients are then applied at Step 2. A figure fixed by the co-legislators as a ceiling is being used as a point of departure, which is a different function; and the effect is visible in the arithmetic, since on the coefficients proposed a 5% basic amount reaches 26.5% of turnover with all six aggravating factors on the corrected reading of Step 2, and 51.5% on the drafted reading.
I do not say the basic amounts are wrong. I say the stated derivation does not support them, and that a reader cannot tell which Article 111 percentage is said to inform which band. Recommendation 2. Publish the correspondence: which Article 111 percentage informs which category, and on what view of the relationship between a statutory ceiling for competent authorities and a starting point for the EBA.
A related observation: Footnote 1 confirms that this methodology does not apply to competent authorities, where Article 111 applies instead. The consequence is not a divergence in ceilings — 12.5% appears both in Article 111 and in Article 131(3) — but a divergence in transparency: the issuer of a significant token will be fined under a published methodology, while the issuer of a materially identical non-significant token faces national practice that is in most Member States unpublished. If the EBA’s object is “a consistent and transparent approach”, the methodology would be worth publishing in a form competent authorities can adopt, or taking forward as convergence work, rather than as an EBA-only instrument.
Q3. Do you agree with the proposed aggravating and mitigating factors? Please justify your answers.
The factors themselves are recognisable from established enforcement practice and I do not object to any of them.
The mapping to Article 131(2), however, is not published, and it does not fully work. Footnote 11 states that the criteria at Article 131(2), points (a), (b), (c), (d), (e), (i), (j) and (k) can be quantified ex ante by coefficients. Reading the tables against the Article: duration and frequency, point (a), is carried by the six-month and repetition coefficients; financial crime, point (b), by the coefficient of 2; serious or systemic weaknesses, point (c), by 2.2; intention, point (d), by 2; previous infringements, point (j), by 1.5; and measures taken after the infringement to prevent repetition, point (k), by 0.6. The level of cooperation, point (i), maps at best partially onto the 0.4 coefficient for bringing the infringement to the EBA’s attention. Point (e), the degree of responsibility of the issuer responsible for the infringement, has no coefficient at all. The 0.7 coefficient, for senior management demonstrating reasonable measures to prevent the breach, is a different matter and is not described as giving effect to point (e).
Recommendation 3. Publish a table mapping each of the eleven criteria in Article 131(2) to the coefficient, or to the paragraph 8 stage, that gives it effect, and say expressly where a criterion is carried by neither. This is a page of drafting and it would forestall the first appeal ground any issuer’s counsel would settle on.
The disgorgement rider, which is the substantive point under this question. Article 131(2)(i) requires the EBA to take into account “the level of cooperation of the issuer responsible for the infringement with EBA, without prejudice to the need to ensure disgorgement of profits gained or losses avoided by that person”. The Regulation therefore provides in terms that credit for cooperation must not be allowed to defeat disgorgement.
The methodology contains no floor. On the corrected reading of the combination rule set out under Question 4, the three mitigating coefficients applied together reduce a basic amount of 5% of turnover to minus 1.5%; applied two at a time they give 0.5%, 1.5% or nil. An issuer whose infringement produced profits of, say, 8% of turnover could therefore cooperate its way to a fine of nothing. Section 4, paragraph 8 does state an aim “to at least disgorge profits made or losses avoided, where these can be determined” — but it states it as part of a discretion the EBA “may” exercise. An aim inside a discretion is not the safeguard that Article 131(2)(i) contemplates, and it is the one place in this methodology where the drafting is in tension with the Level 1 text rather than merely unclear.
Recommendation 4. Provide expressly that, where the profits gained or losses avoided can be determined, the amount resulting from Step 2 shall not be reduced below that figure, subject always to the maxima in Article 131(3) and (4). That gives Article 131(2)(i) effect, costs the EBA nothing in cases where mitigation is genuine and modest, and removes the only outcome in the methodology that would be indefensible on review.
Q4. Do you agree with the proposed coefficients? Please justify your answers.
Section 4, paragraph 3 provides that the EBA “applies the adjustment coefficients set out in the tables below in a cumulative way to the basic amount resulting from the application of Step 1”. Paragraph 4 then provides that where more than one aggravating coefficient applies, “the difference between the basic amount and the sum of the amounts resulting from the application of all applicable aggravating coefficients is added to the basic amount”, and paragraph 5 that where more than one mitigating coefficient applies, that difference “is subtracted from the basic amount”.
Read as the words stand, the basic amount is subtracted once in total from the sum of the coefficient-adjusted amounts. What the methodology evidently intends is that it be subtracted once for each coefficient, so that what accumulates is each coefficient’s increment rather than its whole value. The difference is not presentational. On the mitigating side it reverses the direction of the adjustment.
Taking the basic amount as 100 and the mitigating coefficients in the paper — 0.7 for senior management’s preventive measures, 0.4 for prompt disclosure, 0.6 for voluntary remediation. On the drafting as it stands, an issuer that establishes two mitigating factors is fined more than one that establishes none; an issuer that establishes prompt disclosure and voluntary remediation is in exactly the position of an issuer that did neither; and an issuer that establishes all three is fined 70 per cent more than an issuer that establishes nothing. That is plainly not intended. It is also the kind of defect that stays invisible until the first contested decision, at which point it is the EBA that has to defend it.
Recommendation 5. Redraft paragraphs 4 and 5 so that the adjustment is the sum of the coefficients’ increments rather than the difference between the basic amount and the sum of the adjusted amounts, and state it arithmetically rather than in prose: final amount = basic amount × [1 + Σ (c − 1)], taken over all applicable coefficients. Prose of this kind does not survive translation into practice, and the present sentence is the proof. The corrected formula then needs an express floor, because on it the three mitigating coefficients sum to an increment of −1.3 and produce a negative figure; Recommendation 4 supplies the floor where profits can be determined, and a residual floor of zero is needed where they cannot.
The defect is not confined to mitigation. On the drafted reading, and with a 5% basic amount, every one of the fifteen possible pairs of aggravating factors exceeds the 12.5% maximum, the lowest pair giving 13.0% and the highest 21.0%. On the corrected reading five of the fifteen do. So the drafting error also determines how often the legal maximum binds, which is the subject of Question 6.
Q5. Do you have any alternative ideas for ensuring flexibility in determining the final amount of the fine by the EBA? Please justify your answers.
I note that Question 5 is put in two forms: the box at page 13 asks for “alternative proposals for ensuring flexibility in determining the final amount of the fine by the EBA BoS”, and the overview at page 14 asks for “alternative ideas … by the EBA”. I answer the wider of the two.
The adjustment power at section 4, paragraph 8 is necessary and I do not propose that it be narrowed. A methodology of bands and coefficients cannot anticipate every case, and the power is where Article 131(2)(f), (g) and (h) are given effect. My proposal is that its exercise be made visible.
A methodology of tiers and coefficients presents itself as determinate. If the number actually imposed emerges from an unstructured discretion applied at the end, the appearance of determinacy misleads the issuer and does not assist a reviewing body, which cannot tell whether it is reviewing arithmetic or judgement. This is not an abstract concern here. Article 134 entitles the person concerned to be heard and to have access to the file, and requires that the investigation officer “base its findings only on facts on which the persons concerned have had the opportunity to comment”; Commission Delegated Regulation (EU) 2024/1504 carries the procedural rules and the rights of defence. A right to comment on the facts is of limited value if the calculation built on those facts is not disclosed.
Recommendation 6. Provide that the decision imposing a fine shall state, as separate figures: the basic amount and the band it is taken from; each coefficient applied and the reason for it; the amount resulting from Step 2; the amount after any Article 131(3) or (4) maximum has been applied; and, where the paragraph 8 power is exercised, the amount before and after its exercise together with the reasons. The test is that the addressee should be able to reconstruct the fine from the decision. This costs the EBA nothing where its reasoning is sound, and it is what makes the methodology reviewable rather than merely published.
Q6. Do you have any comments on how the maximum amounts in Article 131(3) and (4) of MiCA are applied under the methodology?
Section 4, paragraph 9 provides that the final amount “is reduced to 12.5% of the annual turnover in the preceding business year of an issuer of an s-ART, or 10% for an issuer of an s-EMT, if it’s higher than these amounts and the EBA is not able to determine the profits gained or losses avoided”. Paragraph 10 provides that where the EBA is able to determine that profits gained or losses avoided are above 12.5% or 10% of turnover, the final amount “is reduced where necessary to twice the amount of profits gained or losses avoided”.
Those two limbs do not meet. Where the profits gained or losses avoided can be determined and are at or below 12.5% or 10% of turnover, paragraph 9 does not apply, because its condition requires that the EBA cannot determine them; and paragraph 10 does not apply, because its condition requires that they exceed the percentage. On the face of the methodology, no maximum applies to that case at all. Since Step 2 can produce 21% of turnover on two aggravating factors, this is not a theoretical gap. It cannot be what is intended, and it is inconsistent with Article 131(3) and (4), which cap the fine without regard to whether profits happen to be determinable.
Recommendation 7. Redraft paragraphs 9 and 10 as a single ceiling rule: the final amount shall not exceed 12.5% of annual turnover for an issuer of an s-ART, or 10% for an issuer of an s-EMT, or, where the profits gained or losses avoided can be determined and twice that amount is the greater, twice that amount. That reproduces Article 131(3) and (4) and closes the gap without changing any figure.
A second point under this question, following from Question 4. Because on the drafted combination rule every pair of aggravating factors already exceeds 12.5%, and on the corrected rule nineteen of the twenty possible triples do, the maximum will bind in most aggravated cases — which are precisely the cases the methodology exists to grade. Two materially different infringements will then attract the same fine, and the addressee cannot tell from the decision whether the assessment was severe or merely truncated. I do not propose that the basic amounts be lowered, which would be a policy judgement for the EBA. I propose, as the lighter alternative, that where a maximum binds, the decision state the uncapped figure produced by the methodology alongside the capped amount imposed. The deterrent signal is preserved, no percentage changes, and the gradation the EBA has designed remains visible.
Note: The figures above use the coefficient values in the tables at section 4, paragraphs 6 and 7, and the basic amounts at paragraph 2. The structural point under Question 4 does not depend on those values: as drafted, the mitigating rule increases the fine whenever the applicable mitigating coefficients sum to more than one, whatever the values are. The worked figures are given so that the drafting point can be tested directly. Two slips, noted only for tidiness: the Step 1 heading in section 4 reads “the basic amount of amount of the fine”, and Question 5 is worded differently at pages 13 and 14 as set out at paragraph 26.